diff --git a/src/pugixml.cpp b/src/pugixml.cpp index fc0fd87..df978d1 100644 --- a/src/pugixml.cpp +++ b/src/pugixml.cpp @@ -16,105 +16,109 @@ #include "pugixml.hpp" -#include -#include -#include #include #include +#include +#include +#include #ifdef PUGIXML_WCHAR_MODE -# include +#include #endif #ifndef PUGIXML_NO_XPATH -# include -# include +#include +#include #endif #ifndef PUGIXML_NO_STL -# include -# include -# include +#include +#include +#include #endif // For placement new #include #ifdef _MSC_VER -# pragma warning(push) -# pragma warning(disable: 4127) // conditional expression is constant -# pragma warning(disable: 4324) // structure was padded due to __declspec(align()) -# pragma warning(disable: 4702) // unreachable code -# pragma warning(disable: 4996) // this function or variable may be unsafe +#pragma warning(push) +#pragma warning(disable : 4127) // conditional expression is constant +#pragma warning(disable : 4324) // structure was padded due to __declspec(align()) +#pragma warning(disable : 4702) // unreachable code +#pragma warning(disable : 4996) // this function or variable may be unsafe #endif #if defined(_MSC_VER) && defined(__c2__) -# pragma clang diagnostic push -# pragma clang diagnostic ignored "-Wdeprecated" // this function or variable may be unsafe +#pragma clang diagnostic push +#pragma clang diagnostic ignored "-Wdeprecated" // this function or variable may be unsafe #endif #ifdef __INTEL_COMPILER -# pragma warning(disable: 177) // function was declared but never referenced -# pragma warning(disable: 279) // controlling expression is constant -# pragma warning(disable: 1478 1786) // function was declared "deprecated" -# pragma warning(disable: 1684) // conversion from pointer to same-sized integral type +#pragma warning(disable : 177) // function was declared but never referenced +#pragma warning(disable : 279) // controlling expression is constant +#pragma warning(disable : 1478 1786) // function was declared "deprecated" +#pragma warning(disable : 1684) // conversion from pointer to same-sized integral type #endif #if defined(__BORLANDC__) && defined(PUGIXML_HEADER_ONLY) -# pragma warn -8080 // symbol is declared but never used; disabling this inside push/pop bracket does not make the warning go away +#pragma warn - 8080 // symbol is declared but never used; disabling this inside push/pop bracket does not make the warning go away #endif #ifdef __BORLANDC__ -# pragma option push -# pragma warn -8008 // condition is always false -# pragma warn -8066 // unreachable code +#pragma option push +#pragma warn - 8008 // condition is always false +#pragma warn - 8066 // unreachable code #endif #ifdef __SNC__ // Using diag_push/diag_pop does not disable the warnings inside templates due to a compiler bug -# pragma diag_suppress=178 // function was declared but never referenced -# pragma diag_suppress=237 // controlling expression is constant +#pragma diag_suppress = 178 // function was declared but never referenced +#pragma diag_suppress = 237 // controlling expression is constant #endif #ifdef __TI_COMPILER_VERSION__ -# pragma diag_suppress 179 // function was declared but never referenced +#pragma diag_suppress 179 // function was declared but never referenced #endif // Inlining controls #if defined(_MSC_VER) && _MSC_VER >= 1300 -# define PUGI__NO_INLINE __declspec(noinline) +#define PUGI__NO_INLINE __declspec(noinline) #elif defined(__GNUC__) -# define PUGI__NO_INLINE __attribute__((noinline)) +#define PUGI__NO_INLINE __attribute__((noinline)) #else -# define PUGI__NO_INLINE +#define PUGI__NO_INLINE #endif // Branch weight controls #if defined(__GNUC__) && !defined(__c2__) -# define PUGI__UNLIKELY(cond) __builtin_expect(cond, 0) +#define PUGI__UNLIKELY(cond) __builtin_expect(cond, 0) #else -# define PUGI__UNLIKELY(cond) (cond) +#define PUGI__UNLIKELY(cond) (cond) #endif // Simple static assertion -#define PUGI__STATIC_ASSERT(cond) { static const char condition_failed[(cond) ? 1 : -1] = {0}; (void)condition_failed[0]; } +#define PUGI__STATIC_ASSERT(cond) \ + { \ + static const char condition_failed[(cond) ? 1 : -1] = {0}; \ + (void)condition_failed[0]; \ + } // Digital Mars C++ bug workaround for passing char loaded from memory via stack #ifdef __DMC__ -# define PUGI__DMC_VOLATILE volatile +#define PUGI__DMC_VOLATILE volatile #else -# define PUGI__DMC_VOLATILE +#define PUGI__DMC_VOLATILE #endif // Integer sanitizer workaround; we only apply this for clang since gcc8 has no_sanitize but not unsigned-integer-overflow and produces "attribute directive ignored" warnings #if defined(__clang__) && defined(__has_attribute) -# if __has_attribute(no_sanitize) -# define PUGI__UNSIGNED_OVERFLOW __attribute__((no_sanitize("unsigned-integer-overflow"))) -# else -# define PUGI__UNSIGNED_OVERFLOW -# endif +#if __has_attribute(no_sanitize) +#define PUGI__UNSIGNED_OVERFLOW __attribute__((no_sanitize("unsigned-integer-overflow"))) #else -# define PUGI__UNSIGNED_OVERFLOW +#define PUGI__UNSIGNED_OVERFLOW +#endif +#else +#define PUGI__UNSIGNED_OVERFLOW #endif // Borland C++ bug workaround for not defining ::memcpy depending on header include order (can't always use std::memcpy because some compilers don't have it at all) @@ -126,4017 +130,4293 @@ using std::memset; // Some MinGW/GCC versions have headers that erroneously omit LLONG_MIN/LLONG_MAX/ULLONG_MAX definitions from limits.h in some configurations #if defined(PUGIXML_HAS_LONG_LONG) && defined(__GNUC__) && !defined(LLONG_MAX) && !defined(LLONG_MIN) && !defined(ULLONG_MAX) -# define LLONG_MIN (-LLONG_MAX - 1LL) -# define LLONG_MAX __LONG_LONG_MAX__ -# define ULLONG_MAX (LLONG_MAX * 2ULL + 1ULL) +#define LLONG_MIN (-LLONG_MAX - 1LL) +#define LLONG_MAX __LONG_LONG_MAX__ +#define ULLONG_MAX (LLONG_MAX * 2ULL + 1ULL) #endif // In some environments MSVC is a compiler but the CRT lacks certain MSVC-specific features #if defined(_MSC_VER) && !defined(__S3E__) -# define PUGI__MSVC_CRT_VERSION _MSC_VER +#define PUGI__MSVC_CRT_VERSION _MSC_VER #endif // Not all platforms have snprintf; we define a wrapper that uses snprintf if possible. This only works with buffers with a known size. #if __cplusplus >= 201103 -# define PUGI__SNPRINTF(buf, ...) snprintf(buf, sizeof(buf), __VA_ARGS__) +#define PUGI__SNPRINTF(buf, ...) snprintf(buf, sizeof(buf), __VA_ARGS__) #elif defined(PUGI__MSVC_CRT_VERSION) && PUGI__MSVC_CRT_VERSION >= 1400 -# define PUGI__SNPRINTF(buf, ...) _snprintf_s(buf, _countof(buf), _TRUNCATE, __VA_ARGS__) +#define PUGI__SNPRINTF(buf, ...) _snprintf_s(buf, _countof(buf), _TRUNCATE, __VA_ARGS__) #else -# define PUGI__SNPRINTF sprintf +#define PUGI__SNPRINTF sprintf #endif // We put implementation details into an anonymous namespace in source mode, but have to keep it in non-anonymous namespace in header-only mode to prevent binary bloat. #ifdef PUGIXML_HEADER_ONLY -# define PUGI__NS_BEGIN namespace pugi { namespace impl { -# define PUGI__NS_END } } -# define PUGI__FN inline -# define PUGI__FN_NO_INLINE inline +#define PUGI__NS_BEGIN \ + namespace pugi \ + { \ + namespace impl \ + { +#define PUGI__NS_END \ + } \ + } +#define PUGI__FN inline +#define PUGI__FN_NO_INLINE inline #else -# if defined(_MSC_VER) && _MSC_VER < 1300 // MSVC6 seems to have an amusing bug with anonymous namespaces inside namespaces -# define PUGI__NS_BEGIN namespace pugi { namespace impl { -# define PUGI__NS_END } } -# else -# define PUGI__NS_BEGIN namespace pugi { namespace impl { namespace { -# define PUGI__NS_END } } } -# endif -# define PUGI__FN -# define PUGI__FN_NO_INLINE PUGI__NO_INLINE +#if defined(_MSC_VER) && _MSC_VER < 1300 // MSVC6 seems to have an amusing bug with anonymous namespaces inside namespaces +#define PUGI__NS_BEGIN \ + namespace pugi \ + { \ + namespace impl \ + { +#define PUGI__NS_END \ + } \ + } +#else +#define PUGI__NS_BEGIN \ + namespace pugi \ + { \ + namespace impl \ + { \ + namespace \ + { +#define PUGI__NS_END \ + } \ + } \ + } +#endif +#define PUGI__FN +#define PUGI__FN_NO_INLINE PUGI__NO_INLINE #endif // uintptr_t #if (defined(_MSC_VER) && _MSC_VER < 1600) || (defined(__BORLANDC__) && __BORLANDC__ < 0x561) namespace pugi { -# ifndef _UINTPTR_T_DEFINED - typedef size_t uintptr_t; -# endif +#ifndef _UINTPTR_T_DEFINED +typedef size_t uintptr_t; +#endif - typedef unsigned __int8 uint8_t; - typedef unsigned __int16 uint16_t; - typedef unsigned __int32 uint32_t; -} +typedef unsigned __int8 uint8_t; +typedef unsigned __int16 uint16_t; +typedef unsigned __int32 uint32_t; +} // namespace pugi #else -# include +#include #endif // Memory allocation PUGI__NS_BEGIN - PUGI__FN void* default_allocate(size_t size) - { - return malloc(size); - } +PUGI__FN void* default_allocate(size_t size) +{ + return malloc(size); +} - PUGI__FN void default_deallocate(void* ptr) - { - free(ptr); - } +PUGI__FN void default_deallocate(void* ptr) +{ + free(ptr); +} - template - struct xml_memory_management_function_storage - { - static allocation_function allocate; - static deallocation_function deallocate; - }; +template +struct xml_memory_management_function_storage +{ + static allocation_function allocate; + static deallocation_function deallocate; +}; - // Global allocation functions are stored in class statics so that in header mode linker deduplicates them - // Without a template<> we'll get multiple definitions of the same static - template allocation_function xml_memory_management_function_storage::allocate = default_allocate; - template deallocation_function xml_memory_management_function_storage::deallocate = default_deallocate; +// Global allocation functions are stored in class statics so that in header mode linker deduplicates them +// Without a template<> we'll get multiple definitions of the same static +template +allocation_function xml_memory_management_function_storage::allocate = default_allocate; +template +deallocation_function xml_memory_management_function_storage::deallocate = default_deallocate; - typedef xml_memory_management_function_storage xml_memory; +typedef xml_memory_management_function_storage xml_memory; PUGI__NS_END // String utilities PUGI__NS_BEGIN - // Get string length - PUGI__FN size_t strlength(const char_t* s) - { - assert(s); +// Get string length +PUGI__FN size_t strlength(const char_t* s) +{ + assert(s); - #ifdef PUGIXML_WCHAR_MODE - return wcslen(s); - #else - return strlen(s); - #endif - } +#ifdef PUGIXML_WCHAR_MODE + return wcslen(s); +#else + return strlen(s); +#endif +} - // Compare two strings - PUGI__FN bool strequal(const char_t* src, const char_t* dst) - { - assert(src && dst); +// Compare two strings +PUGI__FN bool strequal(const char_t* src, const char_t* dst) +{ + assert(src && dst); - #ifdef PUGIXML_WCHAR_MODE - return wcscmp(src, dst) == 0; - #else - return strcmp(src, dst) == 0; - #endif - } +#ifdef PUGIXML_WCHAR_MODE + return wcscmp(src, dst) == 0; +#else + return strcmp(src, dst) == 0; +#endif +} - // Compare lhs with [rhs_begin, rhs_end) - PUGI__FN bool strequalrange(const char_t* lhs, const char_t* rhs, size_t count) - { - for (size_t i = 0; i < count; ++i) - if (lhs[i] != rhs[i]) - return false; +// Compare lhs with [rhs_begin, rhs_end) +PUGI__FN bool strequalrange(const char_t* lhs, const char_t* rhs, size_t count) +{ + for (size_t i = 0; i < count; ++i) + if (lhs[i] != rhs[i]) + return false; - return lhs[count] == 0; - } + return lhs[count] == 0; +} - // Get length of wide string, even if CRT lacks wide character support - PUGI__FN size_t strlength_wide(const wchar_t* s) - { - assert(s); +// Get length of wide string, even if CRT lacks wide character support +PUGI__FN size_t strlength_wide(const wchar_t* s) +{ + assert(s); - #ifdef PUGIXML_WCHAR_MODE - return wcslen(s); - #else - const wchar_t* end = s; - while (*end) end++; - return static_cast(end - s); - #endif - } +#ifdef PUGIXML_WCHAR_MODE + return wcslen(s); +#else + const wchar_t* end = s; + while (*end) + end++; + return static_cast(end - s); +#endif +} PUGI__NS_END // auto_ptr-like object for exception recovery PUGI__NS_BEGIN - template struct auto_deleter +template +struct auto_deleter +{ + typedef void (*D)(T*); + + T* data; + D deleter; + + auto_deleter(T* data_, D deleter_) + : data(data_) + , deleter(deleter_) { - typedef void (*D)(T*); + } - T* data; - D deleter; + ~auto_deleter() + { + if (data) + deleter(data); + } - auto_deleter(T* data_, D deleter_): data(data_), deleter(deleter_) - { - } - - ~auto_deleter() - { - if (data) deleter(data); - } - - T* release() - { - T* result = data; - data = 0; - return result; - } - }; + T* release() + { + T* result = data; + data = 0; + return result; + } +}; PUGI__NS_END #ifdef PUGIXML_COMPACT PUGI__NS_BEGIN - class compact_hash_table +class compact_hash_table +{ + public: + compact_hash_table() + : _items(0) + , _capacity(0) + , _count(0) { - public: - compact_hash_table(): _items(0), _capacity(0), _count(0) - { - } + } - void clear() - { - if (_items) - { - xml_memory::deallocate(_items); - _items = 0; - _capacity = 0; - _count = 0; - } - } - - void* find(const void* key) - { - if (_capacity == 0) return 0; - - item_t* item = get_item(key); - assert(item); - assert(item->key == key || (item->key == 0 && item->value == 0)); - - return item->value; - } - - void insert(const void* key, void* value) - { - assert(_capacity != 0 && _count < _capacity - _capacity / 4); - - item_t* item = get_item(key); - assert(item); - - if (item->key == 0) - { - _count++; - item->key = key; - } - - item->value = value; - } - - bool reserve(size_t extra = 16) - { - if (_count + extra >= _capacity - _capacity / 4) - return rehash(_count + extra); - - return true; - } - - private: - struct item_t - { - const void* key; - void* value; - }; - - item_t* _items; - size_t _capacity; - - size_t _count; - - bool rehash(size_t count); - - item_t* get_item(const void* key) - { - assert(key); - assert(_capacity > 0); - - size_t hashmod = _capacity - 1; - size_t bucket = hash(key) & hashmod; - - for (size_t probe = 0; probe <= hashmod; ++probe) - { - item_t& probe_item = _items[bucket]; - - if (probe_item.key == key || probe_item.key == 0) - return &probe_item; - - // hash collision, quadratic probing - bucket = (bucket + probe + 1) & hashmod; - } - - assert(false && "Hash table is full"); // unreachable - return 0; - } - - static PUGI__UNSIGNED_OVERFLOW unsigned int hash(const void* key) - { - unsigned int h = static_cast(reinterpret_cast(key)); - - // MurmurHash3 32-bit finalizer - h ^= h >> 16; - h *= 0x85ebca6bu; - h ^= h >> 13; - h *= 0xc2b2ae35u; - h ^= h >> 16; - - return h; - } - }; - - PUGI__FN_NO_INLINE bool compact_hash_table::rehash(size_t count) + void clear() { - size_t capacity = 32; - while (count >= capacity - capacity / 4) - capacity *= 2; - - compact_hash_table rt; - rt._capacity = capacity; - rt._items = static_cast(xml_memory::allocate(sizeof(item_t) * capacity)); - - if (!rt._items) - return false; - - memset(rt._items, 0, sizeof(item_t) * capacity); - - for (size_t i = 0; i < _capacity; ++i) - if (_items[i].key) - rt.insert(_items[i].key, _items[i].value); - if (_items) + { xml_memory::deallocate(_items); + _items = 0; + _capacity = 0; + _count = 0; + } + } - _capacity = capacity; - _items = rt._items; + void* find(const void* key) + { + if (_capacity == 0) + return 0; - assert(_count == rt._count); + item_t* item = get_item(key); + assert(item); + assert(item->key == key || (item->key == 0 && item->value == 0)); + + return item->value; + } + + void insert(const void* key, void* value) + { + assert(_capacity != 0 && _count < _capacity - _capacity / 4); + + item_t* item = get_item(key); + assert(item); + + if (item->key == 0) + { + _count++; + item->key = key; + } + + item->value = value; + } + + bool reserve(size_t extra = 16) + { + if (_count + extra >= _capacity - _capacity / 4) + return rehash(_count + extra); return true; } + private: + struct item_t + { + const void* key; + void* value; + }; + + item_t* _items; + size_t _capacity; + + size_t _count; + + bool rehash(size_t count); + + item_t* get_item(const void* key) + { + assert(key); + assert(_capacity > 0); + + size_t hashmod = _capacity - 1; + size_t bucket = hash(key) & hashmod; + + for (size_t probe = 0; probe <= hashmod; ++probe) + { + item_t& probe_item = _items[bucket]; + + if (probe_item.key == key || probe_item.key == 0) + return &probe_item; + + // hash collision, quadratic probing + bucket = (bucket + probe + 1) & hashmod; + } + + assert(false && "Hash table is full"); // unreachable + return 0; + } + + static PUGI__UNSIGNED_OVERFLOW unsigned int hash(const void* key) + { + unsigned int h = static_cast(reinterpret_cast(key)); + + // MurmurHash3 32-bit finalizer + h ^= h >> 16; + h *= 0x85ebca6bu; + h ^= h >> 13; + h *= 0xc2b2ae35u; + h ^= h >> 16; + + return h; + } +}; + +PUGI__FN_NO_INLINE bool compact_hash_table::rehash(size_t count) +{ + size_t capacity = 32; + while (count >= capacity - capacity / 4) + capacity *= 2; + + compact_hash_table rt; + rt._capacity = capacity; + rt._items = static_cast(xml_memory::allocate(sizeof(item_t) * capacity)); + + if (!rt._items) + return false; + + memset(rt._items, 0, sizeof(item_t) * capacity); + + for (size_t i = 0; i < _capacity; ++i) + if (_items[i].key) + rt.insert(_items[i].key, _items[i].value); + + if (_items) + xml_memory::deallocate(_items); + + _capacity = capacity; + _items = rt._items; + + assert(_count == rt._count); + + return true; +} + PUGI__NS_END #endif PUGI__NS_BEGIN #ifdef PUGIXML_COMPACT - static const uintptr_t xml_memory_block_alignment = 4; +static const uintptr_t xml_memory_block_alignment = 4; #else - static const uintptr_t xml_memory_block_alignment = sizeof(void*); +static const uintptr_t xml_memory_block_alignment = sizeof(void*); #endif - // extra metadata bits - static const uintptr_t xml_memory_page_contents_shared_mask = 64; - static const uintptr_t xml_memory_page_name_allocated_mask = 32; - static const uintptr_t xml_memory_page_value_allocated_mask = 16; - static const uintptr_t xml_memory_page_type_mask = 15; +// extra metadata bits +static const uintptr_t xml_memory_page_contents_shared_mask = 64; +static const uintptr_t xml_memory_page_name_allocated_mask = 32; +static const uintptr_t xml_memory_page_value_allocated_mask = 16; +static const uintptr_t xml_memory_page_type_mask = 15; - // combined masks for string uniqueness - static const uintptr_t xml_memory_page_name_allocated_or_shared_mask = xml_memory_page_name_allocated_mask | xml_memory_page_contents_shared_mask; - static const uintptr_t xml_memory_page_value_allocated_or_shared_mask = xml_memory_page_value_allocated_mask | xml_memory_page_contents_shared_mask; +// combined masks for string uniqueness +static const uintptr_t xml_memory_page_name_allocated_or_shared_mask = xml_memory_page_name_allocated_mask | xml_memory_page_contents_shared_mask; +static const uintptr_t xml_memory_page_value_allocated_or_shared_mask = xml_memory_page_value_allocated_mask | xml_memory_page_contents_shared_mask; #ifdef PUGIXML_COMPACT - #define PUGI__GETHEADER_IMPL(object, page, flags) // unused - #define PUGI__GETPAGE_IMPL(header) (header).get_page() +#define PUGI__GETHEADER_IMPL(object, page, flags) // unused +#define PUGI__GETPAGE_IMPL(header) (header).get_page() #else - #define PUGI__GETHEADER_IMPL(object, page, flags) (((reinterpret_cast(object) - reinterpret_cast(page)) << 8) | (flags)) - // this macro casts pointers through void* to avoid 'cast increases required alignment of target type' warnings - #define PUGI__GETPAGE_IMPL(header) static_cast(const_cast(static_cast(reinterpret_cast(&header) - (header >> 8)))) +#define PUGI__GETHEADER_IMPL(object, page, flags) (((reinterpret_cast(object) - reinterpret_cast(page)) << 8) | (flags)) +// this macro casts pointers through void* to avoid 'cast increases required alignment of target type' warnings +#define PUGI__GETPAGE_IMPL(header) static_cast(const_cast(static_cast(reinterpret_cast(&header) - (header >> 8)))) #endif - #define PUGI__GETPAGE(n) PUGI__GETPAGE_IMPL((n)->header) - #define PUGI__NODETYPE(n) static_cast((n)->header & impl::xml_memory_page_type_mask) +#define PUGI__GETPAGE(n) PUGI__GETPAGE_IMPL((n)->header) +#define PUGI__NODETYPE(n) static_cast((n)->header & impl::xml_memory_page_type_mask) - struct xml_allocator; +struct xml_allocator; - struct xml_memory_page +struct xml_memory_page +{ + static xml_memory_page* construct(void* memory) { - static xml_memory_page* construct(void* memory) - { - xml_memory_page* result = static_cast(memory); + xml_memory_page* result = static_cast(memory); - result->allocator = 0; - result->prev = 0; - result->next = 0; - result->busy_size = 0; - result->freed_size = 0; + result->allocator = 0; + result->prev = 0; + result->next = 0; + result->busy_size = 0; + result->freed_size = 0; - #ifdef PUGIXML_COMPACT - result->compact_string_base = 0; - result->compact_shared_parent = 0; - result->compact_page_marker = 0; - #endif +#ifdef PUGIXML_COMPACT + result->compact_string_base = 0; + result->compact_shared_parent = 0; + result->compact_page_marker = 0; +#endif - return result; - } + return result; + } - xml_allocator* allocator; + xml_allocator* allocator; - xml_memory_page* prev; - xml_memory_page* next; + xml_memory_page* prev; + xml_memory_page* next; - size_t busy_size; - size_t freed_size; + size_t busy_size; + size_t freed_size; - #ifdef PUGIXML_COMPACT - char_t* compact_string_base; - void* compact_shared_parent; - uint32_t* compact_page_marker; - #endif - }; +#ifdef PUGIXML_COMPACT + char_t* compact_string_base; + void* compact_shared_parent; + uint32_t* compact_page_marker; +#endif +}; - static const size_t xml_memory_page_size = - #ifdef PUGIXML_MEMORY_PAGE_SIZE - (PUGIXML_MEMORY_PAGE_SIZE) - #else - 32768 - #endif - - sizeof(xml_memory_page); +static const size_t xml_memory_page_size = +#ifdef PUGIXML_MEMORY_PAGE_SIZE + (PUGIXML_MEMORY_PAGE_SIZE) +#else + 32768 +#endif + - sizeof(xml_memory_page); - struct xml_memory_string_header +struct xml_memory_string_header +{ + uint16_t page_offset; // offset from page->data + uint16_t full_size; // 0 if string occupies whole page +}; + +struct xml_allocator +{ + xml_allocator(xml_memory_page* root) + : _root(root) + , _busy_size(root->busy_size) { - uint16_t page_offset; // offset from page->data - uint16_t full_size; // 0 if string occupies whole page - }; +#ifdef PUGIXML_COMPACT + _hash = 0; +#endif + } - struct xml_allocator + xml_memory_page* allocate_page(size_t data_size) { - xml_allocator(xml_memory_page* root): _root(root), _busy_size(root->busy_size) - { - #ifdef PUGIXML_COMPACT - _hash = 0; - #endif - } + size_t size = sizeof(xml_memory_page) + data_size; - xml_memory_page* allocate_page(size_t data_size) - { - size_t size = sizeof(xml_memory_page) + data_size; + // allocate block with some alignment, leaving memory for worst-case padding + void* memory = xml_memory::allocate(size); + if (!memory) + return 0; - // allocate block with some alignment, leaving memory for worst-case padding - void* memory = xml_memory::allocate(size); - if (!memory) return 0; + // prepare page structure + xml_memory_page* page = xml_memory_page::construct(memory); + assert(page); - // prepare page structure - xml_memory_page* page = xml_memory_page::construct(memory); - assert(page); + page->allocator = _root->allocator; - page->allocator = _root->allocator; + return page; + } - return page; - } - - static void deallocate_page(xml_memory_page* page) - { - xml_memory::deallocate(page); - } - - void* allocate_memory_oob(size_t size, xml_memory_page*& out_page); - - void* allocate_memory(size_t size, xml_memory_page*& out_page) - { - if (PUGI__UNLIKELY(_busy_size + size > xml_memory_page_size)) - return allocate_memory_oob(size, out_page); - - void* buf = reinterpret_cast(_root) + sizeof(xml_memory_page) + _busy_size; - - _busy_size += size; - - out_page = _root; - - return buf; - } - - #ifdef PUGIXML_COMPACT - void* allocate_object(size_t size, xml_memory_page*& out_page) - { - void* result = allocate_memory(size + sizeof(uint32_t), out_page); - if (!result) return 0; - - // adjust for marker - ptrdiff_t offset = static_cast(result) - reinterpret_cast(out_page->compact_page_marker); - - if (PUGI__UNLIKELY(static_cast(offset) >= 256 * xml_memory_block_alignment)) - { - // insert new marker - uint32_t* marker = static_cast(result); - - *marker = static_cast(reinterpret_cast(marker) - reinterpret_cast(out_page)); - out_page->compact_page_marker = marker; - - // since we don't reuse the page space until we reallocate it, we can just pretend that we freed the marker block - // this will make sure deallocate_memory correctly tracks the size - out_page->freed_size += sizeof(uint32_t); - - return marker + 1; - } - else - { - // roll back uint32_t part - _busy_size -= sizeof(uint32_t); - - return result; - } - } - #else - void* allocate_object(size_t size, xml_memory_page*& out_page) - { - return allocate_memory(size, out_page); - } - #endif - - void deallocate_memory(void* ptr, size_t size, xml_memory_page* page) - { - if (page == _root) page->busy_size = _busy_size; - - assert(ptr >= reinterpret_cast(page) + sizeof(xml_memory_page) && ptr < reinterpret_cast(page) + sizeof(xml_memory_page) + page->busy_size); - (void)!ptr; - - page->freed_size += size; - assert(page->freed_size <= page->busy_size); - - if (page->freed_size == page->busy_size) - { - if (page->next == 0) - { - assert(_root == page); - - // top page freed, just reset sizes - page->busy_size = 0; - page->freed_size = 0; - - #ifdef PUGIXML_COMPACT - // reset compact state to maximize efficiency - page->compact_string_base = 0; - page->compact_shared_parent = 0; - page->compact_page_marker = 0; - #endif - - _busy_size = 0; - } - else - { - assert(_root != page); - assert(page->prev); - - // remove from the list - page->prev->next = page->next; - page->next->prev = page->prev; - - // deallocate - deallocate_page(page); - } - } - } - - char_t* allocate_string(size_t length) - { - static const size_t max_encoded_offset = (1 << 16) * xml_memory_block_alignment; - - PUGI__STATIC_ASSERT(xml_memory_page_size <= max_encoded_offset); - - // allocate memory for string and header block - size_t size = sizeof(xml_memory_string_header) + length * sizeof(char_t); - - // round size up to block alignment boundary - size_t full_size = (size + (xml_memory_block_alignment - 1)) & ~(xml_memory_block_alignment - 1); - - xml_memory_page* page; - xml_memory_string_header* header = static_cast(allocate_memory(full_size, page)); - - if (!header) return 0; - - // setup header - ptrdiff_t page_offset = reinterpret_cast(header) - reinterpret_cast(page) - sizeof(xml_memory_page); - - assert(page_offset % xml_memory_block_alignment == 0); - assert(page_offset >= 0 && static_cast(page_offset) < max_encoded_offset); - header->page_offset = static_cast(static_cast(page_offset) / xml_memory_block_alignment); - - // full_size == 0 for large strings that occupy the whole page - assert(full_size % xml_memory_block_alignment == 0); - assert(full_size < max_encoded_offset || (page->busy_size == full_size && page_offset == 0)); - header->full_size = static_cast(full_size < max_encoded_offset ? full_size / xml_memory_block_alignment : 0); - - // round-trip through void* to avoid 'cast increases required alignment of target type' warning - // header is guaranteed a pointer-sized alignment, which should be enough for char_t - return static_cast(static_cast(header + 1)); - } - - void deallocate_string(char_t* string) - { - // this function casts pointers through void* to avoid 'cast increases required alignment of target type' warnings - // we're guaranteed the proper (pointer-sized) alignment on the input string if it was allocated via allocate_string - - // get header - xml_memory_string_header* header = static_cast(static_cast(string)) - 1; - assert(header); - - // deallocate - size_t page_offset = sizeof(xml_memory_page) + header->page_offset * xml_memory_block_alignment; - xml_memory_page* page = reinterpret_cast(static_cast(reinterpret_cast(header) - page_offset)); - - // if full_size == 0 then this string occupies the whole page - size_t full_size = header->full_size == 0 ? page->busy_size : header->full_size * xml_memory_block_alignment; - - deallocate_memory(header, full_size, page); - } - - bool reserve() - { - #ifdef PUGIXML_COMPACT - return _hash->reserve(); - #else - return true; - #endif - } - - xml_memory_page* _root; - size_t _busy_size; - - #ifdef PUGIXML_COMPACT - compact_hash_table* _hash; - #endif - }; - - PUGI__FN_NO_INLINE void* xml_allocator::allocate_memory_oob(size_t size, xml_memory_page*& out_page) + static void deallocate_page(xml_memory_page* page) { - const size_t large_allocation_threshold = xml_memory_page_size / 4; + xml_memory::deallocate(page); + } - xml_memory_page* page = allocate_page(size <= large_allocation_threshold ? xml_memory_page_size : size); - out_page = page; + void* allocate_memory_oob(size_t size, xml_memory_page*& out_page); - if (!page) return 0; + void* allocate_memory(size_t size, xml_memory_page*& out_page) + { + if (PUGI__UNLIKELY(_busy_size + size > xml_memory_page_size)) + return allocate_memory_oob(size, out_page); - if (size <= large_allocation_threshold) + void* buf = reinterpret_cast(_root) + sizeof(xml_memory_page) + _busy_size; + + _busy_size += size; + + out_page = _root; + + return buf; + } + +#ifdef PUGIXML_COMPACT + void* allocate_object(size_t size, xml_memory_page*& out_page) + { + void* result = allocate_memory(size + sizeof(uint32_t), out_page); + if (!result) + return 0; + + // adjust for marker + ptrdiff_t offset = static_cast(result) - reinterpret_cast(out_page->compact_page_marker); + + if (PUGI__UNLIKELY(static_cast(offset) >= 256 * xml_memory_block_alignment)) { - _root->busy_size = _busy_size; + // insert new marker + uint32_t* marker = static_cast(result); - // insert page at the end of linked list - page->prev = _root; - _root->next = page; - _root = page; + *marker = static_cast(reinterpret_cast(marker) - reinterpret_cast(out_page)); + out_page->compact_page_marker = marker; - _busy_size = size; + // since we don't reuse the page space until we reallocate it, we can just pretend that we freed the marker block + // this will make sure deallocate_memory correctly tracks the size + out_page->freed_size += sizeof(uint32_t); + + return marker + 1; } else { - // insert page before the end of linked list, so that it is deleted as soon as possible - // the last page is not deleted even if it's empty (see deallocate_memory) - assert(_root->prev); + // roll back uint32_t part + _busy_size -= sizeof(uint32_t); - page->prev = _root->prev; - page->next = _root; - - _root->prev->next = page; - _root->prev = page; - - page->busy_size = size; + return result; } - - return reinterpret_cast(page) + sizeof(xml_memory_page); } +#else + void* allocate_object(size_t size, xml_memory_page*& out_page) + { + return allocate_memory(size, out_page); + } +#endif + + void deallocate_memory(void* ptr, size_t size, xml_memory_page* page) + { + if (page == _root) + page->busy_size = _busy_size; + + assert(ptr >= reinterpret_cast(page) + sizeof(xml_memory_page) && ptr < reinterpret_cast(page) + sizeof(xml_memory_page) + page->busy_size); + (void)!ptr; + + page->freed_size += size; + assert(page->freed_size <= page->busy_size); + + if (page->freed_size == page->busy_size) + { + if (page->next == 0) + { + assert(_root == page); + + // top page freed, just reset sizes + page->busy_size = 0; + page->freed_size = 0; + +#ifdef PUGIXML_COMPACT + // reset compact state to maximize efficiency + page->compact_string_base = 0; + page->compact_shared_parent = 0; + page->compact_page_marker = 0; +#endif + + _busy_size = 0; + } + else + { + assert(_root != page); + assert(page->prev); + + // remove from the list + page->prev->next = page->next; + page->next->prev = page->prev; + + // deallocate + deallocate_page(page); + } + } + } + + char_t* allocate_string(size_t length) + { + static const size_t max_encoded_offset = (1 << 16) * xml_memory_block_alignment; + + PUGI__STATIC_ASSERT(xml_memory_page_size <= max_encoded_offset); + + // allocate memory for string and header block + size_t size = sizeof(xml_memory_string_header) + length * sizeof(char_t); + + // round size up to block alignment boundary + size_t full_size = (size + (xml_memory_block_alignment - 1)) & ~(xml_memory_block_alignment - 1); + + xml_memory_page* page; + xml_memory_string_header* header = static_cast(allocate_memory(full_size, page)); + + if (!header) + return 0; + + // setup header + ptrdiff_t page_offset = reinterpret_cast(header) - reinterpret_cast(page) - sizeof(xml_memory_page); + + assert(page_offset % xml_memory_block_alignment == 0); + assert(page_offset >= 0 && static_cast(page_offset) < max_encoded_offset); + header->page_offset = static_cast(static_cast(page_offset) / xml_memory_block_alignment); + + // full_size == 0 for large strings that occupy the whole page + assert(full_size % xml_memory_block_alignment == 0); + assert(full_size < max_encoded_offset || (page->busy_size == full_size && page_offset == 0)); + header->full_size = static_cast(full_size < max_encoded_offset ? full_size / xml_memory_block_alignment : 0); + + // round-trip through void* to avoid 'cast increases required alignment of target type' warning + // header is guaranteed a pointer-sized alignment, which should be enough for char_t + return static_cast(static_cast(header + 1)); + } + + void deallocate_string(char_t* string) + { + // this function casts pointers through void* to avoid 'cast increases required alignment of target type' warnings + // we're guaranteed the proper (pointer-sized) alignment on the input string if it was allocated via allocate_string + + // get header + xml_memory_string_header* header = static_cast(static_cast(string)) - 1; + assert(header); + + // deallocate + size_t page_offset = sizeof(xml_memory_page) + header->page_offset * xml_memory_block_alignment; + xml_memory_page* page = reinterpret_cast(static_cast(reinterpret_cast(header) - page_offset)); + + // if full_size == 0 then this string occupies the whole page + size_t full_size = header->full_size == 0 ? page->busy_size : header->full_size * xml_memory_block_alignment; + + deallocate_memory(header, full_size, page); + } + + bool reserve() + { +#ifdef PUGIXML_COMPACT + return _hash->reserve(); +#else + return true; +#endif + } + + xml_memory_page* _root; + size_t _busy_size; + +#ifdef PUGIXML_COMPACT + compact_hash_table* _hash; +#endif +}; + +PUGI__FN_NO_INLINE void* xml_allocator::allocate_memory_oob(size_t size, xml_memory_page*& out_page) +{ + const size_t large_allocation_threshold = xml_memory_page_size / 4; + + xml_memory_page* page = allocate_page(size <= large_allocation_threshold ? xml_memory_page_size : size); + out_page = page; + + if (!page) + return 0; + + if (size <= large_allocation_threshold) + { + _root->busy_size = _busy_size; + + // insert page at the end of linked list + page->prev = _root; + _root->next = page; + _root = page; + + _busy_size = size; + } + else + { + // insert page before the end of linked list, so that it is deleted as soon as possible + // the last page is not deleted even if it's empty (see deallocate_memory) + assert(_root->prev); + + page->prev = _root->prev; + page->next = _root; + + _root->prev->next = page; + _root->prev = page; + + page->busy_size = size; + } + + return reinterpret_cast(page) + sizeof(xml_memory_page); +} PUGI__NS_END #ifdef PUGIXML_COMPACT PUGI__NS_BEGIN - static const uintptr_t compact_alignment_log2 = 2; - static const uintptr_t compact_alignment = 1 << compact_alignment_log2; +static const uintptr_t compact_alignment_log2 = 2; +static const uintptr_t compact_alignment = 1 << compact_alignment_log2; - class compact_header +class compact_header +{ + public: + compact_header(xml_memory_page* page, unsigned int flags) { - public: - compact_header(xml_memory_page* page, unsigned int flags) - { - PUGI__STATIC_ASSERT(xml_memory_block_alignment == compact_alignment); + PUGI__STATIC_ASSERT(xml_memory_block_alignment == compact_alignment); - ptrdiff_t offset = (reinterpret_cast(this) - reinterpret_cast(page->compact_page_marker)); - assert(offset % compact_alignment == 0 && static_cast(offset) < 256 * compact_alignment); + ptrdiff_t offset = (reinterpret_cast(this) - reinterpret_cast(page->compact_page_marker)); + assert(offset % compact_alignment == 0 && static_cast(offset) < 256 * compact_alignment); - _page = static_cast(offset >> compact_alignment_log2); - _flags = static_cast(flags); - } - - void operator&=(uintptr_t mod) - { - _flags &= static_cast(mod); - } - - void operator|=(uintptr_t mod) - { - _flags |= static_cast(mod); - } - - uintptr_t operator&(uintptr_t mod) const - { - return _flags & mod; - } - - xml_memory_page* get_page() const - { - // round-trip through void* to silence 'cast increases required alignment of target type' warnings - const char* page_marker = reinterpret_cast(this) - (_page << compact_alignment_log2); - const char* page = page_marker - *reinterpret_cast(static_cast(page_marker)); - - return const_cast(reinterpret_cast(static_cast(page))); - } - - private: - unsigned char _page; - unsigned char _flags; - }; - - PUGI__FN xml_memory_page* compact_get_page(const void* object, int header_offset) - { - const compact_header* header = reinterpret_cast(static_cast(object) - header_offset); - - return header->get_page(); + _page = static_cast(offset >> compact_alignment_log2); + _flags = static_cast(flags); } - template PUGI__FN_NO_INLINE T* compact_get_value(const void* object) + void operator&=(uintptr_t mod) { - return static_cast(compact_get_page(object, header_offset)->allocator->_hash->find(object)); + _flags &= static_cast(mod); } - template PUGI__FN_NO_INLINE void compact_set_value(const void* object, T* value) + void operator|=(uintptr_t mod) { - compact_get_page(object, header_offset)->allocator->_hash->insert(object, value); + _flags |= static_cast(mod); } - template class compact_pointer + uintptr_t operator&(uintptr_t mod) const { - public: - compact_pointer(): _data(0) - { - } + return _flags & mod; + } - void operator=(const compact_pointer& rhs) - { - *this = rhs + 0; - } + xml_memory_page* get_page() const + { + // round-trip through void* to silence 'cast increases required alignment of target type' warnings + const char* page_marker = reinterpret_cast(this) - (_page << compact_alignment_log2); + const char* page = page_marker - *reinterpret_cast(static_cast(page_marker)); - void operator=(T* value) + return const_cast(reinterpret_cast(static_cast(page))); + } + + private: + unsigned char _page; + unsigned char _flags; +}; + +PUGI__FN xml_memory_page* compact_get_page(const void* object, int header_offset) +{ + const compact_header* header = reinterpret_cast(static_cast(object) - header_offset); + + return header->get_page(); +} + +template +PUGI__FN_NO_INLINE T* compact_get_value(const void* object) +{ + return static_cast(compact_get_page(object, header_offset)->allocator->_hash->find(object)); +} + +template +PUGI__FN_NO_INLINE void compact_set_value(const void* object, T* value) +{ + compact_get_page(object, header_offset)->allocator->_hash->insert(object, value); +} + +template +class compact_pointer +{ + public: + compact_pointer() + : _data(0) + { + } + + void operator=(const compact_pointer& rhs) + { + *this = rhs + 0; + } + + void operator=(T* value) + { + if (value) { - if (value) + // value is guaranteed to be compact-aligned; 'this' is not + // our decoding is based on 'this' aligned to compact alignment downwards (see operator T*) + // so for negative offsets (e.g. -3) we need to adjust the diff by compact_alignment - 1 to + // compensate for arithmetic shift rounding for negative values + ptrdiff_t diff = reinterpret_cast(value) - reinterpret_cast(this); + ptrdiff_t offset = ((diff + int(compact_alignment - 1)) >> compact_alignment_log2) - start; + + if (static_cast(offset) <= 253) + _data = static_cast(offset + 1); + else { - // value is guaranteed to be compact-aligned; 'this' is not - // our decoding is based on 'this' aligned to compact alignment downwards (see operator T*) - // so for negative offsets (e.g. -3) we need to adjust the diff by compact_alignment - 1 to - // compensate for arithmetic shift rounding for negative values - ptrdiff_t diff = reinterpret_cast(value) - reinterpret_cast(this); - ptrdiff_t offset = ((diff + int(compact_alignment - 1)) >> compact_alignment_log2) - start; + compact_set_value(this, value); - if (static_cast(offset) <= 253) - _data = static_cast(offset + 1); + _data = 255; + } + } + else + _data = 0; + } + + operator T*() const + { + if (_data) + { + if (_data < 255) + { + uintptr_t base = reinterpret_cast(this) & ~(compact_alignment - 1); + + return reinterpret_cast(base + (_data - 1 + start) * compact_alignment); + } + else + return compact_get_value(this); + } + else + return 0; + } + + T* operator->() const + { + return *this; + } + + private: + unsigned char _data; +}; + +template +class compact_pointer_parent +{ + public: + compact_pointer_parent() + : _data(0) + { + } + + void operator=(const compact_pointer_parent& rhs) + { + *this = rhs + 0; + } + + void operator=(T* value) + { + if (value) + { + // value is guaranteed to be compact-aligned; 'this' is not + // our decoding is based on 'this' aligned to compact alignment downwards (see operator T*) + // so for negative offsets (e.g. -3) we need to adjust the diff by compact_alignment - 1 to + // compensate for arithmetic shift behavior for negative values + ptrdiff_t diff = reinterpret_cast(value) - reinterpret_cast(this); + ptrdiff_t offset = ((diff + int(compact_alignment - 1)) >> compact_alignment_log2) + 65533; + + if (static_cast(offset) <= 65533) + { + _data = static_cast(offset + 1); + } + else + { + xml_memory_page* page = compact_get_page(this, header_offset); + + if (PUGI__UNLIKELY(page->compact_shared_parent == 0)) + page->compact_shared_parent = value; + + if (page->compact_shared_parent == value) + { + _data = 65534; + } else { compact_set_value(this, value); - _data = 255; + _data = 65535; } } - else - _data = 0; } - - operator T*() const + else { - if (_data) - { - if (_data < 255) - { - uintptr_t base = reinterpret_cast(this) & ~(compact_alignment - 1); - - return reinterpret_cast(base + (_data - 1 + start) * compact_alignment); - } - else - return compact_get_value(this); - } - else - return 0; + _data = 0; } + } - T* operator->() const - { - return *this; - } - - private: - unsigned char _data; - }; - - template class compact_pointer_parent + operator T*() const { - public: - compact_pointer_parent(): _data(0) + if (_data) { - } - - void operator=(const compact_pointer_parent& rhs) - { - *this = rhs + 0; - } - - void operator=(T* value) - { - if (value) + if (_data < 65534) { - // value is guaranteed to be compact-aligned; 'this' is not - // our decoding is based on 'this' aligned to compact alignment downwards (see operator T*) - // so for negative offsets (e.g. -3) we need to adjust the diff by compact_alignment - 1 to - // compensate for arithmetic shift behavior for negative values - ptrdiff_t diff = reinterpret_cast(value) - reinterpret_cast(this); - ptrdiff_t offset = ((diff + int(compact_alignment - 1)) >> compact_alignment_log2) + 65533; + uintptr_t base = reinterpret_cast(this) & ~(compact_alignment - 1); - if (static_cast(offset) <= 65533) + return reinterpret_cast(base + (_data - 1 - 65533) * compact_alignment); + } + else if (_data == 65534) + return static_cast(compact_get_page(this, header_offset)->compact_shared_parent); + else + return compact_get_value(this); + } + else + return 0; + } + + T* operator->() const + { + return *this; + } + + private: + uint16_t _data; +}; + +template +class compact_string +{ + public: + compact_string() + : _data(0) + { + } + + void operator=(const compact_string& rhs) + { + *this = rhs + 0; + } + + void operator=(char_t* value) + { + if (value) + { + xml_memory_page* page = compact_get_page(this, header_offset); + + if (PUGI__UNLIKELY(page->compact_string_base == 0)) + page->compact_string_base = value; + + ptrdiff_t offset = value - page->compact_string_base; + + if (static_cast(offset) < (65535 << 7)) + { + // round-trip through void* to silence 'cast increases required alignment of target type' warnings + uint16_t* base = reinterpret_cast(static_cast(reinterpret_cast(this) - base_offset)); + + if (*base == 0) { - _data = static_cast(offset + 1); + *base = static_cast((offset >> 7) + 1); + _data = static_cast((offset & 127) + 1); } else { - xml_memory_page* page = compact_get_page(this, header_offset); + ptrdiff_t remainder = offset - ((*base - 1) << 7); - if (PUGI__UNLIKELY(page->compact_shared_parent == 0)) - page->compact_shared_parent = value; - - if (page->compact_shared_parent == value) + if (static_cast(remainder) <= 253) { - _data = 65534; + _data = static_cast(remainder + 1); } else { compact_set_value(this, value); - _data = 65535; + _data = 255; } } } else { - _data = 0; + compact_set_value(this, value); + + _data = 255; } } - - operator T*() const + else { - if (_data) - { - if (_data < 65534) - { - uintptr_t base = reinterpret_cast(this) & ~(compact_alignment - 1); - - return reinterpret_cast(base + (_data - 1 - 65533) * compact_alignment); - } - else if (_data == 65534) - return static_cast(compact_get_page(this, header_offset)->compact_shared_parent); - else - return compact_get_value(this); - } - else - return 0; + _data = 0; } + } - T* operator->() const - { - return *this; - } - - private: - uint16_t _data; - }; - - template class compact_string + operator char_t*() const { - public: - compact_string(): _data(0) + if (_data) { - } - - void operator=(const compact_string& rhs) - { - *this = rhs + 0; - } - - void operator=(char_t* value) - { - if (value) + if (_data < 255) { xml_memory_page* page = compact_get_page(this, header_offset); - if (PUGI__UNLIKELY(page->compact_string_base == 0)) - page->compact_string_base = value; + // round-trip through void* to silence 'cast increases required alignment of target type' warnings + const uint16_t* base = reinterpret_cast(static_cast(reinterpret_cast(this) - base_offset)); + assert(*base); - ptrdiff_t offset = value - page->compact_string_base; + ptrdiff_t offset = ((*base - 1) << 7) + (_data - 1); - if (static_cast(offset) < (65535 << 7)) - { - // round-trip through void* to silence 'cast increases required alignment of target type' warnings - uint16_t* base = reinterpret_cast(static_cast(reinterpret_cast(this) - base_offset)); - - if (*base == 0) - { - *base = static_cast((offset >> 7) + 1); - _data = static_cast((offset & 127) + 1); - } - else - { - ptrdiff_t remainder = offset - ((*base - 1) << 7); - - if (static_cast(remainder) <= 253) - { - _data = static_cast(remainder + 1); - } - else - { - compact_set_value(this, value); - - _data = 255; - } - } - } - else - { - compact_set_value(this, value); - - _data = 255; - } + return page->compact_string_base + offset; } else { - _data = 0; + return compact_get_value(this); } } + else + return 0; + } - operator char_t*() const - { - if (_data) - { - if (_data < 255) - { - xml_memory_page* page = compact_get_page(this, header_offset); - - // round-trip through void* to silence 'cast increases required alignment of target type' warnings - const uint16_t* base = reinterpret_cast(static_cast(reinterpret_cast(this) - base_offset)); - assert(*base); - - ptrdiff_t offset = ((*base - 1) << 7) + (_data - 1); - - return page->compact_string_base + offset; - } - else - { - return compact_get_value(this); - } - } - else - return 0; - } - - private: - unsigned char _data; - }; + private: + unsigned char _data; +}; PUGI__NS_END #endif #ifdef PUGIXML_COMPACT namespace pugi { - struct xml_attribute_struct +struct xml_attribute_struct +{ + xml_attribute_struct(impl::xml_memory_page* page) + : header(page, 0) + , namevalue_base(0) { - xml_attribute_struct(impl::xml_memory_page* page): header(page, 0), namevalue_base(0) - { - PUGI__STATIC_ASSERT(sizeof(xml_attribute_struct) == 8); - } + PUGI__STATIC_ASSERT(sizeof(xml_attribute_struct) == 8); + } - impl::compact_header header; + impl::compact_header header; - uint16_t namevalue_base; + uint16_t namevalue_base; - impl::compact_string<4, 2> name; - impl::compact_string<5, 3> value; + impl::compact_string<4, 2> name; + impl::compact_string<5, 3> value; - impl::compact_pointer prev_attribute_c; - impl::compact_pointer next_attribute; - }; + impl::compact_pointer prev_attribute_c; + impl::compact_pointer next_attribute; +}; - struct xml_node_struct +struct xml_node_struct +{ + xml_node_struct(impl::xml_memory_page* page, xml_node_type type) + : header(page, type) + , namevalue_base(0) { - xml_node_struct(impl::xml_memory_page* page, xml_node_type type): header(page, type), namevalue_base(0) - { - PUGI__STATIC_ASSERT(sizeof(xml_node_struct) == 12); - } + PUGI__STATIC_ASSERT(sizeof(xml_node_struct) == 12); + } - impl::compact_header header; + impl::compact_header header; - uint16_t namevalue_base; + uint16_t namevalue_base; - impl::compact_string<4, 2> name; - impl::compact_string<5, 3> value; + impl::compact_string<4, 2> name; + impl::compact_string<5, 3> value; - impl::compact_pointer_parent parent; + impl::compact_pointer_parent parent; - impl::compact_pointer first_child; + impl::compact_pointer first_child; - impl::compact_pointer prev_sibling_c; - impl::compact_pointer next_sibling; + impl::compact_pointer prev_sibling_c; + impl::compact_pointer next_sibling; - impl::compact_pointer first_attribute; - }; -} + impl::compact_pointer first_attribute; +}; +} // namespace pugi #else namespace pugi { - struct xml_attribute_struct +struct xml_attribute_struct +{ + xml_attribute_struct(impl::xml_memory_page* page) + : name(0) + , value(0) + , prev_attribute_c(0) + , next_attribute(0) { - xml_attribute_struct(impl::xml_memory_page* page): name(0), value(0), prev_attribute_c(0), next_attribute(0) - { - header = PUGI__GETHEADER_IMPL(this, page, 0); - } + header = PUGI__GETHEADER_IMPL(this, page, 0); + } - uintptr_t header; + uintptr_t header; - char_t* name; - char_t* value; + char_t* name; + char_t* value; - xml_attribute_struct* prev_attribute_c; - xml_attribute_struct* next_attribute; - }; + xml_attribute_struct* prev_attribute_c; + xml_attribute_struct* next_attribute; +}; - struct xml_node_struct +struct xml_node_struct +{ + xml_node_struct(impl::xml_memory_page* page, xml_node_type type) + : name(0) + , value(0) + , parent(0) + , first_child(0) + , prev_sibling_c(0) + , next_sibling(0) + , first_attribute(0) { - xml_node_struct(impl::xml_memory_page* page, xml_node_type type): name(0), value(0), parent(0), first_child(0), prev_sibling_c(0), next_sibling(0), first_attribute(0) - { - header = PUGI__GETHEADER_IMPL(this, page, type); - } + header = PUGI__GETHEADER_IMPL(this, page, type); + } - uintptr_t header; + uintptr_t header; - char_t* name; - char_t* value; + char_t* name; + char_t* value; - xml_node_struct* parent; + xml_node_struct* parent; - xml_node_struct* first_child; + xml_node_struct* first_child; - xml_node_struct* prev_sibling_c; - xml_node_struct* next_sibling; + xml_node_struct* prev_sibling_c; + xml_node_struct* next_sibling; - xml_attribute_struct* first_attribute; - }; -} + xml_attribute_struct* first_attribute; +}; +} // namespace pugi #endif PUGI__NS_BEGIN - struct xml_extra_buffer +struct xml_extra_buffer +{ + char_t* buffer; + xml_extra_buffer* next; +}; + +struct xml_document_struct : public xml_node_struct, public xml_allocator +{ + xml_document_struct(xml_memory_page* page) + : xml_node_struct(page, node_document) + , xml_allocator(page) + , buffer(0) + , extra_buffers(0) { - char_t* buffer; - xml_extra_buffer* next; - }; - - struct xml_document_struct: public xml_node_struct, public xml_allocator - { - xml_document_struct(xml_memory_page* page): xml_node_struct(page, node_document), xml_allocator(page), buffer(0), extra_buffers(0) - { - } - - const char_t* buffer; - - xml_extra_buffer* extra_buffers; - - #ifdef PUGIXML_COMPACT - compact_hash_table hash; - #endif - }; - - template inline xml_allocator& get_allocator(const Object* object) - { - assert(object); - - return *PUGI__GETPAGE(object)->allocator; } - template inline xml_document_struct& get_document(const Object* object) - { - assert(object); + const char_t* buffer; - return *static_cast(PUGI__GETPAGE(object)->allocator); - } + xml_extra_buffer* extra_buffers; + +#ifdef PUGIXML_COMPACT + compact_hash_table hash; +#endif +}; + +template +inline xml_allocator& get_allocator(const Object* object) +{ + assert(object); + + return *PUGI__GETPAGE(object)->allocator; +} + +template +inline xml_document_struct& get_document(const Object* object) +{ + assert(object); + + return *static_cast(PUGI__GETPAGE(object)->allocator); +} PUGI__NS_END // Low-level DOM operations PUGI__NS_BEGIN - inline xml_attribute_struct* allocate_attribute(xml_allocator& alloc) - { - xml_memory_page* page; - void* memory = alloc.allocate_object(sizeof(xml_attribute_struct), page); - if (!memory) return 0; +inline xml_attribute_struct* allocate_attribute(xml_allocator& alloc) +{ + xml_memory_page* page; + void* memory = alloc.allocate_object(sizeof(xml_attribute_struct), page); + if (!memory) + return 0; - return new (memory) xml_attribute_struct(page); + return new (memory) xml_attribute_struct(page); +} + +inline xml_node_struct* allocate_node(xml_allocator& alloc, xml_node_type type) +{ + xml_memory_page* page; + void* memory = alloc.allocate_object(sizeof(xml_node_struct), page); + if (!memory) + return 0; + + return new (memory) xml_node_struct(page, type); +} + +inline void destroy_attribute(xml_attribute_struct* a, xml_allocator& alloc) +{ + if (a->header & impl::xml_memory_page_name_allocated_mask) + alloc.deallocate_string(a->name); + + if (a->header & impl::xml_memory_page_value_allocated_mask) + alloc.deallocate_string(a->value); + + alloc.deallocate_memory(a, sizeof(xml_attribute_struct), PUGI__GETPAGE(a)); +} + +inline void destroy_node(xml_node_struct* n, xml_allocator& alloc) +{ + if (n->header & impl::xml_memory_page_name_allocated_mask) + alloc.deallocate_string(n->name); + + if (n->header & impl::xml_memory_page_value_allocated_mask) + alloc.deallocate_string(n->value); + + for (xml_attribute_struct* attr = n->first_attribute; attr;) + { + xml_attribute_struct* next = attr->next_attribute; + + destroy_attribute(attr, alloc); + + attr = next; } - inline xml_node_struct* allocate_node(xml_allocator& alloc, xml_node_type type) + for (xml_node_struct* child = n->first_child; child;) { - xml_memory_page* page; - void* memory = alloc.allocate_object(sizeof(xml_node_struct), page); - if (!memory) return 0; + xml_node_struct* next = child->next_sibling; - return new (memory) xml_node_struct(page, type); + destroy_node(child, alloc); + + child = next; } - inline void destroy_attribute(xml_attribute_struct* a, xml_allocator& alloc) + alloc.deallocate_memory(n, sizeof(xml_node_struct), PUGI__GETPAGE(n)); +} + +inline void append_node(xml_node_struct* child, xml_node_struct* node) +{ + child->parent = node; + + xml_node_struct* head = node->first_child; + + if (head) { - if (a->header & impl::xml_memory_page_name_allocated_mask) - alloc.deallocate_string(a->name); + xml_node_struct* tail = head->prev_sibling_c; - if (a->header & impl::xml_memory_page_value_allocated_mask) - alloc.deallocate_string(a->value); - - alloc.deallocate_memory(a, sizeof(xml_attribute_struct), PUGI__GETPAGE(a)); + tail->next_sibling = child; + child->prev_sibling_c = tail; + head->prev_sibling_c = child; } - - inline void destroy_node(xml_node_struct* n, xml_allocator& alloc) + else { - if (n->header & impl::xml_memory_page_name_allocated_mask) - alloc.deallocate_string(n->name); - - if (n->header & impl::xml_memory_page_value_allocated_mask) - alloc.deallocate_string(n->value); - - for (xml_attribute_struct* attr = n->first_attribute; attr; ) - { - xml_attribute_struct* next = attr->next_attribute; - - destroy_attribute(attr, alloc); - - attr = next; - } - - for (xml_node_struct* child = n->first_child; child; ) - { - xml_node_struct* next = child->next_sibling; - - destroy_node(child, alloc); - - child = next; - } - - alloc.deallocate_memory(n, sizeof(xml_node_struct), PUGI__GETPAGE(n)); - } - - inline void append_node(xml_node_struct* child, xml_node_struct* node) - { - child->parent = node; - - xml_node_struct* head = node->first_child; - - if (head) - { - xml_node_struct* tail = head->prev_sibling_c; - - tail->next_sibling = child; - child->prev_sibling_c = tail; - head->prev_sibling_c = child; - } - else - { - node->first_child = child; - child->prev_sibling_c = child; - } - } - - inline void prepend_node(xml_node_struct* child, xml_node_struct* node) - { - child->parent = node; - - xml_node_struct* head = node->first_child; - - if (head) - { - child->prev_sibling_c = head->prev_sibling_c; - head->prev_sibling_c = child; - } - else - child->prev_sibling_c = child; - - child->next_sibling = head; node->first_child = child; + child->prev_sibling_c = child; } +} - inline void insert_node_after(xml_node_struct* child, xml_node_struct* node) +inline void prepend_node(xml_node_struct* child, xml_node_struct* node) +{ + child->parent = node; + + xml_node_struct* head = node->first_child; + + if (head) { - xml_node_struct* parent = node->parent; - - child->parent = parent; - - if (node->next_sibling) - node->next_sibling->prev_sibling_c = child; - else - parent->first_child->prev_sibling_c = child; - - child->next_sibling = node->next_sibling; - child->prev_sibling_c = node; - - node->next_sibling = child; + child->prev_sibling_c = head->prev_sibling_c; + head->prev_sibling_c = child; } + else + child->prev_sibling_c = child; - inline void insert_node_before(xml_node_struct* child, xml_node_struct* node) + child->next_sibling = head; + node->first_child = child; +} + +inline void insert_node_after(xml_node_struct* child, xml_node_struct* node) +{ + xml_node_struct* parent = node->parent; + + child->parent = parent; + + if (node->next_sibling) + node->next_sibling->prev_sibling_c = child; + else + parent->first_child->prev_sibling_c = child; + + child->next_sibling = node->next_sibling; + child->prev_sibling_c = node; + + node->next_sibling = child; +} + +inline void insert_node_before(xml_node_struct* child, xml_node_struct* node) +{ + xml_node_struct* parent = node->parent; + + child->parent = parent; + + if (node->prev_sibling_c->next_sibling) + node->prev_sibling_c->next_sibling = child; + else + parent->first_child = child; + + child->prev_sibling_c = node->prev_sibling_c; + child->next_sibling = node; + + node->prev_sibling_c = child; +} + +inline void remove_node(xml_node_struct* node) +{ + xml_node_struct* parent = node->parent; + + if (node->next_sibling) + node->next_sibling->prev_sibling_c = node->prev_sibling_c; + else + parent->first_child->prev_sibling_c = node->prev_sibling_c; + + if (node->prev_sibling_c->next_sibling) + node->prev_sibling_c->next_sibling = node->next_sibling; + else + parent->first_child = node->next_sibling; + + node->parent = 0; + node->prev_sibling_c = 0; + node->next_sibling = 0; +} + +inline void append_attribute(xml_attribute_struct* attr, xml_node_struct* node) +{ + xml_attribute_struct* head = node->first_attribute; + + if (head) { - xml_node_struct* parent = node->parent; + xml_attribute_struct* tail = head->prev_attribute_c; - child->parent = parent; - - if (node->prev_sibling_c->next_sibling) - node->prev_sibling_c->next_sibling = child; - else - parent->first_child = child; - - child->prev_sibling_c = node->prev_sibling_c; - child->next_sibling = node; - - node->prev_sibling_c = child; + tail->next_attribute = attr; + attr->prev_attribute_c = tail; + head->prev_attribute_c = attr; } - - inline void remove_node(xml_node_struct* node) + else { - xml_node_struct* parent = node->parent; - - if (node->next_sibling) - node->next_sibling->prev_sibling_c = node->prev_sibling_c; - else - parent->first_child->prev_sibling_c = node->prev_sibling_c; - - if (node->prev_sibling_c->next_sibling) - node->prev_sibling_c->next_sibling = node->next_sibling; - else - parent->first_child = node->next_sibling; - - node->parent = 0; - node->prev_sibling_c = 0; - node->next_sibling = 0; - } - - inline void append_attribute(xml_attribute_struct* attr, xml_node_struct* node) - { - xml_attribute_struct* head = node->first_attribute; - - if (head) - { - xml_attribute_struct* tail = head->prev_attribute_c; - - tail->next_attribute = attr; - attr->prev_attribute_c = tail; - head->prev_attribute_c = attr; - } - else - { - node->first_attribute = attr; - attr->prev_attribute_c = attr; - } - } - - inline void prepend_attribute(xml_attribute_struct* attr, xml_node_struct* node) - { - xml_attribute_struct* head = node->first_attribute; - - if (head) - { - attr->prev_attribute_c = head->prev_attribute_c; - head->prev_attribute_c = attr; - } - else - attr->prev_attribute_c = attr; - - attr->next_attribute = head; node->first_attribute = attr; + attr->prev_attribute_c = attr; } +} - inline void insert_attribute_after(xml_attribute_struct* attr, xml_attribute_struct* place, xml_node_struct* node) +inline void prepend_attribute(xml_attribute_struct* attr, xml_node_struct* node) +{ + xml_attribute_struct* head = node->first_attribute; + + if (head) { - if (place->next_attribute) - place->next_attribute->prev_attribute_c = attr; - else - node->first_attribute->prev_attribute_c = attr; - - attr->next_attribute = place->next_attribute; - attr->prev_attribute_c = place; - place->next_attribute = attr; + attr->prev_attribute_c = head->prev_attribute_c; + head->prev_attribute_c = attr; } + else + attr->prev_attribute_c = attr; - inline void insert_attribute_before(xml_attribute_struct* attr, xml_attribute_struct* place, xml_node_struct* node) - { - if (place->prev_attribute_c->next_attribute) - place->prev_attribute_c->next_attribute = attr; - else - node->first_attribute = attr; + attr->next_attribute = head; + node->first_attribute = attr; +} - attr->prev_attribute_c = place->prev_attribute_c; - attr->next_attribute = place; - place->prev_attribute_c = attr; - } +inline void insert_attribute_after(xml_attribute_struct* attr, xml_attribute_struct* place, xml_node_struct* node) +{ + if (place->next_attribute) + place->next_attribute->prev_attribute_c = attr; + else + node->first_attribute->prev_attribute_c = attr; - inline void remove_attribute(xml_attribute_struct* attr, xml_node_struct* node) - { - if (attr->next_attribute) - attr->next_attribute->prev_attribute_c = attr->prev_attribute_c; - else - node->first_attribute->prev_attribute_c = attr->prev_attribute_c; + attr->next_attribute = place->next_attribute; + attr->prev_attribute_c = place; + place->next_attribute = attr; +} - if (attr->prev_attribute_c->next_attribute) - attr->prev_attribute_c->next_attribute = attr->next_attribute; - else - node->first_attribute = attr->next_attribute; +inline void insert_attribute_before(xml_attribute_struct* attr, xml_attribute_struct* place, xml_node_struct* node) +{ + if (place->prev_attribute_c->next_attribute) + place->prev_attribute_c->next_attribute = attr; + else + node->first_attribute = attr; - attr->prev_attribute_c = 0; - attr->next_attribute = 0; - } + attr->prev_attribute_c = place->prev_attribute_c; + attr->next_attribute = place; + place->prev_attribute_c = attr; +} - PUGI__FN_NO_INLINE xml_node_struct* append_new_node(xml_node_struct* node, xml_allocator& alloc, xml_node_type type = node_element) - { - if (!alloc.reserve()) return 0; +inline void remove_attribute(xml_attribute_struct* attr, xml_node_struct* node) +{ + if (attr->next_attribute) + attr->next_attribute->prev_attribute_c = attr->prev_attribute_c; + else + node->first_attribute->prev_attribute_c = attr->prev_attribute_c; - xml_node_struct* child = allocate_node(alloc, type); - if (!child) return 0; + if (attr->prev_attribute_c->next_attribute) + attr->prev_attribute_c->next_attribute = attr->next_attribute; + else + node->first_attribute = attr->next_attribute; - append_node(child, node); + attr->prev_attribute_c = 0; + attr->next_attribute = 0; +} - return child; - } +PUGI__FN_NO_INLINE xml_node_struct* append_new_node(xml_node_struct* node, xml_allocator& alloc, xml_node_type type = node_element) +{ + if (!alloc.reserve()) + return 0; - PUGI__FN_NO_INLINE xml_attribute_struct* append_new_attribute(xml_node_struct* node, xml_allocator& alloc) - { - if (!alloc.reserve()) return 0; + xml_node_struct* child = allocate_node(alloc, type); + if (!child) + return 0; - xml_attribute_struct* attr = allocate_attribute(alloc); - if (!attr) return 0; + append_node(child, node); - append_attribute(attr, node); + return child; +} - return attr; - } +PUGI__FN_NO_INLINE xml_attribute_struct* append_new_attribute(xml_node_struct* node, xml_allocator& alloc) +{ + if (!alloc.reserve()) + return 0; + + xml_attribute_struct* attr = allocate_attribute(alloc); + if (!attr) + return 0; + + append_attribute(attr, node); + + return attr; +} PUGI__NS_END // Helper classes for code generation PUGI__NS_BEGIN - struct opt_false +struct opt_false +{ + enum { - enum { value = 0 }; + value = 0 }; +}; - struct opt_true +struct opt_true +{ + enum { - enum { value = 1 }; + value = 1 }; +}; PUGI__NS_END // Unicode utilities PUGI__NS_BEGIN - inline uint16_t endian_swap(uint16_t value) +inline uint16_t endian_swap(uint16_t value) +{ + return static_cast(((value & 0xff) << 8) | (value >> 8)); +} + +inline uint32_t endian_swap(uint32_t value) +{ + return ((value & 0xff) << 24) | ((value & 0xff00) << 8) | ((value & 0xff0000) >> 8) | (value >> 24); +} + +struct utf8_counter +{ + typedef size_t value_type; + + static value_type low(value_type result, uint32_t ch) { - return static_cast(((value & 0xff) << 8) | (value >> 8)); + // U+0000..U+007F + if (ch < 0x80) + return result + 1; + // U+0080..U+07FF + else if (ch < 0x800) + return result + 2; + // U+0800..U+FFFF + else + return result + 3; } - inline uint32_t endian_swap(uint32_t value) + static value_type high(value_type result, uint32_t) { - return ((value & 0xff) << 24) | ((value & 0xff00) << 8) | ((value & 0xff0000) >> 8) | (value >> 24); + // U+10000..U+10FFFF + return result + 4; + } +}; + +struct utf8_writer +{ + typedef uint8_t* value_type; + + static value_type low(value_type result, uint32_t ch) + { + // U+0000..U+007F + if (ch < 0x80) + { + *result = static_cast(ch); + return result + 1; + } + // U+0080..U+07FF + else if (ch < 0x800) + { + result[0] = static_cast(0xC0 | (ch >> 6)); + result[1] = static_cast(0x80 | (ch & 0x3F)); + return result + 2; + } + // U+0800..U+FFFF + else + { + result[0] = static_cast(0xE0 | (ch >> 12)); + result[1] = static_cast(0x80 | ((ch >> 6) & 0x3F)); + result[2] = static_cast(0x80 | (ch & 0x3F)); + return result + 3; + } } - struct utf8_counter + static value_type high(value_type result, uint32_t ch) { - typedef size_t value_type; + // U+10000..U+10FFFF + result[0] = static_cast(0xF0 | (ch >> 18)); + result[1] = static_cast(0x80 | ((ch >> 12) & 0x3F)); + result[2] = static_cast(0x80 | ((ch >> 6) & 0x3F)); + result[3] = static_cast(0x80 | (ch & 0x3F)); + return result + 4; + } - static value_type low(value_type result, uint32_t ch) - { - // U+0000..U+007F - if (ch < 0x80) return result + 1; - // U+0080..U+07FF - else if (ch < 0x800) return result + 2; - // U+0800..U+FFFF - else return result + 3; - } - - static value_type high(value_type result, uint32_t) - { - // U+10000..U+10FFFF - return result + 4; - } - }; - - struct utf8_writer + static value_type any(value_type result, uint32_t ch) { - typedef uint8_t* value_type; + return (ch < 0x10000) ? low(result, ch) : high(result, ch); + } +}; - static value_type low(value_type result, uint32_t ch) +struct utf16_counter +{ + typedef size_t value_type; + + static value_type low(value_type result, uint32_t) + { + return result + 1; + } + + static value_type high(value_type result, uint32_t) + { + return result + 2; + } +}; + +struct utf16_writer +{ + typedef uint16_t* value_type; + + static value_type low(value_type result, uint32_t ch) + { + *result = static_cast(ch); + + return result + 1; + } + + static value_type high(value_type result, uint32_t ch) + { + uint32_t msh = static_cast(ch - 0x10000) >> 10; + uint32_t lsh = static_cast(ch - 0x10000) & 0x3ff; + + result[0] = static_cast(0xD800 + msh); + result[1] = static_cast(0xDC00 + lsh); + + return result + 2; + } + + static value_type any(value_type result, uint32_t ch) + { + return (ch < 0x10000) ? low(result, ch) : high(result, ch); + } +}; + +struct utf32_counter +{ + typedef size_t value_type; + + static value_type low(value_type result, uint32_t) + { + return result + 1; + } + + static value_type high(value_type result, uint32_t) + { + return result + 1; + } +}; + +struct utf32_writer +{ + typedef uint32_t* value_type; + + static value_type low(value_type result, uint32_t ch) + { + *result = ch; + + return result + 1; + } + + static value_type high(value_type result, uint32_t ch) + { + *result = ch; + + return result + 1; + } + + static value_type any(value_type result, uint32_t ch) + { + *result = ch; + + return result + 1; + } +}; + +struct latin1_writer +{ + typedef uint8_t* value_type; + + static value_type low(value_type result, uint32_t ch) + { + *result = static_cast(ch > 255 ? '?' : ch); + + return result + 1; + } + + static value_type high(value_type result, uint32_t ch) + { + (void)ch; + + *result = '?'; + + return result + 1; + } +}; + +struct utf8_decoder +{ + typedef uint8_t type; + + template + static inline typename Traits::value_type process(const uint8_t* data, size_t size, typename Traits::value_type result, Traits) + { + const uint8_t utf8_byte_mask = 0x3f; + + while (size) { - // U+0000..U+007F - if (ch < 0x80) + uint8_t lead = *data; + + // 0xxxxxxx -> U+0000..U+007F + if (lead < 0x80) { - *result = static_cast(ch); - return result + 1; + result = Traits::low(result, lead); + data += 1; + size -= 1; + + // process aligned single-byte (ascii) blocks + if ((reinterpret_cast(data) & 3) == 0) + { + // round-trip through void* to silence 'cast increases required alignment of target type' warnings + while (size >= 4 && (*static_cast(static_cast(data)) & 0x80808080) == 0) + { + result = Traits::low(result, data[0]); + result = Traits::low(result, data[1]); + result = Traits::low(result, data[2]); + result = Traits::low(result, data[3]); + data += 4; + size -= 4; + } + } } - // U+0080..U+07FF - else if (ch < 0x800) + // 110xxxxx -> U+0080..U+07FF + else if (static_cast(lead - 0xC0) < 0x20 && size >= 2 && (data[1] & 0xc0) == 0x80) { - result[0] = static_cast(0xC0 | (ch >> 6)); - result[1] = static_cast(0x80 | (ch & 0x3F)); - return result + 2; + result = Traits::low(result, ((lead & ~0xC0) << 6) | (data[1] & utf8_byte_mask)); + data += 2; + size -= 2; } - // U+0800..U+FFFF + // 1110xxxx -> U+0800-U+FFFF + else if (static_cast(lead - 0xE0) < 0x10 && size >= 3 && (data[1] & 0xc0) == 0x80 && (data[2] & 0xc0) == 0x80) + { + result = Traits::low(result, ((lead & ~0xE0) << 12) | ((data[1] & utf8_byte_mask) << 6) | (data[2] & utf8_byte_mask)); + data += 3; + size -= 3; + } + // 11110xxx -> U+10000..U+10FFFF + else if (static_cast(lead - 0xF0) < 0x08 && size >= 4 && (data[1] & 0xc0) == 0x80 && (data[2] & 0xc0) == 0x80 && (data[3] & 0xc0) == 0x80) + { + result = Traits::high(result, ((lead & ~0xF0) << 18) | ((data[1] & utf8_byte_mask) << 12) | ((data[2] & utf8_byte_mask) << 6) | (data[3] & utf8_byte_mask)); + data += 4; + size -= 4; + } + // 10xxxxxx or 11111xxx -> invalid else { - result[0] = static_cast(0xE0 | (ch >> 12)); - result[1] = static_cast(0x80 | ((ch >> 6) & 0x3F)); - result[2] = static_cast(0x80 | (ch & 0x3F)); - return result + 3; - } - } - - static value_type high(value_type result, uint32_t ch) - { - // U+10000..U+10FFFF - result[0] = static_cast(0xF0 | (ch >> 18)); - result[1] = static_cast(0x80 | ((ch >> 12) & 0x3F)); - result[2] = static_cast(0x80 | ((ch >> 6) & 0x3F)); - result[3] = static_cast(0x80 | (ch & 0x3F)); - return result + 4; - } - - static value_type any(value_type result, uint32_t ch) - { - return (ch < 0x10000) ? low(result, ch) : high(result, ch); - } - }; - - struct utf16_counter - { - typedef size_t value_type; - - static value_type low(value_type result, uint32_t) - { - return result + 1; - } - - static value_type high(value_type result, uint32_t) - { - return result + 2; - } - }; - - struct utf16_writer - { - typedef uint16_t* value_type; - - static value_type low(value_type result, uint32_t ch) - { - *result = static_cast(ch); - - return result + 1; - } - - static value_type high(value_type result, uint32_t ch) - { - uint32_t msh = static_cast(ch - 0x10000) >> 10; - uint32_t lsh = static_cast(ch - 0x10000) & 0x3ff; - - result[0] = static_cast(0xD800 + msh); - result[1] = static_cast(0xDC00 + lsh); - - return result + 2; - } - - static value_type any(value_type result, uint32_t ch) - { - return (ch < 0x10000) ? low(result, ch) : high(result, ch); - } - }; - - struct utf32_counter - { - typedef size_t value_type; - - static value_type low(value_type result, uint32_t) - { - return result + 1; - } - - static value_type high(value_type result, uint32_t) - { - return result + 1; - } - }; - - struct utf32_writer - { - typedef uint32_t* value_type; - - static value_type low(value_type result, uint32_t ch) - { - *result = ch; - - return result + 1; - } - - static value_type high(value_type result, uint32_t ch) - { - *result = ch; - - return result + 1; - } - - static value_type any(value_type result, uint32_t ch) - { - *result = ch; - - return result + 1; - } - }; - - struct latin1_writer - { - typedef uint8_t* value_type; - - static value_type low(value_type result, uint32_t ch) - { - *result = static_cast(ch > 255 ? '?' : ch); - - return result + 1; - } - - static value_type high(value_type result, uint32_t ch) - { - (void)ch; - - *result = '?'; - - return result + 1; - } - }; - - struct utf8_decoder - { - typedef uint8_t type; - - template static inline typename Traits::value_type process(const uint8_t* data, size_t size, typename Traits::value_type result, Traits) - { - const uint8_t utf8_byte_mask = 0x3f; - - while (size) - { - uint8_t lead = *data; - - // 0xxxxxxx -> U+0000..U+007F - if (lead < 0x80) - { - result = Traits::low(result, lead); - data += 1; - size -= 1; - - // process aligned single-byte (ascii) blocks - if ((reinterpret_cast(data) & 3) == 0) - { - // round-trip through void* to silence 'cast increases required alignment of target type' warnings - while (size >= 4 && (*static_cast(static_cast(data)) & 0x80808080) == 0) - { - result = Traits::low(result, data[0]); - result = Traits::low(result, data[1]); - result = Traits::low(result, data[2]); - result = Traits::low(result, data[3]); - data += 4; - size -= 4; - } - } - } - // 110xxxxx -> U+0080..U+07FF - else if (static_cast(lead - 0xC0) < 0x20 && size >= 2 && (data[1] & 0xc0) == 0x80) - { - result = Traits::low(result, ((lead & ~0xC0) << 6) | (data[1] & utf8_byte_mask)); - data += 2; - size -= 2; - } - // 1110xxxx -> U+0800-U+FFFF - else if (static_cast(lead - 0xE0) < 0x10 && size >= 3 && (data[1] & 0xc0) == 0x80 && (data[2] & 0xc0) == 0x80) - { - result = Traits::low(result, ((lead & ~0xE0) << 12) | ((data[1] & utf8_byte_mask) << 6) | (data[2] & utf8_byte_mask)); - data += 3; - size -= 3; - } - // 11110xxx -> U+10000..U+10FFFF - else if (static_cast(lead - 0xF0) < 0x08 && size >= 4 && (data[1] & 0xc0) == 0x80 && (data[2] & 0xc0) == 0x80 && (data[3] & 0xc0) == 0x80) - { - result = Traits::high(result, ((lead & ~0xF0) << 18) | ((data[1] & utf8_byte_mask) << 12) | ((data[2] & utf8_byte_mask) << 6) | (data[3] & utf8_byte_mask)); - data += 4; - size -= 4; - } - // 10xxxxxx or 11111xxx -> invalid - else - { - data += 1; - size -= 1; - } - } - - return result; - } - }; - - template struct utf16_decoder - { - typedef uint16_t type; - - template static inline typename Traits::value_type process(const uint16_t* data, size_t size, typename Traits::value_type result, Traits) - { - while (size) - { - uint16_t lead = opt_swap::value ? endian_swap(*data) : *data; - - // U+0000..U+D7FF - if (lead < 0xD800) - { - result = Traits::low(result, lead); - data += 1; - size -= 1; - } - // U+E000..U+FFFF - else if (static_cast(lead - 0xE000) < 0x2000) - { - result = Traits::low(result, lead); - data += 1; - size -= 1; - } - // surrogate pair lead - else if (static_cast(lead - 0xD800) < 0x400 && size >= 2) - { - uint16_t next = opt_swap::value ? endian_swap(data[1]) : data[1]; - - if (static_cast(next - 0xDC00) < 0x400) - { - result = Traits::high(result, 0x10000 + ((lead & 0x3ff) << 10) + (next & 0x3ff)); - data += 2; - size -= 2; - } - else - { - data += 1; - size -= 1; - } - } - else - { - data += 1; - size -= 1; - } - } - - return result; - } - }; - - template struct utf32_decoder - { - typedef uint32_t type; - - template static inline typename Traits::value_type process(const uint32_t* data, size_t size, typename Traits::value_type result, Traits) - { - while (size) - { - uint32_t lead = opt_swap::value ? endian_swap(*data) : *data; - - // U+0000..U+FFFF - if (lead < 0x10000) - { - result = Traits::low(result, lead); - data += 1; - size -= 1; - } - // U+10000..U+10FFFF - else - { - result = Traits::high(result, lead); - data += 1; - size -= 1; - } - } - - return result; - } - }; - - struct latin1_decoder - { - typedef uint8_t type; - - template static inline typename Traits::value_type process(const uint8_t* data, size_t size, typename Traits::value_type result, Traits) - { - while (size) - { - result = Traits::low(result, *data); data += 1; size -= 1; } - - return result; } - }; - template struct wchar_selector; + return result; + } +}; - template <> struct wchar_selector<2> +template +struct utf16_decoder +{ + typedef uint16_t type; + + template + static inline typename Traits::value_type process(const uint16_t* data, size_t size, typename Traits::value_type result, Traits) { - typedef uint16_t type; - typedef utf16_counter counter; - typedef utf16_writer writer; - typedef utf16_decoder decoder; - }; - - template <> struct wchar_selector<4> - { - typedef uint32_t type; - typedef utf32_counter counter; - typedef utf32_writer writer; - typedef utf32_decoder decoder; - }; - - typedef wchar_selector::counter wchar_counter; - typedef wchar_selector::writer wchar_writer; - - struct wchar_decoder - { - typedef wchar_t type; - - template static inline typename Traits::value_type process(const wchar_t* data, size_t size, typename Traits::value_type result, Traits traits) + while (size) { - typedef wchar_selector::decoder decoder; + uint16_t lead = opt_swap::value ? endian_swap(*data) : *data; - return decoder::process(reinterpret_cast(data), size, result, traits); + // U+0000..U+D7FF + if (lead < 0xD800) + { + result = Traits::low(result, lead); + data += 1; + size -= 1; + } + // U+E000..U+FFFF + else if (static_cast(lead - 0xE000) < 0x2000) + { + result = Traits::low(result, lead); + data += 1; + size -= 1; + } + // surrogate pair lead + else if (static_cast(lead - 0xD800) < 0x400 && size >= 2) + { + uint16_t next = opt_swap::value ? endian_swap(data[1]) : data[1]; + + if (static_cast(next - 0xDC00) < 0x400) + { + result = Traits::high(result, 0x10000 + ((lead & 0x3ff) << 10) + (next & 0x3ff)); + data += 2; + size -= 2; + } + else + { + data += 1; + size -= 1; + } + } + else + { + data += 1; + size -= 1; + } } - }; + + return result; + } +}; + +template +struct utf32_decoder +{ + typedef uint32_t type; + + template + static inline typename Traits::value_type process(const uint32_t* data, size_t size, typename Traits::value_type result, Traits) + { + while (size) + { + uint32_t lead = opt_swap::value ? endian_swap(*data) : *data; + + // U+0000..U+FFFF + if (lead < 0x10000) + { + result = Traits::low(result, lead); + data += 1; + size -= 1; + } + // U+10000..U+10FFFF + else + { + result = Traits::high(result, lead); + data += 1; + size -= 1; + } + } + + return result; + } +}; + +struct latin1_decoder +{ + typedef uint8_t type; + + template + static inline typename Traits::value_type process(const uint8_t* data, size_t size, typename Traits::value_type result, Traits) + { + while (size) + { + result = Traits::low(result, *data); + data += 1; + size -= 1; + } + + return result; + } +}; + +template +struct wchar_selector; + +template <> +struct wchar_selector<2> +{ + typedef uint16_t type; + typedef utf16_counter counter; + typedef utf16_writer writer; + typedef utf16_decoder decoder; +}; + +template <> +struct wchar_selector<4> +{ + typedef uint32_t type; + typedef utf32_counter counter; + typedef utf32_writer writer; + typedef utf32_decoder decoder; +}; + +typedef wchar_selector::counter wchar_counter; +typedef wchar_selector::writer wchar_writer; + +struct wchar_decoder +{ + typedef wchar_t type; + + template + static inline typename Traits::value_type process(const wchar_t* data, size_t size, typename Traits::value_type result, Traits traits) + { + typedef wchar_selector::decoder decoder; + + return decoder::process(reinterpret_cast(data), size, result, traits); + } +}; #ifdef PUGIXML_WCHAR_MODE - PUGI__FN void convert_wchar_endian_swap(wchar_t* result, const wchar_t* data, size_t length) - { - for (size_t i = 0; i < length; ++i) - result[i] = static_cast(endian_swap(static_cast::type>(data[i]))); - } +PUGI__FN void convert_wchar_endian_swap(wchar_t* result, const wchar_t* data, size_t length) +{ + for (size_t i = 0; i < length; ++i) + result[i] = static_cast(endian_swap(static_cast::type>(data[i]))); +} #endif PUGI__NS_END PUGI__NS_BEGIN - enum chartype_t - { - ct_parse_pcdata = 1, // \0, &, \r, < - ct_parse_attr = 2, // \0, &, \r, ', " - ct_parse_attr_ws = 4, // \0, &, \r, ', ", \n, tab - ct_space = 8, // \r, \n, space, tab - ct_parse_cdata = 16, // \0, ], >, \r - ct_parse_comment = 32, // \0, -, >, \r - ct_symbol = 64, // Any symbol > 127, a-z, A-Z, 0-9, _, :, -, . - ct_start_symbol = 128 // Any symbol > 127, a-z, A-Z, _, : - }; +enum chartype_t +{ + ct_parse_pcdata = 1, // \0, &, \r, < + ct_parse_attr = 2, // \0, &, \r, ', " + ct_parse_attr_ws = 4, // \0, &, \r, ', ", \n, tab + ct_space = 8, // \r, \n, space, tab + ct_parse_cdata = 16, // \0, ], >, \r + ct_parse_comment = 32, // \0, -, >, \r + ct_symbol = 64, // Any symbol > 127, a-z, A-Z, 0-9, _, :, -, . + ct_start_symbol = 128 // Any symbol > 127, a-z, A-Z, _, : +}; - static const unsigned char chartype_table[256] = - { - 55, 0, 0, 0, 0, 0, 0, 0, 0, 12, 12, 0, 0, 63, 0, 0, // 0-15 - 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, // 16-31 - 8, 0, 6, 0, 0, 0, 7, 6, 0, 0, 0, 0, 0, 96, 64, 0, // 32-47 - 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 192, 0, 1, 0, 48, 0, // 48-63 - 0, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, // 64-79 - 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 0, 0, 16, 0, 192, // 80-95 - 0, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, // 96-111 - 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 0, 0, 0, 0, 0, // 112-127 +static const unsigned char chartype_table[256] = + { + 55, 0, 0, 0, 0, 0, 0, 0, 0, 12, 12, 0, 0, 63, 0, 0, // 0-15 + 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, // 16-31 + 8, 0, 6, 0, 0, 0, 7, 6, 0, 0, 0, 0, 0, 96, 64, 0, // 32-47 + 64, 64, 64, 64, 64, 64, 64, 64, 64, 64, 192, 0, 1, 0, 48, 0, // 48-63 + 0, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, // 64-79 + 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 0, 0, 16, 0, 192, // 80-95 + 0, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, // 96-111 + 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 0, 0, 0, 0, 0, // 112-127 - 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, // 128+ - 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, - 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, - 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, - 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, - 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, - 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, - 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192 - }; + 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, // 128+ + 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, + 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, + 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, + 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, + 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, + 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, + 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192, 192}; - enum chartypex_t - { - ctx_special_pcdata = 1, // Any symbol >= 0 and < 32 (except \t, \r, \n), &, <, > - ctx_special_attr = 2, // Any symbol >= 0 and < 32 (except \t), &, <, >, " - ctx_start_symbol = 4, // Any symbol > 127, a-z, A-Z, _ - ctx_digit = 8, // 0-9 - ctx_symbol = 16 // Any symbol > 127, a-z, A-Z, 0-9, _, -, . - }; +enum chartypex_t +{ + ctx_special_pcdata = 1, // Any symbol >= 0 and < 32 (except \t, \r, \n), &, <, > + ctx_special_attr = 2, // Any symbol >= 0 and < 32 (except \t), &, <, >, " + ctx_start_symbol = 4, // Any symbol > 127, a-z, A-Z, _ + ctx_digit = 8, // 0-9 + ctx_symbol = 16 // Any symbol > 127, a-z, A-Z, 0-9, _, -, . +}; - static const unsigned char chartypex_table[256] = - { - 3, 3, 3, 3, 3, 3, 3, 3, 3, 0, 2, 3, 3, 2, 3, 3, // 0-15 - 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, // 16-31 - 0, 0, 2, 0, 0, 0, 3, 0, 0, 0, 0, 0, 0, 16, 16, 0, // 32-47 - 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 0, 0, 3, 0, 3, 0, // 48-63 +static const unsigned char chartypex_table[256] = + { + 3, 3, 3, 3, 3, 3, 3, 3, 3, 0, 2, 3, 3, 2, 3, 3, // 0-15 + 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, // 16-31 + 0, 0, 2, 0, 0, 0, 3, 0, 0, 0, 0, 0, 0, 16, 16, 0, // 32-47 + 24, 24, 24, 24, 24, 24, 24, 24, 24, 24, 0, 0, 3, 0, 3, 0, // 48-63 - 0, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, // 64-79 - 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 0, 0, 0, 0, 20, // 80-95 - 0, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, // 96-111 - 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 0, 0, 0, 0, 0, // 112-127 + 0, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, // 64-79 + 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 0, 0, 0, 0, 20, // 80-95 + 0, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, // 96-111 + 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 0, 0, 0, 0, 0, // 112-127 - 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, // 128+ - 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, - 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, - 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, - 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, - 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, - 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, - 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20 - }; + 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, // 128+ + 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, + 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, + 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, + 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, + 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, + 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, + 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20, 20}; #ifdef PUGIXML_WCHAR_MODE - #define PUGI__IS_CHARTYPE_IMPL(c, ct, table) ((static_cast(c) < 128 ? table[static_cast(c)] : table[128]) & (ct)) +#define PUGI__IS_CHARTYPE_IMPL(c, ct, table) ((static_cast(c) < 128 ? table[static_cast(c)] : table[128]) & (ct)) #else - #define PUGI__IS_CHARTYPE_IMPL(c, ct, table) (table[static_cast(c)] & (ct)) +#define PUGI__IS_CHARTYPE_IMPL(c, ct, table) (table[static_cast(c)] & (ct)) #endif - #define PUGI__IS_CHARTYPE(c, ct) PUGI__IS_CHARTYPE_IMPL(c, ct, chartype_table) - #define PUGI__IS_CHARTYPEX(c, ct) PUGI__IS_CHARTYPE_IMPL(c, ct, chartypex_table) +#define PUGI__IS_CHARTYPE(c, ct) PUGI__IS_CHARTYPE_IMPL(c, ct, chartype_table) +#define PUGI__IS_CHARTYPEX(c, ct) PUGI__IS_CHARTYPE_IMPL(c, ct, chartypex_table) - PUGI__FN bool is_little_endian() - { - unsigned int ui = 1; +PUGI__FN bool is_little_endian() +{ + unsigned int ui = 1; - return *reinterpret_cast(&ui) == 1; + return *reinterpret_cast(&ui) == 1; +} + +PUGI__FN xml_encoding get_wchar_encoding() +{ + PUGI__STATIC_ASSERT(sizeof(wchar_t) == 2 || sizeof(wchar_t) == 4); + + if (sizeof(wchar_t) == 2) + return is_little_endian() ? encoding_utf16_le : encoding_utf16_be; + else + return is_little_endian() ? encoding_utf32_le : encoding_utf32_be; +} + +PUGI__FN bool parse_declaration_encoding(const uint8_t* data, size_t size, const uint8_t*& out_encoding, size_t& out_length) +{ +#define PUGI__SCANCHAR(ch) \ + { \ + if (offset >= size || data[offset] != ch) \ + return false; \ + offset++; \ + } +#define PUGI__SCANCHARTYPE(ct) \ + { \ + while (offset < size && PUGI__IS_CHARTYPE(data[offset], ct)) \ + offset++; \ } - PUGI__FN xml_encoding get_wchar_encoding() + // check if we have a non-empty XML declaration + if (size < 6 || !((data[0] == '<') & (data[1] == '?') & (data[2] == 'x') & (data[3] == 'm') & (data[4] == 'l') && PUGI__IS_CHARTYPE(data[5], ct_space))) + return false; + + // scan XML declaration until the encoding field + for (size_t i = 6; i + 1 < size; ++i) { - PUGI__STATIC_ASSERT(sizeof(wchar_t) == 2 || sizeof(wchar_t) == 4); - - if (sizeof(wchar_t) == 2) - return is_little_endian() ? encoding_utf16_le : encoding_utf16_be; - else - return is_little_endian() ? encoding_utf32_le : encoding_utf32_be; - } - - PUGI__FN bool parse_declaration_encoding(const uint8_t* data, size_t size, const uint8_t*& out_encoding, size_t& out_length) - { - #define PUGI__SCANCHAR(ch) { if (offset >= size || data[offset] != ch) return false; offset++; } - #define PUGI__SCANCHARTYPE(ct) { while (offset < size && PUGI__IS_CHARTYPE(data[offset], ct)) offset++; } - - // check if we have a non-empty XML declaration - if (size < 6 || !((data[0] == '<') & (data[1] == '?') & (data[2] == 'x') & (data[3] == 'm') & (data[4] == 'l') && PUGI__IS_CHARTYPE(data[5], ct_space))) + // declaration can not contain ? in quoted values + if (data[i] == '?') return false; - // scan XML declaration until the encoding field - for (size_t i = 6; i + 1 < size; ++i) + if (data[i] == 'e' && data[i + 1] == 'n') { - // declaration can not contain ? in quoted values - if (data[i] == '?') - return false; + size_t offset = i; - if (data[i] == 'e' && data[i + 1] == 'n') - { - size_t offset = i; + // encoding follows the version field which can't contain 'en' so this has to be the encoding if XML is well formed + PUGI__SCANCHAR('e'); + PUGI__SCANCHAR('n'); + PUGI__SCANCHAR('c'); + PUGI__SCANCHAR('o'); + PUGI__SCANCHAR('d'); + PUGI__SCANCHAR('i'); + PUGI__SCANCHAR('n'); + PUGI__SCANCHAR('g'); - // encoding follows the version field which can't contain 'en' so this has to be the encoding if XML is well formed - PUGI__SCANCHAR('e'); PUGI__SCANCHAR('n'); PUGI__SCANCHAR('c'); PUGI__SCANCHAR('o'); - PUGI__SCANCHAR('d'); PUGI__SCANCHAR('i'); PUGI__SCANCHAR('n'); PUGI__SCANCHAR('g'); + // S? = S? + PUGI__SCANCHARTYPE(ct_space); + PUGI__SCANCHAR('='); + PUGI__SCANCHARTYPE(ct_space); - // S? = S? - PUGI__SCANCHARTYPE(ct_space); - PUGI__SCANCHAR('='); - PUGI__SCANCHARTYPE(ct_space); + // the only two valid delimiters are ' and " + uint8_t delimiter = (offset < size && data[offset] == '"') ? '"' : '\''; - // the only two valid delimiters are ' and " - uint8_t delimiter = (offset < size && data[offset] == '"') ? '"' : '\''; + PUGI__SCANCHAR(delimiter); - PUGI__SCANCHAR(delimiter); + size_t start = offset; - size_t start = offset; + out_encoding = data + offset; - out_encoding = data + offset; + PUGI__SCANCHARTYPE(ct_symbol); - PUGI__SCANCHARTYPE(ct_symbol); + out_length = offset - start; - out_length = offset - start; + PUGI__SCANCHAR(delimiter); - PUGI__SCANCHAR(delimiter); - - return true; - } + return true; } - - return false; - - #undef PUGI__SCANCHAR - #undef PUGI__SCANCHARTYPE } - PUGI__FN xml_encoding guess_buffer_encoding(const uint8_t* data, size_t size) - { - // skip encoding autodetection if input buffer is too small - if (size < 4) return encoding_utf8; + return false; - uint8_t d0 = data[0], d1 = data[1], d2 = data[2], d3 = data[3]; - - // look for BOM in first few bytes - if (d0 == 0 && d1 == 0 && d2 == 0xfe && d3 == 0xff) return encoding_utf32_be; - if (d0 == 0xff && d1 == 0xfe && d2 == 0 && d3 == 0) return encoding_utf32_le; - if (d0 == 0xfe && d1 == 0xff) return encoding_utf16_be; - if (d0 == 0xff && d1 == 0xfe) return encoding_utf16_le; - if (d0 == 0xef && d1 == 0xbb && d2 == 0xbf) return encoding_utf8; - - // look for <, (contents); + + return guess_buffer_encoding(data, size); +} + +PUGI__FN bool get_mutable_buffer(char_t*& out_buffer, size_t& out_length, const void* contents, size_t size, bool is_mutable) +{ + size_t length = size / sizeof(char_t); + + if (is_mutable) { - // replace wchar encoding with utf implementation - if (encoding == encoding_wchar) return get_wchar_encoding(); - - // replace utf16 encoding with utf16 with specific endianness - if (encoding == encoding_utf16) return is_little_endian() ? encoding_utf16_le : encoding_utf16_be; - - // replace utf32 encoding with utf32 with specific endianness - if (encoding == encoding_utf32) return is_little_endian() ? encoding_utf32_le : encoding_utf32_be; - - // only do autodetection if no explicit encoding is requested - if (encoding != encoding_auto) return encoding; - - // try to guess encoding (based on XML specification, Appendix F.1) - const uint8_t* data = static_cast(contents); - - return guess_buffer_encoding(data, size); + out_buffer = static_cast(const_cast(contents)); + out_length = length; } - - PUGI__FN bool get_mutable_buffer(char_t*& out_buffer, size_t& out_length, const void* contents, size_t size, bool is_mutable) + else { - size_t length = size / sizeof(char_t); + char_t* buffer = static_cast(xml_memory::allocate((length + 1) * sizeof(char_t))); + if (!buffer) + return false; - if (is_mutable) - { - out_buffer = static_cast(const_cast(contents)); - out_length = length; - } + if (contents) + memcpy(buffer, contents, length * sizeof(char_t)); else - { - char_t* buffer = static_cast(xml_memory::allocate((length + 1) * sizeof(char_t))); - if (!buffer) return false; + assert(length == 0); - if (contents) - memcpy(buffer, contents, length * sizeof(char_t)); - else - assert(length == 0); + buffer[length] = 0; - buffer[length] = 0; - - out_buffer = buffer; - out_length = length + 1; - } - - return true; + out_buffer = buffer; + out_length = length + 1; } + return true; +} + #ifdef PUGIXML_WCHAR_MODE - PUGI__FN bool need_endian_swap_utf(xml_encoding le, xml_encoding re) +PUGI__FN bool need_endian_swap_utf(xml_encoding le, xml_encoding re) +{ + return (le == encoding_utf16_be && re == encoding_utf16_le) || (le == encoding_utf16_le && re == encoding_utf16_be) || + (le == encoding_utf32_be && re == encoding_utf32_le) || (le == encoding_utf32_le && re == encoding_utf32_be); +} + +PUGI__FN bool convert_buffer_endian_swap(char_t*& out_buffer, size_t& out_length, const void* contents, size_t size, bool is_mutable) +{ + const char_t* data = static_cast(contents); + size_t length = size / sizeof(char_t); + + if (is_mutable) { - return (le == encoding_utf16_be && re == encoding_utf16_le) || (le == encoding_utf16_le && re == encoding_utf16_be) || - (le == encoding_utf32_be && re == encoding_utf32_le) || (le == encoding_utf32_le && re == encoding_utf32_be); + char_t* buffer = const_cast(data); + + convert_wchar_endian_swap(buffer, data, length); + + out_buffer = buffer; + out_length = length; } - - PUGI__FN bool convert_buffer_endian_swap(char_t*& out_buffer, size_t& out_length, const void* contents, size_t size, bool is_mutable) + else { - const char_t* data = static_cast(contents); - size_t length = size / sizeof(char_t); - - if (is_mutable) - { - char_t* buffer = const_cast(data); - - convert_wchar_endian_swap(buffer, data, length); - - out_buffer = buffer; - out_length = length; - } - else - { - char_t* buffer = static_cast(xml_memory::allocate((length + 1) * sizeof(char_t))); - if (!buffer) return false; - - convert_wchar_endian_swap(buffer, data, length); - buffer[length] = 0; - - out_buffer = buffer; - out_length = length + 1; - } - - return true; - } - - template PUGI__FN bool convert_buffer_generic(char_t*& out_buffer, size_t& out_length, const void* contents, size_t size, D) - { - const typename D::type* data = static_cast(contents); - size_t data_length = size / sizeof(typename D::type); - - // first pass: get length in wchar_t units - size_t length = D::process(data, data_length, 0, wchar_counter()); - - // allocate buffer of suitable length char_t* buffer = static_cast(xml_memory::allocate((length + 1) * sizeof(char_t))); - if (!buffer) return false; + if (!buffer) + return false; - // second pass: convert utf16 input to wchar_t - wchar_writer::value_type obegin = reinterpret_cast(buffer); - wchar_writer::value_type oend = D::process(data, data_length, obegin, wchar_writer()); - - assert(oend == obegin + length); - *oend = 0; + convert_wchar_endian_swap(buffer, data, length); + buffer[length] = 0; out_buffer = buffer; out_length = length + 1; - - return true; } - PUGI__FN bool convert_buffer(char_t*& out_buffer, size_t& out_length, xml_encoding encoding, const void* contents, size_t size, bool is_mutable) - { - // get native encoding - xml_encoding wchar_encoding = get_wchar_encoding(); + return true; +} - // fast path: no conversion required - if (encoding == wchar_encoding) - return get_mutable_buffer(out_buffer, out_length, contents, size, is_mutable); +template +PUGI__FN bool convert_buffer_generic(char_t*& out_buffer, size_t& out_length, const void* contents, size_t size, D) +{ + const typename D::type* data = static_cast(contents); + size_t data_length = size / sizeof(typename D::type); - // only endian-swapping is required - if (need_endian_swap_utf(encoding, wchar_encoding)) - return convert_buffer_endian_swap(out_buffer, out_length, contents, size, is_mutable); + // first pass: get length in wchar_t units + size_t length = D::process(data, data_length, 0, wchar_counter()); - // source encoding is utf8 - if (encoding == encoding_utf8) - return convert_buffer_generic(out_buffer, out_length, contents, size, utf8_decoder()); - - // source encoding is utf16 - if (encoding == encoding_utf16_be || encoding == encoding_utf16_le) - { - xml_encoding native_encoding = is_little_endian() ? encoding_utf16_le : encoding_utf16_be; - - return (native_encoding == encoding) ? - convert_buffer_generic(out_buffer, out_length, contents, size, utf16_decoder()) : - convert_buffer_generic(out_buffer, out_length, contents, size, utf16_decoder()); - } - - // source encoding is utf32 - if (encoding == encoding_utf32_be || encoding == encoding_utf32_le) - { - xml_encoding native_encoding = is_little_endian() ? encoding_utf32_le : encoding_utf32_be; - - return (native_encoding == encoding) ? - convert_buffer_generic(out_buffer, out_length, contents, size, utf32_decoder()) : - convert_buffer_generic(out_buffer, out_length, contents, size, utf32_decoder()); - } - - // source encoding is latin1 - if (encoding == encoding_latin1) - return convert_buffer_generic(out_buffer, out_length, contents, size, latin1_decoder()); - - assert(false && "Invalid encoding"); // unreachable + // allocate buffer of suitable length + char_t* buffer = static_cast(xml_memory::allocate((length + 1) * sizeof(char_t))); + if (!buffer) return false; + + // second pass: convert utf16 input to wchar_t + wchar_writer::value_type obegin = reinterpret_cast(buffer); + wchar_writer::value_type oend = D::process(data, data_length, obegin, wchar_writer()); + + assert(oend == obegin + length); + *oend = 0; + + out_buffer = buffer; + out_length = length + 1; + + return true; +} + +PUGI__FN bool convert_buffer(char_t*& out_buffer, size_t& out_length, xml_encoding encoding, const void* contents, size_t size, bool is_mutable) +{ + // get native encoding + xml_encoding wchar_encoding = get_wchar_encoding(); + + // fast path: no conversion required + if (encoding == wchar_encoding) + return get_mutable_buffer(out_buffer, out_length, contents, size, is_mutable); + + // only endian-swapping is required + if (need_endian_swap_utf(encoding, wchar_encoding)) + return convert_buffer_endian_swap(out_buffer, out_length, contents, size, is_mutable); + + // source encoding is utf8 + if (encoding == encoding_utf8) + return convert_buffer_generic(out_buffer, out_length, contents, size, utf8_decoder()); + + // source encoding is utf16 + if (encoding == encoding_utf16_be || encoding == encoding_utf16_le) + { + xml_encoding native_encoding = is_little_endian() ? encoding_utf16_le : encoding_utf16_be; + + return (native_encoding == encoding) ? convert_buffer_generic(out_buffer, out_length, contents, size, utf16_decoder()) : convert_buffer_generic(out_buffer, out_length, contents, size, utf16_decoder()); } + + // source encoding is utf32 + if (encoding == encoding_utf32_be || encoding == encoding_utf32_le) + { + xml_encoding native_encoding = is_little_endian() ? encoding_utf32_le : encoding_utf32_be; + + return (native_encoding == encoding) ? convert_buffer_generic(out_buffer, out_length, contents, size, utf32_decoder()) : convert_buffer_generic(out_buffer, out_length, contents, size, utf32_decoder()); + } + + // source encoding is latin1 + if (encoding == encoding_latin1) + return convert_buffer_generic(out_buffer, out_length, contents, size, latin1_decoder()); + + assert(false && "Invalid encoding"); // unreachable + return false; +} #else - template PUGI__FN bool convert_buffer_generic(char_t*& out_buffer, size_t& out_length, const void* contents, size_t size, D) - { - const typename D::type* data = static_cast(contents); - size_t data_length = size / sizeof(typename D::type); +template +PUGI__FN bool convert_buffer_generic(char_t*& out_buffer, size_t& out_length, const void* contents, size_t size, D) +{ + const typename D::type* data = static_cast(contents); + size_t data_length = size / sizeof(typename D::type); - // first pass: get length in utf8 units - size_t length = D::process(data, data_length, 0, utf8_counter()); + // first pass: get length in utf8 units + size_t length = D::process(data, data_length, 0, utf8_counter()); - // allocate buffer of suitable length - char_t* buffer = static_cast(xml_memory::allocate((length + 1) * sizeof(char_t))); - if (!buffer) return false; - - // second pass: convert utf16 input to utf8 - uint8_t* obegin = reinterpret_cast(buffer); - uint8_t* oend = D::process(data, data_length, obegin, utf8_writer()); - - assert(oend == obegin + length); - *oend = 0; - - out_buffer = buffer; - out_length = length + 1; - - return true; - } - - PUGI__FN size_t get_latin1_7bit_prefix_length(const uint8_t* data, size_t size) - { - for (size_t i = 0; i < size; ++i) - if (data[i] > 127) - return i; - - return size; - } - - PUGI__FN bool convert_buffer_latin1(char_t*& out_buffer, size_t& out_length, const void* contents, size_t size, bool is_mutable) - { - const uint8_t* data = static_cast(contents); - size_t data_length = size; - - // get size of prefix that does not need utf8 conversion - size_t prefix_length = get_latin1_7bit_prefix_length(data, data_length); - assert(prefix_length <= data_length); - - const uint8_t* postfix = data + prefix_length; - size_t postfix_length = data_length - prefix_length; - - // if no conversion is needed, just return the original buffer - if (postfix_length == 0) return get_mutable_buffer(out_buffer, out_length, contents, size, is_mutable); - - // first pass: get length in utf8 units - size_t length = prefix_length + latin1_decoder::process(postfix, postfix_length, 0, utf8_counter()); - - // allocate buffer of suitable length - char_t* buffer = static_cast(xml_memory::allocate((length + 1) * sizeof(char_t))); - if (!buffer) return false; - - // second pass: convert latin1 input to utf8 - memcpy(buffer, data, prefix_length); - - uint8_t* obegin = reinterpret_cast(buffer); - uint8_t* oend = latin1_decoder::process(postfix, postfix_length, obegin + prefix_length, utf8_writer()); - - assert(oend == obegin + length); - *oend = 0; - - out_buffer = buffer; - out_length = length + 1; - - return true; - } - - PUGI__FN bool convert_buffer(char_t*& out_buffer, size_t& out_length, xml_encoding encoding, const void* contents, size_t size, bool is_mutable) - { - // fast path: no conversion required - if (encoding == encoding_utf8) - return get_mutable_buffer(out_buffer, out_length, contents, size, is_mutable); - - // source encoding is utf16 - if (encoding == encoding_utf16_be || encoding == encoding_utf16_le) - { - xml_encoding native_encoding = is_little_endian() ? encoding_utf16_le : encoding_utf16_be; - - return (native_encoding == encoding) ? - convert_buffer_generic(out_buffer, out_length, contents, size, utf16_decoder()) : - convert_buffer_generic(out_buffer, out_length, contents, size, utf16_decoder()); - } - - // source encoding is utf32 - if (encoding == encoding_utf32_be || encoding == encoding_utf32_le) - { - xml_encoding native_encoding = is_little_endian() ? encoding_utf32_le : encoding_utf32_be; - - return (native_encoding == encoding) ? - convert_buffer_generic(out_buffer, out_length, contents, size, utf32_decoder()) : - convert_buffer_generic(out_buffer, out_length, contents, size, utf32_decoder()); - } - - // source encoding is latin1 - if (encoding == encoding_latin1) - return convert_buffer_latin1(out_buffer, out_length, contents, size, is_mutable); - - assert(false && "Invalid encoding"); // unreachable + // allocate buffer of suitable length + char_t* buffer = static_cast(xml_memory::allocate((length + 1) * sizeof(char_t))); + if (!buffer) return false; + + // second pass: convert utf16 input to utf8 + uint8_t* obegin = reinterpret_cast(buffer); + uint8_t* oend = D::process(data, data_length, obegin, utf8_writer()); + + assert(oend == obegin + length); + *oend = 0; + + out_buffer = buffer; + out_length = length + 1; + + return true; +} + +PUGI__FN size_t get_latin1_7bit_prefix_length(const uint8_t* data, size_t size) +{ + for (size_t i = 0; i < size; ++i) + if (data[i] > 127) + return i; + + return size; +} + +PUGI__FN bool convert_buffer_latin1(char_t*& out_buffer, size_t& out_length, const void* contents, size_t size, bool is_mutable) +{ + const uint8_t* data = static_cast(contents); + size_t data_length = size; + + // get size of prefix that does not need utf8 conversion + size_t prefix_length = get_latin1_7bit_prefix_length(data, data_length); + assert(prefix_length <= data_length); + + const uint8_t* postfix = data + prefix_length; + size_t postfix_length = data_length - prefix_length; + + // if no conversion is needed, just return the original buffer + if (postfix_length == 0) + return get_mutable_buffer(out_buffer, out_length, contents, size, is_mutable); + + // first pass: get length in utf8 units + size_t length = prefix_length + latin1_decoder::process(postfix, postfix_length, 0, utf8_counter()); + + // allocate buffer of suitable length + char_t* buffer = static_cast(xml_memory::allocate((length + 1) * sizeof(char_t))); + if (!buffer) + return false; + + // second pass: convert latin1 input to utf8 + memcpy(buffer, data, prefix_length); + + uint8_t* obegin = reinterpret_cast(buffer); + uint8_t* oend = latin1_decoder::process(postfix, postfix_length, obegin + prefix_length, utf8_writer()); + + assert(oend == obegin + length); + *oend = 0; + + out_buffer = buffer; + out_length = length + 1; + + return true; +} + +PUGI__FN bool convert_buffer(char_t*& out_buffer, size_t& out_length, xml_encoding encoding, const void* contents, size_t size, bool is_mutable) +{ + // fast path: no conversion required + if (encoding == encoding_utf8) + return get_mutable_buffer(out_buffer, out_length, contents, size, is_mutable); + + // source encoding is utf16 + if (encoding == encoding_utf16_be || encoding == encoding_utf16_le) + { + xml_encoding native_encoding = is_little_endian() ? encoding_utf16_le : encoding_utf16_be; + + return (native_encoding == encoding) ? convert_buffer_generic(out_buffer, out_length, contents, size, utf16_decoder()) : convert_buffer_generic(out_buffer, out_length, contents, size, utf16_decoder()); } + + // source encoding is utf32 + if (encoding == encoding_utf32_be || encoding == encoding_utf32_le) + { + xml_encoding native_encoding = is_little_endian() ? encoding_utf32_le : encoding_utf32_be; + + return (native_encoding == encoding) ? convert_buffer_generic(out_buffer, out_length, contents, size, utf32_decoder()) : convert_buffer_generic(out_buffer, out_length, contents, size, utf32_decoder()); + } + + // source encoding is latin1 + if (encoding == encoding_latin1) + return convert_buffer_latin1(out_buffer, out_length, contents, size, is_mutable); + + assert(false && "Invalid encoding"); // unreachable + return false; +} #endif - PUGI__FN size_t as_utf8_begin(const wchar_t* str, size_t length) - { - // get length in utf8 characters - return wchar_decoder::process(str, length, 0, utf8_counter()); - } +PUGI__FN size_t as_utf8_begin(const wchar_t* str, size_t length) +{ + // get length in utf8 characters + return wchar_decoder::process(str, length, 0, utf8_counter()); +} - PUGI__FN void as_utf8_end(char* buffer, size_t size, const wchar_t* str, size_t length) - { - // convert to utf8 - uint8_t* begin = reinterpret_cast(buffer); - uint8_t* end = wchar_decoder::process(str, length, begin, utf8_writer()); +PUGI__FN void as_utf8_end(char* buffer, size_t size, const wchar_t* str, size_t length) +{ + // convert to utf8 + uint8_t* begin = reinterpret_cast(buffer); + uint8_t* end = wchar_decoder::process(str, length, begin, utf8_writer()); - assert(begin + size == end); - (void)!end; - (void)!size; - } + assert(begin + size == end); + (void)!end; + (void)!size; +} #ifndef PUGIXML_NO_STL - PUGI__FN std::string as_utf8_impl(const wchar_t* str, size_t length) +PUGI__FN std::string as_utf8_impl(const wchar_t* str, size_t length) +{ + // first pass: get length in utf8 characters + size_t size = as_utf8_begin(str, length); + + // allocate resulting string + std::string result; + result.resize(size); + + // second pass: convert to utf8 + if (size > 0) + as_utf8_end(&result[0], size, str, length); + + return result; +} + +PUGI__FN std::basic_string as_wide_impl(const char* str, size_t size) +{ + const uint8_t* data = reinterpret_cast(str); + + // first pass: get length in wchar_t units + size_t length = utf8_decoder::process(data, size, 0, wchar_counter()); + + // allocate resulting string + std::basic_string result; + result.resize(length); + + // second pass: convert to wchar_t + if (length > 0) { - // first pass: get length in utf8 characters - size_t size = as_utf8_begin(str, length); + wchar_writer::value_type begin = reinterpret_cast(&result[0]); + wchar_writer::value_type end = utf8_decoder::process(data, size, begin, wchar_writer()); - // allocate resulting string - std::string result; - result.resize(size); - - // second pass: convert to utf8 - if (size > 0) as_utf8_end(&result[0], size, str, length); - - return result; + assert(begin + length == end); + (void)!end; } - PUGI__FN std::basic_string as_wide_impl(const char* str, size_t size) - { - const uint8_t* data = reinterpret_cast(str); - - // first pass: get length in wchar_t units - size_t length = utf8_decoder::process(data, size, 0, wchar_counter()); - - // allocate resulting string - std::basic_string result; - result.resize(length); - - // second pass: convert to wchar_t - if (length > 0) - { - wchar_writer::value_type begin = reinterpret_cast(&result[0]); - wchar_writer::value_type end = utf8_decoder::process(data, size, begin, wchar_writer()); - - assert(begin + length == end); - (void)!end; - } - - return result; - } + return result; +} #endif - template - inline bool strcpy_insitu_allow(size_t length, const Header& header, uintptr_t header_mask, char_t* target) +template +inline bool strcpy_insitu_allow(size_t length, const Header& header, uintptr_t header_mask, char_t* target) +{ + // never reuse shared memory + if (header & xml_memory_page_contents_shared_mask) + return false; + + size_t target_length = strlength(target); + + // always reuse document buffer memory if possible + if ((header & header_mask) == 0) + return target_length >= length; + + // reuse heap memory if waste is not too great + const size_t reuse_threshold = 32; + + return target_length >= length && (target_length < reuse_threshold || target_length - length < target_length / 2); +} + +template +PUGI__FN bool strcpy_insitu(String& dest, Header& header, uintptr_t header_mask, const char_t* source, size_t source_length) +{ + if (source_length == 0) { - // never reuse shared memory - if (header & xml_memory_page_contents_shared_mask) return false; + // empty string and null pointer are equivalent, so just deallocate old memory + xml_allocator* alloc = PUGI__GETPAGE_IMPL(header)->allocator; - size_t target_length = strlength(target); + if (header & header_mask) + alloc->deallocate_string(dest); - // always reuse document buffer memory if possible - if ((header & header_mask) == 0) return target_length >= length; + // mark the string as not allocated + dest = 0; + header &= ~header_mask; - // reuse heap memory if waste is not too great - const size_t reuse_threshold = 32; + return true; + } + else if (dest && strcpy_insitu_allow(source_length, header, header_mask, dest)) + { + // we can reuse old buffer, so just copy the new data (including zero terminator) + memcpy(dest, source, source_length * sizeof(char_t)); + dest[source_length] = 0; - return target_length >= length && (target_length < reuse_threshold || target_length - length < target_length / 2); + return true; + } + else + { + xml_allocator* alloc = PUGI__GETPAGE_IMPL(header)->allocator; + + if (!alloc->reserve()) + return false; + + // allocate new buffer + char_t* buf = alloc->allocate_string(source_length + 1); + if (!buf) + return false; + + // copy the string (including zero terminator) + memcpy(buf, source, source_length * sizeof(char_t)); + buf[source_length] = 0; + + // deallocate old buffer (*after* the above to protect against overlapping memory and/or allocation failures) + if (header & header_mask) + alloc->deallocate_string(dest); + + // the string is now allocated, so set the flag + dest = buf; + header |= header_mask; + + return true; + } +} + +struct gap +{ + char_t* end; + size_t size; + + gap() + : end(0) + , size(0) + { } - template - PUGI__FN bool strcpy_insitu(String& dest, Header& header, uintptr_t header_mask, const char_t* source, size_t source_length) + // Push new gap, move s count bytes further (skipping the gap). + // Collapse previous gap. + void push(char_t*& s, size_t count) { - if (source_length == 0) + if (end) // there was a gap already; collapse it { - // empty string and null pointer are equivalent, so just deallocate old memory - xml_allocator* alloc = PUGI__GETPAGE_IMPL(header)->allocator; - - if (header & header_mask) alloc->deallocate_string(dest); - - // mark the string as not allocated - dest = 0; - header &= ~header_mask; - - return true; + // Move [old_gap_end, new_gap_start) to [old_gap_start, ...) + assert(s >= end); + memmove(end - size, end, reinterpret_cast(s) - reinterpret_cast(end)); } - else if (dest && strcpy_insitu_allow(source_length, header, header_mask, dest)) - { - // we can reuse old buffer, so just copy the new data (including zero terminator) - memcpy(dest, source, source_length * sizeof(char_t)); - dest[source_length] = 0; - return true; + s += count; // end of current gap + + // "merge" two gaps + end = s; + size += count; + } + + // Collapse all gaps, return past-the-end pointer + char_t* flush(char_t* s) + { + if (end) + { + // Move [old_gap_end, current_pos) to [old_gap_start, ...) + assert(s >= end); + memmove(end - size, end, reinterpret_cast(s) - reinterpret_cast(end)); + + return s - size; } else + return s; + } +}; + +PUGI__FN char_t* strconv_escape(char_t* s, gap& g) +{ + char_t* stre = s + 1; + + switch (*stre) + { + case '#': // &#... + { + unsigned int ucsc = 0; + + if (stre[1] == 'x') // &#x... (hex code) { - xml_allocator* alloc = PUGI__GETPAGE_IMPL(header)->allocator; + stre += 2; - if (!alloc->reserve()) return false; + char_t ch = *stre; - // allocate new buffer - char_t* buf = alloc->allocate_string(source_length + 1); - if (!buf) return false; + if (ch == ';') + return stre; - // copy the string (including zero terminator) - memcpy(buf, source, source_length * sizeof(char_t)); - buf[source_length] = 0; + for (;;) + { + if (static_cast(ch - '0') <= 9) + ucsc = 16 * ucsc + (ch - '0'); + else if (static_cast((ch | ' ') - 'a') <= 5) + ucsc = 16 * ucsc + ((ch | ' ') - 'a' + 10); + else if (ch == ';') + break; + else // cancel + return stre; - // deallocate old buffer (*after* the above to protect against overlapping memory and/or allocation failures) - if (header & header_mask) alloc->deallocate_string(dest); + ch = *++stre; + } - // the string is now allocated, so set the flag - dest = buf; - header |= header_mask; - - return true; + ++stre; } + else // &#... (dec code) + { + char_t ch = *++stre; + + if (ch == ';') + return stre; + + for (;;) + { + if (static_cast(ch - '0') <= 9) + ucsc = 10 * ucsc + (ch - '0'); + else if (ch == ';') + break; + else // cancel + return stre; + + ch = *++stre; + } + + ++stre; + } + +#ifdef PUGIXML_WCHAR_MODE + s = reinterpret_cast(wchar_writer::any(reinterpret_cast(s), ucsc)); +#else + s = reinterpret_cast(utf8_writer::any(reinterpret_cast(s), ucsc)); +#endif + + g.push(s, stre - s); + return stre; } - struct gap + case 'a': // &a { - char_t* end; - size_t size; + ++stre; - gap(): end(0), size(0) + if (*stre == 'm') // &am { - } - - // Push new gap, move s count bytes further (skipping the gap). - // Collapse previous gap. - void push(char_t*& s, size_t count) - { - if (end) // there was a gap already; collapse it + if (*++stre == 'p' && *++stre == ';') // & { - // Move [old_gap_end, new_gap_start) to [old_gap_start, ...) - assert(s >= end); - memmove(end - size, end, reinterpret_cast(s) - reinterpret_cast(end)); - } - - s += count; // end of current gap - - // "merge" two gaps - end = s; - size += count; - } - - // Collapse all gaps, return past-the-end pointer - char_t* flush(char_t* s) - { - if (end) - { - // Move [old_gap_end, current_pos) to [old_gap_start, ...) - assert(s >= end); - memmove(end - size, end, reinterpret_cast(s) - reinterpret_cast(end)); - - return s - size; - } - else return s; - } - }; - - PUGI__FN char_t* strconv_escape(char_t* s, gap& g) - { - char_t* stre = s + 1; - - switch (*stre) - { - case '#': // &#... - { - unsigned int ucsc = 0; - - if (stre[1] == 'x') // &#x... (hex code) - { - stre += 2; - - char_t ch = *stre; - - if (ch == ';') return stre; - - for (;;) - { - if (static_cast(ch - '0') <= 9) - ucsc = 16 * ucsc + (ch - '0'); - else if (static_cast((ch | ' ') - 'a') <= 5) - ucsc = 16 * ucsc + ((ch | ' ') - 'a' + 10); - else if (ch == ';') - break; - else // cancel - return stre; - - ch = *++stre; - } - - ++stre; - } - else // &#... (dec code) - { - char_t ch = *++stre; - - if (ch == ';') return stre; - - for (;;) - { - if (static_cast(ch - '0') <= 9) - ucsc = 10 * ucsc + (ch - '0'); - else if (ch == ';') - break; - else // cancel - return stre; - - ch = *++stre; - } - - ++stre; - } - - #ifdef PUGIXML_WCHAR_MODE - s = reinterpret_cast(wchar_writer::any(reinterpret_cast(s), ucsc)); - #else - s = reinterpret_cast(utf8_writer::any(reinterpret_cast(s), ucsc)); - #endif + *s++ = '&'; + ++stre; g.push(s, stre - s); return stre; } - - case 'a': // &a + } + else if (*stre == 'p') // &ap + { + if (*++stre == 'o' && *++stre == 's' && *++stre == ';') // ' { + *s++ = '\''; ++stre; - if (*stre == 'm') // &am - { - if (*++stre == 'p' && *++stre == ';') // & - { - *s++ = '&'; - ++stre; - - g.push(s, stre - s); - return stre; - } - } - else if (*stre == 'p') // &ap - { - if (*++stre == 'o' && *++stre == 's' && *++stre == ';') // ' - { - *s++ = '\''; - ++stre; - - g.push(s, stre - s); - return stre; - } - } - break; + g.push(s, stre - s); + return stre; } - - case 'g': // &g - { - if (*++stre == 't' && *++stre == ';') // > - { - *s++ = '>'; - ++stre; - - g.push(s, stre - s); - return stre; - } - break; - } - - case 'l': // &l - { - if (*++stre == 't' && *++stre == ';') // < - { - *s++ = '<'; - ++stre; - - g.push(s, stre - s); - return stre; - } - break; - } - - case 'q': // &q - { - if (*++stre == 'u' && *++stre == 'o' && *++stre == 't' && *++stre == ';') // " - { - *s++ = '"'; - ++stre; - - g.push(s, stre - s); - return stre; - } - break; - } - - default: - break; } - - return stre; + break; } - // Parser utilities - #define PUGI__ENDSWITH(c, e) ((c) == (e) || ((c) == 0 && endch == (e))) - #define PUGI__SKIPWS() { while (PUGI__IS_CHARTYPE(*s, ct_space)) ++s; } - #define PUGI__OPTSET(OPT) ( optmsk & (OPT) ) - #define PUGI__PUSHNODE(TYPE) { cursor = append_new_node(cursor, *alloc, TYPE); if (!cursor) PUGI__THROW_ERROR(status_out_of_memory, s); } - #define PUGI__POPNODE() { cursor = cursor->parent; } - #define PUGI__SCANFOR(X) { while (*s != 0 && !(X)) ++s; } - #define PUGI__SCANWHILE(X) { while (X) ++s; } - #define PUGI__SCANWHILE_UNROLL(X) { for (;;) { char_t ss = s[0]; if (PUGI__UNLIKELY(!(X))) { break; } ss = s[1]; if (PUGI__UNLIKELY(!(X))) { s += 1; break; } ss = s[2]; if (PUGI__UNLIKELY(!(X))) { s += 2; break; } ss = s[3]; if (PUGI__UNLIKELY(!(X))) { s += 3; break; } s += 4; } } - #define PUGI__ENDSEG() { ch = *s; *s = 0; ++s; } - #define PUGI__THROW_ERROR(err, m) return error_offset = m, error_status = err, static_cast(0) - #define PUGI__CHECK_ERROR(err, m) { if (*s == 0) PUGI__THROW_ERROR(err, m); } + case 'g': // &g + { + if (*++stre == 't' && *++stre == ';') // > + { + *s++ = '>'; + ++stre; - PUGI__FN char_t* strconv_comment(char_t* s, char_t endch) + g.push(s, stre - s); + return stre; + } + break; + } + + case 'l': // &l + { + if (*++stre == 't' && *++stre == ';') // < + { + *s++ = '<'; + ++stre; + + g.push(s, stre - s); + return stre; + } + break; + } + + case 'q': // &q + { + if (*++stre == 'u' && *++stre == 'o' && *++stre == 't' && *++stre == ';') // " + { + *s++ = '"'; + ++stre; + + g.push(s, stre - s); + return stre; + } + break; + } + + default: + break; + } + + return stre; +} + +// Parser utilities +#define PUGI__ENDSWITH(c, e) ((c) == (e) || ((c) == 0 && endch == (e))) +#define PUGI__SKIPWS() \ + { \ + while (PUGI__IS_CHARTYPE(*s, ct_space)) \ + ++s; \ + } +#define PUGI__OPTSET(OPT) (optmsk & (OPT)) +#define PUGI__PUSHNODE(TYPE) \ + { \ + cursor = append_new_node(cursor, *alloc, TYPE); \ + if (!cursor) \ + PUGI__THROW_ERROR(status_out_of_memory, s); \ + } +#define PUGI__POPNODE() \ + { \ + cursor = cursor->parent; \ + } +#define PUGI__SCANFOR(X) \ + { \ + while (*s != 0 && !(X)) \ + ++s; \ + } +#define PUGI__SCANWHILE(X) \ + { \ + while (X) \ + ++s; \ + } +#define PUGI__SCANWHILE_UNROLL(X) \ + { \ + for (;;) \ + { \ + char_t ss = s[0]; \ + if (PUGI__UNLIKELY(!(X))) \ + { \ + break; \ + } \ + ss = s[1]; \ + if (PUGI__UNLIKELY(!(X))) \ + { \ + s += 1; \ + break; \ + } \ + ss = s[2]; \ + if (PUGI__UNLIKELY(!(X))) \ + { \ + s += 2; \ + break; \ + } \ + ss = s[3]; \ + if (PUGI__UNLIKELY(!(X))) \ + { \ + s += 3; \ + break; \ + } \ + s += 4; \ + } \ + } +#define PUGI__ENDSEG() \ + { \ + ch = *s; \ + *s = 0; \ + ++s; \ + } +#define PUGI__THROW_ERROR(err, m) return error_offset = m, error_status = err, static_cast(0) +#define PUGI__CHECK_ERROR(err, m) \ + { \ + if (*s == 0) \ + PUGI__THROW_ERROR(err, m); \ + } + +PUGI__FN char_t* strconv_comment(char_t* s, char_t endch) +{ + gap g; + + while (true) + { + PUGI__SCANWHILE_UNROLL(!PUGI__IS_CHARTYPE(ss, ct_parse_comment)); + + if (*s == '\r') // Either a single 0x0d or 0x0d 0x0a pair + { + *s++ = '\n'; // replace first one with 0x0a + + if (*s == '\n') + g.push(s, 1); + } + else if (s[0] == '-' && s[1] == '-' && PUGI__ENDSWITH(s[2], '>')) // comment ends here + { + *g.flush(s) = 0; + + return s + (s[2] == '>' ? 3 : 2); + } + else if (*s == 0) + { + return 0; + } + else + ++s; + } +} + +PUGI__FN char_t* strconv_cdata(char_t* s, char_t endch) +{ + gap g; + + while (true) + { + PUGI__SCANWHILE_UNROLL(!PUGI__IS_CHARTYPE(ss, ct_parse_cdata)); + + if (*s == '\r') // Either a single 0x0d or 0x0d 0x0a pair + { + *s++ = '\n'; // replace first one with 0x0a + + if (*s == '\n') + g.push(s, 1); + } + else if (s[0] == ']' && s[1] == ']' && PUGI__ENDSWITH(s[2], '>')) // CDATA ends here + { + *g.flush(s) = 0; + + return s + 1; + } + else if (*s == 0) + { + return 0; + } + else + ++s; + } +} + +typedef char_t* (*strconv_pcdata_t)(char_t*); + +template +struct strconv_pcdata_impl +{ + static char_t* parse(char_t* s) { gap g; + char_t* begin = s; + while (true) { - PUGI__SCANWHILE_UNROLL(!PUGI__IS_CHARTYPE(ss, ct_parse_comment)); + PUGI__SCANWHILE_UNROLL(!PUGI__IS_CHARTYPE(ss, ct_parse_pcdata)); - if (*s == '\r') // Either a single 0x0d or 0x0d 0x0a pair + if (*s == '<') // PCDATA ends here + { + char_t* end = g.flush(s); + + if (opt_trim::value) + while (end > begin && PUGI__IS_CHARTYPE(end[-1], ct_space)) + --end; + + *end = 0; + + return s + 1; + } + else if (opt_eol::value && *s == '\r') // Either a single 0x0d or 0x0d 0x0a pair { *s++ = '\n'; // replace first one with 0x0a - if (*s == '\n') g.push(s, 1); + if (*s == '\n') + g.push(s, 1); } - else if (s[0] == '-' && s[1] == '-' && PUGI__ENDSWITH(s[2], '>')) // comment ends here + else if (opt_escape::value && *s == '&') { - *g.flush(s) = 0; - - return s + (s[2] == '>' ? 3 : 2); + s = strconv_escape(s, g); } else if (*s == 0) + { + char_t* end = g.flush(s); + + if (opt_trim::value) + while (end > begin && PUGI__IS_CHARTYPE(end[-1], ct_space)) + --end; + + *end = 0; + + return s; + } + else + ++s; + } + } +}; + +PUGI__FN strconv_pcdata_t get_strconv_pcdata(unsigned int optmask) +{ + PUGI__STATIC_ASSERT(parse_escapes == 0x10 && parse_eol == 0x20 && parse_trim_pcdata == 0x0800); + + switch (((optmask >> 4) & 3) | ((optmask >> 9) & 4)) // get bitmask for flags (trim eol escapes); this simultaneously checks 3 options from assertion above + { + case 0: + return strconv_pcdata_impl::parse; + case 1: + return strconv_pcdata_impl::parse; + case 2: + return strconv_pcdata_impl::parse; + case 3: + return strconv_pcdata_impl::parse; + case 4: + return strconv_pcdata_impl::parse; + case 5: + return strconv_pcdata_impl::parse; + case 6: + return strconv_pcdata_impl::parse; + case 7: + return strconv_pcdata_impl::parse; + default: + assert(false); + return 0; // unreachable + } +} + +typedef char_t* (*strconv_attribute_t)(char_t*, char_t); + +template +struct strconv_attribute_impl +{ + static char_t* parse_wnorm(char_t* s, char_t end_quote) + { + gap g; + + // trim leading whitespaces + if (PUGI__IS_CHARTYPE(*s, ct_space)) + { + char_t* str = s; + + do + ++str; + while (PUGI__IS_CHARTYPE(*str, ct_space)); + + g.push(s, str - s); + } + + while (true) + { + PUGI__SCANWHILE_UNROLL(!PUGI__IS_CHARTYPE(ss, ct_parse_attr_ws | ct_space)); + + if (*s == end_quote) + { + char_t* str = g.flush(s); + + do + *str-- = 0; + while (PUGI__IS_CHARTYPE(*str, ct_space)); + + return s + 1; + } + else if (PUGI__IS_CHARTYPE(*s, ct_space)) + { + *s++ = ' '; + + if (PUGI__IS_CHARTYPE(*s, ct_space)) + { + char_t* str = s + 1; + while (PUGI__IS_CHARTYPE(*str, ct_space)) + ++str; + + g.push(s, str - s); + } + } + else if (opt_escape::value && *s == '&') + { + s = strconv_escape(s, g); + } + else if (!*s) { return 0; } - else ++s; + else + ++s; } } - PUGI__FN char_t* strconv_cdata(char_t* s, char_t endch) + static char_t* parse_wconv(char_t* s, char_t end_quote) { gap g; while (true) { - PUGI__SCANWHILE_UNROLL(!PUGI__IS_CHARTYPE(ss, ct_parse_cdata)); + PUGI__SCANWHILE_UNROLL(!PUGI__IS_CHARTYPE(ss, ct_parse_attr_ws)); - if (*s == '\r') // Either a single 0x0d or 0x0d 0x0a pair - { - *s++ = '\n'; // replace first one with 0x0a - - if (*s == '\n') g.push(s, 1); - } - else if (s[0] == ']' && s[1] == ']' && PUGI__ENDSWITH(s[2], '>')) // CDATA ends here + if (*s == end_quote) { *g.flush(s) = 0; return s + 1; } - else if (*s == 0) + else if (PUGI__IS_CHARTYPE(*s, ct_space)) { - return 0; - } - else ++s; - } - } - - typedef char_t* (*strconv_pcdata_t)(char_t*); - - template struct strconv_pcdata_impl - { - static char_t* parse(char_t* s) - { - gap g; - - char_t* begin = s; - - while (true) - { - PUGI__SCANWHILE_UNROLL(!PUGI__IS_CHARTYPE(ss, ct_parse_pcdata)); - - if (*s == '<') // PCDATA ends here - { - char_t* end = g.flush(s); - - if (opt_trim::value) - while (end > begin && PUGI__IS_CHARTYPE(end[-1], ct_space)) - --end; - - *end = 0; - - return s + 1; - } - else if (opt_eol::value && *s == '\r') // Either a single 0x0d or 0x0d 0x0a pair - { - *s++ = '\n'; // replace first one with 0x0a - - if (*s == '\n') g.push(s, 1); - } - else if (opt_escape::value && *s == '&') - { - s = strconv_escape(s, g); - } - else if (*s == 0) - { - char_t* end = g.flush(s); - - if (opt_trim::value) - while (end > begin && PUGI__IS_CHARTYPE(end[-1], ct_space)) - --end; - - *end = 0; - - return s; - } - else ++s; - } - } - }; - - PUGI__FN strconv_pcdata_t get_strconv_pcdata(unsigned int optmask) - { - PUGI__STATIC_ASSERT(parse_escapes == 0x10 && parse_eol == 0x20 && parse_trim_pcdata == 0x0800); - - switch (((optmask >> 4) & 3) | ((optmask >> 9) & 4)) // get bitmask for flags (trim eol escapes); this simultaneously checks 3 options from assertion above - { - case 0: return strconv_pcdata_impl::parse; - case 1: return strconv_pcdata_impl::parse; - case 2: return strconv_pcdata_impl::parse; - case 3: return strconv_pcdata_impl::parse; - case 4: return strconv_pcdata_impl::parse; - case 5: return strconv_pcdata_impl::parse; - case 6: return strconv_pcdata_impl::parse; - case 7: return strconv_pcdata_impl::parse; - default: assert(false); return 0; // unreachable - } - } - - typedef char_t* (*strconv_attribute_t)(char_t*, char_t); - - template struct strconv_attribute_impl - { - static char_t* parse_wnorm(char_t* s, char_t end_quote) - { - gap g; - - // trim leading whitespaces - if (PUGI__IS_CHARTYPE(*s, ct_space)) - { - char_t* str = s; - - do ++str; - while (PUGI__IS_CHARTYPE(*str, ct_space)); - - g.push(s, str - s); - } - - while (true) - { - PUGI__SCANWHILE_UNROLL(!PUGI__IS_CHARTYPE(ss, ct_parse_attr_ws | ct_space)); - - if (*s == end_quote) - { - char_t* str = g.flush(s); - - do *str-- = 0; - while (PUGI__IS_CHARTYPE(*str, ct_space)); - - return s + 1; - } - else if (PUGI__IS_CHARTYPE(*s, ct_space)) + if (*s == '\r') { *s++ = ' '; - if (PUGI__IS_CHARTYPE(*s, ct_space)) - { - char_t* str = s + 1; - while (PUGI__IS_CHARTYPE(*str, ct_space)) ++str; - - g.push(s, str - s); - } + if (*s == '\n') + g.push(s, 1); } - else if (opt_escape::value && *s == '&') - { - s = strconv_escape(s, g); - } - else if (!*s) - { - return 0; - } - else ++s; + else + *s++ = ' '; } - } - - static char_t* parse_wconv(char_t* s, char_t end_quote) - { - gap g; - - while (true) + else if (opt_escape::value && *s == '&') { - PUGI__SCANWHILE_UNROLL(!PUGI__IS_CHARTYPE(ss, ct_parse_attr_ws)); - - if (*s == end_quote) - { - *g.flush(s) = 0; - - return s + 1; - } - else if (PUGI__IS_CHARTYPE(*s, ct_space)) - { - if (*s == '\r') - { - *s++ = ' '; - - if (*s == '\n') g.push(s, 1); - } - else *s++ = ' '; - } - else if (opt_escape::value && *s == '&') - { - s = strconv_escape(s, g); - } - else if (!*s) - { - return 0; - } - else ++s; + s = strconv_escape(s, g); } - } - - static char_t* parse_eol(char_t* s, char_t end_quote) - { - gap g; - - while (true) + else if (!*s) { - PUGI__SCANWHILE_UNROLL(!PUGI__IS_CHARTYPE(ss, ct_parse_attr)); - - if (*s == end_quote) - { - *g.flush(s) = 0; - - return s + 1; - } - else if (*s == '\r') - { - *s++ = '\n'; - - if (*s == '\n') g.push(s, 1); - } - else if (opt_escape::value && *s == '&') - { - s = strconv_escape(s, g); - } - else if (!*s) - { - return 0; - } - else ++s; + return 0; } - } - - static char_t* parse_simple(char_t* s, char_t end_quote) - { - gap g; - - while (true) - { - PUGI__SCANWHILE_UNROLL(!PUGI__IS_CHARTYPE(ss, ct_parse_attr)); - - if (*s == end_quote) - { - *g.flush(s) = 0; - - return s + 1; - } - else if (opt_escape::value && *s == '&') - { - s = strconv_escape(s, g); - } - else if (!*s) - { - return 0; - } - else ++s; - } - } - }; - - PUGI__FN strconv_attribute_t get_strconv_attribute(unsigned int optmask) - { - PUGI__STATIC_ASSERT(parse_escapes == 0x10 && parse_eol == 0x20 && parse_wconv_attribute == 0x40 && parse_wnorm_attribute == 0x80); - - switch ((optmask >> 4) & 15) // get bitmask for flags (wnorm wconv eol escapes); this simultaneously checks 4 options from assertion above - { - case 0: return strconv_attribute_impl::parse_simple; - case 1: return strconv_attribute_impl::parse_simple; - case 2: return strconv_attribute_impl::parse_eol; - case 3: return strconv_attribute_impl::parse_eol; - case 4: return strconv_attribute_impl::parse_wconv; - case 5: return strconv_attribute_impl::parse_wconv; - case 6: return strconv_attribute_impl::parse_wconv; - case 7: return strconv_attribute_impl::parse_wconv; - case 8: return strconv_attribute_impl::parse_wnorm; - case 9: return strconv_attribute_impl::parse_wnorm; - case 10: return strconv_attribute_impl::parse_wnorm; - case 11: return strconv_attribute_impl::parse_wnorm; - case 12: return strconv_attribute_impl::parse_wnorm; - case 13: return strconv_attribute_impl::parse_wnorm; - case 14: return strconv_attribute_impl::parse_wnorm; - case 15: return strconv_attribute_impl::parse_wnorm; - default: assert(false); return 0; // unreachable + else + ++s; } } - inline xml_parse_result make_parse_result(xml_parse_status status, ptrdiff_t offset = 0) + static char_t* parse_eol(char_t* s, char_t end_quote) { - xml_parse_result result; - result.status = status; - result.offset = offset; + gap g; - return result; + while (true) + { + PUGI__SCANWHILE_UNROLL(!PUGI__IS_CHARTYPE(ss, ct_parse_attr)); + + if (*s == end_quote) + { + *g.flush(s) = 0; + + return s + 1; + } + else if (*s == '\r') + { + *s++ = '\n'; + + if (*s == '\n') + g.push(s, 1); + } + else if (opt_escape::value && *s == '&') + { + s = strconv_escape(s, g); + } + else if (!*s) + { + return 0; + } + else + ++s; + } } - struct xml_parser + static char_t* parse_simple(char_t* s, char_t end_quote) { - xml_allocator* alloc; - char_t* error_offset; - xml_parse_status error_status; + gap g; - xml_parser(xml_allocator* alloc_): alloc(alloc_), error_offset(0), error_status(status_ok) + while (true) { + PUGI__SCANWHILE_UNROLL(!PUGI__IS_CHARTYPE(ss, ct_parse_attr)); + + if (*s == end_quote) + { + *g.flush(s) = 0; + + return s + 1; + } + else if (opt_escape::value && *s == '&') + { + s = strconv_escape(s, g); + } + else if (!*s) + { + return 0; + } + else + ++s; } + } +}; - // DOCTYPE consists of nested sections of the following possible types: - // , , "...", '...' - // - // - // First group can not contain nested groups - // Second group can contain nested groups of the same type - // Third group can contain all other groups - char_t* parse_doctype_primitive(char_t* s) +PUGI__FN strconv_attribute_t get_strconv_attribute(unsigned int optmask) +{ + PUGI__STATIC_ASSERT(parse_escapes == 0x10 && parse_eol == 0x20 && parse_wconv_attribute == 0x40 && parse_wnorm_attribute == 0x80); + + switch ((optmask >> 4) & 15) // get bitmask for flags (wnorm wconv eol escapes); this simultaneously checks 4 options from assertion above + { + case 0: + return strconv_attribute_impl::parse_simple; + case 1: + return strconv_attribute_impl::parse_simple; + case 2: + return strconv_attribute_impl::parse_eol; + case 3: + return strconv_attribute_impl::parse_eol; + case 4: + return strconv_attribute_impl::parse_wconv; + case 5: + return strconv_attribute_impl::parse_wconv; + case 6: + return strconv_attribute_impl::parse_wconv; + case 7: + return strconv_attribute_impl::parse_wconv; + case 8: + return strconv_attribute_impl::parse_wnorm; + case 9: + return strconv_attribute_impl::parse_wnorm; + case 10: + return strconv_attribute_impl::parse_wnorm; + case 11: + return strconv_attribute_impl::parse_wnorm; + case 12: + return strconv_attribute_impl::parse_wnorm; + case 13: + return strconv_attribute_impl::parse_wnorm; + case 14: + return strconv_attribute_impl::parse_wnorm; + case 15: + return strconv_attribute_impl::parse_wnorm; + default: + assert(false); + return 0; // unreachable + } +} + +inline xml_parse_result make_parse_result(xml_parse_status status, ptrdiff_t offset = 0) +{ + xml_parse_result result; + result.status = status; + result.offset = offset; + + return result; +} + +struct xml_parser +{ + xml_allocator* alloc; + char_t* error_offset; + xml_parse_status error_status; + + xml_parser(xml_allocator* alloc_) + : alloc(alloc_) + , error_offset(0) + , error_status(status_ok) + { + } + + // DOCTYPE consists of nested sections of the following possible types: + // , , "...", '...' + // + // + // First group can not contain nested groups + // Second group can contain nested groups of the same type + // Third group can contain all other groups + char_t* parse_doctype_primitive(char_t* s) + { + if (*s == '"' || *s == '\'') { - if (*s == '"' || *s == '\'') - { - // quoted string - char_t ch = *s++; - PUGI__SCANFOR(*s == ch); - if (!*s) PUGI__THROW_ERROR(status_bad_doctype, s); + // quoted string + char_t ch = *s++; + PUGI__SCANFOR(*s == ch); + if (!*s) + PUGI__THROW_ERROR(status_bad_doctype, s); - s++; - } - else if (s[0] == '<' && s[1] == '?') - { - // - s += 2; - PUGI__SCANFOR(s[0] == '?' && s[1] == '>'); // no need for ENDSWITH because ?> can't terminate proper doctype - if (!*s) PUGI__THROW_ERROR(status_bad_doctype, s); - - s += 2; - } - else if (s[0] == '<' && s[1] == '!' && s[2] == '-' && s[3] == '-') - { - s += 4; - PUGI__SCANFOR(s[0] == '-' && s[1] == '-' && s[2] == '>'); // no need for ENDSWITH because --> can't terminate proper doctype - if (!*s) PUGI__THROW_ERROR(status_bad_doctype, s); - - s += 3; - } - else PUGI__THROW_ERROR(status_bad_doctype, s); - - return s; + s++; } - - char_t* parse_doctype_ignore(char_t* s) + else if (s[0] == '<' && s[1] == '?') { - size_t depth = 0; + // + s += 2; + PUGI__SCANFOR(s[0] == '?' && s[1] == '>'); // no need for ENDSWITH because ?> can't terminate proper doctype + if (!*s) + PUGI__THROW_ERROR(status_bad_doctype, s); + + s += 2; + } + else if (s[0] == '<' && s[1] == '!' && s[2] == '-' && s[3] == '-') + { + s += 4; + PUGI__SCANFOR(s[0] == '-' && s[1] == '-' && s[2] == '>'); // no need for ENDSWITH because --> can't terminate proper doctype + if (!*s) + PUGI__THROW_ERROR(status_bad_doctype, s); - assert(s[0] == '<' && s[1] == '!' && s[2] == '['); s += 3; + } + else + PUGI__THROW_ERROR(status_bad_doctype, s); - while (*s) + return s; + } + + char_t* parse_doctype_ignore(char_t* s) + { + size_t depth = 0; + + assert(s[0] == '<' && s[1] == '!' && s[2] == '['); + s += 3; + + while (*s) + { + if (s[0] == '<' && s[1] == '!' && s[2] == '[') { - if (s[0] == '<' && s[1] == '!' && s[2] == '[') + // nested ignore section + s += 3; + depth++; + } + else if (s[0] == ']' && s[1] == ']' && s[2] == '>') + { + // ignore section end + s += 3; + + if (depth == 0) + return s; + + depth--; + } + else + s++; + } + + PUGI__THROW_ERROR(status_bad_doctype, s); + } + + char_t* parse_doctype_group(char_t* s, char_t endch) + { + size_t depth = 0; + + assert((s[0] == '<' || s[0] == 0) && s[1] == '!'); + s += 2; + + while (*s) + { + if (s[0] == '<' && s[1] == '!' && s[2] != '-') + { + if (s[2] == '[') { - // nested ignore section - s += 3; + // ignore + s = parse_doctype_ignore(s); + if (!s) + return s; + } + else + { + // some control group + s += 2; depth++; } - else if (s[0] == ']' && s[1] == ']' && s[2] == '>') - { - // ignore section end - s += 3; - - if (depth == 0) - return s; - - depth--; - } - else s++; } - - PUGI__THROW_ERROR(status_bad_doctype, s); - } - - char_t* parse_doctype_group(char_t* s, char_t endch) - { - size_t depth = 0; - - assert((s[0] == '<' || s[0] == 0) && s[1] == '!'); - s += 2; - - while (*s) + else if (s[0] == '<' || s[0] == '"' || s[0] == '\'') { - if (s[0] == '<' && s[1] == '!' && s[2] != '-') - { - if (s[2] == '[') - { - // ignore - s = parse_doctype_ignore(s); - if (!s) return s; - } - else - { - // some control group - s += 2; - depth++; - } - } - else if (s[0] == '<' || s[0] == '"' || s[0] == '\'') - { - // unknown tag (forbidden), or some primitive group - s = parse_doctype_primitive(s); - if (!s) return s; - } - else if (*s == '>') - { - if (depth == 0) - return s; - - depth--; - s++; - } - else s++; + // unknown tag (forbidden), or some primitive group + s = parse_doctype_primitive(s); + if (!s) + return s; } + else if (*s == '>') + { + if (depth == 0) + return s; - if (depth != 0 || endch != '>') PUGI__THROW_ERROR(status_bad_doctype, s); - - return s; + depth--; + s++; + } + else + s++; } - char_t* parse_exclamation(char_t* s, xml_node_struct* cursor, unsigned int optmsk, char_t endch) + if (depth != 0 || endch != '>') + PUGI__THROW_ERROR(status_bad_doctype, s); + + return s; + } + + char_t* parse_exclamation(char_t* s, xml_node_struct* cursor, unsigned int optmsk, char_t endch) + { + // parse node contents, starting with exclamation mark + ++s; + + if (*s == '-') // 'value = s; // Save the offset. + } + + if (PUGI__OPTSET(parse_eol) && PUGI__OPTSET(parse_comments)) + { + s = strconv_comment(s, endch); + + if (!s) + PUGI__THROW_ERROR(status_bad_comment, cursor->value); + } + else + { + // Scan for terminating '-->'. + PUGI__SCANFOR(s[0] == '-' && s[1] == '-' && PUGI__ENDSWITH(s[2], '>')); + PUGI__CHECK_ERROR(status_bad_comment, s); if (PUGI__OPTSET(parse_comments)) - { - PUGI__PUSHNODE(node_comment); // Append a new node on the tree. - cursor->value = s; // Save the offset. - } + *s = 0; // Zero-terminate this segment at the first terminating '-'. - if (PUGI__OPTSET(parse_eol) && PUGI__OPTSET(parse_comments)) - { - s = strconv_comment(s, endch); + s += (s[2] == '>' ? 3 : 2); // Step over the '\0->'. + } + } + else + PUGI__THROW_ERROR(status_bad_comment, s); + } + else if (*s == '[') + { + // 'value); + if (PUGI__OPTSET(parse_cdata)) + { + PUGI__PUSHNODE(node_cdata); // Append a new node on the tree. + cursor->value = s; // Save the offset. + + if (PUGI__OPTSET(parse_eol)) + { + s = strconv_cdata(s, endch); + + if (!s) + PUGI__THROW_ERROR(status_bad_cdata, cursor->value); } else - { - // Scan for terminating '-->'. - PUGI__SCANFOR(s[0] == '-' && s[1] == '-' && PUGI__ENDSWITH(s[2], '>')); - PUGI__CHECK_ERROR(status_bad_comment, s); - - if (PUGI__OPTSET(parse_comments)) - *s = 0; // Zero-terminate this segment at the first terminating '-'. - - s += (s[2] == '>' ? 3 : 2); // Step over the '\0->'. - } - } - else PUGI__THROW_ERROR(status_bad_comment, s); - } - else if (*s == '[') - { - // 'value = s; // Save the offset. - - if (PUGI__OPTSET(parse_eol)) - { - s = strconv_cdata(s, endch); - - if (!s) PUGI__THROW_ERROR(status_bad_cdata, cursor->value); - } - else - { - // Scan for terminating ']]>'. - PUGI__SCANFOR(s[0] == ']' && s[1] == ']' && PUGI__ENDSWITH(s[2], '>')); - PUGI__CHECK_ERROR(status_bad_cdata, s); - - *s++ = 0; // Zero-terminate this segment. - } - } - else // Flagged for discard, but we still have to scan for the terminator. { // Scan for terminating ']]>'. PUGI__SCANFOR(s[0] == ']' && s[1] == ']' && PUGI__ENDSWITH(s[2], '>')); PUGI__CHECK_ERROR(status_bad_cdata, s); - ++s; + *s++ = 0; // Zero-terminate this segment. } - - s += (s[1] == '>' ? 2 : 1); // Step over the last ']>'. } - else PUGI__THROW_ERROR(status_bad_cdata, s); - } - else if (s[0] == 'D' && s[1] == 'O' && s[2] == 'C' && s[3] == 'T' && s[4] == 'Y' && s[5] == 'P' && PUGI__ENDSWITH(s[6], 'E')) - { - s -= 2; - - if (cursor->parent) PUGI__THROW_ERROR(status_bad_doctype, s); - - char_t* mark = s + 9; - - s = parse_doctype_group(s, endch); - if (!s) return s; - - assert((*s == 0 && endch == '>') || *s == '>'); - if (*s) *s++ = 0; - - if (PUGI__OPTSET(parse_doctype)) + else // Flagged for discard, but we still have to scan for the terminator. { - while (PUGI__IS_CHARTYPE(*mark, ct_space)) ++mark; + // Scan for terminating ']]>'. + PUGI__SCANFOR(s[0] == ']' && s[1] == ']' && PUGI__ENDSWITH(s[2], '>')); + PUGI__CHECK_ERROR(status_bad_cdata, s); - PUGI__PUSHNODE(node_doctype); - - cursor->value = mark; + ++s; } - } - else if (*s == 0 && endch == '-') PUGI__THROW_ERROR(status_bad_comment, s); - else if (*s == 0 && endch == '[') PUGI__THROW_ERROR(status_bad_cdata, s); - else PUGI__THROW_ERROR(status_unrecognized_tag, s); - return s; + s += (s[1] == '>' ? 2 : 1); // Step over the last ']>'. + } + else + PUGI__THROW_ERROR(status_bad_cdata, s); } - - char_t* parse_question(char_t* s, xml_node_struct*& ref_cursor, unsigned int optmsk, char_t endch) + else if (s[0] == 'D' && s[1] == 'O' && s[2] == 'C' && s[3] == 'T' && s[4] == 'Y' && s[5] == 'P' && PUGI__ENDSWITH(s[6], 'E')) { - // load into registers - xml_node_struct* cursor = ref_cursor; - char_t ch = 0; + s -= 2; - // parse node contents, starting with question mark - ++s; + if (cursor->parent) + PUGI__THROW_ERROR(status_bad_doctype, s); - // read PI target - char_t* target = s; + char_t* mark = s + 9; - if (!PUGI__IS_CHARTYPE(*s, ct_start_symbol)) PUGI__THROW_ERROR(status_bad_pi, s); + s = parse_doctype_group(s, endch); + if (!s) + return s; - PUGI__SCANWHILE(PUGI__IS_CHARTYPE(*s, ct_symbol)); - PUGI__CHECK_ERROR(status_bad_pi, s); + assert((*s == 0 && endch == '>') || *s == '>'); + if (*s) + *s++ = 0; - // determine node type; stricmp / strcasecmp is not portable - bool declaration = (target[0] | ' ') == 'x' && (target[1] | ' ') == 'm' && (target[2] | ' ') == 'l' && target + 3 == s; - - if (declaration ? PUGI__OPTSET(parse_declaration) : PUGI__OPTSET(parse_pi)) + if (PUGI__OPTSET(parse_doctype)) { - if (declaration) - { - // disallow non top-level declarations - if (cursor->parent) PUGI__THROW_ERROR(status_bad_pi, s); + while (PUGI__IS_CHARTYPE(*mark, ct_space)) + ++mark; - PUGI__PUSHNODE(node_declaration); - } - else - { - PUGI__PUSHNODE(node_pi); - } + PUGI__PUSHNODE(node_doctype); - cursor->name = target; + cursor->value = mark; + } + } + else if (*s == 0 && endch == '-') + PUGI__THROW_ERROR(status_bad_comment, s); + else if (*s == 0 && endch == '[') + PUGI__THROW_ERROR(status_bad_cdata, s); + else + PUGI__THROW_ERROR(status_unrecognized_tag, s); - PUGI__ENDSEG(); + return s; + } - // parse value/attributes - if (ch == '?') - { - // empty node - if (!PUGI__ENDSWITH(*s, '>')) PUGI__THROW_ERROR(status_bad_pi, s); - s += (*s == '>'); + char_t* parse_question(char_t* s, xml_node_struct*& ref_cursor, unsigned int optmsk, char_t endch) + { + // load into registers + xml_node_struct* cursor = ref_cursor; + char_t ch = 0; - PUGI__POPNODE(); - } - else if (PUGI__IS_CHARTYPE(ch, ct_space)) - { - PUGI__SKIPWS(); + // parse node contents, starting with question mark + ++s; - // scan for tag end - char_t* value = s; + // read PI target + char_t* target = s; - PUGI__SCANFOR(s[0] == '?' && PUGI__ENDSWITH(s[1], '>')); - PUGI__CHECK_ERROR(status_bad_pi, s); + if (!PUGI__IS_CHARTYPE(*s, ct_start_symbol)) + PUGI__THROW_ERROR(status_bad_pi, s); - if (declaration) - { - // replace ending ? with / so that 'element' terminates properly - *s = '/'; + PUGI__SCANWHILE(PUGI__IS_CHARTYPE(*s, ct_symbol)); + PUGI__CHECK_ERROR(status_bad_pi, s); - // we exit from this function with cursor at node_declaration, which is a signal to parse() to go to LOC_ATTRIBUTES - s = value; - } - else - { - // store value and step over > - cursor->value = value; + // determine node type; stricmp / strcasecmp is not portable + bool declaration = (target[0] | ' ') == 'x' && (target[1] | ' ') == 'm' && (target[2] | ' ') == 'l' && target + 3 == s; - PUGI__POPNODE(); + if (declaration ? PUGI__OPTSET(parse_declaration) : PUGI__OPTSET(parse_pi)) + { + if (declaration) + { + // disallow non top-level declarations + if (cursor->parent) + PUGI__THROW_ERROR(status_bad_pi, s); - PUGI__ENDSEG(); - - s += (*s == '>'); - } - } - else PUGI__THROW_ERROR(status_bad_pi, s); + PUGI__PUSHNODE(node_declaration); } else { + PUGI__PUSHNODE(node_pi); + } + + cursor->name = target; + + PUGI__ENDSEG(); + + // parse value/attributes + if (ch == '?') + { + // empty node + if (!PUGI__ENDSWITH(*s, '>')) + PUGI__THROW_ERROR(status_bad_pi, s); + s += (*s == '>'); + + PUGI__POPNODE(); + } + else if (PUGI__IS_CHARTYPE(ch, ct_space)) + { + PUGI__SKIPWS(); + // scan for tag end + char_t* value = s; + PUGI__SCANFOR(s[0] == '?' && PUGI__ENDSWITH(s[1], '>')); PUGI__CHECK_ERROR(status_bad_pi, s); - s += (s[1] == '>' ? 2 : 1); + if (declaration) + { + // replace ending ? with / so that 'element' terminates properly + *s = '/'; + + // we exit from this function with cursor at node_declaration, which is a signal to parse() to go to LOC_ATTRIBUTES + s = value; + } + else + { + // store value and step over > + cursor->value = value; + + PUGI__POPNODE(); + + PUGI__ENDSEG(); + + s += (*s == '>'); + } } + else + PUGI__THROW_ERROR(status_bad_pi, s); + } + else + { + // scan for tag end + PUGI__SCANFOR(s[0] == '?' && PUGI__ENDSWITH(s[1], '>')); + PUGI__CHECK_ERROR(status_bad_pi, s); - // store from registers - ref_cursor = cursor; - - return s; + s += (s[1] == '>' ? 2 : 1); } - char_t* parse_tree(char_t* s, xml_node_struct* root, unsigned int optmsk, char_t endch) + // store from registers + ref_cursor = cursor; + + return s; + } + + char_t* parse_tree(char_t* s, xml_node_struct* root, unsigned int optmsk, char_t endch) + { + strconv_attribute_t strconv_attribute = get_strconv_attribute(optmsk); + strconv_pcdata_t strconv_pcdata = get_strconv_pcdata(optmsk); + + char_t ch = 0; + xml_node_struct* cursor = root; + char_t* mark = s; + + while (*s != 0) { - strconv_attribute_t strconv_attribute = get_strconv_attribute(optmsk); - strconv_pcdata_t strconv_pcdata = get_strconv_pcdata(optmsk); - - char_t ch = 0; - xml_node_struct* cursor = root; - char_t* mark = s; - - while (*s != 0) + if (*s == '<') { - if (*s == '<') + ++s; + + LOC_TAG: + if (PUGI__IS_CHARTYPE(*s, ct_start_symbol)) // '<#...' { - ++s; + PUGI__PUSHNODE(node_element); // Append a new node to the tree. - LOC_TAG: - if (PUGI__IS_CHARTYPE(*s, ct_start_symbol)) // '<#...' + cursor->name = s; + + PUGI__SCANWHILE_UNROLL(PUGI__IS_CHARTYPE(ss, ct_symbol)); // Scan for a terminator. + PUGI__ENDSEG(); // Save char in 'ch', terminate & step over. + + if (ch == '>') { - PUGI__PUSHNODE(node_element); // Append a new node to the tree. - - cursor->name = s; - - PUGI__SCANWHILE_UNROLL(PUGI__IS_CHARTYPE(ss, ct_symbol)); // Scan for a terminator. - PUGI__ENDSEG(); // Save char in 'ch', terminate & step over. - - if (ch == '>') + // end of tag + } + else if (PUGI__IS_CHARTYPE(ch, ct_space)) + { + LOC_ATTRIBUTES: + while (true) { - // end of tag - } - else if (PUGI__IS_CHARTYPE(ch, ct_space)) - { - LOC_ATTRIBUTES: - while (true) + PUGI__SKIPWS(); // Eat any whitespace. + + if (PUGI__IS_CHARTYPE(*s, ct_start_symbol)) // <... #... { - PUGI__SKIPWS(); // Eat any whitespace. + xml_attribute_struct* a = append_new_attribute(cursor, *alloc); // Make space for this attribute. + if (!a) + PUGI__THROW_ERROR(status_out_of_memory, s); - if (PUGI__IS_CHARTYPE(*s, ct_start_symbol)) // <... #... + a->name = s; // Save the offset. + + PUGI__SCANWHILE_UNROLL(PUGI__IS_CHARTYPE(ss, ct_symbol)); // Scan for a terminator. + PUGI__ENDSEG(); // Save char in 'ch', terminate & step over. + + if (PUGI__IS_CHARTYPE(ch, ct_space)) { - xml_attribute_struct* a = append_new_attribute(cursor, *alloc); // Make space for this attribute. - if (!a) PUGI__THROW_ERROR(status_out_of_memory, s); + PUGI__SKIPWS(); // Eat any whitespace. - a->name = s; // Save the offset. - - PUGI__SCANWHILE_UNROLL(PUGI__IS_CHARTYPE(ss, ct_symbol)); // Scan for a terminator. - PUGI__ENDSEG(); // Save char in 'ch', terminate & step over. - - if (PUGI__IS_CHARTYPE(ch, ct_space)) - { - PUGI__SKIPWS(); // Eat any whitespace. - - ch = *s; - ++s; - } - - if (ch == '=') // '<... #=...' - { - PUGI__SKIPWS(); // Eat any whitespace. - - if (*s == '"' || *s == '\'') // '<... #="...' - { - ch = *s; // Save quote char to avoid breaking on "''" -or- '""'. - ++s; // Step over the quote. - a->value = s; // Save the offset. - - s = strconv_attribute(s, ch); - - if (!s) PUGI__THROW_ERROR(status_bad_attribute, a->value); - - // After this line the loop continues from the start; - // Whitespaces, / and > are ok, symbols and EOF are wrong, - // everything else will be detected - if (PUGI__IS_CHARTYPE(*s, ct_start_symbol)) PUGI__THROW_ERROR(status_bad_attribute, s); - } - else PUGI__THROW_ERROR(status_bad_attribute, s); - } - else PUGI__THROW_ERROR(status_bad_attribute, s); - } - else if (*s == '/') - { + ch = *s; ++s; - - if (*s == '>') - { - PUGI__POPNODE(); - s++; - break; - } - else if (*s == 0 && endch == '>') - { - PUGI__POPNODE(); - break; - } - else PUGI__THROW_ERROR(status_bad_start_element, s); } - else if (*s == '>') - { - ++s; + if (ch == '=') // '<... #=...' + { + PUGI__SKIPWS(); // Eat any whitespace. + + if (*s == '"' || *s == '\'') // '<... #="...' + { + ch = *s; // Save quote char to avoid breaking on "''" -or- '""'. + ++s; // Step over the quote. + a->value = s; // Save the offset. + + s = strconv_attribute(s, ch); + + if (!s) + PUGI__THROW_ERROR(status_bad_attribute, a->value); + + // After this line the loop continues from the start; + // Whitespaces, / and > are ok, symbols and EOF are wrong, + // everything else will be detected + if (PUGI__IS_CHARTYPE(*s, ct_start_symbol)) + PUGI__THROW_ERROR(status_bad_attribute, s); + } + else + PUGI__THROW_ERROR(status_bad_attribute, s); + } + else + PUGI__THROW_ERROR(status_bad_attribute, s); + } + else if (*s == '/') + { + ++s; + + if (*s == '>') + { + PUGI__POPNODE(); + s++; break; } else if (*s == 0 && endch == '>') { + PUGI__POPNODE(); break; } - else PUGI__THROW_ERROR(status_bad_start_element, s); + else + PUGI__THROW_ERROR(status_bad_start_element, s); } + else if (*s == '>') + { + ++s; - // !!! + break; + } + else if (*s == 0 && endch == '>') + { + break; + } + else + PUGI__THROW_ERROR(status_bad_start_element, s); } - else if (ch == '/') // '<#.../' - { - if (!PUGI__ENDSWITH(*s, '>')) PUGI__THROW_ERROR(status_bad_start_element, s); - PUGI__POPNODE(); // Pop. - - s += (*s == '>'); - } - else if (ch == 0) - { - // we stepped over null terminator, backtrack & handle closing tag - --s; - - if (endch != '>') PUGI__THROW_ERROR(status_bad_start_element, s); - } - else PUGI__THROW_ERROR(status_bad_start_element, s); + // !!! } - else if (*s == '/') + else if (ch == '/') // '<#.../' { - ++s; - - mark = s; - - char_t* name = cursor->name; - if (!name) PUGI__THROW_ERROR(status_end_element_mismatch, mark); - - while (PUGI__IS_CHARTYPE(*s, ct_symbol)) - { - if (*s++ != *name++) PUGI__THROW_ERROR(status_end_element_mismatch, mark); - } - - if (*name) - { - if (*s == 0 && name[0] == endch && name[1] == 0) PUGI__THROW_ERROR(status_bad_end_element, s); - else PUGI__THROW_ERROR(status_end_element_mismatch, mark); - } + if (!PUGI__ENDSWITH(*s, '>')) + PUGI__THROW_ERROR(status_bad_start_element, s); PUGI__POPNODE(); // Pop. - PUGI__SKIPWS(); - - if (*s == 0) - { - if (endch != '>') PUGI__THROW_ERROR(status_bad_end_element, s); - } - else - { - if (*s != '>') PUGI__THROW_ERROR(status_bad_end_element, s); - ++s; - } + s += (*s == '>'); } - else if (*s == '?') // '') + PUGI__THROW_ERROR(status_bad_start_element, s); } - else if (*s == '!') // 'name; + if (!name) + PUGI__THROW_ERROR(status_end_element_mismatch, mark); + + while (PUGI__IS_CHARTYPE(*s, ct_symbol)) { - // We skipped some whitespace characters because otherwise we would take the tag branch instead of PCDATA one - assert(mark != s); - - if (!PUGI__OPTSET(parse_ws_pcdata | parse_ws_pcdata_single) || PUGI__OPTSET(parse_trim_pcdata)) - { - continue; - } - else if (PUGI__OPTSET(parse_ws_pcdata_single)) - { - if (s[0] != '<' || s[1] != '/' || cursor->first_child) continue; - } + if (*s++ != *name++) + PUGI__THROW_ERROR(status_end_element_mismatch, mark); } - if (!PUGI__OPTSET(parse_trim_pcdata)) - s = mark; - - if (cursor->parent || PUGI__OPTSET(parse_fragment)) + if (*name) { - if (PUGI__OPTSET(parse_embed_pcdata) && cursor->parent && !cursor->first_child && !cursor->value) - { - cursor->value = s; // Save the offset. - } + if (*s == 0 && name[0] == endch && name[1] == 0) + PUGI__THROW_ERROR(status_bad_end_element, s); else - { - PUGI__PUSHNODE(node_pcdata); // Append a new node on the tree. + PUGI__THROW_ERROR(status_end_element_mismatch, mark); + } - cursor->value = s; // Save the offset. + PUGI__POPNODE(); // Pop. - PUGI__POPNODE(); // Pop since this is a standalone. - } + PUGI__SKIPWS(); - s = strconv_pcdata(s); - - if (!*s) break; + if (*s == 0) + { + if (endch != '>') + PUGI__THROW_ERROR(status_bad_end_element, s); } else { - PUGI__SCANFOR(*s == '<'); // '...<' - if (!*s) break; - + if (*s != '>') + PUGI__THROW_ERROR(status_bad_end_element, s); ++s; } - - // We're after '<' - goto LOC_TAG; } - } - - // check that last tag is closed - if (cursor != root) PUGI__THROW_ERROR(status_end_element_mismatch, s); - - return s; - } - - #ifdef PUGIXML_WCHAR_MODE - static char_t* parse_skip_bom(char_t* s) - { - unsigned int bom = 0xfeff; - return (s[0] == static_cast(bom)) ? s + 1 : s; - } - #else - static char_t* parse_skip_bom(char_t* s) - { - return (s[0] == '\xef' && s[1] == '\xbb' && s[2] == '\xbf') ? s + 3 : s; - } - #endif - - static bool has_element_node_siblings(xml_node_struct* node) - { - while (node) - { - if (PUGI__NODETYPE(node) == node_element) return true; - - node = node->next_sibling; - } - - return false; - } - - static xml_parse_result parse(char_t* buffer, size_t length, xml_document_struct* xmldoc, xml_node_struct* root, unsigned int optmsk) - { - // early-out for empty documents - if (length == 0) - return make_parse_result(PUGI__OPTSET(parse_fragment) ? status_ok : status_no_document_element); - - // get last child of the root before parsing - xml_node_struct* last_root_child = root->first_child ? root->first_child->prev_sibling_c + 0 : 0; - - // create parser on stack - xml_parser parser(static_cast(xmldoc)); - - // save last character and make buffer zero-terminated (speeds up parsing) - char_t endch = buffer[length - 1]; - buffer[length - 1] = 0; - - // skip BOM to make sure it does not end up as part of parse output - char_t* buffer_data = parse_skip_bom(buffer); - - // perform actual parsing - parser.parse_tree(buffer_data, root, optmsk, endch); - - xml_parse_result result = make_parse_result(parser.error_status, parser.error_offset ? parser.error_offset - buffer : 0); - assert(result.offset >= 0 && static_cast(result.offset) <= length); - - if (result) - { - // since we removed last character, we have to handle the only possible false positive (stray <) - if (endch == '<') - return make_parse_result(status_unrecognized_tag, length - 1); - - // check if there are any element nodes parsed - xml_node_struct* first_root_child_parsed = last_root_child ? last_root_child->next_sibling + 0 : root->first_child+ 0; - - if (!PUGI__OPTSET(parse_fragment) && !has_element_node_siblings(first_root_child_parsed)) - return make_parse_result(status_no_document_element, length - 1); - } - else - { - // roll back offset if it occurs on a null terminator in the source buffer - if (result.offset > 0 && static_cast(result.offset) == length - 1 && endch == 0) - result.offset--; - } - - return result; - } - }; - - // Output facilities - PUGI__FN xml_encoding get_write_native_encoding() - { - #ifdef PUGIXML_WCHAR_MODE - return get_wchar_encoding(); - #else - return encoding_utf8; - #endif - } - - PUGI__FN xml_encoding get_write_encoding(xml_encoding encoding) - { - // replace wchar encoding with utf implementation - if (encoding == encoding_wchar) return get_wchar_encoding(); - - // replace utf16 encoding with utf16 with specific endianness - if (encoding == encoding_utf16) return is_little_endian() ? encoding_utf16_le : encoding_utf16_be; - - // replace utf32 encoding with utf32 with specific endianness - if (encoding == encoding_utf32) return is_little_endian() ? encoding_utf32_le : encoding_utf32_be; - - // only do autodetection if no explicit encoding is requested - if (encoding != encoding_auto) return encoding; - - // assume utf8 encoding - return encoding_utf8; - } - - template PUGI__FN size_t convert_buffer_output_generic(typename T::value_type dest, const char_t* data, size_t length, D, T) - { - PUGI__STATIC_ASSERT(sizeof(char_t) == sizeof(typename D::type)); - - typename T::value_type end = D::process(reinterpret_cast(data), length, dest, T()); - - return static_cast(end - dest) * sizeof(*dest); - } - - template PUGI__FN size_t convert_buffer_output_generic(typename T::value_type dest, const char_t* data, size_t length, D, T, bool opt_swap) - { - PUGI__STATIC_ASSERT(sizeof(char_t) == sizeof(typename D::type)); - - typename T::value_type end = D::process(reinterpret_cast(data), length, dest, T()); - - if (opt_swap) - { - for (typename T::value_type i = dest; i != end; ++i) - *i = endian_swap(*i); - } - - return static_cast(end - dest) * sizeof(*dest); - } - -#ifdef PUGIXML_WCHAR_MODE - PUGI__FN size_t get_valid_length(const char_t* data, size_t length) - { - if (length < 1) return 0; - - // discard last character if it's the lead of a surrogate pair - return (sizeof(wchar_t) == 2 && static_cast(static_cast(data[length - 1]) - 0xD800) < 0x400) ? length - 1 : length; - } - - PUGI__FN size_t convert_buffer_output(char_t* r_char, uint8_t* r_u8, uint16_t* r_u16, uint32_t* r_u32, const char_t* data, size_t length, xml_encoding encoding) - { - // only endian-swapping is required - if (need_endian_swap_utf(encoding, get_wchar_encoding())) - { - convert_wchar_endian_swap(r_char, data, length); - - return length * sizeof(char_t); - } - - // convert to utf8 - if (encoding == encoding_utf8) - return convert_buffer_output_generic(r_u8, data, length, wchar_decoder(), utf8_writer()); - - // convert to utf16 - if (encoding == encoding_utf16_be || encoding == encoding_utf16_le) - { - xml_encoding native_encoding = is_little_endian() ? encoding_utf16_le : encoding_utf16_be; - - return convert_buffer_output_generic(r_u16, data, length, wchar_decoder(), utf16_writer(), native_encoding != encoding); - } - - // convert to utf32 - if (encoding == encoding_utf32_be || encoding == encoding_utf32_le) - { - xml_encoding native_encoding = is_little_endian() ? encoding_utf32_le : encoding_utf32_be; - - return convert_buffer_output_generic(r_u32, data, length, wchar_decoder(), utf32_writer(), native_encoding != encoding); - } - - // convert to latin1 - if (encoding == encoding_latin1) - return convert_buffer_output_generic(r_u8, data, length, wchar_decoder(), latin1_writer()); - - assert(false && "Invalid encoding"); // unreachable - return 0; - } -#else - PUGI__FN size_t get_valid_length(const char_t* data, size_t length) - { - if (length < 5) return 0; - - for (size_t i = 1; i <= 4; ++i) - { - uint8_t ch = static_cast(data[length - i]); - - // either a standalone character or a leading one - if ((ch & 0xc0) != 0x80) return length - i; - } - - // there are four non-leading characters at the end, sequence tail is broken so might as well process the whole chunk - return length; - } - - PUGI__FN size_t convert_buffer_output(char_t* /* r_char */, uint8_t* r_u8, uint16_t* r_u16, uint32_t* r_u32, const char_t* data, size_t length, xml_encoding encoding) - { - if (encoding == encoding_utf16_be || encoding == encoding_utf16_le) - { - xml_encoding native_encoding = is_little_endian() ? encoding_utf16_le : encoding_utf16_be; - - return convert_buffer_output_generic(r_u16, data, length, utf8_decoder(), utf16_writer(), native_encoding != encoding); - } - - if (encoding == encoding_utf32_be || encoding == encoding_utf32_le) - { - xml_encoding native_encoding = is_little_endian() ? encoding_utf32_le : encoding_utf32_be; - - return convert_buffer_output_generic(r_u32, data, length, utf8_decoder(), utf32_writer(), native_encoding != encoding); - } - - if (encoding == encoding_latin1) - return convert_buffer_output_generic(r_u8, data, length, utf8_decoder(), latin1_writer()); - - assert(false && "Invalid encoding"); // unreachable - return 0; - } -#endif - - class xml_buffered_writer - { - xml_buffered_writer(const xml_buffered_writer&); - xml_buffered_writer& operator=(const xml_buffered_writer&); - - public: - xml_buffered_writer(xml_writer& writer_, xml_encoding user_encoding): writer(writer_), bufsize(0), encoding(get_write_encoding(user_encoding)) - { - PUGI__STATIC_ASSERT(bufcapacity >= 8); - } - - size_t flush() - { - flush(buffer, bufsize); - bufsize = 0; - return 0; - } - - void flush(const char_t* data, size_t size) - { - if (size == 0) return; - - // fast path, just write data - if (encoding == get_write_native_encoding()) - writer.write(data, size * sizeof(char_t)); - else - { - // convert chunk - size_t result = convert_buffer_output(scratch.data_char, scratch.data_u8, scratch.data_u16, scratch.data_u32, data, size, encoding); - assert(result <= sizeof(scratch)); - - // write data - writer.write(scratch.data_u8, result); - } - } - - void write_direct(const char_t* data, size_t length) - { - // flush the remaining buffer contents - flush(); - - // handle large chunks - if (length > bufcapacity) - { - if (encoding == get_write_native_encoding()) + else if (*s == '?') // 'first_child) + continue; + } } - // need to convert in suitable chunks - while (length > bufcapacity) + if (!PUGI__OPTSET(parse_trim_pcdata)) + s = mark; + + if (cursor->parent || PUGI__OPTSET(parse_fragment)) { - // get chunk size by selecting such number of characters that are guaranteed to fit into scratch buffer - // and form a complete codepoint sequence (i.e. discard start of last codepoint if necessary) - size_t chunk_size = get_valid_length(data, bufcapacity); - assert(chunk_size); + if (PUGI__OPTSET(parse_embed_pcdata) && cursor->parent && !cursor->first_child && !cursor->value) + { + cursor->value = s; // Save the offset. + } + else + { + PUGI__PUSHNODE(node_pcdata); // Append a new node on the tree. - // convert chunk and write - flush(data, chunk_size); + cursor->value = s; // Save the offset. - // iterate - data += chunk_size; - length -= chunk_size; - } + PUGI__POPNODE(); // Pop since this is a standalone. + } - // small tail is copied below - bufsize = 0; - } + s = strconv_pcdata(s); - memcpy(buffer + bufsize, data, length * sizeof(char_t)); - bufsize += length; - } - - void write_buffer(const char_t* data, size_t length) - { - size_t offset = bufsize; - - if (offset + length <= bufcapacity) - { - memcpy(buffer + offset, data, length * sizeof(char_t)); - bufsize = offset + length; - } - else - { - write_direct(data, length); - } - } - - void write_string(const char_t* data) - { - // write the part of the string that fits in the buffer - size_t offset = bufsize; - - while (*data && offset < bufcapacity) - buffer[offset++] = *data++; - - // write the rest - if (offset < bufcapacity) - { - bufsize = offset; - } - else - { - // backtrack a bit if we have split the codepoint - size_t length = offset - bufsize; - size_t extra = length - get_valid_length(data - length, length); - - bufsize = offset - extra; - - write_direct(data - extra, strlength(data) + extra); - } - } - - void write(char_t d0) - { - size_t offset = bufsize; - if (offset > bufcapacity - 1) offset = flush(); - - buffer[offset + 0] = d0; - bufsize = offset + 1; - } - - void write(char_t d0, char_t d1) - { - size_t offset = bufsize; - if (offset > bufcapacity - 2) offset = flush(); - - buffer[offset + 0] = d0; - buffer[offset + 1] = d1; - bufsize = offset + 2; - } - - void write(char_t d0, char_t d1, char_t d2) - { - size_t offset = bufsize; - if (offset > bufcapacity - 3) offset = flush(); - - buffer[offset + 0] = d0; - buffer[offset + 1] = d1; - buffer[offset + 2] = d2; - bufsize = offset + 3; - } - - void write(char_t d0, char_t d1, char_t d2, char_t d3) - { - size_t offset = bufsize; - if (offset > bufcapacity - 4) offset = flush(); - - buffer[offset + 0] = d0; - buffer[offset + 1] = d1; - buffer[offset + 2] = d2; - buffer[offset + 3] = d3; - bufsize = offset + 4; - } - - void write(char_t d0, char_t d1, char_t d2, char_t d3, char_t d4) - { - size_t offset = bufsize; - if (offset > bufcapacity - 5) offset = flush(); - - buffer[offset + 0] = d0; - buffer[offset + 1] = d1; - buffer[offset + 2] = d2; - buffer[offset + 3] = d3; - buffer[offset + 4] = d4; - bufsize = offset + 5; - } - - void write(char_t d0, char_t d1, char_t d2, char_t d3, char_t d4, char_t d5) - { - size_t offset = bufsize; - if (offset > bufcapacity - 6) offset = flush(); - - buffer[offset + 0] = d0; - buffer[offset + 1] = d1; - buffer[offset + 2] = d2; - buffer[offset + 3] = d3; - buffer[offset + 4] = d4; - buffer[offset + 5] = d5; - bufsize = offset + 6; - } - - // utf8 maximum expansion: x4 (-> utf32) - // utf16 maximum expansion: x2 (-> utf32) - // utf32 maximum expansion: x1 - enum - { - bufcapacitybytes = - #ifdef PUGIXML_MEMORY_OUTPUT_STACK - PUGIXML_MEMORY_OUTPUT_STACK - #else - 10240 - #endif - , - bufcapacity = bufcapacitybytes / (sizeof(char_t) + 4) - }; - - char_t buffer[bufcapacity]; - - union - { - uint8_t data_u8[4 * bufcapacity]; - uint16_t data_u16[2 * bufcapacity]; - uint32_t data_u32[bufcapacity]; - char_t data_char[bufcapacity]; - } scratch; - - xml_writer& writer; - size_t bufsize; - xml_encoding encoding; - }; - - PUGI__FN void text_output_escaped(xml_buffered_writer& writer, const char_t* s, chartypex_t type) - { - while (*s) - { - const char_t* prev = s; - - // While *s is a usual symbol - PUGI__SCANWHILE_UNROLL(!PUGI__IS_CHARTYPEX(ss, type)); - - writer.write_buffer(prev, static_cast(s - prev)); - - switch (*s) - { - case 0: break; - case '&': - writer.write('&', 'a', 'm', 'p', ';'); - ++s; - break; - case '<': - writer.write('&', 'l', 't', ';'); - ++s; - break; - case '>': - writer.write('&', 'g', 't', ';'); - ++s; - break; - case '"': - writer.write('&', 'q', 'u', 'o', 't', ';'); - ++s; - break; - default: // s is not a usual symbol - { - unsigned int ch = static_cast(*s++); - assert(ch < 32); - - writer.write('&', '#', static_cast((ch / 10) + '0'), static_cast((ch % 10) + '0'), ';'); - } - } - } - } - - PUGI__FN void text_output(xml_buffered_writer& writer, const char_t* s, chartypex_t type, unsigned int flags) - { - if (flags & format_no_escapes) - writer.write_string(s); - else - text_output_escaped(writer, s, type); - } - - PUGI__FN void text_output_cdata(xml_buffered_writer& writer, const char_t* s) - { - do - { - writer.write('<', '!', '[', 'C', 'D'); - writer.write('A', 'T', 'A', '['); - - const char_t* prev = s; - - // look for ]]> sequence - we can't output it as is since it terminates CDATA - while (*s && !(s[0] == ']' && s[1] == ']' && s[2] == '>')) ++s; - - // skip ]] if we stopped at ]]>, > will go to the next CDATA section - if (*s) s += 2; - - writer.write_buffer(prev, static_cast(s - prev)); - - writer.write(']', ']', '>'); - } - while (*s); - } - - PUGI__FN void text_output_indent(xml_buffered_writer& writer, const char_t* indent, size_t indent_length, unsigned int depth) - { - switch (indent_length) - { - case 1: - { - for (unsigned int i = 0; i < depth; ++i) - writer.write(indent[0]); - break; - } - - case 2: - { - for (unsigned int i = 0; i < depth; ++i) - writer.write(indent[0], indent[1]); - break; - } - - case 3: - { - for (unsigned int i = 0; i < depth; ++i) - writer.write(indent[0], indent[1], indent[2]); - break; - } - - case 4: - { - for (unsigned int i = 0; i < depth; ++i) - writer.write(indent[0], indent[1], indent[2], indent[3]); - break; - } - - default: - { - for (unsigned int i = 0; i < depth; ++i) - writer.write_buffer(indent, indent_length); - } - } - } - - PUGI__FN void node_output_comment(xml_buffered_writer& writer, const char_t* s) - { - writer.write('<', '!', '-', '-'); - - while (*s) - { - const char_t* prev = s; - - // look for -\0 or -- sequence - we can't output it since -- is illegal in comment body - while (*s && !(s[0] == '-' && (s[1] == '-' || s[1] == 0))) ++s; - - writer.write_buffer(prev, static_cast(s - prev)); - - if (*s) - { - assert(*s == '-'); - - writer.write('-', ' '); - ++s; - } - } - - writer.write('-', '-', '>'); - } - - PUGI__FN void node_output_pi_value(xml_buffered_writer& writer, const char_t* s) - { - while (*s) - { - const char_t* prev = s; - - // look for ?> sequence - we can't output it since ?> terminates PI - while (*s && !(s[0] == '?' && s[1] == '>')) ++s; - - writer.write_buffer(prev, static_cast(s - prev)); - - if (*s) - { - assert(s[0] == '?' && s[1] == '>'); - - writer.write('?', ' ', '>'); - s += 2; - } - } - } - - PUGI__FN void node_output_attributes(xml_buffered_writer& writer, xml_node_struct* node, const char_t* indent, size_t indent_length, unsigned int flags, unsigned int depth) - { - const char_t* default_name = PUGIXML_TEXT(":anonymous"); - - for (xml_attribute_struct* a = node->first_attribute; a; a = a->next_attribute) - { - if ((flags & (format_indent_attributes | format_raw)) == format_indent_attributes) - { - writer.write('\n'); - - text_output_indent(writer, indent, indent_length, depth + 1); - } - else - { - writer.write(' '); - } - - writer.write_string(a->name ? a->name + 0 : default_name); - writer.write('=', '"'); - - if (a->value) - text_output(writer, a->value, ctx_special_attr, flags); - - writer.write('"'); - } - } - - PUGI__FN bool node_output_start(xml_buffered_writer& writer, xml_node_struct* node, const char_t* indent, size_t indent_length, unsigned int flags, unsigned int depth) - { - const char_t* default_name = PUGIXML_TEXT(":anonymous"); - const char_t* name = node->name ? node->name + 0 : default_name; - - writer.write('<'); - writer.write_string(name); - - if (node->first_attribute) - node_output_attributes(writer, node, indent, indent_length, flags, depth); - - // element nodes can have value if parse_embed_pcdata was used - if (!node->value) - { - if (!node->first_child) - { - if (flags & format_no_empty_element_tags) - { - writer.write('>', '<', '/'); - writer.write_string(name); - writer.write('>'); - - return false; + if (!*s) + break; } else { - if ((flags & format_raw) == 0) - writer.write(' '); + PUGI__SCANFOR(*s == '<'); // '...<' + if (!*s) + break; - writer.write('/', '>'); - - return false; + ++s; } - } - else - { - writer.write('>'); - return true; + // We're after '<' + goto LOC_TAG; } } + + // check that last tag is closed + if (cursor != root) + PUGI__THROW_ERROR(status_end_element_mismatch, s); + + return s; + } + +#ifdef PUGIXML_WCHAR_MODE + static char_t* parse_skip_bom(char_t* s) + { + unsigned int bom = 0xfeff; + return (s[0] == static_cast(bom)) ? s + 1 : s; + } +#else + static char_t* parse_skip_bom(char_t* s) + { + return (s[0] == '\xef' && s[1] == '\xbb' && s[2] == '\xbf') ? s + 3 : s; + } +#endif + + static bool has_element_node_siblings(xml_node_struct* node) + { + while (node) + { + if (PUGI__NODETYPE(node) == node_element) + return true; + + node = node->next_sibling; + } + + return false; + } + + static xml_parse_result parse(char_t* buffer, size_t length, xml_document_struct* xmldoc, xml_node_struct* root, unsigned int optmsk) + { + // early-out for empty documents + if (length == 0) + return make_parse_result(PUGI__OPTSET(parse_fragment) ? status_ok : status_no_document_element); + + // get last child of the root before parsing + xml_node_struct* last_root_child = root->first_child ? root->first_child->prev_sibling_c + 0 : 0; + + // create parser on stack + xml_parser parser(static_cast(xmldoc)); + + // save last character and make buffer zero-terminated (speeds up parsing) + char_t endch = buffer[length - 1]; + buffer[length - 1] = 0; + + // skip BOM to make sure it does not end up as part of parse output + char_t* buffer_data = parse_skip_bom(buffer); + + // perform actual parsing + parser.parse_tree(buffer_data, root, optmsk, endch); + + xml_parse_result result = make_parse_result(parser.error_status, parser.error_offset ? parser.error_offset - buffer : 0); + assert(result.offset >= 0 && static_cast(result.offset) <= length); + + if (result) + { + // since we removed last character, we have to handle the only possible false positive (stray <) + if (endch == '<') + return make_parse_result(status_unrecognized_tag, length - 1); + + // check if there are any element nodes parsed + xml_node_struct* first_root_child_parsed = last_root_child ? last_root_child->next_sibling + 0 : root->first_child + 0; + + if (!PUGI__OPTSET(parse_fragment) && !has_element_node_siblings(first_root_child_parsed)) + return make_parse_result(status_no_document_element, length - 1); + } else { - writer.write('>'); + // roll back offset if it occurs on a null terminator in the source buffer + if (result.offset > 0 && static_cast(result.offset) == length - 1 && endch == 0) + result.offset--; + } - text_output(writer, node->value, ctx_special_pcdata, flags); + return result; + } +}; - if (!node->first_child) +// Output facilities +PUGI__FN xml_encoding get_write_native_encoding() +{ +#ifdef PUGIXML_WCHAR_MODE + return get_wchar_encoding(); +#else + return encoding_utf8; +#endif +} + +PUGI__FN xml_encoding get_write_encoding(xml_encoding encoding) +{ + // replace wchar encoding with utf implementation + if (encoding == encoding_wchar) + return get_wchar_encoding(); + + // replace utf16 encoding with utf16 with specific endianness + if (encoding == encoding_utf16) + return is_little_endian() ? encoding_utf16_le : encoding_utf16_be; + + // replace utf32 encoding with utf32 with specific endianness + if (encoding == encoding_utf32) + return is_little_endian() ? encoding_utf32_le : encoding_utf32_be; + + // only do autodetection if no explicit encoding is requested + if (encoding != encoding_auto) + return encoding; + + // assume utf8 encoding + return encoding_utf8; +} + +template +PUGI__FN size_t convert_buffer_output_generic(typename T::value_type dest, const char_t* data, size_t length, D, T) +{ + PUGI__STATIC_ASSERT(sizeof(char_t) == sizeof(typename D::type)); + + typename T::value_type end = D::process(reinterpret_cast(data), length, dest, T()); + + return static_cast(end - dest) * sizeof(*dest); +} + +template +PUGI__FN size_t convert_buffer_output_generic(typename T::value_type dest, const char_t* data, size_t length, D, T, bool opt_swap) +{ + PUGI__STATIC_ASSERT(sizeof(char_t) == sizeof(typename D::type)); + + typename T::value_type end = D::process(reinterpret_cast(data), length, dest, T()); + + if (opt_swap) + { + for (typename T::value_type i = dest; i != end; ++i) + *i = endian_swap(*i); + } + + return static_cast(end - dest) * sizeof(*dest); +} + +#ifdef PUGIXML_WCHAR_MODE +PUGI__FN size_t get_valid_length(const char_t* data, size_t length) +{ + if (length < 1) + return 0; + + // discard last character if it's the lead of a surrogate pair + return (sizeof(wchar_t) == 2 && static_cast(static_cast(data[length - 1]) - 0xD800) < 0x400) ? length - 1 : length; +} + +PUGI__FN size_t convert_buffer_output(char_t* r_char, uint8_t* r_u8, uint16_t* r_u16, uint32_t* r_u32, const char_t* data, size_t length, xml_encoding encoding) +{ + // only endian-swapping is required + if (need_endian_swap_utf(encoding, get_wchar_encoding())) + { + convert_wchar_endian_swap(r_char, data, length); + + return length * sizeof(char_t); + } + + // convert to utf8 + if (encoding == encoding_utf8) + return convert_buffer_output_generic(r_u8, data, length, wchar_decoder(), utf8_writer()); + + // convert to utf16 + if (encoding == encoding_utf16_be || encoding == encoding_utf16_le) + { + xml_encoding native_encoding = is_little_endian() ? encoding_utf16_le : encoding_utf16_be; + + return convert_buffer_output_generic(r_u16, data, length, wchar_decoder(), utf16_writer(), native_encoding != encoding); + } + + // convert to utf32 + if (encoding == encoding_utf32_be || encoding == encoding_utf32_le) + { + xml_encoding native_encoding = is_little_endian() ? encoding_utf32_le : encoding_utf32_be; + + return convert_buffer_output_generic(r_u32, data, length, wchar_decoder(), utf32_writer(), native_encoding != encoding); + } + + // convert to latin1 + if (encoding == encoding_latin1) + return convert_buffer_output_generic(r_u8, data, length, wchar_decoder(), latin1_writer()); + + assert(false && "Invalid encoding"); // unreachable + return 0; +} +#else +PUGI__FN size_t get_valid_length(const char_t* data, size_t length) +{ + if (length < 5) + return 0; + + for (size_t i = 1; i <= 4; ++i) + { + uint8_t ch = static_cast(data[length - i]); + + // either a standalone character or a leading one + if ((ch & 0xc0) != 0x80) + return length - i; + } + + // there are four non-leading characters at the end, sequence tail is broken so might as well process the whole chunk + return length; +} + +PUGI__FN size_t convert_buffer_output(char_t* /* r_char */, uint8_t* r_u8, uint16_t* r_u16, uint32_t* r_u32, const char_t* data, size_t length, xml_encoding encoding) +{ + if (encoding == encoding_utf16_be || encoding == encoding_utf16_le) + { + xml_encoding native_encoding = is_little_endian() ? encoding_utf16_le : encoding_utf16_be; + + return convert_buffer_output_generic(r_u16, data, length, utf8_decoder(), utf16_writer(), native_encoding != encoding); + } + + if (encoding == encoding_utf32_be || encoding == encoding_utf32_le) + { + xml_encoding native_encoding = is_little_endian() ? encoding_utf32_le : encoding_utf32_be; + + return convert_buffer_output_generic(r_u32, data, length, utf8_decoder(), utf32_writer(), native_encoding != encoding); + } + + if (encoding == encoding_latin1) + return convert_buffer_output_generic(r_u8, data, length, utf8_decoder(), latin1_writer()); + + assert(false && "Invalid encoding"); // unreachable + return 0; +} +#endif + +class xml_buffered_writer +{ + xml_buffered_writer(const xml_buffered_writer&); + xml_buffered_writer& operator=(const xml_buffered_writer&); + + public: + xml_buffered_writer(xml_writer& writer_, xml_encoding user_encoding) + : writer(writer_) + , bufsize(0) + , encoding(get_write_encoding(user_encoding)) + { + PUGI__STATIC_ASSERT(bufcapacity >= 8); + } + + size_t flush() + { + flush(buffer, bufsize); + bufsize = 0; + return 0; + } + + void flush(const char_t* data, size_t size) + { + if (size == 0) + return; + + // fast path, just write data + if (encoding == get_write_native_encoding()) + writer.write(data, size * sizeof(char_t)); + else + { + // convert chunk + size_t result = convert_buffer_output(scratch.data_char, scratch.data_u8, scratch.data_u16, scratch.data_u32, data, size, encoding); + assert(result <= sizeof(scratch)); + + // write data + writer.write(scratch.data_u8, result); + } + } + + void write_direct(const char_t* data, size_t length) + { + // flush the remaining buffer contents + flush(); + + // handle large chunks + if (length > bufcapacity) + { + if (encoding == get_write_native_encoding()) { - writer.write('<', '/'); + // fast path, can just write data chunk + writer.write(data, length * sizeof(char_t)); + return; + } + + // need to convert in suitable chunks + while (length > bufcapacity) + { + // get chunk size by selecting such number of characters that are guaranteed to fit into scratch buffer + // and form a complete codepoint sequence (i.e. discard start of last codepoint if necessary) + size_t chunk_size = get_valid_length(data, bufcapacity); + assert(chunk_size); + + // convert chunk and write + flush(data, chunk_size); + + // iterate + data += chunk_size; + length -= chunk_size; + } + + // small tail is copied below + bufsize = 0; + } + + memcpy(buffer + bufsize, data, length * sizeof(char_t)); + bufsize += length; + } + + void write_buffer(const char_t* data, size_t length) + { + size_t offset = bufsize; + + if (offset + length <= bufcapacity) + { + memcpy(buffer + offset, data, length * sizeof(char_t)); + bufsize = offset + length; + } + else + { + write_direct(data, length); + } + } + + void write_string(const char_t* data) + { + // write the part of the string that fits in the buffer + size_t offset = bufsize; + + while (*data && offset < bufcapacity) + buffer[offset++] = *data++; + + // write the rest + if (offset < bufcapacity) + { + bufsize = offset; + } + else + { + // backtrack a bit if we have split the codepoint + size_t length = offset - bufsize; + size_t extra = length - get_valid_length(data - length, length); + + bufsize = offset - extra; + + write_direct(data - extra, strlength(data) + extra); + } + } + + void write(char_t d0) + { + size_t offset = bufsize; + if (offset > bufcapacity - 1) + offset = flush(); + + buffer[offset + 0] = d0; + bufsize = offset + 1; + } + + void write(char_t d0, char_t d1) + { + size_t offset = bufsize; + if (offset > bufcapacity - 2) + offset = flush(); + + buffer[offset + 0] = d0; + buffer[offset + 1] = d1; + bufsize = offset + 2; + } + + void write(char_t d0, char_t d1, char_t d2) + { + size_t offset = bufsize; + if (offset > bufcapacity - 3) + offset = flush(); + + buffer[offset + 0] = d0; + buffer[offset + 1] = d1; + buffer[offset + 2] = d2; + bufsize = offset + 3; + } + + void write(char_t d0, char_t d1, char_t d2, char_t d3) + { + size_t offset = bufsize; + if (offset > bufcapacity - 4) + offset = flush(); + + buffer[offset + 0] = d0; + buffer[offset + 1] = d1; + buffer[offset + 2] = d2; + buffer[offset + 3] = d3; + bufsize = offset + 4; + } + + void write(char_t d0, char_t d1, char_t d2, char_t d3, char_t d4) + { + size_t offset = bufsize; + if (offset > bufcapacity - 5) + offset = flush(); + + buffer[offset + 0] = d0; + buffer[offset + 1] = d1; + buffer[offset + 2] = d2; + buffer[offset + 3] = d3; + buffer[offset + 4] = d4; + bufsize = offset + 5; + } + + void write(char_t d0, char_t d1, char_t d2, char_t d3, char_t d4, char_t d5) + { + size_t offset = bufsize; + if (offset > bufcapacity - 6) + offset = flush(); + + buffer[offset + 0] = d0; + buffer[offset + 1] = d1; + buffer[offset + 2] = d2; + buffer[offset + 3] = d3; + buffer[offset + 4] = d4; + buffer[offset + 5] = d5; + bufsize = offset + 6; + } + + // utf8 maximum expansion: x4 (-> utf32) + // utf16 maximum expansion: x2 (-> utf32) + // utf32 maximum expansion: x1 + enum + { + bufcapacitybytes = +#ifdef PUGIXML_MEMORY_OUTPUT_STACK + PUGIXML_MEMORY_OUTPUT_STACK +#else + 10240 +#endif + , + bufcapacity = bufcapacitybytes / (sizeof(char_t) + 4) + }; + + char_t buffer[bufcapacity]; + + union { + uint8_t data_u8[4 * bufcapacity]; + uint16_t data_u16[2 * bufcapacity]; + uint32_t data_u32[bufcapacity]; + char_t data_char[bufcapacity]; + } scratch; + + xml_writer& writer; + size_t bufsize; + xml_encoding encoding; +}; + +PUGI__FN void text_output_escaped(xml_buffered_writer& writer, const char_t* s, chartypex_t type) +{ + while (*s) + { + const char_t* prev = s; + + // While *s is a usual symbol + PUGI__SCANWHILE_UNROLL(!PUGI__IS_CHARTYPEX(ss, type)); + + writer.write_buffer(prev, static_cast(s - prev)); + + switch (*s) + { + case 0: + break; + case '&': + writer.write('&', 'a', 'm', 'p', ';'); + ++s; + break; + case '<': + writer.write('&', 'l', 't', ';'); + ++s; + break; + case '>': + writer.write('&', 'g', 't', ';'); + ++s; + break; + case '"': + writer.write('&', 'q', 'u', 'o', 't', ';'); + ++s; + break; + default: // s is not a usual symbol + { + unsigned int ch = static_cast(*s++); + assert(ch < 32); + + writer.write('&', '#', static_cast((ch / 10) + '0'), static_cast((ch % 10) + '0'), ';'); + } + } + } +} + +PUGI__FN void text_output(xml_buffered_writer& writer, const char_t* s, chartypex_t type, unsigned int flags) +{ + if (flags & format_no_escapes) + writer.write_string(s); + else + text_output_escaped(writer, s, type); +} + +PUGI__FN void text_output_cdata(xml_buffered_writer& writer, const char_t* s) +{ + do + { + writer.write('<', '!', '[', 'C', 'D'); + writer.write('A', 'T', 'A', '['); + + const char_t* prev = s; + + // look for ]]> sequence - we can't output it as is since it terminates CDATA + while (*s && !(s[0] == ']' && s[1] == ']' && s[2] == '>')) + ++s; + + // skip ]] if we stopped at ]]>, > will go to the next CDATA section + if (*s) + s += 2; + + writer.write_buffer(prev, static_cast(s - prev)); + + writer.write(']', ']', '>'); + } while (*s); +} + +PUGI__FN void text_output_indent(xml_buffered_writer& writer, const char_t* indent, size_t indent_length, unsigned int depth) +{ + switch (indent_length) + { + case 1: + { + for (unsigned int i = 0; i < depth; ++i) + writer.write(indent[0]); + break; + } + + case 2: + { + for (unsigned int i = 0; i < depth; ++i) + writer.write(indent[0], indent[1]); + break; + } + + case 3: + { + for (unsigned int i = 0; i < depth; ++i) + writer.write(indent[0], indent[1], indent[2]); + break; + } + + case 4: + { + for (unsigned int i = 0; i < depth; ++i) + writer.write(indent[0], indent[1], indent[2], indent[3]); + break; + } + + default: + { + for (unsigned int i = 0; i < depth; ++i) + writer.write_buffer(indent, indent_length); + } + } +} + +PUGI__FN void node_output_comment(xml_buffered_writer& writer, const char_t* s) +{ + writer.write('<', '!', '-', '-'); + + while (*s) + { + const char_t* prev = s; + + // look for -\0 or -- sequence - we can't output it since -- is illegal in comment body + while (*s && !(s[0] == '-' && (s[1] == '-' || s[1] == 0))) + ++s; + + writer.write_buffer(prev, static_cast(s - prev)); + + if (*s) + { + assert(*s == '-'); + + writer.write('-', ' '); + ++s; + } + } + + writer.write('-', '-', '>'); +} + +PUGI__FN void node_output_pi_value(xml_buffered_writer& writer, const char_t* s) +{ + while (*s) + { + const char_t* prev = s; + + // look for ?> sequence - we can't output it since ?> terminates PI + while (*s && !(s[0] == '?' && s[1] == '>')) + ++s; + + writer.write_buffer(prev, static_cast(s - prev)); + + if (*s) + { + assert(s[0] == '?' && s[1] == '>'); + + writer.write('?', ' ', '>'); + s += 2; + } + } +} + +PUGI__FN void node_output_attributes(xml_buffered_writer& writer, xml_node_struct* node, const char_t* indent, size_t indent_length, unsigned int flags, unsigned int depth) +{ + const char_t* default_name = PUGIXML_TEXT(":anonymous"); + + for (xml_attribute_struct* a = node->first_attribute; a; a = a->next_attribute) + { + if ((flags & (format_indent_attributes | format_raw)) == format_indent_attributes) + { + writer.write('\n'); + + text_output_indent(writer, indent, indent_length, depth + 1); + } + else + { + writer.write(' '); + } + + writer.write_string(a->name ? a->name + 0 : default_name); + writer.write('=', '"'); + + if (a->value) + text_output(writer, a->value, ctx_special_attr, flags); + + writer.write('"'); + } +} + +PUGI__FN bool node_output_start(xml_buffered_writer& writer, xml_node_struct* node, const char_t* indent, size_t indent_length, unsigned int flags, unsigned int depth) +{ + const char_t* default_name = PUGIXML_TEXT(":anonymous"); + const char_t* name = node->name ? node->name + 0 : default_name; + + writer.write('<'); + writer.write_string(name); + + if (node->first_attribute) + node_output_attributes(writer, node, indent, indent_length, flags, depth); + + // element nodes can have value if parse_embed_pcdata was used + if (!node->value) + { + if (!node->first_child) + { + if (flags & format_no_empty_element_tags) + { + writer.write('>', '<', '/'); writer.write_string(name); writer.write('>'); @@ -4144,5392 +4424,5688 @@ PUGI__NS_BEGIN } else { - return true; + if ((flags & format_raw) == 0) + writer.write(' '); + + writer.write('/', '>'); + + return false; } } - } - - PUGI__FN void node_output_end(xml_buffered_writer& writer, xml_node_struct* node) - { - const char_t* default_name = PUGIXML_TEXT(":anonymous"); - const char_t* name = node->name ? node->name + 0 : default_name; - - writer.write('<', '/'); - writer.write_string(name); - writer.write('>'); - } - - PUGI__FN void node_output_simple(xml_buffered_writer& writer, xml_node_struct* node, unsigned int flags) - { - const char_t* default_name = PUGIXML_TEXT(":anonymous"); - - switch (PUGI__NODETYPE(node)) + else { - case node_pcdata: - text_output(writer, node->value ? node->value + 0 : PUGIXML_TEXT(""), ctx_special_pcdata, flags); - break; + writer.write('>'); - case node_cdata: - text_output_cdata(writer, node->value ? node->value + 0 : PUGIXML_TEXT("")); - break; - - case node_comment: - node_output_comment(writer, node->value ? node->value + 0 : PUGIXML_TEXT("")); - break; - - case node_pi: - writer.write('<', '?'); - writer.write_string(node->name ? node->name + 0 : default_name); - - if (node->value) - { - writer.write(' '); - node_output_pi_value(writer, node->value); - } - - writer.write('?', '>'); - break; - - case node_declaration: - writer.write('<', '?'); - writer.write_string(node->name ? node->name + 0 : default_name); - node_output_attributes(writer, node, PUGIXML_TEXT(""), 0, flags | format_raw, 0); - writer.write('?', '>'); - break; - - case node_doctype: - writer.write('<', '!', 'D', 'O', 'C'); - writer.write('T', 'Y', 'P', 'E'); - - if (node->value) - { - writer.write(' '); - writer.write_string(node->value); - } - - writer.write('>'); - break; - - default: - assert(false && "Invalid node type"); // unreachable + return true; } } - - enum indent_flags_t + else { - indent_newline = 1, - indent_indent = 2 - }; + writer.write('>'); - PUGI__FN void node_output(xml_buffered_writer& writer, xml_node_struct* root, const char_t* indent, unsigned int flags, unsigned int depth) - { - size_t indent_length = ((flags & (format_indent | format_indent_attributes)) && (flags & format_raw) == 0) ? strlength(indent) : 0; - unsigned int indent_flags = indent_indent; + text_output(writer, node->value, ctx_special_pcdata, flags); - xml_node_struct* node = root; - - do + if (!node->first_child) { - assert(node); + writer.write('<', '/'); + writer.write_string(name); + writer.write('>'); - // begin writing current node - if (PUGI__NODETYPE(node) == node_pcdata || PUGI__NODETYPE(node) == node_cdata) + return false; + } + else + { + return true; + } + } +} + +PUGI__FN void node_output_end(xml_buffered_writer& writer, xml_node_struct* node) +{ + const char_t* default_name = PUGIXML_TEXT(":anonymous"); + const char_t* name = node->name ? node->name + 0 : default_name; + + writer.write('<', '/'); + writer.write_string(name); + writer.write('>'); +} + +PUGI__FN void node_output_simple(xml_buffered_writer& writer, xml_node_struct* node, unsigned int flags) +{ + const char_t* default_name = PUGIXML_TEXT(":anonymous"); + + switch (PUGI__NODETYPE(node)) + { + case node_pcdata: + text_output(writer, node->value ? node->value + 0 : PUGIXML_TEXT(""), ctx_special_pcdata, flags); + break; + + case node_cdata: + text_output_cdata(writer, node->value ? node->value + 0 : PUGIXML_TEXT("")); + break; + + case node_comment: + node_output_comment(writer, node->value ? node->value + 0 : PUGIXML_TEXT("")); + break; + + case node_pi: + writer.write('<', '?'); + writer.write_string(node->name ? node->name + 0 : default_name); + + if (node->value) + { + writer.write(' '); + node_output_pi_value(writer, node->value); + } + + writer.write('?', '>'); + break; + + case node_declaration: + writer.write('<', '?'); + writer.write_string(node->name ? node->name + 0 : default_name); + node_output_attributes(writer, node, PUGIXML_TEXT(""), 0, flags | format_raw, 0); + writer.write('?', '>'); + break; + + case node_doctype: + writer.write('<', '!', 'D', 'O', 'C'); + writer.write('T', 'Y', 'P', 'E'); + + if (node->value) + { + writer.write(' '); + writer.write_string(node->value); + } + + writer.write('>'); + break; + + default: + assert(false && "Invalid node type"); // unreachable + } +} + +enum indent_flags_t +{ + indent_newline = 1, + indent_indent = 2 +}; + +PUGI__FN void node_output(xml_buffered_writer& writer, xml_node_struct* root, const char_t* indent, unsigned int flags, unsigned int depth) +{ + size_t indent_length = ((flags & (format_indent | format_indent_attributes)) && (flags & format_raw) == 0) ? strlength(indent) : 0; + unsigned int indent_flags = indent_indent; + + xml_node_struct* node = root; + + do + { + assert(node); + + // begin writing current node + if (PUGI__NODETYPE(node) == node_pcdata || PUGI__NODETYPE(node) == node_cdata) + { + node_output_simple(writer, node, flags); + + indent_flags = 0; + } + else + { + if ((indent_flags & indent_newline) && (flags & format_raw) == 0) + writer.write('\n'); + + if ((indent_flags & indent_indent) && indent_length) + text_output_indent(writer, indent, indent_length, depth); + + if (PUGI__NODETYPE(node) == node_element) { - node_output_simple(writer, node, flags); + indent_flags = indent_newline | indent_indent; - indent_flags = 0; + if (node_output_start(writer, node, indent, indent_length, flags, depth)) + { + // element nodes can have value if parse_embed_pcdata was used + if (node->value) + indent_flags = 0; + + node = node->first_child; + depth++; + continue; + } + } + else if (PUGI__NODETYPE(node) == node_document) + { + indent_flags = indent_indent; + + if (node->first_child) + { + node = node->first_child; + continue; + } } else { + node_output_simple(writer, node, flags); + + indent_flags = indent_newline | indent_indent; + } + } + + // continue to the next node + while (node != root) + { + if (node->next_sibling) + { + node = node->next_sibling; + break; + } + + node = node->parent; + + // write closing node + if (PUGI__NODETYPE(node) == node_element) + { + depth--; + if ((indent_flags & indent_newline) && (flags & format_raw) == 0) writer.write('\n'); if ((indent_flags & indent_indent) && indent_length) text_output_indent(writer, indent, indent_length, depth); - if (PUGI__NODETYPE(node) == node_element) - { - indent_flags = indent_newline | indent_indent; + node_output_end(writer, node); - if (node_output_start(writer, node, indent, indent_length, flags, depth)) - { - // element nodes can have value if parse_embed_pcdata was used - if (node->value) - indent_flags = 0; - - node = node->first_child; - depth++; - continue; - } - } - else if (PUGI__NODETYPE(node) == node_document) - { - indent_flags = indent_indent; - - if (node->first_child) - { - node = node->first_child; - continue; - } - } - else - { - node_output_simple(writer, node, flags); - - indent_flags = indent_newline | indent_indent; - } - } - - // continue to the next node - while (node != root) - { - if (node->next_sibling) - { - node = node->next_sibling; - break; - } - - node = node->parent; - - // write closing node - if (PUGI__NODETYPE(node) == node_element) - { - depth--; - - if ((indent_flags & indent_newline) && (flags & format_raw) == 0) - writer.write('\n'); - - if ((indent_flags & indent_indent) && indent_length) - text_output_indent(writer, indent, indent_length, depth); - - node_output_end(writer, node); - - indent_flags = indent_newline | indent_indent; - } + indent_flags = indent_newline | indent_indent; } } - while (node != root); + } while (node != root); - if ((indent_flags & indent_newline) && (flags & format_raw) == 0) - writer.write('\n'); + if ((indent_flags & indent_newline) && (flags & format_raw) == 0) + writer.write('\n'); +} + +PUGI__FN bool has_declaration(xml_node_struct* node) +{ + for (xml_node_struct* child = node->first_child; child; child = child->next_sibling) + { + xml_node_type type = PUGI__NODETYPE(child); + + if (type == node_declaration) + return true; + if (type == node_element) + return false; } - PUGI__FN bool has_declaration(xml_node_struct* node) - { - for (xml_node_struct* child = node->first_child; child; child = child->next_sibling) - { - xml_node_type type = PUGI__NODETYPE(child); + return false; +} - if (type == node_declaration) return true; - if (type == node_element) return false; - } +PUGI__FN bool is_attribute_of(xml_attribute_struct* attr, xml_node_struct* node) +{ + for (xml_attribute_struct* a = node->first_attribute; a; a = a->next_attribute) + if (a == attr) + return true; + return false; +} + +PUGI__FN bool allow_insert_attribute(xml_node_type parent) +{ + return parent == node_element || parent == node_declaration; +} + +PUGI__FN bool allow_insert_child(xml_node_type parent, xml_node_type child) +{ + if (parent != node_document && parent != node_element) return false; - } - - PUGI__FN bool is_attribute_of(xml_attribute_struct* attr, xml_node_struct* node) - { - for (xml_attribute_struct* a = node->first_attribute; a; a = a->next_attribute) - if (a == attr) - return true; - + if (child == node_document || child == node_null) + return false; + if (parent != node_document && (child == node_declaration || child == node_doctype)) return false; - } - PUGI__FN bool allow_insert_attribute(xml_node_type parent) + return true; +} + +PUGI__FN bool allow_move(xml_node parent, xml_node child) +{ + // check that child can be a child of parent + if (!allow_insert_child(parent.type(), child.type())) + return false; + + // check that node is not moved between documents + if (parent.root() != child.root()) + return false; + + // check that new parent is not in the child subtree + xml_node cur = parent; + + while (cur) { - return parent == node_element || parent == node_declaration; - } - - PUGI__FN bool allow_insert_child(xml_node_type parent, xml_node_type child) - { - if (parent != node_document && parent != node_element) return false; - if (child == node_document || child == node_null) return false; - if (parent != node_document && (child == node_declaration || child == node_doctype)) return false; - - return true; - } - - PUGI__FN bool allow_move(xml_node parent, xml_node child) - { - // check that child can be a child of parent - if (!allow_insert_child(parent.type(), child.type())) + if (cur == child) return false; - // check that node is not moved between documents - if (parent.root() != child.root()) - return false; - - // check that new parent is not in the child subtree - xml_node cur = parent; - - while (cur) - { - if (cur == child) - return false; - - cur = cur.parent(); - } - - return true; + cur = cur.parent(); } - template - PUGI__FN void node_copy_string(String& dest, Header& header, uintptr_t header_mask, char_t* source, Header& source_header, xml_allocator* alloc) + return true; +} + +template +PUGI__FN void node_copy_string(String& dest, Header& header, uintptr_t header_mask, char_t* source, Header& source_header, xml_allocator* alloc) +{ + assert(!dest && (header & header_mask) == 0); + + if (source) { - assert(!dest && (header & header_mask) == 0); - - if (source) + if (alloc && (source_header & header_mask) == 0) { - if (alloc && (source_header & header_mask) == 0) - { - dest = source; + dest = source; - // since strcpy_insitu can reuse document buffer memory we need to mark both source and dest as shared - header |= xml_memory_page_contents_shared_mask; - source_header |= xml_memory_page_contents_shared_mask; - } - else - strcpy_insitu(dest, header, header_mask, source, strlength(source)); + // since strcpy_insitu can reuse document buffer memory we need to mark both source and dest as shared + header |= xml_memory_page_contents_shared_mask; + source_header |= xml_memory_page_contents_shared_mask; + } + else + strcpy_insitu(dest, header, header_mask, source, strlength(source)); + } +} + +PUGI__FN void node_copy_contents(xml_node_struct* dn, xml_node_struct* sn, xml_allocator* shared_alloc) +{ + node_copy_string(dn->name, dn->header, xml_memory_page_name_allocated_mask, sn->name, sn->header, shared_alloc); + node_copy_string(dn->value, dn->header, xml_memory_page_value_allocated_mask, sn->value, sn->header, shared_alloc); + + for (xml_attribute_struct* sa = sn->first_attribute; sa; sa = sa->next_attribute) + { + xml_attribute_struct* da = append_new_attribute(dn, get_allocator(dn)); + + if (da) + { + node_copy_string(da->name, da->header, xml_memory_page_name_allocated_mask, sa->name, sa->header, shared_alloc); + node_copy_string(da->value, da->header, xml_memory_page_value_allocated_mask, sa->value, sa->header, shared_alloc); } } +} - PUGI__FN void node_copy_contents(xml_node_struct* dn, xml_node_struct* sn, xml_allocator* shared_alloc) +PUGI__FN void node_copy_tree(xml_node_struct* dn, xml_node_struct* sn) +{ + xml_allocator& alloc = get_allocator(dn); + xml_allocator* shared_alloc = (&alloc == &get_allocator(sn)) ? &alloc : 0; + + node_copy_contents(dn, sn, shared_alloc); + + xml_node_struct* dit = dn; + xml_node_struct* sit = sn->first_child; + + while (sit && sit != sn) { - node_copy_string(dn->name, dn->header, xml_memory_page_name_allocated_mask, sn->name, sn->header, shared_alloc); - node_copy_string(dn->value, dn->header, xml_memory_page_value_allocated_mask, sn->value, sn->header, shared_alloc); - - for (xml_attribute_struct* sa = sn->first_attribute; sa; sa = sa->next_attribute) + // when a tree is copied into one of the descendants, we need to skip that subtree to avoid an infinite loop + if (sit != dn) { - xml_attribute_struct* da = append_new_attribute(dn, get_allocator(dn)); + xml_node_struct* copy = append_new_node(dit, alloc, PUGI__NODETYPE(sit)); - if (da) + if (copy) { - node_copy_string(da->name, da->header, xml_memory_page_name_allocated_mask, sa->name, sa->header, shared_alloc); - node_copy_string(da->value, da->header, xml_memory_page_value_allocated_mask, sa->value, sa->header, shared_alloc); - } - } - } + node_copy_contents(copy, sit, shared_alloc); - PUGI__FN void node_copy_tree(xml_node_struct* dn, xml_node_struct* sn) - { - xml_allocator& alloc = get_allocator(dn); - xml_allocator* shared_alloc = (&alloc == &get_allocator(sn)) ? &alloc : 0; - - node_copy_contents(dn, sn, shared_alloc); - - xml_node_struct* dit = dn; - xml_node_struct* sit = sn->first_child; - - while (sit && sit != sn) - { - // when a tree is copied into one of the descendants, we need to skip that subtree to avoid an infinite loop - if (sit != dn) - { - xml_node_struct* copy = append_new_node(dit, alloc, PUGI__NODETYPE(sit)); - - if (copy) + if (sit->first_child) { - node_copy_contents(copy, sit, shared_alloc); - - if (sit->first_child) - { - dit = copy; - sit = sit->first_child; - continue; - } + dit = copy; + sit = sit->first_child; + continue; } } - - // continue to the next node - do - { - if (sit->next_sibling) - { - sit = sit->next_sibling; - break; - } - - sit = sit->parent; - dit = dit->parent; - } - while (sit != sn); } + + // continue to the next node + do + { + if (sit->next_sibling) + { + sit = sit->next_sibling; + break; + } + + sit = sit->parent; + dit = dit->parent; + } while (sit != sn); } +} - PUGI__FN void node_copy_attribute(xml_attribute_struct* da, xml_attribute_struct* sa) +PUGI__FN void node_copy_attribute(xml_attribute_struct* da, xml_attribute_struct* sa) +{ + xml_allocator& alloc = get_allocator(da); + xml_allocator* shared_alloc = (&alloc == &get_allocator(sa)) ? &alloc : 0; + + node_copy_string(da->name, da->header, xml_memory_page_name_allocated_mask, sa->name, sa->header, shared_alloc); + node_copy_string(da->value, da->header, xml_memory_page_value_allocated_mask, sa->value, sa->header, shared_alloc); +} + +inline bool is_text_node(xml_node_struct* node) +{ + xml_node_type type = PUGI__NODETYPE(node); + + return type == node_pcdata || type == node_cdata; +} + +// get value with conversion functions +template +PUGI__FN PUGI__UNSIGNED_OVERFLOW U string_to_integer(const char_t* value, U minv, U maxv) +{ + U result = 0; + const char_t* s = value; + + while (PUGI__IS_CHARTYPE(*s, ct_space)) + s++; + + bool negative = (*s == '-'); + + s += (*s == '+' || *s == '-'); + + bool overflow = false; + + if (s[0] == '0' && (s[1] | ' ') == 'x') { - xml_allocator& alloc = get_allocator(da); - xml_allocator* shared_alloc = (&alloc == &get_allocator(sa)) ? &alloc : 0; + s += 2; - node_copy_string(da->name, da->header, xml_memory_page_name_allocated_mask, sa->name, sa->header, shared_alloc); - node_copy_string(da->value, da->header, xml_memory_page_value_allocated_mask, sa->value, sa->header, shared_alloc); - } - - inline bool is_text_node(xml_node_struct* node) - { - xml_node_type type = PUGI__NODETYPE(node); - - return type == node_pcdata || type == node_cdata; - } - - // get value with conversion functions - template PUGI__FN PUGI__UNSIGNED_OVERFLOW U string_to_integer(const char_t* value, U minv, U maxv) - { - U result = 0; - const char_t* s = value; - - while (PUGI__IS_CHARTYPE(*s, ct_space)) + // since overflow detection relies on length of the sequence skip leading zeros + while (*s == '0') s++; - bool negative = (*s == '-'); + const char_t* start = s; - s += (*s == '+' || *s == '-'); - - bool overflow = false; - - if (s[0] == '0' && (s[1] | ' ') == 'x') + for (;;) { - s += 2; + if (static_cast(*s - '0') < 10) + result = result * 16 + (*s - '0'); + else if (static_cast((*s | ' ') - 'a') < 6) + result = result * 16 + ((*s | ' ') - 'a' + 10); + else + break; - // since overflow detection relies on length of the sequence skip leading zeros - while (*s == '0') - s++; - - const char_t* start = s; - - for (;;) - { - if (static_cast(*s - '0') < 10) - result = result * 16 + (*s - '0'); - else if (static_cast((*s | ' ') - 'a') < 6) - result = result * 16 + ((*s | ' ') - 'a' + 10); - else - break; - - s++; - } - - size_t digits = static_cast(s - start); - - overflow = digits > sizeof(U) * 2; - } - else - { - // since overflow detection relies on length of the sequence skip leading zeros - while (*s == '0') - s++; - - const char_t* start = s; - - for (;;) - { - if (static_cast(*s - '0') < 10) - result = result * 10 + (*s - '0'); - else - break; - - s++; - } - - size_t digits = static_cast(s - start); - - PUGI__STATIC_ASSERT(sizeof(U) == 8 || sizeof(U) == 4 || sizeof(U) == 2); - - const size_t max_digits10 = sizeof(U) == 8 ? 20 : sizeof(U) == 4 ? 10 : 5; - const char_t max_lead = sizeof(U) == 8 ? '1' : sizeof(U) == 4 ? '4' : '6'; - const size_t high_bit = sizeof(U) * 8 - 1; - - overflow = digits >= max_digits10 && !(digits == max_digits10 && (*start < max_lead || (*start == max_lead && result >> high_bit))); + s++; } - if (negative) + size_t digits = static_cast(s - start); + + overflow = digits > sizeof(U) * 2; + } + else + { + // since overflow detection relies on length of the sequence skip leading zeros + while (*s == '0') + s++; + + const char_t* start = s; + + for (;;) { - // Workaround for crayc++ CC-3059: Expected no overflow in routine. - #ifdef _CRAYC - return (overflow || result > ~minv + 1) ? minv : ~result + 1; - #else - return (overflow || result > 0 - minv) ? minv : 0 - result; - #endif + if (static_cast(*s - '0') < 10) + result = result * 10 + (*s - '0'); + else + break; + + s++; } - else - return (overflow || result > maxv) ? maxv : result; + + size_t digits = static_cast(s - start); + + PUGI__STATIC_ASSERT(sizeof(U) == 8 || sizeof(U) == 4 || sizeof(U) == 2); + + const size_t max_digits10 = sizeof(U) == 8 ? 20 : sizeof(U) == 4 ? 10 : 5; + const char_t max_lead = sizeof(U) == 8 ? '1' : sizeof(U) == 4 ? '4' : '6'; + const size_t high_bit = sizeof(U) * 8 - 1; + + overflow = digits >= max_digits10 && !(digits == max_digits10 && (*start < max_lead || (*start == max_lead && result >> high_bit))); } - PUGI__FN int get_value_int(const char_t* value) + if (negative) { - return string_to_integer(value, static_cast(INT_MIN), INT_MAX); + // Workaround for crayc++ CC-3059: Expected no overflow in routine. +#ifdef _CRAYC + return (overflow || result > ~minv + 1) ? minv : ~result + 1; +#else + return (overflow || result > 0 - minv) ? minv : 0 - result; +#endif } + else + return (overflow || result > maxv) ? maxv : result; +} - PUGI__FN unsigned int get_value_uint(const char_t* value) - { - return string_to_integer(value, 0, UINT_MAX); - } +PUGI__FN int get_value_int(const char_t* value) +{ + return string_to_integer(value, static_cast(INT_MIN), INT_MAX); +} - PUGI__FN double get_value_double(const char_t* value) - { - #ifdef PUGIXML_WCHAR_MODE - return wcstod(value, 0); - #else - return strtod(value, 0); - #endif - } +PUGI__FN unsigned int get_value_uint(const char_t* value) +{ + return string_to_integer(value, 0, UINT_MAX); +} - PUGI__FN float get_value_float(const char_t* value) - { - #ifdef PUGIXML_WCHAR_MODE - return static_cast(wcstod(value, 0)); - #else - return static_cast(strtod(value, 0)); - #endif - } +PUGI__FN double get_value_double(const char_t* value) +{ +#ifdef PUGIXML_WCHAR_MODE + return wcstod(value, 0); +#else + return strtod(value, 0); +#endif +} - PUGI__FN bool get_value_bool(const char_t* value) - { - // only look at first char - char_t first = *value; +PUGI__FN float get_value_float(const char_t* value) +{ +#ifdef PUGIXML_WCHAR_MODE + return static_cast(wcstod(value, 0)); +#else + return static_cast(strtod(value, 0)); +#endif +} - // 1*, t* (true), T* (True), y* (yes), Y* (YES) - return (first == '1' || first == 't' || first == 'T' || first == 'y' || first == 'Y'); - } +PUGI__FN bool get_value_bool(const char_t* value) +{ + // only look at first char + char_t first = *value; + + // 1*, t* (true), T* (True), y* (yes), Y* (YES) + return (first == '1' || first == 't' || first == 'T' || first == 'y' || first == 'Y'); +} #ifdef PUGIXML_HAS_LONG_LONG - PUGI__FN long long get_value_llong(const char_t* value) - { - return string_to_integer(value, static_cast(LLONG_MIN), LLONG_MAX); - } +PUGI__FN long long get_value_llong(const char_t* value) +{ + return string_to_integer(value, static_cast(LLONG_MIN), LLONG_MAX); +} - PUGI__FN unsigned long long get_value_ullong(const char_t* value) +PUGI__FN unsigned long long get_value_ullong(const char_t* value) +{ + return string_to_integer(value, 0, ULLONG_MAX); +} +#endif + +template +PUGI__FN PUGI__UNSIGNED_OVERFLOW char_t* integer_to_string(char_t* begin, char_t* end, U value, bool negative) +{ + char_t* result = end - 1; + U rest = negative ? 0 - value : value; + + do { - return string_to_integer(value, 0, ULLONG_MAX); + *result-- = static_cast('0' + (rest % 10)); + rest /= 10; + } while (rest); + + assert(result >= begin); + (void)begin; + + *result = '-'; + + return result + !negative; +} + +// set value with conversion functions +template +PUGI__FN bool set_value_ascii(String& dest, Header& header, uintptr_t header_mask, char* buf) +{ +#ifdef PUGIXML_WCHAR_MODE + char_t wbuf[128]; + assert(strlen(buf) < sizeof(wbuf) / sizeof(wbuf[0])); + + size_t offset = 0; + for (; buf[offset]; ++offset) + wbuf[offset] = buf[offset]; + + return strcpy_insitu(dest, header, header_mask, wbuf, offset); +#else + return strcpy_insitu(dest, header, header_mask, buf, strlen(buf)); +#endif +} + +template +PUGI__FN bool set_value_integer(String& dest, Header& header, uintptr_t header_mask, U value, bool negative) +{ + char_t buf[64]; + char_t* end = buf + sizeof(buf) / sizeof(buf[0]); + char_t* begin = integer_to_string(buf, end, value, negative); + + return strcpy_insitu(dest, header, header_mask, begin, end - begin); +} + +template +PUGI__FN bool set_value_convert(String& dest, Header& header, uintptr_t header_mask, float value) +{ + char buf[128]; + PUGI__SNPRINTF(buf, "%.9g", value); + + return set_value_ascii(dest, header, header_mask, buf); +} + +template +PUGI__FN bool set_value_convert(String& dest, Header& header, uintptr_t header_mask, double value) +{ + char buf[128]; + PUGI__SNPRINTF(buf, "%.17g", value); + + return set_value_ascii(dest, header, header_mask, buf); +} + +template +PUGI__FN bool set_value_bool(String& dest, Header& header, uintptr_t header_mask, bool value) +{ + return strcpy_insitu(dest, header, header_mask, value ? PUGIXML_TEXT("true") : PUGIXML_TEXT("false"), value ? 4 : 5); +} + +PUGI__FN xml_parse_result load_buffer_impl(xml_document_struct* doc, xml_node_struct* root, void* contents, size_t size, unsigned int options, xml_encoding encoding, bool is_mutable, bool own, char_t** out_buffer) +{ + // check input buffer + if (!contents && size) + return make_parse_result(status_io_error); + + // get actual encoding + xml_encoding buffer_encoding = impl::get_buffer_encoding(encoding, contents, size); + + // get private buffer + char_t* buffer = 0; + size_t length = 0; + + if (!impl::convert_buffer(buffer, length, buffer_encoding, contents, size, is_mutable)) + return impl::make_parse_result(status_out_of_memory); + + // delete original buffer if we performed a conversion + if (own && buffer != contents && contents) + impl::xml_memory::deallocate(contents); + + // grab onto buffer if it's our buffer, user is responsible for deallocating contents himself + if (own || buffer != contents) + *out_buffer = buffer; + + // store buffer for offset_debug + doc->buffer = buffer; + + // parse + xml_parse_result res = impl::xml_parser::parse(buffer, length, doc, root, options); + + // remember encoding + res.encoding = buffer_encoding; + + return res; +} + +// we need to get length of entire file to load it in memory; the only (relatively) sane way to do it is via seek/tell trick +PUGI__FN xml_parse_status get_file_size(FILE* file, size_t& out_result) +{ +#if defined(PUGI__MSVC_CRT_VERSION) && PUGI__MSVC_CRT_VERSION >= 1400 && !defined(_WIN32_WCE) + // there are 64-bit versions of fseek/ftell, let's use them + typedef __int64 length_type; + + _fseeki64(file, 0, SEEK_END); + length_type length = _ftelli64(file); + _fseeki64(file, 0, SEEK_SET); +#elif defined(__MINGW32__) && !defined(__NO_MINGW_LFS) && (!defined(__STRICT_ANSI__) || defined(__MINGW64_VERSION_MAJOR)) + // there are 64-bit versions of fseek/ftell, let's use them + typedef off64_t length_type; + + fseeko64(file, 0, SEEK_END); + length_type length = ftello64(file); + fseeko64(file, 0, SEEK_SET); +#else + // if this is a 32-bit OS, long is enough; if this is a unix system, long is 64-bit, which is enough; otherwise we can't do anything anyway. + typedef long length_type; + + fseek(file, 0, SEEK_END); + length_type length = ftell(file); + fseek(file, 0, SEEK_SET); +#endif + + // check for I/O errors + if (length < 0) + return status_io_error; + + // check for overflow + size_t result = static_cast(length); + + if (static_cast(result) != length) + return status_out_of_memory; + + // finalize + out_result = result; + + return status_ok; +} + +// This function assumes that buffer has extra sizeof(char_t) writable bytes after size +PUGI__FN size_t zero_terminate_buffer(void* buffer, size_t size, xml_encoding encoding) +{ + // We only need to zero-terminate if encoding conversion does not do it for us +#ifdef PUGIXML_WCHAR_MODE + xml_encoding wchar_encoding = get_wchar_encoding(); + + if (encoding == wchar_encoding || need_endian_swap_utf(encoding, wchar_encoding)) + { + size_t length = size / sizeof(char_t); + + static_cast(buffer)[length] = 0; + return (length + 1) * sizeof(char_t); + } +#else + if (encoding == encoding_utf8) + { + static_cast(buffer)[size] = 0; + return size + 1; } #endif - template PUGI__FN PUGI__UNSIGNED_OVERFLOW char_t* integer_to_string(char_t* begin, char_t* end, U value, bool negative) + return size; +} + +PUGI__FN xml_parse_result load_file_impl(xml_document_struct* doc, FILE* file, unsigned int options, xml_encoding encoding, char_t** out_buffer) +{ + if (!file) + return make_parse_result(status_file_not_found); + + // get file size (can result in I/O errors) + size_t size = 0; + xml_parse_status size_status = get_file_size(file, size); + if (size_status != status_ok) + return make_parse_result(size_status); + + size_t max_suffix_size = sizeof(char_t); + + // allocate buffer for the whole file + char* contents = static_cast(xml_memory::allocate(size + max_suffix_size)); + if (!contents) + return make_parse_result(status_out_of_memory); + + // read file in memory + size_t read_size = fread(contents, 1, size, file); + + if (read_size != size) { - char_t* result = end - 1; - U rest = negative ? 0 - value : value; - - do - { - *result-- = static_cast('0' + (rest % 10)); - rest /= 10; - } - while (rest); - - assert(result >= begin); - (void)begin; - - *result = '-'; - - return result + !negative; + xml_memory::deallocate(contents); + return make_parse_result(status_io_error); } - // set value with conversion functions - template - PUGI__FN bool set_value_ascii(String& dest, Header& header, uintptr_t header_mask, char* buf) - { - #ifdef PUGIXML_WCHAR_MODE - char_t wbuf[128]; - assert(strlen(buf) < sizeof(wbuf) / sizeof(wbuf[0])); + xml_encoding real_encoding = get_buffer_encoding(encoding, contents, size); - size_t offset = 0; - for (; buf[offset]; ++offset) wbuf[offset] = buf[offset]; + return load_buffer_impl(doc, doc, contents, zero_terminate_buffer(contents, size, real_encoding), options, real_encoding, true, true, out_buffer); +} - return strcpy_insitu(dest, header, header_mask, wbuf, offset); - #else - return strcpy_insitu(dest, header, header_mask, buf, strlen(buf)); - #endif - } - - template - PUGI__FN bool set_value_integer(String& dest, Header& header, uintptr_t header_mask, U value, bool negative) - { - char_t buf[64]; - char_t* end = buf + sizeof(buf) / sizeof(buf[0]); - char_t* begin = integer_to_string(buf, end, value, negative); - - return strcpy_insitu(dest, header, header_mask, begin, end - begin); - } - - template - PUGI__FN bool set_value_convert(String& dest, Header& header, uintptr_t header_mask, float value) - { - char buf[128]; - PUGI__SNPRINTF(buf, "%.9g", value); - - return set_value_ascii(dest, header, header_mask, buf); - } - - template - PUGI__FN bool set_value_convert(String& dest, Header& header, uintptr_t header_mask, double value) - { - char buf[128]; - PUGI__SNPRINTF(buf, "%.17g", value); - - return set_value_ascii(dest, header, header_mask, buf); - } - - template - PUGI__FN bool set_value_bool(String& dest, Header& header, uintptr_t header_mask, bool value) - { - return strcpy_insitu(dest, header, header_mask, value ? PUGIXML_TEXT("true") : PUGIXML_TEXT("false"), value ? 4 : 5); - } - - PUGI__FN xml_parse_result load_buffer_impl(xml_document_struct* doc, xml_node_struct* root, void* contents, size_t size, unsigned int options, xml_encoding encoding, bool is_mutable, bool own, char_t** out_buffer) - { - // check input buffer - if (!contents && size) return make_parse_result(status_io_error); - - // get actual encoding - xml_encoding buffer_encoding = impl::get_buffer_encoding(encoding, contents, size); - - // get private buffer - char_t* buffer = 0; - size_t length = 0; - - if (!impl::convert_buffer(buffer, length, buffer_encoding, contents, size, is_mutable)) return impl::make_parse_result(status_out_of_memory); - - // delete original buffer if we performed a conversion - if (own && buffer != contents && contents) impl::xml_memory::deallocate(contents); - - // grab onto buffer if it's our buffer, user is responsible for deallocating contents himself - if (own || buffer != contents) *out_buffer = buffer; - - // store buffer for offset_debug - doc->buffer = buffer; - - // parse - xml_parse_result res = impl::xml_parser::parse(buffer, length, doc, root, options); - - // remember encoding - res.encoding = buffer_encoding; - - return res; - } - - // we need to get length of entire file to load it in memory; the only (relatively) sane way to do it is via seek/tell trick - PUGI__FN xml_parse_status get_file_size(FILE* file, size_t& out_result) - { - #if defined(PUGI__MSVC_CRT_VERSION) && PUGI__MSVC_CRT_VERSION >= 1400 && !defined(_WIN32_WCE) - // there are 64-bit versions of fseek/ftell, let's use them - typedef __int64 length_type; - - _fseeki64(file, 0, SEEK_END); - length_type length = _ftelli64(file); - _fseeki64(file, 0, SEEK_SET); - #elif defined(__MINGW32__) && !defined(__NO_MINGW_LFS) && (!defined(__STRICT_ANSI__) || defined(__MINGW64_VERSION_MAJOR)) - // there are 64-bit versions of fseek/ftell, let's use them - typedef off64_t length_type; - - fseeko64(file, 0, SEEK_END); - length_type length = ftello64(file); - fseeko64(file, 0, SEEK_SET); - #else - // if this is a 32-bit OS, long is enough; if this is a unix system, long is 64-bit, which is enough; otherwise we can't do anything anyway. - typedef long length_type; - - fseek(file, 0, SEEK_END); - length_type length = ftell(file); - fseek(file, 0, SEEK_SET); - #endif - - // check for I/O errors - if (length < 0) return status_io_error; - - // check for overflow - size_t result = static_cast(length); - - if (static_cast(result) != length) return status_out_of_memory; - - // finalize - out_result = result; - - return status_ok; - } - - // This function assumes that buffer has extra sizeof(char_t) writable bytes after size - PUGI__FN size_t zero_terminate_buffer(void* buffer, size_t size, xml_encoding encoding) - { - // We only need to zero-terminate if encoding conversion does not do it for us - #ifdef PUGIXML_WCHAR_MODE - xml_encoding wchar_encoding = get_wchar_encoding(); - - if (encoding == wchar_encoding || need_endian_swap_utf(encoding, wchar_encoding)) - { - size_t length = size / sizeof(char_t); - - static_cast(buffer)[length] = 0; - return (length + 1) * sizeof(char_t); - } - #else - if (encoding == encoding_utf8) - { - static_cast(buffer)[size] = 0; - return size + 1; - } - #endif - - return size; - } - - PUGI__FN xml_parse_result load_file_impl(xml_document_struct* doc, FILE* file, unsigned int options, xml_encoding encoding, char_t** out_buffer) - { - if (!file) return make_parse_result(status_file_not_found); - - // get file size (can result in I/O errors) - size_t size = 0; - xml_parse_status size_status = get_file_size(file, size); - if (size_status != status_ok) return make_parse_result(size_status); - - size_t max_suffix_size = sizeof(char_t); - - // allocate buffer for the whole file - char* contents = static_cast(xml_memory::allocate(size + max_suffix_size)); - if (!contents) return make_parse_result(status_out_of_memory); - - // read file in memory - size_t read_size = fread(contents, 1, size, file); - - if (read_size != size) - { - xml_memory::deallocate(contents); - return make_parse_result(status_io_error); - } - - xml_encoding real_encoding = get_buffer_encoding(encoding, contents, size); - - return load_buffer_impl(doc, doc, contents, zero_terminate_buffer(contents, size, real_encoding), options, real_encoding, true, true, out_buffer); - } - - PUGI__FN void close_file(FILE* file) - { - fclose(file); - } +PUGI__FN void close_file(FILE* file) +{ + fclose(file); +} #ifndef PUGIXML_NO_STL - template struct xml_stream_chunk +template +struct xml_stream_chunk +{ + static xml_stream_chunk* create() { - static xml_stream_chunk* create() - { - void* memory = xml_memory::allocate(sizeof(xml_stream_chunk)); - if (!memory) return 0; + void* memory = xml_memory::allocate(sizeof(xml_stream_chunk)); + if (!memory) + return 0; - return new (memory) xml_stream_chunk(); - } - - static void destroy(xml_stream_chunk* chunk) - { - // free chunk chain - while (chunk) - { - xml_stream_chunk* next_ = chunk->next; - - xml_memory::deallocate(chunk); - - chunk = next_; - } - } - - xml_stream_chunk(): next(0), size(0) - { - } - - xml_stream_chunk* next; - size_t size; - - T data[xml_memory_page_size / sizeof(T)]; - }; - - template PUGI__FN xml_parse_status load_stream_data_noseek(std::basic_istream& stream, void** out_buffer, size_t* out_size) - { - auto_deleter > chunks(0, xml_stream_chunk::destroy); - - // read file to a chunk list - size_t total = 0; - xml_stream_chunk* last = 0; - - while (!stream.eof()) - { - // allocate new chunk - xml_stream_chunk* chunk = xml_stream_chunk::create(); - if (!chunk) return status_out_of_memory; - - // append chunk to list - if (last) last = last->next = chunk; - else chunks.data = last = chunk; - - // read data to chunk - stream.read(chunk->data, static_cast(sizeof(chunk->data) / sizeof(T))); - chunk->size = static_cast(stream.gcount()) * sizeof(T); - - // read may set failbit | eofbit in case gcount() is less than read length, so check for other I/O errors - if (stream.bad() || (!stream.eof() && stream.fail())) return status_io_error; - - // guard against huge files (chunk size is small enough to make this overflow check work) - if (total + chunk->size < total) return status_out_of_memory; - total += chunk->size; - } - - size_t max_suffix_size = sizeof(char_t); - - // copy chunk list to a contiguous buffer - char* buffer = static_cast(xml_memory::allocate(total + max_suffix_size)); - if (!buffer) return status_out_of_memory; - - char* write = buffer; - - for (xml_stream_chunk* chunk = chunks.data; chunk; chunk = chunk->next) - { - assert(write + chunk->size <= buffer + total); - memcpy(write, chunk->data, chunk->size); - write += chunk->size; - } - - assert(write == buffer + total); - - // return buffer - *out_buffer = buffer; - *out_size = total; - - return status_ok; + return new (memory) xml_stream_chunk(); } - template PUGI__FN xml_parse_status load_stream_data_seek(std::basic_istream& stream, void** out_buffer, size_t* out_size) + static void destroy(xml_stream_chunk* chunk) { - // get length of remaining data in stream - typename std::basic_istream::pos_type pos = stream.tellg(); - stream.seekg(0, std::ios::end); - std::streamoff length = stream.tellg() - pos; - stream.seekg(pos); + // free chunk chain + while (chunk) + { + xml_stream_chunk* next_ = chunk->next; - if (stream.fail() || pos < 0) return status_io_error; + xml_memory::deallocate(chunk); - // guard against huge files - size_t read_length = static_cast(length); - - if (static_cast(read_length) != length || length < 0) return status_out_of_memory; - - size_t max_suffix_size = sizeof(char_t); - - // read stream data into memory (guard against stream exceptions with buffer holder) - auto_deleter buffer(xml_memory::allocate(read_length * sizeof(T) + max_suffix_size), xml_memory::deallocate); - if (!buffer.data) return status_out_of_memory; - - stream.read(static_cast(buffer.data), static_cast(read_length)); - - // read may set failbit | eofbit in case gcount() is less than read_length (i.e. line ending conversion), so check for other I/O errors - if (stream.bad() || (!stream.eof() && stream.fail())) return status_io_error; - - // return buffer - size_t actual_length = static_cast(stream.gcount()); - assert(actual_length <= read_length); - - *out_buffer = buffer.release(); - *out_size = actual_length * sizeof(T); - - return status_ok; + chunk = next_; + } } - template PUGI__FN xml_parse_result load_stream_impl(xml_document_struct* doc, std::basic_istream& stream, unsigned int options, xml_encoding encoding, char_t** out_buffer) + xml_stream_chunk() + : next(0) + , size(0) { - void* buffer = 0; - size_t size = 0; - xml_parse_status status = status_ok; + } - // if stream has an error bit set, bail out (otherwise tellg() can fail and we'll clear error bits) - if (stream.fail()) return make_parse_result(status_io_error); + xml_stream_chunk* next; + size_t size; - // load stream to memory (using seek-based implementation if possible, since it's faster and takes less memory) - if (stream.tellg() < 0) - { - stream.clear(); // clear error flags that could be set by a failing tellg - status = load_stream_data_noseek(stream, &buffer, &size); - } + T data[xml_memory_page_size / sizeof(T)]; +}; + +template +PUGI__FN xml_parse_status load_stream_data_noseek(std::basic_istream& stream, void** out_buffer, size_t* out_size) +{ + auto_deleter> chunks(0, xml_stream_chunk::destroy); + + // read file to a chunk list + size_t total = 0; + xml_stream_chunk* last = 0; + + while (!stream.eof()) + { + // allocate new chunk + xml_stream_chunk* chunk = xml_stream_chunk::create(); + if (!chunk) + return status_out_of_memory; + + // append chunk to list + if (last) + last = last->next = chunk; else - status = load_stream_data_seek(stream, &buffer, &size); + chunks.data = last = chunk; - if (status != status_ok) return make_parse_result(status); + // read data to chunk + stream.read(chunk->data, static_cast(sizeof(chunk->data) / sizeof(T))); + chunk->size = static_cast(stream.gcount()) * sizeof(T); - xml_encoding real_encoding = get_buffer_encoding(encoding, buffer, size); + // read may set failbit | eofbit in case gcount() is less than read length, so check for other I/O errors + if (stream.bad() || (!stream.eof() && stream.fail())) + return status_io_error; - return load_buffer_impl(doc, doc, buffer, zero_terminate_buffer(buffer, size, real_encoding), options, real_encoding, true, true, out_buffer); + // guard against huge files (chunk size is small enough to make this overflow check work) + if (total + chunk->size < total) + return status_out_of_memory; + total += chunk->size; } + + size_t max_suffix_size = sizeof(char_t); + + // copy chunk list to a contiguous buffer + char* buffer = static_cast(xml_memory::allocate(total + max_suffix_size)); + if (!buffer) + return status_out_of_memory; + + char* write = buffer; + + for (xml_stream_chunk* chunk = chunks.data; chunk; chunk = chunk->next) + { + assert(write + chunk->size <= buffer + total); + memcpy(write, chunk->data, chunk->size); + write += chunk->size; + } + + assert(write == buffer + total); + + // return buffer + *out_buffer = buffer; + *out_size = total; + + return status_ok; +} + +template +PUGI__FN xml_parse_status load_stream_data_seek(std::basic_istream& stream, void** out_buffer, size_t* out_size) +{ + // get length of remaining data in stream + typename std::basic_istream::pos_type pos = stream.tellg(); + stream.seekg(0, std::ios::end); + std::streamoff length = stream.tellg() - pos; + stream.seekg(pos); + + if (stream.fail() || pos < 0) + return status_io_error; + + // guard against huge files + size_t read_length = static_cast(length); + + if (static_cast(read_length) != length || length < 0) + return status_out_of_memory; + + size_t max_suffix_size = sizeof(char_t); + + // read stream data into memory (guard against stream exceptions with buffer holder) + auto_deleter buffer(xml_memory::allocate(read_length * sizeof(T) + max_suffix_size), xml_memory::deallocate); + if (!buffer.data) + return status_out_of_memory; + + stream.read(static_cast(buffer.data), static_cast(read_length)); + + // read may set failbit | eofbit in case gcount() is less than read_length (i.e. line ending conversion), so check for other I/O errors + if (stream.bad() || (!stream.eof() && stream.fail())) + return status_io_error; + + // return buffer + size_t actual_length = static_cast(stream.gcount()); + assert(actual_length <= read_length); + + *out_buffer = buffer.release(); + *out_size = actual_length * sizeof(T); + + return status_ok; +} + +template +PUGI__FN xml_parse_result load_stream_impl(xml_document_struct* doc, std::basic_istream& stream, unsigned int options, xml_encoding encoding, char_t** out_buffer) +{ + void* buffer = 0; + size_t size = 0; + xml_parse_status status = status_ok; + + // if stream has an error bit set, bail out (otherwise tellg() can fail and we'll clear error bits) + if (stream.fail()) + return make_parse_result(status_io_error); + + // load stream to memory (using seek-based implementation if possible, since it's faster and takes less memory) + if (stream.tellg() < 0) + { + stream.clear(); // clear error flags that could be set by a failing tellg + status = load_stream_data_noseek(stream, &buffer, &size); + } + else + status = load_stream_data_seek(stream, &buffer, &size); + + if (status != status_ok) + return make_parse_result(status); + + xml_encoding real_encoding = get_buffer_encoding(encoding, buffer, size); + + return load_buffer_impl(doc, doc, buffer, zero_terminate_buffer(buffer, size, real_encoding), options, real_encoding, true, true, out_buffer); +} #endif #if defined(PUGI__MSVC_CRT_VERSION) || defined(__BORLANDC__) || (defined(__MINGW32__) && (!defined(__STRICT_ANSI__) || defined(__MINGW64_VERSION_MAJOR))) - PUGI__FN FILE* open_file_wide(const wchar_t* path, const wchar_t* mode) - { - return _wfopen(path, mode); - } +PUGI__FN FILE* open_file_wide(const wchar_t* path, const wchar_t* mode) +{ + return _wfopen(path, mode); +} #else - PUGI__FN char* convert_path_heap(const wchar_t* str) - { - assert(str); +PUGI__FN char* convert_path_heap(const wchar_t* str) +{ + assert(str); - // first pass: get length in utf8 characters - size_t length = strlength_wide(str); - size_t size = as_utf8_begin(str, length); + // first pass: get length in utf8 characters + size_t length = strlength_wide(str); + size_t size = as_utf8_begin(str, length); - // allocate resulting string - char* result = static_cast(xml_memory::allocate(size + 1)); - if (!result) return 0; + // allocate resulting string + char* result = static_cast(xml_memory::allocate(size + 1)); + if (!result) + return 0; - // second pass: convert to utf8 - as_utf8_end(result, size, str, length); + // second pass: convert to utf8 + as_utf8_end(result, size, str, length); - // zero-terminate - result[size] = 0; + // zero-terminate + result[size] = 0; - return result; - } + return result; +} - PUGI__FN FILE* open_file_wide(const wchar_t* path, const wchar_t* mode) - { - // there is no standard function to open wide paths, so our best bet is to try utf8 path - char* path_utf8 = convert_path_heap(path); - if (!path_utf8) return 0; +PUGI__FN FILE* open_file_wide(const wchar_t* path, const wchar_t* mode) +{ + // there is no standard function to open wide paths, so our best bet is to try utf8 path + char* path_utf8 = convert_path_heap(path); + if (!path_utf8) + return 0; - // convert mode to ASCII (we mirror _wfopen interface) - char mode_ascii[4] = {0}; - for (size_t i = 0; mode[i]; ++i) mode_ascii[i] = static_cast(mode[i]); + // convert mode to ASCII (we mirror _wfopen interface) + char mode_ascii[4] = {0}; + for (size_t i = 0; mode[i]; ++i) + mode_ascii[i] = static_cast(mode[i]); - // try to open the utf8 path - FILE* result = fopen(path_utf8, mode_ascii); + // try to open the utf8 path + FILE* result = fopen(path_utf8, mode_ascii); - // free dummy buffer - xml_memory::deallocate(path_utf8); + // free dummy buffer + xml_memory::deallocate(path_utf8); - return result; - } + return result; +} #endif - PUGI__FN bool save_file_impl(const xml_document& doc, FILE* file, const char_t* indent, unsigned int flags, xml_encoding encoding) +PUGI__FN bool save_file_impl(const xml_document& doc, FILE* file, const char_t* indent, unsigned int flags, xml_encoding encoding) +{ + if (!file) + return false; + + xml_writer_file writer(file); + doc.save(writer, indent, flags, encoding); + + return ferror(file) == 0; +} + +struct name_null_sentry +{ + xml_node_struct* node; + char_t* name; + + name_null_sentry(xml_node_struct* node_) + : node(node_) + , name(node_->name) { - if (!file) return false; - - xml_writer_file writer(file); - doc.save(writer, indent, flags, encoding); - - return ferror(file) == 0; + node->name = 0; } - struct name_null_sentry + ~name_null_sentry() { - xml_node_struct* node; - char_t* name; - - name_null_sentry(xml_node_struct* node_): node(node_), name(node_->name) - { - node->name = 0; - } - - ~name_null_sentry() - { - node->name = name; - } - }; + node->name = name; + } +}; PUGI__NS_END namespace pugi { - PUGI__FN xml_writer_file::xml_writer_file(void* file_): file(file_) - { - } +PUGI__FN xml_writer_file::xml_writer_file(void* file_) + : file(file_) +{ +} - PUGI__FN void xml_writer_file::write(const void* data, size_t size) - { - size_t result = fwrite(data, 1, size, static_cast(file)); - (void)!result; // unfortunately we can't do proper error handling here - } +PUGI__FN void xml_writer_file::write(const void* data, size_t size) +{ + size_t result = fwrite(data, 1, size, static_cast(file)); + (void)!result; // unfortunately we can't do proper error handling here +} #ifndef PUGIXML_NO_STL - PUGI__FN xml_writer_stream::xml_writer_stream(std::basic_ostream >& stream): narrow_stream(&stream), wide_stream(0) - { - } +PUGI__FN xml_writer_stream::xml_writer_stream(std::basic_ostream>& stream) + : narrow_stream(&stream) + , wide_stream(0) +{ +} - PUGI__FN xml_writer_stream::xml_writer_stream(std::basic_ostream >& stream): narrow_stream(0), wide_stream(&stream) - { - } +PUGI__FN xml_writer_stream::xml_writer_stream(std::basic_ostream>& stream) + : narrow_stream(0) + , wide_stream(&stream) +{ +} - PUGI__FN void xml_writer_stream::write(const void* data, size_t size) +PUGI__FN void xml_writer_stream::write(const void* data, size_t size) +{ + if (narrow_stream) { - if (narrow_stream) - { - assert(!wide_stream); - narrow_stream->write(reinterpret_cast(data), static_cast(size)); - } - else - { - assert(wide_stream); - assert(size % sizeof(wchar_t) == 0); - - wide_stream->write(reinterpret_cast(data), static_cast(size / sizeof(wchar_t))); - } + assert(!wide_stream); + narrow_stream->write(reinterpret_cast(data), static_cast(size)); } + else + { + assert(wide_stream); + assert(size % sizeof(wchar_t) == 0); + + wide_stream->write(reinterpret_cast(data), static_cast(size / sizeof(wchar_t))); + } +} #endif - PUGI__FN xml_tree_walker::xml_tree_walker(): _depth(0) - { - } +PUGI__FN xml_tree_walker::xml_tree_walker() + : _depth(0) +{ +} - PUGI__FN xml_tree_walker::~xml_tree_walker() - { - } +PUGI__FN xml_tree_walker::~xml_tree_walker() +{ +} - PUGI__FN int xml_tree_walker::depth() const - { - return _depth; - } +PUGI__FN int xml_tree_walker::depth() const +{ + return _depth; +} - PUGI__FN bool xml_tree_walker::begin(xml_node&) - { - return true; - } +PUGI__FN bool xml_tree_walker::begin(xml_node&) +{ + return true; +} - PUGI__FN bool xml_tree_walker::end(xml_node&) - { - return true; - } +PUGI__FN bool xml_tree_walker::end(xml_node&) +{ + return true; +} - PUGI__FN xml_attribute::xml_attribute(): _attr(0) - { - } +PUGI__FN xml_attribute::xml_attribute() + : _attr(0) +{ +} - PUGI__FN xml_attribute::xml_attribute(xml_attribute_struct* attr): _attr(attr) - { - } +PUGI__FN xml_attribute::xml_attribute(xml_attribute_struct* attr) + : _attr(attr) +{ +} - PUGI__FN static void unspecified_bool_xml_attribute(xml_attribute***) - { - } +PUGI__FN static void unspecified_bool_xml_attribute(xml_attribute***) +{ +} - PUGI__FN xml_attribute::operator xml_attribute::unspecified_bool_type() const - { - return _attr ? unspecified_bool_xml_attribute : 0; - } +PUGI__FN xml_attribute::operator xml_attribute::unspecified_bool_type() const +{ + return _attr ? unspecified_bool_xml_attribute : 0; +} - PUGI__FN bool xml_attribute::operator!() const - { - return !_attr; - } +PUGI__FN bool xml_attribute::operator!() const +{ + return !_attr; +} - PUGI__FN bool xml_attribute::operator==(const xml_attribute& r) const - { - return (_attr == r._attr); - } +PUGI__FN bool xml_attribute::operator==(const xml_attribute& r) const +{ + return (_attr == r._attr); +} - PUGI__FN bool xml_attribute::operator!=(const xml_attribute& r) const - { - return (_attr != r._attr); - } +PUGI__FN bool xml_attribute::operator!=(const xml_attribute& r) const +{ + return (_attr != r._attr); +} - PUGI__FN bool xml_attribute::operator<(const xml_attribute& r) const - { - return (_attr < r._attr); - } +PUGI__FN bool xml_attribute::operator<(const xml_attribute& r) const +{ + return (_attr < r._attr); +} - PUGI__FN bool xml_attribute::operator>(const xml_attribute& r) const - { - return (_attr > r._attr); - } +PUGI__FN bool xml_attribute::operator>(const xml_attribute& r) const +{ + return (_attr > r._attr); +} - PUGI__FN bool xml_attribute::operator<=(const xml_attribute& r) const - { - return (_attr <= r._attr); - } +PUGI__FN bool xml_attribute::operator<=(const xml_attribute& r) const +{ + return (_attr <= r._attr); +} - PUGI__FN bool xml_attribute::operator>=(const xml_attribute& r) const - { - return (_attr >= r._attr); - } +PUGI__FN bool xml_attribute::operator>=(const xml_attribute& r) const +{ + return (_attr >= r._attr); +} - PUGI__FN xml_attribute xml_attribute::next_attribute() const - { - return _attr ? xml_attribute(_attr->next_attribute) : xml_attribute(); - } +PUGI__FN xml_attribute xml_attribute::next_attribute() const +{ + return _attr ? xml_attribute(_attr->next_attribute) : xml_attribute(); +} - PUGI__FN xml_attribute xml_attribute::previous_attribute() const - { - return _attr && _attr->prev_attribute_c->next_attribute ? xml_attribute(_attr->prev_attribute_c) : xml_attribute(); - } +PUGI__FN xml_attribute xml_attribute::previous_attribute() const +{ + return _attr && _attr->prev_attribute_c->next_attribute ? xml_attribute(_attr->prev_attribute_c) : xml_attribute(); +} - PUGI__FN const char_t* xml_attribute::as_string(const char_t* def) const - { - return (_attr && _attr->value) ? _attr->value + 0 : def; - } +PUGI__FN const char_t* xml_attribute::as_string(const char_t* def) const +{ + return (_attr && _attr->value) ? _attr->value + 0 : def; +} - PUGI__FN int xml_attribute::as_int(int def) const - { - return (_attr && _attr->value) ? impl::get_value_int(_attr->value) : def; - } +PUGI__FN int xml_attribute::as_int(int def) const +{ + return (_attr && _attr->value) ? impl::get_value_int(_attr->value) : def; +} - PUGI__FN unsigned int xml_attribute::as_uint(unsigned int def) const - { - return (_attr && _attr->value) ? impl::get_value_uint(_attr->value) : def; - } +PUGI__FN unsigned int xml_attribute::as_uint(unsigned int def) const +{ + return (_attr && _attr->value) ? impl::get_value_uint(_attr->value) : def; +} - PUGI__FN double xml_attribute::as_double(double def) const - { - return (_attr && _attr->value) ? impl::get_value_double(_attr->value) : def; - } +PUGI__FN double xml_attribute::as_double(double def) const +{ + return (_attr && _attr->value) ? impl::get_value_double(_attr->value) : def; +} - PUGI__FN float xml_attribute::as_float(float def) const - { - return (_attr && _attr->value) ? impl::get_value_float(_attr->value) : def; - } +PUGI__FN float xml_attribute::as_float(float def) const +{ + return (_attr && _attr->value) ? impl::get_value_float(_attr->value) : def; +} - PUGI__FN bool xml_attribute::as_bool(bool def) const - { - return (_attr && _attr->value) ? impl::get_value_bool(_attr->value) : def; - } +PUGI__FN bool xml_attribute::as_bool(bool def) const +{ + return (_attr && _attr->value) ? impl::get_value_bool(_attr->value) : def; +} #ifdef PUGIXML_HAS_LONG_LONG - PUGI__FN long long xml_attribute::as_llong(long long def) const - { - return (_attr && _attr->value) ? impl::get_value_llong(_attr->value) : def; - } +PUGI__FN long long xml_attribute::as_llong(long long def) const +{ + return (_attr && _attr->value) ? impl::get_value_llong(_attr->value) : def; +} - PUGI__FN unsigned long long xml_attribute::as_ullong(unsigned long long def) const - { - return (_attr && _attr->value) ? impl::get_value_ullong(_attr->value) : def; - } +PUGI__FN unsigned long long xml_attribute::as_ullong(unsigned long long def) const +{ + return (_attr && _attr->value) ? impl::get_value_ullong(_attr->value) : def; +} #endif - PUGI__FN bool xml_attribute::empty() const - { - return !_attr; - } +PUGI__FN bool xml_attribute::empty() const +{ + return !_attr; +} - PUGI__FN const char_t* xml_attribute::name() const - { - return (_attr && _attr->name) ? _attr->name + 0 : PUGIXML_TEXT(""); - } +PUGI__FN const char_t* xml_attribute::name() const +{ + return (_attr && _attr->name) ? _attr->name + 0 : PUGIXML_TEXT(""); +} - PUGI__FN const char_t* xml_attribute::value() const - { - return (_attr && _attr->value) ? _attr->value + 0 : PUGIXML_TEXT(""); - } +PUGI__FN const char_t* xml_attribute::value() const +{ + return (_attr && _attr->value) ? _attr->value + 0 : PUGIXML_TEXT(""); +} - PUGI__FN size_t xml_attribute::hash_value() const - { - return static_cast(reinterpret_cast(_attr) / sizeof(xml_attribute_struct)); - } +PUGI__FN size_t xml_attribute::hash_value() const +{ + return static_cast(reinterpret_cast(_attr) / sizeof(xml_attribute_struct)); +} - PUGI__FN xml_attribute_struct* xml_attribute::internal_object() const - { - return _attr; - } +PUGI__FN xml_attribute_struct* xml_attribute::internal_object() const +{ + return _attr; +} - PUGI__FN xml_attribute& xml_attribute::operator=(const char_t* rhs) - { - set_value(rhs); - return *this; - } +PUGI__FN xml_attribute& xml_attribute::operator=(const char_t* rhs) +{ + set_value(rhs); + return *this; +} - PUGI__FN xml_attribute& xml_attribute::operator=(int rhs) - { - set_value(rhs); - return *this; - } +PUGI__FN xml_attribute& xml_attribute::operator=(int rhs) +{ + set_value(rhs); + return *this; +} - PUGI__FN xml_attribute& xml_attribute::operator=(unsigned int rhs) - { - set_value(rhs); - return *this; - } +PUGI__FN xml_attribute& xml_attribute::operator=(unsigned int rhs) +{ + set_value(rhs); + return *this; +} - PUGI__FN xml_attribute& xml_attribute::operator=(long rhs) - { - set_value(rhs); - return *this; - } +PUGI__FN xml_attribute& xml_attribute::operator=(long rhs) +{ + set_value(rhs); + return *this; +} - PUGI__FN xml_attribute& xml_attribute::operator=(unsigned long rhs) - { - set_value(rhs); - return *this; - } +PUGI__FN xml_attribute& xml_attribute::operator=(unsigned long rhs) +{ + set_value(rhs); + return *this; +} - PUGI__FN xml_attribute& xml_attribute::operator=(double rhs) - { - set_value(rhs); - return *this; - } +PUGI__FN xml_attribute& xml_attribute::operator=(double rhs) +{ + set_value(rhs); + return *this; +} - PUGI__FN xml_attribute& xml_attribute::operator=(float rhs) - { - set_value(rhs); - return *this; - } +PUGI__FN xml_attribute& xml_attribute::operator=(float rhs) +{ + set_value(rhs); + return *this; +} - PUGI__FN xml_attribute& xml_attribute::operator=(bool rhs) - { - set_value(rhs); - return *this; - } +PUGI__FN xml_attribute& xml_attribute::operator=(bool rhs) +{ + set_value(rhs); + return *this; +} #ifdef PUGIXML_HAS_LONG_LONG - PUGI__FN xml_attribute& xml_attribute::operator=(long long rhs) - { - set_value(rhs); - return *this; - } +PUGI__FN xml_attribute& xml_attribute::operator=(long long rhs) +{ + set_value(rhs); + return *this; +} - PUGI__FN xml_attribute& xml_attribute::operator=(unsigned long long rhs) - { - set_value(rhs); - return *this; - } +PUGI__FN xml_attribute& xml_attribute::operator=(unsigned long long rhs) +{ + set_value(rhs); + return *this; +} #endif - PUGI__FN bool xml_attribute::set_name(const char_t* rhs) - { - if (!_attr) return false; +PUGI__FN bool xml_attribute::set_name(const char_t* rhs) +{ + if (!_attr) + return false; - return impl::strcpy_insitu(_attr->name, _attr->header, impl::xml_memory_page_name_allocated_mask, rhs, impl::strlength(rhs)); - } + return impl::strcpy_insitu(_attr->name, _attr->header, impl::xml_memory_page_name_allocated_mask, rhs, impl::strlength(rhs)); +} - PUGI__FN bool xml_attribute::set_value(const char_t* rhs) - { - if (!_attr) return false; +PUGI__FN bool xml_attribute::set_value(const char_t* rhs) +{ + if (!_attr) + return false; - return impl::strcpy_insitu(_attr->value, _attr->header, impl::xml_memory_page_value_allocated_mask, rhs, impl::strlength(rhs)); - } + return impl::strcpy_insitu(_attr->value, _attr->header, impl::xml_memory_page_value_allocated_mask, rhs, impl::strlength(rhs)); +} - PUGI__FN bool xml_attribute::set_value(int rhs) - { - if (!_attr) return false; +PUGI__FN bool xml_attribute::set_value(int rhs) +{ + if (!_attr) + return false; - return impl::set_value_integer(_attr->value, _attr->header, impl::xml_memory_page_value_allocated_mask, rhs, rhs < 0); - } + return impl::set_value_integer(_attr->value, _attr->header, impl::xml_memory_page_value_allocated_mask, rhs, rhs < 0); +} - PUGI__FN bool xml_attribute::set_value(unsigned int rhs) - { - if (!_attr) return false; +PUGI__FN bool xml_attribute::set_value(unsigned int rhs) +{ + if (!_attr) + return false; - return impl::set_value_integer(_attr->value, _attr->header, impl::xml_memory_page_value_allocated_mask, rhs, false); - } + return impl::set_value_integer(_attr->value, _attr->header, impl::xml_memory_page_value_allocated_mask, rhs, false); +} - PUGI__FN bool xml_attribute::set_value(long rhs) - { - if (!_attr) return false; +PUGI__FN bool xml_attribute::set_value(long rhs) +{ + if (!_attr) + return false; - return impl::set_value_integer(_attr->value, _attr->header, impl::xml_memory_page_value_allocated_mask, rhs, rhs < 0); - } + return impl::set_value_integer(_attr->value, _attr->header, impl::xml_memory_page_value_allocated_mask, rhs, rhs < 0); +} - PUGI__FN bool xml_attribute::set_value(unsigned long rhs) - { - if (!_attr) return false; +PUGI__FN bool xml_attribute::set_value(unsigned long rhs) +{ + if (!_attr) + return false; - return impl::set_value_integer(_attr->value, _attr->header, impl::xml_memory_page_value_allocated_mask, rhs, false); - } + return impl::set_value_integer(_attr->value, _attr->header, impl::xml_memory_page_value_allocated_mask, rhs, false); +} - PUGI__FN bool xml_attribute::set_value(double rhs) - { - if (!_attr) return false; +PUGI__FN bool xml_attribute::set_value(double rhs) +{ + if (!_attr) + return false; - return impl::set_value_convert(_attr->value, _attr->header, impl::xml_memory_page_value_allocated_mask, rhs); - } + return impl::set_value_convert(_attr->value, _attr->header, impl::xml_memory_page_value_allocated_mask, rhs); +} - PUGI__FN bool xml_attribute::set_value(float rhs) - { - if (!_attr) return false; +PUGI__FN bool xml_attribute::set_value(float rhs) +{ + if (!_attr) + return false; - return impl::set_value_convert(_attr->value, _attr->header, impl::xml_memory_page_value_allocated_mask, rhs); - } + return impl::set_value_convert(_attr->value, _attr->header, impl::xml_memory_page_value_allocated_mask, rhs); +} - PUGI__FN bool xml_attribute::set_value(bool rhs) - { - if (!_attr) return false; +PUGI__FN bool xml_attribute::set_value(bool rhs) +{ + if (!_attr) + return false; - return impl::set_value_bool(_attr->value, _attr->header, impl::xml_memory_page_value_allocated_mask, rhs); - } + return impl::set_value_bool(_attr->value, _attr->header, impl::xml_memory_page_value_allocated_mask, rhs); +} #ifdef PUGIXML_HAS_LONG_LONG - PUGI__FN bool xml_attribute::set_value(long long rhs) - { - if (!_attr) return false; +PUGI__FN bool xml_attribute::set_value(long long rhs) +{ + if (!_attr) + return false; - return impl::set_value_integer(_attr->value, _attr->header, impl::xml_memory_page_value_allocated_mask, rhs, rhs < 0); - } + return impl::set_value_integer(_attr->value, _attr->header, impl::xml_memory_page_value_allocated_mask, rhs, rhs < 0); +} - PUGI__FN bool xml_attribute::set_value(unsigned long long rhs) - { - if (!_attr) return false; +PUGI__FN bool xml_attribute::set_value(unsigned long long rhs) +{ + if (!_attr) + return false; - return impl::set_value_integer(_attr->value, _attr->header, impl::xml_memory_page_value_allocated_mask, rhs, false); - } + return impl::set_value_integer(_attr->value, _attr->header, impl::xml_memory_page_value_allocated_mask, rhs, false); +} #endif #ifdef __BORLANDC__ - PUGI__FN bool operator&&(const xml_attribute& lhs, bool rhs) - { - return (bool)lhs && rhs; - } +PUGI__FN bool operator&&(const xml_attribute& lhs, bool rhs) +{ + return (bool)lhs && rhs; +} - PUGI__FN bool operator||(const xml_attribute& lhs, bool rhs) - { - return (bool)lhs || rhs; - } +PUGI__FN bool operator||(const xml_attribute& lhs, bool rhs) +{ + return (bool)lhs || rhs; +} #endif - PUGI__FN xml_node::xml_node(): _root(0) - { - } +PUGI__FN xml_node::xml_node() + : _root(0) +{ +} - PUGI__FN xml_node::xml_node(xml_node_struct* p): _root(p) - { - } +PUGI__FN xml_node::xml_node(xml_node_struct* p) + : _root(p) +{ +} - PUGI__FN static void unspecified_bool_xml_node(xml_node***) - { - } +PUGI__FN static void unspecified_bool_xml_node(xml_node***) +{ +} - PUGI__FN xml_node::operator xml_node::unspecified_bool_type() const - { - return _root ? unspecified_bool_xml_node : 0; - } +PUGI__FN xml_node::operator xml_node::unspecified_bool_type() const +{ + return _root ? unspecified_bool_xml_node : 0; +} - PUGI__FN bool xml_node::operator!() const - { - return !_root; - } +PUGI__FN bool xml_node::operator!() const +{ + return !_root; +} - PUGI__FN xml_node::iterator xml_node::begin() const - { - return iterator(_root ? _root->first_child + 0 : 0, _root); - } +PUGI__FN xml_node::iterator xml_node::begin() const +{ + return iterator(_root ? _root->first_child + 0 : 0, _root); +} - PUGI__FN xml_node::iterator xml_node::end() const - { - return iterator(0, _root); - } +PUGI__FN xml_node::iterator xml_node::end() const +{ + return iterator(0, _root); +} - PUGI__FN xml_node::attribute_iterator xml_node::attributes_begin() const - { - return attribute_iterator(_root ? _root->first_attribute + 0 : 0, _root); - } +PUGI__FN xml_node::attribute_iterator xml_node::attributes_begin() const +{ + return attribute_iterator(_root ? _root->first_attribute + 0 : 0, _root); +} - PUGI__FN xml_node::attribute_iterator xml_node::attributes_end() const - { - return attribute_iterator(0, _root); - } +PUGI__FN xml_node::attribute_iterator xml_node::attributes_end() const +{ + return attribute_iterator(0, _root); +} - PUGI__FN xml_object_range xml_node::children() const - { - return xml_object_range(begin(), end()); - } +PUGI__FN xml_object_range xml_node::children() const +{ + return xml_object_range(begin(), end()); +} - PUGI__FN xml_object_range xml_node::children(const char_t* name_) const - { - return xml_object_range(xml_named_node_iterator(child(name_)._root, _root, name_), xml_named_node_iterator(0, _root, name_)); - } +PUGI__FN xml_object_range xml_node::children(const char_t* name_) const +{ + return xml_object_range(xml_named_node_iterator(child(name_)._root, _root, name_), xml_named_node_iterator(0, _root, name_)); +} - PUGI__FN xml_object_range xml_node::attributes() const - { - return xml_object_range(attributes_begin(), attributes_end()); - } +PUGI__FN xml_object_range xml_node::attributes() const +{ + return xml_object_range(attributes_begin(), attributes_end()); +} - PUGI__FN bool xml_node::operator==(const xml_node& r) const - { - return (_root == r._root); - } +PUGI__FN bool xml_node::operator==(const xml_node& r) const +{ + return (_root == r._root); +} - PUGI__FN bool xml_node::operator!=(const xml_node& r) const - { - return (_root != r._root); - } +PUGI__FN bool xml_node::operator!=(const xml_node& r) const +{ + return (_root != r._root); +} - PUGI__FN bool xml_node::operator<(const xml_node& r) const - { - return (_root < r._root); - } +PUGI__FN bool xml_node::operator<(const xml_node& r) const +{ + return (_root < r._root); +} - PUGI__FN bool xml_node::operator>(const xml_node& r) const - { - return (_root > r._root); - } +PUGI__FN bool xml_node::operator>(const xml_node& r) const +{ + return (_root > r._root); +} - PUGI__FN bool xml_node::operator<=(const xml_node& r) const - { - return (_root <= r._root); - } +PUGI__FN bool xml_node::operator<=(const xml_node& r) const +{ + return (_root <= r._root); +} - PUGI__FN bool xml_node::operator>=(const xml_node& r) const - { - return (_root >= r._root); - } +PUGI__FN bool xml_node::operator>=(const xml_node& r) const +{ + return (_root >= r._root); +} - PUGI__FN bool xml_node::empty() const - { - return !_root; - } +PUGI__FN bool xml_node::empty() const +{ + return !_root; +} - PUGI__FN const char_t* xml_node::name() const - { - return (_root && _root->name) ? _root->name + 0 : PUGIXML_TEXT(""); - } +PUGI__FN const char_t* xml_node::name() const +{ + return (_root && _root->name) ? _root->name + 0 : PUGIXML_TEXT(""); +} - PUGI__FN xml_node_type xml_node::type() const - { - return _root ? PUGI__NODETYPE(_root) : node_null; - } +PUGI__FN xml_node_type xml_node::type() const +{ + return _root ? PUGI__NODETYPE(_root) : node_null; +} - PUGI__FN const char_t* xml_node::value() const - { - return (_root && _root->value) ? _root->value + 0 : PUGIXML_TEXT(""); - } - - PUGI__FN xml_node xml_node::child(const char_t* name_) const - { - if (!_root) return xml_node(); - - for (xml_node_struct* i = _root->first_child; i; i = i->next_sibling) - if (i->name && impl::strequal(name_, i->name)) return xml_node(i); +PUGI__FN const char_t* xml_node::value() const +{ + return (_root && _root->value) ? _root->value + 0 : PUGIXML_TEXT(""); +} +PUGI__FN xml_node xml_node::child(const char_t* name_) const +{ + if (!_root) return xml_node(); - } - PUGI__FN xml_attribute xml_node::attribute(const char_t* name_) const - { - if (!_root) return xml_attribute(); + for (xml_node_struct* i = _root->first_child; i; i = i->next_sibling) + if (i->name && impl::strequal(name_, i->name)) + return xml_node(i); - for (xml_attribute_struct* i = _root->first_attribute; i; i = i->next_attribute) - if (i->name && impl::strequal(name_, i->name)) - return xml_attribute(i); + return xml_node(); +} +PUGI__FN xml_attribute xml_node::attribute(const char_t* name_) const +{ + if (!_root) return xml_attribute(); - } - PUGI__FN xml_node xml_node::next_sibling(const char_t* name_) const - { - if (!_root) return xml_node(); + for (xml_attribute_struct* i = _root->first_attribute; i; i = i->next_attribute) + if (i->name && impl::strequal(name_, i->name)) + return xml_attribute(i); - for (xml_node_struct* i = _root->next_sibling; i; i = i->next_sibling) - if (i->name && impl::strequal(name_, i->name)) return xml_node(i); + return xml_attribute(); +} +PUGI__FN xml_node xml_node::next_sibling(const char_t* name_) const +{ + if (!_root) return xml_node(); - } - PUGI__FN xml_node xml_node::next_sibling() const - { - return _root ? xml_node(_root->next_sibling) : xml_node(); - } + for (xml_node_struct* i = _root->next_sibling; i; i = i->next_sibling) + if (i->name && impl::strequal(name_, i->name)) + return xml_node(i); - PUGI__FN xml_node xml_node::previous_sibling(const char_t* name_) const - { - if (!_root) return xml_node(); + return xml_node(); +} - for (xml_node_struct* i = _root->prev_sibling_c; i->next_sibling; i = i->prev_sibling_c) - if (i->name && impl::strequal(name_, i->name)) return xml_node(i); +PUGI__FN xml_node xml_node::next_sibling() const +{ + return _root ? xml_node(_root->next_sibling) : xml_node(); +} +PUGI__FN xml_node xml_node::previous_sibling(const char_t* name_) const +{ + if (!_root) return xml_node(); - } - PUGI__FN xml_attribute xml_node::attribute(const char_t* name_, xml_attribute& hint_) const - { - xml_attribute_struct* hint = hint_._attr; + for (xml_node_struct* i = _root->prev_sibling_c; i->next_sibling; i = i->prev_sibling_c) + if (i->name && impl::strequal(name_, i->name)) + return xml_node(i); - // if hint is not an attribute of node, behavior is not defined - assert(!hint || (_root && impl::is_attribute_of(hint, _root))); + return xml_node(); +} - if (!_root) return xml_attribute(); +PUGI__FN xml_attribute xml_node::attribute(const char_t* name_, xml_attribute& hint_) const +{ + xml_attribute_struct* hint = hint_._attr; - // optimistically search from hint up until the end - for (xml_attribute_struct* i = hint; i; i = i->next_attribute) - if (i->name && impl::strequal(name_, i->name)) - { - // update hint to maximize efficiency of searching for consecutive attributes - hint_._attr = i->next_attribute; - - return xml_attribute(i); - } - - // wrap around and search from the first attribute until the hint - // 'j' null pointer check is technically redundant, but it prevents a crash in case the assertion above fails - for (xml_attribute_struct* j = _root->first_attribute; j && j != hint; j = j->next_attribute) - if (j->name && impl::strequal(name_, j->name)) - { - // update hint to maximize efficiency of searching for consecutive attributes - hint_._attr = j->next_attribute; - - return xml_attribute(j); - } + // if hint is not an attribute of node, behavior is not defined + assert(!hint || (_root && impl::is_attribute_of(hint, _root))); + if (!_root) return xml_attribute(); - } - PUGI__FN xml_node xml_node::previous_sibling() const - { - if (!_root) return xml_node(); + // optimistically search from hint up until the end + for (xml_attribute_struct* i = hint; i; i = i->next_attribute) + if (i->name && impl::strequal(name_, i->name)) + { + // update hint to maximize efficiency of searching for consecutive attributes + hint_._attr = i->next_attribute; - if (_root->prev_sibling_c->next_sibling) return xml_node(_root->prev_sibling_c); - else return xml_node(); - } + return xml_attribute(i); + } - PUGI__FN xml_node xml_node::parent() const - { - return _root ? xml_node(_root->parent) : xml_node(); - } + // wrap around and search from the first attribute until the hint + // 'j' null pointer check is technically redundant, but it prevents a crash in case the assertion above fails + for (xml_attribute_struct* j = _root->first_attribute; j && j != hint; j = j->next_attribute) + if (j->name && impl::strequal(name_, j->name)) + { + // update hint to maximize efficiency of searching for consecutive attributes + hint_._attr = j->next_attribute; - PUGI__FN xml_node xml_node::root() const - { - return _root ? xml_node(&impl::get_document(_root)) : xml_node(); - } + return xml_attribute(j); + } - PUGI__FN xml_text xml_node::text() const - { - return xml_text(_root); - } + return xml_attribute(); +} - PUGI__FN const char_t* xml_node::child_value() const - { - if (!_root) return PUGIXML_TEXT(""); +PUGI__FN xml_node xml_node::previous_sibling() const +{ + if (!_root) + return xml_node(); - // element nodes can have value if parse_embed_pcdata was used - if (PUGI__NODETYPE(_root) == node_element && _root->value) - return _root->value; + if (_root->prev_sibling_c->next_sibling) + return xml_node(_root->prev_sibling_c); + else + return xml_node(); +} - for (xml_node_struct* i = _root->first_child; i; i = i->next_sibling) - if (impl::is_text_node(i) && i->value) - return i->value; +PUGI__FN xml_node xml_node::parent() const +{ + return _root ? xml_node(_root->parent) : xml_node(); +} +PUGI__FN xml_node xml_node::root() const +{ + return _root ? xml_node(&impl::get_document(_root)) : xml_node(); +} + +PUGI__FN xml_text xml_node::text() const +{ + return xml_text(_root); +} + +PUGI__FN const char_t* xml_node::child_value() const +{ + if (!_root) return PUGIXML_TEXT(""); - } - PUGI__FN const char_t* xml_node::child_value(const char_t* name_) const - { - return child(name_).child_value(); - } + // element nodes can have value if parse_embed_pcdata was used + if (PUGI__NODETYPE(_root) == node_element && _root->value) + return _root->value; - PUGI__FN xml_attribute xml_node::first_attribute() const - { - return _root ? xml_attribute(_root->first_attribute) : xml_attribute(); - } + for (xml_node_struct* i = _root->first_child; i; i = i->next_sibling) + if (impl::is_text_node(i) && i->value) + return i->value; - PUGI__FN xml_attribute xml_node::last_attribute() const - { - return _root && _root->first_attribute ? xml_attribute(_root->first_attribute->prev_attribute_c) : xml_attribute(); - } + return PUGIXML_TEXT(""); +} - PUGI__FN xml_node xml_node::first_child() const - { - return _root ? xml_node(_root->first_child) : xml_node(); - } +PUGI__FN const char_t* xml_node::child_value(const char_t* name_) const +{ + return child(name_).child_value(); +} - PUGI__FN xml_node xml_node::last_child() const - { - return _root && _root->first_child ? xml_node(_root->first_child->prev_sibling_c) : xml_node(); - } +PUGI__FN xml_attribute xml_node::first_attribute() const +{ + return _root ? xml_attribute(_root->first_attribute) : xml_attribute(); +} - PUGI__FN bool xml_node::set_name(const char_t* rhs) - { - xml_node_type type_ = _root ? PUGI__NODETYPE(_root) : node_null; +PUGI__FN xml_attribute xml_node::last_attribute() const +{ + return _root && _root->first_attribute ? xml_attribute(_root->first_attribute->prev_attribute_c) : xml_attribute(); +} - if (type_ != node_element && type_ != node_pi && type_ != node_declaration) - return false; +PUGI__FN xml_node xml_node::first_child() const +{ + return _root ? xml_node(_root->first_child) : xml_node(); +} - return impl::strcpy_insitu(_root->name, _root->header, impl::xml_memory_page_name_allocated_mask, rhs, impl::strlength(rhs)); - } +PUGI__FN xml_node xml_node::last_child() const +{ + return _root && _root->first_child ? xml_node(_root->first_child->prev_sibling_c) : xml_node(); +} - PUGI__FN bool xml_node::set_value(const char_t* rhs) - { - xml_node_type type_ = _root ? PUGI__NODETYPE(_root) : node_null; +PUGI__FN bool xml_node::set_name(const char_t* rhs) +{ + xml_node_type type_ = _root ? PUGI__NODETYPE(_root) : node_null; - if (type_ != node_pcdata && type_ != node_cdata && type_ != node_comment && type_ != node_pi && type_ != node_doctype) - return false; + if (type_ != node_element && type_ != node_pi && type_ != node_declaration) + return false; - return impl::strcpy_insitu(_root->value, _root->header, impl::xml_memory_page_value_allocated_mask, rhs, impl::strlength(rhs)); - } + return impl::strcpy_insitu(_root->name, _root->header, impl::xml_memory_page_name_allocated_mask, rhs, impl::strlength(rhs)); +} - PUGI__FN xml_attribute xml_node::append_attribute(const char_t* name_) - { - if (!impl::allow_insert_attribute(type())) return xml_attribute(); +PUGI__FN bool xml_node::set_value(const char_t* rhs) +{ + xml_node_type type_ = _root ? PUGI__NODETYPE(_root) : node_null; - impl::xml_allocator& alloc = impl::get_allocator(_root); - if (!alloc.reserve()) return xml_attribute(); + if (type_ != node_pcdata && type_ != node_cdata && type_ != node_comment && type_ != node_pi && type_ != node_doctype) + return false; - xml_attribute a(impl::allocate_attribute(alloc)); - if (!a) return xml_attribute(); + return impl::strcpy_insitu(_root->value, _root->header, impl::xml_memory_page_value_allocated_mask, rhs, impl::strlength(rhs)); +} - impl::append_attribute(a._attr, _root); +PUGI__FN xml_attribute xml_node::append_attribute(const char_t* name_) +{ + if (!impl::allow_insert_attribute(type())) + return xml_attribute(); - a.set_name(name_); + impl::xml_allocator& alloc = impl::get_allocator(_root); + if (!alloc.reserve()) + return xml_attribute(); - return a; - } + xml_attribute a(impl::allocate_attribute(alloc)); + if (!a) + return xml_attribute(); - PUGI__FN xml_attribute xml_node::prepend_attribute(const char_t* name_) - { - if (!impl::allow_insert_attribute(type())) return xml_attribute(); + impl::append_attribute(a._attr, _root); - impl::xml_allocator& alloc = impl::get_allocator(_root); - if (!alloc.reserve()) return xml_attribute(); + a.set_name(name_); - xml_attribute a(impl::allocate_attribute(alloc)); - if (!a) return xml_attribute(); + return a; +} - impl::prepend_attribute(a._attr, _root); +PUGI__FN xml_attribute xml_node::prepend_attribute(const char_t* name_) +{ + if (!impl::allow_insert_attribute(type())) + return xml_attribute(); - a.set_name(name_); + impl::xml_allocator& alloc = impl::get_allocator(_root); + if (!alloc.reserve()) + return xml_attribute(); - return a; - } + xml_attribute a(impl::allocate_attribute(alloc)); + if (!a) + return xml_attribute(); - PUGI__FN xml_attribute xml_node::insert_attribute_after(const char_t* name_, const xml_attribute& attr) - { - if (!impl::allow_insert_attribute(type())) return xml_attribute(); - if (!attr || !impl::is_attribute_of(attr._attr, _root)) return xml_attribute(); + impl::prepend_attribute(a._attr, _root); - impl::xml_allocator& alloc = impl::get_allocator(_root); - if (!alloc.reserve()) return xml_attribute(); + a.set_name(name_); - xml_attribute a(impl::allocate_attribute(alloc)); - if (!a) return xml_attribute(); + return a; +} - impl::insert_attribute_after(a._attr, attr._attr, _root); +PUGI__FN xml_attribute xml_node::insert_attribute_after(const char_t* name_, const xml_attribute& attr) +{ + if (!impl::allow_insert_attribute(type())) + return xml_attribute(); + if (!attr || !impl::is_attribute_of(attr._attr, _root)) + return xml_attribute(); - a.set_name(name_); + impl::xml_allocator& alloc = impl::get_allocator(_root); + if (!alloc.reserve()) + return xml_attribute(); - return a; - } + xml_attribute a(impl::allocate_attribute(alloc)); + if (!a) + return xml_attribute(); - PUGI__FN xml_attribute xml_node::insert_attribute_before(const char_t* name_, const xml_attribute& attr) - { - if (!impl::allow_insert_attribute(type())) return xml_attribute(); - if (!attr || !impl::is_attribute_of(attr._attr, _root)) return xml_attribute(); + impl::insert_attribute_after(a._attr, attr._attr, _root); - impl::xml_allocator& alloc = impl::get_allocator(_root); - if (!alloc.reserve()) return xml_attribute(); + a.set_name(name_); - xml_attribute a(impl::allocate_attribute(alloc)); - if (!a) return xml_attribute(); + return a; +} - impl::insert_attribute_before(a._attr, attr._attr, _root); +PUGI__FN xml_attribute xml_node::insert_attribute_before(const char_t* name_, const xml_attribute& attr) +{ + if (!impl::allow_insert_attribute(type())) + return xml_attribute(); + if (!attr || !impl::is_attribute_of(attr._attr, _root)) + return xml_attribute(); - a.set_name(name_); + impl::xml_allocator& alloc = impl::get_allocator(_root); + if (!alloc.reserve()) + return xml_attribute(); - return a; - } + xml_attribute a(impl::allocate_attribute(alloc)); + if (!a) + return xml_attribute(); - PUGI__FN xml_attribute xml_node::append_copy(const xml_attribute& proto) - { - if (!proto) return xml_attribute(); - if (!impl::allow_insert_attribute(type())) return xml_attribute(); + impl::insert_attribute_before(a._attr, attr._attr, _root); - impl::xml_allocator& alloc = impl::get_allocator(_root); - if (!alloc.reserve()) return xml_attribute(); + a.set_name(name_); - xml_attribute a(impl::allocate_attribute(alloc)); - if (!a) return xml_attribute(); + return a; +} - impl::append_attribute(a._attr, _root); - impl::node_copy_attribute(a._attr, proto._attr); +PUGI__FN xml_attribute xml_node::append_copy(const xml_attribute& proto) +{ + if (!proto) + return xml_attribute(); + if (!impl::allow_insert_attribute(type())) + return xml_attribute(); - return a; - } + impl::xml_allocator& alloc = impl::get_allocator(_root); + if (!alloc.reserve()) + return xml_attribute(); - PUGI__FN xml_attribute xml_node::prepend_copy(const xml_attribute& proto) - { - if (!proto) return xml_attribute(); - if (!impl::allow_insert_attribute(type())) return xml_attribute(); + xml_attribute a(impl::allocate_attribute(alloc)); + if (!a) + return xml_attribute(); - impl::xml_allocator& alloc = impl::get_allocator(_root); - if (!alloc.reserve()) return xml_attribute(); + impl::append_attribute(a._attr, _root); + impl::node_copy_attribute(a._attr, proto._attr); - xml_attribute a(impl::allocate_attribute(alloc)); - if (!a) return xml_attribute(); + return a; +} - impl::prepend_attribute(a._attr, _root); - impl::node_copy_attribute(a._attr, proto._attr); +PUGI__FN xml_attribute xml_node::prepend_copy(const xml_attribute& proto) +{ + if (!proto) + return xml_attribute(); + if (!impl::allow_insert_attribute(type())) + return xml_attribute(); - return a; - } + impl::xml_allocator& alloc = impl::get_allocator(_root); + if (!alloc.reserve()) + return xml_attribute(); - PUGI__FN xml_attribute xml_node::insert_copy_after(const xml_attribute& proto, const xml_attribute& attr) - { - if (!proto) return xml_attribute(); - if (!impl::allow_insert_attribute(type())) return xml_attribute(); - if (!attr || !impl::is_attribute_of(attr._attr, _root)) return xml_attribute(); + xml_attribute a(impl::allocate_attribute(alloc)); + if (!a) + return xml_attribute(); - impl::xml_allocator& alloc = impl::get_allocator(_root); - if (!alloc.reserve()) return xml_attribute(); + impl::prepend_attribute(a._attr, _root); + impl::node_copy_attribute(a._attr, proto._attr); - xml_attribute a(impl::allocate_attribute(alloc)); - if (!a) return xml_attribute(); + return a; +} - impl::insert_attribute_after(a._attr, attr._attr, _root); - impl::node_copy_attribute(a._attr, proto._attr); +PUGI__FN xml_attribute xml_node::insert_copy_after(const xml_attribute& proto, const xml_attribute& attr) +{ + if (!proto) + return xml_attribute(); + if (!impl::allow_insert_attribute(type())) + return xml_attribute(); + if (!attr || !impl::is_attribute_of(attr._attr, _root)) + return xml_attribute(); - return a; - } + impl::xml_allocator& alloc = impl::get_allocator(_root); + if (!alloc.reserve()) + return xml_attribute(); + + xml_attribute a(impl::allocate_attribute(alloc)); + if (!a) + return xml_attribute(); + + impl::insert_attribute_after(a._attr, attr._attr, _root); + impl::node_copy_attribute(a._attr, proto._attr); - PUGI__FN xml_attribute xml_node::insert_copy_before(const xml_attribute& proto, const xml_attribute& attr) - { - if (!proto) return xml_attribute(); - if (!impl::allow_insert_attribute(type())) return xml_attribute(); - if (!attr || !impl::is_attribute_of(attr._attr, _root)) return xml_attribute(); + return a; +} - impl::xml_allocator& alloc = impl::get_allocator(_root); - if (!alloc.reserve()) return xml_attribute(); - - xml_attribute a(impl::allocate_attribute(alloc)); - if (!a) return xml_attribute(); - - impl::insert_attribute_before(a._attr, attr._attr, _root); - impl::node_copy_attribute(a._attr, proto._attr); - - return a; - } - - PUGI__FN xml_node xml_node::append_child(xml_node_type type_) - { - if (!impl::allow_insert_child(type(), type_)) return xml_node(); - - impl::xml_allocator& alloc = impl::get_allocator(_root); - if (!alloc.reserve()) return xml_node(); - - xml_node n(impl::allocate_node(alloc, type_)); - if (!n) return xml_node(); - - impl::append_node(n._root, _root); - - if (type_ == node_declaration) n.set_name(PUGIXML_TEXT("xml")); - - return n; - } - - PUGI__FN xml_node xml_node::prepend_child(xml_node_type type_) - { - if (!impl::allow_insert_child(type(), type_)) return xml_node(); - - impl::xml_allocator& alloc = impl::get_allocator(_root); - if (!alloc.reserve()) return xml_node(); - - xml_node n(impl::allocate_node(alloc, type_)); - if (!n) return xml_node(); - - impl::prepend_node(n._root, _root); - - if (type_ == node_declaration) n.set_name(PUGIXML_TEXT("xml")); - - return n; - } - - PUGI__FN xml_node xml_node::insert_child_before(xml_node_type type_, const xml_node& node) - { - if (!impl::allow_insert_child(type(), type_)) return xml_node(); - if (!node._root || node._root->parent != _root) return xml_node(); - - impl::xml_allocator& alloc = impl::get_allocator(_root); - if (!alloc.reserve()) return xml_node(); - - xml_node n(impl::allocate_node(alloc, type_)); - if (!n) return xml_node(); - - impl::insert_node_before(n._root, node._root); - - if (type_ == node_declaration) n.set_name(PUGIXML_TEXT("xml")); - - return n; - } - - PUGI__FN xml_node xml_node::insert_child_after(xml_node_type type_, const xml_node& node) - { - if (!impl::allow_insert_child(type(), type_)) return xml_node(); - if (!node._root || node._root->parent != _root) return xml_node(); - - impl::xml_allocator& alloc = impl::get_allocator(_root); - if (!alloc.reserve()) return xml_node(); - - xml_node n(impl::allocate_node(alloc, type_)); - if (!n) return xml_node(); - - impl::insert_node_after(n._root, node._root); - - if (type_ == node_declaration) n.set_name(PUGIXML_TEXT("xml")); - - return n; - } - - PUGI__FN xml_node xml_node::append_child(const char_t* name_) - { - xml_node result = append_child(node_element); - - result.set_name(name_); - - return result; - } - - PUGI__FN xml_node xml_node::prepend_child(const char_t* name_) - { - xml_node result = prepend_child(node_element); - - result.set_name(name_); - - return result; - } - - PUGI__FN xml_node xml_node::insert_child_after(const char_t* name_, const xml_node& node) - { - xml_node result = insert_child_after(node_element, node); - - result.set_name(name_); - - return result; - } - - PUGI__FN xml_node xml_node::insert_child_before(const char_t* name_, const xml_node& node) - { - xml_node result = insert_child_before(node_element, node); - - result.set_name(name_); - - return result; - } - - PUGI__FN xml_node xml_node::append_copy(const xml_node& proto) - { - xml_node_type type_ = proto.type(); - if (!impl::allow_insert_child(type(), type_)) return xml_node(); - - impl::xml_allocator& alloc = impl::get_allocator(_root); - if (!alloc.reserve()) return xml_node(); - - xml_node n(impl::allocate_node(alloc, type_)); - if (!n) return xml_node(); - - impl::append_node(n._root, _root); - impl::node_copy_tree(n._root, proto._root); - - return n; - } - - PUGI__FN xml_node xml_node::prepend_copy(const xml_node& proto) - { - xml_node_type type_ = proto.type(); - if (!impl::allow_insert_child(type(), type_)) return xml_node(); - - impl::xml_allocator& alloc = impl::get_allocator(_root); - if (!alloc.reserve()) return xml_node(); - - xml_node n(impl::allocate_node(alloc, type_)); - if (!n) return xml_node(); - - impl::prepend_node(n._root, _root); - impl::node_copy_tree(n._root, proto._root); - - return n; - } - - PUGI__FN xml_node xml_node::insert_copy_after(const xml_node& proto, const xml_node& node) - { - xml_node_type type_ = proto.type(); - if (!impl::allow_insert_child(type(), type_)) return xml_node(); - if (!node._root || node._root->parent != _root) return xml_node(); - - impl::xml_allocator& alloc = impl::get_allocator(_root); - if (!alloc.reserve()) return xml_node(); - - xml_node n(impl::allocate_node(alloc, type_)); - if (!n) return xml_node(); - - impl::insert_node_after(n._root, node._root); - impl::node_copy_tree(n._root, proto._root); - - return n; - } - - PUGI__FN xml_node xml_node::insert_copy_before(const xml_node& proto, const xml_node& node) - { - xml_node_type type_ = proto.type(); - if (!impl::allow_insert_child(type(), type_)) return xml_node(); - if (!node._root || node._root->parent != _root) return xml_node(); - - impl::xml_allocator& alloc = impl::get_allocator(_root); - if (!alloc.reserve()) return xml_node(); - - xml_node n(impl::allocate_node(alloc, type_)); - if (!n) return xml_node(); - - impl::insert_node_before(n._root, node._root); - impl::node_copy_tree(n._root, proto._root); - - return n; - } - - PUGI__FN xml_node xml_node::append_move(const xml_node& moved) - { - if (!impl::allow_move(*this, moved)) return xml_node(); - - impl::xml_allocator& alloc = impl::get_allocator(_root); - if (!alloc.reserve()) return xml_node(); - - // disable document_buffer_order optimization since moving nodes around changes document order without changing buffer pointers - impl::get_document(_root).header |= impl::xml_memory_page_contents_shared_mask; - - impl::remove_node(moved._root); - impl::append_node(moved._root, _root); - - return moved; - } - - PUGI__FN xml_node xml_node::prepend_move(const xml_node& moved) - { - if (!impl::allow_move(*this, moved)) return xml_node(); - - impl::xml_allocator& alloc = impl::get_allocator(_root); - if (!alloc.reserve()) return xml_node(); - - // disable document_buffer_order optimization since moving nodes around changes document order without changing buffer pointers - impl::get_document(_root).header |= impl::xml_memory_page_contents_shared_mask; - - impl::remove_node(moved._root); - impl::prepend_node(moved._root, _root); - - return moved; - } - - PUGI__FN xml_node xml_node::insert_move_after(const xml_node& moved, const xml_node& node) - { - if (!impl::allow_move(*this, moved)) return xml_node(); - if (!node._root || node._root->parent != _root) return xml_node(); - if (moved._root == node._root) return xml_node(); - - impl::xml_allocator& alloc = impl::get_allocator(_root); - if (!alloc.reserve()) return xml_node(); - - // disable document_buffer_order optimization since moving nodes around changes document order without changing buffer pointers - impl::get_document(_root).header |= impl::xml_memory_page_contents_shared_mask; - - impl::remove_node(moved._root); - impl::insert_node_after(moved._root, node._root); - - return moved; - } - - PUGI__FN xml_node xml_node::insert_move_before(const xml_node& moved, const xml_node& node) - { - if (!impl::allow_move(*this, moved)) return xml_node(); - if (!node._root || node._root->parent != _root) return xml_node(); - if (moved._root == node._root) return xml_node(); - - impl::xml_allocator& alloc = impl::get_allocator(_root); - if (!alloc.reserve()) return xml_node(); - - // disable document_buffer_order optimization since moving nodes around changes document order without changing buffer pointers - impl::get_document(_root).header |= impl::xml_memory_page_contents_shared_mask; - - impl::remove_node(moved._root); - impl::insert_node_before(moved._root, node._root); - - return moved; - } - - PUGI__FN bool xml_node::remove_attribute(const char_t* name_) - { - return remove_attribute(attribute(name_)); - } - - PUGI__FN bool xml_node::remove_attribute(const xml_attribute& a) - { - if (!_root || !a._attr) return false; - if (!impl::is_attribute_of(a._attr, _root)) return false; - - impl::xml_allocator& alloc = impl::get_allocator(_root); - if (!alloc.reserve()) return false; - - impl::remove_attribute(a._attr, _root); - impl::destroy_attribute(a._attr, alloc); - - return true; - } - - PUGI__FN bool xml_node::remove_child(const char_t* name_) - { - return remove_child(child(name_)); - } - - PUGI__FN bool xml_node::remove_child(const xml_node& n) - { - if (!_root || !n._root || n._root->parent != _root) return false; - - impl::xml_allocator& alloc = impl::get_allocator(_root); - if (!alloc.reserve()) return false; - - impl::remove_node(n._root); - impl::destroy_node(n._root, alloc); - - return true; - } - - PUGI__FN xml_parse_result xml_node::append_buffer(const void* contents, size_t size, unsigned int options, xml_encoding encoding) - { - // append_buffer is only valid for elements/documents - if (!impl::allow_insert_child(type(), node_element)) return impl::make_parse_result(status_append_invalid_root); - - // get document node - impl::xml_document_struct* doc = &impl::get_document(_root); - - // disable document_buffer_order optimization since in a document with multiple buffers comparing buffer pointers does not make sense - doc->header |= impl::xml_memory_page_contents_shared_mask; - - // get extra buffer element (we'll store the document fragment buffer there so that we can deallocate it later) - impl::xml_memory_page* page = 0; - impl::xml_extra_buffer* extra = static_cast(doc->allocate_memory(sizeof(impl::xml_extra_buffer) + sizeof(void*), page)); - (void)page; - - if (!extra) return impl::make_parse_result(status_out_of_memory); - - #ifdef PUGIXML_COMPACT - // align the memory block to a pointer boundary; this is required for compact mode where memory allocations are only 4b aligned - // note that this requires up to sizeof(void*)-1 additional memory, which the allocation above takes into account - extra = reinterpret_cast((reinterpret_cast(extra) + (sizeof(void*) - 1)) & ~(sizeof(void*) - 1)); - #endif - - // add extra buffer to the list - extra->buffer = 0; - extra->next = doc->extra_buffers; - doc->extra_buffers = extra; - - // name of the root has to be NULL before parsing - otherwise closing node mismatches will not be detected at the top level - impl::name_null_sentry sentry(_root); - - return impl::load_buffer_impl(doc, _root, const_cast(contents), size, options, encoding, false, false, &extra->buffer); - } - - PUGI__FN xml_node xml_node::find_child_by_attribute(const char_t* name_, const char_t* attr_name, const char_t* attr_value) const - { - if (!_root) return xml_node(); - - for (xml_node_struct* i = _root->first_child; i; i = i->next_sibling) - if (i->name && impl::strequal(name_, i->name)) - { - for (xml_attribute_struct* a = i->first_attribute; a; a = a->next_attribute) - if (a->name && impl::strequal(attr_name, a->name) && impl::strequal(attr_value, a->value ? a->value + 0 : PUGIXML_TEXT(""))) - return xml_node(i); - } +PUGI__FN xml_attribute xml_node::insert_copy_before(const xml_attribute& proto, const xml_attribute& attr) +{ + if (!proto) + return xml_attribute(); + if (!impl::allow_insert_attribute(type())) + return xml_attribute(); + if (!attr || !impl::is_attribute_of(attr._attr, _root)) + return xml_attribute(); + + impl::xml_allocator& alloc = impl::get_allocator(_root); + if (!alloc.reserve()) + return xml_attribute(); + + xml_attribute a(impl::allocate_attribute(alloc)); + if (!a) + return xml_attribute(); + + impl::insert_attribute_before(a._attr, attr._attr, _root); + impl::node_copy_attribute(a._attr, proto._attr); + return a; +} + +PUGI__FN xml_node xml_node::append_child(xml_node_type type_) +{ + if (!impl::allow_insert_child(type(), type_)) return xml_node(); - } - PUGI__FN xml_node xml_node::find_child_by_attribute(const char_t* attr_name, const char_t* attr_value) const - { - if (!_root) return xml_node(); + impl::xml_allocator& alloc = impl::get_allocator(_root); + if (!alloc.reserve()) + return xml_node(); - for (xml_node_struct* i = _root->first_child; i; i = i->next_sibling) + xml_node n(impl::allocate_node(alloc, type_)); + if (!n) + return xml_node(); + + impl::append_node(n._root, _root); + + if (type_ == node_declaration) + n.set_name(PUGIXML_TEXT("xml")); + + return n; +} + +PUGI__FN xml_node xml_node::prepend_child(xml_node_type type_) +{ + if (!impl::allow_insert_child(type(), type_)) + return xml_node(); + + impl::xml_allocator& alloc = impl::get_allocator(_root); + if (!alloc.reserve()) + return xml_node(); + + xml_node n(impl::allocate_node(alloc, type_)); + if (!n) + return xml_node(); + + impl::prepend_node(n._root, _root); + + if (type_ == node_declaration) + n.set_name(PUGIXML_TEXT("xml")); + + return n; +} + +PUGI__FN xml_node xml_node::insert_child_before(xml_node_type type_, const xml_node& node) +{ + if (!impl::allow_insert_child(type(), type_)) + return xml_node(); + if (!node._root || node._root->parent != _root) + return xml_node(); + + impl::xml_allocator& alloc = impl::get_allocator(_root); + if (!alloc.reserve()) + return xml_node(); + + xml_node n(impl::allocate_node(alloc, type_)); + if (!n) + return xml_node(); + + impl::insert_node_before(n._root, node._root); + + if (type_ == node_declaration) + n.set_name(PUGIXML_TEXT("xml")); + + return n; +} + +PUGI__FN xml_node xml_node::insert_child_after(xml_node_type type_, const xml_node& node) +{ + if (!impl::allow_insert_child(type(), type_)) + return xml_node(); + if (!node._root || node._root->parent != _root) + return xml_node(); + + impl::xml_allocator& alloc = impl::get_allocator(_root); + if (!alloc.reserve()) + return xml_node(); + + xml_node n(impl::allocate_node(alloc, type_)); + if (!n) + return xml_node(); + + impl::insert_node_after(n._root, node._root); + + if (type_ == node_declaration) + n.set_name(PUGIXML_TEXT("xml")); + + return n; +} + +PUGI__FN xml_node xml_node::append_child(const char_t* name_) +{ + xml_node result = append_child(node_element); + + result.set_name(name_); + + return result; +} + +PUGI__FN xml_node xml_node::prepend_child(const char_t* name_) +{ + xml_node result = prepend_child(node_element); + + result.set_name(name_); + + return result; +} + +PUGI__FN xml_node xml_node::insert_child_after(const char_t* name_, const xml_node& node) +{ + xml_node result = insert_child_after(node_element, node); + + result.set_name(name_); + + return result; +} + +PUGI__FN xml_node xml_node::insert_child_before(const char_t* name_, const xml_node& node) +{ + xml_node result = insert_child_before(node_element, node); + + result.set_name(name_); + + return result; +} + +PUGI__FN xml_node xml_node::append_copy(const xml_node& proto) +{ + xml_node_type type_ = proto.type(); + if (!impl::allow_insert_child(type(), type_)) + return xml_node(); + + impl::xml_allocator& alloc = impl::get_allocator(_root); + if (!alloc.reserve()) + return xml_node(); + + xml_node n(impl::allocate_node(alloc, type_)); + if (!n) + return xml_node(); + + impl::append_node(n._root, _root); + impl::node_copy_tree(n._root, proto._root); + + return n; +} + +PUGI__FN xml_node xml_node::prepend_copy(const xml_node& proto) +{ + xml_node_type type_ = proto.type(); + if (!impl::allow_insert_child(type(), type_)) + return xml_node(); + + impl::xml_allocator& alloc = impl::get_allocator(_root); + if (!alloc.reserve()) + return xml_node(); + + xml_node n(impl::allocate_node(alloc, type_)); + if (!n) + return xml_node(); + + impl::prepend_node(n._root, _root); + impl::node_copy_tree(n._root, proto._root); + + return n; +} + +PUGI__FN xml_node xml_node::insert_copy_after(const xml_node& proto, const xml_node& node) +{ + xml_node_type type_ = proto.type(); + if (!impl::allow_insert_child(type(), type_)) + return xml_node(); + if (!node._root || node._root->parent != _root) + return xml_node(); + + impl::xml_allocator& alloc = impl::get_allocator(_root); + if (!alloc.reserve()) + return xml_node(); + + xml_node n(impl::allocate_node(alloc, type_)); + if (!n) + return xml_node(); + + impl::insert_node_after(n._root, node._root); + impl::node_copy_tree(n._root, proto._root); + + return n; +} + +PUGI__FN xml_node xml_node::insert_copy_before(const xml_node& proto, const xml_node& node) +{ + xml_node_type type_ = proto.type(); + if (!impl::allow_insert_child(type(), type_)) + return xml_node(); + if (!node._root || node._root->parent != _root) + return xml_node(); + + impl::xml_allocator& alloc = impl::get_allocator(_root); + if (!alloc.reserve()) + return xml_node(); + + xml_node n(impl::allocate_node(alloc, type_)); + if (!n) + return xml_node(); + + impl::insert_node_before(n._root, node._root); + impl::node_copy_tree(n._root, proto._root); + + return n; +} + +PUGI__FN xml_node xml_node::append_move(const xml_node& moved) +{ + if (!impl::allow_move(*this, moved)) + return xml_node(); + + impl::xml_allocator& alloc = impl::get_allocator(_root); + if (!alloc.reserve()) + return xml_node(); + + // disable document_buffer_order optimization since moving nodes around changes document order without changing buffer pointers + impl::get_document(_root).header |= impl::xml_memory_page_contents_shared_mask; + + impl::remove_node(moved._root); + impl::append_node(moved._root, _root); + + return moved; +} + +PUGI__FN xml_node xml_node::prepend_move(const xml_node& moved) +{ + if (!impl::allow_move(*this, moved)) + return xml_node(); + + impl::xml_allocator& alloc = impl::get_allocator(_root); + if (!alloc.reserve()) + return xml_node(); + + // disable document_buffer_order optimization since moving nodes around changes document order without changing buffer pointers + impl::get_document(_root).header |= impl::xml_memory_page_contents_shared_mask; + + impl::remove_node(moved._root); + impl::prepend_node(moved._root, _root); + + return moved; +} + +PUGI__FN xml_node xml_node::insert_move_after(const xml_node& moved, const xml_node& node) +{ + if (!impl::allow_move(*this, moved)) + return xml_node(); + if (!node._root || node._root->parent != _root) + return xml_node(); + if (moved._root == node._root) + return xml_node(); + + impl::xml_allocator& alloc = impl::get_allocator(_root); + if (!alloc.reserve()) + return xml_node(); + + // disable document_buffer_order optimization since moving nodes around changes document order without changing buffer pointers + impl::get_document(_root).header |= impl::xml_memory_page_contents_shared_mask; + + impl::remove_node(moved._root); + impl::insert_node_after(moved._root, node._root); + + return moved; +} + +PUGI__FN xml_node xml_node::insert_move_before(const xml_node& moved, const xml_node& node) +{ + if (!impl::allow_move(*this, moved)) + return xml_node(); + if (!node._root || node._root->parent != _root) + return xml_node(); + if (moved._root == node._root) + return xml_node(); + + impl::xml_allocator& alloc = impl::get_allocator(_root); + if (!alloc.reserve()) + return xml_node(); + + // disable document_buffer_order optimization since moving nodes around changes document order without changing buffer pointers + impl::get_document(_root).header |= impl::xml_memory_page_contents_shared_mask; + + impl::remove_node(moved._root); + impl::insert_node_before(moved._root, node._root); + + return moved; +} + +PUGI__FN bool xml_node::remove_attribute(const char_t* name_) +{ + return remove_attribute(attribute(name_)); +} + +PUGI__FN bool xml_node::remove_attribute(const xml_attribute& a) +{ + if (!_root || !a._attr) + return false; + if (!impl::is_attribute_of(a._attr, _root)) + return false; + + impl::xml_allocator& alloc = impl::get_allocator(_root); + if (!alloc.reserve()) + return false; + + impl::remove_attribute(a._attr, _root); + impl::destroy_attribute(a._attr, alloc); + + return true; +} + +PUGI__FN bool xml_node::remove_child(const char_t* name_) +{ + return remove_child(child(name_)); +} + +PUGI__FN bool xml_node::remove_child(const xml_node& n) +{ + if (!_root || !n._root || n._root->parent != _root) + return false; + + impl::xml_allocator& alloc = impl::get_allocator(_root); + if (!alloc.reserve()) + return false; + + impl::remove_node(n._root); + impl::destroy_node(n._root, alloc); + + return true; +} + +PUGI__FN xml_parse_result xml_node::append_buffer(const void* contents, size_t size, unsigned int options, xml_encoding encoding) +{ + // append_buffer is only valid for elements/documents + if (!impl::allow_insert_child(type(), node_element)) + return impl::make_parse_result(status_append_invalid_root); + + // get document node + impl::xml_document_struct* doc = &impl::get_document(_root); + + // disable document_buffer_order optimization since in a document with multiple buffers comparing buffer pointers does not make sense + doc->header |= impl::xml_memory_page_contents_shared_mask; + + // get extra buffer element (we'll store the document fragment buffer there so that we can deallocate it later) + impl::xml_memory_page* page = 0; + impl::xml_extra_buffer* extra = static_cast(doc->allocate_memory(sizeof(impl::xml_extra_buffer) + sizeof(void*), page)); + (void)page; + + if (!extra) + return impl::make_parse_result(status_out_of_memory); + +#ifdef PUGIXML_COMPACT + // align the memory block to a pointer boundary; this is required for compact mode where memory allocations are only 4b aligned + // note that this requires up to sizeof(void*)-1 additional memory, which the allocation above takes into account + extra = reinterpret_cast((reinterpret_cast(extra) + (sizeof(void*) - 1)) & ~(sizeof(void*) - 1)); +#endif + + // add extra buffer to the list + extra->buffer = 0; + extra->next = doc->extra_buffers; + doc->extra_buffers = extra; + + // name of the root has to be NULL before parsing - otherwise closing node mismatches will not be detected at the top level + impl::name_null_sentry sentry(_root); + + return impl::load_buffer_impl(doc, _root, const_cast(contents), size, options, encoding, false, false, &extra->buffer); +} + +PUGI__FN xml_node xml_node::find_child_by_attribute(const char_t* name_, const char_t* attr_name, const char_t* attr_value) const +{ + if (!_root) + return xml_node(); + + for (xml_node_struct* i = _root->first_child; i; i = i->next_sibling) + if (i->name && impl::strequal(name_, i->name)) + { for (xml_attribute_struct* a = i->first_attribute; a; a = a->next_attribute) if (a->name && impl::strequal(attr_name, a->name) && impl::strequal(attr_value, a->value ? a->value + 0 : PUGIXML_TEXT(""))) return xml_node(i); + } + return xml_node(); +} + +PUGI__FN xml_node xml_node::find_child_by_attribute(const char_t* attr_name, const char_t* attr_value) const +{ + if (!_root) return xml_node(); - } -#ifndef PUGIXML_NO_STL - PUGI__FN string_t xml_node::path(char_t delimiter) const - { - if (!_root) return string_t(); - - size_t offset = 0; - - for (xml_node_struct* i = _root; i; i = i->parent) - { - offset += (i != _root); - offset += i->name ? impl::strlength(i->name) : 0; - } - - string_t result; - result.resize(offset); - - for (xml_node_struct* j = _root; j; j = j->parent) - { - if (j != _root) - result[--offset] = delimiter; - - if (j->name) - { - size_t length = impl::strlength(j->name); - - offset -= length; - memcpy(&result[offset], j->name, length * sizeof(char_t)); - } - } - - assert(offset == 0); - - return result; - } -#endif - - PUGI__FN xml_node xml_node::first_element_by_path(const char_t* path_, char_t delimiter) const - { - xml_node found = *this; // Current search context. - - if (!_root || !path_[0]) return found; - - if (path_[0] == delimiter) - { - // Absolute path; e.g. '/foo/bar' - found = found.root(); - ++path_; - } - - const char_t* path_segment = path_; - - while (*path_segment == delimiter) ++path_segment; - - const char_t* path_segment_end = path_segment; - - while (*path_segment_end && *path_segment_end != delimiter) ++path_segment_end; - - if (path_segment == path_segment_end) return found; - - const char_t* next_segment = path_segment_end; - - while (*next_segment == delimiter) ++next_segment; - - if (*path_segment == '.' && path_segment + 1 == path_segment_end) - return found.first_element_by_path(next_segment, delimiter); - else if (*path_segment == '.' && *(path_segment+1) == '.' && path_segment + 2 == path_segment_end) - return found.parent().first_element_by_path(next_segment, delimiter); - else - { - for (xml_node_struct* j = found._root->first_child; j; j = j->next_sibling) - { - if (j->name && impl::strequalrange(j->name, path_segment, static_cast(path_segment_end - path_segment))) - { - xml_node subsearch = xml_node(j).first_element_by_path(next_segment, delimiter); - - if (subsearch) return subsearch; - } - } - - return xml_node(); - } - } - - PUGI__FN bool xml_node::traverse(xml_tree_walker& walker) - { - walker._depth = -1; - - xml_node arg_begin(_root); - if (!walker.begin(arg_begin)) return false; - - xml_node_struct* cur = _root ? _root->first_child + 0 : 0; - - if (cur) - { - ++walker._depth; - - do - { - xml_node arg_for_each(cur); - if (!walker.for_each(arg_for_each)) - return false; - - if (cur->first_child) - { - ++walker._depth; - cur = cur->first_child; - } - else if (cur->next_sibling) - cur = cur->next_sibling; - else - { - while (!cur->next_sibling && cur != _root && cur->parent) - { - --walker._depth; - cur = cur->parent; - } - - if (cur != _root) - cur = cur->next_sibling; - } - } - while (cur && cur != _root); - } - - assert(walker._depth == -1); - - xml_node arg_end(_root); - return walker.end(arg_end); - } - - PUGI__FN size_t xml_node::hash_value() const - { - return static_cast(reinterpret_cast(_root) / sizeof(xml_node_struct)); - } - - PUGI__FN xml_node_struct* xml_node::internal_object() const - { - return _root; - } - - PUGI__FN void xml_node::print(xml_writer& writer, const char_t* indent, unsigned int flags, xml_encoding encoding, unsigned int depth) const - { - if (!_root) return; - - impl::xml_buffered_writer buffered_writer(writer, encoding); - - impl::node_output(buffered_writer, _root, indent, flags, depth); - - buffered_writer.flush(); - } - -#ifndef PUGIXML_NO_STL - PUGI__FN void xml_node::print(std::basic_ostream >& stream, const char_t* indent, unsigned int flags, xml_encoding encoding, unsigned int depth) const - { - xml_writer_stream writer(stream); - - print(writer, indent, flags, encoding, depth); - } - - PUGI__FN void xml_node::print(std::basic_ostream >& stream, const char_t* indent, unsigned int flags, unsigned int depth) const - { - xml_writer_stream writer(stream); - - print(writer, indent, flags, encoding_wchar, depth); - } -#endif - - PUGI__FN ptrdiff_t xml_node::offset_debug() const - { - if (!_root) return -1; - - impl::xml_document_struct& doc = impl::get_document(_root); - - // we can determine the offset reliably only if there is exactly once parse buffer - if (!doc.buffer || doc.extra_buffers) return -1; - - switch (type()) - { - case node_document: - return 0; - - case node_element: - case node_declaration: - case node_pi: - return _root->name && (_root->header & impl::xml_memory_page_name_allocated_or_shared_mask) == 0 ? _root->name - doc.buffer : -1; - - case node_pcdata: - case node_cdata: - case node_comment: - case node_doctype: - return _root->value && (_root->header & impl::xml_memory_page_value_allocated_or_shared_mask) == 0 ? _root->value - doc.buffer : -1; - - default: - assert(false && "Invalid node type"); // unreachable - return -1; - } - } - -#ifdef __BORLANDC__ - PUGI__FN bool operator&&(const xml_node& lhs, bool rhs) - { - return (bool)lhs && rhs; - } - - PUGI__FN bool operator||(const xml_node& lhs, bool rhs) - { - return (bool)lhs || rhs; - } -#endif - - PUGI__FN xml_text::xml_text(xml_node_struct* root): _root(root) - { - } - - PUGI__FN xml_node_struct* xml_text::_data() const - { - if (!_root || impl::is_text_node(_root)) return _root; - - // element nodes can have value if parse_embed_pcdata was used - if (PUGI__NODETYPE(_root) == node_element && _root->value) - return _root; - - for (xml_node_struct* node = _root->first_child; node; node = node->next_sibling) - if (impl::is_text_node(node)) - return node; - - return 0; - } - - PUGI__FN xml_node_struct* xml_text::_data_new() - { - xml_node_struct* d = _data(); - if (d) return d; - - return xml_node(_root).append_child(node_pcdata).internal_object(); - } - - PUGI__FN xml_text::xml_text(): _root(0) - { - } - - PUGI__FN static void unspecified_bool_xml_text(xml_text***) - { - } - - PUGI__FN xml_text::operator xml_text::unspecified_bool_type() const - { - return _data() ? unspecified_bool_xml_text : 0; - } - - PUGI__FN bool xml_text::operator!() const - { - return !_data(); - } - - PUGI__FN bool xml_text::empty() const - { - return _data() == 0; - } - - PUGI__FN const char_t* xml_text::get() const - { - xml_node_struct* d = _data(); - - return (d && d->value) ? d->value + 0 : PUGIXML_TEXT(""); - } - - PUGI__FN const char_t* xml_text::as_string(const char_t* def) const - { - xml_node_struct* d = _data(); - - return (d && d->value) ? d->value + 0 : def; - } - - PUGI__FN int xml_text::as_int(int def) const - { - xml_node_struct* d = _data(); - - return (d && d->value) ? impl::get_value_int(d->value) : def; - } - - PUGI__FN unsigned int xml_text::as_uint(unsigned int def) const - { - xml_node_struct* d = _data(); - - return (d && d->value) ? impl::get_value_uint(d->value) : def; - } - - PUGI__FN double xml_text::as_double(double def) const - { - xml_node_struct* d = _data(); - - return (d && d->value) ? impl::get_value_double(d->value) : def; - } - - PUGI__FN float xml_text::as_float(float def) const - { - xml_node_struct* d = _data(); - - return (d && d->value) ? impl::get_value_float(d->value) : def; - } - - PUGI__FN bool xml_text::as_bool(bool def) const - { - xml_node_struct* d = _data(); - - return (d && d->value) ? impl::get_value_bool(d->value) : def; - } - -#ifdef PUGIXML_HAS_LONG_LONG - PUGI__FN long long xml_text::as_llong(long long def) const - { - xml_node_struct* d = _data(); - - return (d && d->value) ? impl::get_value_llong(d->value) : def; - } - - PUGI__FN unsigned long long xml_text::as_ullong(unsigned long long def) const - { - xml_node_struct* d = _data(); - - return (d && d->value) ? impl::get_value_ullong(d->value) : def; - } -#endif - - PUGI__FN bool xml_text::set(const char_t* rhs) - { - xml_node_struct* dn = _data_new(); - - return dn ? impl::strcpy_insitu(dn->value, dn->header, impl::xml_memory_page_value_allocated_mask, rhs, impl::strlength(rhs)) : false; - } - - PUGI__FN bool xml_text::set(int rhs) - { - xml_node_struct* dn = _data_new(); - - return dn ? impl::set_value_integer(dn->value, dn->header, impl::xml_memory_page_value_allocated_mask, rhs, rhs < 0) : false; - } - - PUGI__FN bool xml_text::set(unsigned int rhs) - { - xml_node_struct* dn = _data_new(); - - return dn ? impl::set_value_integer(dn->value, dn->header, impl::xml_memory_page_value_allocated_mask, rhs, false) : false; - } - - PUGI__FN bool xml_text::set(long rhs) - { - xml_node_struct* dn = _data_new(); - - return dn ? impl::set_value_integer(dn->value, dn->header, impl::xml_memory_page_value_allocated_mask, rhs, rhs < 0) : false; - } - - PUGI__FN bool xml_text::set(unsigned long rhs) - { - xml_node_struct* dn = _data_new(); - - return dn ? impl::set_value_integer(dn->value, dn->header, impl::xml_memory_page_value_allocated_mask, rhs, false) : false; - } - - PUGI__FN bool xml_text::set(float rhs) - { - xml_node_struct* dn = _data_new(); - - return dn ? impl::set_value_convert(dn->value, dn->header, impl::xml_memory_page_value_allocated_mask, rhs) : false; - } - - PUGI__FN bool xml_text::set(double rhs) - { - xml_node_struct* dn = _data_new(); - - return dn ? impl::set_value_convert(dn->value, dn->header, impl::xml_memory_page_value_allocated_mask, rhs) : false; - } - - PUGI__FN bool xml_text::set(bool rhs) - { - xml_node_struct* dn = _data_new(); - - return dn ? impl::set_value_bool(dn->value, dn->header, impl::xml_memory_page_value_allocated_mask, rhs) : false; - } - -#ifdef PUGIXML_HAS_LONG_LONG - PUGI__FN bool xml_text::set(long long rhs) - { - xml_node_struct* dn = _data_new(); - - return dn ? impl::set_value_integer(dn->value, dn->header, impl::xml_memory_page_value_allocated_mask, rhs, rhs < 0) : false; - } - - PUGI__FN bool xml_text::set(unsigned long long rhs) - { - xml_node_struct* dn = _data_new(); - - return dn ? impl::set_value_integer(dn->value, dn->header, impl::xml_memory_page_value_allocated_mask, rhs, false) : false; - } -#endif - - PUGI__FN xml_text& xml_text::operator=(const char_t* rhs) - { - set(rhs); - return *this; - } - - PUGI__FN xml_text& xml_text::operator=(int rhs) - { - set(rhs); - return *this; - } - - PUGI__FN xml_text& xml_text::operator=(unsigned int rhs) - { - set(rhs); - return *this; - } - - PUGI__FN xml_text& xml_text::operator=(long rhs) - { - set(rhs); - return *this; - } - - PUGI__FN xml_text& xml_text::operator=(unsigned long rhs) - { - set(rhs); - return *this; - } - - PUGI__FN xml_text& xml_text::operator=(double rhs) - { - set(rhs); - return *this; - } - - PUGI__FN xml_text& xml_text::operator=(float rhs) - { - set(rhs); - return *this; - } - - PUGI__FN xml_text& xml_text::operator=(bool rhs) - { - set(rhs); - return *this; - } - -#ifdef PUGIXML_HAS_LONG_LONG - PUGI__FN xml_text& xml_text::operator=(long long rhs) - { - set(rhs); - return *this; - } - - PUGI__FN xml_text& xml_text::operator=(unsigned long long rhs) - { - set(rhs); - return *this; - } -#endif - - PUGI__FN xml_node xml_text::data() const - { - return xml_node(_data()); - } - -#ifdef __BORLANDC__ - PUGI__FN bool operator&&(const xml_text& lhs, bool rhs) - { - return (bool)lhs && rhs; - } - - PUGI__FN bool operator||(const xml_text& lhs, bool rhs) - { - return (bool)lhs || rhs; - } -#endif - - PUGI__FN xml_node_iterator::xml_node_iterator() - { - } - - PUGI__FN xml_node_iterator::xml_node_iterator(const xml_node& node): _wrap(node), _parent(node.parent()) - { - } - - PUGI__FN xml_node_iterator::xml_node_iterator(xml_node_struct* ref, xml_node_struct* parent): _wrap(ref), _parent(parent) - { - } - - PUGI__FN bool xml_node_iterator::operator==(const xml_node_iterator& rhs) const - { - return _wrap._root == rhs._wrap._root && _parent._root == rhs._parent._root; - } - - PUGI__FN bool xml_node_iterator::operator!=(const xml_node_iterator& rhs) const - { - return _wrap._root != rhs._wrap._root || _parent._root != rhs._parent._root; - } - - PUGI__FN xml_node& xml_node_iterator::operator*() const - { - assert(_wrap._root); - return _wrap; - } - - PUGI__FN xml_node* xml_node_iterator::operator->() const - { - assert(_wrap._root); - return const_cast(&_wrap); // BCC5 workaround - } - - PUGI__FN const xml_node_iterator& xml_node_iterator::operator++() - { - assert(_wrap._root); - _wrap._root = _wrap._root->next_sibling; - return *this; - } - - PUGI__FN xml_node_iterator xml_node_iterator::operator++(int) - { - xml_node_iterator temp = *this; - ++*this; - return temp; - } - - PUGI__FN const xml_node_iterator& xml_node_iterator::operator--() - { - _wrap = _wrap._root ? _wrap.previous_sibling() : _parent.last_child(); - return *this; - } - - PUGI__FN xml_node_iterator xml_node_iterator::operator--(int) - { - xml_node_iterator temp = *this; - --*this; - return temp; - } - - PUGI__FN xml_attribute_iterator::xml_attribute_iterator() - { - } - - PUGI__FN xml_attribute_iterator::xml_attribute_iterator(const xml_attribute& attr, const xml_node& parent): _wrap(attr), _parent(parent) - { - } - - PUGI__FN xml_attribute_iterator::xml_attribute_iterator(xml_attribute_struct* ref, xml_node_struct* parent): _wrap(ref), _parent(parent) - { - } - - PUGI__FN bool xml_attribute_iterator::operator==(const xml_attribute_iterator& rhs) const - { - return _wrap._attr == rhs._wrap._attr && _parent._root == rhs._parent._root; - } - - PUGI__FN bool xml_attribute_iterator::operator!=(const xml_attribute_iterator& rhs) const - { - return _wrap._attr != rhs._wrap._attr || _parent._root != rhs._parent._root; - } - - PUGI__FN xml_attribute& xml_attribute_iterator::operator*() const - { - assert(_wrap._attr); - return _wrap; - } - - PUGI__FN xml_attribute* xml_attribute_iterator::operator->() const - { - assert(_wrap._attr); - return const_cast(&_wrap); // BCC5 workaround - } - - PUGI__FN const xml_attribute_iterator& xml_attribute_iterator::operator++() - { - assert(_wrap._attr); - _wrap._attr = _wrap._attr->next_attribute; - return *this; - } - - PUGI__FN xml_attribute_iterator xml_attribute_iterator::operator++(int) - { - xml_attribute_iterator temp = *this; - ++*this; - return temp; - } - - PUGI__FN const xml_attribute_iterator& xml_attribute_iterator::operator--() - { - _wrap = _wrap._attr ? _wrap.previous_attribute() : _parent.last_attribute(); - return *this; - } - - PUGI__FN xml_attribute_iterator xml_attribute_iterator::operator--(int) - { - xml_attribute_iterator temp = *this; - --*this; - return temp; - } - - PUGI__FN xml_named_node_iterator::xml_named_node_iterator(): _name(0) - { - } - - PUGI__FN xml_named_node_iterator::xml_named_node_iterator(const xml_node& node, const char_t* name): _wrap(node), _parent(node.parent()), _name(name) - { - } - - PUGI__FN xml_named_node_iterator::xml_named_node_iterator(xml_node_struct* ref, xml_node_struct* parent, const char_t* name): _wrap(ref), _parent(parent), _name(name) - { - } - - PUGI__FN bool xml_named_node_iterator::operator==(const xml_named_node_iterator& rhs) const - { - return _wrap._root == rhs._wrap._root && _parent._root == rhs._parent._root; - } - - PUGI__FN bool xml_named_node_iterator::operator!=(const xml_named_node_iterator& rhs) const - { - return _wrap._root != rhs._wrap._root || _parent._root != rhs._parent._root; - } - - PUGI__FN xml_node& xml_named_node_iterator::operator*() const - { - assert(_wrap._root); - return _wrap; - } - - PUGI__FN xml_node* xml_named_node_iterator::operator->() const - { - assert(_wrap._root); - return const_cast(&_wrap); // BCC5 workaround - } - - PUGI__FN const xml_named_node_iterator& xml_named_node_iterator::operator++() - { - assert(_wrap._root); - _wrap = _wrap.next_sibling(_name); - return *this; - } - - PUGI__FN xml_named_node_iterator xml_named_node_iterator::operator++(int) - { - xml_named_node_iterator temp = *this; - ++*this; - return temp; - } - - PUGI__FN const xml_named_node_iterator& xml_named_node_iterator::operator--() - { - if (_wrap._root) - _wrap = _wrap.previous_sibling(_name); - else - { - _wrap = _parent.last_child(); - - if (!impl::strequal(_wrap.name(), _name)) - _wrap = _wrap.previous_sibling(_name); - } - - return *this; - } - - PUGI__FN xml_named_node_iterator xml_named_node_iterator::operator--(int) - { - xml_named_node_iterator temp = *this; - --*this; - return temp; - } - - PUGI__FN xml_parse_result::xml_parse_result(): status(status_internal_error), offset(0), encoding(encoding_auto) - { - } - - PUGI__FN xml_parse_result::operator bool() const - { - return status == status_ok; - } - - PUGI__FN const char* xml_parse_result::description() const - { - switch (status) - { - case status_ok: return "No error"; - - case status_file_not_found: return "File was not found"; - case status_io_error: return "Error reading from file/stream"; - case status_out_of_memory: return "Could not allocate memory"; - case status_internal_error: return "Internal error occurred"; - - case status_unrecognized_tag: return "Could not determine tag type"; - - case status_bad_pi: return "Error parsing document declaration/processing instruction"; - case status_bad_comment: return "Error parsing comment"; - case status_bad_cdata: return "Error parsing CDATA section"; - case status_bad_doctype: return "Error parsing document type declaration"; - case status_bad_pcdata: return "Error parsing PCDATA section"; - case status_bad_start_element: return "Error parsing start element tag"; - case status_bad_attribute: return "Error parsing element attribute"; - case status_bad_end_element: return "Error parsing end element tag"; - case status_end_element_mismatch: return "Start-end tags mismatch"; - - case status_append_invalid_root: return "Unable to append nodes: root is not an element or document"; - - case status_no_document_element: return "No document element found"; - - default: return "Unknown error"; - } - } - - PUGI__FN xml_document::xml_document(): _buffer(0) - { - _create(); - } - - PUGI__FN xml_document::~xml_document() - { - _destroy(); - } - -#ifdef PUGIXML_HAS_MOVE - PUGI__FN xml_document::xml_document(xml_document&& rhs) PUGIXML_NOEXCEPT_IF_NOT_COMPACT: _buffer(0) - { - _create(); - _move(rhs); - } - - PUGI__FN xml_document& xml_document::operator=(xml_document&& rhs) PUGIXML_NOEXCEPT_IF_NOT_COMPACT - { - if (this == &rhs) return *this; - - _destroy(); - _create(); - _move(rhs); - - return *this; - } -#endif - - PUGI__FN void xml_document::reset() - { - _destroy(); - _create(); - } - - PUGI__FN void xml_document::reset(const xml_document& proto) - { - reset(); - - for (xml_node cur = proto.first_child(); cur; cur = cur.next_sibling()) - append_copy(cur); - } - - PUGI__FN void xml_document::_create() - { - assert(!_root); - - #ifdef PUGIXML_COMPACT - // space for page marker for the first page (uint32_t), rounded up to pointer size; assumes pointers are at least 32-bit - const size_t page_offset = sizeof(void*); - #else - const size_t page_offset = 0; - #endif - - // initialize sentinel page - PUGI__STATIC_ASSERT(sizeof(impl::xml_memory_page) + sizeof(impl::xml_document_struct) + page_offset <= sizeof(_memory)); - - // prepare page structure - impl::xml_memory_page* page = impl::xml_memory_page::construct(_memory); - assert(page); - - page->busy_size = impl::xml_memory_page_size; - - // setup first page marker - #ifdef PUGIXML_COMPACT - // round-trip through void* to avoid 'cast increases required alignment of target type' warning - page->compact_page_marker = reinterpret_cast(static_cast(reinterpret_cast(page) + sizeof(impl::xml_memory_page))); - *page->compact_page_marker = sizeof(impl::xml_memory_page); - #endif - - // allocate new root - _root = new (reinterpret_cast(page) + sizeof(impl::xml_memory_page) + page_offset) impl::xml_document_struct(page); - _root->prev_sibling_c = _root; - - // setup sentinel page - page->allocator = static_cast(_root); - - // setup hash table pointer in allocator - #ifdef PUGIXML_COMPACT - page->allocator->_hash = &static_cast(_root)->hash; - #endif - - // verify the document allocation - assert(reinterpret_cast(_root) + sizeof(impl::xml_document_struct) <= _memory + sizeof(_memory)); - } - - PUGI__FN void xml_document::_destroy() - { - assert(_root); - - // destroy static storage - if (_buffer) - { - impl::xml_memory::deallocate(_buffer); - _buffer = 0; - } - - // destroy extra buffers (note: no need to destroy linked list nodes, they're allocated using document allocator) - for (impl::xml_extra_buffer* extra = static_cast(_root)->extra_buffers; extra; extra = extra->next) - { - if (extra->buffer) impl::xml_memory::deallocate(extra->buffer); - } - - // destroy dynamic storage, leave sentinel page (it's in static memory) - impl::xml_memory_page* root_page = PUGI__GETPAGE(_root); - assert(root_page && !root_page->prev); - assert(reinterpret_cast(root_page) >= _memory && reinterpret_cast(root_page) < _memory + sizeof(_memory)); - - for (impl::xml_memory_page* page = root_page->next; page; ) - { - impl::xml_memory_page* next = page->next; - - impl::xml_allocator::deallocate_page(page); - - page = next; - } - - #ifdef PUGIXML_COMPACT - // destroy hash table - static_cast(_root)->hash.clear(); - #endif - - _root = 0; - } - -#ifdef PUGIXML_HAS_MOVE - PUGI__FN void xml_document::_move(xml_document& rhs) PUGIXML_NOEXCEPT_IF_NOT_COMPACT - { - impl::xml_document_struct* doc = static_cast(_root); - impl::xml_document_struct* other = static_cast(rhs._root); - - // save first child pointer for later; this needs hash access - xml_node_struct* other_first_child = other->first_child; - - #ifdef PUGIXML_COMPACT - // reserve space for the hash table up front; this is the only operation that can fail - // if it does, we have no choice but to throw (if we have exceptions) - if (other_first_child) - { - size_t other_children = 0; - for (xml_node_struct* node = other_first_child; node; node = node->next_sibling) - other_children++; - - // in compact mode, each pointer assignment could result in a hash table request - // during move, we have to relocate document first_child and parents of all children - // normally there's just one child and its parent has a pointerless encoding but - // we assume the worst here - if (!other->_hash->reserve(other_children + 1)) - { - #ifdef PUGIXML_NO_EXCEPTIONS - return; - #else - throw std::bad_alloc(); - #endif - } - } - #endif - - // move allocation state - doc->_root = other->_root; - doc->_busy_size = other->_busy_size; - - // move buffer state - doc->buffer = other->buffer; - doc->extra_buffers = other->extra_buffers; - _buffer = rhs._buffer; - - #ifdef PUGIXML_COMPACT - // move compact hash; note that the hash table can have pointers to other but they will be "inactive", similarly to nodes removed with remove_child - doc->hash = other->hash; - doc->_hash = &doc->hash; - - // make sure we don't access other hash up until the end when we reinitialize other document - other->_hash = 0; - #endif - - // move page structure - impl::xml_memory_page* doc_page = PUGI__GETPAGE(doc); - assert(doc_page && !doc_page->prev && !doc_page->next); - - impl::xml_memory_page* other_page = PUGI__GETPAGE(other); - assert(other_page && !other_page->prev); - - // relink pages since root page is embedded into xml_document - if (impl::xml_memory_page* page = other_page->next) - { - assert(page->prev == other_page); - - page->prev = doc_page; - - doc_page->next = page; - other_page->next = 0; - } - - // make sure pages point to the correct document state - for (impl::xml_memory_page* page = doc_page->next; page; page = page->next) - { - assert(page->allocator == other); - - page->allocator = doc; - - #ifdef PUGIXML_COMPACT - // this automatically migrates most children between documents and prevents ->parent assignment from allocating - if (page->compact_shared_parent == other) - page->compact_shared_parent = doc; - #endif - } - - // move tree structure - assert(!doc->first_child); - - doc->first_child = other_first_child; - - for (xml_node_struct* node = other_first_child; node; node = node->next_sibling) - { - #ifdef PUGIXML_COMPACT - // most children will have migrated when we reassigned compact_shared_parent - assert(node->parent == other || node->parent == doc); - - node->parent = doc; - #else - assert(node->parent == other); - node->parent = doc; - #endif - } - - // reset other document - new (other) impl::xml_document_struct(PUGI__GETPAGE(other)); - rhs._buffer = 0; - } -#endif - -#ifndef PUGIXML_NO_STL - PUGI__FN xml_parse_result xml_document::load(std::basic_istream >& stream, unsigned int options, xml_encoding encoding) - { - reset(); - - return impl::load_stream_impl(static_cast(_root), stream, options, encoding, &_buffer); - } - - PUGI__FN xml_parse_result xml_document::load(std::basic_istream >& stream, unsigned int options) - { - reset(); - - return impl::load_stream_impl(static_cast(_root), stream, options, encoding_wchar, &_buffer); - } -#endif - - PUGI__FN xml_parse_result xml_document::load_string(const char_t* contents, unsigned int options) - { - // Force native encoding (skip autodetection) - #ifdef PUGIXML_WCHAR_MODE - xml_encoding encoding = encoding_wchar; - #else - xml_encoding encoding = encoding_utf8; - #endif - - return load_buffer(contents, impl::strlength(contents) * sizeof(char_t), options, encoding); - } - - PUGI__FN xml_parse_result xml_document::load(const char_t* contents, unsigned int options) - { - return load_string(contents, options); - } - - PUGI__FN xml_parse_result xml_document::load_file(const char* path_, unsigned int options, xml_encoding encoding) - { - reset(); - - using impl::auto_deleter; // MSVC7 workaround - auto_deleter file(fopen(path_, "rb"), impl::close_file); - - return impl::load_file_impl(static_cast(_root), file.data, options, encoding, &_buffer); - } - - PUGI__FN xml_parse_result xml_document::load_file(const wchar_t* path_, unsigned int options, xml_encoding encoding) - { - reset(); - - using impl::auto_deleter; // MSVC7 workaround - auto_deleter file(impl::open_file_wide(path_, L"rb"), impl::close_file); - - return impl::load_file_impl(static_cast(_root), file.data, options, encoding, &_buffer); - } - - PUGI__FN xml_parse_result xml_document::load_buffer(const void* contents, size_t size, unsigned int options, xml_encoding encoding) - { - reset(); - - return impl::load_buffer_impl(static_cast(_root), _root, const_cast(contents), size, options, encoding, false, false, &_buffer); - } - - PUGI__FN xml_parse_result xml_document::load_buffer_inplace(void* contents, size_t size, unsigned int options, xml_encoding encoding) - { - reset(); - - return impl::load_buffer_impl(static_cast(_root), _root, contents, size, options, encoding, true, false, &_buffer); - } - - PUGI__FN xml_parse_result xml_document::load_buffer_inplace_own(void* contents, size_t size, unsigned int options, xml_encoding encoding) - { - reset(); - - return impl::load_buffer_impl(static_cast(_root), _root, contents, size, options, encoding, true, true, &_buffer); - } - - PUGI__FN void xml_document::save(xml_writer& writer, const char_t* indent, unsigned int flags, xml_encoding encoding) const - { - impl::xml_buffered_writer buffered_writer(writer, encoding); - - if ((flags & format_write_bom) && encoding != encoding_latin1) - { - // BOM always represents the codepoint U+FEFF, so just write it in native encoding - #ifdef PUGIXML_WCHAR_MODE - unsigned int bom = 0xfeff; - buffered_writer.write(static_cast(bom)); - #else - buffered_writer.write('\xef', '\xbb', '\xbf'); - #endif - } - - if (!(flags & format_no_declaration) && !impl::has_declaration(_root)) - { - buffered_writer.write_string(PUGIXML_TEXT("'); - if (!(flags & format_raw)) buffered_writer.write('\n'); - } - - impl::node_output(buffered_writer, _root, indent, flags, 0); - - buffered_writer.flush(); - } - -#ifndef PUGIXML_NO_STL - PUGI__FN void xml_document::save(std::basic_ostream >& stream, const char_t* indent, unsigned int flags, xml_encoding encoding) const - { - xml_writer_stream writer(stream); - - save(writer, indent, flags, encoding); - } - - PUGI__FN void xml_document::save(std::basic_ostream >& stream, const char_t* indent, unsigned int flags) const - { - xml_writer_stream writer(stream); - - save(writer, indent, flags, encoding_wchar); - } -#endif - - PUGI__FN bool xml_document::save_file(const char* path_, const char_t* indent, unsigned int flags, xml_encoding encoding) const - { - using impl::auto_deleter; // MSVC7 workaround - auto_deleter file(fopen(path_, (flags & format_save_file_text) ? "w" : "wb"), impl::close_file); - - return impl::save_file_impl(*this, file.data, indent, flags, encoding); - } - - PUGI__FN bool xml_document::save_file(const wchar_t* path_, const char_t* indent, unsigned int flags, xml_encoding encoding) const - { - using impl::auto_deleter; // MSVC7 workaround - auto_deleter file(impl::open_file_wide(path_, (flags & format_save_file_text) ? L"w" : L"wb"), impl::close_file); - - return impl::save_file_impl(*this, file.data, indent, flags, encoding); - } - - PUGI__FN xml_node xml_document::document_element() const - { - assert(_root); - - for (xml_node_struct* i = _root->first_child; i; i = i->next_sibling) - if (PUGI__NODETYPE(i) == node_element) + for (xml_node_struct* i = _root->first_child; i; i = i->next_sibling) + for (xml_attribute_struct* a = i->first_attribute; a; a = a->next_attribute) + if (a->name && impl::strequal(attr_name, a->name) && impl::strequal(attr_value, a->value ? a->value + 0 : PUGIXML_TEXT(""))) return xml_node(i); - return xml_node(); - } + return xml_node(); +} #ifndef PUGIXML_NO_STL - PUGI__FN std::string PUGIXML_FUNCTION as_utf8(const wchar_t* str) - { - assert(str); +PUGI__FN string_t xml_node::path(char_t delimiter) const +{ + if (!_root) + return string_t(); - return impl::as_utf8_impl(str, impl::strlength_wide(str)); + size_t offset = 0; + + for (xml_node_struct* i = _root; i; i = i->parent) + { + offset += (i != _root); + offset += i->name ? impl::strlength(i->name) : 0; } - PUGI__FN std::string PUGIXML_FUNCTION as_utf8(const std::basic_string& str) + string_t result; + result.resize(offset); + + for (xml_node_struct* j = _root; j; j = j->parent) { - return impl::as_utf8_impl(str.c_str(), str.size()); + if (j != _root) + result[--offset] = delimiter; + + if (j->name) + { + size_t length = impl::strlength(j->name); + + offset -= length; + memcpy(&result[offset], j->name, length * sizeof(char_t)); + } } - PUGI__FN std::basic_string PUGIXML_FUNCTION as_wide(const char* str) - { - assert(str); + assert(offset == 0); - return impl::as_wide_impl(str, strlen(str)); + return result; +} +#endif + +PUGI__FN xml_node xml_node::first_element_by_path(const char_t* path_, char_t delimiter) const +{ + xml_node found = *this; // Current search context. + + if (!_root || !path_[0]) + return found; + + if (path_[0] == delimiter) + { + // Absolute path; e.g. '/foo/bar' + found = found.root(); + ++path_; } - PUGI__FN std::basic_string PUGIXML_FUNCTION as_wide(const std::string& str) + const char_t* path_segment = path_; + + while (*path_segment == delimiter) + ++path_segment; + + const char_t* path_segment_end = path_segment; + + while (*path_segment_end && *path_segment_end != delimiter) + ++path_segment_end; + + if (path_segment == path_segment_end) + return found; + + const char_t* next_segment = path_segment_end; + + while (*next_segment == delimiter) + ++next_segment; + + if (*path_segment == '.' && path_segment + 1 == path_segment_end) + return found.first_element_by_path(next_segment, delimiter); + else if (*path_segment == '.' && *(path_segment + 1) == '.' && path_segment + 2 == path_segment_end) + return found.parent().first_element_by_path(next_segment, delimiter); + else { - return impl::as_wide_impl(str.c_str(), str.size()); + for (xml_node_struct* j = found._root->first_child; j; j = j->next_sibling) + { + if (j->name && impl::strequalrange(j->name, path_segment, static_cast(path_segment_end - path_segment))) + { + xml_node subsearch = xml_node(j).first_element_by_path(next_segment, delimiter); + + if (subsearch) + return subsearch; + } + } + + return xml_node(); + } +} + +PUGI__FN bool xml_node::traverse(xml_tree_walker& walker) +{ + walker._depth = -1; + + xml_node arg_begin(_root); + if (!walker.begin(arg_begin)) + return false; + + xml_node_struct* cur = _root ? _root->first_child + 0 : 0; + + if (cur) + { + ++walker._depth; + + do + { + xml_node arg_for_each(cur); + if (!walker.for_each(arg_for_each)) + return false; + + if (cur->first_child) + { + ++walker._depth; + cur = cur->first_child; + } + else if (cur->next_sibling) + cur = cur->next_sibling; + else + { + while (!cur->next_sibling && cur != _root && cur->parent) + { + --walker._depth; + cur = cur->parent; + } + + if (cur != _root) + cur = cur->next_sibling; + } + } while (cur && cur != _root); + } + + assert(walker._depth == -1); + + xml_node arg_end(_root); + return walker.end(arg_end); +} + +PUGI__FN size_t xml_node::hash_value() const +{ + return static_cast(reinterpret_cast(_root) / sizeof(xml_node_struct)); +} + +PUGI__FN xml_node_struct* xml_node::internal_object() const +{ + return _root; +} + +PUGI__FN void xml_node::print(xml_writer& writer, const char_t* indent, unsigned int flags, xml_encoding encoding, unsigned int depth) const +{ + if (!_root) + return; + + impl::xml_buffered_writer buffered_writer(writer, encoding); + + impl::node_output(buffered_writer, _root, indent, flags, depth); + + buffered_writer.flush(); +} + +#ifndef PUGIXML_NO_STL +PUGI__FN void xml_node::print(std::basic_ostream>& stream, const char_t* indent, unsigned int flags, xml_encoding encoding, unsigned int depth) const +{ + xml_writer_stream writer(stream); + + print(writer, indent, flags, encoding, depth); +} + +PUGI__FN void xml_node::print(std::basic_ostream>& stream, const char_t* indent, unsigned int flags, unsigned int depth) const +{ + xml_writer_stream writer(stream); + + print(writer, indent, flags, encoding_wchar, depth); +} +#endif + +PUGI__FN ptrdiff_t xml_node::offset_debug() const +{ + if (!_root) + return -1; + + impl::xml_document_struct& doc = impl::get_document(_root); + + // we can determine the offset reliably only if there is exactly once parse buffer + if (!doc.buffer || doc.extra_buffers) + return -1; + + switch (type()) + { + case node_document: + return 0; + + case node_element: + case node_declaration: + case node_pi: + return _root->name && (_root->header & impl::xml_memory_page_name_allocated_or_shared_mask) == 0 ? _root->name - doc.buffer : -1; + + case node_pcdata: + case node_cdata: + case node_comment: + case node_doctype: + return _root->value && (_root->header & impl::xml_memory_page_value_allocated_or_shared_mask) == 0 ? _root->value - doc.buffer : -1; + + default: + assert(false && "Invalid node type"); // unreachable + return -1; + } +} + +#ifdef __BORLANDC__ +PUGI__FN bool operator&&(const xml_node& lhs, bool rhs) +{ + return (bool)lhs && rhs; +} + +PUGI__FN bool operator||(const xml_node& lhs, bool rhs) +{ + return (bool)lhs || rhs; +} +#endif + +PUGI__FN xml_text::xml_text(xml_node_struct* root) + : _root(root) +{ +} + +PUGI__FN xml_node_struct* xml_text::_data() const +{ + if (!_root || impl::is_text_node(_root)) + return _root; + + // element nodes can have value if parse_embed_pcdata was used + if (PUGI__NODETYPE(_root) == node_element && _root->value) + return _root; + + for (xml_node_struct* node = _root->first_child; node; node = node->next_sibling) + if (impl::is_text_node(node)) + return node; + + return 0; +} + +PUGI__FN xml_node_struct* xml_text::_data_new() +{ + xml_node_struct* d = _data(); + if (d) + return d; + + return xml_node(_root).append_child(node_pcdata).internal_object(); +} + +PUGI__FN xml_text::xml_text() + : _root(0) +{ +} + +PUGI__FN static void unspecified_bool_xml_text(xml_text***) +{ +} + +PUGI__FN xml_text::operator xml_text::unspecified_bool_type() const +{ + return _data() ? unspecified_bool_xml_text : 0; +} + +PUGI__FN bool xml_text::operator!() const +{ + return !_data(); +} + +PUGI__FN bool xml_text::empty() const +{ + return _data() == 0; +} + +PUGI__FN const char_t* xml_text::get() const +{ + xml_node_struct* d = _data(); + + return (d && d->value) ? d->value + 0 : PUGIXML_TEXT(""); +} + +PUGI__FN const char_t* xml_text::as_string(const char_t* def) const +{ + xml_node_struct* d = _data(); + + return (d && d->value) ? d->value + 0 : def; +} + +PUGI__FN int xml_text::as_int(int def) const +{ + xml_node_struct* d = _data(); + + return (d && d->value) ? impl::get_value_int(d->value) : def; +} + +PUGI__FN unsigned int xml_text::as_uint(unsigned int def) const +{ + xml_node_struct* d = _data(); + + return (d && d->value) ? impl::get_value_uint(d->value) : def; +} + +PUGI__FN double xml_text::as_double(double def) const +{ + xml_node_struct* d = _data(); + + return (d && d->value) ? impl::get_value_double(d->value) : def; +} + +PUGI__FN float xml_text::as_float(float def) const +{ + xml_node_struct* d = _data(); + + return (d && d->value) ? impl::get_value_float(d->value) : def; +} + +PUGI__FN bool xml_text::as_bool(bool def) const +{ + xml_node_struct* d = _data(); + + return (d && d->value) ? impl::get_value_bool(d->value) : def; +} + +#ifdef PUGIXML_HAS_LONG_LONG +PUGI__FN long long xml_text::as_llong(long long def) const +{ + xml_node_struct* d = _data(); + + return (d && d->value) ? impl::get_value_llong(d->value) : def; +} + +PUGI__FN unsigned long long xml_text::as_ullong(unsigned long long def) const +{ + xml_node_struct* d = _data(); + + return (d && d->value) ? impl::get_value_ullong(d->value) : def; +} +#endif + +PUGI__FN bool xml_text::set(const char_t* rhs) +{ + xml_node_struct* dn = _data_new(); + + return dn ? impl::strcpy_insitu(dn->value, dn->header, impl::xml_memory_page_value_allocated_mask, rhs, impl::strlength(rhs)) : false; +} + +PUGI__FN bool xml_text::set(int rhs) +{ + xml_node_struct* dn = _data_new(); + + return dn ? impl::set_value_integer(dn->value, dn->header, impl::xml_memory_page_value_allocated_mask, rhs, rhs < 0) : false; +} + +PUGI__FN bool xml_text::set(unsigned int rhs) +{ + xml_node_struct* dn = _data_new(); + + return dn ? impl::set_value_integer(dn->value, dn->header, impl::xml_memory_page_value_allocated_mask, rhs, false) : false; +} + +PUGI__FN bool xml_text::set(long rhs) +{ + xml_node_struct* dn = _data_new(); + + return dn ? impl::set_value_integer(dn->value, dn->header, impl::xml_memory_page_value_allocated_mask, rhs, rhs < 0) : false; +} + +PUGI__FN bool xml_text::set(unsigned long rhs) +{ + xml_node_struct* dn = _data_new(); + + return dn ? impl::set_value_integer(dn->value, dn->header, impl::xml_memory_page_value_allocated_mask, rhs, false) : false; +} + +PUGI__FN bool xml_text::set(float rhs) +{ + xml_node_struct* dn = _data_new(); + + return dn ? impl::set_value_convert(dn->value, dn->header, impl::xml_memory_page_value_allocated_mask, rhs) : false; +} + +PUGI__FN bool xml_text::set(double rhs) +{ + xml_node_struct* dn = _data_new(); + + return dn ? impl::set_value_convert(dn->value, dn->header, impl::xml_memory_page_value_allocated_mask, rhs) : false; +} + +PUGI__FN bool xml_text::set(bool rhs) +{ + xml_node_struct* dn = _data_new(); + + return dn ? impl::set_value_bool(dn->value, dn->header, impl::xml_memory_page_value_allocated_mask, rhs) : false; +} + +#ifdef PUGIXML_HAS_LONG_LONG +PUGI__FN bool xml_text::set(long long rhs) +{ + xml_node_struct* dn = _data_new(); + + return dn ? impl::set_value_integer(dn->value, dn->header, impl::xml_memory_page_value_allocated_mask, rhs, rhs < 0) : false; +} + +PUGI__FN bool xml_text::set(unsigned long long rhs) +{ + xml_node_struct* dn = _data_new(); + + return dn ? impl::set_value_integer(dn->value, dn->header, impl::xml_memory_page_value_allocated_mask, rhs, false) : false; +} +#endif + +PUGI__FN xml_text& xml_text::operator=(const char_t* rhs) +{ + set(rhs); + return *this; +} + +PUGI__FN xml_text& xml_text::operator=(int rhs) +{ + set(rhs); + return *this; +} + +PUGI__FN xml_text& xml_text::operator=(unsigned int rhs) +{ + set(rhs); + return *this; +} + +PUGI__FN xml_text& xml_text::operator=(long rhs) +{ + set(rhs); + return *this; +} + +PUGI__FN xml_text& xml_text::operator=(unsigned long rhs) +{ + set(rhs); + return *this; +} + +PUGI__FN xml_text& xml_text::operator=(double rhs) +{ + set(rhs); + return *this; +} + +PUGI__FN xml_text& xml_text::operator=(float rhs) +{ + set(rhs); + return *this; +} + +PUGI__FN xml_text& xml_text::operator=(bool rhs) +{ + set(rhs); + return *this; +} + +#ifdef PUGIXML_HAS_LONG_LONG +PUGI__FN xml_text& xml_text::operator=(long long rhs) +{ + set(rhs); + return *this; +} + +PUGI__FN xml_text& xml_text::operator=(unsigned long long rhs) +{ + set(rhs); + return *this; +} +#endif + +PUGI__FN xml_node xml_text::data() const +{ + return xml_node(_data()); +} + +#ifdef __BORLANDC__ +PUGI__FN bool operator&&(const xml_text& lhs, bool rhs) +{ + return (bool)lhs && rhs; +} + +PUGI__FN bool operator||(const xml_text& lhs, bool rhs) +{ + return (bool)lhs || rhs; +} +#endif + +PUGI__FN xml_node_iterator::xml_node_iterator() +{ +} + +PUGI__FN xml_node_iterator::xml_node_iterator(const xml_node& node) + : _wrap(node) + , _parent(node.parent()) +{ +} + +PUGI__FN xml_node_iterator::xml_node_iterator(xml_node_struct* ref, xml_node_struct* parent) + : _wrap(ref) + , _parent(parent) +{ +} + +PUGI__FN bool xml_node_iterator::operator==(const xml_node_iterator& rhs) const +{ + return _wrap._root == rhs._wrap._root && _parent._root == rhs._parent._root; +} + +PUGI__FN bool xml_node_iterator::operator!=(const xml_node_iterator& rhs) const +{ + return _wrap._root != rhs._wrap._root || _parent._root != rhs._parent._root; +} + +PUGI__FN xml_node& xml_node_iterator::operator*() const +{ + assert(_wrap._root); + return _wrap; +} + +PUGI__FN xml_node* xml_node_iterator::operator->() const +{ + assert(_wrap._root); + return const_cast(&_wrap); // BCC5 workaround +} + +PUGI__FN const xml_node_iterator& xml_node_iterator::operator++() +{ + assert(_wrap._root); + _wrap._root = _wrap._root->next_sibling; + return *this; +} + +PUGI__FN xml_node_iterator xml_node_iterator::operator++(int) +{ + xml_node_iterator temp = *this; + ++*this; + return temp; +} + +PUGI__FN const xml_node_iterator& xml_node_iterator::operator--() +{ + _wrap = _wrap._root ? _wrap.previous_sibling() : _parent.last_child(); + return *this; +} + +PUGI__FN xml_node_iterator xml_node_iterator::operator--(int) +{ + xml_node_iterator temp = *this; + --*this; + return temp; +} + +PUGI__FN xml_attribute_iterator::xml_attribute_iterator() +{ +} + +PUGI__FN xml_attribute_iterator::xml_attribute_iterator(const xml_attribute& attr, const xml_node& parent) + : _wrap(attr) + , _parent(parent) +{ +} + +PUGI__FN xml_attribute_iterator::xml_attribute_iterator(xml_attribute_struct* ref, xml_node_struct* parent) + : _wrap(ref) + , _parent(parent) +{ +} + +PUGI__FN bool xml_attribute_iterator::operator==(const xml_attribute_iterator& rhs) const +{ + return _wrap._attr == rhs._wrap._attr && _parent._root == rhs._parent._root; +} + +PUGI__FN bool xml_attribute_iterator::operator!=(const xml_attribute_iterator& rhs) const +{ + return _wrap._attr != rhs._wrap._attr || _parent._root != rhs._parent._root; +} + +PUGI__FN xml_attribute& xml_attribute_iterator::operator*() const +{ + assert(_wrap._attr); + return _wrap; +} + +PUGI__FN xml_attribute* xml_attribute_iterator::operator->() const +{ + assert(_wrap._attr); + return const_cast(&_wrap); // BCC5 workaround +} + +PUGI__FN const xml_attribute_iterator& xml_attribute_iterator::operator++() +{ + assert(_wrap._attr); + _wrap._attr = _wrap._attr->next_attribute; + return *this; +} + +PUGI__FN xml_attribute_iterator xml_attribute_iterator::operator++(int) +{ + xml_attribute_iterator temp = *this; + ++*this; + return temp; +} + +PUGI__FN const xml_attribute_iterator& xml_attribute_iterator::operator--() +{ + _wrap = _wrap._attr ? _wrap.previous_attribute() : _parent.last_attribute(); + return *this; +} + +PUGI__FN xml_attribute_iterator xml_attribute_iterator::operator--(int) +{ + xml_attribute_iterator temp = *this; + --*this; + return temp; +} + +PUGI__FN xml_named_node_iterator::xml_named_node_iterator() + : _name(0) +{ +} + +PUGI__FN xml_named_node_iterator::xml_named_node_iterator(const xml_node& node, const char_t* name) + : _wrap(node) + , _parent(node.parent()) + , _name(name) +{ +} + +PUGI__FN xml_named_node_iterator::xml_named_node_iterator(xml_node_struct* ref, xml_node_struct* parent, const char_t* name) + : _wrap(ref) + , _parent(parent) + , _name(name) +{ +} + +PUGI__FN bool xml_named_node_iterator::operator==(const xml_named_node_iterator& rhs) const +{ + return _wrap._root == rhs._wrap._root && _parent._root == rhs._parent._root; +} + +PUGI__FN bool xml_named_node_iterator::operator!=(const xml_named_node_iterator& rhs) const +{ + return _wrap._root != rhs._wrap._root || _parent._root != rhs._parent._root; +} + +PUGI__FN xml_node& xml_named_node_iterator::operator*() const +{ + assert(_wrap._root); + return _wrap; +} + +PUGI__FN xml_node* xml_named_node_iterator::operator->() const +{ + assert(_wrap._root); + return const_cast(&_wrap); // BCC5 workaround +} + +PUGI__FN const xml_named_node_iterator& xml_named_node_iterator::operator++() +{ + assert(_wrap._root); + _wrap = _wrap.next_sibling(_name); + return *this; +} + +PUGI__FN xml_named_node_iterator xml_named_node_iterator::operator++(int) +{ + xml_named_node_iterator temp = *this; + ++*this; + return temp; +} + +PUGI__FN const xml_named_node_iterator& xml_named_node_iterator::operator--() +{ + if (_wrap._root) + _wrap = _wrap.previous_sibling(_name); + else + { + _wrap = _parent.last_child(); + + if (!impl::strequal(_wrap.name(), _name)) + _wrap = _wrap.previous_sibling(_name); + } + + return *this; +} + +PUGI__FN xml_named_node_iterator xml_named_node_iterator::operator--(int) +{ + xml_named_node_iterator temp = *this; + --*this; + return temp; +} + +PUGI__FN xml_parse_result::xml_parse_result() + : status(status_internal_error) + , offset(0) + , encoding(encoding_auto) +{ +} + +PUGI__FN xml_parse_result::operator bool() const +{ + return status == status_ok; +} + +PUGI__FN const char* xml_parse_result::description() const +{ + switch (status) + { + case status_ok: + return "No error"; + + case status_file_not_found: + return "File was not found"; + case status_io_error: + return "Error reading from file/stream"; + case status_out_of_memory: + return "Could not allocate memory"; + case status_internal_error: + return "Internal error occurred"; + + case status_unrecognized_tag: + return "Could not determine tag type"; + + case status_bad_pi: + return "Error parsing document declaration/processing instruction"; + case status_bad_comment: + return "Error parsing comment"; + case status_bad_cdata: + return "Error parsing CDATA section"; + case status_bad_doctype: + return "Error parsing document type declaration"; + case status_bad_pcdata: + return "Error parsing PCDATA section"; + case status_bad_start_element: + return "Error parsing start element tag"; + case status_bad_attribute: + return "Error parsing element attribute"; + case status_bad_end_element: + return "Error parsing end element tag"; + case status_end_element_mismatch: + return "Start-end tags mismatch"; + + case status_append_invalid_root: + return "Unable to append nodes: root is not an element or document"; + + case status_no_document_element: + return "No document element found"; + + default: + return "Unknown error"; + } +} + +PUGI__FN xml_document::xml_document() + : _buffer(0) +{ + _create(); +} + +PUGI__FN xml_document::~xml_document() +{ + _destroy(); +} + +#ifdef PUGIXML_HAS_MOVE +PUGI__FN xml_document::xml_document(xml_document&& rhs) PUGIXML_NOEXCEPT_IF_NOT_COMPACT : _buffer(0) +{ + _create(); + _move(rhs); +} + +PUGI__FN xml_document& xml_document::operator=(xml_document&& rhs) PUGIXML_NOEXCEPT_IF_NOT_COMPACT +{ + if (this == &rhs) + return *this; + + _destroy(); + _create(); + _move(rhs); + + return *this; +} +#endif + +PUGI__FN void xml_document::reset() +{ + _destroy(); + _create(); +} + +PUGI__FN void xml_document::reset(const xml_document& proto) +{ + reset(); + + for (xml_node cur = proto.first_child(); cur; cur = cur.next_sibling()) + append_copy(cur); +} + +PUGI__FN void xml_document::_create() +{ + assert(!_root); + +#ifdef PUGIXML_COMPACT + // space for page marker for the first page (uint32_t), rounded up to pointer size; assumes pointers are at least 32-bit + const size_t page_offset = sizeof(void*); +#else + const size_t page_offset = 0; +#endif + + // initialize sentinel page + PUGI__STATIC_ASSERT(sizeof(impl::xml_memory_page) + sizeof(impl::xml_document_struct) + page_offset <= sizeof(_memory)); + + // prepare page structure + impl::xml_memory_page* page = impl::xml_memory_page::construct(_memory); + assert(page); + + page->busy_size = impl::xml_memory_page_size; + + // setup first page marker +#ifdef PUGIXML_COMPACT + // round-trip through void* to avoid 'cast increases required alignment of target type' warning + page->compact_page_marker = reinterpret_cast(static_cast(reinterpret_cast(page) + sizeof(impl::xml_memory_page))); + *page->compact_page_marker = sizeof(impl::xml_memory_page); +#endif + + // allocate new root + _root = new (reinterpret_cast(page) + sizeof(impl::xml_memory_page) + page_offset) impl::xml_document_struct(page); + _root->prev_sibling_c = _root; + + // setup sentinel page + page->allocator = static_cast(_root); + + // setup hash table pointer in allocator +#ifdef PUGIXML_COMPACT + page->allocator->_hash = &static_cast(_root)->hash; +#endif + + // verify the document allocation + assert(reinterpret_cast(_root) + sizeof(impl::xml_document_struct) <= _memory + sizeof(_memory)); +} + +PUGI__FN void xml_document::_destroy() +{ + assert(_root); + + // destroy static storage + if (_buffer) + { + impl::xml_memory::deallocate(_buffer); + _buffer = 0; + } + + // destroy extra buffers (note: no need to destroy linked list nodes, they're allocated using document allocator) + for (impl::xml_extra_buffer* extra = static_cast(_root)->extra_buffers; extra; extra = extra->next) + { + if (extra->buffer) + impl::xml_memory::deallocate(extra->buffer); + } + + // destroy dynamic storage, leave sentinel page (it's in static memory) + impl::xml_memory_page* root_page = PUGI__GETPAGE(_root); + assert(root_page && !root_page->prev); + assert(reinterpret_cast(root_page) >= _memory && reinterpret_cast(root_page) < _memory + sizeof(_memory)); + + for (impl::xml_memory_page* page = root_page->next; page;) + { + impl::xml_memory_page* next = page->next; + + impl::xml_allocator::deallocate_page(page); + + page = next; + } + +#ifdef PUGIXML_COMPACT + // destroy hash table + static_cast(_root)->hash.clear(); +#endif + + _root = 0; +} + +#ifdef PUGIXML_HAS_MOVE +PUGI__FN void xml_document::_move(xml_document& rhs) PUGIXML_NOEXCEPT_IF_NOT_COMPACT +{ + impl::xml_document_struct* doc = static_cast(_root); + impl::xml_document_struct* other = static_cast(rhs._root); + + // save first child pointer for later; this needs hash access + xml_node_struct* other_first_child = other->first_child; + +#ifdef PUGIXML_COMPACT + // reserve space for the hash table up front; this is the only operation that can fail + // if it does, we have no choice but to throw (if we have exceptions) + if (other_first_child) + { + size_t other_children = 0; + for (xml_node_struct* node = other_first_child; node; node = node->next_sibling) + other_children++; + + // in compact mode, each pointer assignment could result in a hash table request + // during move, we have to relocate document first_child and parents of all children + // normally there's just one child and its parent has a pointerless encoding but + // we assume the worst here + if (!other->_hash->reserve(other_children + 1)) + { +#ifdef PUGIXML_NO_EXCEPTIONS + return; +#else + throw std::bad_alloc(); +#endif + } } #endif - PUGI__FN void PUGIXML_FUNCTION set_memory_management_functions(allocation_function allocate, deallocation_function deallocate) + // move allocation state + doc->_root = other->_root; + doc->_busy_size = other->_busy_size; + + // move buffer state + doc->buffer = other->buffer; + doc->extra_buffers = other->extra_buffers; + _buffer = rhs._buffer; + +#ifdef PUGIXML_COMPACT + // move compact hash; note that the hash table can have pointers to other but they will be "inactive", similarly to nodes removed with remove_child + doc->hash = other->hash; + doc->_hash = &doc->hash; + + // make sure we don't access other hash up until the end when we reinitialize other document + other->_hash = 0; +#endif + + // move page structure + impl::xml_memory_page* doc_page = PUGI__GETPAGE(doc); + assert(doc_page && !doc_page->prev && !doc_page->next); + + impl::xml_memory_page* other_page = PUGI__GETPAGE(other); + assert(other_page && !other_page->prev); + + // relink pages since root page is embedded into xml_document + if (impl::xml_memory_page* page = other_page->next) { - impl::xml_memory::allocate = allocate; - impl::xml_memory::deallocate = deallocate; + assert(page->prev == other_page); + + page->prev = doc_page; + + doc_page->next = page; + other_page->next = 0; } - PUGI__FN allocation_function PUGIXML_FUNCTION get_memory_allocation_function() + // make sure pages point to the correct document state + for (impl::xml_memory_page* page = doc_page->next; page; page = page->next) { - return impl::xml_memory::allocate; + assert(page->allocator == other); + + page->allocator = doc; + +#ifdef PUGIXML_COMPACT + // this automatically migrates most children between documents and prevents ->parent assignment from allocating + if (page->compact_shared_parent == other) + page->compact_shared_parent = doc; +#endif } - PUGI__FN deallocation_function PUGIXML_FUNCTION get_memory_deallocation_function() + // move tree structure + assert(!doc->first_child); + + doc->first_child = other_first_child; + + for (xml_node_struct* node = other_first_child; node; node = node->next_sibling) { - return impl::xml_memory::deallocate; +#ifdef PUGIXML_COMPACT + // most children will have migrated when we reassigned compact_shared_parent + assert(node->parent == other || node->parent == doc); + + node->parent = doc; +#else + assert(node->parent == other); + node->parent = doc; +#endif } + + // reset other document + new (other) impl::xml_document_struct(PUGI__GETPAGE(other)); + rhs._buffer = 0; } +#endif + +#ifndef PUGIXML_NO_STL +PUGI__FN xml_parse_result xml_document::load(std::basic_istream>& stream, unsigned int options, xml_encoding encoding) +{ + reset(); + + return impl::load_stream_impl(static_cast(_root), stream, options, encoding, &_buffer); +} + +PUGI__FN xml_parse_result xml_document::load(std::basic_istream>& stream, unsigned int options) +{ + reset(); + + return impl::load_stream_impl(static_cast(_root), stream, options, encoding_wchar, &_buffer); +} +#endif + +PUGI__FN xml_parse_result xml_document::load_string(const char_t* contents, unsigned int options) +{ + // Force native encoding (skip autodetection) +#ifdef PUGIXML_WCHAR_MODE + xml_encoding encoding = encoding_wchar; +#else + xml_encoding encoding = encoding_utf8; +#endif + + return load_buffer(contents, impl::strlength(contents) * sizeof(char_t), options, encoding); +} + +PUGI__FN xml_parse_result xml_document::load(const char_t* contents, unsigned int options) +{ + return load_string(contents, options); +} + +PUGI__FN xml_parse_result xml_document::load_file(const char* path_, unsigned int options, xml_encoding encoding) +{ + reset(); + + using impl::auto_deleter; // MSVC7 workaround + auto_deleter file(fopen(path_, "rb"), impl::close_file); + + return impl::load_file_impl(static_cast(_root), file.data, options, encoding, &_buffer); +} + +PUGI__FN xml_parse_result xml_document::load_file(const wchar_t* path_, unsigned int options, xml_encoding encoding) +{ + reset(); + + using impl::auto_deleter; // MSVC7 workaround + auto_deleter file(impl::open_file_wide(path_, L"rb"), impl::close_file); + + return impl::load_file_impl(static_cast(_root), file.data, options, encoding, &_buffer); +} + +PUGI__FN xml_parse_result xml_document::load_buffer(const void* contents, size_t size, unsigned int options, xml_encoding encoding) +{ + reset(); + + return impl::load_buffer_impl(static_cast(_root), _root, const_cast(contents), size, options, encoding, false, false, &_buffer); +} + +PUGI__FN xml_parse_result xml_document::load_buffer_inplace(void* contents, size_t size, unsigned int options, xml_encoding encoding) +{ + reset(); + + return impl::load_buffer_impl(static_cast(_root), _root, contents, size, options, encoding, true, false, &_buffer); +} + +PUGI__FN xml_parse_result xml_document::load_buffer_inplace_own(void* contents, size_t size, unsigned int options, xml_encoding encoding) +{ + reset(); + + return impl::load_buffer_impl(static_cast(_root), _root, contents, size, options, encoding, true, true, &_buffer); +} + +PUGI__FN void xml_document::save(xml_writer& writer, const char_t* indent, unsigned int flags, xml_encoding encoding) const +{ + impl::xml_buffered_writer buffered_writer(writer, encoding); + + if ((flags & format_write_bom) && encoding != encoding_latin1) + { + // BOM always represents the codepoint U+FEFF, so just write it in native encoding +#ifdef PUGIXML_WCHAR_MODE + unsigned int bom = 0xfeff; + buffered_writer.write(static_cast(bom)); +#else + buffered_writer.write('\xef', '\xbb', '\xbf'); +#endif + } + + if (!(flags & format_no_declaration) && !impl::has_declaration(_root)) + { + buffered_writer.write_string(PUGIXML_TEXT("'); + if (!(flags & format_raw)) + buffered_writer.write('\n'); + } + + impl::node_output(buffered_writer, _root, indent, flags, 0); + + buffered_writer.flush(); +} + +#ifndef PUGIXML_NO_STL +PUGI__FN void xml_document::save(std::basic_ostream>& stream, const char_t* indent, unsigned int flags, xml_encoding encoding) const +{ + xml_writer_stream writer(stream); + + save(writer, indent, flags, encoding); +} + +PUGI__FN void xml_document::save(std::basic_ostream>& stream, const char_t* indent, unsigned int flags) const +{ + xml_writer_stream writer(stream); + + save(writer, indent, flags, encoding_wchar); +} +#endif + +PUGI__FN bool xml_document::save_file(const char* path_, const char_t* indent, unsigned int flags, xml_encoding encoding) const +{ + using impl::auto_deleter; // MSVC7 workaround + auto_deleter file(fopen(path_, (flags & format_save_file_text) ? "w" : "wb"), impl::close_file); + + return impl::save_file_impl(*this, file.data, indent, flags, encoding); +} + +PUGI__FN bool xml_document::save_file(const wchar_t* path_, const char_t* indent, unsigned int flags, xml_encoding encoding) const +{ + using impl::auto_deleter; // MSVC7 workaround + auto_deleter file(impl::open_file_wide(path_, (flags & format_save_file_text) ? L"w" : L"wb"), impl::close_file); + + return impl::save_file_impl(*this, file.data, indent, flags, encoding); +} + +PUGI__FN xml_node xml_document::document_element() const +{ + assert(_root); + + for (xml_node_struct* i = _root->first_child; i; i = i->next_sibling) + if (PUGI__NODETYPE(i) == node_element) + return xml_node(i); + + return xml_node(); +} + +#ifndef PUGIXML_NO_STL +PUGI__FN std::string PUGIXML_FUNCTION as_utf8(const wchar_t* str) +{ + assert(str); + + return impl::as_utf8_impl(str, impl::strlength_wide(str)); +} + +PUGI__FN std::string PUGIXML_FUNCTION as_utf8(const std::basic_string& str) +{ + return impl::as_utf8_impl(str.c_str(), str.size()); +} + +PUGI__FN std::basic_string PUGIXML_FUNCTION as_wide(const char* str) +{ + assert(str); + + return impl::as_wide_impl(str, strlen(str)); +} + +PUGI__FN std::basic_string PUGIXML_FUNCTION as_wide(const std::string& str) +{ + return impl::as_wide_impl(str.c_str(), str.size()); +} +#endif + +PUGI__FN void PUGIXML_FUNCTION set_memory_management_functions(allocation_function allocate, deallocation_function deallocate) +{ + impl::xml_memory::allocate = allocate; + impl::xml_memory::deallocate = deallocate; +} + +PUGI__FN allocation_function PUGIXML_FUNCTION get_memory_allocation_function() +{ + return impl::xml_memory::allocate; +} + +PUGI__FN deallocation_function PUGIXML_FUNCTION get_memory_deallocation_function() +{ + return impl::xml_memory::deallocate; +} +} // namespace pugi #if !defined(PUGIXML_NO_STL) && (defined(_MSC_VER) || defined(__ICC)) namespace std { - // Workarounds for (non-standard) iterator category detection for older versions (MSVC7/IC8 and earlier) - PUGI__FN std::bidirectional_iterator_tag _Iter_cat(const pugi::xml_node_iterator&) - { - return std::bidirectional_iterator_tag(); - } - - PUGI__FN std::bidirectional_iterator_tag _Iter_cat(const pugi::xml_attribute_iterator&) - { - return std::bidirectional_iterator_tag(); - } - - PUGI__FN std::bidirectional_iterator_tag _Iter_cat(const pugi::xml_named_node_iterator&) - { - return std::bidirectional_iterator_tag(); - } +// Workarounds for (non-standard) iterator category detection for older versions (MSVC7/IC8 and earlier) +PUGI__FN std::bidirectional_iterator_tag _Iter_cat(const pugi::xml_node_iterator&) +{ + return std::bidirectional_iterator_tag(); } + +PUGI__FN std::bidirectional_iterator_tag _Iter_cat(const pugi::xml_attribute_iterator&) +{ + return std::bidirectional_iterator_tag(); +} + +PUGI__FN std::bidirectional_iterator_tag _Iter_cat(const pugi::xml_named_node_iterator&) +{ + return std::bidirectional_iterator_tag(); +} +} // namespace std #endif #if !defined(PUGIXML_NO_STL) && defined(__SUNPRO_CC) namespace std { - // Workarounds for (non-standard) iterator category detection - PUGI__FN std::bidirectional_iterator_tag __iterator_category(const pugi::xml_node_iterator&) - { - return std::bidirectional_iterator_tag(); - } - - PUGI__FN std::bidirectional_iterator_tag __iterator_category(const pugi::xml_attribute_iterator&) - { - return std::bidirectional_iterator_tag(); - } - - PUGI__FN std::bidirectional_iterator_tag __iterator_category(const pugi::xml_named_node_iterator&) - { - return std::bidirectional_iterator_tag(); - } +// Workarounds for (non-standard) iterator category detection +PUGI__FN std::bidirectional_iterator_tag __iterator_category(const pugi::xml_node_iterator&) +{ + return std::bidirectional_iterator_tag(); } + +PUGI__FN std::bidirectional_iterator_tag __iterator_category(const pugi::xml_attribute_iterator&) +{ + return std::bidirectional_iterator_tag(); +} + +PUGI__FN std::bidirectional_iterator_tag __iterator_category(const pugi::xml_named_node_iterator&) +{ + return std::bidirectional_iterator_tag(); +} +} // namespace std #endif #ifndef PUGIXML_NO_XPATH // STL replacements PUGI__NS_BEGIN - struct equal_to +struct equal_to +{ + template + bool operator()(const T& lhs, const T& rhs) const { - template bool operator()(const T& lhs, const T& rhs) const - { - return lhs == rhs; - } - }; + return lhs == rhs; + } +}; - struct not_equal_to +struct not_equal_to +{ + template + bool operator()(const T& lhs, const T& rhs) const { - template bool operator()(const T& lhs, const T& rhs) const - { - return lhs != rhs; - } - }; + return lhs != rhs; + } +}; - struct less +struct less +{ + template + bool operator()(const T& lhs, const T& rhs) const { - template bool operator()(const T& lhs, const T& rhs) const - { - return lhs < rhs; - } - }; + return lhs < rhs; + } +}; - struct less_equal +struct less_equal +{ + template + bool operator()(const T& lhs, const T& rhs) const { - template bool operator()(const T& lhs, const T& rhs) const - { - return lhs <= rhs; - } - }; + return lhs <= rhs; + } +}; - template void swap(T& lhs, T& rhs) +template +void swap(T& lhs, T& rhs) +{ + T temp = lhs; + lhs = rhs; + rhs = temp; +} + +template +I min_element(I begin, I end, const Pred& pred) +{ + I result = begin; + + for (I it = begin + 1; it != end; ++it) + if (pred(*it, *result)) + result = it; + + return result; +} + +template +void reverse(I begin, I end) +{ + while (end - begin > 1) + swap(*begin++, *--end); +} + +template +I unique(I begin, I end) +{ + // fast skip head + while (end - begin > 1 && *begin != *(begin + 1)) + begin++; + + if (begin == end) + return begin; + + // last written element + I write = begin++; + + // merge unique elements + while (begin != end) { - T temp = lhs; - lhs = rhs; - rhs = temp; + if (*begin != *write) + *++write = *begin++; + else + begin++; } - template I min_element(I begin, I end, const Pred& pred) + // past-the-end (write points to live element) + return write + 1; +} + +template +void insertion_sort(T* begin, T* end, const Pred& pred) +{ + if (begin == end) + return; + + for (T* it = begin + 1; it != end; ++it) { - I result = begin; + T val = *it; + T* hole = it; - for (I it = begin + 1; it != end; ++it) - if (pred(*it, *result)) - result = it; - - return result; - } - - template void reverse(I begin, I end) - { - while (end - begin > 1) swap(*begin++, *--end); - } - - template I unique(I begin, I end) - { - // fast skip head - while (end - begin > 1 && *begin != *(begin + 1)) begin++; - - if (begin == end) return begin; - - // last written element - I write = begin++; - - // merge unique elements - while (begin != end) + // move hole backwards + while (hole > begin && pred(val, *(hole - 1))) { - if (*begin != *write) - *++write = *begin++; - else - begin++; + *hole = *(hole - 1); + hole--; } - // past-the-end (write points to live element) - return write + 1; + // fill hole with element + *hole = val; + } +} + +template +I median3(I first, I middle, I last, const Pred& pred) +{ + if (pred(*middle, *first)) + swap(middle, first); + if (pred(*last, *middle)) + swap(last, middle); + if (pred(*middle, *first)) + swap(middle, first); + + return middle; +} + +template +void partition3(T* begin, T* end, T pivot, const Pred& pred, T** out_eqbeg, T** out_eqend) +{ + // invariant: array is split into 4 groups: = < ? > (each variable denotes the boundary between the groups) + T* eq = begin; + T* lt = begin; + T* gt = end; + + while (lt < gt) + { + if (pred(*lt, pivot)) + lt++; + else if (*lt == pivot) + swap(*eq++, *lt++); + else + swap(*lt, *--gt); } - template void insertion_sort(T* begin, T* end, const Pred& pred) + // we now have just 4 groups: = < >; move equal elements to the middle + T* eqbeg = gt; + + for (T* it = begin; it != eq; ++it) + swap(*it, *--eqbeg); + + *out_eqbeg = eqbeg; + *out_eqend = gt; +} + +template +void sort(I begin, I end, const Pred& pred) +{ + // sort large chunks + while (end - begin > 16) { - if (begin == end) - return; + // find median element + I middle = begin + (end - begin) / 2; + I median = median3(begin, middle, end - 1, pred); - for (T* it = begin + 1; it != end; ++it) + // partition in three chunks (< = >) + I eqbeg, eqend; + partition3(begin, end, *median, pred, &eqbeg, &eqend); + + // loop on larger half + if (eqbeg - begin > end - eqend) { - T val = *it; - T* hole = it; - - // move hole backwards - while (hole > begin && pred(val, *(hole - 1))) - { - *hole = *(hole - 1); - hole--; - } - - // fill hole with element - *hole = val; + sort(eqend, end, pred); + end = eqbeg; + } + else + { + sort(begin, eqbeg, pred); + begin = eqend; } } - template I median3(I first, I middle, I last, const Pred& pred) - { - if (pred(*middle, *first)) swap(middle, first); - if (pred(*last, *middle)) swap(last, middle); - if (pred(*middle, *first)) swap(middle, first); - - return middle; - } - - template void partition3(T* begin, T* end, T pivot, const Pred& pred, T** out_eqbeg, T** out_eqend) - { - // invariant: array is split into 4 groups: = < ? > (each variable denotes the boundary between the groups) - T* eq = begin; - T* lt = begin; - T* gt = end; - - while (lt < gt) - { - if (pred(*lt, pivot)) - lt++; - else if (*lt == pivot) - swap(*eq++, *lt++); - else - swap(*lt, *--gt); - } - - // we now have just 4 groups: = < >; move equal elements to the middle - T* eqbeg = gt; - - for (T* it = begin; it != eq; ++it) - swap(*it, *--eqbeg); - - *out_eqbeg = eqbeg; - *out_eqend = gt; - } - - template void sort(I begin, I end, const Pred& pred) - { - // sort large chunks - while (end - begin > 16) - { - // find median element - I middle = begin + (end - begin) / 2; - I median = median3(begin, middle, end - 1, pred); - - // partition in three chunks (< = >) - I eqbeg, eqend; - partition3(begin, end, *median, pred, &eqbeg, &eqend); - - // loop on larger half - if (eqbeg - begin > end - eqend) - { - sort(eqend, end, pred); - end = eqbeg; - } - else - { - sort(begin, eqbeg, pred); - begin = eqend; - } - } - - // insertion sort small chunk - insertion_sort(begin, end, pred); - } + // insertion sort small chunk + insertion_sort(begin, end, pred); +} PUGI__NS_END // Allocator used for AST and evaluation stacks PUGI__NS_BEGIN - static const size_t xpath_memory_page_size = - #ifdef PUGIXML_MEMORY_XPATH_PAGE_SIZE - PUGIXML_MEMORY_XPATH_PAGE_SIZE - #else - 4096 - #endif - ; +static const size_t xpath_memory_page_size = +#ifdef PUGIXML_MEMORY_XPATH_PAGE_SIZE + PUGIXML_MEMORY_XPATH_PAGE_SIZE +#else + 4096 +#endif + ; - static const uintptr_t xpath_memory_block_alignment = sizeof(double) > sizeof(void*) ? sizeof(double) : sizeof(void*); +static const uintptr_t xpath_memory_block_alignment = sizeof(double) > sizeof(void*) ? sizeof(double) : sizeof(void*); - struct xpath_memory_block - { - xpath_memory_block* next; - size_t capacity; +struct xpath_memory_block +{ + xpath_memory_block* next; + size_t capacity; - union - { - char data[xpath_memory_page_size]; - double alignment; - }; + union { + char data[xpath_memory_page_size]; + double alignment; }; +}; - struct xpath_allocator +struct xpath_allocator +{ + xpath_memory_block* _root; + size_t _root_size; + bool* _error; + + xpath_allocator(xpath_memory_block* root, bool* error = 0) + : _root(root) + , _root_size(0) + , _error(error) { - xpath_memory_block* _root; - size_t _root_size; - bool* _error; + } - xpath_allocator(xpath_memory_block* root, bool* error = 0): _root(root), _root_size(0), _error(error) + void* allocate(size_t size) + { + // round size up to block alignment boundary + size = (size + xpath_memory_block_alignment - 1) & ~(xpath_memory_block_alignment - 1); + + if (_root_size + size <= _root->capacity) { + void* buf = &_root->data[0] + _root_size; + _root_size += size; + return buf; + } + else + { + // make sure we have at least 1/4th of the page free after allocation to satisfy subsequent allocation requests + size_t block_capacity_base = sizeof(_root->data); + size_t block_capacity_req = size + block_capacity_base / 4; + size_t block_capacity = (block_capacity_base > block_capacity_req) ? block_capacity_base : block_capacity_req; + + size_t block_size = block_capacity + offsetof(xpath_memory_block, data); + + xpath_memory_block* block = static_cast(xml_memory::allocate(block_size)); + if (!block) + { + if (_error) + *_error = true; + return 0; + } + + block->next = _root; + block->capacity = block_capacity; + + _root = block; + _root_size = size; + + return block->data; + } + } + + void* reallocate(void* ptr, size_t old_size, size_t new_size) + { + // round size up to block alignment boundary + old_size = (old_size + xpath_memory_block_alignment - 1) & ~(xpath_memory_block_alignment - 1); + new_size = (new_size + xpath_memory_block_alignment - 1) & ~(xpath_memory_block_alignment - 1); + + // we can only reallocate the last object + assert(ptr == 0 || static_cast(ptr) + old_size == &_root->data[0] + _root_size); + + // try to reallocate the object inplace + if (ptr && _root_size - old_size + new_size <= _root->capacity) + { + _root_size = _root_size - old_size + new_size; + return ptr; } - void* allocate(size_t size) + // allocate a new block + void* result = allocate(new_size); + if (!result) + return 0; + + // we have a new block + if (ptr) { - // round size up to block alignment boundary - size = (size + xpath_memory_block_alignment - 1) & ~(xpath_memory_block_alignment - 1); + // copy old data (we only support growing) + assert(new_size >= old_size); + memcpy(result, ptr, old_size); - if (_root_size + size <= _root->capacity) + // free the previous page if it had no other objects + assert(_root->data == result); + assert(_root->next); + + if (_root->next->data == ptr) { - void* buf = &_root->data[0] + _root_size; - _root_size += size; - return buf; - } - else - { - // make sure we have at least 1/4th of the page free after allocation to satisfy subsequent allocation requests - size_t block_capacity_base = sizeof(_root->data); - size_t block_capacity_req = size + block_capacity_base / 4; - size_t block_capacity = (block_capacity_base > block_capacity_req) ? block_capacity_base : block_capacity_req; + // deallocate the whole page, unless it was the first one + xpath_memory_block* next = _root->next->next; - size_t block_size = block_capacity + offsetof(xpath_memory_block, data); - - xpath_memory_block* block = static_cast(xml_memory::allocate(block_size)); - if (!block) + if (next) { - if (_error) *_error = true; - return 0; - } - - block->next = _root; - block->capacity = block_capacity; - - _root = block; - _root_size = size; - - return block->data; - } - } - - void* reallocate(void* ptr, size_t old_size, size_t new_size) - { - // round size up to block alignment boundary - old_size = (old_size + xpath_memory_block_alignment - 1) & ~(xpath_memory_block_alignment - 1); - new_size = (new_size + xpath_memory_block_alignment - 1) & ~(xpath_memory_block_alignment - 1); - - // we can only reallocate the last object - assert(ptr == 0 || static_cast(ptr) + old_size == &_root->data[0] + _root_size); - - // try to reallocate the object inplace - if (ptr && _root_size - old_size + new_size <= _root->capacity) - { - _root_size = _root_size - old_size + new_size; - return ptr; - } - - // allocate a new block - void* result = allocate(new_size); - if (!result) return 0; - - // we have a new block - if (ptr) - { - // copy old data (we only support growing) - assert(new_size >= old_size); - memcpy(result, ptr, old_size); - - // free the previous page if it had no other objects - assert(_root->data == result); - assert(_root->next); - - if (_root->next->data == ptr) - { - // deallocate the whole page, unless it was the first one - xpath_memory_block* next = _root->next->next; - - if (next) - { - xml_memory::deallocate(_root->next); - _root->next = next; - } + xml_memory::deallocate(_root->next); + _root->next = next; } } - - return result; } - void revert(const xpath_allocator& state) - { - // free all new pages - xpath_memory_block* cur = _root; + return result; + } - while (cur != state._root) - { - xpath_memory_block* next = cur->next; - - xml_memory::deallocate(cur); - - cur = next; - } - - // restore state - _root = state._root; - _root_size = state._root_size; - } - - void release() - { - xpath_memory_block* cur = _root; - assert(cur); - - while (cur->next) - { - xpath_memory_block* next = cur->next; - - xml_memory::deallocate(cur); - - cur = next; - } - } - }; - - struct xpath_allocator_capture + void revert(const xpath_allocator& state) { - xpath_allocator_capture(xpath_allocator* alloc): _target(alloc), _state(*alloc) + // free all new pages + xpath_memory_block* cur = _root; + + while (cur != state._root) { + xpath_memory_block* next = cur->next; + + xml_memory::deallocate(cur); + + cur = next; } - ~xpath_allocator_capture() - { - _target->revert(_state); - } + // restore state + _root = state._root; + _root_size = state._root_size; + } - xpath_allocator* _target; - xpath_allocator _state; - }; - - struct xpath_stack + void release() { - xpath_allocator* result; - xpath_allocator* temp; - }; + xpath_memory_block* cur = _root; + assert(cur); - struct xpath_stack_data + while (cur->next) + { + xpath_memory_block* next = cur->next; + + xml_memory::deallocate(cur); + + cur = next; + } + } +}; + +struct xpath_allocator_capture +{ + xpath_allocator_capture(xpath_allocator* alloc) + : _target(alloc) + , _state(*alloc) { - xpath_memory_block blocks[2]; - xpath_allocator result; - xpath_allocator temp; - xpath_stack stack; - bool oom; + } - xpath_stack_data(): result(blocks + 0, &oom), temp(blocks + 1, &oom), oom(false) - { - blocks[0].next = blocks[1].next = 0; - blocks[0].capacity = blocks[1].capacity = sizeof(blocks[0].data); + ~xpath_allocator_capture() + { + _target->revert(_state); + } - stack.result = &result; - stack.temp = &temp; - } + xpath_allocator* _target; + xpath_allocator _state; +}; - ~xpath_stack_data() - { - result.release(); - temp.release(); - } - }; +struct xpath_stack +{ + xpath_allocator* result; + xpath_allocator* temp; +}; + +struct xpath_stack_data +{ + xpath_memory_block blocks[2]; + xpath_allocator result; + xpath_allocator temp; + xpath_stack stack; + bool oom; + + xpath_stack_data() + : result(blocks + 0, &oom) + , temp(blocks + 1, &oom) + , oom(false) + { + blocks[0].next = blocks[1].next = 0; + blocks[0].capacity = blocks[1].capacity = sizeof(blocks[0].data); + + stack.result = &result; + stack.temp = &temp; + } + + ~xpath_stack_data() + { + result.release(); + temp.release(); + } +}; PUGI__NS_END // String class PUGI__NS_BEGIN - class xpath_string +class xpath_string +{ + const char_t* _buffer; + bool _uses_heap; + size_t _length_heap; + + static char_t* duplicate_string(const char_t* string, size_t length, xpath_allocator* alloc) { - const char_t* _buffer; - bool _uses_heap; - size_t _length_heap; + char_t* result = static_cast(alloc->allocate((length + 1) * sizeof(char_t))); + if (!result) + return 0; - static char_t* duplicate_string(const char_t* string, size_t length, xpath_allocator* alloc) + memcpy(result, string, length * sizeof(char_t)); + result[length] = 0; + + return result; + } + + xpath_string(const char_t* buffer, bool uses_heap_, size_t length_heap) + : _buffer(buffer) + , _uses_heap(uses_heap_) + , _length_heap(length_heap) + { + } + + public: + static xpath_string from_const(const char_t* str) + { + return xpath_string(str, false, 0); + } + + static xpath_string from_heap_preallocated(const char_t* begin, const char_t* end) + { + assert(begin <= end && *end == 0); + + return xpath_string(begin, true, static_cast(end - begin)); + } + + static xpath_string from_heap(const char_t* begin, const char_t* end, xpath_allocator* alloc) + { + assert(begin <= end); + + if (begin == end) + return xpath_string(); + + size_t length = static_cast(end - begin); + const char_t* data = duplicate_string(begin, length, alloc); + + return data ? xpath_string(data, true, length) : xpath_string(); + } + + xpath_string() + : _buffer(PUGIXML_TEXT("")) + , _uses_heap(false) + , _length_heap(0) + { + } + + void append(const xpath_string& o, xpath_allocator* alloc) + { + // skip empty sources + if (!*o._buffer) + return; + + // fast append for constant empty target and constant source + if (!*_buffer && !_uses_heap && !o._uses_heap) { - char_t* result = static_cast(alloc->allocate((length + 1) * sizeof(char_t))); - if (!result) return 0; + _buffer = o._buffer; + } + else + { + // need to make heap copy + size_t target_length = length(); + size_t source_length = o.length(); + size_t result_length = target_length + source_length; - memcpy(result, string, length * sizeof(char_t)); - result[length] = 0; + // allocate new buffer + char_t* result = static_cast(alloc->reallocate(_uses_heap ? const_cast(_buffer) : 0, (target_length + 1) * sizeof(char_t), (result_length + 1) * sizeof(char_t))); + if (!result) + return; + + // append first string to the new buffer in case there was no reallocation + if (!_uses_heap) + memcpy(result, _buffer, target_length * sizeof(char_t)); + + // append second string to the new buffer + memcpy(result + target_length, o._buffer, source_length * sizeof(char_t)); + result[result_length] = 0; + + // finalize + _buffer = result; + _uses_heap = true; + _length_heap = result_length; + } + } + + const char_t* c_str() const + { + return _buffer; + } + + size_t length() const + { + return _uses_heap ? _length_heap : strlength(_buffer); + } + + char_t* data(xpath_allocator* alloc) + { + // make private heap copy + if (!_uses_heap) + { + size_t length_ = strlength(_buffer); + const char_t* data_ = duplicate_string(_buffer, length_, alloc); + + if (!data_) + return 0; + + _buffer = data_; + _uses_heap = true; + _length_heap = length_; + } + + return const_cast(_buffer); + } + + bool empty() const + { + return *_buffer == 0; + } + + bool operator==(const xpath_string& o) const + { + return strequal(_buffer, o._buffer); + } + + bool operator!=(const xpath_string& o) const + { + return !strequal(_buffer, o._buffer); + } + + bool uses_heap() const + { + return _uses_heap; + } +}; +PUGI__NS_END + +PUGI__NS_BEGIN +PUGI__FN bool starts_with(const char_t* string, const char_t* pattern) +{ + while (*pattern && *string == *pattern) + { + string++; + pattern++; + } + + return *pattern == 0; +} + +PUGI__FN const char_t* find_char(const char_t* s, char_t c) +{ +#ifdef PUGIXML_WCHAR_MODE + return wcschr(s, c); +#else + return strchr(s, c); +#endif +} + +PUGI__FN const char_t* find_substring(const char_t* s, const char_t* p) +{ +#ifdef PUGIXML_WCHAR_MODE + // MSVC6 wcsstr bug workaround (if s is empty it always returns 0) + return (*p == 0) ? s : wcsstr(s, p); +#else + return strstr(s, p); +#endif +} + +// Converts symbol to lower case, if it is an ASCII one +PUGI__FN char_t tolower_ascii(char_t ch) +{ + return static_cast(ch - 'A') < 26 ? static_cast(ch | ' ') : ch; +} + +PUGI__FN xpath_string string_value(const xpath_node& na, xpath_allocator* alloc) +{ + if (na.attribute()) + return xpath_string::from_const(na.attribute().value()); + else + { + xml_node n = na.node(); + + switch (n.type()) + { + case node_pcdata: + case node_cdata: + case node_comment: + case node_pi: + return xpath_string::from_const(n.value()); + + case node_document: + case node_element: + { + xpath_string result; + + // element nodes can have value if parse_embed_pcdata was used + if (n.value()[0]) + result.append(xpath_string::from_const(n.value()), alloc); + + xml_node cur = n.first_child(); + + while (cur && cur != n) + { + if (cur.type() == node_pcdata || cur.type() == node_cdata) + result.append(xpath_string::from_const(cur.value()), alloc); + + if (cur.first_child()) + cur = cur.first_child(); + else if (cur.next_sibling()) + cur = cur.next_sibling(); + else + { + while (!cur.next_sibling() && cur != n) + cur = cur.parent(); + + if (cur != n) + cur = cur.next_sibling(); + } + } return result; } - xpath_string(const char_t* buffer, bool uses_heap_, size_t length_heap): _buffer(buffer), _uses_heap(uses_heap_), _length_heap(length_heap) - { - } - - public: - static xpath_string from_const(const char_t* str) - { - return xpath_string(str, false, 0); - } - - static xpath_string from_heap_preallocated(const char_t* begin, const char_t* end) - { - assert(begin <= end && *end == 0); - - return xpath_string(begin, true, static_cast(end - begin)); - } - - static xpath_string from_heap(const char_t* begin, const char_t* end, xpath_allocator* alloc) - { - assert(begin <= end); - - if (begin == end) - return xpath_string(); - - size_t length = static_cast(end - begin); - const char_t* data = duplicate_string(begin, length, alloc); - - return data ? xpath_string(data, true, length) : xpath_string(); - } - - xpath_string(): _buffer(PUGIXML_TEXT("")), _uses_heap(false), _length_heap(0) - { - } - - void append(const xpath_string& o, xpath_allocator* alloc) - { - // skip empty sources - if (!*o._buffer) return; - - // fast append for constant empty target and constant source - if (!*_buffer && !_uses_heap && !o._uses_heap) - { - _buffer = o._buffer; - } - else - { - // need to make heap copy - size_t target_length = length(); - size_t source_length = o.length(); - size_t result_length = target_length + source_length; - - // allocate new buffer - char_t* result = static_cast(alloc->reallocate(_uses_heap ? const_cast(_buffer) : 0, (target_length + 1) * sizeof(char_t), (result_length + 1) * sizeof(char_t))); - if (!result) return; - - // append first string to the new buffer in case there was no reallocation - if (!_uses_heap) memcpy(result, _buffer, target_length * sizeof(char_t)); - - // append second string to the new buffer - memcpy(result + target_length, o._buffer, source_length * sizeof(char_t)); - result[result_length] = 0; - - // finalize - _buffer = result; - _uses_heap = true; - _length_heap = result_length; - } - } - - const char_t* c_str() const - { - return _buffer; - } - - size_t length() const - { - return _uses_heap ? _length_heap : strlength(_buffer); - } - - char_t* data(xpath_allocator* alloc) - { - // make private heap copy - if (!_uses_heap) - { - size_t length_ = strlength(_buffer); - const char_t* data_ = duplicate_string(_buffer, length_, alloc); - - if (!data_) return 0; - - _buffer = data_; - _uses_heap = true; - _length_heap = length_; - } - - return const_cast(_buffer); - } - - bool empty() const - { - return *_buffer == 0; - } - - bool operator==(const xpath_string& o) const - { - return strequal(_buffer, o._buffer); - } - - bool operator!=(const xpath_string& o) const - { - return !strequal(_buffer, o._buffer); - } - - bool uses_heap() const - { - return _uses_heap; - } - }; -PUGI__NS_END - -PUGI__NS_BEGIN - PUGI__FN bool starts_with(const char_t* string, const char_t* pattern) - { - while (*pattern && *string == *pattern) - { - string++; - pattern++; - } - - return *pattern == 0; - } - - PUGI__FN const char_t* find_char(const char_t* s, char_t c) - { - #ifdef PUGIXML_WCHAR_MODE - return wcschr(s, c); - #else - return strchr(s, c); - #endif - } - - PUGI__FN const char_t* find_substring(const char_t* s, const char_t* p) - { - #ifdef PUGIXML_WCHAR_MODE - // MSVC6 wcsstr bug workaround (if s is empty it always returns 0) - return (*p == 0) ? s : wcsstr(s, p); - #else - return strstr(s, p); - #endif - } - - // Converts symbol to lower case, if it is an ASCII one - PUGI__FN char_t tolower_ascii(char_t ch) - { - return static_cast(ch - 'A') < 26 ? static_cast(ch | ' ') : ch; - } - - PUGI__FN xpath_string string_value(const xpath_node& na, xpath_allocator* alloc) - { - if (na.attribute()) - return xpath_string::from_const(na.attribute().value()); - else - { - xml_node n = na.node(); - - switch (n.type()) - { - case node_pcdata: - case node_cdata: - case node_comment: - case node_pi: - return xpath_string::from_const(n.value()); - - case node_document: - case node_element: - { - xpath_string result; - - // element nodes can have value if parse_embed_pcdata was used - if (n.value()[0]) - result.append(xpath_string::from_const(n.value()), alloc); - - xml_node cur = n.first_child(); - - while (cur && cur != n) - { - if (cur.type() == node_pcdata || cur.type() == node_cdata) - result.append(xpath_string::from_const(cur.value()), alloc); - - if (cur.first_child()) - cur = cur.first_child(); - else if (cur.next_sibling()) - cur = cur.next_sibling(); - else - { - while (!cur.next_sibling() && cur != n) - cur = cur.parent(); - - if (cur != n) cur = cur.next_sibling(); - } - } - - return result; - } - - default: - return xpath_string(); - } - } - } - - PUGI__FN bool node_is_before_sibling(xml_node_struct* ln, xml_node_struct* rn) - { - assert(ln->parent == rn->parent); - - // there is no common ancestor (the shared parent is null), nodes are from different documents - if (!ln->parent) return ln < rn; - - // determine sibling order - xml_node_struct* ls = ln; - xml_node_struct* rs = rn; - - while (ls && rs) - { - if (ls == rn) return true; - if (rs == ln) return false; - - ls = ls->next_sibling; - rs = rs->next_sibling; - } - - // if rn sibling chain ended ln must be before rn - return !rs; - } - - PUGI__FN bool node_is_before(xml_node_struct* ln, xml_node_struct* rn) - { - // find common ancestor at the same depth, if any - xml_node_struct* lp = ln; - xml_node_struct* rp = rn; - - while (lp && rp && lp->parent != rp->parent) - { - lp = lp->parent; - rp = rp->parent; - } - - // parents are the same! - if (lp && rp) return node_is_before_sibling(lp, rp); - - // nodes are at different depths, need to normalize heights - bool left_higher = !lp; - - while (lp) - { - lp = lp->parent; - ln = ln->parent; - } - - while (rp) - { - rp = rp->parent; - rn = rn->parent; - } - - // one node is the ancestor of the other - if (ln == rn) return left_higher; - - // find common ancestor... again - while (ln->parent != rn->parent) - { - ln = ln->parent; - rn = rn->parent; - } - - return node_is_before_sibling(ln, rn); - } - - PUGI__FN bool node_is_ancestor(xml_node_struct* parent, xml_node_struct* node) - { - while (node && node != parent) node = node->parent; - - return parent && node == parent; - } - - PUGI__FN const void* document_buffer_order(const xpath_node& xnode) - { - xml_node_struct* node = xnode.node().internal_object(); - - if (node) - { - if ((get_document(node).header & xml_memory_page_contents_shared_mask) == 0) - { - if (node->name && (node->header & impl::xml_memory_page_name_allocated_or_shared_mask) == 0) return node->name; - if (node->value && (node->header & impl::xml_memory_page_value_allocated_or_shared_mask) == 0) return node->value; - } - - return 0; - } - - xml_attribute_struct* attr = xnode.attribute().internal_object(); - - if (attr) - { - if ((get_document(attr).header & xml_memory_page_contents_shared_mask) == 0) - { - if ((attr->header & impl::xml_memory_page_name_allocated_or_shared_mask) == 0) return attr->name; - if ((attr->header & impl::xml_memory_page_value_allocated_or_shared_mask) == 0) return attr->value; - } - - return 0; - } - - return 0; - } - - struct document_order_comparator - { - bool operator()(const xpath_node& lhs, const xpath_node& rhs) const - { - // optimized document order based check - const void* lo = document_buffer_order(lhs); - const void* ro = document_buffer_order(rhs); - - if (lo && ro) return lo < ro; - - // slow comparison - xml_node ln = lhs.node(), rn = rhs.node(); - - // compare attributes - if (lhs.attribute() && rhs.attribute()) - { - // shared parent - if (lhs.parent() == rhs.parent()) - { - // determine sibling order - for (xml_attribute a = lhs.attribute(); a; a = a.next_attribute()) - if (a == rhs.attribute()) - return true; - - return false; - } - - // compare attribute parents - ln = lhs.parent(); - rn = rhs.parent(); - } - else if (lhs.attribute()) - { - // attributes go after the parent element - if (lhs.parent() == rhs.node()) return false; - - ln = lhs.parent(); - } - else if (rhs.attribute()) - { - // attributes go after the parent element - if (rhs.parent() == lhs.node()) return true; - - rn = rhs.parent(); - } - - if (ln == rn) return false; - - if (!ln || !rn) return ln < rn; - - return node_is_before(ln.internal_object(), rn.internal_object()); - } - }; - - struct duplicate_comparator - { - bool operator()(const xpath_node& lhs, const xpath_node& rhs) const - { - if (lhs.attribute()) return rhs.attribute() ? lhs.attribute() < rhs.attribute() : true; - else return rhs.attribute() ? false : lhs.node() < rhs.node(); - } - }; - - PUGI__FN double gen_nan() - { - #if defined(__STDC_IEC_559__) || ((FLT_RADIX - 0 == 2) && (FLT_MAX_EXP - 0 == 128) && (FLT_MANT_DIG - 0 == 24)) - PUGI__STATIC_ASSERT(sizeof(float) == sizeof(uint32_t)); - typedef uint32_t UI; // BCC5 workaround - union { float f; UI i; } u; - u.i = 0x7fc00000; - return u.f; - #else - // fallback - const volatile double zero = 0.0; - return zero / zero; - #endif - } - - PUGI__FN bool is_nan(double value) - { - #if defined(PUGI__MSVC_CRT_VERSION) || defined(__BORLANDC__) - return !!_isnan(value); - #elif defined(fpclassify) && defined(FP_NAN) - return fpclassify(value) == FP_NAN; - #else - // fallback - const volatile double v = value; - return v != v; - #endif - } - - PUGI__FN const char_t* convert_number_to_string_special(double value) - { - #if defined(PUGI__MSVC_CRT_VERSION) || defined(__BORLANDC__) - if (_finite(value)) return (value == 0) ? PUGIXML_TEXT("0") : 0; - if (_isnan(value)) return PUGIXML_TEXT("NaN"); - return value > 0 ? PUGIXML_TEXT("Infinity") : PUGIXML_TEXT("-Infinity"); - #elif defined(fpclassify) && defined(FP_NAN) && defined(FP_INFINITE) && defined(FP_ZERO) - switch (fpclassify(value)) - { - case FP_NAN: - return PUGIXML_TEXT("NaN"); - - case FP_INFINITE: - return value > 0 ? PUGIXML_TEXT("Infinity") : PUGIXML_TEXT("-Infinity"); - - case FP_ZERO: - return PUGIXML_TEXT("0"); - default: - return 0; + return xpath_string(); + } + } +} + +PUGI__FN bool node_is_before_sibling(xml_node_struct* ln, xml_node_struct* rn) +{ + assert(ln->parent == rn->parent); + + // there is no common ancestor (the shared parent is null), nodes are from different documents + if (!ln->parent) + return ln < rn; + + // determine sibling order + xml_node_struct* ls = ln; + xml_node_struct* rs = rn; + + while (ls && rs) + { + if (ls == rn) + return true; + if (rs == ln) + return false; + + ls = ls->next_sibling; + rs = rs->next_sibling; + } + + // if rn sibling chain ended ln must be before rn + return !rs; +} + +PUGI__FN bool node_is_before(xml_node_struct* ln, xml_node_struct* rn) +{ + // find common ancestor at the same depth, if any + xml_node_struct* lp = ln; + xml_node_struct* rp = rn; + + while (lp && rp && lp->parent != rp->parent) + { + lp = lp->parent; + rp = rp->parent; + } + + // parents are the same! + if (lp && rp) + return node_is_before_sibling(lp, rp); + + // nodes are at different depths, need to normalize heights + bool left_higher = !lp; + + while (lp) + { + lp = lp->parent; + ln = ln->parent; + } + + while (rp) + { + rp = rp->parent; + rn = rn->parent; + } + + // one node is the ancestor of the other + if (ln == rn) + return left_higher; + + // find common ancestor... again + while (ln->parent != rn->parent) + { + ln = ln->parent; + rn = rn->parent; + } + + return node_is_before_sibling(ln, rn); +} + +PUGI__FN bool node_is_ancestor(xml_node_struct* parent, xml_node_struct* node) +{ + while (node && node != parent) + node = node->parent; + + return parent && node == parent; +} + +PUGI__FN const void* document_buffer_order(const xpath_node& xnode) +{ + xml_node_struct* node = xnode.node().internal_object(); + + if (node) + { + if ((get_document(node).header & xml_memory_page_contents_shared_mask) == 0) + { + if (node->name && (node->header & impl::xml_memory_page_name_allocated_or_shared_mask) == 0) + return node->name; + if (node->value && (node->header & impl::xml_memory_page_value_allocated_or_shared_mask) == 0) + return node->value; } - #else - // fallback - const volatile double v = value; - if (v == 0) return PUGIXML_TEXT("0"); - if (v != v) return PUGIXML_TEXT("NaN"); - if (v * 2 == v) return value > 0 ? PUGIXML_TEXT("Infinity") : PUGIXML_TEXT("-Infinity"); return 0; - #endif } - PUGI__FN bool convert_number_to_boolean(double value) + xml_attribute_struct* attr = xnode.attribute().internal_object(); + + if (attr) { - return (value != 0 && !is_nan(value)); + if ((get_document(attr).header & xml_memory_page_contents_shared_mask) == 0) + { + if ((attr->header & impl::xml_memory_page_name_allocated_or_shared_mask) == 0) + return attr->name; + if ((attr->header & impl::xml_memory_page_value_allocated_or_shared_mask) == 0) + return attr->value; + } + + return 0; } - PUGI__FN void truncate_zeros(char* begin, char* end) - { - while (begin != end && end[-1] == '0') end--; + return 0; +} - *end = 0; +struct document_order_comparator +{ + bool operator()(const xpath_node& lhs, const xpath_node& rhs) const + { + // optimized document order based check + const void* lo = document_buffer_order(lhs); + const void* ro = document_buffer_order(rhs); + + if (lo && ro) + return lo < ro; + + // slow comparison + xml_node ln = lhs.node(), rn = rhs.node(); + + // compare attributes + if (lhs.attribute() && rhs.attribute()) + { + // shared parent + if (lhs.parent() == rhs.parent()) + { + // determine sibling order + for (xml_attribute a = lhs.attribute(); a; a = a.next_attribute()) + if (a == rhs.attribute()) + return true; + + return false; + } + + // compare attribute parents + ln = lhs.parent(); + rn = rhs.parent(); + } + else if (lhs.attribute()) + { + // attributes go after the parent element + if (lhs.parent() == rhs.node()) + return false; + + ln = lhs.parent(); + } + else if (rhs.attribute()) + { + // attributes go after the parent element + if (rhs.parent() == lhs.node()) + return true; + + rn = rhs.parent(); + } + + if (ln == rn) + return false; + + if (!ln || !rn) + return ln < rn; + + return node_is_before(ln.internal_object(), rn.internal_object()); } +}; - // gets mantissa digits in the form of 0.xxxxx with 0. implied and the exponent -#if defined(PUGI__MSVC_CRT_VERSION) && PUGI__MSVC_CRT_VERSION >= 1400 && !defined(_WIN32_WCE) - PUGI__FN void convert_number_to_mantissa_exponent(double value, char (&buffer)[32], char** out_mantissa, int* out_exponent) +struct duplicate_comparator +{ + bool operator()(const xpath_node& lhs, const xpath_node& rhs) const { - // get base values - int sign, exponent; - _ecvt_s(buffer, sizeof(buffer), value, DBL_DIG + 1, &exponent, &sign); + if (lhs.attribute()) + return rhs.attribute() ? lhs.attribute() < rhs.attribute() : true; + else + return rhs.attribute() ? false : lhs.node() < rhs.node(); + } +}; - // truncate redundant zeros - truncate_zeros(buffer, buffer + strlen(buffer)); +PUGI__FN double gen_nan() +{ +#if defined(__STDC_IEC_559__) || ((FLT_RADIX - 0 == 2) && (FLT_MAX_EXP - 0 == 128) && (FLT_MANT_DIG - 0 == 24)) + PUGI__STATIC_ASSERT(sizeof(float) == sizeof(uint32_t)); + typedef uint32_t UI; // BCC5 workaround + union { + float f; + UI i; + } u; + u.i = 0x7fc00000; + return u.f; +#else + // fallback + const volatile double zero = 0.0; + return zero / zero; +#endif +} - // fill results - *out_mantissa = buffer; - *out_exponent = exponent; +PUGI__FN bool is_nan(double value) +{ +#if defined(PUGI__MSVC_CRT_VERSION) || defined(__BORLANDC__) + return !!_isnan(value); +#elif defined(fpclassify) && defined(FP_NAN) + return fpclassify(value) == FP_NAN; +#else + // fallback + const volatile double v = value; + return v != v; +#endif +} + +PUGI__FN const char_t* convert_number_to_string_special(double value) +{ +#if defined(PUGI__MSVC_CRT_VERSION) || defined(__BORLANDC__) + if (_finite(value)) + return (value == 0) ? PUGIXML_TEXT("0") : 0; + if (_isnan(value)) + return PUGIXML_TEXT("NaN"); + return value > 0 ? PUGIXML_TEXT("Infinity") : PUGIXML_TEXT("-Infinity"); +#elif defined(fpclassify) && defined(FP_NAN) && defined(FP_INFINITE) && defined(FP_ZERO) + switch (fpclassify(value)) + { + case FP_NAN: + return PUGIXML_TEXT("NaN"); + + case FP_INFINITE: + return value > 0 ? PUGIXML_TEXT("Infinity") : PUGIXML_TEXT("-Infinity"); + + case FP_ZERO: + return PUGIXML_TEXT("0"); + + default: + return 0; } #else - PUGI__FN void convert_number_to_mantissa_exponent(double value, char (&buffer)[32], char** out_mantissa, int* out_exponent) - { - // get a scientific notation value with IEEE DBL_DIG decimals - PUGI__SNPRINTF(buffer, "%.*e", DBL_DIG, value); + // fallback + const volatile double v = value; - // get the exponent (possibly negative) - char* exponent_string = strchr(buffer, 'e'); - assert(exponent_string); + if (v == 0) + return PUGIXML_TEXT("0"); + if (v != v) + return PUGIXML_TEXT("NaN"); + if (v * 2 == v) + return value > 0 ? PUGIXML_TEXT("Infinity") : PUGIXML_TEXT("-Infinity"); + return 0; +#endif +} - int exponent = atoi(exponent_string + 1); +PUGI__FN bool convert_number_to_boolean(double value) +{ + return (value != 0 && !is_nan(value)); +} - // extract mantissa string: skip sign - char* mantissa = buffer[0] == '-' ? buffer + 1 : buffer; - assert(mantissa[0] != '0' && mantissa[1] == '.'); +PUGI__FN void truncate_zeros(char* begin, char* end) +{ + while (begin != end && end[-1] == '0') + end--; - // divide mantissa by 10 to eliminate integer part - mantissa[1] = mantissa[0]; - mantissa++; - exponent++; + *end = 0; +} - // remove extra mantissa digits and zero-terminate mantissa - truncate_zeros(mantissa, exponent_string); +// gets mantissa digits in the form of 0.xxxxx with 0. implied and the exponent +#if defined(PUGI__MSVC_CRT_VERSION) && PUGI__MSVC_CRT_VERSION >= 1400 && !defined(_WIN32_WCE) +PUGI__FN void convert_number_to_mantissa_exponent(double value, char (&buffer)[32], char** out_mantissa, int* out_exponent) +{ + // get base values + int sign, exponent; + _ecvt_s(buffer, sizeof(buffer), value, DBL_DIG + 1, &exponent, &sign); - // fill results - *out_mantissa = mantissa; - *out_exponent = exponent; - } + // truncate redundant zeros + truncate_zeros(buffer, buffer + strlen(buffer)); + + // fill results + *out_mantissa = buffer; + *out_exponent = exponent; +} +#else +PUGI__FN void convert_number_to_mantissa_exponent(double value, char (&buffer)[32], char** out_mantissa, int* out_exponent) +{ + // get a scientific notation value with IEEE DBL_DIG decimals + PUGI__SNPRINTF(buffer, "%.*e", DBL_DIG, value); + + // get the exponent (possibly negative) + char* exponent_string = strchr(buffer, 'e'); + assert(exponent_string); + + int exponent = atoi(exponent_string + 1); + + // extract mantissa string: skip sign + char* mantissa = buffer[0] == '-' ? buffer + 1 : buffer; + assert(mantissa[0] != '0' && mantissa[1] == '.'); + + // divide mantissa by 10 to eliminate integer part + mantissa[1] = mantissa[0]; + mantissa++; + exponent++; + + // remove extra mantissa digits and zero-terminate mantissa + truncate_zeros(mantissa, exponent_string); + + // fill results + *out_mantissa = mantissa; + *out_exponent = exponent; +} #endif - PUGI__FN xpath_string convert_number_to_string(double value, xpath_allocator* alloc) +PUGI__FN xpath_string convert_number_to_string(double value, xpath_allocator* alloc) +{ + // try special number conversion + const char_t* special = convert_number_to_string_special(value); + if (special) + return xpath_string::from_const(special); + + // get mantissa + exponent form + char mantissa_buffer[32]; + + char* mantissa; + int exponent; + convert_number_to_mantissa_exponent(value, mantissa_buffer, &mantissa, &exponent); + + // allocate a buffer of suitable length for the number + size_t result_size = strlen(mantissa_buffer) + (exponent > 0 ? exponent : -exponent) + 4; + char_t* result = static_cast(alloc->allocate(sizeof(char_t) * result_size)); + if (!result) + return xpath_string(); + + // make the number! + char_t* s = result; + + // sign + if (value < 0) + *s++ = '-'; + + // integer part + if (exponent <= 0) { - // try special number conversion - const char_t* special = convert_number_to_string_special(value); - if (special) return xpath_string::from_const(special); + *s++ = '0'; + } + else + { + while (exponent > 0) + { + assert(*mantissa == 0 || static_cast(*mantissa - '0') <= 9); + *s++ = *mantissa ? *mantissa++ : '0'; + exponent--; + } + } - // get mantissa + exponent form - char mantissa_buffer[32]; + // fractional part + if (*mantissa) + { + // decimal point + *s++ = '.'; - char* mantissa; - int exponent; - convert_number_to_mantissa_exponent(value, mantissa_buffer, &mantissa, &exponent); - - // allocate a buffer of suitable length for the number - size_t result_size = strlen(mantissa_buffer) + (exponent > 0 ? exponent : -exponent) + 4; - char_t* result = static_cast(alloc->allocate(sizeof(char_t) * result_size)); - if (!result) return xpath_string(); - - // make the number! - char_t* s = result; - - // sign - if (value < 0) *s++ = '-'; - - // integer part - if (exponent <= 0) + // extra zeroes from negative exponent + while (exponent < 0) { *s++ = '0'; + exponent++; + } + + // extra mantissa digits + while (*mantissa) + { + assert(static_cast(*mantissa - '0') <= 9); + *s++ = *mantissa++; + } + } + + // zero-terminate + assert(s < result + result_size); + *s = 0; + + return xpath_string::from_heap_preallocated(result, s); +} + +PUGI__FN bool check_string_to_number_format(const char_t* string) +{ + // parse leading whitespace + while (PUGI__IS_CHARTYPE(*string, ct_space)) + ++string; + + // parse sign + if (*string == '-') + ++string; + + if (!*string) + return false; + + // if there is no integer part, there should be a decimal part with at least one digit + if (!PUGI__IS_CHARTYPEX(string[0], ctx_digit) && (string[0] != '.' || !PUGI__IS_CHARTYPEX(string[1], ctx_digit))) + return false; + + // parse integer part + while (PUGI__IS_CHARTYPEX(*string, ctx_digit)) + ++string; + + // parse decimal part + if (*string == '.') + { + ++string; + + while (PUGI__IS_CHARTYPEX(*string, ctx_digit)) + ++string; + } + + // parse trailing whitespace + while (PUGI__IS_CHARTYPE(*string, ct_space)) + ++string; + + return *string == 0; +} + +PUGI__FN double convert_string_to_number(const char_t* string) +{ + // check string format + if (!check_string_to_number_format(string)) + return gen_nan(); + + // parse string +#ifdef PUGIXML_WCHAR_MODE + return wcstod(string, 0); +#else + return strtod(string, 0); +#endif +} + +PUGI__FN bool convert_string_to_number_scratch(char_t (&buffer)[32], const char_t* begin, const char_t* end, double* out_result) +{ + size_t length = static_cast(end - begin); + char_t* scratch = buffer; + + if (length >= sizeof(buffer) / sizeof(buffer[0])) + { + // need to make dummy on-heap copy + scratch = static_cast(xml_memory::allocate((length + 1) * sizeof(char_t))); + if (!scratch) + return false; + } + + // copy string to zero-terminated buffer and perform conversion + memcpy(scratch, begin, length * sizeof(char_t)); + scratch[length] = 0; + + *out_result = convert_string_to_number(scratch); + + // free dummy buffer + if (scratch != buffer) + xml_memory::deallocate(scratch); + + return true; +} + +PUGI__FN double round_nearest(double value) +{ + return floor(value + 0.5); +} + +PUGI__FN double round_nearest_nzero(double value) +{ + // same as round_nearest, but returns -0 for [-0.5, -0] + // ceil is used to differentiate between +0 and -0 (we return -0 for [-0.5, -0] and +0 for +0) + return (value >= -0.5 && value <= 0) ? ceil(value) : floor(value + 0.5); +} + +PUGI__FN const char_t* qualified_name(const xpath_node& node) +{ + return node.attribute() ? node.attribute().name() : node.node().name(); +} + +PUGI__FN const char_t* local_name(const xpath_node& node) +{ + const char_t* name = qualified_name(node); + const char_t* p = find_char(name, ':'); + + return p ? p + 1 : name; +} + +struct namespace_uri_predicate +{ + const char_t* prefix; + size_t prefix_length; + + namespace_uri_predicate(const char_t* name) + { + const char_t* pos = find_char(name, ':'); + + prefix = pos ? name : 0; + prefix_length = pos ? static_cast(pos - name) : 0; + } + + bool operator()(xml_attribute a) const + { + const char_t* name = a.name(); + + if (!starts_with(name, PUGIXML_TEXT("xmlns"))) + return false; + + return prefix ? name[5] == ':' && strequalrange(name + 6, prefix, prefix_length) : name[5] == 0; + } +}; + +PUGI__FN const char_t* namespace_uri(xml_node node) +{ + namespace_uri_predicate pred = node.name(); + + xml_node p = node; + + while (p) + { + xml_attribute a = p.find_attribute(pred); + + if (a) + return a.value(); + + p = p.parent(); + } + + return PUGIXML_TEXT(""); +} + +PUGI__FN const char_t* namespace_uri(xml_attribute attr, xml_node parent) +{ + namespace_uri_predicate pred = attr.name(); + + // Default namespace does not apply to attributes + if (!pred.prefix) + return PUGIXML_TEXT(""); + + xml_node p = parent; + + while (p) + { + xml_attribute a = p.find_attribute(pred); + + if (a) + return a.value(); + + p = p.parent(); + } + + return PUGIXML_TEXT(""); +} + +PUGI__FN const char_t* namespace_uri(const xpath_node& node) +{ + return node.attribute() ? namespace_uri(node.attribute(), node.parent()) : namespace_uri(node.node()); +} + +PUGI__FN char_t* normalize_space(char_t* buffer) +{ + char_t* write = buffer; + + for (char_t* it = buffer; *it;) + { + char_t ch = *it++; + + if (PUGI__IS_CHARTYPE(ch, ct_space)) + { + // replace whitespace sequence with single space + while (PUGI__IS_CHARTYPE(*it, ct_space)) + it++; + + // avoid leading spaces + if (write != buffer) + *write++ = ' '; + } + else + *write++ = ch; + } + + // remove trailing space + if (write != buffer && PUGI__IS_CHARTYPE(write[-1], ct_space)) + write--; + + // zero-terminate + *write = 0; + + return write; +} + +PUGI__FN char_t* translate(char_t* buffer, const char_t* from, const char_t* to, size_t to_length) +{ + char_t* write = buffer; + + while (*buffer) + { + PUGI__DMC_VOLATILE char_t ch = *buffer++; + + const char_t* pos = find_char(from, ch); + + if (!pos) + *write++ = ch; // do not process + else if (static_cast(pos - from) < to_length) + *write++ = to[pos - from]; // replace + } + + // zero-terminate + *write = 0; + + return write; +} + +PUGI__FN unsigned char* translate_table_generate(xpath_allocator* alloc, const char_t* from, const char_t* to) +{ + unsigned char table[128] = {0}; + + while (*from) + { + unsigned int fc = static_cast(*from); + unsigned int tc = static_cast(*to); + + if (fc >= 128 || tc >= 128) + return 0; + + // code=128 means "skip character" + if (!table[fc]) + table[fc] = static_cast(tc ? tc : 128); + + from++; + if (tc) + to++; + } + + for (int i = 0; i < 128; ++i) + if (!table[i]) + table[i] = static_cast(i); + + void* result = alloc->allocate(sizeof(table)); + if (!result) + return 0; + + memcpy(result, table, sizeof(table)); + + return static_cast(result); +} + +PUGI__FN char_t* translate_table(char_t* buffer, const unsigned char* table) +{ + char_t* write = buffer; + + while (*buffer) + { + char_t ch = *buffer++; + unsigned int index = static_cast(ch); + + if (index < 128) + { + unsigned char code = table[index]; + + // code=128 means "skip character" (table size is 128 so 128 can be a special value) + // this code skips these characters without extra branches + *write = static_cast(code); + write += 1 - (code >> 7); } else { - while (exponent > 0) - { - assert(*mantissa == 0 || static_cast(*mantissa - '0') <= 9); - *s++ = *mantissa ? *mantissa++ : '0'; - exponent--; - } - } - - // fractional part - if (*mantissa) - { - // decimal point - *s++ = '.'; - - // extra zeroes from negative exponent - while (exponent < 0) - { - *s++ = '0'; - exponent++; - } - - // extra mantissa digits - while (*mantissa) - { - assert(static_cast(*mantissa - '0') <= 9); - *s++ = *mantissa++; - } - } - - // zero-terminate - assert(s < result + result_size); - *s = 0; - - return xpath_string::from_heap_preallocated(result, s); - } - - PUGI__FN bool check_string_to_number_format(const char_t* string) - { - // parse leading whitespace - while (PUGI__IS_CHARTYPE(*string, ct_space)) ++string; - - // parse sign - if (*string == '-') ++string; - - if (!*string) return false; - - // if there is no integer part, there should be a decimal part with at least one digit - if (!PUGI__IS_CHARTYPEX(string[0], ctx_digit) && (string[0] != '.' || !PUGI__IS_CHARTYPEX(string[1], ctx_digit))) return false; - - // parse integer part - while (PUGI__IS_CHARTYPEX(*string, ctx_digit)) ++string; - - // parse decimal part - if (*string == '.') - { - ++string; - - while (PUGI__IS_CHARTYPEX(*string, ctx_digit)) ++string; - } - - // parse trailing whitespace - while (PUGI__IS_CHARTYPE(*string, ct_space)) ++string; - - return *string == 0; - } - - PUGI__FN double convert_string_to_number(const char_t* string) - { - // check string format - if (!check_string_to_number_format(string)) return gen_nan(); - - // parse string - #ifdef PUGIXML_WCHAR_MODE - return wcstod(string, 0); - #else - return strtod(string, 0); - #endif - } - - PUGI__FN bool convert_string_to_number_scratch(char_t (&buffer)[32], const char_t* begin, const char_t* end, double* out_result) - { - size_t length = static_cast(end - begin); - char_t* scratch = buffer; - - if (length >= sizeof(buffer) / sizeof(buffer[0])) - { - // need to make dummy on-heap copy - scratch = static_cast(xml_memory::allocate((length + 1) * sizeof(char_t))); - if (!scratch) return false; - } - - // copy string to zero-terminated buffer and perform conversion - memcpy(scratch, begin, length * sizeof(char_t)); - scratch[length] = 0; - - *out_result = convert_string_to_number(scratch); - - // free dummy buffer - if (scratch != buffer) xml_memory::deallocate(scratch); - - return true; - } - - PUGI__FN double round_nearest(double value) - { - return floor(value + 0.5); - } - - PUGI__FN double round_nearest_nzero(double value) - { - // same as round_nearest, but returns -0 for [-0.5, -0] - // ceil is used to differentiate between +0 and -0 (we return -0 for [-0.5, -0] and +0 for +0) - return (value >= -0.5 && value <= 0) ? ceil(value) : floor(value + 0.5); - } - - PUGI__FN const char_t* qualified_name(const xpath_node& node) - { - return node.attribute() ? node.attribute().name() : node.node().name(); - } - - PUGI__FN const char_t* local_name(const xpath_node& node) - { - const char_t* name = qualified_name(node); - const char_t* p = find_char(name, ':'); - - return p ? p + 1 : name; - } - - struct namespace_uri_predicate - { - const char_t* prefix; - size_t prefix_length; - - namespace_uri_predicate(const char_t* name) - { - const char_t* pos = find_char(name, ':'); - - prefix = pos ? name : 0; - prefix_length = pos ? static_cast(pos - name) : 0; - } - - bool operator()(xml_attribute a) const - { - const char_t* name = a.name(); - - if (!starts_with(name, PUGIXML_TEXT("xmlns"))) return false; - - return prefix ? name[5] == ':' && strequalrange(name + 6, prefix, prefix_length) : name[5] == 0; - } - }; - - PUGI__FN const char_t* namespace_uri(xml_node node) - { - namespace_uri_predicate pred = node.name(); - - xml_node p = node; - - while (p) - { - xml_attribute a = p.find_attribute(pred); - - if (a) return a.value(); - - p = p.parent(); - } - - return PUGIXML_TEXT(""); - } - - PUGI__FN const char_t* namespace_uri(xml_attribute attr, xml_node parent) - { - namespace_uri_predicate pred = attr.name(); - - // Default namespace does not apply to attributes - if (!pred.prefix) return PUGIXML_TEXT(""); - - xml_node p = parent; - - while (p) - { - xml_attribute a = p.find_attribute(pred); - - if (a) return a.value(); - - p = p.parent(); - } - - return PUGIXML_TEXT(""); - } - - PUGI__FN const char_t* namespace_uri(const xpath_node& node) - { - return node.attribute() ? namespace_uri(node.attribute(), node.parent()) : namespace_uri(node.node()); - } - - PUGI__FN char_t* normalize_space(char_t* buffer) - { - char_t* write = buffer; - - for (char_t* it = buffer; *it; ) - { - char_t ch = *it++; - - if (PUGI__IS_CHARTYPE(ch, ct_space)) - { - // replace whitespace sequence with single space - while (PUGI__IS_CHARTYPE(*it, ct_space)) it++; - - // avoid leading spaces - if (write != buffer) *write++ = ' '; - } - else *write++ = ch; - } - - // remove trailing space - if (write != buffer && PUGI__IS_CHARTYPE(write[-1], ct_space)) write--; - - // zero-terminate - *write = 0; - - return write; - } - - PUGI__FN char_t* translate(char_t* buffer, const char_t* from, const char_t* to, size_t to_length) - { - char_t* write = buffer; - - while (*buffer) - { - PUGI__DMC_VOLATILE char_t ch = *buffer++; - - const char_t* pos = find_char(from, ch); - - if (!pos) - *write++ = ch; // do not process - else if (static_cast(pos - from) < to_length) - *write++ = to[pos - from]; // replace - } - - // zero-terminate - *write = 0; - - return write; - } - - PUGI__FN unsigned char* translate_table_generate(xpath_allocator* alloc, const char_t* from, const char_t* to) - { - unsigned char table[128] = {0}; - - while (*from) - { - unsigned int fc = static_cast(*from); - unsigned int tc = static_cast(*to); - - if (fc >= 128 || tc >= 128) - return 0; - - // code=128 means "skip character" - if (!table[fc]) - table[fc] = static_cast(tc ? tc : 128); - - from++; - if (tc) to++; - } - - for (int i = 0; i < 128; ++i) - if (!table[i]) - table[i] = static_cast(i); - - void* result = alloc->allocate(sizeof(table)); - if (!result) return 0; - - memcpy(result, table, sizeof(table)); - - return static_cast(result); - } - - PUGI__FN char_t* translate_table(char_t* buffer, const unsigned char* table) - { - char_t* write = buffer; - - while (*buffer) - { - char_t ch = *buffer++; - unsigned int index = static_cast(ch); - - if (index < 128) - { - unsigned char code = table[index]; - - // code=128 means "skip character" (table size is 128 so 128 can be a special value) - // this code skips these characters without extra branches - *write = static_cast(code); - write += 1 - (code >> 7); - } - else - { - *write++ = ch; - } - } - - // zero-terminate - *write = 0; - - return write; - } - - inline bool is_xpath_attribute(const char_t* name) - { - return !(starts_with(name, PUGIXML_TEXT("xmlns")) && (name[5] == 0 || name[5] == ':')); - } - - struct xpath_variable_boolean: xpath_variable - { - xpath_variable_boolean(): xpath_variable(xpath_type_boolean), value(false) - { - } - - bool value; - char_t name[1]; - }; - - struct xpath_variable_number: xpath_variable - { - xpath_variable_number(): xpath_variable(xpath_type_number), value(0) - { - } - - double value; - char_t name[1]; - }; - - struct xpath_variable_string: xpath_variable - { - xpath_variable_string(): xpath_variable(xpath_type_string), value(0) - { - } - - ~xpath_variable_string() - { - if (value) xml_memory::deallocate(value); - } - - char_t* value; - char_t name[1]; - }; - - struct xpath_variable_node_set: xpath_variable - { - xpath_variable_node_set(): xpath_variable(xpath_type_node_set) - { - } - - xpath_node_set value; - char_t name[1]; - }; - - static const xpath_node_set dummy_node_set; - - PUGI__FN PUGI__UNSIGNED_OVERFLOW unsigned int hash_string(const char_t* str) - { - // Jenkins one-at-a-time hash (http://en.wikipedia.org/wiki/Jenkins_hash_function#one-at-a-time) - unsigned int result = 0; - - while (*str) - { - result += static_cast(*str++); - result += result << 10; - result ^= result >> 6; - } - - result += result << 3; - result ^= result >> 11; - result += result << 15; - - return result; - } - - template PUGI__FN T* new_xpath_variable(const char_t* name) - { - size_t length = strlength(name); - if (length == 0) return 0; // empty variable names are invalid - - // $$ we can't use offsetof(T, name) because T is non-POD, so we just allocate additional length characters - void* memory = xml_memory::allocate(sizeof(T) + length * sizeof(char_t)); - if (!memory) return 0; - - T* result = new (memory) T(); - - memcpy(result->name, name, (length + 1) * sizeof(char_t)); - - return result; - } - - PUGI__FN xpath_variable* new_xpath_variable(xpath_value_type type, const char_t* name) - { - switch (type) - { - case xpath_type_node_set: - return new_xpath_variable(name); - - case xpath_type_number: - return new_xpath_variable(name); - - case xpath_type_string: - return new_xpath_variable(name); - - case xpath_type_boolean: - return new_xpath_variable(name); - - default: - return 0; + *write++ = ch; } } - template PUGI__FN void delete_xpath_variable(T* var) + // zero-terminate + *write = 0; + + return write; +} + +inline bool is_xpath_attribute(const char_t* name) +{ + return !(starts_with(name, PUGIXML_TEXT("xmlns")) && (name[5] == 0 || name[5] == ':')); +} + +struct xpath_variable_boolean : xpath_variable +{ + xpath_variable_boolean() + : xpath_variable(xpath_type_boolean) + , value(false) { - var->~T(); - xml_memory::deallocate(var); } - PUGI__FN void delete_xpath_variable(xpath_value_type type, xpath_variable* var) + bool value; + char_t name[1]; +}; + +struct xpath_variable_number : xpath_variable +{ + xpath_variable_number() + : xpath_variable(xpath_type_number) + , value(0) { - switch (type) - { - case xpath_type_node_set: - delete_xpath_variable(static_cast(var)); - break; - - case xpath_type_number: - delete_xpath_variable(static_cast(var)); - break; - - case xpath_type_string: - delete_xpath_variable(static_cast(var)); - break; - - case xpath_type_boolean: - delete_xpath_variable(static_cast(var)); - break; - - default: - assert(false && "Invalid variable type"); // unreachable - } } - PUGI__FN bool copy_xpath_variable(xpath_variable* lhs, const xpath_variable* rhs) + double value; + char_t name[1]; +}; + +struct xpath_variable_string : xpath_variable +{ + xpath_variable_string() + : xpath_variable(xpath_type_string) + , value(0) { - switch (rhs->type()) - { - case xpath_type_node_set: - return lhs->set(static_cast(rhs)->value); + } - case xpath_type_number: - return lhs->set(static_cast(rhs)->value); + ~xpath_variable_string() + { + if (value) + xml_memory::deallocate(value); + } - case xpath_type_string: - return lhs->set(static_cast(rhs)->value); + char_t* value; + char_t name[1]; +}; - case xpath_type_boolean: - return lhs->set(static_cast(rhs)->value); +struct xpath_variable_node_set : xpath_variable +{ + xpath_variable_node_set() + : xpath_variable(xpath_type_node_set) + { + } - default: - assert(false && "Invalid variable type"); // unreachable + xpath_node_set value; + char_t name[1]; +}; + +static const xpath_node_set dummy_node_set; + +PUGI__FN PUGI__UNSIGNED_OVERFLOW unsigned int hash_string(const char_t* str) +{ + // Jenkins one-at-a-time hash (http://en.wikipedia.org/wiki/Jenkins_hash_function#one-at-a-time) + unsigned int result = 0; + + while (*str) + { + result += static_cast(*str++); + result += result << 10; + result ^= result >> 6; + } + + result += result << 3; + result ^= result >> 11; + result += result << 15; + + return result; +} + +template +PUGI__FN T* new_xpath_variable(const char_t* name) +{ + size_t length = strlength(name); + if (length == 0) + return 0; // empty variable names are invalid + + // $$ we can't use offsetof(T, name) because T is non-POD, so we just allocate additional length characters + void* memory = xml_memory::allocate(sizeof(T) + length * sizeof(char_t)); + if (!memory) + return 0; + + T* result = new (memory) T(); + + memcpy(result->name, name, (length + 1) * sizeof(char_t)); + + return result; +} + +PUGI__FN xpath_variable* new_xpath_variable(xpath_value_type type, const char_t* name) +{ + switch (type) + { + case xpath_type_node_set: + return new_xpath_variable(name); + + case xpath_type_number: + return new_xpath_variable(name); + + case xpath_type_string: + return new_xpath_variable(name); + + case xpath_type_boolean: + return new_xpath_variable(name); + + default: + return 0; + } +} + +template +PUGI__FN void delete_xpath_variable(T* var) +{ + var->~T(); + xml_memory::deallocate(var); +} + +PUGI__FN void delete_xpath_variable(xpath_value_type type, xpath_variable* var) +{ + switch (type) + { + case xpath_type_node_set: + delete_xpath_variable(static_cast(var)); + break; + + case xpath_type_number: + delete_xpath_variable(static_cast(var)); + break; + + case xpath_type_string: + delete_xpath_variable(static_cast(var)); + break; + + case xpath_type_boolean: + delete_xpath_variable(static_cast(var)); + break; + + default: + assert(false && "Invalid variable type"); // unreachable + } +} + +PUGI__FN bool copy_xpath_variable(xpath_variable* lhs, const xpath_variable* rhs) +{ + switch (rhs->type()) + { + case xpath_type_node_set: + return lhs->set(static_cast(rhs)->value); + + case xpath_type_number: + return lhs->set(static_cast(rhs)->value); + + case xpath_type_string: + return lhs->set(static_cast(rhs)->value); + + case xpath_type_boolean: + return lhs->set(static_cast(rhs)->value); + + default: + assert(false && "Invalid variable type"); // unreachable + return false; + } +} + +PUGI__FN bool get_variable_scratch(char_t (&buffer)[32], xpath_variable_set* set, const char_t* begin, const char_t* end, xpath_variable** out_result) +{ + size_t length = static_cast(end - begin); + char_t* scratch = buffer; + + if (length >= sizeof(buffer) / sizeof(buffer[0])) + { + // need to make dummy on-heap copy + scratch = static_cast(xml_memory::allocate((length + 1) * sizeof(char_t))); + if (!scratch) return false; - } } - PUGI__FN bool get_variable_scratch(char_t (&buffer)[32], xpath_variable_set* set, const char_t* begin, const char_t* end, xpath_variable** out_result) - { - size_t length = static_cast(end - begin); - char_t* scratch = buffer; + // copy string to zero-terminated buffer and perform lookup + memcpy(scratch, begin, length * sizeof(char_t)); + scratch[length] = 0; - if (length >= sizeof(buffer) / sizeof(buffer[0])) - { - // need to make dummy on-heap copy - scratch = static_cast(xml_memory::allocate((length + 1) * sizeof(char_t))); - if (!scratch) return false; - } + *out_result = set->get(scratch); - // copy string to zero-terminated buffer and perform lookup - memcpy(scratch, begin, length * sizeof(char_t)); - scratch[length] = 0; + // free dummy buffer + if (scratch != buffer) + xml_memory::deallocate(scratch); - *out_result = set->get(scratch); - - // free dummy buffer - if (scratch != buffer) xml_memory::deallocate(scratch); - - return true; - } + return true; +} PUGI__NS_END // Internal node set class PUGI__NS_BEGIN - PUGI__FN xpath_node_set::type_t xpath_get_order(const xpath_node* begin, const xpath_node* end) +PUGI__FN xpath_node_set::type_t xpath_get_order(const xpath_node* begin, const xpath_node* end) +{ + if (end - begin < 2) + return xpath_node_set::type_sorted; + + document_order_comparator cmp; + + bool first = cmp(begin[0], begin[1]); + + for (const xpath_node* it = begin + 1; it + 1 < end; ++it) + if (cmp(it[0], it[1]) != first) + return xpath_node_set::type_unsorted; + + return first ? xpath_node_set::type_sorted : xpath_node_set::type_sorted_reverse; +} + +PUGI__FN xpath_node_set::type_t xpath_sort(xpath_node* begin, xpath_node* end, xpath_node_set::type_t type, bool rev) +{ + xpath_node_set::type_t order = rev ? xpath_node_set::type_sorted_reverse : xpath_node_set::type_sorted; + + if (type == xpath_node_set::type_unsorted) { - if (end - begin < 2) - return xpath_node_set::type_sorted; + xpath_node_set::type_t sorted = xpath_get_order(begin, end); - document_order_comparator cmp; + if (sorted == xpath_node_set::type_unsorted) + { + sort(begin, end, document_order_comparator()); - bool first = cmp(begin[0], begin[1]); - - for (const xpath_node* it = begin + 1; it + 1 < end; ++it) - if (cmp(it[0], it[1]) != first) - return xpath_node_set::type_unsorted; - - return first ? xpath_node_set::type_sorted : xpath_node_set::type_sorted_reverse; + type = xpath_node_set::type_sorted; + } + else + type = sorted; } - PUGI__FN xpath_node_set::type_t xpath_sort(xpath_node* begin, xpath_node* end, xpath_node_set::type_t type, bool rev) + if (type != order) + reverse(begin, end); + + return order; +} + +PUGI__FN xpath_node xpath_first(const xpath_node* begin, const xpath_node* end, xpath_node_set::type_t type) +{ + if (begin == end) + return xpath_node(); + + switch (type) { - xpath_node_set::type_t order = rev ? xpath_node_set::type_sorted_reverse : xpath_node_set::type_sorted; + case xpath_node_set::type_sorted: + return *begin; - if (type == xpath_node_set::type_unsorted) - { - xpath_node_set::type_t sorted = xpath_get_order(begin, end); + case xpath_node_set::type_sorted_reverse: + return *(end - 1); - if (sorted == xpath_node_set::type_unsorted) - { - sort(begin, end, document_order_comparator()); + case xpath_node_set::type_unsorted: + return *min_element(begin, end, document_order_comparator()); - type = xpath_node_set::type_sorted; - } - else - type = sorted; - } + default: + assert(false && "Invalid node set type"); // unreachable + return xpath_node(); + } +} - if (type != order) reverse(begin, end); +class xpath_node_set_raw +{ + xpath_node_set::type_t _type; - return order; + xpath_node* _begin; + xpath_node* _end; + xpath_node* _eos; + + public: + xpath_node_set_raw() + : _type(xpath_node_set::type_unsorted) + , _begin(0) + , _end(0) + , _eos(0) + { } - PUGI__FN xpath_node xpath_first(const xpath_node* begin, const xpath_node* end, xpath_node_set::type_t type) + xpath_node* begin() const { - if (begin == end) return xpath_node(); - - switch (type) - { - case xpath_node_set::type_sorted: - return *begin; - - case xpath_node_set::type_sorted_reverse: - return *(end - 1); - - case xpath_node_set::type_unsorted: - return *min_element(begin, end, document_order_comparator()); - - default: - assert(false && "Invalid node set type"); // unreachable - return xpath_node(); - } + return _begin; } - class xpath_node_set_raw + xpath_node* end() const { - xpath_node_set::type_t _type; + return _end; + } - xpath_node* _begin; - xpath_node* _end; - xpath_node* _eos; - - public: - xpath_node_set_raw(): _type(xpath_node_set::type_unsorted), _begin(0), _end(0), _eos(0) - { - } - - xpath_node* begin() const - { - return _begin; - } - - xpath_node* end() const - { - return _end; - } - - bool empty() const - { - return _begin == _end; - } - - size_t size() const - { - return static_cast(_end - _begin); - } - - xpath_node first() const - { - return xpath_first(_begin, _end, _type); - } - - void push_back_grow(const xpath_node& node, xpath_allocator* alloc); - - void push_back(const xpath_node& node, xpath_allocator* alloc) - { - if (_end != _eos) - *_end++ = node; - else - push_back_grow(node, alloc); - } - - void append(const xpath_node* begin_, const xpath_node* end_, xpath_allocator* alloc) - { - if (begin_ == end_) return; - - size_t size_ = static_cast(_end - _begin); - size_t capacity = static_cast(_eos - _begin); - size_t count = static_cast(end_ - begin_); - - if (size_ + count > capacity) - { - // reallocate the old array or allocate a new one - xpath_node* data = static_cast(alloc->reallocate(_begin, capacity * sizeof(xpath_node), (size_ + count) * sizeof(xpath_node))); - if (!data) return; - - // finalize - _begin = data; - _end = data + size_; - _eos = data + size_ + count; - } - - memcpy(_end, begin_, count * sizeof(xpath_node)); - _end += count; - } - - void sort_do() - { - _type = xpath_sort(_begin, _end, _type, false); - } - - void truncate(xpath_node* pos) - { - assert(_begin <= pos && pos <= _end); - - _end = pos; - } - - void remove_duplicates() - { - if (_type == xpath_node_set::type_unsorted) - sort(_begin, _end, duplicate_comparator()); - - _end = unique(_begin, _end); - } - - xpath_node_set::type_t type() const - { - return _type; - } - - void set_type(xpath_node_set::type_t value) - { - _type = value; - } - }; - - PUGI__FN_NO_INLINE void xpath_node_set_raw::push_back_grow(const xpath_node& node, xpath_allocator* alloc) + bool empty() const { + return _begin == _end; + } + + size_t size() const + { + return static_cast(_end - _begin); + } + + xpath_node first() const + { + return xpath_first(_begin, _end, _type); + } + + void push_back_grow(const xpath_node& node, xpath_allocator* alloc); + + void push_back(const xpath_node& node, xpath_allocator* alloc) + { + if (_end != _eos) + *_end++ = node; + else + push_back_grow(node, alloc); + } + + void append(const xpath_node* begin_, const xpath_node* end_, xpath_allocator* alloc) + { + if (begin_ == end_) + return; + + size_t size_ = static_cast(_end - _begin); size_t capacity = static_cast(_eos - _begin); + size_t count = static_cast(end_ - begin_); - // get new capacity (1.5x rule) - size_t new_capacity = capacity + capacity / 2 + 1; + if (size_ + count > capacity) + { + // reallocate the old array or allocate a new one + xpath_node* data = static_cast(alloc->reallocate(_begin, capacity * sizeof(xpath_node), (size_ + count) * sizeof(xpath_node))); + if (!data) + return; - // reallocate the old array or allocate a new one - xpath_node* data = static_cast(alloc->reallocate(_begin, capacity * sizeof(xpath_node), new_capacity * sizeof(xpath_node))); - if (!data) return; + // finalize + _begin = data; + _end = data + size_; + _eos = data + size_ + count; + } - // finalize - _begin = data; - _end = data + capacity; - _eos = data + new_capacity; - - // push - *_end++ = node; + memcpy(_end, begin_, count * sizeof(xpath_node)); + _end += count; } + + void sort_do() + { + _type = xpath_sort(_begin, _end, _type, false); + } + + void truncate(xpath_node* pos) + { + assert(_begin <= pos && pos <= _end); + + _end = pos; + } + + void remove_duplicates() + { + if (_type == xpath_node_set::type_unsorted) + sort(_begin, _end, duplicate_comparator()); + + _end = unique(_begin, _end); + } + + xpath_node_set::type_t type() const + { + return _type; + } + + void set_type(xpath_node_set::type_t value) + { + _type = value; + } +}; + +PUGI__FN_NO_INLINE void xpath_node_set_raw::push_back_grow(const xpath_node& node, xpath_allocator* alloc) +{ + size_t capacity = static_cast(_eos - _begin); + + // get new capacity (1.5x rule) + size_t new_capacity = capacity + capacity / 2 + 1; + + // reallocate the old array or allocate a new one + xpath_node* data = static_cast(alloc->reallocate(_begin, capacity * sizeof(xpath_node), new_capacity * sizeof(xpath_node))); + if (!data) + return; + + // finalize + _begin = data; + _end = data + capacity; + _eos = data + new_capacity; + + // push + *_end++ = node; +} PUGI__NS_END PUGI__NS_BEGIN - struct xpath_context +struct xpath_context +{ + xpath_node n; + size_t position, size; + + xpath_context(const xpath_node& n_, size_t position_, size_t size_) + : n(n_) + , position(position_) + , size(size_) { - xpath_node n; - size_t position, size; + } +}; - xpath_context(const xpath_node& n_, size_t position_, size_t size_): n(n_), position(position_), size(size_) - { - } - }; +enum lexeme_t +{ + lex_none = 0, + lex_equal, + lex_not_equal, + lex_less, + lex_greater, + lex_less_or_equal, + lex_greater_or_equal, + lex_plus, + lex_minus, + lex_multiply, + lex_union, + lex_var_ref, + lex_open_brace, + lex_close_brace, + lex_quoted_string, + lex_number, + lex_slash, + lex_double_slash, + lex_open_square_brace, + lex_close_square_brace, + lex_string, + lex_comma, + lex_axis_attribute, + lex_dot, + lex_double_dot, + lex_double_colon, + lex_eof +}; - enum lexeme_t +struct xpath_lexer_string +{ + const char_t* begin; + const char_t* end; + + xpath_lexer_string() + : begin(0) + , end(0) { - lex_none = 0, - lex_equal, - lex_not_equal, - lex_less, - lex_greater, - lex_less_or_equal, - lex_greater_or_equal, - lex_plus, - lex_minus, - lex_multiply, - lex_union, - lex_var_ref, - lex_open_brace, - lex_close_brace, - lex_quoted_string, - lex_number, - lex_slash, - lex_double_slash, - lex_open_square_brace, - lex_close_square_brace, - lex_string, - lex_comma, - lex_axis_attribute, - lex_dot, - lex_double_dot, - lex_double_colon, - lex_eof - }; + } - struct xpath_lexer_string + bool operator==(const char_t* other) const { - const char_t* begin; - const char_t* end; + size_t length = static_cast(end - begin); - xpath_lexer_string(): begin(0), end(0) - { - } + return strequalrange(other, begin, length); + } +}; - bool operator==(const char_t* other) const - { - size_t length = static_cast(end - begin); +class xpath_lexer +{ + const char_t* _cur; + const char_t* _cur_lexeme_pos; + xpath_lexer_string _cur_lexeme_contents; - return strequalrange(other, begin, length); - } - }; + lexeme_t _cur_lexeme; - class xpath_lexer + public: + explicit xpath_lexer(const char_t* query) + : _cur(query) { - const char_t* _cur; - const char_t* _cur_lexeme_pos; - xpath_lexer_string _cur_lexeme_contents; + next(); + } - lexeme_t _cur_lexeme; + const char_t* state() const + { + return _cur; + } - public: - explicit xpath_lexer(const char_t* query): _cur(query) + void next() + { + const char_t* cur = _cur; + + while (PUGI__IS_CHARTYPE(*cur, ct_space)) + ++cur; + + // save lexeme position for error reporting + _cur_lexeme_pos = cur; + + switch (*cur) { - next(); - } + case 0: + _cur_lexeme = lex_eof; + break; - const char_t* state() const - { - return _cur; - } - - void next() - { - const char_t* cur = _cur; - - while (PUGI__IS_CHARTYPE(*cur, ct_space)) ++cur; - - // save lexeme position for error reporting - _cur_lexeme_pos = cur; - - switch (*cur) + case '>': + if (*(cur + 1) == '=') { - case 0: - _cur_lexeme = lex_eof; - break; - - case '>': - if (*(cur+1) == '=') - { - cur += 2; - _cur_lexeme = lex_greater_or_equal; - } - else - { - cur += 1; - _cur_lexeme = lex_greater; - } - break; - - case '<': - if (*(cur+1) == '=') - { - cur += 2; - _cur_lexeme = lex_less_or_equal; - } - else - { - cur += 1; - _cur_lexeme = lex_less; - } - break; - - case '!': - if (*(cur+1) == '=') - { - cur += 2; - _cur_lexeme = lex_not_equal; - } - else - { - _cur_lexeme = lex_none; - } - break; - - case '=': - cur += 1; - _cur_lexeme = lex_equal; - - break; - - case '+': - cur += 1; - _cur_lexeme = lex_plus; - - break; - - case '-': - cur += 1; - _cur_lexeme = lex_minus; - - break; - - case '*': - cur += 1; - _cur_lexeme = lex_multiply; - - break; - - case '|': - cur += 1; - _cur_lexeme = lex_union; - - break; - - case '$': - cur += 1; - - if (PUGI__IS_CHARTYPEX(*cur, ctx_start_symbol)) - { - _cur_lexeme_contents.begin = cur; - - while (PUGI__IS_CHARTYPEX(*cur, ctx_symbol)) cur++; - - if (cur[0] == ':' && PUGI__IS_CHARTYPEX(cur[1], ctx_symbol)) // qname - { - cur++; // : - - while (PUGI__IS_CHARTYPEX(*cur, ctx_symbol)) cur++; - } - - _cur_lexeme_contents.end = cur; - - _cur_lexeme = lex_var_ref; - } - else - { - _cur_lexeme = lex_none; - } - - break; - - case '(': - cur += 1; - _cur_lexeme = lex_open_brace; - - break; - - case ')': - cur += 1; - _cur_lexeme = lex_close_brace; - - break; - - case '[': - cur += 1; - _cur_lexeme = lex_open_square_brace; - - break; - - case ']': - cur += 1; - _cur_lexeme = lex_close_square_brace; - - break; - - case ',': - cur += 1; - _cur_lexeme = lex_comma; - - break; - - case '/': - if (*(cur+1) == '/') - { - cur += 2; - _cur_lexeme = lex_double_slash; - } - else - { - cur += 1; - _cur_lexeme = lex_slash; - } - break; - - case '.': - if (*(cur+1) == '.') - { - cur += 2; - _cur_lexeme = lex_double_dot; - } - else if (PUGI__IS_CHARTYPEX(*(cur+1), ctx_digit)) - { - _cur_lexeme_contents.begin = cur; // . - - ++cur; - - while (PUGI__IS_CHARTYPEX(*cur, ctx_digit)) cur++; - - _cur_lexeme_contents.end = cur; - - _cur_lexeme = lex_number; - } - else - { - cur += 1; - _cur_lexeme = lex_dot; - } - break; - - case '@': - cur += 1; - _cur_lexeme = lex_axis_attribute; - - break; - - case '"': - case '\'': - { - char_t terminator = *cur; - - ++cur; - - _cur_lexeme_contents.begin = cur; - while (*cur && *cur != terminator) cur++; - _cur_lexeme_contents.end = cur; - - if (!*cur) - _cur_lexeme = lex_none; - else - { - cur += 1; - _cur_lexeme = lex_quoted_string; - } - - break; - } - - case ':': - if (*(cur+1) == ':') - { - cur += 2; - _cur_lexeme = lex_double_colon; - } - else - { - _cur_lexeme = lex_none; - } - break; - - default: - if (PUGI__IS_CHARTYPEX(*cur, ctx_digit)) - { - _cur_lexeme_contents.begin = cur; - - while (PUGI__IS_CHARTYPEX(*cur, ctx_digit)) cur++; - - if (*cur == '.') - { - cur++; - - while (PUGI__IS_CHARTYPEX(*cur, ctx_digit)) cur++; - } - - _cur_lexeme_contents.end = cur; - - _cur_lexeme = lex_number; - } - else if (PUGI__IS_CHARTYPEX(*cur, ctx_start_symbol)) - { - _cur_lexeme_contents.begin = cur; - - while (PUGI__IS_CHARTYPEX(*cur, ctx_symbol)) cur++; - - if (cur[0] == ':') - { - if (cur[1] == '*') // namespace test ncname:* - { - cur += 2; // :* - } - else if (PUGI__IS_CHARTYPEX(cur[1], ctx_symbol)) // namespace test qname - { - cur++; // : - - while (PUGI__IS_CHARTYPEX(*cur, ctx_symbol)) cur++; - } - } - - _cur_lexeme_contents.end = cur; - - _cur_lexeme = lex_string; - } - else - { - _cur_lexeme = lex_none; - } - } - - _cur = cur; - } - - lexeme_t current() const - { - return _cur_lexeme; - } - - const char_t* current_pos() const - { - return _cur_lexeme_pos; - } - - const xpath_lexer_string& contents() const - { - assert(_cur_lexeme == lex_var_ref || _cur_lexeme == lex_number || _cur_lexeme == lex_string || _cur_lexeme == lex_quoted_string); - - return _cur_lexeme_contents; - } - }; - - enum ast_type_t - { - ast_unknown, - ast_op_or, // left or right - ast_op_and, // left and right - ast_op_equal, // left = right - ast_op_not_equal, // left != right - ast_op_less, // left < right - ast_op_greater, // left > right - ast_op_less_or_equal, // left <= right - ast_op_greater_or_equal, // left >= right - ast_op_add, // left + right - ast_op_subtract, // left - right - ast_op_multiply, // left * right - ast_op_divide, // left / right - ast_op_mod, // left % right - ast_op_negate, // left - right - ast_op_union, // left | right - ast_predicate, // apply predicate to set; next points to next predicate - ast_filter, // select * from left where right - ast_string_constant, // string constant - ast_number_constant, // number constant - ast_variable, // variable - ast_func_last, // last() - ast_func_position, // position() - ast_func_count, // count(left) - ast_func_id, // id(left) - ast_func_local_name_0, // local-name() - ast_func_local_name_1, // local-name(left) - ast_func_namespace_uri_0, // namespace-uri() - ast_func_namespace_uri_1, // namespace-uri(left) - ast_func_name_0, // name() - ast_func_name_1, // name(left) - ast_func_string_0, // string() - ast_func_string_1, // string(left) - ast_func_concat, // concat(left, right, siblings) - ast_func_starts_with, // starts_with(left, right) - ast_func_contains, // contains(left, right) - ast_func_substring_before, // substring-before(left, right) - ast_func_substring_after, // substring-after(left, right) - ast_func_substring_2, // substring(left, right) - ast_func_substring_3, // substring(left, right, third) - ast_func_string_length_0, // string-length() - ast_func_string_length_1, // string-length(left) - ast_func_normalize_space_0, // normalize-space() - ast_func_normalize_space_1, // normalize-space(left) - ast_func_translate, // translate(left, right, third) - ast_func_boolean, // boolean(left) - ast_func_not, // not(left) - ast_func_true, // true() - ast_func_false, // false() - ast_func_lang, // lang(left) - ast_func_number_0, // number() - ast_func_number_1, // number(left) - ast_func_sum, // sum(left) - ast_func_floor, // floor(left) - ast_func_ceiling, // ceiling(left) - ast_func_round, // round(left) - ast_step, // process set left with step - ast_step_root, // select root node - - ast_opt_translate_table, // translate(left, right, third) where right/third are constants - ast_opt_compare_attribute // @name = 'string' - }; - - enum axis_t - { - axis_ancestor, - axis_ancestor_or_self, - axis_attribute, - axis_child, - axis_descendant, - axis_descendant_or_self, - axis_following, - axis_following_sibling, - axis_namespace, - axis_parent, - axis_preceding, - axis_preceding_sibling, - axis_self - }; - - enum nodetest_t - { - nodetest_none, - nodetest_name, - nodetest_type_node, - nodetest_type_comment, - nodetest_type_pi, - nodetest_type_text, - nodetest_pi, - nodetest_all, - nodetest_all_in_namespace - }; - - enum predicate_t - { - predicate_default, - predicate_posinv, - predicate_constant, - predicate_constant_one - }; - - enum nodeset_eval_t - { - nodeset_eval_all, - nodeset_eval_any, - nodeset_eval_first - }; - - template struct axis_to_type - { - static const axis_t axis; - }; - - template const axis_t axis_to_type::axis = N; - - class xpath_ast_node - { - private: - // node type - char _type; - char _rettype; - - // for ast_step - char _axis; - - // for ast_step/ast_predicate/ast_filter - char _test; - - // tree node structure - xpath_ast_node* _left; - xpath_ast_node* _right; - xpath_ast_node* _next; - - union - { - // value for ast_string_constant - const char_t* string; - // value for ast_number_constant - double number; - // variable for ast_variable - xpath_variable* variable; - // node test for ast_step (node name/namespace/node type/pi target) - const char_t* nodetest; - // table for ast_opt_translate_table - const unsigned char* table; - } _data; - - xpath_ast_node(const xpath_ast_node&); - xpath_ast_node& operator=(const xpath_ast_node&); - - template static bool compare_eq(xpath_ast_node* lhs, xpath_ast_node* rhs, const xpath_context& c, const xpath_stack& stack, const Comp& comp) - { - xpath_value_type lt = lhs->rettype(), rt = rhs->rettype(); - - if (lt != xpath_type_node_set && rt != xpath_type_node_set) - { - if (lt == xpath_type_boolean || rt == xpath_type_boolean) - return comp(lhs->eval_boolean(c, stack), rhs->eval_boolean(c, stack)); - else if (lt == xpath_type_number || rt == xpath_type_number) - return comp(lhs->eval_number(c, stack), rhs->eval_number(c, stack)); - else if (lt == xpath_type_string || rt == xpath_type_string) - { - xpath_allocator_capture cr(stack.result); - - xpath_string ls = lhs->eval_string(c, stack); - xpath_string rs = rhs->eval_string(c, stack); - - return comp(ls, rs); - } - } - else if (lt == xpath_type_node_set && rt == xpath_type_node_set) - { - xpath_allocator_capture cr(stack.result); - - xpath_node_set_raw ls = lhs->eval_node_set(c, stack, nodeset_eval_all); - xpath_node_set_raw rs = rhs->eval_node_set(c, stack, nodeset_eval_all); - - for (const xpath_node* li = ls.begin(); li != ls.end(); ++li) - for (const xpath_node* ri = rs.begin(); ri != rs.end(); ++ri) - { - xpath_allocator_capture cri(stack.result); - - if (comp(string_value(*li, stack.result), string_value(*ri, stack.result))) - return true; - } - - return false; + cur += 2; + _cur_lexeme = lex_greater_or_equal; } else { - if (lt == xpath_type_node_set) + cur += 1; + _cur_lexeme = lex_greater; + } + break; + + case '<': + if (*(cur + 1) == '=') + { + cur += 2; + _cur_lexeme = lex_less_or_equal; + } + else + { + cur += 1; + _cur_lexeme = lex_less; + } + break; + + case '!': + if (*(cur + 1) == '=') + { + cur += 2; + _cur_lexeme = lex_not_equal; + } + else + { + _cur_lexeme = lex_none; + } + break; + + case '=': + cur += 1; + _cur_lexeme = lex_equal; + + break; + + case '+': + cur += 1; + _cur_lexeme = lex_plus; + + break; + + case '-': + cur += 1; + _cur_lexeme = lex_minus; + + break; + + case '*': + cur += 1; + _cur_lexeme = lex_multiply; + + break; + + case '|': + cur += 1; + _cur_lexeme = lex_union; + + break; + + case '$': + cur += 1; + + if (PUGI__IS_CHARTYPEX(*cur, ctx_start_symbol)) + { + _cur_lexeme_contents.begin = cur; + + while (PUGI__IS_CHARTYPEX(*cur, ctx_symbol)) + cur++; + + if (cur[0] == ':' && PUGI__IS_CHARTYPEX(cur[1], ctx_symbol)) // qname { - swap(lhs, rhs); - swap(lt, rt); + cur++; // : + + while (PUGI__IS_CHARTYPEX(*cur, ctx_symbol)) + cur++; } - if (lt == xpath_type_boolean) - return comp(lhs->eval_boolean(c, stack), rhs->eval_boolean(c, stack)); - else if (lt == xpath_type_number) - { - xpath_allocator_capture cr(stack.result); + _cur_lexeme_contents.end = cur; - double l = lhs->eval_number(c, stack); - xpath_node_set_raw rs = rhs->eval_node_set(c, stack, nodeset_eval_all); - - for (const xpath_node* ri = rs.begin(); ri != rs.end(); ++ri) - { - xpath_allocator_capture cri(stack.result); - - if (comp(l, convert_string_to_number(string_value(*ri, stack.result).c_str()))) - return true; - } - - return false; - } - else if (lt == xpath_type_string) - { - xpath_allocator_capture cr(stack.result); - - xpath_string l = lhs->eval_string(c, stack); - xpath_node_set_raw rs = rhs->eval_node_set(c, stack, nodeset_eval_all); - - for (const xpath_node* ri = rs.begin(); ri != rs.end(); ++ri) - { - xpath_allocator_capture cri(stack.result); - - if (comp(l, string_value(*ri, stack.result))) - return true; - } - - return false; - } + _cur_lexeme = lex_var_ref; + } + else + { + _cur_lexeme = lex_none; } - assert(false && "Wrong types"); // unreachable - return false; + break; + + case '(': + cur += 1; + _cur_lexeme = lex_open_brace; + + break; + + case ')': + cur += 1; + _cur_lexeme = lex_close_brace; + + break; + + case '[': + cur += 1; + _cur_lexeme = lex_open_square_brace; + + break; + + case ']': + cur += 1; + _cur_lexeme = lex_close_square_brace; + + break; + + case ',': + cur += 1; + _cur_lexeme = lex_comma; + + break; + + case '/': + if (*(cur + 1) == '/') + { + cur += 2; + _cur_lexeme = lex_double_slash; + } + else + { + cur += 1; + _cur_lexeme = lex_slash; + } + break; + + case '.': + if (*(cur + 1) == '.') + { + cur += 2; + _cur_lexeme = lex_double_dot; + } + else if (PUGI__IS_CHARTYPEX(*(cur + 1), ctx_digit)) + { + _cur_lexeme_contents.begin = cur; // . + + ++cur; + + while (PUGI__IS_CHARTYPEX(*cur, ctx_digit)) + cur++; + + _cur_lexeme_contents.end = cur; + + _cur_lexeme = lex_number; + } + else + { + cur += 1; + _cur_lexeme = lex_dot; + } + break; + + case '@': + cur += 1; + _cur_lexeme = lex_axis_attribute; + + break; + + case '"': + case '\'': + { + char_t terminator = *cur; + + ++cur; + + _cur_lexeme_contents.begin = cur; + while (*cur && *cur != terminator) + cur++; + _cur_lexeme_contents.end = cur; + + if (!*cur) + _cur_lexeme = lex_none; + else + { + cur += 1; + _cur_lexeme = lex_quoted_string; + } + + break; } - static bool eval_once(xpath_node_set::type_t type, nodeset_eval_t eval) - { - return type == xpath_node_set::type_sorted ? eval != nodeset_eval_all : eval == nodeset_eval_any; + case ':': + if (*(cur + 1) == ':') + { + cur += 2; + _cur_lexeme = lex_double_colon; + } + else + { + _cur_lexeme = lex_none; + } + break; + + default: + if (PUGI__IS_CHARTYPEX(*cur, ctx_digit)) + { + _cur_lexeme_contents.begin = cur; + + while (PUGI__IS_CHARTYPEX(*cur, ctx_digit)) + cur++; + + if (*cur == '.') + { + cur++; + + while (PUGI__IS_CHARTYPEX(*cur, ctx_digit)) + cur++; + } + + _cur_lexeme_contents.end = cur; + + _cur_lexeme = lex_number; + } + else if (PUGI__IS_CHARTYPEX(*cur, ctx_start_symbol)) + { + _cur_lexeme_contents.begin = cur; + + while (PUGI__IS_CHARTYPEX(*cur, ctx_symbol)) + cur++; + + if (cur[0] == ':') + { + if (cur[1] == '*') // namespace test ncname:* + { + cur += 2; // :* + } + else if (PUGI__IS_CHARTYPEX(cur[1], ctx_symbol)) // namespace test qname + { + cur++; // : + + while (PUGI__IS_CHARTYPEX(*cur, ctx_symbol)) + cur++; + } + } + + _cur_lexeme_contents.end = cur; + + _cur_lexeme = lex_string; + } + else + { + _cur_lexeme = lex_none; + } } - template static bool compare_rel(xpath_ast_node* lhs, xpath_ast_node* rhs, const xpath_context& c, const xpath_stack& stack, const Comp& comp) - { - xpath_value_type lt = lhs->rettype(), rt = rhs->rettype(); + _cur = cur; + } - if (lt != xpath_type_node_set && rt != xpath_type_node_set) + lexeme_t current() const + { + return _cur_lexeme; + } + + const char_t* current_pos() const + { + return _cur_lexeme_pos; + } + + const xpath_lexer_string& contents() const + { + assert(_cur_lexeme == lex_var_ref || _cur_lexeme == lex_number || _cur_lexeme == lex_string || _cur_lexeme == lex_quoted_string); + + return _cur_lexeme_contents; + } +}; + +enum ast_type_t +{ + ast_unknown, + ast_op_or, // left or right + ast_op_and, // left and right + ast_op_equal, // left = right + ast_op_not_equal, // left != right + ast_op_less, // left < right + ast_op_greater, // left > right + ast_op_less_or_equal, // left <= right + ast_op_greater_or_equal, // left >= right + ast_op_add, // left + right + ast_op_subtract, // left - right + ast_op_multiply, // left * right + ast_op_divide, // left / right + ast_op_mod, // left % right + ast_op_negate, // left - right + ast_op_union, // left | right + ast_predicate, // apply predicate to set; next points to next predicate + ast_filter, // select * from left where right + ast_string_constant, // string constant + ast_number_constant, // number constant + ast_variable, // variable + ast_func_last, // last() + ast_func_position, // position() + ast_func_count, // count(left) + ast_func_id, // id(left) + ast_func_local_name_0, // local-name() + ast_func_local_name_1, // local-name(left) + ast_func_namespace_uri_0, // namespace-uri() + ast_func_namespace_uri_1, // namespace-uri(left) + ast_func_name_0, // name() + ast_func_name_1, // name(left) + ast_func_string_0, // string() + ast_func_string_1, // string(left) + ast_func_concat, // concat(left, right, siblings) + ast_func_starts_with, // starts_with(left, right) + ast_func_contains, // contains(left, right) + ast_func_substring_before, // substring-before(left, right) + ast_func_substring_after, // substring-after(left, right) + ast_func_substring_2, // substring(left, right) + ast_func_substring_3, // substring(left, right, third) + ast_func_string_length_0, // string-length() + ast_func_string_length_1, // string-length(left) + ast_func_normalize_space_0, // normalize-space() + ast_func_normalize_space_1, // normalize-space(left) + ast_func_translate, // translate(left, right, third) + ast_func_boolean, // boolean(left) + ast_func_not, // not(left) + ast_func_true, // true() + ast_func_false, // false() + ast_func_lang, // lang(left) + ast_func_number_0, // number() + ast_func_number_1, // number(left) + ast_func_sum, // sum(left) + ast_func_floor, // floor(left) + ast_func_ceiling, // ceiling(left) + ast_func_round, // round(left) + ast_step, // process set left with step + ast_step_root, // select root node + + ast_opt_translate_table, // translate(left, right, third) where right/third are constants + ast_opt_compare_attribute // @name = 'string' +}; + +enum axis_t +{ + axis_ancestor, + axis_ancestor_or_self, + axis_attribute, + axis_child, + axis_descendant, + axis_descendant_or_self, + axis_following, + axis_following_sibling, + axis_namespace, + axis_parent, + axis_preceding, + axis_preceding_sibling, + axis_self +}; + +enum nodetest_t +{ + nodetest_none, + nodetest_name, + nodetest_type_node, + nodetest_type_comment, + nodetest_type_pi, + nodetest_type_text, + nodetest_pi, + nodetest_all, + nodetest_all_in_namespace +}; + +enum predicate_t +{ + predicate_default, + predicate_posinv, + predicate_constant, + predicate_constant_one +}; + +enum nodeset_eval_t +{ + nodeset_eval_all, + nodeset_eval_any, + nodeset_eval_first +}; + +template +struct axis_to_type +{ + static const axis_t axis; +}; + +template +const axis_t axis_to_type::axis = N; + +class xpath_ast_node +{ + private: + // node type + char _type; + char _rettype; + + // for ast_step + char _axis; + + // for ast_step/ast_predicate/ast_filter + char _test; + + // tree node structure + xpath_ast_node* _left; + xpath_ast_node* _right; + xpath_ast_node* _next; + + union { + // value for ast_string_constant + const char_t* string; + // value for ast_number_constant + double number; + // variable for ast_variable + xpath_variable* variable; + // node test for ast_step (node name/namespace/node type/pi target) + const char_t* nodetest; + // table for ast_opt_translate_table + const unsigned char* table; + } _data; + + xpath_ast_node(const xpath_ast_node&); + xpath_ast_node& operator=(const xpath_ast_node&); + + template + static bool compare_eq(xpath_ast_node* lhs, xpath_ast_node* rhs, const xpath_context& c, const xpath_stack& stack, const Comp& comp) + { + xpath_value_type lt = lhs->rettype(), rt = rhs->rettype(); + + if (lt != xpath_type_node_set && rt != xpath_type_node_set) + { + if (lt == xpath_type_boolean || rt == xpath_type_boolean) + return comp(lhs->eval_boolean(c, stack), rhs->eval_boolean(c, stack)); + else if (lt == xpath_type_number || rt == xpath_type_number) return comp(lhs->eval_number(c, stack), rhs->eval_number(c, stack)); - else if (lt == xpath_type_node_set && rt == xpath_type_node_set) + else if (lt == xpath_type_string || rt == xpath_type_string) { xpath_allocator_capture cr(stack.result); - xpath_node_set_raw ls = lhs->eval_node_set(c, stack, nodeset_eval_all); - xpath_node_set_raw rs = rhs->eval_node_set(c, stack, nodeset_eval_all); + xpath_string ls = lhs->eval_string(c, stack); + xpath_string rs = rhs->eval_string(c, stack); - for (const xpath_node* li = ls.begin(); li != ls.end(); ++li) + return comp(ls, rs); + } + } + else if (lt == xpath_type_node_set && rt == xpath_type_node_set) + { + xpath_allocator_capture cr(stack.result); + + xpath_node_set_raw ls = lhs->eval_node_set(c, stack, nodeset_eval_all); + xpath_node_set_raw rs = rhs->eval_node_set(c, stack, nodeset_eval_all); + + for (const xpath_node* li = ls.begin(); li != ls.end(); ++li) + for (const xpath_node* ri = rs.begin(); ri != rs.end(); ++ri) { xpath_allocator_capture cri(stack.result); - double l = convert_string_to_number(string_value(*li, stack.result).c_str()); - - for (const xpath_node* ri = rs.begin(); ri != rs.end(); ++ri) - { - xpath_allocator_capture crii(stack.result); - - if (comp(l, convert_string_to_number(string_value(*ri, stack.result).c_str()))) - return true; - } + if (comp(string_value(*li, stack.result), string_value(*ri, stack.result))) + return true; } - return false; + return false; + } + else + { + if (lt == xpath_type_node_set) + { + swap(lhs, rhs); + swap(lt, rt); } - else if (lt != xpath_type_node_set && rt == xpath_type_node_set) + + if (lt == xpath_type_boolean) + return comp(lhs->eval_boolean(c, stack), rhs->eval_boolean(c, stack)); + else if (lt == xpath_type_number) { xpath_allocator_capture cr(stack.result); @@ -9546,3186 +10122,3427 @@ PUGI__NS_BEGIN return false; } - else if (lt == xpath_type_node_set && rt != xpath_type_node_set) + else if (lt == xpath_type_string) { xpath_allocator_capture cr(stack.result); - xpath_node_set_raw ls = lhs->eval_node_set(c, stack, nodeset_eval_all); - double r = rhs->eval_number(c, stack); + xpath_string l = lhs->eval_string(c, stack); + xpath_node_set_raw rs = rhs->eval_node_set(c, stack, nodeset_eval_all); - for (const xpath_node* li = ls.begin(); li != ls.end(); ++li) + for (const xpath_node* ri = rs.begin(); ri != rs.end(); ++ri) { xpath_allocator_capture cri(stack.result); - if (comp(convert_string_to_number(string_value(*li, stack.result).c_str()), r)) + if (comp(l, string_value(*ri, stack.result))) return true; } return false; } - else + } + + assert(false && "Wrong types"); // unreachable + return false; + } + + static bool eval_once(xpath_node_set::type_t type, nodeset_eval_t eval) + { + return type == xpath_node_set::type_sorted ? eval != nodeset_eval_all : eval == nodeset_eval_any; + } + + template + static bool compare_rel(xpath_ast_node* lhs, xpath_ast_node* rhs, const xpath_context& c, const xpath_stack& stack, const Comp& comp) + { + xpath_value_type lt = lhs->rettype(), rt = rhs->rettype(); + + if (lt != xpath_type_node_set && rt != xpath_type_node_set) + return comp(lhs->eval_number(c, stack), rhs->eval_number(c, stack)); + else if (lt == xpath_type_node_set && rt == xpath_type_node_set) + { + xpath_allocator_capture cr(stack.result); + + xpath_node_set_raw ls = lhs->eval_node_set(c, stack, nodeset_eval_all); + xpath_node_set_raw rs = rhs->eval_node_set(c, stack, nodeset_eval_all); + + for (const xpath_node* li = ls.begin(); li != ls.end(); ++li) { - assert(false && "Wrong types"); // unreachable + xpath_allocator_capture cri(stack.result); + + double l = convert_string_to_number(string_value(*li, stack.result).c_str()); + + for (const xpath_node* ri = rs.begin(); ri != rs.end(); ++ri) + { + xpath_allocator_capture crii(stack.result); + + if (comp(l, convert_string_to_number(string_value(*ri, stack.result).c_str()))) + return true; + } + } + + return false; + } + else if (lt != xpath_type_node_set && rt == xpath_type_node_set) + { + xpath_allocator_capture cr(stack.result); + + double l = lhs->eval_number(c, stack); + xpath_node_set_raw rs = rhs->eval_node_set(c, stack, nodeset_eval_all); + + for (const xpath_node* ri = rs.begin(); ri != rs.end(); ++ri) + { + xpath_allocator_capture cri(stack.result); + + if (comp(l, convert_string_to_number(string_value(*ri, stack.result).c_str()))) + return true; + } + + return false; + } + else if (lt == xpath_type_node_set && rt != xpath_type_node_set) + { + xpath_allocator_capture cr(stack.result); + + xpath_node_set_raw ls = lhs->eval_node_set(c, stack, nodeset_eval_all); + double r = rhs->eval_number(c, stack); + + for (const xpath_node* li = ls.begin(); li != ls.end(); ++li) + { + xpath_allocator_capture cri(stack.result); + + if (comp(convert_string_to_number(string_value(*li, stack.result).c_str()), r)) + return true; + } + + return false; + } + else + { + assert(false && "Wrong types"); // unreachable + return false; + } + } + + static void apply_predicate_boolean(xpath_node_set_raw& ns, size_t first, xpath_ast_node* expr, const xpath_stack& stack, bool once) + { + assert(ns.size() >= first); + assert(expr->rettype() != xpath_type_number); + + size_t i = 1; + size_t size = ns.size() - first; + + xpath_node* last = ns.begin() + first; + + // remove_if... or well, sort of + for (xpath_node* it = last; it != ns.end(); ++it, ++i) + { + xpath_context c(*it, i, size); + + if (expr->eval_boolean(c, stack)) + { + *last++ = *it; + + if (once) + break; + } + } + + ns.truncate(last); + } + + static void apply_predicate_number(xpath_node_set_raw& ns, size_t first, xpath_ast_node* expr, const xpath_stack& stack, bool once) + { + assert(ns.size() >= first); + assert(expr->rettype() == xpath_type_number); + + size_t i = 1; + size_t size = ns.size() - first; + + xpath_node* last = ns.begin() + first; + + // remove_if... or well, sort of + for (xpath_node* it = last; it != ns.end(); ++it, ++i) + { + xpath_context c(*it, i, size); + + if (expr->eval_number(c, stack) == i) + { + *last++ = *it; + + if (once) + break; + } + } + + ns.truncate(last); + } + + static void apply_predicate_number_const(xpath_node_set_raw& ns, size_t first, xpath_ast_node* expr, const xpath_stack& stack) + { + assert(ns.size() >= first); + assert(expr->rettype() == xpath_type_number); + + size_t size = ns.size() - first; + + xpath_node* last = ns.begin() + first; + + xpath_context c(xpath_node(), 1, size); + + double er = expr->eval_number(c, stack); + + if (er >= 1.0 && er <= size) + { + size_t eri = static_cast(er); + + if (er == eri) + { + xpath_node r = last[eri - 1]; + + *last++ = r; + } + } + + ns.truncate(last); + } + + void apply_predicate(xpath_node_set_raw& ns, size_t first, const xpath_stack& stack, bool once) + { + if (ns.size() == first) + return; + + assert(_type == ast_filter || _type == ast_predicate); + + if (_test == predicate_constant || _test == predicate_constant_one) + apply_predicate_number_const(ns, first, _right, stack); + else if (_right->rettype() == xpath_type_number) + apply_predicate_number(ns, first, _right, stack, once); + else + apply_predicate_boolean(ns, first, _right, stack, once); + } + + void apply_predicates(xpath_node_set_raw& ns, size_t first, const xpath_stack& stack, nodeset_eval_t eval) + { + if (ns.size() == first) + return; + + bool last_once = eval_once(ns.type(), eval); + + for (xpath_ast_node* pred = _right; pred; pred = pred->_next) + pred->apply_predicate(ns, first, stack, !pred->_next && last_once); + } + + bool step_push(xpath_node_set_raw& ns, xml_attribute_struct* a, xml_node_struct* parent, xpath_allocator* alloc) + { + assert(a); + + const char_t* name = a->name ? a->name + 0 : PUGIXML_TEXT(""); + + switch (_test) + { + case nodetest_name: + if (strequal(name, _data.nodetest) && is_xpath_attribute(name)) + { + ns.push_back(xpath_node(xml_attribute(a), xml_node(parent)), alloc); + return true; + } + break; + + case nodetest_type_node: + case nodetest_all: + if (is_xpath_attribute(name)) + { + ns.push_back(xpath_node(xml_attribute(a), xml_node(parent)), alloc); + return true; + } + break; + + case nodetest_all_in_namespace: + if (starts_with(name, _data.nodetest) && is_xpath_attribute(name)) + { + ns.push_back(xpath_node(xml_attribute(a), xml_node(parent)), alloc); + return true; + } + break; + + default:; + } + + return false; + } + + bool step_push(xpath_node_set_raw& ns, xml_node_struct* n, xpath_allocator* alloc) + { + assert(n); + + xml_node_type type = PUGI__NODETYPE(n); + + switch (_test) + { + case nodetest_name: + if (type == node_element && n->name && strequal(n->name, _data.nodetest)) + { + ns.push_back(xml_node(n), alloc); + return true; + } + break; + + case nodetest_type_node: + ns.push_back(xml_node(n), alloc); + return true; + + case nodetest_type_comment: + if (type == node_comment) + { + ns.push_back(xml_node(n), alloc); + return true; + } + break; + + case nodetest_type_text: + if (type == node_pcdata || type == node_cdata) + { + ns.push_back(xml_node(n), alloc); + return true; + } + break; + + case nodetest_type_pi: + if (type == node_pi) + { + ns.push_back(xml_node(n), alloc); + return true; + } + break; + + case nodetest_pi: + if (type == node_pi && n->name && strequal(n->name, _data.nodetest)) + { + ns.push_back(xml_node(n), alloc); + return true; + } + break; + + case nodetest_all: + if (type == node_element) + { + ns.push_back(xml_node(n), alloc); + return true; + } + break; + + case nodetest_all_in_namespace: + if (type == node_element && n->name && starts_with(n->name, _data.nodetest)) + { + ns.push_back(xml_node(n), alloc); + return true; + } + break; + + default: + assert(false && "Unknown axis"); // unreachable + } + + return false; + } + + template + void step_fill(xpath_node_set_raw& ns, xml_node_struct* n, xpath_allocator* alloc, bool once, T) + { + const axis_t axis = T::axis; + + switch (axis) + { + case axis_attribute: + { + for (xml_attribute_struct* a = n->first_attribute; a; a = a->next_attribute) + if (step_push(ns, a, n, alloc) & once) + return; + + break; + } + + case axis_child: + { + for (xml_node_struct* c = n->first_child; c; c = c->next_sibling) + if (step_push(ns, c, alloc) & once) + return; + + break; + } + + case axis_descendant: + case axis_descendant_or_self: + { + if (axis == axis_descendant_or_self) + if (step_push(ns, n, alloc) & once) + return; + + xml_node_struct* cur = n->first_child; + + while (cur) + { + if (step_push(ns, cur, alloc) & once) + return; + + if (cur->first_child) + cur = cur->first_child; + else + { + while (!cur->next_sibling) + { + cur = cur->parent; + + if (cur == n) + return; + } + + cur = cur->next_sibling; + } + } + + break; + } + + case axis_following_sibling: + { + for (xml_node_struct* c = n->next_sibling; c; c = c->next_sibling) + if (step_push(ns, c, alloc) & once) + return; + + break; + } + + case axis_preceding_sibling: + { + for (xml_node_struct* c = n->prev_sibling_c; c->next_sibling; c = c->prev_sibling_c) + if (step_push(ns, c, alloc) & once) + return; + + break; + } + + case axis_following: + { + xml_node_struct* cur = n; + + // exit from this node so that we don't include descendants + while (!cur->next_sibling) + { + cur = cur->parent; + + if (!cur) + return; + } + + cur = cur->next_sibling; + + while (cur) + { + if (step_push(ns, cur, alloc) & once) + return; + + if (cur->first_child) + cur = cur->first_child; + else + { + while (!cur->next_sibling) + { + cur = cur->parent; + + if (!cur) + return; + } + + cur = cur->next_sibling; + } + } + + break; + } + + case axis_preceding: + { + xml_node_struct* cur = n; + + // exit from this node so that we don't include descendants + while (!cur->prev_sibling_c->next_sibling) + { + cur = cur->parent; + + if (!cur) + return; + } + + cur = cur->prev_sibling_c; + + while (cur) + { + if (cur->first_child) + cur = cur->first_child->prev_sibling_c; + else + { + // leaf node, can't be ancestor + if (step_push(ns, cur, alloc) & once) + return; + + while (!cur->prev_sibling_c->next_sibling) + { + cur = cur->parent; + + if (!cur) + return; + + if (!node_is_ancestor(cur, n)) + if (step_push(ns, cur, alloc) & once) + return; + } + + cur = cur->prev_sibling_c; + } + } + + break; + } + + case axis_ancestor: + case axis_ancestor_or_self: + { + if (axis == axis_ancestor_or_self) + if (step_push(ns, n, alloc) & once) + return; + + xml_node_struct* cur = n->parent; + + while (cur) + { + if (step_push(ns, cur, alloc) & once) + return; + + cur = cur->parent; + } + + break; + } + + case axis_self: + { + step_push(ns, n, alloc); + + break; + } + + case axis_parent: + { + if (n->parent) + step_push(ns, n->parent, alloc); + + break; + } + + default: + assert(false && "Unimplemented axis"); // unreachable + } + } + + template + void step_fill(xpath_node_set_raw& ns, xml_attribute_struct* a, xml_node_struct* p, xpath_allocator* alloc, bool once, T v) + { + const axis_t axis = T::axis; + + switch (axis) + { + case axis_ancestor: + case axis_ancestor_or_self: + { + if (axis == axis_ancestor_or_self && _test == nodetest_type_node) // reject attributes based on principal node type test + if (step_push(ns, a, p, alloc) & once) + return; + + xml_node_struct* cur = p; + + while (cur) + { + if (step_push(ns, cur, alloc) & once) + return; + + cur = cur->parent; + } + + break; + } + + case axis_descendant_or_self: + case axis_self: + { + if (_test == nodetest_type_node) // reject attributes based on principal node type test + step_push(ns, a, p, alloc); + + break; + } + + case axis_following: + { + xml_node_struct* cur = p; + + while (cur) + { + if (cur->first_child) + cur = cur->first_child; + else + { + while (!cur->next_sibling) + { + cur = cur->parent; + + if (!cur) + return; + } + + cur = cur->next_sibling; + } + + if (step_push(ns, cur, alloc) & once) + return; + } + + break; + } + + case axis_parent: + { + step_push(ns, p, alloc); + + break; + } + + case axis_preceding: + { + // preceding:: axis does not include attribute nodes and attribute ancestors (they are the same as parent's ancestors), so we can reuse node preceding + step_fill(ns, p, alloc, once, v); + break; + } + + default: + assert(false && "Unimplemented axis"); // unreachable + } + } + + template + void step_fill(xpath_node_set_raw& ns, const xpath_node& xn, xpath_allocator* alloc, bool once, T v) + { + const axis_t axis = T::axis; + const bool axis_has_attributes = (axis == axis_ancestor || axis == axis_ancestor_or_self || axis == axis_descendant_or_self || axis == axis_following || axis == axis_parent || axis == axis_preceding || axis == axis_self); + + if (xn.node()) + step_fill(ns, xn.node().internal_object(), alloc, once, v); + else if (axis_has_attributes && xn.attribute() && xn.parent()) + step_fill(ns, xn.attribute().internal_object(), xn.parent().internal_object(), alloc, once, v); + } + + template + xpath_node_set_raw step_do(const xpath_context& c, const xpath_stack& stack, nodeset_eval_t eval, T v) + { + const axis_t axis = T::axis; + const bool axis_reverse = (axis == axis_ancestor || axis == axis_ancestor_or_self || axis == axis_preceding || axis == axis_preceding_sibling); + const xpath_node_set::type_t axis_type = axis_reverse ? xpath_node_set::type_sorted_reverse : xpath_node_set::type_sorted; + + bool once = + (axis == axis_attribute && _test == nodetest_name) || + (!_right && eval_once(axis_type, eval)) || + (_right && !_right->_next && _right->_test == predicate_constant_one); + + xpath_node_set_raw ns; + ns.set_type(axis_type); + + if (_left) + { + xpath_node_set_raw s = _left->eval_node_set(c, stack, nodeset_eval_all); + + // self axis preserves the original order + if (axis == axis_self) + ns.set_type(s.type()); + + for (const xpath_node* it = s.begin(); it != s.end(); ++it) + { + size_t size = ns.size(); + + // in general, all axes generate elements in a particular order, but there is no order guarantee if axis is applied to two nodes + if (axis != axis_self && size != 0) + ns.set_type(xpath_node_set::type_unsorted); + + step_fill(ns, *it, stack.result, once, v); + if (_right) + apply_predicates(ns, size, stack, eval); + } + } + else + { + step_fill(ns, c.n, stack.result, once, v); + if (_right) + apply_predicates(ns, 0, stack, eval); + } + + // child, attribute and self axes always generate unique set of nodes + // for other axis, if the set stayed sorted, it stayed unique because the traversal algorithms do not visit the same node twice + if (axis != axis_child && axis != axis_attribute && axis != axis_self && ns.type() == xpath_node_set::type_unsorted) + ns.remove_duplicates(); + + return ns; + } + + public: + xpath_ast_node(ast_type_t type, xpath_value_type rettype_, const char_t* value) + : _type(static_cast(type)) + , _rettype(static_cast(rettype_)) + , _axis(0) + , _test(0) + , _left(0) + , _right(0) + , _next(0) + { + assert(type == ast_string_constant); + _data.string = value; + } + + xpath_ast_node(ast_type_t type, xpath_value_type rettype_, double value) + : _type(static_cast(type)) + , _rettype(static_cast(rettype_)) + , _axis(0) + , _test(0) + , _left(0) + , _right(0) + , _next(0) + { + assert(type == ast_number_constant); + _data.number = value; + } + + xpath_ast_node(ast_type_t type, xpath_value_type rettype_, xpath_variable* value) + : _type(static_cast(type)) + , _rettype(static_cast(rettype_)) + , _axis(0) + , _test(0) + , _left(0) + , _right(0) + , _next(0) + { + assert(type == ast_variable); + _data.variable = value; + } + + xpath_ast_node(ast_type_t type, xpath_value_type rettype_, xpath_ast_node* left = 0, xpath_ast_node* right = 0) + : _type(static_cast(type)) + , _rettype(static_cast(rettype_)) + , _axis(0) + , _test(0) + , _left(left) + , _right(right) + , _next(0) + { + } + + xpath_ast_node(ast_type_t type, xpath_ast_node* left, axis_t axis, nodetest_t test, const char_t* contents) + : _type(static_cast(type)) + , _rettype(xpath_type_node_set) + , _axis(static_cast(axis)) + , _test(static_cast(test)) + , _left(left) + , _right(0) + , _next(0) + { + assert(type == ast_step); + _data.nodetest = contents; + } + + xpath_ast_node(ast_type_t type, xpath_ast_node* left, xpath_ast_node* right, predicate_t test) + : _type(static_cast(type)) + , _rettype(xpath_type_node_set) + , _axis(0) + , _test(static_cast(test)) + , _left(left) + , _right(right) + , _next(0) + { + assert(type == ast_filter || type == ast_predicate); + } + + void set_next(xpath_ast_node* value) + { + _next = value; + } + + void set_right(xpath_ast_node* value) + { + _right = value; + } + + bool eval_boolean(const xpath_context& c, const xpath_stack& stack) + { + switch (_type) + { + case ast_op_or: + return _left->eval_boolean(c, stack) || _right->eval_boolean(c, stack); + + case ast_op_and: + return _left->eval_boolean(c, stack) && _right->eval_boolean(c, stack); + + case ast_op_equal: + return compare_eq(_left, _right, c, stack, equal_to()); + + case ast_op_not_equal: + return compare_eq(_left, _right, c, stack, not_equal_to()); + + case ast_op_less: + return compare_rel(_left, _right, c, stack, less()); + + case ast_op_greater: + return compare_rel(_right, _left, c, stack, less()); + + case ast_op_less_or_equal: + return compare_rel(_left, _right, c, stack, less_equal()); + + case ast_op_greater_or_equal: + return compare_rel(_right, _left, c, stack, less_equal()); + + case ast_func_starts_with: + { + xpath_allocator_capture cr(stack.result); + + xpath_string lr = _left->eval_string(c, stack); + xpath_string rr = _right->eval_string(c, stack); + + return starts_with(lr.c_str(), rr.c_str()); + } + + case ast_func_contains: + { + xpath_allocator_capture cr(stack.result); + + xpath_string lr = _left->eval_string(c, stack); + xpath_string rr = _right->eval_string(c, stack); + + return find_substring(lr.c_str(), rr.c_str()) != 0; + } + + case ast_func_boolean: + return _left->eval_boolean(c, stack); + + case ast_func_not: + return !_left->eval_boolean(c, stack); + + case ast_func_true: + return true; + + case ast_func_false: + return false; + + case ast_func_lang: + { + if (c.n.attribute()) + return false; + + xpath_allocator_capture cr(stack.result); + + xpath_string lang = _left->eval_string(c, stack); + + for (xml_node n = c.n.node(); n; n = n.parent()) + { + xml_attribute a = n.attribute(PUGIXML_TEXT("xml:lang")); + + if (a) + { + const char_t* value = a.value(); + + // strnicmp / strncasecmp is not portable + for (const char_t* lit = lang.c_str(); *lit; ++lit) + { + if (tolower_ascii(*lit) != tolower_ascii(*value)) + return false; + ++value; + } + + return *value == 0 || *value == '-'; + } + } + + return false; + } + + case ast_opt_compare_attribute: + { + const char_t* value = (_right->_type == ast_string_constant) ? _right->_data.string : _right->_data.variable->get_string(); + + xml_attribute attr = c.n.node().attribute(_left->_data.nodetest); + + return attr && strequal(attr.value(), value) && is_xpath_attribute(attr.name()); + } + + case ast_variable: + { + assert(_rettype == _data.variable->type()); + + if (_rettype == xpath_type_boolean) + return _data.variable->get_boolean(); + } + + // fallthrough + default: + { + switch (_rettype) + { + case xpath_type_number: + return convert_number_to_boolean(eval_number(c, stack)); + + case xpath_type_string: + { + xpath_allocator_capture cr(stack.result); + + return !eval_string(c, stack).empty(); + } + + case xpath_type_node_set: + { + xpath_allocator_capture cr(stack.result); + + return !eval_node_set(c, stack, nodeset_eval_any).empty(); + } + + default: + assert(false && "Wrong expression for return type boolean"); // unreachable return false; } } + } + } - static void apply_predicate_boolean(xpath_node_set_raw& ns, size_t first, xpath_ast_node* expr, const xpath_stack& stack, bool once) + double eval_number(const xpath_context& c, const xpath_stack& stack) + { + switch (_type) { - assert(ns.size() >= first); - assert(expr->rettype() != xpath_type_number); + case ast_op_add: + return _left->eval_number(c, stack) + _right->eval_number(c, stack); - size_t i = 1; - size_t size = ns.size() - first; + case ast_op_subtract: + return _left->eval_number(c, stack) - _right->eval_number(c, stack); - xpath_node* last = ns.begin() + first; + case ast_op_multiply: + return _left->eval_number(c, stack) * _right->eval_number(c, stack); - // remove_if... or well, sort of - for (xpath_node* it = last; it != ns.end(); ++it, ++i) + case ast_op_divide: + return _left->eval_number(c, stack) / _right->eval_number(c, stack); + + case ast_op_mod: + return fmod(_left->eval_number(c, stack), _right->eval_number(c, stack)); + + case ast_op_negate: + return -_left->eval_number(c, stack); + + case ast_number_constant: + return _data.number; + + case ast_func_last: + return static_cast(c.size); + + case ast_func_position: + return static_cast(c.position); + + case ast_func_count: + { + xpath_allocator_capture cr(stack.result); + + return static_cast(_left->eval_node_set(c, stack, nodeset_eval_all).size()); + } + + case ast_func_string_length_0: + { + xpath_allocator_capture cr(stack.result); + + return static_cast(string_value(c.n, stack.result).length()); + } + + case ast_func_string_length_1: + { + xpath_allocator_capture cr(stack.result); + + return static_cast(_left->eval_string(c, stack).length()); + } + + case ast_func_number_0: + { + xpath_allocator_capture cr(stack.result); + + return convert_string_to_number(string_value(c.n, stack.result).c_str()); + } + + case ast_func_number_1: + return _left->eval_number(c, stack); + + case ast_func_sum: + { + xpath_allocator_capture cr(stack.result); + + double r = 0; + + xpath_node_set_raw ns = _left->eval_node_set(c, stack, nodeset_eval_all); + + for (const xpath_node* it = ns.begin(); it != ns.end(); ++it) { - xpath_context c(*it, i, size); + xpath_allocator_capture cri(stack.result); - if (expr->eval_boolean(c, stack)) - { - *last++ = *it; - - if (once) break; - } + r += convert_string_to_number(string_value(*it, stack.result).c_str()); } - ns.truncate(last); + return r; } - static void apply_predicate_number(xpath_node_set_raw& ns, size_t first, xpath_ast_node* expr, const xpath_stack& stack, bool once) + case ast_func_floor: { - assert(ns.size() >= first); - assert(expr->rettype() == xpath_type_number); + double r = _left->eval_number(c, stack); - size_t i = 1; - size_t size = ns.size() - first; + return r == r ? floor(r) : r; + } - xpath_node* last = ns.begin() + first; + case ast_func_ceiling: + { + double r = _left->eval_number(c, stack); - // remove_if... or well, sort of - for (xpath_node* it = last; it != ns.end(); ++it, ++i) + return r == r ? ceil(r) : r; + } + + case ast_func_round: + return round_nearest_nzero(_left->eval_number(c, stack)); + + case ast_variable: + { + assert(_rettype == _data.variable->type()); + + if (_rettype == xpath_type_number) + return _data.variable->get_number(); + } + + // fallthrough + default: + { + switch (_rettype) { - xpath_context c(*it, i, size); + case xpath_type_boolean: + return eval_boolean(c, stack) ? 1 : 0; - if (expr->eval_number(c, stack) == i) - { - *last++ = *it; + case xpath_type_string: + { + xpath_allocator_capture cr(stack.result); - if (once) break; - } + return convert_string_to_number(eval_string(c, stack).c_str()); } - ns.truncate(last); - } - - static void apply_predicate_number_const(xpath_node_set_raw& ns, size_t first, xpath_ast_node* expr, const xpath_stack& stack) - { - assert(ns.size() >= first); - assert(expr->rettype() == xpath_type_number); - - size_t size = ns.size() - first; - - xpath_node* last = ns.begin() + first; - - xpath_context c(xpath_node(), 1, size); - - double er = expr->eval_number(c, stack); - - if (er >= 1.0 && er <= size) + case xpath_type_node_set: { - size_t eri = static_cast(er); + xpath_allocator_capture cr(stack.result); - if (er == eri) - { - xpath_node r = last[eri - 1]; - - *last++ = r; - } + return convert_string_to_number(eval_string(c, stack).c_str()); } - ns.truncate(last); - } - - void apply_predicate(xpath_node_set_raw& ns, size_t first, const xpath_stack& stack, bool once) - { - if (ns.size() == first) return; - - assert(_type == ast_filter || _type == ast_predicate); - - if (_test == predicate_constant || _test == predicate_constant_one) - apply_predicate_number_const(ns, first, _right, stack); - else if (_right->rettype() == xpath_type_number) - apply_predicate_number(ns, first, _right, stack, once); - else - apply_predicate_boolean(ns, first, _right, stack, once); - } - - void apply_predicates(xpath_node_set_raw& ns, size_t first, const xpath_stack& stack, nodeset_eval_t eval) - { - if (ns.size() == first) return; - - bool last_once = eval_once(ns.type(), eval); - - for (xpath_ast_node* pred = _right; pred; pred = pred->_next) - pred->apply_predicate(ns, first, stack, !pred->_next && last_once); - } - - bool step_push(xpath_node_set_raw& ns, xml_attribute_struct* a, xml_node_struct* parent, xpath_allocator* alloc) - { - assert(a); - - const char_t* name = a->name ? a->name + 0 : PUGIXML_TEXT(""); - - switch (_test) - { - case nodetest_name: - if (strequal(name, _data.nodetest) && is_xpath_attribute(name)) - { - ns.push_back(xpath_node(xml_attribute(a), xml_node(parent)), alloc); - return true; - } - break; - - case nodetest_type_node: - case nodetest_all: - if (is_xpath_attribute(name)) - { - ns.push_back(xpath_node(xml_attribute(a), xml_node(parent)), alloc); - return true; - } - break; - - case nodetest_all_in_namespace: - if (starts_with(name, _data.nodetest) && is_xpath_attribute(name)) - { - ns.push_back(xpath_node(xml_attribute(a), xml_node(parent)), alloc); - return true; - } - break; - default: - ; + assert(false && "Wrong expression for return type number"); // unreachable + return 0; } + } + } + } - return false; + xpath_string eval_string_concat(const xpath_context& c, const xpath_stack& stack) + { + assert(_type == ast_func_concat); + + xpath_allocator_capture ct(stack.temp); + + // count the string number + size_t count = 1; + for (xpath_ast_node* nc = _right; nc; nc = nc->_next) + count++; + + // allocate a buffer for temporary string objects + xpath_string* buffer = static_cast(stack.temp->allocate(count * sizeof(xpath_string))); + if (!buffer) + return xpath_string(); + + // evaluate all strings to temporary stack + xpath_stack swapped_stack = {stack.temp, stack.result}; + + buffer[0] = _left->eval_string(c, swapped_stack); + + size_t pos = 1; + for (xpath_ast_node* n = _right; n; n = n->_next, ++pos) + buffer[pos] = n->eval_string(c, swapped_stack); + assert(pos == count); + + // get total length + size_t length = 0; + for (size_t i = 0; i < count; ++i) + length += buffer[i].length(); + + // create final string + char_t* result = static_cast(stack.result->allocate((length + 1) * sizeof(char_t))); + if (!result) + return xpath_string(); + + char_t* ri = result; + + for (size_t j = 0; j < count; ++j) + for (const char_t* bi = buffer[j].c_str(); *bi; ++bi) + *ri++ = *bi; + + *ri = 0; + + return xpath_string::from_heap_preallocated(result, ri); + } + + xpath_string eval_string(const xpath_context& c, const xpath_stack& stack) + { + switch (_type) + { + case ast_string_constant: + return xpath_string::from_const(_data.string); + + case ast_func_local_name_0: + { + xpath_node na = c.n; + + return xpath_string::from_const(local_name(na)); } - bool step_push(xpath_node_set_raw& ns, xml_node_struct* n, xpath_allocator* alloc) + case ast_func_local_name_1: { - assert(n); + xpath_allocator_capture cr(stack.result); - xml_node_type type = PUGI__NODETYPE(n); + xpath_node_set_raw ns = _left->eval_node_set(c, stack, nodeset_eval_first); + xpath_node na = ns.first(); - switch (_test) + return xpath_string::from_const(local_name(na)); + } + + case ast_func_name_0: + { + xpath_node na = c.n; + + return xpath_string::from_const(qualified_name(na)); + } + + case ast_func_name_1: + { + xpath_allocator_capture cr(stack.result); + + xpath_node_set_raw ns = _left->eval_node_set(c, stack, nodeset_eval_first); + xpath_node na = ns.first(); + + return xpath_string::from_const(qualified_name(na)); + } + + case ast_func_namespace_uri_0: + { + xpath_node na = c.n; + + return xpath_string::from_const(namespace_uri(na)); + } + + case ast_func_namespace_uri_1: + { + xpath_allocator_capture cr(stack.result); + + xpath_node_set_raw ns = _left->eval_node_set(c, stack, nodeset_eval_first); + xpath_node na = ns.first(); + + return xpath_string::from_const(namespace_uri(na)); + } + + case ast_func_string_0: + return string_value(c.n, stack.result); + + case ast_func_string_1: + return _left->eval_string(c, stack); + + case ast_func_concat: + return eval_string_concat(c, stack); + + case ast_func_substring_before: + { + xpath_allocator_capture cr(stack.temp); + + xpath_stack swapped_stack = {stack.temp, stack.result}; + + xpath_string s = _left->eval_string(c, swapped_stack); + xpath_string p = _right->eval_string(c, swapped_stack); + + const char_t* pos = find_substring(s.c_str(), p.c_str()); + + return pos ? xpath_string::from_heap(s.c_str(), pos, stack.result) : xpath_string(); + } + + case ast_func_substring_after: + { + xpath_allocator_capture cr(stack.temp); + + xpath_stack swapped_stack = {stack.temp, stack.result}; + + xpath_string s = _left->eval_string(c, swapped_stack); + xpath_string p = _right->eval_string(c, swapped_stack); + + const char_t* pos = find_substring(s.c_str(), p.c_str()); + if (!pos) + return xpath_string(); + + const char_t* rbegin = pos + p.length(); + const char_t* rend = s.c_str() + s.length(); + + return s.uses_heap() ? xpath_string::from_heap(rbegin, rend, stack.result) : xpath_string::from_const(rbegin); + } + + case ast_func_substring_2: + { + xpath_allocator_capture cr(stack.temp); + + xpath_stack swapped_stack = {stack.temp, stack.result}; + + xpath_string s = _left->eval_string(c, swapped_stack); + size_t s_length = s.length(); + + double first = round_nearest(_right->eval_number(c, stack)); + + if (is_nan(first)) + return xpath_string(); // NaN + else if (first >= s_length + 1) + return xpath_string(); + + size_t pos = first < 1 ? 1 : static_cast(first); + assert(1 <= pos && pos <= s_length + 1); + + const char_t* rbegin = s.c_str() + (pos - 1); + const char_t* rend = s.c_str() + s.length(); + + return s.uses_heap() ? xpath_string::from_heap(rbegin, rend, stack.result) : xpath_string::from_const(rbegin); + } + + case ast_func_substring_3: + { + xpath_allocator_capture cr(stack.temp); + + xpath_stack swapped_stack = {stack.temp, stack.result}; + + xpath_string s = _left->eval_string(c, swapped_stack); + size_t s_length = s.length(); + + double first = round_nearest(_right->eval_number(c, stack)); + double last = first + round_nearest(_right->_next->eval_number(c, stack)); + + if (is_nan(first) || is_nan(last)) + return xpath_string(); + else if (first >= s_length + 1) + return xpath_string(); + else if (first >= last) + return xpath_string(); + else if (last < 1) + return xpath_string(); + + size_t pos = first < 1 ? 1 : static_cast(first); + size_t end = last >= s_length + 1 ? s_length + 1 : static_cast(last); + + assert(1 <= pos && pos <= end && end <= s_length + 1); + const char_t* rbegin = s.c_str() + (pos - 1); + const char_t* rend = s.c_str() + (end - 1); + + return (end == s_length + 1 && !s.uses_heap()) ? xpath_string::from_const(rbegin) : xpath_string::from_heap(rbegin, rend, stack.result); + } + + case ast_func_normalize_space_0: + { + xpath_string s = string_value(c.n, stack.result); + + char_t* begin = s.data(stack.result); + if (!begin) + return xpath_string(); + + char_t* end = normalize_space(begin); + + return xpath_string::from_heap_preallocated(begin, end); + } + + case ast_func_normalize_space_1: + { + xpath_string s = _left->eval_string(c, stack); + + char_t* begin = s.data(stack.result); + if (!begin) + return xpath_string(); + + char_t* end = normalize_space(begin); + + return xpath_string::from_heap_preallocated(begin, end); + } + + case ast_func_translate: + { + xpath_allocator_capture cr(stack.temp); + + xpath_stack swapped_stack = {stack.temp, stack.result}; + + xpath_string s = _left->eval_string(c, stack); + xpath_string from = _right->eval_string(c, swapped_stack); + xpath_string to = _right->_next->eval_string(c, swapped_stack); + + char_t* begin = s.data(stack.result); + if (!begin) + return xpath_string(); + + char_t* end = translate(begin, from.c_str(), to.c_str(), to.length()); + + return xpath_string::from_heap_preallocated(begin, end); + } + + case ast_opt_translate_table: + { + xpath_string s = _left->eval_string(c, stack); + + char_t* begin = s.data(stack.result); + if (!begin) + return xpath_string(); + + char_t* end = translate_table(begin, _data.table); + + return xpath_string::from_heap_preallocated(begin, end); + } + + case ast_variable: + { + assert(_rettype == _data.variable->type()); + + if (_rettype == xpath_type_string) + return xpath_string::from_const(_data.variable->get_string()); + } + + // fallthrough + default: + { + switch (_rettype) { - case nodetest_name: - if (type == node_element && n->name && strequal(n->name, _data.nodetest)) - { - ns.push_back(xml_node(n), alloc); - return true; - } - break; + case xpath_type_boolean: + return xpath_string::from_const(eval_boolean(c, stack) ? PUGIXML_TEXT("true") : PUGIXML_TEXT("false")); - case nodetest_type_node: - ns.push_back(xml_node(n), alloc); - return true; + case xpath_type_number: + return convert_number_to_string(eval_number(c, stack), stack.result); - case nodetest_type_comment: - if (type == node_comment) - { - ns.push_back(xml_node(n), alloc); - return true; - } - break; + case xpath_type_node_set: + { + xpath_allocator_capture cr(stack.temp); - case nodetest_type_text: - if (type == node_pcdata || type == node_cdata) - { - ns.push_back(xml_node(n), alloc); - return true; - } - break; + xpath_stack swapped_stack = {stack.temp, stack.result}; - case nodetest_type_pi: - if (type == node_pi) - { - ns.push_back(xml_node(n), alloc); - return true; - } - break; + xpath_node_set_raw ns = eval_node_set(c, swapped_stack, nodeset_eval_first); + return ns.empty() ? xpath_string() : string_value(ns.first(), stack.result); + } - case nodetest_pi: - if (type == node_pi && n->name && strequal(n->name, _data.nodetest)) - { - ns.push_back(xml_node(n), alloc); - return true; - } - break; + default: + assert(false && "Wrong expression for return type string"); // unreachable + return xpath_string(); + } + } + } + } - case nodetest_all: - if (type == node_element) - { - ns.push_back(xml_node(n), alloc); - return true; - } - break; + xpath_node_set_raw eval_node_set(const xpath_context& c, const xpath_stack& stack, nodeset_eval_t eval) + { + switch (_type) + { + case ast_op_union: + { + xpath_allocator_capture cr(stack.temp); - case nodetest_all_in_namespace: - if (type == node_element && n->name && starts_with(n->name, _data.nodetest)) - { - ns.push_back(xml_node(n), alloc); - return true; - } - break; + xpath_stack swapped_stack = {stack.temp, stack.result}; + + xpath_node_set_raw ls = _left->eval_node_set(c, swapped_stack, eval); + xpath_node_set_raw rs = _right->eval_node_set(c, stack, eval); + + // we can optimize merging two sorted sets, but this is a very rare operation, so don't bother + rs.set_type(xpath_node_set::type_unsorted); + + rs.append(ls.begin(), ls.end(), stack.result); + rs.remove_duplicates(); + + return rs; + } + + case ast_filter: + { + xpath_node_set_raw set = _left->eval_node_set(c, stack, _test == predicate_constant_one ? nodeset_eval_first : nodeset_eval_all); + + // either expression is a number or it contains position() call; sort by document order + if (_test != predicate_posinv) + set.sort_do(); + + bool once = eval_once(set.type(), eval); + + apply_predicate(set, 0, stack, once); + + return set; + } + + case ast_func_id: + return xpath_node_set_raw(); + + case ast_step: + { + switch (_axis) + { + case axis_ancestor: + return step_do(c, stack, eval, axis_to_type()); + + case axis_ancestor_or_self: + return step_do(c, stack, eval, axis_to_type()); + + case axis_attribute: + return step_do(c, stack, eval, axis_to_type()); + + case axis_child: + return step_do(c, stack, eval, axis_to_type()); + + case axis_descendant: + return step_do(c, stack, eval, axis_to_type()); + + case axis_descendant_or_self: + return step_do(c, stack, eval, axis_to_type()); + + case axis_following: + return step_do(c, stack, eval, axis_to_type()); + + case axis_following_sibling: + return step_do(c, stack, eval, axis_to_type()); + + case axis_namespace: + // namespaced axis is not supported + return xpath_node_set_raw(); + + case axis_parent: + return step_do(c, stack, eval, axis_to_type()); + + case axis_preceding: + return step_do(c, stack, eval, axis_to_type()); + + case axis_preceding_sibling: + return step_do(c, stack, eval, axis_to_type()); + + case axis_self: + return step_do(c, stack, eval, axis_to_type()); default: assert(false && "Unknown axis"); // unreachable - } - - return false; - } - - template void step_fill(xpath_node_set_raw& ns, xml_node_struct* n, xpath_allocator* alloc, bool once, T) - { - const axis_t axis = T::axis; - - switch (axis) - { - case axis_attribute: - { - for (xml_attribute_struct* a = n->first_attribute; a; a = a->next_attribute) - if (step_push(ns, a, n, alloc) & once) - return; - - break; - } - - case axis_child: - { - for (xml_node_struct* c = n->first_child; c; c = c->next_sibling) - if (step_push(ns, c, alloc) & once) - return; - - break; - } - - case axis_descendant: - case axis_descendant_or_self: - { - if (axis == axis_descendant_or_self) - if (step_push(ns, n, alloc) & once) - return; - - xml_node_struct* cur = n->first_child; - - while (cur) - { - if (step_push(ns, cur, alloc) & once) - return; - - if (cur->first_child) - cur = cur->first_child; - else - { - while (!cur->next_sibling) - { - cur = cur->parent; - - if (cur == n) return; - } - - cur = cur->next_sibling; - } - } - - break; - } - - case axis_following_sibling: - { - for (xml_node_struct* c = n->next_sibling; c; c = c->next_sibling) - if (step_push(ns, c, alloc) & once) - return; - - break; - } - - case axis_preceding_sibling: - { - for (xml_node_struct* c = n->prev_sibling_c; c->next_sibling; c = c->prev_sibling_c) - if (step_push(ns, c, alloc) & once) - return; - - break; - } - - case axis_following: - { - xml_node_struct* cur = n; - - // exit from this node so that we don't include descendants - while (!cur->next_sibling) - { - cur = cur->parent; - - if (!cur) return; - } - - cur = cur->next_sibling; - - while (cur) - { - if (step_push(ns, cur, alloc) & once) - return; - - if (cur->first_child) - cur = cur->first_child; - else - { - while (!cur->next_sibling) - { - cur = cur->parent; - - if (!cur) return; - } - - cur = cur->next_sibling; - } - } - - break; - } - - case axis_preceding: - { - xml_node_struct* cur = n; - - // exit from this node so that we don't include descendants - while (!cur->prev_sibling_c->next_sibling) - { - cur = cur->parent; - - if (!cur) return; - } - - cur = cur->prev_sibling_c; - - while (cur) - { - if (cur->first_child) - cur = cur->first_child->prev_sibling_c; - else - { - // leaf node, can't be ancestor - if (step_push(ns, cur, alloc) & once) - return; - - while (!cur->prev_sibling_c->next_sibling) - { - cur = cur->parent; - - if (!cur) return; - - if (!node_is_ancestor(cur, n)) - if (step_push(ns, cur, alloc) & once) - return; - } - - cur = cur->prev_sibling_c; - } - } - - break; - } - - case axis_ancestor: - case axis_ancestor_or_self: - { - if (axis == axis_ancestor_or_self) - if (step_push(ns, n, alloc) & once) - return; - - xml_node_struct* cur = n->parent; - - while (cur) - { - if (step_push(ns, cur, alloc) & once) - return; - - cur = cur->parent; - } - - break; - } - - case axis_self: - { - step_push(ns, n, alloc); - - break; - } - - case axis_parent: - { - if (n->parent) - step_push(ns, n->parent, alloc); - - break; - } - - default: - assert(false && "Unimplemented axis"); // unreachable + return xpath_node_set_raw(); } } - template void step_fill(xpath_node_set_raw& ns, xml_attribute_struct* a, xml_node_struct* p, xpath_allocator* alloc, bool once, T v) + case ast_step_root: { - const axis_t axis = T::axis; - - switch (axis) - { - case axis_ancestor: - case axis_ancestor_or_self: - { - if (axis == axis_ancestor_or_self && _test == nodetest_type_node) // reject attributes based on principal node type test - if (step_push(ns, a, p, alloc) & once) - return; - - xml_node_struct* cur = p; - - while (cur) - { - if (step_push(ns, cur, alloc) & once) - return; - - cur = cur->parent; - } - - break; - } - - case axis_descendant_or_self: - case axis_self: - { - if (_test == nodetest_type_node) // reject attributes based on principal node type test - step_push(ns, a, p, alloc); - - break; - } - - case axis_following: - { - xml_node_struct* cur = p; - - while (cur) - { - if (cur->first_child) - cur = cur->first_child; - else - { - while (!cur->next_sibling) - { - cur = cur->parent; - - if (!cur) return; - } - - cur = cur->next_sibling; - } - - if (step_push(ns, cur, alloc) & once) - return; - } - - break; - } - - case axis_parent: - { - step_push(ns, p, alloc); - - break; - } - - case axis_preceding: - { - // preceding:: axis does not include attribute nodes and attribute ancestors (they are the same as parent's ancestors), so we can reuse node preceding - step_fill(ns, p, alloc, once, v); - break; - } - - default: - assert(false && "Unimplemented axis"); // unreachable - } - } - - template void step_fill(xpath_node_set_raw& ns, const xpath_node& xn, xpath_allocator* alloc, bool once, T v) - { - const axis_t axis = T::axis; - const bool axis_has_attributes = (axis == axis_ancestor || axis == axis_ancestor_or_self || axis == axis_descendant_or_self || axis == axis_following || axis == axis_parent || axis == axis_preceding || axis == axis_self); - - if (xn.node()) - step_fill(ns, xn.node().internal_object(), alloc, once, v); - else if (axis_has_attributes && xn.attribute() && xn.parent()) - step_fill(ns, xn.attribute().internal_object(), xn.parent().internal_object(), alloc, once, v); - } - - template xpath_node_set_raw step_do(const xpath_context& c, const xpath_stack& stack, nodeset_eval_t eval, T v) - { - const axis_t axis = T::axis; - const bool axis_reverse = (axis == axis_ancestor || axis == axis_ancestor_or_self || axis == axis_preceding || axis == axis_preceding_sibling); - const xpath_node_set::type_t axis_type = axis_reverse ? xpath_node_set::type_sorted_reverse : xpath_node_set::type_sorted; - - bool once = - (axis == axis_attribute && _test == nodetest_name) || - (!_right && eval_once(axis_type, eval)) || - (_right && !_right->_next && _right->_test == predicate_constant_one); + assert(!_right); // root step can't have any predicates xpath_node_set_raw ns; - ns.set_type(axis_type); - if (_left) - { - xpath_node_set_raw s = _left->eval_node_set(c, stack, nodeset_eval_all); + ns.set_type(xpath_node_set::type_sorted); - // self axis preserves the original order - if (axis == axis_self) ns.set_type(s.type()); - - for (const xpath_node* it = s.begin(); it != s.end(); ++it) - { - size_t size = ns.size(); - - // in general, all axes generate elements in a particular order, but there is no order guarantee if axis is applied to two nodes - if (axis != axis_self && size != 0) ns.set_type(xpath_node_set::type_unsorted); - - step_fill(ns, *it, stack.result, once, v); - if (_right) apply_predicates(ns, size, stack, eval); - } - } - else - { - step_fill(ns, c.n, stack.result, once, v); - if (_right) apply_predicates(ns, 0, stack, eval); - } - - // child, attribute and self axes always generate unique set of nodes - // for other axis, if the set stayed sorted, it stayed unique because the traversal algorithms do not visit the same node twice - if (axis != axis_child && axis != axis_attribute && axis != axis_self && ns.type() == xpath_node_set::type_unsorted) - ns.remove_duplicates(); + if (c.n.node()) + ns.push_back(c.n.node().root(), stack.result); + else if (c.n.attribute()) + ns.push_back(c.n.parent().root(), stack.result); return ns; } - public: - xpath_ast_node(ast_type_t type, xpath_value_type rettype_, const char_t* value): - _type(static_cast(type)), _rettype(static_cast(rettype_)), _axis(0), _test(0), _left(0), _right(0), _next(0) + case ast_variable: { - assert(type == ast_string_constant); - _data.string = value; - } + assert(_rettype == _data.variable->type()); - xpath_ast_node(ast_type_t type, xpath_value_type rettype_, double value): - _type(static_cast(type)), _rettype(static_cast(rettype_)), _axis(0), _test(0), _left(0), _right(0), _next(0) - { - assert(type == ast_number_constant); - _data.number = value; - } - - xpath_ast_node(ast_type_t type, xpath_value_type rettype_, xpath_variable* value): - _type(static_cast(type)), _rettype(static_cast(rettype_)), _axis(0), _test(0), _left(0), _right(0), _next(0) - { - assert(type == ast_variable); - _data.variable = value; - } - - xpath_ast_node(ast_type_t type, xpath_value_type rettype_, xpath_ast_node* left = 0, xpath_ast_node* right = 0): - _type(static_cast(type)), _rettype(static_cast(rettype_)), _axis(0), _test(0), _left(left), _right(right), _next(0) - { - } - - xpath_ast_node(ast_type_t type, xpath_ast_node* left, axis_t axis, nodetest_t test, const char_t* contents): - _type(static_cast(type)), _rettype(xpath_type_node_set), _axis(static_cast(axis)), _test(static_cast(test)), _left(left), _right(0), _next(0) - { - assert(type == ast_step); - _data.nodetest = contents; - } - - xpath_ast_node(ast_type_t type, xpath_ast_node* left, xpath_ast_node* right, predicate_t test): - _type(static_cast(type)), _rettype(xpath_type_node_set), _axis(0), _test(static_cast(test)), _left(left), _right(right), _next(0) - { - assert(type == ast_filter || type == ast_predicate); - } - - void set_next(xpath_ast_node* value) - { - _next = value; - } - - void set_right(xpath_ast_node* value) - { - _right = value; - } - - bool eval_boolean(const xpath_context& c, const xpath_stack& stack) - { - switch (_type) + if (_rettype == xpath_type_node_set) { - case ast_op_or: - return _left->eval_boolean(c, stack) || _right->eval_boolean(c, stack); - - case ast_op_and: - return _left->eval_boolean(c, stack) && _right->eval_boolean(c, stack); - - case ast_op_equal: - return compare_eq(_left, _right, c, stack, equal_to()); - - case ast_op_not_equal: - return compare_eq(_left, _right, c, stack, not_equal_to()); - - case ast_op_less: - return compare_rel(_left, _right, c, stack, less()); - - case ast_op_greater: - return compare_rel(_right, _left, c, stack, less()); - - case ast_op_less_or_equal: - return compare_rel(_left, _right, c, stack, less_equal()); - - case ast_op_greater_or_equal: - return compare_rel(_right, _left, c, stack, less_equal()); - - case ast_func_starts_with: - { - xpath_allocator_capture cr(stack.result); - - xpath_string lr = _left->eval_string(c, stack); - xpath_string rr = _right->eval_string(c, stack); - - return starts_with(lr.c_str(), rr.c_str()); - } - - case ast_func_contains: - { - xpath_allocator_capture cr(stack.result); - - xpath_string lr = _left->eval_string(c, stack); - xpath_string rr = _right->eval_string(c, stack); - - return find_substring(lr.c_str(), rr.c_str()) != 0; - } - - case ast_func_boolean: - return _left->eval_boolean(c, stack); - - case ast_func_not: - return !_left->eval_boolean(c, stack); - - case ast_func_true: - return true; - - case ast_func_false: - return false; - - case ast_func_lang: - { - if (c.n.attribute()) return false; - - xpath_allocator_capture cr(stack.result); - - xpath_string lang = _left->eval_string(c, stack); - - for (xml_node n = c.n.node(); n; n = n.parent()) - { - xml_attribute a = n.attribute(PUGIXML_TEXT("xml:lang")); - - if (a) - { - const char_t* value = a.value(); - - // strnicmp / strncasecmp is not portable - for (const char_t* lit = lang.c_str(); *lit; ++lit) - { - if (tolower_ascii(*lit) != tolower_ascii(*value)) return false; - ++value; - } - - return *value == 0 || *value == '-'; - } - } - - return false; - } - - case ast_opt_compare_attribute: - { - const char_t* value = (_right->_type == ast_string_constant) ? _right->_data.string : _right->_data.variable->get_string(); - - xml_attribute attr = c.n.node().attribute(_left->_data.nodetest); - - return attr && strequal(attr.value(), value) && is_xpath_attribute(attr.name()); - } - - case ast_variable: - { - assert(_rettype == _data.variable->type()); - - if (_rettype == xpath_type_boolean) - return _data.variable->get_boolean(); - } - - // fallthrough - default: - { - switch (_rettype) - { - case xpath_type_number: - return convert_number_to_boolean(eval_number(c, stack)); - - case xpath_type_string: - { - xpath_allocator_capture cr(stack.result); - - return !eval_string(c, stack).empty(); - } - - case xpath_type_node_set: - { - xpath_allocator_capture cr(stack.result); - - return !eval_node_set(c, stack, nodeset_eval_any).empty(); - } - - default: - assert(false && "Wrong expression for return type boolean"); // unreachable - return false; - } - } - } - } - - double eval_number(const xpath_context& c, const xpath_stack& stack) - { - switch (_type) - { - case ast_op_add: - return _left->eval_number(c, stack) + _right->eval_number(c, stack); - - case ast_op_subtract: - return _left->eval_number(c, stack) - _right->eval_number(c, stack); - - case ast_op_multiply: - return _left->eval_number(c, stack) * _right->eval_number(c, stack); - - case ast_op_divide: - return _left->eval_number(c, stack) / _right->eval_number(c, stack); - - case ast_op_mod: - return fmod(_left->eval_number(c, stack), _right->eval_number(c, stack)); - - case ast_op_negate: - return -_left->eval_number(c, stack); - - case ast_number_constant: - return _data.number; - - case ast_func_last: - return static_cast(c.size); - - case ast_func_position: - return static_cast(c.position); - - case ast_func_count: - { - xpath_allocator_capture cr(stack.result); - - return static_cast(_left->eval_node_set(c, stack, nodeset_eval_all).size()); - } - - case ast_func_string_length_0: - { - xpath_allocator_capture cr(stack.result); - - return static_cast(string_value(c.n, stack.result).length()); - } - - case ast_func_string_length_1: - { - xpath_allocator_capture cr(stack.result); - - return static_cast(_left->eval_string(c, stack).length()); - } - - case ast_func_number_0: - { - xpath_allocator_capture cr(stack.result); - - return convert_string_to_number(string_value(c.n, stack.result).c_str()); - } - - case ast_func_number_1: - return _left->eval_number(c, stack); - - case ast_func_sum: - { - xpath_allocator_capture cr(stack.result); - - double r = 0; - - xpath_node_set_raw ns = _left->eval_node_set(c, stack, nodeset_eval_all); - - for (const xpath_node* it = ns.begin(); it != ns.end(); ++it) - { - xpath_allocator_capture cri(stack.result); - - r += convert_string_to_number(string_value(*it, stack.result).c_str()); - } - - return r; - } - - case ast_func_floor: - { - double r = _left->eval_number(c, stack); - - return r == r ? floor(r) : r; - } - - case ast_func_ceiling: - { - double r = _left->eval_number(c, stack); - - return r == r ? ceil(r) : r; - } - - case ast_func_round: - return round_nearest_nzero(_left->eval_number(c, stack)); - - case ast_variable: - { - assert(_rettype == _data.variable->type()); - - if (_rettype == xpath_type_number) - return _data.variable->get_number(); - } - - // fallthrough - default: - { - switch (_rettype) - { - case xpath_type_boolean: - return eval_boolean(c, stack) ? 1 : 0; - - case xpath_type_string: - { - xpath_allocator_capture cr(stack.result); - - return convert_string_to_number(eval_string(c, stack).c_str()); - } - - case xpath_type_node_set: - { - xpath_allocator_capture cr(stack.result); - - return convert_string_to_number(eval_string(c, stack).c_str()); - } - - default: - assert(false && "Wrong expression for return type number"); // unreachable - return 0; - } - - } - } - } - - xpath_string eval_string_concat(const xpath_context& c, const xpath_stack& stack) - { - assert(_type == ast_func_concat); - - xpath_allocator_capture ct(stack.temp); - - // count the string number - size_t count = 1; - for (xpath_ast_node* nc = _right; nc; nc = nc->_next) count++; - - // allocate a buffer for temporary string objects - xpath_string* buffer = static_cast(stack.temp->allocate(count * sizeof(xpath_string))); - if (!buffer) return xpath_string(); - - // evaluate all strings to temporary stack - xpath_stack swapped_stack = {stack.temp, stack.result}; - - buffer[0] = _left->eval_string(c, swapped_stack); - - size_t pos = 1; - for (xpath_ast_node* n = _right; n; n = n->_next, ++pos) buffer[pos] = n->eval_string(c, swapped_stack); - assert(pos == count); - - // get total length - size_t length = 0; - for (size_t i = 0; i < count; ++i) length += buffer[i].length(); - - // create final string - char_t* result = static_cast(stack.result->allocate((length + 1) * sizeof(char_t))); - if (!result) return xpath_string(); - - char_t* ri = result; - - for (size_t j = 0; j < count; ++j) - for (const char_t* bi = buffer[j].c_str(); *bi; ++bi) - *ri++ = *bi; - - *ri = 0; - - return xpath_string::from_heap_preallocated(result, ri); - } - - xpath_string eval_string(const xpath_context& c, const xpath_stack& stack) - { - switch (_type) - { - case ast_string_constant: - return xpath_string::from_const(_data.string); - - case ast_func_local_name_0: - { - xpath_node na = c.n; - - return xpath_string::from_const(local_name(na)); - } - - case ast_func_local_name_1: - { - xpath_allocator_capture cr(stack.result); - - xpath_node_set_raw ns = _left->eval_node_set(c, stack, nodeset_eval_first); - xpath_node na = ns.first(); - - return xpath_string::from_const(local_name(na)); - } - - case ast_func_name_0: - { - xpath_node na = c.n; - - return xpath_string::from_const(qualified_name(na)); - } - - case ast_func_name_1: - { - xpath_allocator_capture cr(stack.result); - - xpath_node_set_raw ns = _left->eval_node_set(c, stack, nodeset_eval_first); - xpath_node na = ns.first(); - - return xpath_string::from_const(qualified_name(na)); - } - - case ast_func_namespace_uri_0: - { - xpath_node na = c.n; - - return xpath_string::from_const(namespace_uri(na)); - } - - case ast_func_namespace_uri_1: - { - xpath_allocator_capture cr(stack.result); - - xpath_node_set_raw ns = _left->eval_node_set(c, stack, nodeset_eval_first); - xpath_node na = ns.first(); - - return xpath_string::from_const(namespace_uri(na)); - } - - case ast_func_string_0: - return string_value(c.n, stack.result); - - case ast_func_string_1: - return _left->eval_string(c, stack); - - case ast_func_concat: - return eval_string_concat(c, stack); - - case ast_func_substring_before: - { - xpath_allocator_capture cr(stack.temp); - - xpath_stack swapped_stack = {stack.temp, stack.result}; - - xpath_string s = _left->eval_string(c, swapped_stack); - xpath_string p = _right->eval_string(c, swapped_stack); - - const char_t* pos = find_substring(s.c_str(), p.c_str()); - - return pos ? xpath_string::from_heap(s.c_str(), pos, stack.result) : xpath_string(); - } - - case ast_func_substring_after: - { - xpath_allocator_capture cr(stack.temp); - - xpath_stack swapped_stack = {stack.temp, stack.result}; - - xpath_string s = _left->eval_string(c, swapped_stack); - xpath_string p = _right->eval_string(c, swapped_stack); - - const char_t* pos = find_substring(s.c_str(), p.c_str()); - if (!pos) return xpath_string(); - - const char_t* rbegin = pos + p.length(); - const char_t* rend = s.c_str() + s.length(); - - return s.uses_heap() ? xpath_string::from_heap(rbegin, rend, stack.result) : xpath_string::from_const(rbegin); - } - - case ast_func_substring_2: - { - xpath_allocator_capture cr(stack.temp); - - xpath_stack swapped_stack = {stack.temp, stack.result}; - - xpath_string s = _left->eval_string(c, swapped_stack); - size_t s_length = s.length(); - - double first = round_nearest(_right->eval_number(c, stack)); - - if (is_nan(first)) return xpath_string(); // NaN - else if (first >= s_length + 1) return xpath_string(); - - size_t pos = first < 1 ? 1 : static_cast(first); - assert(1 <= pos && pos <= s_length + 1); - - const char_t* rbegin = s.c_str() + (pos - 1); - const char_t* rend = s.c_str() + s.length(); - - return s.uses_heap() ? xpath_string::from_heap(rbegin, rend, stack.result) : xpath_string::from_const(rbegin); - } - - case ast_func_substring_3: - { - xpath_allocator_capture cr(stack.temp); - - xpath_stack swapped_stack = {stack.temp, stack.result}; - - xpath_string s = _left->eval_string(c, swapped_stack); - size_t s_length = s.length(); - - double first = round_nearest(_right->eval_number(c, stack)); - double last = first + round_nearest(_right->_next->eval_number(c, stack)); - - if (is_nan(first) || is_nan(last)) return xpath_string(); - else if (first >= s_length + 1) return xpath_string(); - else if (first >= last) return xpath_string(); - else if (last < 1) return xpath_string(); - - size_t pos = first < 1 ? 1 : static_cast(first); - size_t end = last >= s_length + 1 ? s_length + 1 : static_cast(last); - - assert(1 <= pos && pos <= end && end <= s_length + 1); - const char_t* rbegin = s.c_str() + (pos - 1); - const char_t* rend = s.c_str() + (end - 1); - - return (end == s_length + 1 && !s.uses_heap()) ? xpath_string::from_const(rbegin) : xpath_string::from_heap(rbegin, rend, stack.result); - } - - case ast_func_normalize_space_0: - { - xpath_string s = string_value(c.n, stack.result); - - char_t* begin = s.data(stack.result); - if (!begin) return xpath_string(); - - char_t* end = normalize_space(begin); - - return xpath_string::from_heap_preallocated(begin, end); - } - - case ast_func_normalize_space_1: - { - xpath_string s = _left->eval_string(c, stack); - - char_t* begin = s.data(stack.result); - if (!begin) return xpath_string(); - - char_t* end = normalize_space(begin); - - return xpath_string::from_heap_preallocated(begin, end); - } - - case ast_func_translate: - { - xpath_allocator_capture cr(stack.temp); - - xpath_stack swapped_stack = {stack.temp, stack.result}; - - xpath_string s = _left->eval_string(c, stack); - xpath_string from = _right->eval_string(c, swapped_stack); - xpath_string to = _right->_next->eval_string(c, swapped_stack); - - char_t* begin = s.data(stack.result); - if (!begin) return xpath_string(); - - char_t* end = translate(begin, from.c_str(), to.c_str(), to.length()); - - return xpath_string::from_heap_preallocated(begin, end); - } - - case ast_opt_translate_table: - { - xpath_string s = _left->eval_string(c, stack); - - char_t* begin = s.data(stack.result); - if (!begin) return xpath_string(); - - char_t* end = translate_table(begin, _data.table); - - return xpath_string::from_heap_preallocated(begin, end); - } - - case ast_variable: - { - assert(_rettype == _data.variable->type()); - - if (_rettype == xpath_type_string) - return xpath_string::from_const(_data.variable->get_string()); - } - - // fallthrough - default: - { - switch (_rettype) - { - case xpath_type_boolean: - return xpath_string::from_const(eval_boolean(c, stack) ? PUGIXML_TEXT("true") : PUGIXML_TEXT("false")); - - case xpath_type_number: - return convert_number_to_string(eval_number(c, stack), stack.result); - - case xpath_type_node_set: - { - xpath_allocator_capture cr(stack.temp); - - xpath_stack swapped_stack = {stack.temp, stack.result}; - - xpath_node_set_raw ns = eval_node_set(c, swapped_stack, nodeset_eval_first); - return ns.empty() ? xpath_string() : string_value(ns.first(), stack.result); - } - - default: - assert(false && "Wrong expression for return type string"); // unreachable - return xpath_string(); - } - } - } - } - - xpath_node_set_raw eval_node_set(const xpath_context& c, const xpath_stack& stack, nodeset_eval_t eval) - { - switch (_type) - { - case ast_op_union: - { - xpath_allocator_capture cr(stack.temp); - - xpath_stack swapped_stack = {stack.temp, stack.result}; - - xpath_node_set_raw ls = _left->eval_node_set(c, swapped_stack, eval); - xpath_node_set_raw rs = _right->eval_node_set(c, stack, eval); - - // we can optimize merging two sorted sets, but this is a very rare operation, so don't bother - rs.set_type(xpath_node_set::type_unsorted); - - rs.append(ls.begin(), ls.end(), stack.result); - rs.remove_duplicates(); - - return rs; - } - - case ast_filter: - { - xpath_node_set_raw set = _left->eval_node_set(c, stack, _test == predicate_constant_one ? nodeset_eval_first : nodeset_eval_all); - - // either expression is a number or it contains position() call; sort by document order - if (_test != predicate_posinv) set.sort_do(); - - bool once = eval_once(set.type(), eval); - - apply_predicate(set, 0, stack, once); - - return set; - } - - case ast_func_id: - return xpath_node_set_raw(); - - case ast_step: - { - switch (_axis) - { - case axis_ancestor: - return step_do(c, stack, eval, axis_to_type()); - - case axis_ancestor_or_self: - return step_do(c, stack, eval, axis_to_type()); - - case axis_attribute: - return step_do(c, stack, eval, axis_to_type()); - - case axis_child: - return step_do(c, stack, eval, axis_to_type()); - - case axis_descendant: - return step_do(c, stack, eval, axis_to_type()); - - case axis_descendant_or_self: - return step_do(c, stack, eval, axis_to_type()); - - case axis_following: - return step_do(c, stack, eval, axis_to_type()); - - case axis_following_sibling: - return step_do(c, stack, eval, axis_to_type()); - - case axis_namespace: - // namespaced axis is not supported - return xpath_node_set_raw(); - - case axis_parent: - return step_do(c, stack, eval, axis_to_type()); - - case axis_preceding: - return step_do(c, stack, eval, axis_to_type()); - - case axis_preceding_sibling: - return step_do(c, stack, eval, axis_to_type()); - - case axis_self: - return step_do(c, stack, eval, axis_to_type()); - - default: - assert(false && "Unknown axis"); // unreachable - return xpath_node_set_raw(); - } - } - - case ast_step_root: - { - assert(!_right); // root step can't have any predicates + const xpath_node_set& s = _data.variable->get_node_set(); xpath_node_set_raw ns; - ns.set_type(xpath_node_set::type_sorted); - - if (c.n.node()) ns.push_back(c.n.node().root(), stack.result); - else if (c.n.attribute()) ns.push_back(c.n.parent().root(), stack.result); + ns.set_type(s.type()); + ns.append(s.begin(), s.end(), stack.result); return ns; } + } - case ast_variable: + // fallthrough + default: + assert(false && "Wrong expression for return type node set"); // unreachable + return xpath_node_set_raw(); + } + } + + void optimize(xpath_allocator* alloc) + { + if (_left) + _left->optimize(alloc); + + if (_right) + _right->optimize(alloc); + + if (_next) + _next->optimize(alloc); + + optimize_self(alloc); + } + + void optimize_self(xpath_allocator* alloc) + { + // Rewrite [position()=expr] with [expr] + // Note that this step has to go before classification to recognize [position()=1] + if ((_type == ast_filter || _type == ast_predicate) && + _right && // workaround for clang static analyzer (_right is never null for ast_filter/ast_predicate) + _right->_type == ast_op_equal && _right->_left->_type == ast_func_position && _right->_right->_rettype == xpath_type_number) + { + _right = _right->_right; + } + + // Classify filter/predicate ops to perform various optimizations during evaluation + if ((_type == ast_filter || _type == ast_predicate) && _right) // workaround for clang static analyzer (_right is never null for ast_filter/ast_predicate) + { + assert(_test == predicate_default); + + if (_right->_type == ast_number_constant && _right->_data.number == 1.0) + _test = predicate_constant_one; + else if (_right->_rettype == xpath_type_number && (_right->_type == ast_number_constant || _right->_type == ast_variable || _right->_type == ast_func_last)) + _test = predicate_constant; + else if (_right->_rettype != xpath_type_number && _right->is_posinv_expr()) + _test = predicate_posinv; + } + + // Rewrite descendant-or-self::node()/child::foo with descendant::foo + // The former is a full form of //foo, the latter is much faster since it executes the node test immediately + // Do a similar kind of rewrite for self/descendant/descendant-or-self axes + // Note that we only rewrite positionally invariant steps (//foo[1] != /descendant::foo[1]) + if (_type == ast_step && (_axis == axis_child || _axis == axis_self || _axis == axis_descendant || _axis == axis_descendant_or_self) && + _left && _left->_type == ast_step && _left->_axis == axis_descendant_or_self && _left->_test == nodetest_type_node && !_left->_right && + is_posinv_step()) + { + if (_axis == axis_child || _axis == axis_descendant) + _axis = axis_descendant; + else + _axis = axis_descendant_or_self; + + _left = _left->_left; + } + + // Use optimized lookup table implementation for translate() with constant arguments + if (_type == ast_func_translate && + _right && // workaround for clang static analyzer (_right is never null for ast_func_translate) + _right->_type == ast_string_constant && _right->_next->_type == ast_string_constant) + { + unsigned char* table = translate_table_generate(alloc, _right->_data.string, _right->_next->_data.string); + + if (table) { - assert(_rettype == _data.variable->type()); - - if (_rettype == xpath_type_node_set) - { - const xpath_node_set& s = _data.variable->get_node_set(); - - xpath_node_set_raw ns; - - ns.set_type(s.type()); - ns.append(s.begin(), s.end(), stack.result); - - return ns; - } - } - - // fallthrough - default: - assert(false && "Wrong expression for return type node set"); // unreachable - return xpath_node_set_raw(); + _type = ast_opt_translate_table; + _data.table = table; } } - void optimize(xpath_allocator* alloc) + // Use optimized path for @attr = 'value' or @attr = $value + if (_type == ast_op_equal && + _left && // workaround for clang static analyzer (_left is never null for ast_op_equal) + _left->_type == ast_step && _left->_axis == axis_attribute && _left->_test == nodetest_name && !_left->_left && !_left->_right && + (_right->_type == ast_string_constant || (_right->_type == ast_variable && _right->_rettype == xpath_type_string))) { - if (_left) - _left->optimize(alloc); - - if (_right) - _right->optimize(alloc); - - if (_next) - _next->optimize(alloc); - - optimize_self(alloc); + _type = ast_opt_compare_attribute; } + } - void optimize_self(xpath_allocator* alloc) + bool is_posinv_expr() const + { + switch (_type) { - // Rewrite [position()=expr] with [expr] - // Note that this step has to go before classification to recognize [position()=1] - if ((_type == ast_filter || _type == ast_predicate) && - _right && // workaround for clang static analyzer (_right is never null for ast_filter/ast_predicate) - _right->_type == ast_op_equal && _right->_left->_type == ast_func_position && _right->_right->_rettype == xpath_type_number) - { - _right = _right->_right; - } + case ast_func_position: + case ast_func_last: + return false; - // Classify filter/predicate ops to perform various optimizations during evaluation - if ((_type == ast_filter || _type == ast_predicate) && _right) // workaround for clang static analyzer (_right is never null for ast_filter/ast_predicate) - { - assert(_test == predicate_default); + case ast_string_constant: + case ast_number_constant: + case ast_variable: + return true; - if (_right->_type == ast_number_constant && _right->_data.number == 1.0) - _test = predicate_constant_one; - else if (_right->_rettype == xpath_type_number && (_right->_type == ast_number_constant || _right->_type == ast_variable || _right->_type == ast_func_last)) - _test = predicate_constant; - else if (_right->_rettype != xpath_type_number && _right->is_posinv_expr()) - _test = predicate_posinv; - } + case ast_step: + case ast_step_root: + return true; - // Rewrite descendant-or-self::node()/child::foo with descendant::foo - // The former is a full form of //foo, the latter is much faster since it executes the node test immediately - // Do a similar kind of rewrite for self/descendant/descendant-or-self axes - // Note that we only rewrite positionally invariant steps (//foo[1] != /descendant::foo[1]) - if (_type == ast_step && (_axis == axis_child || _axis == axis_self || _axis == axis_descendant || _axis == axis_descendant_or_self) && - _left && _left->_type == ast_step && _left->_axis == axis_descendant_or_self && _left->_test == nodetest_type_node && !_left->_right && - is_posinv_step()) - { - if (_axis == axis_child || _axis == axis_descendant) - _axis = axis_descendant; - else - _axis = axis_descendant_or_self; + case ast_predicate: + case ast_filter: + return true; - _left = _left->_left; - } - - // Use optimized lookup table implementation for translate() with constant arguments - if (_type == ast_func_translate && - _right && // workaround for clang static analyzer (_right is never null for ast_func_translate) - _right->_type == ast_string_constant && _right->_next->_type == ast_string_constant) - { - unsigned char* table = translate_table_generate(alloc, _right->_data.string, _right->_next->_data.string); - - if (table) - { - _type = ast_opt_translate_table; - _data.table = table; - } - } - - // Use optimized path for @attr = 'value' or @attr = $value - if (_type == ast_op_equal && - _left && // workaround for clang static analyzer (_left is never null for ast_op_equal) - _left->_type == ast_step && _left->_axis == axis_attribute && _left->_test == nodetest_name && !_left->_left && !_left->_right && - (_right->_type == ast_string_constant || (_right->_type == ast_variable && _right->_rettype == xpath_type_string))) - { - _type = ast_opt_compare_attribute; - } - } - - bool is_posinv_expr() const - { - switch (_type) - { - case ast_func_position: - case ast_func_last: + default: + if (_left && !_left->is_posinv_expr()) return false; - case ast_string_constant: - case ast_number_constant: - case ast_variable: - return true; - - case ast_step: - case ast_step_root: - return true; - - case ast_predicate: - case ast_filter: - return true; - - default: - if (_left && !_left->is_posinv_expr()) return false; - - for (xpath_ast_node* n = _right; n; n = n->_next) - if (!n->is_posinv_expr()) return false; - - return true; - } - } - - bool is_posinv_step() const - { - assert(_type == ast_step); - for (xpath_ast_node* n = _right; n; n = n->_next) - { - assert(n->_type == ast_predicate); - - if (n->_test != predicate_posinv) + if (!n->is_posinv_expr()) return false; - } return true; } + } - xpath_value_type rettype() const - { - return static_cast(_rettype); - } - }; - - struct xpath_parser + bool is_posinv_step() const { - xpath_allocator* _alloc; - xpath_lexer _lexer; + assert(_type == ast_step); - const char_t* _query; - xpath_variable_set* _variables; - - xpath_parse_result* _result; - - char_t _scratch[32]; - - xpath_ast_node* error(const char* message) + for (xpath_ast_node* n = _right; n; n = n->_next) { - _result->error = message; - _result->offset = _lexer.current_pos() - _query; + assert(n->_type == ast_predicate); + if (n->_test != predicate_posinv) + return false; + } + + return true; + } + + xpath_value_type rettype() const + { + return static_cast(_rettype); + } +}; + +struct xpath_parser +{ + xpath_allocator* _alloc; + xpath_lexer _lexer; + + const char_t* _query; + xpath_variable_set* _variables; + + xpath_parse_result* _result; + + char_t _scratch[32]; + + xpath_ast_node* error(const char* message) + { + _result->error = message; + _result->offset = _lexer.current_pos() - _query; + + return 0; + } + + xpath_ast_node* error_oom() + { + assert(_alloc->_error); + *_alloc->_error = true; + + return 0; + } + + void* alloc_node() + { + return _alloc->allocate(sizeof(xpath_ast_node)); + } + + xpath_ast_node* alloc_node(ast_type_t type, xpath_value_type rettype, const char_t* value) + { + void* memory = alloc_node(); + return memory ? new (memory) xpath_ast_node(type, rettype, value) : 0; + } + + xpath_ast_node* alloc_node(ast_type_t type, xpath_value_type rettype, double value) + { + void* memory = alloc_node(); + return memory ? new (memory) xpath_ast_node(type, rettype, value) : 0; + } + + xpath_ast_node* alloc_node(ast_type_t type, xpath_value_type rettype, xpath_variable* value) + { + void* memory = alloc_node(); + return memory ? new (memory) xpath_ast_node(type, rettype, value) : 0; + } + + xpath_ast_node* alloc_node(ast_type_t type, xpath_value_type rettype, xpath_ast_node* left = 0, xpath_ast_node* right = 0) + { + void* memory = alloc_node(); + return memory ? new (memory) xpath_ast_node(type, rettype, left, right) : 0; + } + + xpath_ast_node* alloc_node(ast_type_t type, xpath_ast_node* left, axis_t axis, nodetest_t test, const char_t* contents) + { + void* memory = alloc_node(); + return memory ? new (memory) xpath_ast_node(type, left, axis, test, contents) : 0; + } + + xpath_ast_node* alloc_node(ast_type_t type, xpath_ast_node* left, xpath_ast_node* right, predicate_t test) + { + void* memory = alloc_node(); + return memory ? new (memory) xpath_ast_node(type, left, right, test) : 0; + } + + const char_t* alloc_string(const xpath_lexer_string& value) + { + if (!value.begin) + return PUGIXML_TEXT(""); + + size_t length = static_cast(value.end - value.begin); + + char_t* c = static_cast(_alloc->allocate((length + 1) * sizeof(char_t))); + if (!c) return 0; - } - xpath_ast_node* error_oom() + memcpy(c, value.begin, length * sizeof(char_t)); + c[length] = 0; + + return c; + } + + xpath_ast_node* parse_function(const xpath_lexer_string& name, size_t argc, xpath_ast_node* args[2]) + { + switch (name.begin[0]) { - assert(_alloc->_error); - *_alloc->_error = true; + case 'b': + if (name == PUGIXML_TEXT("boolean") && argc == 1) + return alloc_node(ast_func_boolean, xpath_type_boolean, args[0]); - return 0; - } + break; - void* alloc_node() - { - return _alloc->allocate(sizeof(xpath_ast_node)); - } - - xpath_ast_node* alloc_node(ast_type_t type, xpath_value_type rettype, const char_t* value) - { - void* memory = alloc_node(); - return memory ? new (memory) xpath_ast_node(type, rettype, value) : 0; - } - - xpath_ast_node* alloc_node(ast_type_t type, xpath_value_type rettype, double value) - { - void* memory = alloc_node(); - return memory ? new (memory) xpath_ast_node(type, rettype, value) : 0; - } - - xpath_ast_node* alloc_node(ast_type_t type, xpath_value_type rettype, xpath_variable* value) - { - void* memory = alloc_node(); - return memory ? new (memory) xpath_ast_node(type, rettype, value) : 0; - } - - xpath_ast_node* alloc_node(ast_type_t type, xpath_value_type rettype, xpath_ast_node* left = 0, xpath_ast_node* right = 0) - { - void* memory = alloc_node(); - return memory ? new (memory) xpath_ast_node(type, rettype, left, right) : 0; - } - - xpath_ast_node* alloc_node(ast_type_t type, xpath_ast_node* left, axis_t axis, nodetest_t test, const char_t* contents) - { - void* memory = alloc_node(); - return memory ? new (memory) xpath_ast_node(type, left, axis, test, contents) : 0; - } - - xpath_ast_node* alloc_node(ast_type_t type, xpath_ast_node* left, xpath_ast_node* right, predicate_t test) - { - void* memory = alloc_node(); - return memory ? new (memory) xpath_ast_node(type, left, right, test) : 0; - } - - const char_t* alloc_string(const xpath_lexer_string& value) - { - if (!value.begin) - return PUGIXML_TEXT(""); - - size_t length = static_cast(value.end - value.begin); - - char_t* c = static_cast(_alloc->allocate((length + 1) * sizeof(char_t))); - if (!c) return 0; - - memcpy(c, value.begin, length * sizeof(char_t)); - c[length] = 0; - - return c; - } - - xpath_ast_node* parse_function(const xpath_lexer_string& name, size_t argc, xpath_ast_node* args[2]) - { - switch (name.begin[0]) + case 'c': + if (name == PUGIXML_TEXT("count") && argc == 1) { - case 'b': - if (name == PUGIXML_TEXT("boolean") && argc == 1) - return alloc_node(ast_func_boolean, xpath_type_boolean, args[0]); + if (args[0]->rettype() != xpath_type_node_set) + return error("Function has to be applied to node set"); + return alloc_node(ast_func_count, xpath_type_number, args[0]); + } + else if (name == PUGIXML_TEXT("contains") && argc == 2) + return alloc_node(ast_func_contains, xpath_type_boolean, args[0], args[1]); + else if (name == PUGIXML_TEXT("concat") && argc >= 2) + return alloc_node(ast_func_concat, xpath_type_string, args[0], args[1]); + else if (name == PUGIXML_TEXT("ceiling") && argc == 1) + return alloc_node(ast_func_ceiling, xpath_type_number, args[0]); - break; + break; - case 'c': - if (name == PUGIXML_TEXT("count") && argc == 1) - { - if (args[0]->rettype() != xpath_type_node_set) return error("Function has to be applied to node set"); - return alloc_node(ast_func_count, xpath_type_number, args[0]); - } - else if (name == PUGIXML_TEXT("contains") && argc == 2) - return alloc_node(ast_func_contains, xpath_type_boolean, args[0], args[1]); - else if (name == PUGIXML_TEXT("concat") && argc >= 2) - return alloc_node(ast_func_concat, xpath_type_string, args[0], args[1]); - else if (name == PUGIXML_TEXT("ceiling") && argc == 1) - return alloc_node(ast_func_ceiling, xpath_type_number, args[0]); + case 'f': + if (name == PUGIXML_TEXT("false") && argc == 0) + return alloc_node(ast_func_false, xpath_type_boolean); + else if (name == PUGIXML_TEXT("floor") && argc == 1) + return alloc_node(ast_func_floor, xpath_type_number, args[0]); - break; + break; - case 'f': - if (name == PUGIXML_TEXT("false") && argc == 0) - return alloc_node(ast_func_false, xpath_type_boolean); - else if (name == PUGIXML_TEXT("floor") && argc == 1) - return alloc_node(ast_func_floor, xpath_type_number, args[0]); + case 'i': + if (name == PUGIXML_TEXT("id") && argc == 1) + return alloc_node(ast_func_id, xpath_type_node_set, args[0]); - break; + break; - case 'i': - if (name == PUGIXML_TEXT("id") && argc == 1) - return alloc_node(ast_func_id, xpath_type_node_set, args[0]); - - break; - - case 'l': - if (name == PUGIXML_TEXT("last") && argc == 0) - return alloc_node(ast_func_last, xpath_type_number); - else if (name == PUGIXML_TEXT("lang") && argc == 1) - return alloc_node(ast_func_lang, xpath_type_boolean, args[0]); - else if (name == PUGIXML_TEXT("local-name") && argc <= 1) - { - if (argc == 1 && args[0]->rettype() != xpath_type_node_set) return error("Function has to be applied to node set"); - return alloc_node(argc == 0 ? ast_func_local_name_0 : ast_func_local_name_1, xpath_type_string, args[0]); - } - - break; - - case 'n': - if (name == PUGIXML_TEXT("name") && argc <= 1) - { - if (argc == 1 && args[0]->rettype() != xpath_type_node_set) return error("Function has to be applied to node set"); - return alloc_node(argc == 0 ? ast_func_name_0 : ast_func_name_1, xpath_type_string, args[0]); - } - else if (name == PUGIXML_TEXT("namespace-uri") && argc <= 1) - { - if (argc == 1 && args[0]->rettype() != xpath_type_node_set) return error("Function has to be applied to node set"); - return alloc_node(argc == 0 ? ast_func_namespace_uri_0 : ast_func_namespace_uri_1, xpath_type_string, args[0]); - } - else if (name == PUGIXML_TEXT("normalize-space") && argc <= 1) - return alloc_node(argc == 0 ? ast_func_normalize_space_0 : ast_func_normalize_space_1, xpath_type_string, args[0], args[1]); - else if (name == PUGIXML_TEXT("not") && argc == 1) - return alloc_node(ast_func_not, xpath_type_boolean, args[0]); - else if (name == PUGIXML_TEXT("number") && argc <= 1) - return alloc_node(argc == 0 ? ast_func_number_0 : ast_func_number_1, xpath_type_number, args[0]); - - break; - - case 'p': - if (name == PUGIXML_TEXT("position") && argc == 0) - return alloc_node(ast_func_position, xpath_type_number); - - break; - - case 'r': - if (name == PUGIXML_TEXT("round") && argc == 1) - return alloc_node(ast_func_round, xpath_type_number, args[0]); - - break; - - case 's': - if (name == PUGIXML_TEXT("string") && argc <= 1) - return alloc_node(argc == 0 ? ast_func_string_0 : ast_func_string_1, xpath_type_string, args[0]); - else if (name == PUGIXML_TEXT("string-length") && argc <= 1) - return alloc_node(argc == 0 ? ast_func_string_length_0 : ast_func_string_length_1, xpath_type_number, args[0]); - else if (name == PUGIXML_TEXT("starts-with") && argc == 2) - return alloc_node(ast_func_starts_with, xpath_type_boolean, args[0], args[1]); - else if (name == PUGIXML_TEXT("substring-before") && argc == 2) - return alloc_node(ast_func_substring_before, xpath_type_string, args[0], args[1]); - else if (name == PUGIXML_TEXT("substring-after") && argc == 2) - return alloc_node(ast_func_substring_after, xpath_type_string, args[0], args[1]); - else if (name == PUGIXML_TEXT("substring") && (argc == 2 || argc == 3)) - return alloc_node(argc == 2 ? ast_func_substring_2 : ast_func_substring_3, xpath_type_string, args[0], args[1]); - else if (name == PUGIXML_TEXT("sum") && argc == 1) - { - if (args[0]->rettype() != xpath_type_node_set) return error("Function has to be applied to node set"); - return alloc_node(ast_func_sum, xpath_type_number, args[0]); - } - - break; - - case 't': - if (name == PUGIXML_TEXT("translate") && argc == 3) - return alloc_node(ast_func_translate, xpath_type_string, args[0], args[1]); - else if (name == PUGIXML_TEXT("true") && argc == 0) - return alloc_node(ast_func_true, xpath_type_boolean); - - break; - - default: - break; + case 'l': + if (name == PUGIXML_TEXT("last") && argc == 0) + return alloc_node(ast_func_last, xpath_type_number); + else if (name == PUGIXML_TEXT("lang") && argc == 1) + return alloc_node(ast_func_lang, xpath_type_boolean, args[0]); + else if (name == PUGIXML_TEXT("local-name") && argc <= 1) + { + if (argc == 1 && args[0]->rettype() != xpath_type_node_set) + return error("Function has to be applied to node set"); + return alloc_node(argc == 0 ? ast_func_local_name_0 : ast_func_local_name_1, xpath_type_string, args[0]); } - return error("Unrecognized function or wrong parameter count"); - } + break; - axis_t parse_axis_name(const xpath_lexer_string& name, bool& specified) - { - specified = true; - - switch (name.begin[0]) + case 'n': + if (name == PUGIXML_TEXT("name") && argc <= 1) { - case 'a': - if (name == PUGIXML_TEXT("ancestor")) - return axis_ancestor; - else if (name == PUGIXML_TEXT("ancestor-or-self")) - return axis_ancestor_or_self; - else if (name == PUGIXML_TEXT("attribute")) - return axis_attribute; + if (argc == 1 && args[0]->rettype() != xpath_type_node_set) + return error("Function has to be applied to node set"); + return alloc_node(argc == 0 ? ast_func_name_0 : ast_func_name_1, xpath_type_string, args[0]); + } + else if (name == PUGIXML_TEXT("namespace-uri") && argc <= 1) + { + if (argc == 1 && args[0]->rettype() != xpath_type_node_set) + return error("Function has to be applied to node set"); + return alloc_node(argc == 0 ? ast_func_namespace_uri_0 : ast_func_namespace_uri_1, xpath_type_string, args[0]); + } + else if (name == PUGIXML_TEXT("normalize-space") && argc <= 1) + return alloc_node(argc == 0 ? ast_func_normalize_space_0 : ast_func_normalize_space_1, xpath_type_string, args[0], args[1]); + else if (name == PUGIXML_TEXT("not") && argc == 1) + return alloc_node(ast_func_not, xpath_type_boolean, args[0]); + else if (name == PUGIXML_TEXT("number") && argc <= 1) + return alloc_node(argc == 0 ? ast_func_number_0 : ast_func_number_1, xpath_type_number, args[0]); - break; + break; - case 'c': - if (name == PUGIXML_TEXT("child")) - return axis_child; + case 'p': + if (name == PUGIXML_TEXT("position") && argc == 0) + return alloc_node(ast_func_position, xpath_type_number); - break; + break; - case 'd': - if (name == PUGIXML_TEXT("descendant")) - return axis_descendant; - else if (name == PUGIXML_TEXT("descendant-or-self")) - return axis_descendant_or_self; + case 'r': + if (name == PUGIXML_TEXT("round") && argc == 1) + return alloc_node(ast_func_round, xpath_type_number, args[0]); - break; + break; - case 'f': - if (name == PUGIXML_TEXT("following")) - return axis_following; - else if (name == PUGIXML_TEXT("following-sibling")) - return axis_following_sibling; - - break; - - case 'n': - if (name == PUGIXML_TEXT("namespace")) - return axis_namespace; - - break; - - case 'p': - if (name == PUGIXML_TEXT("parent")) - return axis_parent; - else if (name == PUGIXML_TEXT("preceding")) - return axis_preceding; - else if (name == PUGIXML_TEXT("preceding-sibling")) - return axis_preceding_sibling; - - break; - - case 's': - if (name == PUGIXML_TEXT("self")) - return axis_self; - - break; - - default: - break; + case 's': + if (name == PUGIXML_TEXT("string") && argc <= 1) + return alloc_node(argc == 0 ? ast_func_string_0 : ast_func_string_1, xpath_type_string, args[0]); + else if (name == PUGIXML_TEXT("string-length") && argc <= 1) + return alloc_node(argc == 0 ? ast_func_string_length_0 : ast_func_string_length_1, xpath_type_number, args[0]); + else if (name == PUGIXML_TEXT("starts-with") && argc == 2) + return alloc_node(ast_func_starts_with, xpath_type_boolean, args[0], args[1]); + else if (name == PUGIXML_TEXT("substring-before") && argc == 2) + return alloc_node(ast_func_substring_before, xpath_type_string, args[0], args[1]); + else if (name == PUGIXML_TEXT("substring-after") && argc == 2) + return alloc_node(ast_func_substring_after, xpath_type_string, args[0], args[1]); + else if (name == PUGIXML_TEXT("substring") && (argc == 2 || argc == 3)) + return alloc_node(argc == 2 ? ast_func_substring_2 : ast_func_substring_3, xpath_type_string, args[0], args[1]); + else if (name == PUGIXML_TEXT("sum") && argc == 1) + { + if (args[0]->rettype() != xpath_type_node_set) + return error("Function has to be applied to node set"); + return alloc_node(ast_func_sum, xpath_type_number, args[0]); } - specified = false; - return axis_child; + break; + + case 't': + if (name == PUGIXML_TEXT("translate") && argc == 3) + return alloc_node(ast_func_translate, xpath_type_string, args[0], args[1]); + else if (name == PUGIXML_TEXT("true") && argc == 0) + return alloc_node(ast_func_true, xpath_type_boolean); + + break; + + default: + break; } - nodetest_t parse_node_test_type(const xpath_lexer_string& name) + return error("Unrecognized function or wrong parameter count"); + } + + axis_t parse_axis_name(const xpath_lexer_string& name, bool& specified) + { + specified = true; + + switch (name.begin[0]) { - switch (name.begin[0]) - { - case 'c': - if (name == PUGIXML_TEXT("comment")) - return nodetest_type_comment; + case 'a': + if (name == PUGIXML_TEXT("ancestor")) + return axis_ancestor; + else if (name == PUGIXML_TEXT("ancestor-or-self")) + return axis_ancestor_or_self; + else if (name == PUGIXML_TEXT("attribute")) + return axis_attribute; - break; + break; - case 'n': - if (name == PUGIXML_TEXT("node")) - return nodetest_type_node; + case 'c': + if (name == PUGIXML_TEXT("child")) + return axis_child; - break; + break; - case 'p': - if (name == PUGIXML_TEXT("processing-instruction")) - return nodetest_type_pi; + case 'd': + if (name == PUGIXML_TEXT("descendant")) + return axis_descendant; + else if (name == PUGIXML_TEXT("descendant-or-self")) + return axis_descendant_or_self; - break; + break; - case 't': - if (name == PUGIXML_TEXT("text")) - return nodetest_type_text; + case 'f': + if (name == PUGIXML_TEXT("following")) + return axis_following; + else if (name == PUGIXML_TEXT("following-sibling")) + return axis_following_sibling; - break; + break; - default: - break; - } + case 'n': + if (name == PUGIXML_TEXT("namespace")) + return axis_namespace; - return nodetest_none; + break; + + case 'p': + if (name == PUGIXML_TEXT("parent")) + return axis_parent; + else if (name == PUGIXML_TEXT("preceding")) + return axis_preceding; + else if (name == PUGIXML_TEXT("preceding-sibling")) + return axis_preceding_sibling; + + break; + + case 's': + if (name == PUGIXML_TEXT("self")) + return axis_self; + + break; + + default: + break; } - // PrimaryExpr ::= VariableReference | '(' Expr ')' | Literal | Number | FunctionCall - xpath_ast_node* parse_primary_expression() + specified = false; + return axis_child; + } + + nodetest_t parse_node_test_type(const xpath_lexer_string& name) + { + switch (name.begin[0]) { - switch (_lexer.current()) + case 'c': + if (name == PUGIXML_TEXT("comment")) + return nodetest_type_comment; + + break; + + case 'n': + if (name == PUGIXML_TEXT("node")) + return nodetest_type_node; + + break; + + case 'p': + if (name == PUGIXML_TEXT("processing-instruction")) + return nodetest_type_pi; + + break; + + case 't': + if (name == PUGIXML_TEXT("text")) + return nodetest_type_text; + + break; + + default: + break; + } + + return nodetest_none; + } + + // PrimaryExpr ::= VariableReference | '(' Expr ')' | Literal | Number | FunctionCall + xpath_ast_node* parse_primary_expression() + { + switch (_lexer.current()) + { + case lex_var_ref: + { + xpath_lexer_string name = _lexer.contents(); + + if (!_variables) + return error("Unknown variable: variable set is not provided"); + + xpath_variable* var = 0; + if (!get_variable_scratch(_scratch, _variables, name.begin, name.end, &var)) + return error_oom(); + + if (!var) + return error("Unknown variable: variable set does not contain the given name"); + + _lexer.next(); + + return alloc_node(ast_variable, var->type(), var); + } + + case lex_open_brace: + { + _lexer.next(); + + xpath_ast_node* n = parse_expression(); + if (!n) + return 0; + + if (_lexer.current() != lex_close_brace) + return error("Expected ')' to match an opening '('"); + + _lexer.next(); + + return n; + } + + case lex_quoted_string: + { + const char_t* value = alloc_string(_lexer.contents()); + if (!value) + return 0; + + _lexer.next(); + + return alloc_node(ast_string_constant, xpath_type_string, value); + } + + case lex_number: + { + double value = 0; + + if (!convert_string_to_number_scratch(_scratch, _lexer.contents().begin, _lexer.contents().end, &value)) + return error_oom(); + + _lexer.next(); + + return alloc_node(ast_number_constant, xpath_type_number, value); + } + + case lex_string: + { + xpath_ast_node* args[2] = {0}; + size_t argc = 0; + + xpath_lexer_string function = _lexer.contents(); + _lexer.next(); + + xpath_ast_node* last_arg = 0; + + if (_lexer.current() != lex_open_brace) + return error("Unrecognized function call"); + _lexer.next(); + + while (_lexer.current() != lex_close_brace) { - case lex_var_ref: - { - xpath_lexer_string name = _lexer.contents(); - - if (!_variables) - return error("Unknown variable: variable set is not provided"); - - xpath_variable* var = 0; - if (!get_variable_scratch(_scratch, _variables, name.begin, name.end, &var)) - return error_oom(); - - if (!var) - return error("Unknown variable: variable set does not contain the given name"); - - _lexer.next(); - - return alloc_node(ast_variable, var->type(), var); - } - - case lex_open_brace: - { - _lexer.next(); + if (argc > 0) + { + if (_lexer.current() != lex_comma) + return error("No comma between function arguments"); + _lexer.next(); + } xpath_ast_node* n = parse_expression(); - if (!n) return 0; + if (!n) + return 0; - if (_lexer.current() != lex_close_brace) - return error("Expected ')' to match an opening '('"); + if (argc < 2) + args[argc] = n; + else + last_arg->set_next(n); - _lexer.next(); - - return n; + argc++; + last_arg = n; } - case lex_quoted_string: + _lexer.next(); + + return parse_function(function, argc, args); + } + + default: + return error("Unrecognizable primary expression"); + } + } + + // FilterExpr ::= PrimaryExpr | FilterExpr Predicate + // Predicate ::= '[' PredicateExpr ']' + // PredicateExpr ::= Expr + xpath_ast_node* parse_filter_expression() + { + xpath_ast_node* n = parse_primary_expression(); + if (!n) + return 0; + + while (_lexer.current() == lex_open_square_brace) + { + _lexer.next(); + + if (n->rettype() != xpath_type_node_set) + return error("Predicate has to be applied to node set"); + + xpath_ast_node* expr = parse_expression(); + if (!expr) + return 0; + + n = alloc_node(ast_filter, n, expr, predicate_default); + if (!n) + return 0; + + if (_lexer.current() != lex_close_square_brace) + return error("Expected ']' to match an opening '['"); + + _lexer.next(); + } + + return n; + } + + // Step ::= AxisSpecifier NodeTest Predicate* | AbbreviatedStep + // AxisSpecifier ::= AxisName '::' | '@'? + // NodeTest ::= NameTest | NodeType '(' ')' | 'processing-instruction' '(' Literal ')' + // NameTest ::= '*' | NCName ':' '*' | QName + // AbbreviatedStep ::= '.' | '..' + xpath_ast_node* parse_step(xpath_ast_node* set) + { + if (set && set->rettype() != xpath_type_node_set) + return error("Step has to be applied to node set"); + + bool axis_specified = false; + axis_t axis = axis_child; // implied child axis + + if (_lexer.current() == lex_axis_attribute) + { + axis = axis_attribute; + axis_specified = true; + + _lexer.next(); + } + else if (_lexer.current() == lex_dot) + { + _lexer.next(); + + if (_lexer.current() == lex_open_square_brace) + return error("Predicates are not allowed after an abbreviated step"); + + return alloc_node(ast_step, set, axis_self, nodetest_type_node, 0); + } + else if (_lexer.current() == lex_double_dot) + { + _lexer.next(); + + if (_lexer.current() == lex_open_square_brace) + return error("Predicates are not allowed after an abbreviated step"); + + return alloc_node(ast_step, set, axis_parent, nodetest_type_node, 0); + } + + nodetest_t nt_type = nodetest_none; + xpath_lexer_string nt_name; + + if (_lexer.current() == lex_string) + { + // node name test + nt_name = _lexer.contents(); + _lexer.next(); + + // was it an axis name? + if (_lexer.current() == lex_double_colon) { - const char_t* value = alloc_string(_lexer.contents()); - if (!value) return 0; + // parse axis name + if (axis_specified) + return error("Two axis specifiers in one step"); + axis = parse_axis_name(nt_name, axis_specified); + + if (!axis_specified) + return error("Unknown axis"); + + // read actual node test _lexer.next(); - return alloc_node(ast_string_constant, xpath_type_string, value); - } - - case lex_number: - { - double value = 0; - - if (!convert_string_to_number_scratch(_scratch, _lexer.contents().begin, _lexer.contents().end, &value)) - return error_oom(); - - _lexer.next(); - - return alloc_node(ast_number_constant, xpath_type_number, value); - } - - case lex_string: - { - xpath_ast_node* args[2] = {0}; - size_t argc = 0; - - xpath_lexer_string function = _lexer.contents(); - _lexer.next(); - - xpath_ast_node* last_arg = 0; - - if (_lexer.current() != lex_open_brace) - return error("Unrecognized function call"); - _lexer.next(); - - while (_lexer.current() != lex_close_brace) + if (_lexer.current() == lex_multiply) { - if (argc > 0) - { - if (_lexer.current() != lex_comma) - return error("No comma between function arguments"); - _lexer.next(); - } - - xpath_ast_node* n = parse_expression(); - if (!n) return 0; - - if (argc < 2) args[argc] = n; - else last_arg->set_next(n); - - argc++; - last_arg = n; + nt_type = nodetest_all; + nt_name = xpath_lexer_string(); + _lexer.next(); } - - _lexer.next(); - - return parse_function(function, argc, args); - } - - default: - return error("Unrecognizable primary expression"); - } - } - - // FilterExpr ::= PrimaryExpr | FilterExpr Predicate - // Predicate ::= '[' PredicateExpr ']' - // PredicateExpr ::= Expr - xpath_ast_node* parse_filter_expression() - { - xpath_ast_node* n = parse_primary_expression(); - if (!n) return 0; - - while (_lexer.current() == lex_open_square_brace) - { - _lexer.next(); - - if (n->rettype() != xpath_type_node_set) - return error("Predicate has to be applied to node set"); - - xpath_ast_node* expr = parse_expression(); - if (!expr) return 0; - - n = alloc_node(ast_filter, n, expr, predicate_default); - if (!n) return 0; - - if (_lexer.current() != lex_close_square_brace) - return error("Expected ']' to match an opening '['"); - - _lexer.next(); - } - - return n; - } - - // Step ::= AxisSpecifier NodeTest Predicate* | AbbreviatedStep - // AxisSpecifier ::= AxisName '::' | '@'? - // NodeTest ::= NameTest | NodeType '(' ')' | 'processing-instruction' '(' Literal ')' - // NameTest ::= '*' | NCName ':' '*' | QName - // AbbreviatedStep ::= '.' | '..' - xpath_ast_node* parse_step(xpath_ast_node* set) - { - if (set && set->rettype() != xpath_type_node_set) - return error("Step has to be applied to node set"); - - bool axis_specified = false; - axis_t axis = axis_child; // implied child axis - - if (_lexer.current() == lex_axis_attribute) - { - axis = axis_attribute; - axis_specified = true; - - _lexer.next(); - } - else if (_lexer.current() == lex_dot) - { - _lexer.next(); - - if (_lexer.current() == lex_open_square_brace) - return error("Predicates are not allowed after an abbreviated step"); - - return alloc_node(ast_step, set, axis_self, nodetest_type_node, 0); - } - else if (_lexer.current() == lex_double_dot) - { - _lexer.next(); - - if (_lexer.current() == lex_open_square_brace) - return error("Predicates are not allowed after an abbreviated step"); - - return alloc_node(ast_step, set, axis_parent, nodetest_type_node, 0); - } - - nodetest_t nt_type = nodetest_none; - xpath_lexer_string nt_name; - - if (_lexer.current() == lex_string) - { - // node name test - nt_name = _lexer.contents(); - _lexer.next(); - - // was it an axis name? - if (_lexer.current() == lex_double_colon) + else if (_lexer.current() == lex_string) { - // parse axis name - if (axis_specified) - return error("Two axis specifiers in one step"); + nt_name = _lexer.contents(); + _lexer.next(); + } + else + { + return error("Unrecognized node test"); + } + } - axis = parse_axis_name(nt_name, axis_specified); - - if (!axis_specified) - return error("Unknown axis"); - - // read actual node test + if (nt_type == nodetest_none) + { + // node type test or processing-instruction + if (_lexer.current() == lex_open_brace) + { _lexer.next(); - if (_lexer.current() == lex_multiply) + if (_lexer.current() == lex_close_brace) { - nt_type = nodetest_all; - nt_name = xpath_lexer_string(); _lexer.next(); + + nt_type = parse_node_test_type(nt_name); + + if (nt_type == nodetest_none) + return error("Unrecognized node type"); + + nt_name = xpath_lexer_string(); } - else if (_lexer.current() == lex_string) + else if (nt_name == PUGIXML_TEXT("processing-instruction")) { + if (_lexer.current() != lex_quoted_string) + return error("Only literals are allowed as arguments to processing-instruction()"); + + nt_type = nodetest_pi; nt_name = _lexer.contents(); _lexer.next(); - } - else - { - return error("Unrecognized node test"); - } - } - if (nt_type == nodetest_none) - { - // node type test or processing-instruction - if (_lexer.current() == lex_open_brace) - { + if (_lexer.current() != lex_close_brace) + return error("Unmatched brace near processing-instruction()"); _lexer.next(); - - if (_lexer.current() == lex_close_brace) - { - _lexer.next(); - - nt_type = parse_node_test_type(nt_name); - - if (nt_type == nodetest_none) - return error("Unrecognized node type"); - - nt_name = xpath_lexer_string(); - } - else if (nt_name == PUGIXML_TEXT("processing-instruction")) - { - if (_lexer.current() != lex_quoted_string) - return error("Only literals are allowed as arguments to processing-instruction()"); - - nt_type = nodetest_pi; - nt_name = _lexer.contents(); - _lexer.next(); - - if (_lexer.current() != lex_close_brace) - return error("Unmatched brace near processing-instruction()"); - _lexer.next(); - } - else - { - return error("Unmatched brace near node type test"); - } } - // QName or NCName:* else { - if (nt_name.end - nt_name.begin > 2 && nt_name.end[-2] == ':' && nt_name.end[-1] == '*') // NCName:* - { - nt_name.end--; // erase * + return error("Unmatched brace near node type test"); + } + } + // QName or NCName:* + else + { + if (nt_name.end - nt_name.begin > 2 && nt_name.end[-2] == ':' && nt_name.end[-1] == '*') // NCName:* + { + nt_name.end--; // erase * - nt_type = nodetest_all_in_namespace; - } - else - { - nt_type = nodetest_name; - } + nt_type = nodetest_all_in_namespace; + } + else + { + nt_type = nodetest_name; } } } - else if (_lexer.current() == lex_multiply) - { - nt_type = nodetest_all; - _lexer.next(); - } - else - { - return error("Unrecognized node test"); - } - - const char_t* nt_name_copy = alloc_string(nt_name); - if (!nt_name_copy) return 0; - - xpath_ast_node* n = alloc_node(ast_step, set, axis, nt_type, nt_name_copy); - if (!n) return 0; - - xpath_ast_node* last = 0; - - while (_lexer.current() == lex_open_square_brace) - { - _lexer.next(); - - xpath_ast_node* expr = parse_expression(); - if (!expr) return 0; - - xpath_ast_node* pred = alloc_node(ast_predicate, 0, expr, predicate_default); - if (!pred) return 0; - - if (_lexer.current() != lex_close_square_brace) - return error("Expected ']' to match an opening '['"); - _lexer.next(); - - if (last) last->set_next(pred); - else n->set_right(pred); - - last = pred; - } - - return n; + } + else if (_lexer.current() == lex_multiply) + { + nt_type = nodetest_all; + _lexer.next(); + } + else + { + return error("Unrecognized node test"); } - // RelativeLocationPath ::= Step | RelativeLocationPath '/' Step | RelativeLocationPath '//' Step - xpath_ast_node* parse_relative_location_path(xpath_ast_node* set) - { - xpath_ast_node* n = parse_step(set); - if (!n) return 0; + const char_t* nt_name_copy = alloc_string(nt_name); + if (!nt_name_copy) + return 0; - while (_lexer.current() == lex_slash || _lexer.current() == lex_double_slash) + xpath_ast_node* n = alloc_node(ast_step, set, axis, nt_type, nt_name_copy); + if (!n) + return 0; + + xpath_ast_node* last = 0; + + while (_lexer.current() == lex_open_square_brace) + { + _lexer.next(); + + xpath_ast_node* expr = parse_expression(); + if (!expr) + return 0; + + xpath_ast_node* pred = alloc_node(ast_predicate, 0, expr, predicate_default); + if (!pred) + return 0; + + if (_lexer.current() != lex_close_square_brace) + return error("Expected ']' to match an opening '['"); + _lexer.next(); + + if (last) + last->set_next(pred); + else + n->set_right(pred); + + last = pred; + } + + return n; + } + + // RelativeLocationPath ::= Step | RelativeLocationPath '/' Step | RelativeLocationPath '//' Step + xpath_ast_node* parse_relative_location_path(xpath_ast_node* set) + { + xpath_ast_node* n = parse_step(set); + if (!n) + return 0; + + while (_lexer.current() == lex_slash || _lexer.current() == lex_double_slash) + { + lexeme_t l = _lexer.current(); + _lexer.next(); + + if (l == lex_double_slash) + { + n = alloc_node(ast_step, n, axis_descendant_or_self, nodetest_type_node, 0); + if (!n) + return 0; + } + + n = parse_step(n); + if (!n) + return 0; + } + + return n; + } + + // LocationPath ::= RelativeLocationPath | AbsoluteLocationPath + // AbsoluteLocationPath ::= '/' RelativeLocationPath? | '//' RelativeLocationPath + xpath_ast_node* parse_location_path() + { + if (_lexer.current() == lex_slash) + { + _lexer.next(); + + xpath_ast_node* n = alloc_node(ast_step_root, xpath_type_node_set); + if (!n) + return 0; + + // relative location path can start from axis_attribute, dot, double_dot, multiply and string lexemes; any other lexeme means standalone root path + lexeme_t l = _lexer.current(); + + if (l == lex_string || l == lex_axis_attribute || l == lex_dot || l == lex_double_dot || l == lex_multiply) + return parse_relative_location_path(n); + else + return n; + } + else if (_lexer.current() == lex_double_slash) + { + _lexer.next(); + + xpath_ast_node* n = alloc_node(ast_step_root, xpath_type_node_set); + if (!n) + return 0; + + n = alloc_node(ast_step, n, axis_descendant_or_self, nodetest_type_node, 0); + if (!n) + return 0; + + return parse_relative_location_path(n); + } + + // else clause moved outside of if because of bogus warning 'control may reach end of non-void function being inlined' in gcc 4.0.1 + return parse_relative_location_path(0); + } + + // PathExpr ::= LocationPath + // | FilterExpr + // | FilterExpr '/' RelativeLocationPath + // | FilterExpr '//' RelativeLocationPath + // UnionExpr ::= PathExpr | UnionExpr '|' PathExpr + // UnaryExpr ::= UnionExpr | '-' UnaryExpr + xpath_ast_node* parse_path_or_unary_expression() + { + // Clarification. + // PathExpr begins with either LocationPath or FilterExpr. + // FilterExpr begins with PrimaryExpr + // PrimaryExpr begins with '$' in case of it being a variable reference, + // '(' in case of it being an expression, string literal, number constant or + // function call. + if (_lexer.current() == lex_var_ref || _lexer.current() == lex_open_brace || + _lexer.current() == lex_quoted_string || _lexer.current() == lex_number || + _lexer.current() == lex_string) + { + if (_lexer.current() == lex_string) + { + // This is either a function call, or not - if not, we shall proceed with location path + const char_t* state = _lexer.state(); + + while (PUGI__IS_CHARTYPE(*state, ct_space)) + ++state; + + if (*state != '(') + return parse_location_path(); + + // This looks like a function call; however this still can be a node-test. Check it. + if (parse_node_test_type(_lexer.contents()) != nodetest_none) + return parse_location_path(); + } + + xpath_ast_node* n = parse_filter_expression(); + if (!n) + return 0; + + if (_lexer.current() == lex_slash || _lexer.current() == lex_double_slash) { lexeme_t l = _lexer.current(); _lexer.next(); if (l == lex_double_slash) { + if (n->rettype() != xpath_type_node_set) + return error("Step has to be applied to node set"); + n = alloc_node(ast_step, n, axis_descendant_or_self, nodetest_type_node, 0); - if (!n) return 0; + if (!n) + return 0; } - n = parse_step(n); - if (!n) return 0; - } - - return n; - } - - // LocationPath ::= RelativeLocationPath | AbsoluteLocationPath - // AbsoluteLocationPath ::= '/' RelativeLocationPath? | '//' RelativeLocationPath - xpath_ast_node* parse_location_path() - { - if (_lexer.current() == lex_slash) - { - _lexer.next(); - - xpath_ast_node* n = alloc_node(ast_step_root, xpath_type_node_set); - if (!n) return 0; - - // relative location path can start from axis_attribute, dot, double_dot, multiply and string lexemes; any other lexeme means standalone root path - lexeme_t l = _lexer.current(); - - if (l == lex_string || l == lex_axis_attribute || l == lex_dot || l == lex_double_dot || l == lex_multiply) - return parse_relative_location_path(n); - else - return n; - } - else if (_lexer.current() == lex_double_slash) - { - _lexer.next(); - - xpath_ast_node* n = alloc_node(ast_step_root, xpath_type_node_set); - if (!n) return 0; - - n = alloc_node(ast_step, n, axis_descendant_or_self, nodetest_type_node, 0); - if (!n) return 0; - + // select from location path return parse_relative_location_path(n); } - // else clause moved outside of if because of bogus warning 'control may reach end of non-void function being inlined' in gcc 4.0.1 - return parse_relative_location_path(0); - } - - // PathExpr ::= LocationPath - // | FilterExpr - // | FilterExpr '/' RelativeLocationPath - // | FilterExpr '//' RelativeLocationPath - // UnionExpr ::= PathExpr | UnionExpr '|' PathExpr - // UnaryExpr ::= UnionExpr | '-' UnaryExpr - xpath_ast_node* parse_path_or_unary_expression() - { - // Clarification. - // PathExpr begins with either LocationPath or FilterExpr. - // FilterExpr begins with PrimaryExpr - // PrimaryExpr begins with '$' in case of it being a variable reference, - // '(' in case of it being an expression, string literal, number constant or - // function call. - if (_lexer.current() == lex_var_ref || _lexer.current() == lex_open_brace || - _lexer.current() == lex_quoted_string || _lexer.current() == lex_number || - _lexer.current() == lex_string) - { - if (_lexer.current() == lex_string) - { - // This is either a function call, or not - if not, we shall proceed with location path - const char_t* state = _lexer.state(); - - while (PUGI__IS_CHARTYPE(*state, ct_space)) ++state; - - if (*state != '(') - return parse_location_path(); - - // This looks like a function call; however this still can be a node-test. Check it. - if (parse_node_test_type(_lexer.contents()) != nodetest_none) - return parse_location_path(); - } - - xpath_ast_node* n = parse_filter_expression(); - if (!n) return 0; - - if (_lexer.current() == lex_slash || _lexer.current() == lex_double_slash) - { - lexeme_t l = _lexer.current(); - _lexer.next(); - - if (l == lex_double_slash) - { - if (n->rettype() != xpath_type_node_set) - return error("Step has to be applied to node set"); - - n = alloc_node(ast_step, n, axis_descendant_or_self, nodetest_type_node, 0); - if (!n) return 0; - } - - // select from location path - return parse_relative_location_path(n); - } - - return n; - } - else if (_lexer.current() == lex_minus) - { - _lexer.next(); - - // precedence 7+ - only parses union expressions - xpath_ast_node* n = parse_expression(7); - if (!n) return 0; - - return alloc_node(ast_op_negate, xpath_type_number, n); - } - else - { - return parse_location_path(); - } - } - - struct binary_op_t - { - ast_type_t asttype; - xpath_value_type rettype; - int precedence; - - binary_op_t(): asttype(ast_unknown), rettype(xpath_type_none), precedence(0) - { - } - - binary_op_t(ast_type_t asttype_, xpath_value_type rettype_, int precedence_): asttype(asttype_), rettype(rettype_), precedence(precedence_) - { - } - - static binary_op_t parse(xpath_lexer& lexer) - { - switch (lexer.current()) - { - case lex_string: - if (lexer.contents() == PUGIXML_TEXT("or")) - return binary_op_t(ast_op_or, xpath_type_boolean, 1); - else if (lexer.contents() == PUGIXML_TEXT("and")) - return binary_op_t(ast_op_and, xpath_type_boolean, 2); - else if (lexer.contents() == PUGIXML_TEXT("div")) - return binary_op_t(ast_op_divide, xpath_type_number, 6); - else if (lexer.contents() == PUGIXML_TEXT("mod")) - return binary_op_t(ast_op_mod, xpath_type_number, 6); - else - return binary_op_t(); - - case lex_equal: - return binary_op_t(ast_op_equal, xpath_type_boolean, 3); - - case lex_not_equal: - return binary_op_t(ast_op_not_equal, xpath_type_boolean, 3); - - case lex_less: - return binary_op_t(ast_op_less, xpath_type_boolean, 4); - - case lex_greater: - return binary_op_t(ast_op_greater, xpath_type_boolean, 4); - - case lex_less_or_equal: - return binary_op_t(ast_op_less_or_equal, xpath_type_boolean, 4); - - case lex_greater_or_equal: - return binary_op_t(ast_op_greater_or_equal, xpath_type_boolean, 4); - - case lex_plus: - return binary_op_t(ast_op_add, xpath_type_number, 5); - - case lex_minus: - return binary_op_t(ast_op_subtract, xpath_type_number, 5); - - case lex_multiply: - return binary_op_t(ast_op_multiply, xpath_type_number, 6); - - case lex_union: - return binary_op_t(ast_op_union, xpath_type_node_set, 7); - - default: - return binary_op_t(); - } - } - }; - - xpath_ast_node* parse_expression_rec(xpath_ast_node* lhs, int limit) - { - binary_op_t op = binary_op_t::parse(_lexer); - - while (op.asttype != ast_unknown && op.precedence >= limit) - { - _lexer.next(); - - xpath_ast_node* rhs = parse_path_or_unary_expression(); - if (!rhs) return 0; - - binary_op_t nextop = binary_op_t::parse(_lexer); - - while (nextop.asttype != ast_unknown && nextop.precedence > op.precedence) - { - rhs = parse_expression_rec(rhs, nextop.precedence); - if (!rhs) return 0; - - nextop = binary_op_t::parse(_lexer); - } - - if (op.asttype == ast_op_union && (lhs->rettype() != xpath_type_node_set || rhs->rettype() != xpath_type_node_set)) - return error("Union operator has to be applied to node sets"); - - lhs = alloc_node(op.asttype, op.rettype, lhs, rhs); - if (!lhs) return 0; - - op = binary_op_t::parse(_lexer); - } - - return lhs; - } - - // Expr ::= OrExpr - // OrExpr ::= AndExpr | OrExpr 'or' AndExpr - // AndExpr ::= EqualityExpr | AndExpr 'and' EqualityExpr - // EqualityExpr ::= RelationalExpr - // | EqualityExpr '=' RelationalExpr - // | EqualityExpr '!=' RelationalExpr - // RelationalExpr ::= AdditiveExpr - // | RelationalExpr '<' AdditiveExpr - // | RelationalExpr '>' AdditiveExpr - // | RelationalExpr '<=' AdditiveExpr - // | RelationalExpr '>=' AdditiveExpr - // AdditiveExpr ::= MultiplicativeExpr - // | AdditiveExpr '+' MultiplicativeExpr - // | AdditiveExpr '-' MultiplicativeExpr - // MultiplicativeExpr ::= UnaryExpr - // | MultiplicativeExpr '*' UnaryExpr - // | MultiplicativeExpr 'div' UnaryExpr - // | MultiplicativeExpr 'mod' UnaryExpr - xpath_ast_node* parse_expression(int limit = 0) - { - xpath_ast_node* n = parse_path_or_unary_expression(); - if (!n) return 0; - - return parse_expression_rec(n, limit); - } - - xpath_parser(const char_t* query, xpath_variable_set* variables, xpath_allocator* alloc, xpath_parse_result* result): _alloc(alloc), _lexer(query), _query(query), _variables(variables), _result(result) - { - } - - xpath_ast_node* parse() - { - xpath_ast_node* n = parse_expression(); - if (!n) return 0; - - // check if there are unparsed tokens left - if (_lexer.current() != lex_eof) - return error("Incorrect query"); - return n; } - - static xpath_ast_node* parse(const char_t* query, xpath_variable_set* variables, xpath_allocator* alloc, xpath_parse_result* result) + else if (_lexer.current() == lex_minus) { - xpath_parser parser(query, variables, alloc, result); + _lexer.next(); - return parser.parse(); + // precedence 7+ - only parses union expressions + xpath_ast_node* n = parse_expression(7); + if (!n) + return 0; + + return alloc_node(ast_op_negate, xpath_type_number, n); } - }; - - struct xpath_query_impl - { - static xpath_query_impl* create() + else { - void* memory = xml_memory::allocate(sizeof(xpath_query_impl)); - if (!memory) return 0; - - return new (memory) xpath_query_impl(); + return parse_location_path(); } - - static void destroy(xpath_query_impl* impl) - { - // free all allocated pages - impl->alloc.release(); - - // free allocator memory (with the first page) - xml_memory::deallocate(impl); - } - - xpath_query_impl(): root(0), alloc(&block, &oom), oom(false) - { - block.next = 0; - block.capacity = sizeof(block.data); - } - - xpath_ast_node* root; - xpath_allocator alloc; - xpath_memory_block block; - bool oom; - }; - - PUGI__FN impl::xpath_ast_node* evaluate_node_set_prepare(xpath_query_impl* impl) - { - if (!impl) return 0; - - if (impl->root->rettype() != xpath_type_node_set) - { - #ifdef PUGIXML_NO_EXCEPTIONS - return 0; - #else - xpath_parse_result res; - res.error = "Expression does not evaluate to node set"; - - throw xpath_exception(res); - #endif - } - - return impl->root; } + + struct binary_op_t + { + ast_type_t asttype; + xpath_value_type rettype; + int precedence; + + binary_op_t() + : asttype(ast_unknown) + , rettype(xpath_type_none) + , precedence(0) + { + } + + binary_op_t(ast_type_t asttype_, xpath_value_type rettype_, int precedence_) + : asttype(asttype_) + , rettype(rettype_) + , precedence(precedence_) + { + } + + static binary_op_t parse(xpath_lexer& lexer) + { + switch (lexer.current()) + { + case lex_string: + if (lexer.contents() == PUGIXML_TEXT("or")) + return binary_op_t(ast_op_or, xpath_type_boolean, 1); + else if (lexer.contents() == PUGIXML_TEXT("and")) + return binary_op_t(ast_op_and, xpath_type_boolean, 2); + else if (lexer.contents() == PUGIXML_TEXT("div")) + return binary_op_t(ast_op_divide, xpath_type_number, 6); + else if (lexer.contents() == PUGIXML_TEXT("mod")) + return binary_op_t(ast_op_mod, xpath_type_number, 6); + else + return binary_op_t(); + + case lex_equal: + return binary_op_t(ast_op_equal, xpath_type_boolean, 3); + + case lex_not_equal: + return binary_op_t(ast_op_not_equal, xpath_type_boolean, 3); + + case lex_less: + return binary_op_t(ast_op_less, xpath_type_boolean, 4); + + case lex_greater: + return binary_op_t(ast_op_greater, xpath_type_boolean, 4); + + case lex_less_or_equal: + return binary_op_t(ast_op_less_or_equal, xpath_type_boolean, 4); + + case lex_greater_or_equal: + return binary_op_t(ast_op_greater_or_equal, xpath_type_boolean, 4); + + case lex_plus: + return binary_op_t(ast_op_add, xpath_type_number, 5); + + case lex_minus: + return binary_op_t(ast_op_subtract, xpath_type_number, 5); + + case lex_multiply: + return binary_op_t(ast_op_multiply, xpath_type_number, 6); + + case lex_union: + return binary_op_t(ast_op_union, xpath_type_node_set, 7); + + default: + return binary_op_t(); + } + } + }; + + xpath_ast_node* parse_expression_rec(xpath_ast_node* lhs, int limit) + { + binary_op_t op = binary_op_t::parse(_lexer); + + while (op.asttype != ast_unknown && op.precedence >= limit) + { + _lexer.next(); + + xpath_ast_node* rhs = parse_path_or_unary_expression(); + if (!rhs) + return 0; + + binary_op_t nextop = binary_op_t::parse(_lexer); + + while (nextop.asttype != ast_unknown && nextop.precedence > op.precedence) + { + rhs = parse_expression_rec(rhs, nextop.precedence); + if (!rhs) + return 0; + + nextop = binary_op_t::parse(_lexer); + } + + if (op.asttype == ast_op_union && (lhs->rettype() != xpath_type_node_set || rhs->rettype() != xpath_type_node_set)) + return error("Union operator has to be applied to node sets"); + + lhs = alloc_node(op.asttype, op.rettype, lhs, rhs); + if (!lhs) + return 0; + + op = binary_op_t::parse(_lexer); + } + + return lhs; + } + + // Expr ::= OrExpr + // OrExpr ::= AndExpr | OrExpr 'or' AndExpr + // AndExpr ::= EqualityExpr | AndExpr 'and' EqualityExpr + // EqualityExpr ::= RelationalExpr + // | EqualityExpr '=' RelationalExpr + // | EqualityExpr '!=' RelationalExpr + // RelationalExpr ::= AdditiveExpr + // | RelationalExpr '<' AdditiveExpr + // | RelationalExpr '>' AdditiveExpr + // | RelationalExpr '<=' AdditiveExpr + // | RelationalExpr '>=' AdditiveExpr + // AdditiveExpr ::= MultiplicativeExpr + // | AdditiveExpr '+' MultiplicativeExpr + // | AdditiveExpr '-' MultiplicativeExpr + // MultiplicativeExpr ::= UnaryExpr + // | MultiplicativeExpr '*' UnaryExpr + // | MultiplicativeExpr 'div' UnaryExpr + // | MultiplicativeExpr 'mod' UnaryExpr + xpath_ast_node* parse_expression(int limit = 0) + { + xpath_ast_node* n = parse_path_or_unary_expression(); + if (!n) + return 0; + + return parse_expression_rec(n, limit); + } + + xpath_parser(const char_t* query, xpath_variable_set* variables, xpath_allocator* alloc, xpath_parse_result* result) + : _alloc(alloc) + , _lexer(query) + , _query(query) + , _variables(variables) + , _result(result) + { + } + + xpath_ast_node* parse() + { + xpath_ast_node* n = parse_expression(); + if (!n) + return 0; + + // check if there are unparsed tokens left + if (_lexer.current() != lex_eof) + return error("Incorrect query"); + + return n; + } + + static xpath_ast_node* parse(const char_t* query, xpath_variable_set* variables, xpath_allocator* alloc, xpath_parse_result* result) + { + xpath_parser parser(query, variables, alloc, result); + + return parser.parse(); + } +}; + +struct xpath_query_impl +{ + static xpath_query_impl* create() + { + void* memory = xml_memory::allocate(sizeof(xpath_query_impl)); + if (!memory) + return 0; + + return new (memory) xpath_query_impl(); + } + + static void destroy(xpath_query_impl* impl) + { + // free all allocated pages + impl->alloc.release(); + + // free allocator memory (with the first page) + xml_memory::deallocate(impl); + } + + xpath_query_impl() + : root(0) + , alloc(&block, &oom) + , oom(false) + { + block.next = 0; + block.capacity = sizeof(block.data); + } + + xpath_ast_node* root; + xpath_allocator alloc; + xpath_memory_block block; + bool oom; +}; + +PUGI__FN impl::xpath_ast_node* evaluate_node_set_prepare(xpath_query_impl* impl) +{ + if (!impl) + return 0; + + if (impl->root->rettype() != xpath_type_node_set) + { +#ifdef PUGIXML_NO_EXCEPTIONS + return 0; +#else + xpath_parse_result res; + res.error = "Expression does not evaluate to node set"; + + throw xpath_exception(res); +#endif + } + + return impl->root; +} PUGI__NS_END namespace pugi { #ifndef PUGIXML_NO_EXCEPTIONS - PUGI__FN xpath_exception::xpath_exception(const xpath_parse_result& result_): _result(result_) - { - assert(_result.error); - } +PUGI__FN xpath_exception::xpath_exception(const xpath_parse_result& result_) + : _result(result_) +{ + assert(_result.error); +} - PUGI__FN const char* xpath_exception::what() const throw() - { - return _result.error; - } +PUGI__FN const char* xpath_exception::what() const throw() +{ + return _result.error; +} - PUGI__FN const xpath_parse_result& xpath_exception::result() const - { - return _result; - } +PUGI__FN const xpath_parse_result& xpath_exception::result() const +{ + return _result; +} #endif - PUGI__FN xpath_node::xpath_node() - { - } +PUGI__FN xpath_node::xpath_node() +{ +} - PUGI__FN xpath_node::xpath_node(const xml_node& node_): _node(node_) - { - } +PUGI__FN xpath_node::xpath_node(const xml_node& node_) + : _node(node_) +{ +} - PUGI__FN xpath_node::xpath_node(const xml_attribute& attribute_, const xml_node& parent_): _node(attribute_ ? parent_ : xml_node()), _attribute(attribute_) - { - } +PUGI__FN xpath_node::xpath_node(const xml_attribute& attribute_, const xml_node& parent_) + : _node(attribute_ ? parent_ : xml_node()) + , _attribute(attribute_) +{ +} - PUGI__FN xml_node xpath_node::node() const - { - return _attribute ? xml_node() : _node; - } +PUGI__FN xml_node xpath_node::node() const +{ + return _attribute ? xml_node() : _node; +} - PUGI__FN xml_attribute xpath_node::attribute() const - { - return _attribute; - } +PUGI__FN xml_attribute xpath_node::attribute() const +{ + return _attribute; +} - PUGI__FN xml_node xpath_node::parent() const - { - return _attribute ? _node : _node.parent(); - } +PUGI__FN xml_node xpath_node::parent() const +{ + return _attribute ? _node : _node.parent(); +} - PUGI__FN static void unspecified_bool_xpath_node(xpath_node***) - { - } +PUGI__FN static void unspecified_bool_xpath_node(xpath_node***) +{ +} - PUGI__FN xpath_node::operator xpath_node::unspecified_bool_type() const - { - return (_node || _attribute) ? unspecified_bool_xpath_node : 0; - } +PUGI__FN xpath_node::operator xpath_node::unspecified_bool_type() const +{ + return (_node || _attribute) ? unspecified_bool_xpath_node : 0; +} - PUGI__FN bool xpath_node::operator!() const - { - return !(_node || _attribute); - } +PUGI__FN bool xpath_node::operator!() const +{ + return !(_node || _attribute); +} - PUGI__FN bool xpath_node::operator==(const xpath_node& n) const - { - return _node == n._node && _attribute == n._attribute; - } +PUGI__FN bool xpath_node::operator==(const xpath_node& n) const +{ + return _node == n._node && _attribute == n._attribute; +} - PUGI__FN bool xpath_node::operator!=(const xpath_node& n) const - { - return _node != n._node || _attribute != n._attribute; - } +PUGI__FN bool xpath_node::operator!=(const xpath_node& n) const +{ + return _node != n._node || _attribute != n._attribute; +} #ifdef __BORLANDC__ - PUGI__FN bool operator&&(const xpath_node& lhs, bool rhs) - { - return (bool)lhs && rhs; - } +PUGI__FN bool operator&&(const xpath_node& lhs, bool rhs) +{ + return (bool)lhs && rhs; +} - PUGI__FN bool operator||(const xpath_node& lhs, bool rhs) - { - return (bool)lhs || rhs; - } +PUGI__FN bool operator||(const xpath_node& lhs, bool rhs) +{ + return (bool)lhs || rhs; +} #endif - PUGI__FN void xpath_node_set::_assign(const_iterator begin_, const_iterator end_, type_t type_) +PUGI__FN void xpath_node_set::_assign(const_iterator begin_, const_iterator end_, type_t type_) +{ + assert(begin_ <= end_); + + size_t size_ = static_cast(end_ - begin_); + + // use internal buffer for 0 or 1 elements, heap buffer otherwise + xpath_node* storage = (size_ <= 1) ? _storage : static_cast(impl::xml_memory::allocate(size_ * sizeof(xpath_node))); + + if (!storage) { - assert(begin_ <= end_); - - size_t size_ = static_cast(end_ - begin_); - - // use internal buffer for 0 or 1 elements, heap buffer otherwise - xpath_node* storage = (size_ <= 1) ? _storage : static_cast(impl::xml_memory::allocate(size_ * sizeof(xpath_node))); - - if (!storage) - { - #ifdef PUGIXML_NO_EXCEPTIONS - return; - #else - throw std::bad_alloc(); - #endif - } - - // deallocate old buffer - if (_begin != _storage) - impl::xml_memory::deallocate(_begin); - - // size check is necessary because for begin_ = end_ = nullptr, memcpy is UB - if (size_) - memcpy(storage, begin_, size_ * sizeof(xpath_node)); - - _begin = storage; - _end = storage + size_; - _type = type_; +#ifdef PUGIXML_NO_EXCEPTIONS + return; +#else + throw std::bad_alloc(); +#endif } + // deallocate old buffer + if (_begin != _storage) + impl::xml_memory::deallocate(_begin); + + // size check is necessary because for begin_ = end_ = nullptr, memcpy is UB + if (size_) + memcpy(storage, begin_, size_ * sizeof(xpath_node)); + + _begin = storage; + _end = storage + size_; + _type = type_; +} + #ifdef PUGIXML_HAS_MOVE - PUGI__FN void xpath_node_set::_move(xpath_node_set& rhs) PUGIXML_NOEXCEPT - { - _type = rhs._type; - _storage[0] = rhs._storage[0]; - _begin = (rhs._begin == rhs._storage) ? _storage : rhs._begin; - _end = _begin + (rhs._end - rhs._begin); +PUGI__FN void xpath_node_set::_move(xpath_node_set& rhs) PUGIXML_NOEXCEPT +{ + _type = rhs._type; + _storage[0] = rhs._storage[0]; + _begin = (rhs._begin == rhs._storage) ? _storage : rhs._begin; + _end = _begin + (rhs._end - rhs._begin); - rhs._type = type_unsorted; - rhs._begin = rhs._storage; - rhs._end = rhs._storage; - } + rhs._type = type_unsorted; + rhs._begin = rhs._storage; + rhs._end = rhs._storage; +} #endif - PUGI__FN xpath_node_set::xpath_node_set(): _type(type_unsorted), _begin(_storage), _end(_storage) - { - } +PUGI__FN xpath_node_set::xpath_node_set() + : _type(type_unsorted) + , _begin(_storage) + , _end(_storage) +{ +} - PUGI__FN xpath_node_set::xpath_node_set(const_iterator begin_, const_iterator end_, type_t type_): _type(type_unsorted), _begin(_storage), _end(_storage) - { - _assign(begin_, end_, type_); - } +PUGI__FN xpath_node_set::xpath_node_set(const_iterator begin_, const_iterator end_, type_t type_) + : _type(type_unsorted) + , _begin(_storage) + , _end(_storage) +{ + _assign(begin_, end_, type_); +} - PUGI__FN xpath_node_set::~xpath_node_set() - { - if (_begin != _storage) - impl::xml_memory::deallocate(_begin); - } +PUGI__FN xpath_node_set::~xpath_node_set() +{ + if (_begin != _storage) + impl::xml_memory::deallocate(_begin); +} - PUGI__FN xpath_node_set::xpath_node_set(const xpath_node_set& ns): _type(type_unsorted), _begin(_storage), _end(_storage) - { - _assign(ns._begin, ns._end, ns._type); - } - - PUGI__FN xpath_node_set& xpath_node_set::operator=(const xpath_node_set& ns) - { - if (this == &ns) return *this; - - _assign(ns._begin, ns._end, ns._type); +PUGI__FN xpath_node_set::xpath_node_set(const xpath_node_set& ns) + : _type(type_unsorted) + , _begin(_storage) + , _end(_storage) +{ + _assign(ns._begin, ns._end, ns._type); +} +PUGI__FN xpath_node_set& xpath_node_set::operator=(const xpath_node_set& ns) +{ + if (this == &ns) return *this; - } + + _assign(ns._begin, ns._end, ns._type); + + return *this; +} #ifdef PUGIXML_HAS_MOVE - PUGI__FN xpath_node_set::xpath_node_set(xpath_node_set&& rhs) PUGIXML_NOEXCEPT: _type(type_unsorted), _begin(_storage), _end(_storage) - { - _move(rhs); - } - - PUGI__FN xpath_node_set& xpath_node_set::operator=(xpath_node_set&& rhs) PUGIXML_NOEXCEPT - { - if (this == &rhs) return *this; - - if (_begin != _storage) - impl::xml_memory::deallocate(_begin); - - _move(rhs); +PUGI__FN xpath_node_set::xpath_node_set(xpath_node_set&& rhs) PUGIXML_NOEXCEPT : _type(type_unsorted), _begin(_storage), _end(_storage) +{ + _move(rhs); +} +PUGI__FN xpath_node_set& xpath_node_set::operator=(xpath_node_set&& rhs) PUGIXML_NOEXCEPT +{ + if (this == &rhs) return *this; - } + + if (_begin != _storage) + impl::xml_memory::deallocate(_begin); + + _move(rhs); + + return *this; +} #endif - PUGI__FN xpath_node_set::type_t xpath_node_set::type() const +PUGI__FN xpath_node_set::type_t xpath_node_set::type() const +{ + return _type; +} + +PUGI__FN size_t xpath_node_set::size() const +{ + return _end - _begin; +} + +PUGI__FN bool xpath_node_set::empty() const +{ + return _begin == _end; +} + +PUGI__FN const xpath_node& xpath_node_set::operator[](size_t index) const +{ + assert(index < size()); + return _begin[index]; +} + +PUGI__FN xpath_node_set::const_iterator xpath_node_set::begin() const +{ + return _begin; +} + +PUGI__FN xpath_node_set::const_iterator xpath_node_set::end() const +{ + return _end; +} + +PUGI__FN void xpath_node_set::sort(bool reverse) +{ + _type = impl::xpath_sort(_begin, _end, _type, reverse); +} + +PUGI__FN xpath_node xpath_node_set::first() const +{ + return impl::xpath_first(_begin, _end, _type); +} + +PUGI__FN xpath_parse_result::xpath_parse_result() + : error("Internal error") + , offset(0) +{ +} + +PUGI__FN xpath_parse_result::operator bool() const +{ + return error == 0; +} + +PUGI__FN const char* xpath_parse_result::description() const +{ + return error ? error : "No error"; +} + +PUGI__FN xpath_variable::xpath_variable(xpath_value_type type_) + : _type(type_) + , _next(0) +{ +} + +PUGI__FN const char_t* xpath_variable::name() const +{ + switch (_type) { - return _type; - } + case xpath_type_node_set: + return static_cast(this)->name; - PUGI__FN size_t xpath_node_set::size() const - { - return _end - _begin; - } + case xpath_type_number: + return static_cast(this)->name; - PUGI__FN bool xpath_node_set::empty() const - { - return _begin == _end; - } + case xpath_type_string: + return static_cast(this)->name; - PUGI__FN const xpath_node& xpath_node_set::operator[](size_t index) const - { - assert(index < size()); - return _begin[index]; - } - - PUGI__FN xpath_node_set::const_iterator xpath_node_set::begin() const - { - return _begin; - } - - PUGI__FN xpath_node_set::const_iterator xpath_node_set::end() const - { - return _end; - } - - PUGI__FN void xpath_node_set::sort(bool reverse) - { - _type = impl::xpath_sort(_begin, _end, _type, reverse); - } - - PUGI__FN xpath_node xpath_node_set::first() const - { - return impl::xpath_first(_begin, _end, _type); - } - - PUGI__FN xpath_parse_result::xpath_parse_result(): error("Internal error"), offset(0) - { - } - - PUGI__FN xpath_parse_result::operator bool() const - { - return error == 0; - } - - PUGI__FN const char* xpath_parse_result::description() const - { - return error ? error : "No error"; - } - - PUGI__FN xpath_variable::xpath_variable(xpath_value_type type_): _type(type_), _next(0) - { - } - - PUGI__FN const char_t* xpath_variable::name() const - { - switch (_type) - { - case xpath_type_node_set: - return static_cast(this)->name; - - case xpath_type_number: - return static_cast(this)->name; - - case xpath_type_string: - return static_cast(this)->name; - - case xpath_type_boolean: - return static_cast(this)->name; - - default: - assert(false && "Invalid variable type"); // unreachable - return 0; - } - } - - PUGI__FN xpath_value_type xpath_variable::type() const - { - return _type; - } - - PUGI__FN bool xpath_variable::get_boolean() const - { - return (_type == xpath_type_boolean) ? static_cast(this)->value : false; - } - - PUGI__FN double xpath_variable::get_number() const - { - return (_type == xpath_type_number) ? static_cast(this)->value : impl::gen_nan(); - } - - PUGI__FN const char_t* xpath_variable::get_string() const - { - const char_t* value = (_type == xpath_type_string) ? static_cast(this)->value : 0; - return value ? value : PUGIXML_TEXT(""); - } - - PUGI__FN const xpath_node_set& xpath_variable::get_node_set() const - { - return (_type == xpath_type_node_set) ? static_cast(this)->value : impl::dummy_node_set; - } - - PUGI__FN bool xpath_variable::set(bool value) - { - if (_type != xpath_type_boolean) return false; - - static_cast(this)->value = value; - return true; - } - - PUGI__FN bool xpath_variable::set(double value) - { - if (_type != xpath_type_number) return false; - - static_cast(this)->value = value; - return true; - } - - PUGI__FN bool xpath_variable::set(const char_t* value) - { - if (_type != xpath_type_string) return false; - - impl::xpath_variable_string* var = static_cast(this); - - // duplicate string - size_t size = (impl::strlength(value) + 1) * sizeof(char_t); - - char_t* copy = static_cast(impl::xml_memory::allocate(size)); - if (!copy) return false; - - memcpy(copy, value, size); - - // replace old string - if (var->value) impl::xml_memory::deallocate(var->value); - var->value = copy; - - return true; - } - - PUGI__FN bool xpath_variable::set(const xpath_node_set& value) - { - if (_type != xpath_type_node_set) return false; - - static_cast(this)->value = value; - return true; - } - - PUGI__FN xpath_variable_set::xpath_variable_set() - { - for (size_t i = 0; i < sizeof(_data) / sizeof(_data[0]); ++i) - _data[i] = 0; - } - - PUGI__FN xpath_variable_set::~xpath_variable_set() - { - for (size_t i = 0; i < sizeof(_data) / sizeof(_data[0]); ++i) - _destroy(_data[i]); - } - - PUGI__FN xpath_variable_set::xpath_variable_set(const xpath_variable_set& rhs) - { - for (size_t i = 0; i < sizeof(_data) / sizeof(_data[0]); ++i) - _data[i] = 0; - - _assign(rhs); - } - - PUGI__FN xpath_variable_set& xpath_variable_set::operator=(const xpath_variable_set& rhs) - { - if (this == &rhs) return *this; - - _assign(rhs); - - return *this; - } - -#ifdef PUGIXML_HAS_MOVE - PUGI__FN xpath_variable_set::xpath_variable_set(xpath_variable_set&& rhs) PUGIXML_NOEXCEPT - { - for (size_t i = 0; i < sizeof(_data) / sizeof(_data[0]); ++i) - { - _data[i] = rhs._data[i]; - rhs._data[i] = 0; - } - } - - PUGI__FN xpath_variable_set& xpath_variable_set::operator=(xpath_variable_set&& rhs) PUGIXML_NOEXCEPT - { - for (size_t i = 0; i < sizeof(_data) / sizeof(_data[0]); ++i) - { - _destroy(_data[i]); - - _data[i] = rhs._data[i]; - rhs._data[i] = 0; - } - - return *this; - } -#endif - - PUGI__FN void xpath_variable_set::_assign(const xpath_variable_set& rhs) - { - xpath_variable_set temp; - - for (size_t i = 0; i < sizeof(_data) / sizeof(_data[0]); ++i) - if (rhs._data[i] && !_clone(rhs._data[i], &temp._data[i])) - return; - - _swap(temp); - } - - PUGI__FN void xpath_variable_set::_swap(xpath_variable_set& rhs) - { - for (size_t i = 0; i < sizeof(_data) / sizeof(_data[0]); ++i) - { - xpath_variable* chain = _data[i]; - - _data[i] = rhs._data[i]; - rhs._data[i] = chain; - } - } - - PUGI__FN xpath_variable* xpath_variable_set::_find(const char_t* name) const - { - const size_t hash_size = sizeof(_data) / sizeof(_data[0]); - size_t hash = impl::hash_string(name) % hash_size; - - // look for existing variable - for (xpath_variable* var = _data[hash]; var; var = var->_next) - if (impl::strequal(var->name(), name)) - return var; + case xpath_type_boolean: + return static_cast(this)->name; + default: + assert(false && "Invalid variable type"); // unreachable return 0; } - - PUGI__FN bool xpath_variable_set::_clone(xpath_variable* var, xpath_variable** out_result) - { - xpath_variable* last = 0; - - while (var) - { - // allocate storage for new variable - xpath_variable* nvar = impl::new_xpath_variable(var->_type, var->name()); - if (!nvar) return false; - - // link the variable to the result immediately to handle failures gracefully - if (last) - last->_next = nvar; - else - *out_result = nvar; - - last = nvar; - - // copy the value; this can fail due to out-of-memory conditions - if (!impl::copy_xpath_variable(nvar, var)) return false; - - var = var->_next; - } - - return true; - } - - PUGI__FN void xpath_variable_set::_destroy(xpath_variable* var) - { - while (var) - { - xpath_variable* next = var->_next; - - impl::delete_xpath_variable(var->_type, var); - - var = next; - } - } - - PUGI__FN xpath_variable* xpath_variable_set::add(const char_t* name, xpath_value_type type) - { - const size_t hash_size = sizeof(_data) / sizeof(_data[0]); - size_t hash = impl::hash_string(name) % hash_size; - - // look for existing variable - for (xpath_variable* var = _data[hash]; var; var = var->_next) - if (impl::strequal(var->name(), name)) - return var->type() == type ? var : 0; - - // add new variable - xpath_variable* result = impl::new_xpath_variable(type, name); - - if (result) - { - result->_next = _data[hash]; - - _data[hash] = result; - } - - return result; - } - - PUGI__FN bool xpath_variable_set::set(const char_t* name, bool value) - { - xpath_variable* var = add(name, xpath_type_boolean); - return var ? var->set(value) : false; - } - - PUGI__FN bool xpath_variable_set::set(const char_t* name, double value) - { - xpath_variable* var = add(name, xpath_type_number); - return var ? var->set(value) : false; - } - - PUGI__FN bool xpath_variable_set::set(const char_t* name, const char_t* value) - { - xpath_variable* var = add(name, xpath_type_string); - return var ? var->set(value) : false; - } - - PUGI__FN bool xpath_variable_set::set(const char_t* name, const xpath_node_set& value) - { - xpath_variable* var = add(name, xpath_type_node_set); - return var ? var->set(value) : false; - } - - PUGI__FN xpath_variable* xpath_variable_set::get(const char_t* name) - { - return _find(name); - } - - PUGI__FN const xpath_variable* xpath_variable_set::get(const char_t* name) const - { - return _find(name); - } - - PUGI__FN xpath_query::xpath_query(const char_t* query, xpath_variable_set* variables): _impl(0) - { - impl::xpath_query_impl* qimpl = impl::xpath_query_impl::create(); - - if (!qimpl) - { - #ifdef PUGIXML_NO_EXCEPTIONS - _result.error = "Out of memory"; - #else - throw std::bad_alloc(); - #endif - } - else - { - using impl::auto_deleter; // MSVC7 workaround - auto_deleter impl(qimpl, impl::xpath_query_impl::destroy); - - qimpl->root = impl::xpath_parser::parse(query, variables, &qimpl->alloc, &_result); - - if (qimpl->root) - { - qimpl->root->optimize(&qimpl->alloc); - - _impl = impl.release(); - _result.error = 0; - } - else - { - #ifdef PUGIXML_NO_EXCEPTIONS - if (qimpl->oom) _result.error = "Out of memory"; - #else - if (qimpl->oom) throw std::bad_alloc(); - throw xpath_exception(_result); - #endif - } - } - } - - PUGI__FN xpath_query::xpath_query(): _impl(0) - { - } - - PUGI__FN xpath_query::~xpath_query() - { - if (_impl) - impl::xpath_query_impl::destroy(static_cast(_impl)); - } - -#ifdef PUGIXML_HAS_MOVE - PUGI__FN xpath_query::xpath_query(xpath_query&& rhs) PUGIXML_NOEXCEPT - { - _impl = rhs._impl; - _result = rhs._result; - rhs._impl = 0; - rhs._result = xpath_parse_result(); - } - - PUGI__FN xpath_query& xpath_query::operator=(xpath_query&& rhs) PUGIXML_NOEXCEPT - { - if (this == &rhs) return *this; - - if (_impl) - impl::xpath_query_impl::destroy(static_cast(_impl)); - - _impl = rhs._impl; - _result = rhs._result; - rhs._impl = 0; - rhs._result = xpath_parse_result(); - - return *this; - } -#endif - - PUGI__FN xpath_value_type xpath_query::return_type() const - { - if (!_impl) return xpath_type_none; - - return static_cast(_impl)->root->rettype(); - } - - PUGI__FN bool xpath_query::evaluate_boolean(const xpath_node& n) const - { - if (!_impl) return false; - - impl::xpath_context c(n, 1, 1); - impl::xpath_stack_data sd; - - bool r = static_cast(_impl)->root->eval_boolean(c, sd.stack); - - if (sd.oom) - { - #ifdef PUGIXML_NO_EXCEPTIONS - return false; - #else - throw std::bad_alloc(); - #endif - } - - return r; - } - - PUGI__FN double xpath_query::evaluate_number(const xpath_node& n) const - { - if (!_impl) return impl::gen_nan(); - - impl::xpath_context c(n, 1, 1); - impl::xpath_stack_data sd; - - double r = static_cast(_impl)->root->eval_number(c, sd.stack); - - if (sd.oom) - { - #ifdef PUGIXML_NO_EXCEPTIONS - return impl::gen_nan(); - #else - throw std::bad_alloc(); - #endif - } - - return r; - } - -#ifndef PUGIXML_NO_STL - PUGI__FN string_t xpath_query::evaluate_string(const xpath_node& n) const - { - if (!_impl) return string_t(); - - impl::xpath_context c(n, 1, 1); - impl::xpath_stack_data sd; - - impl::xpath_string r = static_cast(_impl)->root->eval_string(c, sd.stack); - - if (sd.oom) - { - #ifdef PUGIXML_NO_EXCEPTIONS - return string_t(); - #else - throw std::bad_alloc(); - #endif - } - - return string_t(r.c_str(), r.length()); - } -#endif - - PUGI__FN size_t xpath_query::evaluate_string(char_t* buffer, size_t capacity, const xpath_node& n) const - { - impl::xpath_context c(n, 1, 1); - impl::xpath_stack_data sd; - - impl::xpath_string r = _impl ? static_cast(_impl)->root->eval_string(c, sd.stack) : impl::xpath_string(); - - if (sd.oom) - { - #ifdef PUGIXML_NO_EXCEPTIONS - r = impl::xpath_string(); - #else - throw std::bad_alloc(); - #endif - } - - size_t full_size = r.length() + 1; - - if (capacity > 0) - { - size_t size = (full_size < capacity) ? full_size : capacity; - assert(size > 0); - - memcpy(buffer, r.c_str(), (size - 1) * sizeof(char_t)); - buffer[size - 1] = 0; - } - - return full_size; - } - - PUGI__FN xpath_node_set xpath_query::evaluate_node_set(const xpath_node& n) const - { - impl::xpath_ast_node* root = impl::evaluate_node_set_prepare(static_cast(_impl)); - if (!root) return xpath_node_set(); - - impl::xpath_context c(n, 1, 1); - impl::xpath_stack_data sd; - - impl::xpath_node_set_raw r = root->eval_node_set(c, sd.stack, impl::nodeset_eval_all); - - if (sd.oom) - { - #ifdef PUGIXML_NO_EXCEPTIONS - return xpath_node_set(); - #else - throw std::bad_alloc(); - #endif - } - - return xpath_node_set(r.begin(), r.end(), r.type()); - } - - PUGI__FN xpath_node xpath_query::evaluate_node(const xpath_node& n) const - { - impl::xpath_ast_node* root = impl::evaluate_node_set_prepare(static_cast(_impl)); - if (!root) return xpath_node(); - - impl::xpath_context c(n, 1, 1); - impl::xpath_stack_data sd; - - impl::xpath_node_set_raw r = root->eval_node_set(c, sd.stack, impl::nodeset_eval_first); - - if (sd.oom) - { - #ifdef PUGIXML_NO_EXCEPTIONS - return xpath_node(); - #else - throw std::bad_alloc(); - #endif - } - - return r.first(); - } - - PUGI__FN const xpath_parse_result& xpath_query::result() const - { - return _result; - } - - PUGI__FN static void unspecified_bool_xpath_query(xpath_query***) - { - } - - PUGI__FN xpath_query::operator xpath_query::unspecified_bool_type() const - { - return _impl ? unspecified_bool_xpath_query : 0; - } - - PUGI__FN bool xpath_query::operator!() const - { - return !_impl; - } - - PUGI__FN xpath_node xml_node::select_node(const char_t* query, xpath_variable_set* variables) const - { - xpath_query q(query, variables); - return q.evaluate_node(*this); - } - - PUGI__FN xpath_node xml_node::select_node(const xpath_query& query) const - { - return query.evaluate_node(*this); - } - - PUGI__FN xpath_node_set xml_node::select_nodes(const char_t* query, xpath_variable_set* variables) const - { - xpath_query q(query, variables); - return q.evaluate_node_set(*this); - } - - PUGI__FN xpath_node_set xml_node::select_nodes(const xpath_query& query) const - { - return query.evaluate_node_set(*this); - } - - PUGI__FN xpath_node xml_node::select_single_node(const char_t* query, xpath_variable_set* variables) const - { - xpath_query q(query, variables); - return q.evaluate_node(*this); - } - - PUGI__FN xpath_node xml_node::select_single_node(const xpath_query& query) const - { - return query.evaluate_node(*this); - } } +PUGI__FN xpath_value_type xpath_variable::type() const +{ + return _type; +} + +PUGI__FN bool xpath_variable::get_boolean() const +{ + return (_type == xpath_type_boolean) ? static_cast(this)->value : false; +} + +PUGI__FN double xpath_variable::get_number() const +{ + return (_type == xpath_type_number) ? static_cast(this)->value : impl::gen_nan(); +} + +PUGI__FN const char_t* xpath_variable::get_string() const +{ + const char_t* value = (_type == xpath_type_string) ? static_cast(this)->value : 0; + return value ? value : PUGIXML_TEXT(""); +} + +PUGI__FN const xpath_node_set& xpath_variable::get_node_set() const +{ + return (_type == xpath_type_node_set) ? static_cast(this)->value : impl::dummy_node_set; +} + +PUGI__FN bool xpath_variable::set(bool value) +{ + if (_type != xpath_type_boolean) + return false; + + static_cast(this)->value = value; + return true; +} + +PUGI__FN bool xpath_variable::set(double value) +{ + if (_type != xpath_type_number) + return false; + + static_cast(this)->value = value; + return true; +} + +PUGI__FN bool xpath_variable::set(const char_t* value) +{ + if (_type != xpath_type_string) + return false; + + impl::xpath_variable_string* var = static_cast(this); + + // duplicate string + size_t size = (impl::strlength(value) + 1) * sizeof(char_t); + + char_t* copy = static_cast(impl::xml_memory::allocate(size)); + if (!copy) + return false; + + memcpy(copy, value, size); + + // replace old string + if (var->value) + impl::xml_memory::deallocate(var->value); + var->value = copy; + + return true; +} + +PUGI__FN bool xpath_variable::set(const xpath_node_set& value) +{ + if (_type != xpath_type_node_set) + return false; + + static_cast(this)->value = value; + return true; +} + +PUGI__FN xpath_variable_set::xpath_variable_set() +{ + for (size_t i = 0; i < sizeof(_data) / sizeof(_data[0]); ++i) + _data[i] = 0; +} + +PUGI__FN xpath_variable_set::~xpath_variable_set() +{ + for (size_t i = 0; i < sizeof(_data) / sizeof(_data[0]); ++i) + _destroy(_data[i]); +} + +PUGI__FN xpath_variable_set::xpath_variable_set(const xpath_variable_set& rhs) +{ + for (size_t i = 0; i < sizeof(_data) / sizeof(_data[0]); ++i) + _data[i] = 0; + + _assign(rhs); +} + +PUGI__FN xpath_variable_set& xpath_variable_set::operator=(const xpath_variable_set& rhs) +{ + if (this == &rhs) + return *this; + + _assign(rhs); + + return *this; +} + +#ifdef PUGIXML_HAS_MOVE +PUGI__FN xpath_variable_set::xpath_variable_set(xpath_variable_set&& rhs) PUGIXML_NOEXCEPT +{ + for (size_t i = 0; i < sizeof(_data) / sizeof(_data[0]); ++i) + { + _data[i] = rhs._data[i]; + rhs._data[i] = 0; + } +} + +PUGI__FN xpath_variable_set& xpath_variable_set::operator=(xpath_variable_set&& rhs) PUGIXML_NOEXCEPT +{ + for (size_t i = 0; i < sizeof(_data) / sizeof(_data[0]); ++i) + { + _destroy(_data[i]); + + _data[i] = rhs._data[i]; + rhs._data[i] = 0; + } + + return *this; +} +#endif + +PUGI__FN void xpath_variable_set::_assign(const xpath_variable_set& rhs) +{ + xpath_variable_set temp; + + for (size_t i = 0; i < sizeof(_data) / sizeof(_data[0]); ++i) + if (rhs._data[i] && !_clone(rhs._data[i], &temp._data[i])) + return; + + _swap(temp); +} + +PUGI__FN void xpath_variable_set::_swap(xpath_variable_set& rhs) +{ + for (size_t i = 0; i < sizeof(_data) / sizeof(_data[0]); ++i) + { + xpath_variable* chain = _data[i]; + + _data[i] = rhs._data[i]; + rhs._data[i] = chain; + } +} + +PUGI__FN xpath_variable* xpath_variable_set::_find(const char_t* name) const +{ + const size_t hash_size = sizeof(_data) / sizeof(_data[0]); + size_t hash = impl::hash_string(name) % hash_size; + + // look for existing variable + for (xpath_variable* var = _data[hash]; var; var = var->_next) + if (impl::strequal(var->name(), name)) + return var; + + return 0; +} + +PUGI__FN bool xpath_variable_set::_clone(xpath_variable* var, xpath_variable** out_result) +{ + xpath_variable* last = 0; + + while (var) + { + // allocate storage for new variable + xpath_variable* nvar = impl::new_xpath_variable(var->_type, var->name()); + if (!nvar) + return false; + + // link the variable to the result immediately to handle failures gracefully + if (last) + last->_next = nvar; + else + *out_result = nvar; + + last = nvar; + + // copy the value; this can fail due to out-of-memory conditions + if (!impl::copy_xpath_variable(nvar, var)) + return false; + + var = var->_next; + } + + return true; +} + +PUGI__FN void xpath_variable_set::_destroy(xpath_variable* var) +{ + while (var) + { + xpath_variable* next = var->_next; + + impl::delete_xpath_variable(var->_type, var); + + var = next; + } +} + +PUGI__FN xpath_variable* xpath_variable_set::add(const char_t* name, xpath_value_type type) +{ + const size_t hash_size = sizeof(_data) / sizeof(_data[0]); + size_t hash = impl::hash_string(name) % hash_size; + + // look for existing variable + for (xpath_variable* var = _data[hash]; var; var = var->_next) + if (impl::strequal(var->name(), name)) + return var->type() == type ? var : 0; + + // add new variable + xpath_variable* result = impl::new_xpath_variable(type, name); + + if (result) + { + result->_next = _data[hash]; + + _data[hash] = result; + } + + return result; +} + +PUGI__FN bool xpath_variable_set::set(const char_t* name, bool value) +{ + xpath_variable* var = add(name, xpath_type_boolean); + return var ? var->set(value) : false; +} + +PUGI__FN bool xpath_variable_set::set(const char_t* name, double value) +{ + xpath_variable* var = add(name, xpath_type_number); + return var ? var->set(value) : false; +} + +PUGI__FN bool xpath_variable_set::set(const char_t* name, const char_t* value) +{ + xpath_variable* var = add(name, xpath_type_string); + return var ? var->set(value) : false; +} + +PUGI__FN bool xpath_variable_set::set(const char_t* name, const xpath_node_set& value) +{ + xpath_variable* var = add(name, xpath_type_node_set); + return var ? var->set(value) : false; +} + +PUGI__FN xpath_variable* xpath_variable_set::get(const char_t* name) +{ + return _find(name); +} + +PUGI__FN const xpath_variable* xpath_variable_set::get(const char_t* name) const +{ + return _find(name); +} + +PUGI__FN xpath_query::xpath_query(const char_t* query, xpath_variable_set* variables) + : _impl(0) +{ + impl::xpath_query_impl* qimpl = impl::xpath_query_impl::create(); + + if (!qimpl) + { +#ifdef PUGIXML_NO_EXCEPTIONS + _result.error = "Out of memory"; +#else + throw std::bad_alloc(); +#endif + } + else + { + using impl::auto_deleter; // MSVC7 workaround + auto_deleter impl(qimpl, impl::xpath_query_impl::destroy); + + qimpl->root = impl::xpath_parser::parse(query, variables, &qimpl->alloc, &_result); + + if (qimpl->root) + { + qimpl->root->optimize(&qimpl->alloc); + + _impl = impl.release(); + _result.error = 0; + } + else + { +#ifdef PUGIXML_NO_EXCEPTIONS + if (qimpl->oom) + _result.error = "Out of memory"; +#else + if (qimpl->oom) + throw std::bad_alloc(); + throw xpath_exception(_result); +#endif + } + } +} + +PUGI__FN xpath_query::xpath_query() + : _impl(0) +{ +} + +PUGI__FN xpath_query::~xpath_query() +{ + if (_impl) + impl::xpath_query_impl::destroy(static_cast(_impl)); +} + +#ifdef PUGIXML_HAS_MOVE +PUGI__FN xpath_query::xpath_query(xpath_query&& rhs) PUGIXML_NOEXCEPT +{ + _impl = rhs._impl; + _result = rhs._result; + rhs._impl = 0; + rhs._result = xpath_parse_result(); +} + +PUGI__FN xpath_query& xpath_query::operator=(xpath_query&& rhs) PUGIXML_NOEXCEPT +{ + if (this == &rhs) + return *this; + + if (_impl) + impl::xpath_query_impl::destroy(static_cast(_impl)); + + _impl = rhs._impl; + _result = rhs._result; + rhs._impl = 0; + rhs._result = xpath_parse_result(); + + return *this; +} +#endif + +PUGI__FN xpath_value_type xpath_query::return_type() const +{ + if (!_impl) + return xpath_type_none; + + return static_cast(_impl)->root->rettype(); +} + +PUGI__FN bool xpath_query::evaluate_boolean(const xpath_node& n) const +{ + if (!_impl) + return false; + + impl::xpath_context c(n, 1, 1); + impl::xpath_stack_data sd; + + bool r = static_cast(_impl)->root->eval_boolean(c, sd.stack); + + if (sd.oom) + { +#ifdef PUGIXML_NO_EXCEPTIONS + return false; +#else + throw std::bad_alloc(); +#endif + } + + return r; +} + +PUGI__FN double xpath_query::evaluate_number(const xpath_node& n) const +{ + if (!_impl) + return impl::gen_nan(); + + impl::xpath_context c(n, 1, 1); + impl::xpath_stack_data sd; + + double r = static_cast(_impl)->root->eval_number(c, sd.stack); + + if (sd.oom) + { +#ifdef PUGIXML_NO_EXCEPTIONS + return impl::gen_nan(); +#else + throw std::bad_alloc(); +#endif + } + + return r; +} + +#ifndef PUGIXML_NO_STL +PUGI__FN string_t xpath_query::evaluate_string(const xpath_node& n) const +{ + if (!_impl) + return string_t(); + + impl::xpath_context c(n, 1, 1); + impl::xpath_stack_data sd; + + impl::xpath_string r = static_cast(_impl)->root->eval_string(c, sd.stack); + + if (sd.oom) + { +#ifdef PUGIXML_NO_EXCEPTIONS + return string_t(); +#else + throw std::bad_alloc(); +#endif + } + + return string_t(r.c_str(), r.length()); +} +#endif + +PUGI__FN size_t xpath_query::evaluate_string(char_t* buffer, size_t capacity, const xpath_node& n) const +{ + impl::xpath_context c(n, 1, 1); + impl::xpath_stack_data sd; + + impl::xpath_string r = _impl ? static_cast(_impl)->root->eval_string(c, sd.stack) : impl::xpath_string(); + + if (sd.oom) + { +#ifdef PUGIXML_NO_EXCEPTIONS + r = impl::xpath_string(); +#else + throw std::bad_alloc(); +#endif + } + + size_t full_size = r.length() + 1; + + if (capacity > 0) + { + size_t size = (full_size < capacity) ? full_size : capacity; + assert(size > 0); + + memcpy(buffer, r.c_str(), (size - 1) * sizeof(char_t)); + buffer[size - 1] = 0; + } + + return full_size; +} + +PUGI__FN xpath_node_set xpath_query::evaluate_node_set(const xpath_node& n) const +{ + impl::xpath_ast_node* root = impl::evaluate_node_set_prepare(static_cast(_impl)); + if (!root) + return xpath_node_set(); + + impl::xpath_context c(n, 1, 1); + impl::xpath_stack_data sd; + + impl::xpath_node_set_raw r = root->eval_node_set(c, sd.stack, impl::nodeset_eval_all); + + if (sd.oom) + { +#ifdef PUGIXML_NO_EXCEPTIONS + return xpath_node_set(); +#else + throw std::bad_alloc(); +#endif + } + + return xpath_node_set(r.begin(), r.end(), r.type()); +} + +PUGI__FN xpath_node xpath_query::evaluate_node(const xpath_node& n) const +{ + impl::xpath_ast_node* root = impl::evaluate_node_set_prepare(static_cast(_impl)); + if (!root) + return xpath_node(); + + impl::xpath_context c(n, 1, 1); + impl::xpath_stack_data sd; + + impl::xpath_node_set_raw r = root->eval_node_set(c, sd.stack, impl::nodeset_eval_first); + + if (sd.oom) + { +#ifdef PUGIXML_NO_EXCEPTIONS + return xpath_node(); +#else + throw std::bad_alloc(); +#endif + } + + return r.first(); +} + +PUGI__FN const xpath_parse_result& xpath_query::result() const +{ + return _result; +} + +PUGI__FN static void unspecified_bool_xpath_query(xpath_query***) +{ +} + +PUGI__FN xpath_query::operator xpath_query::unspecified_bool_type() const +{ + return _impl ? unspecified_bool_xpath_query : 0; +} + +PUGI__FN bool xpath_query::operator!() const +{ + return !_impl; +} + +PUGI__FN xpath_node xml_node::select_node(const char_t* query, xpath_variable_set* variables) const +{ + xpath_query q(query, variables); + return q.evaluate_node(*this); +} + +PUGI__FN xpath_node xml_node::select_node(const xpath_query& query) const +{ + return query.evaluate_node(*this); +} + +PUGI__FN xpath_node_set xml_node::select_nodes(const char_t* query, xpath_variable_set* variables) const +{ + xpath_query q(query, variables); + return q.evaluate_node_set(*this); +} + +PUGI__FN xpath_node_set xml_node::select_nodes(const xpath_query& query) const +{ + return query.evaluate_node_set(*this); +} + +PUGI__FN xpath_node xml_node::select_single_node(const char_t* query, xpath_variable_set* variables) const +{ + xpath_query q(query, variables); + return q.evaluate_node(*this); +} + +PUGI__FN xpath_node xml_node::select_single_node(const xpath_query& query) const +{ + return query.evaluate_node(*this); +} +} // namespace pugi + #endif #ifdef __BORLANDC__ -# pragma option pop +#pragma option pop #endif // Intel C++ does not properly keep warning state for function templates, // so popping warning state at the end of translation unit leads to warnings in the middle. #if defined(_MSC_VER) && !defined(__INTEL_COMPILER) -# pragma warning(pop) +#pragma warning(pop) #endif #if defined(_MSC_VER) && defined(__c2__) -# pragma clang diagnostic pop +#pragma clang diagnostic pop #endif // Undefine all local macros (makes sure we're not leaking macros in header-only mode) diff --git a/src/pugixml.hpp b/src/pugixml.hpp index c0dacb7..784f8aa 100644 --- a/src/pugixml.hpp +++ b/src/pugixml.hpp @@ -13,7 +13,7 @@ #ifndef PUGIXML_VERSION // Define version macro; evaluates to major * 100 + minor * 10 + patch so that it's safe to use in less-than comparisons -# define PUGIXML_VERSION 190 +#define PUGIXML_VERSION 190 #endif // Include user configuration file (this can define various configuration macros) @@ -27,1412 +27,1426 @@ // Include exception header for XPath #if !defined(PUGIXML_NO_XPATH) && !defined(PUGIXML_NO_EXCEPTIONS) -# include +#include #endif // Include STL headers #ifndef PUGIXML_NO_STL -# include -# include -# include +#include +#include +#include #endif // Macro for deprecated features #ifndef PUGIXML_DEPRECATED -# if defined(__GNUC__) -# define PUGIXML_DEPRECATED __attribute__((deprecated)) -# elif defined(_MSC_VER) && _MSC_VER >= 1300 -# define PUGIXML_DEPRECATED __declspec(deprecated) -# else -# define PUGIXML_DEPRECATED -# endif +#if defined(__GNUC__) +#define PUGIXML_DEPRECATED __attribute__((deprecated)) +#elif defined(_MSC_VER) && _MSC_VER >= 1300 +#define PUGIXML_DEPRECATED __declspec(deprecated) +#else +#define PUGIXML_DEPRECATED +#endif #endif // If no API is defined, assume default #ifndef PUGIXML_API -# define PUGIXML_API +#define PUGIXML_API #endif // If no API for classes is defined, assume default #ifndef PUGIXML_CLASS -# define PUGIXML_CLASS PUGIXML_API +#define PUGIXML_CLASS PUGIXML_API #endif // If no API for functions is defined, assume default #ifndef PUGIXML_FUNCTION -# define PUGIXML_FUNCTION PUGIXML_API +#define PUGIXML_FUNCTION PUGIXML_API #endif // If the platform is known to have long long support, enable long long functions #ifndef PUGIXML_HAS_LONG_LONG -# if __cplusplus >= 201103 -# define PUGIXML_HAS_LONG_LONG -# elif defined(_MSC_VER) && _MSC_VER >= 1400 -# define PUGIXML_HAS_LONG_LONG -# endif +#if __cplusplus >= 201103 +#define PUGIXML_HAS_LONG_LONG +#elif defined(_MSC_VER) && _MSC_VER >= 1400 +#define PUGIXML_HAS_LONG_LONG +#endif #endif // If the platform is known to have move semantics support, compile move ctor/operator implementation #ifndef PUGIXML_HAS_MOVE -# if __cplusplus >= 201103 -# define PUGIXML_HAS_MOVE -# elif defined(_MSC_VER) && _MSC_VER >= 1600 -# define PUGIXML_HAS_MOVE -# endif +#if __cplusplus >= 201103 +#define PUGIXML_HAS_MOVE +#elif defined(_MSC_VER) && _MSC_VER >= 1600 +#define PUGIXML_HAS_MOVE +#endif #endif // If C++ is 2011 or higher, add 'noexcept' specifiers #ifndef PUGIXML_NOEXCEPT -# if __cplusplus >= 201103 -# define PUGIXML_NOEXCEPT noexcept -# elif defined(_MSC_VER) && _MSC_VER >= 1900 -# define PUGIXML_NOEXCEPT noexcept -# else -# define PUGIXML_NOEXCEPT -# endif +#if __cplusplus >= 201103 +#define PUGIXML_NOEXCEPT noexcept +#elif defined(_MSC_VER) && _MSC_VER >= 1900 +#define PUGIXML_NOEXCEPT noexcept +#else +#define PUGIXML_NOEXCEPT +#endif #endif // Some functions can not be noexcept in compact mode #ifdef PUGIXML_COMPACT -# define PUGIXML_NOEXCEPT_IF_NOT_COMPACT +#define PUGIXML_NOEXCEPT_IF_NOT_COMPACT #else -# define PUGIXML_NOEXCEPT_IF_NOT_COMPACT PUGIXML_NOEXCEPT +#define PUGIXML_NOEXCEPT_IF_NOT_COMPACT PUGIXML_NOEXCEPT #endif // If C++ is 2011 or higher, add 'override' qualifiers #ifndef PUGIXML_OVERRIDE -# if __cplusplus >= 201103 -# define PUGIXML_OVERRIDE override -# elif defined(_MSC_VER) && _MSC_VER >= 1700 -# define PUGIXML_OVERRIDE override -# else -# define PUGIXML_OVERRIDE -# endif +#if __cplusplus >= 201103 +#define PUGIXML_OVERRIDE override +#elif defined(_MSC_VER) && _MSC_VER >= 1700 +#define PUGIXML_OVERRIDE override +#else +#define PUGIXML_OVERRIDE +#endif #endif // Character interface macros #ifdef PUGIXML_WCHAR_MODE -# define PUGIXML_TEXT(t) L ## t -# define PUGIXML_CHAR wchar_t +#define PUGIXML_TEXT(t) L##t +#define PUGIXML_CHAR wchar_t #else -# define PUGIXML_TEXT(t) t -# define PUGIXML_CHAR char +#define PUGIXML_TEXT(t) t +#define PUGIXML_CHAR char #endif namespace pugi { - // Character type used for all internal storage and operations; depends on PUGIXML_WCHAR_MODE - typedef PUGIXML_CHAR char_t; +// Character type used for all internal storage and operations; depends on PUGIXML_WCHAR_MODE +typedef PUGIXML_CHAR char_t; #ifndef PUGIXML_NO_STL - // String type used for operations that work with STL string; depends on PUGIXML_WCHAR_MODE - typedef std::basic_string, std::allocator > string_t; +// String type used for operations that work with STL string; depends on PUGIXML_WCHAR_MODE +typedef std::basic_string, std::allocator> string_t; #endif -} +} // namespace pugi // The PugiXML namespace namespace pugi { - // Tree node types - enum xml_node_type - { - node_null, // Empty (null) node handle - node_document, // A document tree's absolute root - node_element, // Element tag, i.e. '' - node_pcdata, // Plain character data, i.e. 'text' - node_cdata, // Character data, i.e. '' - node_comment, // Comment tag, i.e. '' - node_pi, // Processing instruction, i.e. '' - node_declaration, // Document declaration, i.e. '' - node_doctype // Document type declaration, i.e. '' - }; - - // Parsing options - - // Minimal parsing mode (equivalent to turning all other flags off). - // Only elements and PCDATA sections are added to the DOM tree, no text conversions are performed. - const unsigned int parse_minimal = 0x0000; - - // This flag determines if processing instructions (node_pi) are added to the DOM tree. This flag is off by default. - const unsigned int parse_pi = 0x0001; +// Tree node types +enum xml_node_type +{ + node_null, // Empty (null) node handle + node_document, // A document tree's absolute root + node_element, // Element tag, i.e. '' + node_pcdata, // Plain character data, i.e. 'text' + node_cdata, // Character data, i.e. '' + node_comment, // Comment tag, i.e. '' + node_pi, // Processing instruction, i.e. '' + node_declaration, // Document declaration, i.e. '' + node_doctype // Document type declaration, i.e. '' +}; + +// Parsing options + +// Minimal parsing mode (equivalent to turning all other flags off). +// Only elements and PCDATA sections are added to the DOM tree, no text conversions are performed. +const unsigned int parse_minimal = 0x0000; + +// This flag determines if processing instructions (node_pi) are added to the DOM tree. This flag is off by default. +const unsigned int parse_pi = 0x0001; - // This flag determines if comments (node_comment) are added to the DOM tree. This flag is off by default. - const unsigned int parse_comments = 0x0002; - - // This flag determines if CDATA sections (node_cdata) are added to the DOM tree. This flag is on by default. - const unsigned int parse_cdata = 0x0004; - - // This flag determines if plain character data (node_pcdata) that consist only of whitespace are added to the DOM tree. - // This flag is off by default; turning it on usually results in slower parsing and more memory consumption. - const unsigned int parse_ws_pcdata = 0x0008; +// This flag determines if comments (node_comment) are added to the DOM tree. This flag is off by default. +const unsigned int parse_comments = 0x0002; + +// This flag determines if CDATA sections (node_cdata) are added to the DOM tree. This flag is on by default. +const unsigned int parse_cdata = 0x0004; + +// This flag determines if plain character data (node_pcdata) that consist only of whitespace are added to the DOM tree. +// This flag is off by default; turning it on usually results in slower parsing and more memory consumption. +const unsigned int parse_ws_pcdata = 0x0008; - // This flag determines if character and entity references are expanded during parsing. This flag is on by default. - const unsigned int parse_escapes = 0x0010; +// This flag determines if character and entity references are expanded during parsing. This flag is on by default. +const unsigned int parse_escapes = 0x0010; - // This flag determines if EOL characters are normalized (converted to #xA) during parsing. This flag is on by default. - const unsigned int parse_eol = 0x0020; +// This flag determines if EOL characters are normalized (converted to #xA) during parsing. This flag is on by default. +const unsigned int parse_eol = 0x0020; - // This flag determines if attribute values are normalized using CDATA normalization rules during parsing. This flag is on by default. - const unsigned int parse_wconv_attribute = 0x0040; +// This flag determines if attribute values are normalized using CDATA normalization rules during parsing. This flag is on by default. +const unsigned int parse_wconv_attribute = 0x0040; - // This flag determines if attribute values are normalized using NMTOKENS normalization rules during parsing. This flag is off by default. - const unsigned int parse_wnorm_attribute = 0x0080; +// This flag determines if attribute values are normalized using NMTOKENS normalization rules during parsing. This flag is off by default. +const unsigned int parse_wnorm_attribute = 0x0080; - // This flag determines if document declaration (node_declaration) is added to the DOM tree. This flag is off by default. - const unsigned int parse_declaration = 0x0100; +// This flag determines if document declaration (node_declaration) is added to the DOM tree. This flag is off by default. +const unsigned int parse_declaration = 0x0100; - // This flag determines if document type declaration (node_doctype) is added to the DOM tree. This flag is off by default. - const unsigned int parse_doctype = 0x0200; +// This flag determines if document type declaration (node_doctype) is added to the DOM tree. This flag is off by default. +const unsigned int parse_doctype = 0x0200; - // This flag determines if plain character data (node_pcdata) that is the only child of the parent node and that consists only - // of whitespace is added to the DOM tree. - // This flag is off by default; turning it on may result in slower parsing and more memory consumption. - const unsigned int parse_ws_pcdata_single = 0x0400; - - // This flag determines if leading and trailing whitespace is to be removed from plain character data. This flag is off by default. - const unsigned int parse_trim_pcdata = 0x0800; +// This flag determines if plain character data (node_pcdata) that is the only child of the parent node and that consists only +// of whitespace is added to the DOM tree. +// This flag is off by default; turning it on may result in slower parsing and more memory consumption. +const unsigned int parse_ws_pcdata_single = 0x0400; - // This flag determines if plain character data that does not have a parent node is added to the DOM tree, and if an empty document - // is a valid document. This flag is off by default. - const unsigned int parse_fragment = 0x1000; +// This flag determines if leading and trailing whitespace is to be removed from plain character data. This flag is off by default. +const unsigned int parse_trim_pcdata = 0x0800; - // This flag determines if plain character data is be stored in the parent element's value. This significantly changes the structure of - // the document; this flag is only recommended for parsing documents with many PCDATA nodes in memory-constrained environments. - // This flag is off by default. - const unsigned int parse_embed_pcdata = 0x2000; - - // The default parsing mode. - // Elements, PCDATA and CDATA sections are added to the DOM tree, character/reference entities are expanded, - // End-of-Line characters are normalized, attribute values are normalized using CDATA normalization rules. - const unsigned int parse_default = parse_cdata | parse_escapes | parse_wconv_attribute | parse_eol; - - // The full parsing mode. - // Nodes of all types are added to the DOM tree, character/reference entities are expanded, - // End-of-Line characters are normalized, attribute values are normalized using CDATA normalization rules. - const unsigned int parse_full = parse_default | parse_pi | parse_comments | parse_declaration | parse_doctype; - - // These flags determine the encoding of input data for XML document - enum xml_encoding - { - encoding_auto, // Auto-detect input encoding using BOM or < / class xml_object_range - { - public: - typedef It const_iterator; - typedef It iterator; - - xml_object_range(It b, It e): _begin(b), _end(e) - { - } - - It begin() const { return _begin; } - It end() const { return _end; } - - private: - It _begin, _end; - }; - - // Writer interface for node printing (see xml_node::print) - class PUGIXML_CLASS xml_writer - { - public: - virtual ~xml_writer() {} - - // Write memory chunk into stream/file/whatever - virtual void write(const void* data, size_t size) = 0; - }; - - // xml_writer implementation for FILE* - class PUGIXML_CLASS xml_writer_file: public xml_writer - { - public: - // Construct writer from a FILE* object; void* is used to avoid header dependencies on stdio - xml_writer_file(void* file); - - virtual void write(const void* data, size_t size) PUGIXML_OVERRIDE; - - private: - void* file; - }; - - #ifndef PUGIXML_NO_STL - // xml_writer implementation for streams - class PUGIXML_CLASS xml_writer_stream: public xml_writer - { - public: - // Construct writer from an output stream object - xml_writer_stream(std::basic_ostream >& stream); - xml_writer_stream(std::basic_ostream >& stream); - - virtual void write(const void* data, size_t size) PUGIXML_OVERRIDE; - - private: - std::basic_ostream >* narrow_stream; - std::basic_ostream >* wide_stream; - }; - #endif - - // A light-weight handle for manipulating attributes in DOM tree - class PUGIXML_CLASS xml_attribute - { - friend class xml_attribute_iterator; - friend class xml_node; - - private: - xml_attribute_struct* _attr; - - typedef void (*unspecified_bool_type)(xml_attribute***); - - public: - // Default constructor. Constructs an empty attribute. - xml_attribute(); - - // Constructs attribute from internal pointer - explicit xml_attribute(xml_attribute_struct* attr); - - // Safe bool conversion operator - operator unspecified_bool_type() const; - - // Borland C++ workaround - bool operator!() const; - - // Comparison operators (compares wrapped attribute pointers) - bool operator==(const xml_attribute& r) const; - bool operator!=(const xml_attribute& r) const; - bool operator<(const xml_attribute& r) const; - bool operator>(const xml_attribute& r) const; - bool operator<=(const xml_attribute& r) const; - bool operator>=(const xml_attribute& r) const; - - // Check if attribute is empty - bool empty() const; - - // Get attribute name/value, or "" if attribute is empty - const char_t* name() const; - const char_t* value() const; - - // Get attribute value, or the default value if attribute is empty - const char_t* as_string(const char_t* def = PUGIXML_TEXT("")) const; - - // Get attribute value as a number, or the default value if conversion did not succeed or attribute is empty - int as_int(int def = 0) const; - unsigned int as_uint(unsigned int def = 0) const; - double as_double(double def = 0) const; - float as_float(float def = 0) const; - - #ifdef PUGIXML_HAS_LONG_LONG - long long as_llong(long long def = 0) const; - unsigned long long as_ullong(unsigned long long def = 0) const; - #endif - - // Get attribute value as bool (returns true if first character is in '1tTyY' set), or the default value if attribute is empty - bool as_bool(bool def = false) const; - - // Set attribute name/value (returns false if attribute is empty or there is not enough memory) - bool set_name(const char_t* rhs); - bool set_value(const char_t* rhs); - - // Set attribute value with type conversion (numbers are converted to strings, boolean is converted to "true"/"false") - bool set_value(int rhs); - bool set_value(unsigned int rhs); - bool set_value(long rhs); - bool set_value(unsigned long rhs); - bool set_value(double rhs); - bool set_value(float rhs); - bool set_value(bool rhs); - - #ifdef PUGIXML_HAS_LONG_LONG - bool set_value(long long rhs); - bool set_value(unsigned long long rhs); - #endif - - // Set attribute value (equivalent to set_value without error checking) - xml_attribute& operator=(const char_t* rhs); - xml_attribute& operator=(int rhs); - xml_attribute& operator=(unsigned int rhs); - xml_attribute& operator=(long rhs); - xml_attribute& operator=(unsigned long rhs); - xml_attribute& operator=(double rhs); - xml_attribute& operator=(float rhs); - xml_attribute& operator=(bool rhs); - - #ifdef PUGIXML_HAS_LONG_LONG - xml_attribute& operator=(long long rhs); - xml_attribute& operator=(unsigned long long rhs); - #endif - - // Get next/previous attribute in the attribute list of the parent node - xml_attribute next_attribute() const; - xml_attribute previous_attribute() const; - - // Get hash value (unique for handles to the same object) - size_t hash_value() const; - - // Get internal pointer - xml_attribute_struct* internal_object() const; - }; - -#ifdef __BORLANDC__ - // Borland C++ workaround - bool PUGIXML_FUNCTION operator&&(const xml_attribute& lhs, bool rhs); - bool PUGIXML_FUNCTION operator||(const xml_attribute& lhs, bool rhs); -#endif - - // A light-weight handle for manipulating nodes in DOM tree - class PUGIXML_CLASS xml_node - { - friend class xml_attribute_iterator; - friend class xml_node_iterator; - friend class xml_named_node_iterator; - - protected: - xml_node_struct* _root; - - typedef void (*unspecified_bool_type)(xml_node***); - - public: - // Default constructor. Constructs an empty node. - xml_node(); - - // Constructs node from internal pointer - explicit xml_node(xml_node_struct* p); - - // Safe bool conversion operator - operator unspecified_bool_type() const; - - // Borland C++ workaround - bool operator!() const; - - // Comparison operators (compares wrapped node pointers) - bool operator==(const xml_node& r) const; - bool operator!=(const xml_node& r) const; - bool operator<(const xml_node& r) const; - bool operator>(const xml_node& r) const; - bool operator<=(const xml_node& r) const; - bool operator>=(const xml_node& r) const; - - // Check if node is empty. - bool empty() const; - - // Get node type - xml_node_type type() const; - - // Get node name, or "" if node is empty or it has no name - const char_t* name() const; - - // Get node value, or "" if node is empty or it has no value - // Note: For text node.value() does not return "text"! Use child_value() or text() methods to access text inside nodes. - const char_t* value() const; - - // Get attribute list - xml_attribute first_attribute() const; - xml_attribute last_attribute() const; - - // Get children list - xml_node first_child() const; - xml_node last_child() const; - - // Get next/previous sibling in the children list of the parent node - xml_node next_sibling() const; - xml_node previous_sibling() const; - - // Get parent node - xml_node parent() const; - - // Get root of DOM tree this node belongs to - xml_node root() const; - - // Get text object for the current node - xml_text text() const; - - // Get child, attribute or next/previous sibling with the specified name - xml_node child(const char_t* name) const; - xml_attribute attribute(const char_t* name) const; - xml_node next_sibling(const char_t* name) const; - xml_node previous_sibling(const char_t* name) const; - - // Get attribute, starting the search from a hint (and updating hint so that searching for a sequence of attributes is fast) - xml_attribute attribute(const char_t* name, xml_attribute& hint) const; - - // Get child value of current node; that is, value of the first child node of type PCDATA/CDATA - const char_t* child_value() const; - - // Get child value of child with specified name. Equivalent to child(name).child_value(). - const char_t* child_value(const char_t* name) const; - - // Set node name/value (returns false if node is empty, there is not enough memory, or node can not have name/value) - bool set_name(const char_t* rhs); - bool set_value(const char_t* rhs); - - // Add attribute with specified name. Returns added attribute, or empty attribute on errors. - xml_attribute append_attribute(const char_t* name); - xml_attribute prepend_attribute(const char_t* name); - xml_attribute insert_attribute_after(const char_t* name, const xml_attribute& attr); - xml_attribute insert_attribute_before(const char_t* name, const xml_attribute& attr); - - // Add a copy of the specified attribute. Returns added attribute, or empty attribute on errors. - xml_attribute append_copy(const xml_attribute& proto); - xml_attribute prepend_copy(const xml_attribute& proto); - xml_attribute insert_copy_after(const xml_attribute& proto, const xml_attribute& attr); - xml_attribute insert_copy_before(const xml_attribute& proto, const xml_attribute& attr); - - // Add child node with specified type. Returns added node, or empty node on errors. - xml_node append_child(xml_node_type type = node_element); - xml_node prepend_child(xml_node_type type = node_element); - xml_node insert_child_after(xml_node_type type, const xml_node& node); - xml_node insert_child_before(xml_node_type type, const xml_node& node); - - // Add child element with specified name. Returns added node, or empty node on errors. - xml_node append_child(const char_t* name); - xml_node prepend_child(const char_t* name); - xml_node insert_child_after(const char_t* name, const xml_node& node); - xml_node insert_child_before(const char_t* name, const xml_node& node); - - // Add a copy of the specified node as a child. Returns added node, or empty node on errors. - xml_node append_copy(const xml_node& proto); - xml_node prepend_copy(const xml_node& proto); - xml_node insert_copy_after(const xml_node& proto, const xml_node& node); - xml_node insert_copy_before(const xml_node& proto, const xml_node& node); +// Omit default XML declaration even if there is no declaration in the document. This flag is off by default. +const unsigned int format_no_declaration = 0x08; - // Move the specified node to become a child of this node. Returns moved node, or empty node on errors. - xml_node append_move(const xml_node& moved); - xml_node prepend_move(const xml_node& moved); - xml_node insert_move_after(const xml_node& moved, const xml_node& node); - xml_node insert_move_before(const xml_node& moved, const xml_node& node); - - // Remove specified attribute - bool remove_attribute(const xml_attribute& a); - bool remove_attribute(const char_t* name); - - // Remove specified child - bool remove_child(const xml_node& n); - bool remove_child(const char_t* name); - - // Parses buffer as an XML document fragment and appends all nodes as children of the current node. - // Copies/converts the buffer, so it may be deleted or changed after the function returns. - // Note: append_buffer allocates memory that has the lifetime of the owning document; removing the appended nodes does not immediately reclaim that memory. - xml_parse_result append_buffer(const void* contents, size_t size, unsigned int options = parse_default, xml_encoding encoding = encoding_auto); +// Don't escape attribute values and PCDATA contents. This flag is off by default. +const unsigned int format_no_escapes = 0x10; - // Find attribute using predicate. Returns first attribute for which predicate returned true. - template xml_attribute find_attribute(Predicate pred) const - { - if (!_root) return xml_attribute(); - - for (xml_attribute attrib = first_attribute(); attrib; attrib = attrib.next_attribute()) - if (pred(attrib)) - return attrib; - - return xml_attribute(); - } - - // Find child node using predicate. Returns first child for which predicate returned true. - template xml_node find_child(Predicate pred) const - { - if (!_root) return xml_node(); - - for (xml_node node = first_child(); node; node = node.next_sibling()) - if (pred(node)) - return node; - - return xml_node(); - } - - // Find node from subtree using predicate. Returns first node from subtree (depth-first), for which predicate returned true. - template xml_node find_node(Predicate pred) const - { - if (!_root) return xml_node(); +// Open file using text mode in xml_document::save_file. This enables special character (i.e. new-line) conversions on some systems. This flag is off by default. +const unsigned int format_save_file_text = 0x20; - xml_node cur = first_child(); +// Write every attribute on a new line with appropriate indentation. This flag is off by default. +const unsigned int format_indent_attributes = 0x40; - while (cur._root && cur._root != _root) - { - if (pred(cur)) return cur; +// Don't output empty element tags, instead writing an explicit start and end tag even if there are no children. This flag is off by default. +const unsigned int format_no_empty_element_tags = 0x80; - if (cur.first_child()) cur = cur.first_child(); - else if (cur.next_sibling()) cur = cur.next_sibling(); - else - { - while (!cur.next_sibling() && cur._root != _root) cur = cur.parent(); +// The default set of formatting flags. +// Nodes are indented depending on their depth in DOM tree, a default declaration is output if document has none. +const unsigned int format_default = format_indent; - if (cur._root != _root) cur = cur.next_sibling(); - } - } +// Forward declarations +struct xml_attribute_struct; +struct xml_node_struct; - return xml_node(); - } +class xml_node_iterator; +class xml_attribute_iterator; +class xml_named_node_iterator; - // Find child node by attribute name/value - xml_node find_child_by_attribute(const char_t* name, const char_t* attr_name, const char_t* attr_value) const; - xml_node find_child_by_attribute(const char_t* attr_name, const char_t* attr_value) const; +class xml_tree_walker; - #ifndef PUGIXML_NO_STL - // Get the absolute node path from root as a text string. - string_t path(char_t delimiter = '/') const; - #endif +struct xml_parse_result; - // Search for a node by path consisting of node names and . or .. elements. - xml_node first_element_by_path(const char_t* path, char_t delimiter = '/') const; +class xml_node; - // Recursively traverse subtree with xml_tree_walker - bool traverse(xml_tree_walker& walker); - - #ifndef PUGIXML_NO_XPATH - // Select single node by evaluating XPath query. Returns first node from the resulting node set. - xpath_node select_node(const char_t* query, xpath_variable_set* variables = 0) const; - xpath_node select_node(const xpath_query& query) const; - - // Select node set by evaluating XPath query - xpath_node_set select_nodes(const char_t* query, xpath_variable_set* variables = 0) const; - xpath_node_set select_nodes(const xpath_query& query) const; - - // (deprecated: use select_node instead) Select single node by evaluating XPath query. - PUGIXML_DEPRECATED xpath_node select_single_node(const char_t* query, xpath_variable_set* variables = 0) const; - PUGIXML_DEPRECATED xpath_node select_single_node(const xpath_query& query) const; - - #endif - - // Print subtree using a writer object - void print(xml_writer& writer, const char_t* indent = PUGIXML_TEXT("\t"), unsigned int flags = format_default, xml_encoding encoding = encoding_auto, unsigned int depth = 0) const; - - #ifndef PUGIXML_NO_STL - // Print subtree to stream - void print(std::basic_ostream >& os, const char_t* indent = PUGIXML_TEXT("\t"), unsigned int flags = format_default, xml_encoding encoding = encoding_auto, unsigned int depth = 0) const; - void print(std::basic_ostream >& os, const char_t* indent = PUGIXML_TEXT("\t"), unsigned int flags = format_default, unsigned int depth = 0) const; - #endif - - // Child nodes iterators - typedef xml_node_iterator iterator; - - iterator begin() const; - iterator end() const; - - // Attribute iterators - typedef xml_attribute_iterator attribute_iterator; - - attribute_iterator attributes_begin() const; - attribute_iterator attributes_end() const; - - // Range-based for support - xml_object_range children() const; - xml_object_range children(const char_t* name) const; - xml_object_range attributes() const; - - // Get node offset in parsed file/string (in char_t units) for debugging purposes - ptrdiff_t offset_debug() const; - - // Get hash value (unique for handles to the same object) - size_t hash_value() const; - - // Get internal pointer - xml_node_struct* internal_object() const; - }; - -#ifdef __BORLANDC__ - // Borland C++ workaround - bool PUGIXML_FUNCTION operator&&(const xml_node& lhs, bool rhs); - bool PUGIXML_FUNCTION operator||(const xml_node& lhs, bool rhs); -#endif - - // A helper for working with text inside PCDATA nodes - class PUGIXML_CLASS xml_text - { - friend class xml_node; - - xml_node_struct* _root; - - typedef void (*unspecified_bool_type)(xml_text***); - - explicit xml_text(xml_node_struct* root); - - xml_node_struct* _data_new(); - xml_node_struct* _data() const; - - public: - // Default constructor. Constructs an empty object. - xml_text(); - - // Safe bool conversion operator - operator unspecified_bool_type() const; - - // Borland C++ workaround - bool operator!() const; - - // Check if text object is empty - bool empty() const; - - // Get text, or "" if object is empty - const char_t* get() const; - - // Get text, or the default value if object is empty - const char_t* as_string(const char_t* def = PUGIXML_TEXT("")) const; - - // Get text as a number, or the default value if conversion did not succeed or object is empty - int as_int(int def = 0) const; - unsigned int as_uint(unsigned int def = 0) const; - double as_double(double def = 0) const; - float as_float(float def = 0) const; - - #ifdef PUGIXML_HAS_LONG_LONG - long long as_llong(long long def = 0) const; - unsigned long long as_ullong(unsigned long long def = 0) const; - #endif - - // Get text as bool (returns true if first character is in '1tTyY' set), or the default value if object is empty - bool as_bool(bool def = false) const; - - // Set text (returns false if object is empty or there is not enough memory) - bool set(const char_t* rhs); - - // Set text with type conversion (numbers are converted to strings, boolean is converted to "true"/"false") - bool set(int rhs); - bool set(unsigned int rhs); - bool set(long rhs); - bool set(unsigned long rhs); - bool set(double rhs); - bool set(float rhs); - bool set(bool rhs); - - #ifdef PUGIXML_HAS_LONG_LONG - bool set(long long rhs); - bool set(unsigned long long rhs); - #endif - - // Set text (equivalent to set without error checking) - xml_text& operator=(const char_t* rhs); - xml_text& operator=(int rhs); - xml_text& operator=(unsigned int rhs); - xml_text& operator=(long rhs); - xml_text& operator=(unsigned long rhs); - xml_text& operator=(double rhs); - xml_text& operator=(float rhs); - xml_text& operator=(bool rhs); - - #ifdef PUGIXML_HAS_LONG_LONG - xml_text& operator=(long long rhs); - xml_text& operator=(unsigned long long rhs); - #endif - - // Get the data node (node_pcdata or node_cdata) for this object - xml_node data() const; - }; - -#ifdef __BORLANDC__ - // Borland C++ workaround - bool PUGIXML_FUNCTION operator&&(const xml_text& lhs, bool rhs); - bool PUGIXML_FUNCTION operator||(const xml_text& lhs, bool rhs); -#endif - - // Child node iterator (a bidirectional iterator over a collection of xml_node) - class PUGIXML_CLASS xml_node_iterator - { - friend class xml_node; - - private: - mutable xml_node _wrap; - xml_node _parent; - - xml_node_iterator(xml_node_struct* ref, xml_node_struct* parent); - - public: - // Iterator traits - typedef ptrdiff_t difference_type; - typedef xml_node value_type; - typedef xml_node* pointer; - typedef xml_node& reference; - - #ifndef PUGIXML_NO_STL - typedef std::bidirectional_iterator_tag iterator_category; - #endif - - // Default constructor - xml_node_iterator(); - - // Construct an iterator which points to the specified node - xml_node_iterator(const xml_node& node); - - // Iterator operators - bool operator==(const xml_node_iterator& rhs) const; - bool operator!=(const xml_node_iterator& rhs) const; - - xml_node& operator*() const; - xml_node* operator->() const; - - const xml_node_iterator& operator++(); - xml_node_iterator operator++(int); - - const xml_node_iterator& operator--(); - xml_node_iterator operator--(int); - }; - - // Attribute iterator (a bidirectional iterator over a collection of xml_attribute) - class PUGIXML_CLASS xml_attribute_iterator - { - friend class xml_node; - - private: - mutable xml_attribute _wrap; - xml_node _parent; - - xml_attribute_iterator(xml_attribute_struct* ref, xml_node_struct* parent); - - public: - // Iterator traits - typedef ptrdiff_t difference_type; - typedef xml_attribute value_type; - typedef xml_attribute* pointer; - typedef xml_attribute& reference; - - #ifndef PUGIXML_NO_STL - typedef std::bidirectional_iterator_tag iterator_category; - #endif - - // Default constructor - xml_attribute_iterator(); - - // Construct an iterator which points to the specified attribute - xml_attribute_iterator(const xml_attribute& attr, const xml_node& parent); - - // Iterator operators - bool operator==(const xml_attribute_iterator& rhs) const; - bool operator!=(const xml_attribute_iterator& rhs) const; - - xml_attribute& operator*() const; - xml_attribute* operator->() const; - - const xml_attribute_iterator& operator++(); - xml_attribute_iterator operator++(int); - - const xml_attribute_iterator& operator--(); - xml_attribute_iterator operator--(int); - }; - - // Named node range helper - class PUGIXML_CLASS xml_named_node_iterator - { - friend class xml_node; - - public: - // Iterator traits - typedef ptrdiff_t difference_type; - typedef xml_node value_type; - typedef xml_node* pointer; - typedef xml_node& reference; - - #ifndef PUGIXML_NO_STL - typedef std::bidirectional_iterator_tag iterator_category; - #endif - - // Default constructor - xml_named_node_iterator(); - - // Construct an iterator which points to the specified node - xml_named_node_iterator(const xml_node& node, const char_t* name); - - // Iterator operators - bool operator==(const xml_named_node_iterator& rhs) const; - bool operator!=(const xml_named_node_iterator& rhs) const; - - xml_node& operator*() const; - xml_node* operator->() const; - - const xml_named_node_iterator& operator++(); - xml_named_node_iterator operator++(int); - - const xml_named_node_iterator& operator--(); - xml_named_node_iterator operator--(int); - - private: - mutable xml_node _wrap; - xml_node _parent; - const char_t* _name; - - xml_named_node_iterator(xml_node_struct* ref, xml_node_struct* parent, const char_t* name); - }; - - // Abstract tree walker class (see xml_node::traverse) - class PUGIXML_CLASS xml_tree_walker - { - friend class xml_node; - - private: - int _depth; - - protected: - // Get current traversal depth - int depth() const; - - public: - xml_tree_walker(); - virtual ~xml_tree_walker(); - - // Callback that is called when traversal begins - virtual bool begin(xml_node& node); - - // Callback that is called for each node traversed - virtual bool for_each(xml_node& node) = 0; - - // Callback that is called when traversal ends - virtual bool end(xml_node& node); - }; - - // Parsing status, returned as part of xml_parse_result object - enum xml_parse_status - { - status_ok = 0, // No error - - status_file_not_found, // File was not found during load_file() - status_io_error, // Error reading from file/stream - status_out_of_memory, // Could not allocate memory - status_internal_error, // Internal error occurred - - status_unrecognized_tag, // Parser could not determine tag type - - status_bad_pi, // Parsing error occurred while parsing document declaration/processing instruction - status_bad_comment, // Parsing error occurred while parsing comment - status_bad_cdata, // Parsing error occurred while parsing CDATA section - status_bad_doctype, // Parsing error occurred while parsing document type declaration - status_bad_pcdata, // Parsing error occurred while parsing PCDATA section - status_bad_start_element, // Parsing error occurred while parsing start element tag - status_bad_attribute, // Parsing error occurred while parsing element attribute - status_bad_end_element, // Parsing error occurred while parsing end element tag - status_end_element_mismatch,// There was a mismatch of start-end tags (closing tag had incorrect name, some tag was not closed or there was an excessive closing tag) - - status_append_invalid_root, // Unable to append nodes since root type is not node_element or node_document (exclusive to xml_node::append_buffer) - - status_no_document_element // Parsing resulted in a document without element nodes - }; - - // Parsing result - struct PUGIXML_CLASS xml_parse_result - { - // Parsing status (see xml_parse_status) - xml_parse_status status; - - // Last parsed offset (in char_t units from start of input data) - ptrdiff_t offset; - - // Source document encoding - xml_encoding encoding; - - // Default constructor, initializes object to failed state - xml_parse_result(); - - // Cast to bool operator - operator bool() const; - - // Get error description - const char* description() const; - }; - - // Document class (DOM tree root) - class PUGIXML_CLASS xml_document: public xml_node - { - private: - char_t* _buffer; - - char _memory[192]; - - // Non-copyable semantics - xml_document(const xml_document&); - xml_document& operator=(const xml_document&); - - void _create(); - void _destroy(); - void _move(xml_document& rhs) PUGIXML_NOEXCEPT_IF_NOT_COMPACT; - - public: - // Default constructor, makes empty document - xml_document(); - - // Destructor, invalidates all node/attribute handles to this document - ~xml_document(); - - #ifdef PUGIXML_HAS_MOVE - // Move semantics support - xml_document(xml_document&& rhs) PUGIXML_NOEXCEPT_IF_NOT_COMPACT; - xml_document& operator=(xml_document&& rhs) PUGIXML_NOEXCEPT_IF_NOT_COMPACT; - #endif - - // Removes all nodes, leaving the empty document - void reset(); - - // Removes all nodes, then copies the entire contents of the specified document - void reset(const xml_document& proto); - - #ifndef PUGIXML_NO_STL - // Load document from stream. - xml_parse_result load(std::basic_istream >& stream, unsigned int options = parse_default, xml_encoding encoding = encoding_auto); - xml_parse_result load(std::basic_istream >& stream, unsigned int options = parse_default); - #endif - - // (deprecated: use load_string instead) Load document from zero-terminated string. No encoding conversions are applied. - PUGIXML_DEPRECATED xml_parse_result load(const char_t* contents, unsigned int options = parse_default); - - // Load document from zero-terminated string. No encoding conversions are applied. - xml_parse_result load_string(const char_t* contents, unsigned int options = parse_default); - - // Load document from file - xml_parse_result load_file(const char* path, unsigned int options = parse_default, xml_encoding encoding = encoding_auto); - xml_parse_result load_file(const wchar_t* path, unsigned int options = parse_default, xml_encoding encoding = encoding_auto); - - // Load document from buffer. Copies/converts the buffer, so it may be deleted or changed after the function returns. - xml_parse_result load_buffer(const void* contents, size_t size, unsigned int options = parse_default, xml_encoding encoding = encoding_auto); - - // Load document from buffer, using the buffer for in-place parsing (the buffer is modified and used for storage of document data). - // You should ensure that buffer data will persist throughout the document's lifetime, and free the buffer memory manually once document is destroyed. - xml_parse_result load_buffer_inplace(void* contents, size_t size, unsigned int options = parse_default, xml_encoding encoding = encoding_auto); - - // Load document from buffer, using the buffer for in-place parsing (the buffer is modified and used for storage of document data). - // You should allocate the buffer with pugixml allocation function; document will free the buffer when it is no longer needed (you can't use it anymore). - xml_parse_result load_buffer_inplace_own(void* contents, size_t size, unsigned int options = parse_default, xml_encoding encoding = encoding_auto); - - // Save XML document to writer (semantics is slightly different from xml_node::print, see documentation for details). - void save(xml_writer& writer, const char_t* indent = PUGIXML_TEXT("\t"), unsigned int flags = format_default, xml_encoding encoding = encoding_auto) const; - - #ifndef PUGIXML_NO_STL - // Save XML document to stream (semantics is slightly different from xml_node::print, see documentation for details). - void save(std::basic_ostream >& stream, const char_t* indent = PUGIXML_TEXT("\t"), unsigned int flags = format_default, xml_encoding encoding = encoding_auto) const; - void save(std::basic_ostream >& stream, const char_t* indent = PUGIXML_TEXT("\t"), unsigned int flags = format_default) const; - #endif - - // Save XML to file - bool save_file(const char* path, const char_t* indent = PUGIXML_TEXT("\t"), unsigned int flags = format_default, xml_encoding encoding = encoding_auto) const; - bool save_file(const wchar_t* path, const char_t* indent = PUGIXML_TEXT("\t"), unsigned int flags = format_default, xml_encoding encoding = encoding_auto) const; - - // Get document element - xml_node document_element() const; - }; +class xml_text; #ifndef PUGIXML_NO_XPATH - // XPath query return type - enum xpath_value_type - { - xpath_type_none, // Unknown type (query failed to compile) - xpath_type_node_set, // Node set (xpath_node_set) - xpath_type_number, // Number - xpath_type_string, // String - xpath_type_boolean // Boolean - }; - - // XPath parsing result - struct PUGIXML_CLASS xpath_parse_result - { - // Error message (0 if no error) - const char* error; - - // Last parsed offset (in char_t units from string start) - ptrdiff_t offset; - - // Default constructor, initializes object to failed state - xpath_parse_result(); - - // Cast to bool operator - operator bool() const; - - // Get error description - const char* description() const; - }; - - // A single XPath variable - class PUGIXML_CLASS xpath_variable - { - friend class xpath_variable_set; - - protected: - xpath_value_type _type; - xpath_variable* _next; - - xpath_variable(xpath_value_type type); - - // Non-copyable semantics - xpath_variable(const xpath_variable&); - xpath_variable& operator=(const xpath_variable&); - - public: - // Get variable name - const char_t* name() const; - - // Get variable type - xpath_value_type type() const; - - // Get variable value; no type conversion is performed, default value (false, NaN, empty string, empty node set) is returned on type mismatch error - bool get_boolean() const; - double get_number() const; - const char_t* get_string() const; - const xpath_node_set& get_node_set() const; - - // Set variable value; no type conversion is performed, false is returned on type mismatch error - bool set(bool value); - bool set(double value); - bool set(const char_t* value); - bool set(const xpath_node_set& value); - }; - - // A set of XPath variables - class PUGIXML_CLASS xpath_variable_set - { - private: - xpath_variable* _data[64]; - - void _assign(const xpath_variable_set& rhs); - void _swap(xpath_variable_set& rhs); - - xpath_variable* _find(const char_t* name) const; - - static bool _clone(xpath_variable* var, xpath_variable** out_result); - static void _destroy(xpath_variable* var); - - public: - // Default constructor/destructor - xpath_variable_set(); - ~xpath_variable_set(); - - // Copy constructor/assignment operator - xpath_variable_set(const xpath_variable_set& rhs); - xpath_variable_set& operator=(const xpath_variable_set& rhs); - - #ifdef PUGIXML_HAS_MOVE - // Move semantics support - xpath_variable_set(xpath_variable_set&& rhs) PUGIXML_NOEXCEPT; - xpath_variable_set& operator=(xpath_variable_set&& rhs) PUGIXML_NOEXCEPT; - #endif - - // Add a new variable or get the existing one, if the types match - xpath_variable* add(const char_t* name, xpath_value_type type); - - // Set value of an existing variable; no type conversion is performed, false is returned if there is no such variable or if types mismatch - bool set(const char_t* name, bool value); - bool set(const char_t* name, double value); - bool set(const char_t* name, const char_t* value); - bool set(const char_t* name, const xpath_node_set& value); - - // Get existing variable by name - xpath_variable* get(const char_t* name); - const xpath_variable* get(const char_t* name) const; - }; - - // A compiled XPath query object - class PUGIXML_CLASS xpath_query - { - private: - void* _impl; - xpath_parse_result _result; - - typedef void (*unspecified_bool_type)(xpath_query***); - - // Non-copyable semantics - xpath_query(const xpath_query&); - xpath_query& operator=(const xpath_query&); - - public: - // Construct a compiled object from XPath expression. - // If PUGIXML_NO_EXCEPTIONS is not defined, throws xpath_exception on compilation errors. - explicit xpath_query(const char_t* query, xpath_variable_set* variables = 0); - - // Constructor - xpath_query(); - - // Destructor - ~xpath_query(); - - #ifdef PUGIXML_HAS_MOVE - // Move semantics support - xpath_query(xpath_query&& rhs) PUGIXML_NOEXCEPT; - xpath_query& operator=(xpath_query&& rhs) PUGIXML_NOEXCEPT; - #endif - - // Get query expression return type - xpath_value_type return_type() const; - - // Evaluate expression as boolean value in the specified context; performs type conversion if necessary. - // If PUGIXML_NO_EXCEPTIONS is not defined, throws std::bad_alloc on out of memory errors. - bool evaluate_boolean(const xpath_node& n) const; - - // Evaluate expression as double value in the specified context; performs type conversion if necessary. - // If PUGIXML_NO_EXCEPTIONS is not defined, throws std::bad_alloc on out of memory errors. - double evaluate_number(const xpath_node& n) const; - - #ifndef PUGIXML_NO_STL - // Evaluate expression as string value in the specified context; performs type conversion if necessary. - // If PUGIXML_NO_EXCEPTIONS is not defined, throws std::bad_alloc on out of memory errors. - string_t evaluate_string(const xpath_node& n) const; - #endif - - // Evaluate expression as string value in the specified context; performs type conversion if necessary. - // At most capacity characters are written to the destination buffer, full result size is returned (includes terminating zero). - // If PUGIXML_NO_EXCEPTIONS is not defined, throws std::bad_alloc on out of memory errors. - // If PUGIXML_NO_EXCEPTIONS is defined, returns empty set instead. - size_t evaluate_string(char_t* buffer, size_t capacity, const xpath_node& n) const; - - // Evaluate expression as node set in the specified context. - // If PUGIXML_NO_EXCEPTIONS is not defined, throws xpath_exception on type mismatch and std::bad_alloc on out of memory errors. - // If PUGIXML_NO_EXCEPTIONS is defined, returns empty node set instead. - xpath_node_set evaluate_node_set(const xpath_node& n) const; - - // Evaluate expression as node set in the specified context. - // Return first node in document order, or empty node if node set is empty. - // If PUGIXML_NO_EXCEPTIONS is not defined, throws xpath_exception on type mismatch and std::bad_alloc on out of memory errors. - // If PUGIXML_NO_EXCEPTIONS is defined, returns empty node instead. - xpath_node evaluate_node(const xpath_node& n) const; - - // Get parsing result (used to get compilation errors in PUGIXML_NO_EXCEPTIONS mode) - const xpath_parse_result& result() const; - - // Safe bool conversion operator - operator unspecified_bool_type() const; - - // Borland C++ workaround - bool operator!() const; - }; - - #ifndef PUGIXML_NO_EXCEPTIONS - #if defined(_MSC_VER) - // C4275 can be ignored in Visual C++ if you are deriving - // from a type in the Standard C++ Library - #pragma warning(push) - #pragma warning(disable: 4275) - #endif - // XPath exception class - class PUGIXML_CLASS xpath_exception: public std::exception - { - private: - xpath_parse_result _result; - - public: - // Construct exception from parse result - explicit xpath_exception(const xpath_parse_result& result); - - // Get error message - virtual const char* what() const throw() PUGIXML_OVERRIDE; - - // Get parse result - const xpath_parse_result& result() const; - }; - #if defined(_MSC_VER) - #pragma warning(pop) - #endif - #endif - - // XPath node class (either xml_node or xml_attribute) - class PUGIXML_CLASS xpath_node - { - private: - xml_node _node; - xml_attribute _attribute; - - typedef void (*unspecified_bool_type)(xpath_node***); - - public: - // Default constructor; constructs empty XPath node - xpath_node(); - - // Construct XPath node from XML node/attribute - xpath_node(const xml_node& node); - xpath_node(const xml_attribute& attribute, const xml_node& parent); - - // Get node/attribute, if any - xml_node node() const; - xml_attribute attribute() const; - - // Get parent of contained node/attribute - xml_node parent() const; - - // Safe bool conversion operator - operator unspecified_bool_type() const; - - // Borland C++ workaround - bool operator!() const; - - // Comparison operators - bool operator==(const xpath_node& n) const; - bool operator!=(const xpath_node& n) const; - }; - -#ifdef __BORLANDC__ - // Borland C++ workaround - bool PUGIXML_FUNCTION operator&&(const xpath_node& lhs, bool rhs); - bool PUGIXML_FUNCTION operator||(const xpath_node& lhs, bool rhs); +class xpath_node; +class xpath_node_set; +class xpath_query; +class xpath_variable_set; #endif - // A fixed-size collection of XPath nodes - class PUGIXML_CLASS xpath_node_set +// Range-based for loop support +template +class xml_object_range +{ + public: + typedef It const_iterator; + typedef It iterator; + + xml_object_range(It b, It e) + : _begin(b) + , _end(e) { - public: - // Collection type - enum type_t + } + + It begin() const { return _begin; } + It end() const { return _end; } + + private: + It _begin, _end; +}; + +// Writer interface for node printing (see xml_node::print) +class PUGIXML_CLASS xml_writer +{ + public: + virtual ~xml_writer() {} + + // Write memory chunk into stream/file/whatever + virtual void write(const void* data, size_t size) = 0; +}; + +// xml_writer implementation for FILE* +class PUGIXML_CLASS xml_writer_file : public xml_writer +{ + public: + // Construct writer from a FILE* object; void* is used to avoid header dependencies on stdio + xml_writer_file(void* file); + + virtual void write(const void* data, size_t size) PUGIXML_OVERRIDE; + + private: + void* file; +}; + +#ifndef PUGIXML_NO_STL +// xml_writer implementation for streams +class PUGIXML_CLASS xml_writer_stream : public xml_writer +{ + public: + // Construct writer from an output stream object + xml_writer_stream(std::basic_ostream>& stream); + xml_writer_stream(std::basic_ostream>& stream); + + virtual void write(const void* data, size_t size) PUGIXML_OVERRIDE; + + private: + std::basic_ostream>* narrow_stream; + std::basic_ostream>* wide_stream; +}; +#endif + +// A light-weight handle for manipulating attributes in DOM tree +class PUGIXML_CLASS xml_attribute +{ + friend class xml_attribute_iterator; + friend class xml_node; + + private: + xml_attribute_struct* _attr; + + typedef void (*unspecified_bool_type)(xml_attribute***); + + public: + // Default constructor. Constructs an empty attribute. + xml_attribute(); + + // Constructs attribute from internal pointer + explicit xml_attribute(xml_attribute_struct* attr); + + // Safe bool conversion operator + operator unspecified_bool_type() const; + + // Borland C++ workaround + bool operator!() const; + + // Comparison operators (compares wrapped attribute pointers) + bool operator==(const xml_attribute& r) const; + bool operator!=(const xml_attribute& r) const; + bool operator<(const xml_attribute& r) const; + bool operator>(const xml_attribute& r) const; + bool operator<=(const xml_attribute& r) const; + bool operator>=(const xml_attribute& r) const; + + // Check if attribute is empty + bool empty() const; + + // Get attribute name/value, or "" if attribute is empty + const char_t* name() const; + const char_t* value() const; + + // Get attribute value, or the default value if attribute is empty + const char_t* as_string(const char_t* def = PUGIXML_TEXT("")) const; + + // Get attribute value as a number, or the default value if conversion did not succeed or attribute is empty + int as_int(int def = 0) const; + unsigned int as_uint(unsigned int def = 0) const; + double as_double(double def = 0) const; + float as_float(float def = 0) const; + +#ifdef PUGIXML_HAS_LONG_LONG + long long as_llong(long long def = 0) const; + unsigned long long as_ullong(unsigned long long def = 0) const; +#endif + + // Get attribute value as bool (returns true if first character is in '1tTyY' set), or the default value if attribute is empty + bool as_bool(bool def = false) const; + + // Set attribute name/value (returns false if attribute is empty or there is not enough memory) + bool set_name(const char_t* rhs); + bool set_value(const char_t* rhs); + + // Set attribute value with type conversion (numbers are converted to strings, boolean is converted to "true"/"false") + bool set_value(int rhs); + bool set_value(unsigned int rhs); + bool set_value(long rhs); + bool set_value(unsigned long rhs); + bool set_value(double rhs); + bool set_value(float rhs); + bool set_value(bool rhs); + +#ifdef PUGIXML_HAS_LONG_LONG + bool set_value(long long rhs); + bool set_value(unsigned long long rhs); +#endif + + // Set attribute value (equivalent to set_value without error checking) + xml_attribute& operator=(const char_t* rhs); + xml_attribute& operator=(int rhs); + xml_attribute& operator=(unsigned int rhs); + xml_attribute& operator=(long rhs); + xml_attribute& operator=(unsigned long rhs); + xml_attribute& operator=(double rhs); + xml_attribute& operator=(float rhs); + xml_attribute& operator=(bool rhs); + +#ifdef PUGIXML_HAS_LONG_LONG + xml_attribute& operator=(long long rhs); + xml_attribute& operator=(unsigned long long rhs); +#endif + + // Get next/previous attribute in the attribute list of the parent node + xml_attribute next_attribute() const; + xml_attribute previous_attribute() const; + + // Get hash value (unique for handles to the same object) + size_t hash_value() const; + + // Get internal pointer + xml_attribute_struct* internal_object() const; +}; + +#ifdef __BORLANDC__ +// Borland C++ workaround +bool PUGIXML_FUNCTION operator&&(const xml_attribute& lhs, bool rhs); +bool PUGIXML_FUNCTION operator||(const xml_attribute& lhs, bool rhs); +#endif + +// A light-weight handle for manipulating nodes in DOM tree +class PUGIXML_CLASS xml_node +{ + friend class xml_attribute_iterator; + friend class xml_node_iterator; + friend class xml_named_node_iterator; + + protected: + xml_node_struct* _root; + + typedef void (*unspecified_bool_type)(xml_node***); + + public: + // Default constructor. Constructs an empty node. + xml_node(); + + // Constructs node from internal pointer + explicit xml_node(xml_node_struct* p); + + // Safe bool conversion operator + operator unspecified_bool_type() const; + + // Borland C++ workaround + bool operator!() const; + + // Comparison operators (compares wrapped node pointers) + bool operator==(const xml_node& r) const; + bool operator!=(const xml_node& r) const; + bool operator<(const xml_node& r) const; + bool operator>(const xml_node& r) const; + bool operator<=(const xml_node& r) const; + bool operator>=(const xml_node& r) const; + + // Check if node is empty. + bool empty() const; + + // Get node type + xml_node_type type() const; + + // Get node name, or "" if node is empty or it has no name + const char_t* name() const; + + // Get node value, or "" if node is empty or it has no value + // Note: For text node.value() does not return "text"! Use child_value() or text() methods to access text inside nodes. + const char_t* value() const; + + // Get attribute list + xml_attribute first_attribute() const; + xml_attribute last_attribute() const; + + // Get children list + xml_node first_child() const; + xml_node last_child() const; + + // Get next/previous sibling in the children list of the parent node + xml_node next_sibling() const; + xml_node previous_sibling() const; + + // Get parent node + xml_node parent() const; + + // Get root of DOM tree this node belongs to + xml_node root() const; + + // Get text object for the current node + xml_text text() const; + + // Get child, attribute or next/previous sibling with the specified name + xml_node child(const char_t* name) const; + xml_attribute attribute(const char_t* name) const; + xml_node next_sibling(const char_t* name) const; + xml_node previous_sibling(const char_t* name) const; + + // Get attribute, starting the search from a hint (and updating hint so that searching for a sequence of attributes is fast) + xml_attribute attribute(const char_t* name, xml_attribute& hint) const; + + // Get child value of current node; that is, value of the first child node of type PCDATA/CDATA + const char_t* child_value() const; + + // Get child value of child with specified name. Equivalent to child(name).child_value(). + const char_t* child_value(const char_t* name) const; + + // Set node name/value (returns false if node is empty, there is not enough memory, or node can not have name/value) + bool set_name(const char_t* rhs); + bool set_value(const char_t* rhs); + + // Add attribute with specified name. Returns added attribute, or empty attribute on errors. + xml_attribute append_attribute(const char_t* name); + xml_attribute prepend_attribute(const char_t* name); + xml_attribute insert_attribute_after(const char_t* name, const xml_attribute& attr); + xml_attribute insert_attribute_before(const char_t* name, const xml_attribute& attr); + + // Add a copy of the specified attribute. Returns added attribute, or empty attribute on errors. + xml_attribute append_copy(const xml_attribute& proto); + xml_attribute prepend_copy(const xml_attribute& proto); + xml_attribute insert_copy_after(const xml_attribute& proto, const xml_attribute& attr); + xml_attribute insert_copy_before(const xml_attribute& proto, const xml_attribute& attr); + + // Add child node with specified type. Returns added node, or empty node on errors. + xml_node append_child(xml_node_type type = node_element); + xml_node prepend_child(xml_node_type type = node_element); + xml_node insert_child_after(xml_node_type type, const xml_node& node); + xml_node insert_child_before(xml_node_type type, const xml_node& node); + + // Add child element with specified name. Returns added node, or empty node on errors. + xml_node append_child(const char_t* name); + xml_node prepend_child(const char_t* name); + xml_node insert_child_after(const char_t* name, const xml_node& node); + xml_node insert_child_before(const char_t* name, const xml_node& node); + + // Add a copy of the specified node as a child. Returns added node, or empty node on errors. + xml_node append_copy(const xml_node& proto); + xml_node prepend_copy(const xml_node& proto); + xml_node insert_copy_after(const xml_node& proto, const xml_node& node); + xml_node insert_copy_before(const xml_node& proto, const xml_node& node); + + // Move the specified node to become a child of this node. Returns moved node, or empty node on errors. + xml_node append_move(const xml_node& moved); + xml_node prepend_move(const xml_node& moved); + xml_node insert_move_after(const xml_node& moved, const xml_node& node); + xml_node insert_move_before(const xml_node& moved, const xml_node& node); + + // Remove specified attribute + bool remove_attribute(const xml_attribute& a); + bool remove_attribute(const char_t* name); + + // Remove specified child + bool remove_child(const xml_node& n); + bool remove_child(const char_t* name); + + // Parses buffer as an XML document fragment and appends all nodes as children of the current node. + // Copies/converts the buffer, so it may be deleted or changed after the function returns. + // Note: append_buffer allocates memory that has the lifetime of the owning document; removing the appended nodes does not immediately reclaim that memory. + xml_parse_result append_buffer(const void* contents, size_t size, unsigned int options = parse_default, xml_encoding encoding = encoding_auto); + + // Find attribute using predicate. Returns first attribute for which predicate returned true. + template + xml_attribute find_attribute(Predicate pred) const + { + if (!_root) + return xml_attribute(); + + for (xml_attribute attrib = first_attribute(); attrib; attrib = attrib.next_attribute()) + if (pred(attrib)) + return attrib; + + return xml_attribute(); + } + + // Find child node using predicate. Returns first child for which predicate returned true. + template + xml_node find_child(Predicate pred) const + { + if (!_root) + return xml_node(); + + for (xml_node node = first_child(); node; node = node.next_sibling()) + if (pred(node)) + return node; + + return xml_node(); + } + + // Find node from subtree using predicate. Returns first node from subtree (depth-first), for which predicate returned true. + template + xml_node find_node(Predicate pred) const + { + if (!_root) + return xml_node(); + + xml_node cur = first_child(); + + while (cur._root && cur._root != _root) { - type_unsorted, // Not ordered - type_sorted, // Sorted by document order (ascending) - type_sorted_reverse // Sorted by document order (descending) - }; + if (pred(cur)) + return cur; - // Constant iterator type - typedef const xpath_node* const_iterator; + if (cur.first_child()) + cur = cur.first_child(); + else if (cur.next_sibling()) + cur = cur.next_sibling(); + else + { + while (!cur.next_sibling() && cur._root != _root) + cur = cur.parent(); - // We define non-constant iterator to be the same as constant iterator so that various generic algorithms (i.e. boost foreach) work - typedef const xpath_node* iterator; + if (cur._root != _root) + cur = cur.next_sibling(); + } + } - // Default constructor. Constructs empty set. - xpath_node_set(); + return xml_node(); + } - // Constructs a set from iterator range; data is not checked for duplicates and is not sorted according to provided type, so be careful - xpath_node_set(const_iterator begin, const_iterator end, type_t type = type_unsorted); + // Find child node by attribute name/value + xml_node find_child_by_attribute(const char_t* name, const char_t* attr_name, const char_t* attr_value) const; + xml_node find_child_by_attribute(const char_t* attr_name, const char_t* attr_value) const; - // Destructor - ~xpath_node_set(); +#ifndef PUGIXML_NO_STL + // Get the absolute node path from root as a text string. + string_t path(char_t delimiter = '/') const; +#endif - // Copy constructor/assignment operator - xpath_node_set(const xpath_node_set& ns); - xpath_node_set& operator=(const xpath_node_set& ns); + // Search for a node by path consisting of node names and . or .. elements. + xml_node first_element_by_path(const char_t* path, char_t delimiter = '/') const; - #ifdef PUGIXML_HAS_MOVE - // Move semantics support - xpath_node_set(xpath_node_set&& rhs) PUGIXML_NOEXCEPT; - xpath_node_set& operator=(xpath_node_set&& rhs) PUGIXML_NOEXCEPT; - #endif + // Recursively traverse subtree with xml_tree_walker + bool traverse(xml_tree_walker& walker); - // Get collection type - type_t type() const; +#ifndef PUGIXML_NO_XPATH + // Select single node by evaluating XPath query. Returns first node from the resulting node set. + xpath_node select_node(const char_t* query, xpath_variable_set* variables = 0) const; + xpath_node select_node(const xpath_query& query) const; - // Get collection size - size_t size() const; + // Select node set by evaluating XPath query + xpath_node_set select_nodes(const char_t* query, xpath_variable_set* variables = 0) const; + xpath_node_set select_nodes(const xpath_query& query) const; - // Indexing operator - const xpath_node& operator[](size_t index) const; + // (deprecated: use select_node instead) Select single node by evaluating XPath query. + PUGIXML_DEPRECATED xpath_node select_single_node(const char_t* query, xpath_variable_set* variables = 0) const; + PUGIXML_DEPRECATED xpath_node select_single_node(const xpath_query& query) const; - // Collection iterators - const_iterator begin() const; - const_iterator end() const; +#endif - // Sort the collection in ascending/descending order by document order - void sort(bool reverse = false); + // Print subtree using a writer object + void print(xml_writer& writer, const char_t* indent = PUGIXML_TEXT("\t"), unsigned int flags = format_default, xml_encoding encoding = encoding_auto, unsigned int depth = 0) const; - // Get first node in the collection by document order - xpath_node first() const; +#ifndef PUGIXML_NO_STL + // Print subtree to stream + void print(std::basic_ostream>& os, const char_t* indent = PUGIXML_TEXT("\t"), unsigned int flags = format_default, xml_encoding encoding = encoding_auto, unsigned int depth = 0) const; + void print(std::basic_ostream>& os, const char_t* indent = PUGIXML_TEXT("\t"), unsigned int flags = format_default, unsigned int depth = 0) const; +#endif - // Check if collection is empty - bool empty() const; + // Child nodes iterators + typedef xml_node_iterator iterator; - private: - type_t _type; + iterator begin() const; + iterator end() const; - xpath_node _storage[1]; + // Attribute iterators + typedef xml_attribute_iterator attribute_iterator; - xpath_node* _begin; - xpath_node* _end; + attribute_iterator attributes_begin() const; + attribute_iterator attributes_end() const; - void _assign(const_iterator begin, const_iterator end, type_t type); - void _move(xpath_node_set& rhs) PUGIXML_NOEXCEPT; + // Range-based for support + xml_object_range children() const; + xml_object_range children(const char_t* name) const; + xml_object_range attributes() const; + + // Get node offset in parsed file/string (in char_t units) for debugging purposes + ptrdiff_t offset_debug() const; + + // Get hash value (unique for handles to the same object) + size_t hash_value() const; + + // Get internal pointer + xml_node_struct* internal_object() const; +}; + +#ifdef __BORLANDC__ +// Borland C++ workaround +bool PUGIXML_FUNCTION operator&&(const xml_node& lhs, bool rhs); +bool PUGIXML_FUNCTION operator||(const xml_node& lhs, bool rhs); +#endif + +// A helper for working with text inside PCDATA nodes +class PUGIXML_CLASS xml_text +{ + friend class xml_node; + + xml_node_struct* _root; + + typedef void (*unspecified_bool_type)(xml_text***); + + explicit xml_text(xml_node_struct* root); + + xml_node_struct* _data_new(); + xml_node_struct* _data() const; + + public: + // Default constructor. Constructs an empty object. + xml_text(); + + // Safe bool conversion operator + operator unspecified_bool_type() const; + + // Borland C++ workaround + bool operator!() const; + + // Check if text object is empty + bool empty() const; + + // Get text, or "" if object is empty + const char_t* get() const; + + // Get text, or the default value if object is empty + const char_t* as_string(const char_t* def = PUGIXML_TEXT("")) const; + + // Get text as a number, or the default value if conversion did not succeed or object is empty + int as_int(int def = 0) const; + unsigned int as_uint(unsigned int def = 0) const; + double as_double(double def = 0) const; + float as_float(float def = 0) const; + +#ifdef PUGIXML_HAS_LONG_LONG + long long as_llong(long long def = 0) const; + unsigned long long as_ullong(unsigned long long def = 0) const; +#endif + + // Get text as bool (returns true if first character is in '1tTyY' set), or the default value if object is empty + bool as_bool(bool def = false) const; + + // Set text (returns false if object is empty or there is not enough memory) + bool set(const char_t* rhs); + + // Set text with type conversion (numbers are converted to strings, boolean is converted to "true"/"false") + bool set(int rhs); + bool set(unsigned int rhs); + bool set(long rhs); + bool set(unsigned long rhs); + bool set(double rhs); + bool set(float rhs); + bool set(bool rhs); + +#ifdef PUGIXML_HAS_LONG_LONG + bool set(long long rhs); + bool set(unsigned long long rhs); +#endif + + // Set text (equivalent to set without error checking) + xml_text& operator=(const char_t* rhs); + xml_text& operator=(int rhs); + xml_text& operator=(unsigned int rhs); + xml_text& operator=(long rhs); + xml_text& operator=(unsigned long rhs); + xml_text& operator=(double rhs); + xml_text& operator=(float rhs); + xml_text& operator=(bool rhs); + +#ifdef PUGIXML_HAS_LONG_LONG + xml_text& operator=(long long rhs); + xml_text& operator=(unsigned long long rhs); +#endif + + // Get the data node (node_pcdata or node_cdata) for this object + xml_node data() const; +}; + +#ifdef __BORLANDC__ +// Borland C++ workaround +bool PUGIXML_FUNCTION operator&&(const xml_text& lhs, bool rhs); +bool PUGIXML_FUNCTION operator||(const xml_text& lhs, bool rhs); +#endif + +// Child node iterator (a bidirectional iterator over a collection of xml_node) +class PUGIXML_CLASS xml_node_iterator +{ + friend class xml_node; + + private: + mutable xml_node _wrap; + xml_node _parent; + + xml_node_iterator(xml_node_struct* ref, xml_node_struct* parent); + + public: + // Iterator traits + typedef ptrdiff_t difference_type; + typedef xml_node value_type; + typedef xml_node* pointer; + typedef xml_node& reference; + +#ifndef PUGIXML_NO_STL + typedef std::bidirectional_iterator_tag iterator_category; +#endif + + // Default constructor + xml_node_iterator(); + + // Construct an iterator which points to the specified node + xml_node_iterator(const xml_node& node); + + // Iterator operators + bool operator==(const xml_node_iterator& rhs) const; + bool operator!=(const xml_node_iterator& rhs) const; + + xml_node& operator*() const; + xml_node* operator->() const; + + const xml_node_iterator& operator++(); + xml_node_iterator operator++(int); + + const xml_node_iterator& operator--(); + xml_node_iterator operator--(int); +}; + +// Attribute iterator (a bidirectional iterator over a collection of xml_attribute) +class PUGIXML_CLASS xml_attribute_iterator +{ + friend class xml_node; + + private: + mutable xml_attribute _wrap; + xml_node _parent; + + xml_attribute_iterator(xml_attribute_struct* ref, xml_node_struct* parent); + + public: + // Iterator traits + typedef ptrdiff_t difference_type; + typedef xml_attribute value_type; + typedef xml_attribute* pointer; + typedef xml_attribute& reference; + +#ifndef PUGIXML_NO_STL + typedef std::bidirectional_iterator_tag iterator_category; +#endif + + // Default constructor + xml_attribute_iterator(); + + // Construct an iterator which points to the specified attribute + xml_attribute_iterator(const xml_attribute& attr, const xml_node& parent); + + // Iterator operators + bool operator==(const xml_attribute_iterator& rhs) const; + bool operator!=(const xml_attribute_iterator& rhs) const; + + xml_attribute& operator*() const; + xml_attribute* operator->() const; + + const xml_attribute_iterator& operator++(); + xml_attribute_iterator operator++(int); + + const xml_attribute_iterator& operator--(); + xml_attribute_iterator operator--(int); +}; + +// Named node range helper +class PUGIXML_CLASS xml_named_node_iterator +{ + friend class xml_node; + + public: + // Iterator traits + typedef ptrdiff_t difference_type; + typedef xml_node value_type; + typedef xml_node* pointer; + typedef xml_node& reference; + +#ifndef PUGIXML_NO_STL + typedef std::bidirectional_iterator_tag iterator_category; +#endif + + // Default constructor + xml_named_node_iterator(); + + // Construct an iterator which points to the specified node + xml_named_node_iterator(const xml_node& node, const char_t* name); + + // Iterator operators + bool operator==(const xml_named_node_iterator& rhs) const; + bool operator!=(const xml_named_node_iterator& rhs) const; + + xml_node& operator*() const; + xml_node* operator->() const; + + const xml_named_node_iterator& operator++(); + xml_named_node_iterator operator++(int); + + const xml_named_node_iterator& operator--(); + xml_named_node_iterator operator--(int); + + private: + mutable xml_node _wrap; + xml_node _parent; + const char_t* _name; + + xml_named_node_iterator(xml_node_struct* ref, xml_node_struct* parent, const char_t* name); +}; + +// Abstract tree walker class (see xml_node::traverse) +class PUGIXML_CLASS xml_tree_walker +{ + friend class xml_node; + + private: + int _depth; + + protected: + // Get current traversal depth + int depth() const; + + public: + xml_tree_walker(); + virtual ~xml_tree_walker(); + + // Callback that is called when traversal begins + virtual bool begin(xml_node& node); + + // Callback that is called for each node traversed + virtual bool for_each(xml_node& node) = 0; + + // Callback that is called when traversal ends + virtual bool end(xml_node& node); +}; + +// Parsing status, returned as part of xml_parse_result object +enum xml_parse_status +{ + status_ok = 0, // No error + + status_file_not_found, // File was not found during load_file() + status_io_error, // Error reading from file/stream + status_out_of_memory, // Could not allocate memory + status_internal_error, // Internal error occurred + + status_unrecognized_tag, // Parser could not determine tag type + + status_bad_pi, // Parsing error occurred while parsing document declaration/processing instruction + status_bad_comment, // Parsing error occurred while parsing comment + status_bad_cdata, // Parsing error occurred while parsing CDATA section + status_bad_doctype, // Parsing error occurred while parsing document type declaration + status_bad_pcdata, // Parsing error occurred while parsing PCDATA section + status_bad_start_element, // Parsing error occurred while parsing start element tag + status_bad_attribute, // Parsing error occurred while parsing element attribute + status_bad_end_element, // Parsing error occurred while parsing end element tag + status_end_element_mismatch, // There was a mismatch of start-end tags (closing tag had incorrect name, some tag was not closed or there was an excessive closing tag) + + status_append_invalid_root, // Unable to append nodes since root type is not node_element or node_document (exclusive to xml_node::append_buffer) + + status_no_document_element // Parsing resulted in a document without element nodes +}; + +// Parsing result +struct PUGIXML_CLASS xml_parse_result +{ + // Parsing status (see xml_parse_status) + xml_parse_status status; + + // Last parsed offset (in char_t units from start of input data) + ptrdiff_t offset; + + // Source document encoding + xml_encoding encoding; + + // Default constructor, initializes object to failed state + xml_parse_result(); + + // Cast to bool operator + operator bool() const; + + // Get error description + const char* description() const; +}; + +// Document class (DOM tree root) +class PUGIXML_CLASS xml_document : public xml_node +{ + private: + char_t* _buffer; + + char _memory[192]; + + // Non-copyable semantics + xml_document(const xml_document&); + xml_document& operator=(const xml_document&); + + void _create(); + void _destroy(); + void _move(xml_document& rhs) PUGIXML_NOEXCEPT_IF_NOT_COMPACT; + + public: + // Default constructor, makes empty document + xml_document(); + + // Destructor, invalidates all node/attribute handles to this document + ~xml_document(); + +#ifdef PUGIXML_HAS_MOVE + // Move semantics support + xml_document(xml_document&& rhs) PUGIXML_NOEXCEPT_IF_NOT_COMPACT; + xml_document& operator=(xml_document&& rhs) PUGIXML_NOEXCEPT_IF_NOT_COMPACT; +#endif + + // Removes all nodes, leaving the empty document + void reset(); + + // Removes all nodes, then copies the entire contents of the specified document + void reset(const xml_document& proto); + +#ifndef PUGIXML_NO_STL + // Load document from stream. + xml_parse_result load(std::basic_istream>& stream, unsigned int options = parse_default, xml_encoding encoding = encoding_auto); + xml_parse_result load(std::basic_istream>& stream, unsigned int options = parse_default); +#endif + + // (deprecated: use load_string instead) Load document from zero-terminated string. No encoding conversions are applied. + PUGIXML_DEPRECATED xml_parse_result load(const char_t* contents, unsigned int options = parse_default); + + // Load document from zero-terminated string. No encoding conversions are applied. + xml_parse_result load_string(const char_t* contents, unsigned int options = parse_default); + + // Load document from file + xml_parse_result load_file(const char* path, unsigned int options = parse_default, xml_encoding encoding = encoding_auto); + xml_parse_result load_file(const wchar_t* path, unsigned int options = parse_default, xml_encoding encoding = encoding_auto); + + // Load document from buffer. Copies/converts the buffer, so it may be deleted or changed after the function returns. + xml_parse_result load_buffer(const void* contents, size_t size, unsigned int options = parse_default, xml_encoding encoding = encoding_auto); + + // Load document from buffer, using the buffer for in-place parsing (the buffer is modified and used for storage of document data). + // You should ensure that buffer data will persist throughout the document's lifetime, and free the buffer memory manually once document is destroyed. + xml_parse_result load_buffer_inplace(void* contents, size_t size, unsigned int options = parse_default, xml_encoding encoding = encoding_auto); + + // Load document from buffer, using the buffer for in-place parsing (the buffer is modified and used for storage of document data). + // You should allocate the buffer with pugixml allocation function; document will free the buffer when it is no longer needed (you can't use it anymore). + xml_parse_result load_buffer_inplace_own(void* contents, size_t size, unsigned int options = parse_default, xml_encoding encoding = encoding_auto); + + // Save XML document to writer (semantics is slightly different from xml_node::print, see documentation for details). + void save(xml_writer& writer, const char_t* indent = PUGIXML_TEXT("\t"), unsigned int flags = format_default, xml_encoding encoding = encoding_auto) const; + +#ifndef PUGIXML_NO_STL + // Save XML document to stream (semantics is slightly different from xml_node::print, see documentation for details). + void save(std::basic_ostream>& stream, const char_t* indent = PUGIXML_TEXT("\t"), unsigned int flags = format_default, xml_encoding encoding = encoding_auto) const; + void save(std::basic_ostream>& stream, const char_t* indent = PUGIXML_TEXT("\t"), unsigned int flags = format_default) const; +#endif + + // Save XML to file + bool save_file(const char* path, const char_t* indent = PUGIXML_TEXT("\t"), unsigned int flags = format_default, xml_encoding encoding = encoding_auto) const; + bool save_file(const wchar_t* path, const char_t* indent = PUGIXML_TEXT("\t"), unsigned int flags = format_default, xml_encoding encoding = encoding_auto) const; + + // Get document element + xml_node document_element() const; +}; + +#ifndef PUGIXML_NO_XPATH +// XPath query return type +enum xpath_value_type +{ + xpath_type_none, // Unknown type (query failed to compile) + xpath_type_node_set, // Node set (xpath_node_set) + xpath_type_number, // Number + xpath_type_string, // String + xpath_type_boolean // Boolean +}; + +// XPath parsing result +struct PUGIXML_CLASS xpath_parse_result +{ + // Error message (0 if no error) + const char* error; + + // Last parsed offset (in char_t units from string start) + ptrdiff_t offset; + + // Default constructor, initializes object to failed state + xpath_parse_result(); + + // Cast to bool operator + operator bool() const; + + // Get error description + const char* description() const; +}; + +// A single XPath variable +class PUGIXML_CLASS xpath_variable +{ + friend class xpath_variable_set; + + protected: + xpath_value_type _type; + xpath_variable* _next; + + xpath_variable(xpath_value_type type); + + // Non-copyable semantics + xpath_variable(const xpath_variable&); + xpath_variable& operator=(const xpath_variable&); + + public: + // Get variable name + const char_t* name() const; + + // Get variable type + xpath_value_type type() const; + + // Get variable value; no type conversion is performed, default value (false, NaN, empty string, empty node set) is returned on type mismatch error + bool get_boolean() const; + double get_number() const; + const char_t* get_string() const; + const xpath_node_set& get_node_set() const; + + // Set variable value; no type conversion is performed, false is returned on type mismatch error + bool set(bool value); + bool set(double value); + bool set(const char_t* value); + bool set(const xpath_node_set& value); +}; + +// A set of XPath variables +class PUGIXML_CLASS xpath_variable_set +{ + private: + xpath_variable* _data[64]; + + void _assign(const xpath_variable_set& rhs); + void _swap(xpath_variable_set& rhs); + + xpath_variable* _find(const char_t* name) const; + + static bool _clone(xpath_variable* var, xpath_variable** out_result); + static void _destroy(xpath_variable* var); + + public: + // Default constructor/destructor + xpath_variable_set(); + ~xpath_variable_set(); + + // Copy constructor/assignment operator + xpath_variable_set(const xpath_variable_set& rhs); + xpath_variable_set& operator=(const xpath_variable_set& rhs); + +#ifdef PUGIXML_HAS_MOVE + // Move semantics support + xpath_variable_set(xpath_variable_set&& rhs) PUGIXML_NOEXCEPT; + xpath_variable_set& operator=(xpath_variable_set&& rhs) PUGIXML_NOEXCEPT; +#endif + + // Add a new variable or get the existing one, if the types match + xpath_variable* add(const char_t* name, xpath_value_type type); + + // Set value of an existing variable; no type conversion is performed, false is returned if there is no such variable or if types mismatch + bool set(const char_t* name, bool value); + bool set(const char_t* name, double value); + bool set(const char_t* name, const char_t* value); + bool set(const char_t* name, const xpath_node_set& value); + + // Get existing variable by name + xpath_variable* get(const char_t* name); + const xpath_variable* get(const char_t* name) const; +}; + +// A compiled XPath query object +class PUGIXML_CLASS xpath_query +{ + private: + void* _impl; + xpath_parse_result _result; + + typedef void (*unspecified_bool_type)(xpath_query***); + + // Non-copyable semantics + xpath_query(const xpath_query&); + xpath_query& operator=(const xpath_query&); + + public: + // Construct a compiled object from XPath expression. + // If PUGIXML_NO_EXCEPTIONS is not defined, throws xpath_exception on compilation errors. + explicit xpath_query(const char_t* query, xpath_variable_set* variables = 0); + + // Constructor + xpath_query(); + + // Destructor + ~xpath_query(); + +#ifdef PUGIXML_HAS_MOVE + // Move semantics support + xpath_query(xpath_query&& rhs) PUGIXML_NOEXCEPT; + xpath_query& operator=(xpath_query&& rhs) PUGIXML_NOEXCEPT; +#endif + + // Get query expression return type + xpath_value_type return_type() const; + + // Evaluate expression as boolean value in the specified context; performs type conversion if necessary. + // If PUGIXML_NO_EXCEPTIONS is not defined, throws std::bad_alloc on out of memory errors. + bool evaluate_boolean(const xpath_node& n) const; + + // Evaluate expression as double value in the specified context; performs type conversion if necessary. + // If PUGIXML_NO_EXCEPTIONS is not defined, throws std::bad_alloc on out of memory errors. + double evaluate_number(const xpath_node& n) const; + +#ifndef PUGIXML_NO_STL + // Evaluate expression as string value in the specified context; performs type conversion if necessary. + // If PUGIXML_NO_EXCEPTIONS is not defined, throws std::bad_alloc on out of memory errors. + string_t evaluate_string(const xpath_node& n) const; +#endif + + // Evaluate expression as string value in the specified context; performs type conversion if necessary. + // At most capacity characters are written to the destination buffer, full result size is returned (includes terminating zero). + // If PUGIXML_NO_EXCEPTIONS is not defined, throws std::bad_alloc on out of memory errors. + // If PUGIXML_NO_EXCEPTIONS is defined, returns empty set instead. + size_t evaluate_string(char_t* buffer, size_t capacity, const xpath_node& n) const; + + // Evaluate expression as node set in the specified context. + // If PUGIXML_NO_EXCEPTIONS is not defined, throws xpath_exception on type mismatch and std::bad_alloc on out of memory errors. + // If PUGIXML_NO_EXCEPTIONS is defined, returns empty node set instead. + xpath_node_set evaluate_node_set(const xpath_node& n) const; + + // Evaluate expression as node set in the specified context. + // Return first node in document order, or empty node if node set is empty. + // If PUGIXML_NO_EXCEPTIONS is not defined, throws xpath_exception on type mismatch and std::bad_alloc on out of memory errors. + // If PUGIXML_NO_EXCEPTIONS is defined, returns empty node instead. + xpath_node evaluate_node(const xpath_node& n) const; + + // Get parsing result (used to get compilation errors in PUGIXML_NO_EXCEPTIONS mode) + const xpath_parse_result& result() const; + + // Safe bool conversion operator + operator unspecified_bool_type() const; + + // Borland C++ workaround + bool operator!() const; +}; + +#ifndef PUGIXML_NO_EXCEPTIONS +#if defined(_MSC_VER) +// C4275 can be ignored in Visual C++ if you are deriving +// from a type in the Standard C++ Library +#pragma warning(push) +#pragma warning(disable : 4275) +#endif +// XPath exception class +class PUGIXML_CLASS xpath_exception : public std::exception +{ + private: + xpath_parse_result _result; + + public: + // Construct exception from parse result + explicit xpath_exception(const xpath_parse_result& result); + + // Get error message + virtual const char* what() const throw() PUGIXML_OVERRIDE; + + // Get parse result + const xpath_parse_result& result() const; +}; +#if defined(_MSC_VER) +#pragma warning(pop) +#endif +#endif + +// XPath node class (either xml_node or xml_attribute) +class PUGIXML_CLASS xpath_node +{ + private: + xml_node _node; + xml_attribute _attribute; + + typedef void (*unspecified_bool_type)(xpath_node***); + + public: + // Default constructor; constructs empty XPath node + xpath_node(); + + // Construct XPath node from XML node/attribute + xpath_node(const xml_node& node); + xpath_node(const xml_attribute& attribute, const xml_node& parent); + + // Get node/attribute, if any + xml_node node() const; + xml_attribute attribute() const; + + // Get parent of contained node/attribute + xml_node parent() const; + + // Safe bool conversion operator + operator unspecified_bool_type() const; + + // Borland C++ workaround + bool operator!() const; + + // Comparison operators + bool operator==(const xpath_node& n) const; + bool operator!=(const xpath_node& n) const; +}; + +#ifdef __BORLANDC__ +// Borland C++ workaround +bool PUGIXML_FUNCTION operator&&(const xpath_node& lhs, bool rhs); +bool PUGIXML_FUNCTION operator||(const xpath_node& lhs, bool rhs); +#endif + +// A fixed-size collection of XPath nodes +class PUGIXML_CLASS xpath_node_set +{ + public: + // Collection type + enum type_t + { + type_unsorted, // Not ordered + type_sorted, // Sorted by document order (ascending) + type_sorted_reverse // Sorted by document order (descending) }; + + // Constant iterator type + typedef const xpath_node* const_iterator; + + // We define non-constant iterator to be the same as constant iterator so that various generic algorithms (i.e. boost foreach) work + typedef const xpath_node* iterator; + + // Default constructor. Constructs empty set. + xpath_node_set(); + + // Constructs a set from iterator range; data is not checked for duplicates and is not sorted according to provided type, so be careful + xpath_node_set(const_iterator begin, const_iterator end, type_t type = type_unsorted); + + // Destructor + ~xpath_node_set(); + + // Copy constructor/assignment operator + xpath_node_set(const xpath_node_set& ns); + xpath_node_set& operator=(const xpath_node_set& ns); + +#ifdef PUGIXML_HAS_MOVE + // Move semantics support + xpath_node_set(xpath_node_set&& rhs) PUGIXML_NOEXCEPT; + xpath_node_set& operator=(xpath_node_set&& rhs) PUGIXML_NOEXCEPT; +#endif + + // Get collection type + type_t type() const; + + // Get collection size + size_t size() const; + + // Indexing operator + const xpath_node& operator[](size_t index) const; + + // Collection iterators + const_iterator begin() const; + const_iterator end() const; + + // Sort the collection in ascending/descending order by document order + void sort(bool reverse = false); + + // Get first node in the collection by document order + xpath_node first() const; + + // Check if collection is empty + bool empty() const; + + private: + type_t _type; + + xpath_node _storage[1]; + + xpath_node* _begin; + xpath_node* _end; + + void _assign(const_iterator begin, const_iterator end, type_t type); + void _move(xpath_node_set& rhs) PUGIXML_NOEXCEPT; +}; #endif #ifndef PUGIXML_NO_STL - // Convert wide string to UTF8 - std::basic_string, std::allocator > PUGIXML_FUNCTION as_utf8(const wchar_t* str); - std::basic_string, std::allocator > PUGIXML_FUNCTION as_utf8(const std::basic_string, std::allocator >& str); +// Convert wide string to UTF8 +std::basic_string, std::allocator> PUGIXML_FUNCTION as_utf8(const wchar_t* str); +std::basic_string, std::allocator> PUGIXML_FUNCTION as_utf8(const std::basic_string, std::allocator>& str); - // Convert UTF8 to wide string - std::basic_string, std::allocator > PUGIXML_FUNCTION as_wide(const char* str); - std::basic_string, std::allocator > PUGIXML_FUNCTION as_wide(const std::basic_string, std::allocator >& str); +// Convert UTF8 to wide string +std::basic_string, std::allocator> PUGIXML_FUNCTION as_wide(const char* str); +std::basic_string, std::allocator> PUGIXML_FUNCTION as_wide(const std::basic_string, std::allocator>& str); #endif - // Memory allocation function interface; returns pointer to allocated memory or NULL on failure - typedef void* (*allocation_function)(size_t size); +// Memory allocation function interface; returns pointer to allocated memory or NULL on failure +typedef void* (*allocation_function)(size_t size); - // Memory deallocation function interface - typedef void (*deallocation_function)(void* ptr); +// Memory deallocation function interface +typedef void (*deallocation_function)(void* ptr); - // Override default memory management functions. All subsequent allocations/deallocations will be performed via supplied functions. - void PUGIXML_FUNCTION set_memory_management_functions(allocation_function allocate, deallocation_function deallocate); +// Override default memory management functions. All subsequent allocations/deallocations will be performed via supplied functions. +void PUGIXML_FUNCTION set_memory_management_functions(allocation_function allocate, deallocation_function deallocate); - // Get current memory management functions - allocation_function PUGIXML_FUNCTION get_memory_allocation_function(); - deallocation_function PUGIXML_FUNCTION get_memory_deallocation_function(); -} +// Get current memory management functions +allocation_function PUGIXML_FUNCTION get_memory_allocation_function(); +deallocation_function PUGIXML_FUNCTION get_memory_deallocation_function(); +} // namespace pugi #if !defined(PUGIXML_NO_STL) && (defined(_MSC_VER) || defined(__ICC)) namespace std { - // Workarounds for (non-standard) iterator category detection for older versions (MSVC7/IC8 and earlier) - std::bidirectional_iterator_tag PUGIXML_FUNCTION _Iter_cat(const pugi::xml_node_iterator&); - std::bidirectional_iterator_tag PUGIXML_FUNCTION _Iter_cat(const pugi::xml_attribute_iterator&); - std::bidirectional_iterator_tag PUGIXML_FUNCTION _Iter_cat(const pugi::xml_named_node_iterator&); -} +// Workarounds for (non-standard) iterator category detection for older versions (MSVC7/IC8 and earlier) +std::bidirectional_iterator_tag PUGIXML_FUNCTION _Iter_cat(const pugi::xml_node_iterator&); +std::bidirectional_iterator_tag PUGIXML_FUNCTION _Iter_cat(const pugi::xml_attribute_iterator&); +std::bidirectional_iterator_tag PUGIXML_FUNCTION _Iter_cat(const pugi::xml_named_node_iterator&); +} // namespace std #endif #if !defined(PUGIXML_NO_STL) && defined(__SUNPRO_CC) namespace std { - // Workarounds for (non-standard) iterator category detection - std::bidirectional_iterator_tag PUGIXML_FUNCTION __iterator_category(const pugi::xml_node_iterator&); - std::bidirectional_iterator_tag PUGIXML_FUNCTION __iterator_category(const pugi::xml_attribute_iterator&); - std::bidirectional_iterator_tag PUGIXML_FUNCTION __iterator_category(const pugi::xml_named_node_iterator&); -} +// Workarounds for (non-standard) iterator category detection +std::bidirectional_iterator_tag PUGIXML_FUNCTION __iterator_category(const pugi::xml_node_iterator&); +std::bidirectional_iterator_tag PUGIXML_FUNCTION __iterator_category(const pugi::xml_attribute_iterator&); +std::bidirectional_iterator_tag PUGIXML_FUNCTION __iterator_category(const pugi::xml_named_node_iterator&); +} // namespace std #endif #endif @@ -1440,8 +1454,8 @@ namespace std // Make sure implementation is included in header-only mode // Use macro expansion in #include to work around QMake (QTBUG-11923) #if defined(PUGIXML_HEADER_ONLY) && !defined(PUGIXML_SOURCE) -# define PUGIXML_SOURCE "pugixml.cpp" -# include PUGIXML_SOURCE +#define PUGIXML_SOURCE "pugixml.cpp" +#include PUGIXML_SOURCE #endif /** diff --git a/tests/allocator.cpp b/tests/allocator.cpp index 588e456..f7e1711 100644 --- a/tests/allocator.cpp +++ b/tests/allocator.cpp @@ -1,144 +1,146 @@ #include "allocator.hpp" -#include #include #include +#include // Address sanitizer #if defined(__has_feature) -# define ADDRESS_SANITIZER __has_feature(address_sanitizer) +#define ADDRESS_SANITIZER __has_feature(address_sanitizer) #else -# if defined(__SANITIZE_ADDRESS__) -# define ADDRESS_SANITIZER 1 -# else -# define ADDRESS_SANITIZER 0 -# endif +#if defined(__SANITIZE_ADDRESS__) +#define ADDRESS_SANITIZER 1 +#else +#define ADDRESS_SANITIZER 0 +#endif #endif // Low-level allocation functions #if defined(_WIN32) || defined(_WIN64) -# ifdef __MWERKS__ -# pragma ANSI_strict off // disable ANSI strictness to include windows.h -# pragma cpp_extensions on // enable some extensions to include windows.h -# endif +#ifdef __MWERKS__ +#pragma ANSI_strict off // disable ANSI strictness to include windows.h +#pragma cpp_extensions on // enable some extensions to include windows.h +#endif -# if defined(_MSC_VER) -# pragma warning(disable: 4201) // nonstandard extension used: nameless struct/union -# endif +#if defined(_MSC_VER) +#pragma warning(disable : 4201) // nonstandard extension used: nameless struct/union +#endif -# ifdef _XBOX_VER -# define NOD3D -# include -# else -# include -# endif +#ifdef _XBOX_VER +#define NOD3D +#include +#else +#include +#endif namespace { - const size_t page_size = 4096; +const size_t page_size = 4096; - size_t align_to_page(size_t value) - { - return (value + page_size - 1) & ~(page_size - 1); - } - - void* allocate_page_aligned(size_t size) - { - // We can't use VirtualAlloc because it has 64Kb granularity so we run out of address space quickly - // We can't use malloc because of occasional problems with CW on CRT termination - static HANDLE heap = HeapCreate(0, 0, 0); - - void* result = HeapAlloc(heap, 0, size + page_size); - - return reinterpret_cast(align_to_page(reinterpret_cast(result))); - } - - void* allocate(size_t size) - { - size_t aligned_size = align_to_page(size); - - void* ptr = allocate_page_aligned(aligned_size + page_size); - if (!ptr) return 0; - - char* end = static_cast(ptr) + aligned_size; - - DWORD old_flags; - VirtualProtect(end, page_size, PAGE_NOACCESS, &old_flags); - - return end - size; - } - - void deallocate(void* ptr, size_t size) - { - size_t aligned_size = align_to_page(size); - - void* rptr = static_cast(ptr) + size - aligned_size; - - DWORD old_flags; - VirtualProtect(rptr, aligned_size + page_size, PAGE_NOACCESS, &old_flags); - } +size_t align_to_page(size_t value) +{ + return (value + page_size - 1) & ~(page_size - 1); } + +void* allocate_page_aligned(size_t size) +{ + // We can't use VirtualAlloc because it has 64Kb granularity so we run out of address space quickly + // We can't use malloc because of occasional problems with CW on CRT termination + static HANDLE heap = HeapCreate(0, 0, 0); + + void* result = HeapAlloc(heap, 0, size + page_size); + + return reinterpret_cast(align_to_page(reinterpret_cast(result))); +} + +void* allocate(size_t size) +{ + size_t aligned_size = align_to_page(size); + + void* ptr = allocate_page_aligned(aligned_size + page_size); + if (!ptr) + return 0; + + char* end = static_cast(ptr) + aligned_size; + + DWORD old_flags; + VirtualProtect(end, page_size, PAGE_NOACCESS, &old_flags); + + return end - size; +} + +void deallocate(void* ptr, size_t size) +{ + size_t aligned_size = align_to_page(size); + + void* rptr = static_cast(ptr) + size - aligned_size; + + DWORD old_flags; + VirtualProtect(rptr, aligned_size + page_size, PAGE_NOACCESS, &old_flags); +} +} // namespace #elif (defined(__APPLE__) || defined(__linux__)) && (defined(__i386) || defined(__x86_64)) && !ADDRESS_SANITIZER -# include +#include namespace { - const size_t page_size = 4096; +const size_t page_size = 4096; - size_t align_to_page(size_t value) - { - return (value + page_size - 1) & ~(page_size - 1); - } - - void* allocate_page_aligned(size_t size) - { - void* result = malloc(size + page_size); - - return reinterpret_cast(align_to_page(reinterpret_cast(result))); - } - - void* allocate(size_t size) - { - size_t aligned_size = align_to_page(size); - - void* ptr = allocate_page_aligned(aligned_size + page_size); - if (!ptr) return 0; - - char* end = static_cast(ptr) + aligned_size; - - int res = mprotect(end, page_size, PROT_NONE); - assert(res == 0); - (void)!res; - - return end - size; - } - - void deallocate(void* ptr, size_t size) - { - size_t aligned_size = align_to_page(size); - - void* rptr = static_cast(ptr) + size - aligned_size; - - int res = mprotect(rptr, aligned_size + page_size, PROT_NONE); - assert(res == 0); - (void)!res; - } +size_t align_to_page(size_t value) +{ + return (value + page_size - 1) & ~(page_size - 1); } + +void* allocate_page_aligned(size_t size) +{ + void* result = malloc(size + page_size); + + return reinterpret_cast(align_to_page(reinterpret_cast(result))); +} + +void* allocate(size_t size) +{ + size_t aligned_size = align_to_page(size); + + void* ptr = allocate_page_aligned(aligned_size + page_size); + if (!ptr) + return 0; + + char* end = static_cast(ptr) + aligned_size; + + int res = mprotect(end, page_size, PROT_NONE); + assert(res == 0); + (void)!res; + + return end - size; +} + +void deallocate(void* ptr, size_t size) +{ + size_t aligned_size = align_to_page(size); + + void* rptr = static_cast(ptr) + size - aligned_size; + + int res = mprotect(rptr, aligned_size + page_size, PROT_NONE); + assert(res == 0); + (void)!res; +} +} // namespace #else namespace { - void* allocate(size_t size) - { - return malloc(size); - } - - void deallocate(void* ptr, size_t size) - { - (void)size; - - free(ptr); - } +void* allocate(size_t size) +{ + return malloc(size); } + +void deallocate(void* ptr, size_t size) +{ + (void)size; + + free(ptr); +} +} // namespace #endif // High-level allocation functions @@ -147,7 +149,8 @@ const size_t memory_alignment = sizeof(double) > sizeof(void*) ? sizeof(double) void* memory_allocate(size_t size) { void* result = allocate(size + memory_alignment); - if (!result) return 0; + if (!result) + return 0; memcpy(result, &size, sizeof(size_t)); @@ -166,10 +169,10 @@ size_t memory_size(void* ptr) void memory_deallocate(void* ptr) { - if (!ptr) return; + if (!ptr) + return; size_t size = memory_size(ptr); deallocate(static_cast(ptr) - memory_alignment, size + memory_alignment); } - diff --git a/tests/fuzz_parse.cpp b/tests/fuzz_parse.cpp index 94c610a..5cecb9c 100644 --- a/tests/fuzz_parse.cpp +++ b/tests/fuzz_parse.cpp @@ -2,9 +2,9 @@ #include -extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) +extern "C" int LLVMFuzzerTestOneInput(const uint8_t* Data, size_t Size) { - pugi::xml_document doc; + pugi::xml_document doc; doc.load_buffer(Data, Size); doc.load_buffer(Data, Size, pugi::parse_minimal); diff --git a/tests/fuzz_xpath.cpp b/tests/fuzz_xpath.cpp index c7ff4cd..745cf0d 100644 --- a/tests/fuzz_xpath.cpp +++ b/tests/fuzz_xpath.cpp @@ -3,7 +3,7 @@ #include #include -extern "C" int LLVMFuzzerTestOneInput(const uint8_t *Data, size_t Size) +extern "C" int LLVMFuzzerTestOneInput(const uint8_t* Data, size_t Size) { char* text = new char[Size + 1]; memcpy(text, Data, Size); diff --git a/tests/helpers.hpp b/tests/helpers.hpp index 6dfdc4b..48381a7 100644 --- a/tests/helpers.hpp +++ b/tests/helpers.hpp @@ -5,7 +5,8 @@ #include -template static void generic_bool_ops_test(const T& obj) +template +static void generic_bool_ops_test(const T& obj) { T null; @@ -14,14 +15,14 @@ template static void generic_bool_ops_test(const T& obj) CHECK(!!obj); #ifdef _MSC_VER -# pragma warning(push) -# pragma warning(disable: 4800) // forcing value to bool 'true' or 'false' (performance warning) - we really want to just cast to bool instead of !! +#pragma warning(push) +#pragma warning(disable : 4800) // forcing value to bool 'true' or 'false' (performance warning) - we really want to just cast to bool instead of !! #endif bool b1 = null, b2 = obj; #ifdef _MSC_VER -# pragma warning(pop) +#pragma warning(pop) #endif CHECK(!b1); @@ -33,7 +34,8 @@ template static void generic_bool_ops_test(const T& obj) CHECK(obj || obj); } -template static void generic_eq_ops_test(const T& obj1, const T& obj2) +template +static void generic_eq_ops_test(const T& obj1, const T& obj2) { T null = T(); @@ -54,7 +56,8 @@ template static void generic_eq_ops_test(const T& obj1, const T& ob CHECK(!(T(obj1) != obj1)); } -template static void generic_rel_ops_test(T obj1, T obj2) +template +static void generic_rel_ops_test(T obj1, T obj2) { T null = T(); @@ -103,7 +106,8 @@ template static void generic_rel_ops_test(T obj1, T obj2) CHECK(!(obj1 >= obj2)); } -template static void generic_empty_test(const T& obj) +template +static void generic_empty_test(const T& obj) { T null; diff --git a/tests/main.cpp b/tests/main.cpp index 352b58b..be6599a 100644 --- a/tests/main.cpp +++ b/tests/main.cpp @@ -1,21 +1,21 @@ -#include "test.hpp" #include "allocator.hpp" +#include "test.hpp" +#include +#include #include #include -#include -#include #include #ifndef PUGIXML_NO_EXCEPTIONS -# include +#include #endif #ifdef _WIN32_WCE -# undef DebugBreak -# pragma warning(disable: 4201) // nonstandard extension used: nameless struct/union -# include +#undef DebugBreak +#pragma warning(disable : 4201) // nonstandard extension used: nameless struct/union +#include #endif test_runner* test_runner::_tests = 0; @@ -41,7 +41,8 @@ static void* custom_allocate(size_t size) else { void* ptr = memory_allocate(size); - if (!ptr) return 0; + if (!ptr) + return 0; g_memory_total_size += memory_size(ptr); g_memory_total_count++; @@ -89,9 +90,9 @@ static void replace_memory_management() namespace std { - _CRTIMP2 _Prhand _Raise_handler; - _CRTIMP2 void __cdecl _Throw(const exception&) {} -} +_CRTIMP2 _Prhand _Raise_handler; +_CRTIMP2 void __cdecl _Throw(const exception&) {} +} // namespace std #endif static bool run_test(test_runner* test, const char* test_name, pugi::allocation_function allocate) @@ -109,15 +110,15 @@ static bool run_test(test_runner* test, const char* test_name, pugi::allocation_ pugi::set_memory_management_functions(allocate, custom_deallocate); #ifdef _MSC_VER -# pragma warning(push) -# pragma warning(disable: 4611) // interaction between _setjmp and C++ object destruction is non-portable -# pragma warning(disable: 4793) // function compiled as native: presence of '_setjmp' makes a function unmanaged +#pragma warning(push) +#pragma warning(disable : 4611) // interaction between _setjmp and C++ object destruction is non-portable +#pragma warning(disable : 4793) // function compiled as native: presence of '_setjmp' makes a function unmanaged #endif volatile int result = setjmp(test_runner::_failure_buffer); #ifdef _MSC_VER -# pragma warning(pop) +#pragma warning(pop) #endif if (result) @@ -193,10 +194,10 @@ int main(int, char** argv) if (g_memory_fail_triggered) { // run tests that trigger memory failures twice - with an allocator that returns NULL and with an allocator that throws - #ifndef PUGIXML_NO_EXCEPTIONS +#ifndef PUGIXML_NO_EXCEPTIONS total++; passed += run_test(test, (test->_name + std::string(" (throw)")).c_str(), custom_allocate_throw); - #endif +#endif } } diff --git a/tests/test.cpp b/tests/test.cpp index 6347984..1ae29df 100644 --- a/tests/test.cpp +++ b/tests/test.cpp @@ -5,8 +5,8 @@ #include "writer_string.hpp" -#include #include +#include #include #include @@ -33,10 +33,12 @@ static void build_document_order(std::vector& result, pugi::xm cur = cur.next_sibling(); else { - while (cur && !cur.next_sibling()) cur = cur.parent(); + while (cur && !cur.next_sibling()) + cur = cur.parent(); cur = cur.next_sibling(); - if (!cur) break; + if (!cur) + break; } } } @@ -45,11 +47,11 @@ static void build_document_order(std::vector& result, pugi::xm bool test_string_equal(const pugi::char_t* lhs, const pugi::char_t* rhs) { return (!lhs || !rhs) ? lhs == rhs : - #ifdef PUGIXML_WCHAR_MODE - wcscmp(lhs, rhs) == 0; - #else - strcmp(lhs, rhs) == 0; - #endif +#ifdef PUGIXML_WCHAR_MODE + wcscmp(lhs, rhs) == 0; +#else + strcmp(lhs, rhs) == 0; +#endif } bool test_node(const pugi::xml_node& node, const pugi::char_t* contents, const pugi::char_t* indent, unsigned int flags) @@ -83,7 +85,8 @@ static size_t strlength(const pugi::char_t* s) bool test_xpath_string(const pugi::xpath_node& node, const pugi::char_t* query, pugi::xpath_variable_set* variables, const pugi::char_t* expected) { pugi::xpath_query q(query, variables); - if (!q) return false; + if (!q) + return false; const size_t capacity = 64; pugi::char_t result[capacity]; @@ -104,7 +107,8 @@ bool test_xpath_string(const pugi::xpath_node& node, const pugi::char_t* query, bool test_xpath_boolean(const pugi::xpath_node& node, const pugi::char_t* query, pugi::xpath_variable_set* variables, bool expected) { pugi::xpath_query q(query, variables); - if (!q) return false; + if (!q) + return false; return q.evaluate_boolean(node) == expected; } @@ -112,7 +116,8 @@ bool test_xpath_boolean(const pugi::xpath_node& node, const pugi::char_t* query, bool test_xpath_number(const pugi::xpath_node& node, const pugi::char_t* query, pugi::xpath_variable_set* variables, double expected) { pugi::xpath_query q(query, variables); - if (!q) return false; + if (!q) + return false; double value = q.evaluate_number(node); double absolute_error = fabs(value - expected); @@ -124,7 +129,8 @@ bool test_xpath_number(const pugi::xpath_node& node, const pugi::char_t* query, bool test_xpath_number_nan(const pugi::xpath_node& node, const pugi::char_t* query, pugi::xpath_variable_set* variables) { pugi::xpath_query q(query, variables); - if (!q) return false; + if (!q) + return false; return test_double_nan(q.evaluate_number(node)); } @@ -155,12 +161,15 @@ void xpath_node_set_tester::check(bool condition) } } -xpath_node_set_tester::xpath_node_set_tester(const pugi::xpath_node_set& set, const char* message_): last(0), message(message_) +xpath_node_set_tester::xpath_node_set_tester(const pugi::xpath_node_set& set, const char* message_) + : last(0) + , message(message_) { result = set; // only sort unsorted sets so that we're able to verify reverse order for some axes - if (result.type() == pugi::xpath_node_set::type_unsorted) result.sort(); + if (result.type() == pugi::xpath_node_set::type_unsorted) + result.sort(); if (result.empty()) { diff --git a/tests/test.hpp b/tests/test.hpp index dd14af6..3785113 100644 --- a/tests/test.hpp +++ b/tests/test.hpp @@ -36,7 +36,8 @@ struct test_runner bool test_string_equal(const pugi::char_t* lhs, const pugi::char_t* rhs); -template inline bool test_node_name_value(const Node& node, const pugi::char_t* name, const pugi::char_t* value) +template +inline bool test_node_name_value(const Node& node, const pugi::char_t* name, const pugi::char_t* value) { return test_string_equal(node.name(), name) && test_string_equal(node.value(), value); } @@ -71,17 +72,22 @@ struct xpath_node_set_tester #endif -struct dummy_fixture {}; +struct dummy_fixture +{ +}; #define TEST_FIXTURE(name, fixture) \ - struct test_runner_helper_##name: fixture \ + struct test_runner_helper_##name : fixture \ { \ void run(); \ }; \ - static struct test_runner_##name: test_runner \ + static struct test_runner_##name : test_runner \ { \ - test_runner_##name(): test_runner(#name) {} \ - \ + test_runner_##name() \ + : test_runner(#name) \ + { \ + } \ +\ virtual void run() PUGIXML_OVERRIDE \ { \ test_runner_helper_##name helper; \ @@ -96,30 +102,34 @@ struct dummy_fixture {}; struct test_fixture_##name \ { \ pugi::xml_document doc; \ - \ +\ test_fixture_##name() \ { \ CHECK(doc.load_string(PUGIXML_TEXT(xml), flags)); \ } \ - \ - private: \ +\ + private: \ test_fixture_##name(const test_fixture_##name&); \ test_fixture_##name& operator=(const test_fixture_##name&); \ }; \ - \ +\ TEST_FIXTURE(name, test_fixture_##name) #define TEST_XML(name, xml) TEST_XML_FLAGS(name, xml, pugi::parse_default) #define CHECK_JOIN(text, file, line) text " at " file ":" #line #define CHECK_JOIN2(text, file, line) CHECK_JOIN(text, file, line) -#define CHECK_TEXT(condition, text) if (condition) ; else test_runner::_failure_message = CHECK_JOIN2(text, __FILE__, __LINE__), longjmp(test_runner::_failure_buffer, 1) +#define CHECK_TEXT(condition, text) \ + if (condition) \ + ; \ + else \ + test_runner::_failure_message = CHECK_JOIN2(text, __FILE__, __LINE__), longjmp(test_runner::_failure_buffer, 1) #define CHECK_FORCE_FAIL(text) test_runner::_failure_message = CHECK_JOIN2(text, __FILE__, __LINE__), longjmp(test_runner::_failure_buffer, 1) #if (defined(_MSC_VER) && _MSC_VER == 1200) || defined(__MWERKS__) || (defined(__BORLANDC__) && __BORLANDC__ <= 0x540) -# define STRINGIZE(value) "??" // Some compilers have issues with stringizing expressions that contain strings w/escaping inside +#define STRINGIZE(value) "??" // Some compilers have issues with stringizing expressions that contain strings w/escaping inside #else -# define STRINGIZE(value) #value +#define STRINGIZE(value) #value #endif #define CHECK(condition) CHECK_TEXT(condition, STRINGIZE(condition) " is false") @@ -147,9 +157,29 @@ struct dummy_fixture {}; #endif #ifdef PUGIXML_NO_EXCEPTIONS -#define CHECK_ALLOC_FAIL(code) do { CHECK(!test_runner::_memory_fail_triggered); code; CHECK(test_runner::_memory_fail_triggered); test_runner::_memory_fail_triggered = false; } while (test_runner::_memory_fail_triggered) +#define CHECK_ALLOC_FAIL(code) \ + do \ + { \ + CHECK(!test_runner::_memory_fail_triggered); \ + code; \ + CHECK(test_runner::_memory_fail_triggered); \ + test_runner::_memory_fail_triggered = false; \ + } while (test_runner::_memory_fail_triggered) #else -#define CHECK_ALLOC_FAIL(code) do { CHECK(!test_runner::_memory_fail_triggered); try { code; } catch (std::bad_alloc&) {} CHECK(test_runner::_memory_fail_triggered); test_runner::_memory_fail_triggered = false; } while (test_runner::_memory_fail_triggered) +#define CHECK_ALLOC_FAIL(code) \ + do \ + { \ + CHECK(!test_runner::_memory_fail_triggered); \ + try \ + { \ + code; \ + } \ + catch (std::bad_alloc&) \ + { \ + } \ + CHECK(test_runner::_memory_fail_triggered); \ + test_runner::_memory_fail_triggered = false; \ + } while (test_runner::_memory_fail_triggered) #endif #define STR(text) PUGIXML_TEXT(text) @@ -161,7 +191,7 @@ struct dummy_fixture {}; #if (defined(_MSC_VER) && _MSC_VER == 1200) || (defined(__INTEL_COMPILER) && __INTEL_COMPILER == 800) || defined(__BORLANDC__) // NaN comparison on MSVC6 is incorrect, see http://www.nabble.com/assertDoubleEquals,-NaN---Microsoft-Visual-Studio-6-td9137859.html // IC8 and BCC are also affected by the same bug -# define MSVC6_NAN_BUG +#define MSVC6_NAN_BUG #endif inline wchar_t wchar_cast(unsigned int value) diff --git a/tests/test_document.cpp b/tests/test_document.cpp index 310eba9..c5bebb2 100644 --- a/tests/test_document.cpp +++ b/tests/test_document.cpp @@ -15,19 +15,19 @@ #include #include -#include #include +#include #ifndef PUGIXML_NO_EXCEPTIONS -# include +#include #endif #ifdef __MINGW32__ -# include // for unlink in C++0x mode +#include // for unlink in C++0x mode #endif #if defined(__CELLOS_LV2__) || defined(ANDROID) || defined(_GLIBCXX_HAVE_UNISTD_H) || defined(__APPLE__) -# include // for unlink +#include // for unlink #endif using namespace pugi; @@ -35,7 +35,8 @@ using namespace pugi; static bool load_file_in_memory(const char* path, char*& data, size_t& size) { FILE* file = fopen(path, "rb"); - if (!file) return false; + if (!file) + return false; fseek(file, 0, SEEK_END); long length = ftell(file); @@ -53,15 +54,16 @@ static bool load_file_in_memory(const char* path, char*& data, size_t& size) static bool test_file_contents(const char* path, const char* data, size_t size) { - char* fdata; - size_t fsize; - if (!load_file_in_memory(path, fdata, fsize)) return false; + char* fdata; + size_t fsize; + if (!load_file_in_memory(path, fdata, fsize)) + return false; - bool result = (size == fsize && memcmp(data, fdata, size) == 0); + bool result = (size == fsize && memcmp(data, fdata, size) == 0); - delete[] fdata; + delete[] fdata; - return result; + return result; } TEST(document_create_empty) @@ -199,114 +201,120 @@ TEST(document_load_stream_wide_error_previous) CHECK(!doc.first_child()); } -template class char_array_buffer: public std::basic_streambuf +template +class char_array_buffer : public std::basic_streambuf { -public: - char_array_buffer(T* begin, T* end) - { - this->setg(begin, begin, end); - } + public: + char_array_buffer(T* begin, T* end) + { + this->setg(begin, begin, end); + } }; TEST(document_load_stream_nonseekable) { - char contents[] = ""; - char_array_buffer buffer(contents, contents + sizeof(contents) / sizeof(contents[0])); - std::istream in(&buffer); + char contents[] = ""; + char_array_buffer buffer(contents, contents + sizeof(contents) / sizeof(contents[0])); + std::istream in(&buffer); - xml_document doc; - CHECK(doc.load(in)); - CHECK_NODE(doc, STR("")); + xml_document doc; + CHECK(doc.load(in)); + CHECK_NODE(doc, STR("")); } TEST(document_load_stream_wide_nonseekable) { - wchar_t contents[] = L""; - char_array_buffer buffer(contents, contents + sizeof(contents) / sizeof(contents[0])); - std::basic_istream in(&buffer); + wchar_t contents[] = L""; + char_array_buffer buffer(contents, contents + sizeof(contents) / sizeof(contents[0])); + std::basic_istream in(&buffer); - xml_document doc; - CHECK(doc.load(in)); - CHECK_NODE(doc, STR("")); + xml_document doc; + CHECK(doc.load(in)); + CHECK_NODE(doc, STR("")); } TEST(document_load_stream_nonseekable_large) { std::basic_string str; str += STR(""); - for (int i = 0; i < 10000; ++i) str += STR(""); + for (int i = 0; i < 10000; ++i) + str += STR(""); str += STR(""); - char_array_buffer buffer(&str[0], &str[0] + str.length()); - std::basic_istream in(&buffer); + char_array_buffer buffer(&str[0], &str[0] + str.length()); + std::basic_istream in(&buffer); - xml_document doc; - CHECK(doc.load(in)); - CHECK_NODE(doc, str.c_str()); + xml_document doc; + CHECK(doc.load(in)); + CHECK_NODE(doc, str.c_str()); } TEST(document_load_stream_nonseekable_out_of_memory) { - char contents[] = ""; - char_array_buffer buffer(contents, contents + sizeof(contents) / sizeof(contents[0])); - std::istream in(&buffer); + char contents[] = ""; + char_array_buffer buffer(contents, contents + sizeof(contents) / sizeof(contents[0])); + std::istream in(&buffer); - test_runner::_memory_fail_threshold = 1; + test_runner::_memory_fail_threshold = 1; - xml_document doc; - CHECK_ALLOC_FAIL(CHECK(doc.load(in).status == status_out_of_memory)); + xml_document doc; + CHECK_ALLOC_FAIL(CHECK(doc.load(in).status == status_out_of_memory)); } TEST(document_load_stream_wide_nonseekable_out_of_memory) { - wchar_t contents[] = L""; - char_array_buffer buffer(contents, contents + sizeof(contents) / sizeof(contents[0])); - std::basic_istream in(&buffer); + wchar_t contents[] = L""; + char_array_buffer buffer(contents, contents + sizeof(contents) / sizeof(contents[0])); + std::basic_istream in(&buffer); - test_runner::_memory_fail_threshold = 1; + test_runner::_memory_fail_threshold = 1; - xml_document doc; - CHECK_ALLOC_FAIL(CHECK(doc.load(in).status == status_out_of_memory)); + xml_document doc; + CHECK_ALLOC_FAIL(CHECK(doc.load(in).status == status_out_of_memory)); } TEST(document_load_stream_nonseekable_out_of_memory_large) { std::basic_string str; str += ""; - for (int i = 0; i < 10000; ++i) str += ""; + for (int i = 0; i < 10000; ++i) + str += ""; str += ""; - char_array_buffer buffer(&str[0], &str[0] + str.length()); - std::basic_istream in(&buffer); + char_array_buffer buffer(&str[0], &str[0] + str.length()); + std::basic_istream in(&buffer); - test_runner::_memory_fail_threshold = 32768 * 3 + 4096; + test_runner::_memory_fail_threshold = 32768 * 3 + 4096; - xml_document doc; - CHECK_ALLOC_FAIL(CHECK(doc.load(in).status == status_out_of_memory)); + xml_document doc; + CHECK_ALLOC_FAIL(CHECK(doc.load(in).status == status_out_of_memory)); } TEST(document_load_stream_wide_nonseekable_out_of_memory_large) { std::basic_string str; str += L""; - for (int i = 0; i < 10000; ++i) str += L""; + for (int i = 0; i < 10000; ++i) + str += L""; str += L""; - char_array_buffer buffer(&str[0], &str[0] + str.length()); - std::basic_istream in(&buffer); + char_array_buffer buffer(&str[0], &str[0] + str.length()); + std::basic_istream in(&buffer); - test_runner::_memory_fail_threshold = 32768 * 3 * sizeof(wchar_t) + 4096; + test_runner::_memory_fail_threshold = 32768 * 3 * sizeof(wchar_t) + 4096; - xml_document doc; - CHECK_ALLOC_FAIL(CHECK(doc.load(in).status == status_out_of_memory)); + xml_document doc; + CHECK_ALLOC_FAIL(CHECK(doc.load(in).status == status_out_of_memory)); } -template class seek_fail_buffer: public std::basic_streambuf +template +class seek_fail_buffer : public std::basic_streambuf { -public: + public: int seeks; - seek_fail_buffer(): seeks(0) + seek_fail_buffer() + : seeks(0) { } @@ -321,28 +329,30 @@ public: TEST(document_load_stream_seekable_fail_seek) { - seek_fail_buffer buffer; - std::basic_istream in(&buffer); + seek_fail_buffer buffer; + std::basic_istream in(&buffer); - xml_document doc; - CHECK(doc.load(in).status == status_io_error); + xml_document doc; + CHECK(doc.load(in).status == status_io_error); } TEST(document_load_stream_wide_seekable_fail_seek) { - seek_fail_buffer buffer; - std::basic_istream in(&buffer); + seek_fail_buffer buffer; + std::basic_istream in(&buffer); - xml_document doc; - CHECK(doc.load(in).status == status_io_error); + xml_document doc; + CHECK(doc.load(in).status == status_io_error); } -template class tell_fail_buffer: public std::basic_streambuf +template +class tell_fail_buffer : public std::basic_streambuf { -public: + public: int seeks; - tell_fail_buffer(): seeks(0) + tell_fail_buffer() + : seeks(0) { } @@ -361,80 +371,89 @@ public: TEST(document_load_stream_seekable_fail_tell) { - tell_fail_buffer buffer; - std::basic_istream in(&buffer); + tell_fail_buffer buffer; + std::basic_istream in(&buffer); - xml_document doc; - CHECK(doc.load(in).status == status_io_error); + xml_document doc; + CHECK(doc.load(in).status == status_io_error); } TEST(document_load_stream_wide_seekable_fail_tell) { - tell_fail_buffer buffer; - std::basic_istream in(&buffer); + tell_fail_buffer buffer; + std::basic_istream in(&buffer); - xml_document doc; - CHECK(doc.load(in).status == status_io_error); + xml_document doc; + CHECK(doc.load(in).status == status_io_error); } #ifndef PUGIXML_NO_EXCEPTIONS -template class read_fail_buffer: public std::basic_streambuf +template +class read_fail_buffer : public std::basic_streambuf { -public: - typename std::basic_streambuf::int_type underflow() PUGIXML_OVERRIDE + public: + typename std::basic_streambuf::int_type underflow() PUGIXML_OVERRIDE { throw std::runtime_error("underflow failed"); - #ifdef __DMC__ +#ifdef __DMC__ return 0; - #endif +#endif } }; TEST(document_load_stream_nonseekable_fail_read) { - read_fail_buffer buffer; - std::basic_istream in(&buffer); + read_fail_buffer buffer; + std::basic_istream in(&buffer); - xml_document doc; - CHECK(doc.load(in).status == status_io_error); + xml_document doc; + CHECK(doc.load(in).status == status_io_error); } TEST(document_load_stream_wide_nonseekable_fail_read) { - read_fail_buffer buffer; - std::basic_istream in(&buffer); + read_fail_buffer buffer; + std::basic_istream in(&buffer); - xml_document doc; - CHECK(doc.load(in).status == status_io_error); + xml_document doc; + CHECK(doc.load(in).status == status_io_error); } -template class read_fail_seekable_buffer: public std::basic_streambuf +template +class read_fail_seekable_buffer : public std::basic_streambuf { -public: + public: typename std::basic_streambuf::pos_type offset; - read_fail_seekable_buffer(): offset(0) - { - } + read_fail_seekable_buffer() + : offset(0) + { + } - typename std::basic_streambuf::int_type underflow() PUGIXML_OVERRIDE + typename std::basic_streambuf::int_type underflow() PUGIXML_OVERRIDE { throw std::runtime_error("underflow failed"); - #ifdef __DMC__ +#ifdef __DMC__ return 0; - #endif +#endif } typename std::basic_streambuf::pos_type seekoff(typename std::basic_streambuf::off_type off, std::ios_base::seekdir dir, std::ios_base::openmode) PUGIXML_OVERRIDE { switch (dir) { - case std::ios_base::beg: offset = off; break; - case std::ios_base::cur: offset += off; break; - case std::ios_base::end: offset = 16 + off; break; - default: ; + case std::ios_base::beg: + offset = off; + break; + case std::ios_base::cur: + offset += off; + break; + case std::ios_base::end: + offset = 16 + off; + break; + default:; } return offset; } @@ -448,20 +467,20 @@ public: TEST(document_load_stream_seekable_fail_read) { - read_fail_seekable_buffer buffer; - std::basic_istream in(&buffer); + read_fail_seekable_buffer buffer; + std::basic_istream in(&buffer); - xml_document doc; - CHECK(doc.load(in).status == status_io_error); + xml_document doc; + CHECK(doc.load(in).status == status_io_error); } TEST(document_load_stream_wide_seekable_fail_read) { - read_fail_seekable_buffer buffer; - std::basic_istream in(&buffer); + read_fail_seekable_buffer buffer; + std::basic_istream in(&buffer); - xml_document doc; - CHECK(doc.load(in).status == status_io_error); + xml_document doc; + CHECK(doc.load(in).status == status_io_error); } #endif #endif @@ -498,7 +517,8 @@ TEST(document_load_file_large) std::basic_string str; str += STR(""); - for (int i = 0; i < 10000; ++i) str += STR(""); + for (int i = 0; i < 10000; ++i) + str += STR(""); str += STR(""); CHECK_NODE(doc, str.c_str()); @@ -744,9 +764,9 @@ struct temp_file ~temp_file() { - #ifndef _WIN32_WCE +#ifndef _WIN32_WCE CHECK(unlink(path) == 0); - #endif +#endif } }; @@ -784,10 +804,10 @@ TEST_XML(document_save_file_text, "") temp_file f; CHECK(doc.save_file(f.path, STR(""), format_no_declaration | format_save_file_text)); - CHECK(test_file_contents(f.path, "\n", 9) || test_file_contents(f.path, "\r\n", 10)); + CHECK(test_file_contents(f.path, "\n", 9) || test_file_contents(f.path, "\r\n", 10)); CHECK(doc.save_file(f.path, STR(""), format_no_declaration)); - CHECK(test_file_contents(f.path, "\n", 9)); + CHECK(test_file_contents(f.path, "\n", 9)); } TEST_XML(document_save_file_wide_text, "") @@ -799,10 +819,10 @@ TEST_XML(document_save_file_wide_text, "") std::copy(f.path, f.path + strlen(f.path) + 1, wpath + 0); CHECK(doc.save_file(wpath, STR(""), format_no_declaration | format_save_file_text)); - CHECK(test_file_contents(f.path, "\n", 9) || test_file_contents(f.path, "\r\n", 10)); + CHECK(test_file_contents(f.path, "\n", 9) || test_file_contents(f.path, "\r\n", 10)); CHECK(doc.save_file(wpath, STR(""), format_no_declaration)); - CHECK(test_file_contents(f.path, "\n", 9)); + CHECK(test_file_contents(f.path, "\n", 9)); } TEST_XML(document_save_file_leak, "") @@ -926,32 +946,30 @@ inline void check_utftest_document(const xml_document& doc) TEST(document_load_file_convert_auto) { const char* files[] = - { - "tests/data/utftest_utf16_be.xml", - "tests/data/utftest_utf16_be_bom.xml", - "tests/data/utftest_utf16_be_nodecl.xml", - "tests/data/utftest_utf16_le.xml", - "tests/data/utftest_utf16_le_bom.xml", - "tests/data/utftest_utf16_le_nodecl.xml", - "tests/data/utftest_utf32_be.xml", - "tests/data/utftest_utf32_be_bom.xml", - "tests/data/utftest_utf32_be_nodecl.xml", - "tests/data/utftest_utf32_le.xml", - "tests/data/utftest_utf32_le_bom.xml", - "tests/data/utftest_utf32_le_nodecl.xml", - "tests/data/utftest_utf8.xml", - "tests/data/utftest_utf8_bom.xml", - "tests/data/utftest_utf8_nodecl.xml" - }; + { + "tests/data/utftest_utf16_be.xml", + "tests/data/utftest_utf16_be_bom.xml", + "tests/data/utftest_utf16_be_nodecl.xml", + "tests/data/utftest_utf16_le.xml", + "tests/data/utftest_utf16_le_bom.xml", + "tests/data/utftest_utf16_le_nodecl.xml", + "tests/data/utftest_utf32_be.xml", + "tests/data/utftest_utf32_be_bom.xml", + "tests/data/utftest_utf32_be_nodecl.xml", + "tests/data/utftest_utf32_le.xml", + "tests/data/utftest_utf32_le_bom.xml", + "tests/data/utftest_utf32_le_nodecl.xml", + "tests/data/utftest_utf8.xml", + "tests/data/utftest_utf8_bom.xml", + "tests/data/utftest_utf8_nodecl.xml"}; xml_encoding encodings[] = - { - encoding_utf16_be, encoding_utf16_be, encoding_utf16_be, - encoding_utf16_le, encoding_utf16_le, encoding_utf16_le, - encoding_utf32_be, encoding_utf32_be, encoding_utf32_be, - encoding_utf32_le, encoding_utf32_le, encoding_utf32_le, - encoding_utf8, encoding_utf8, encoding_utf8 - }; + { + encoding_utf16_be, encoding_utf16_be, encoding_utf16_be, + encoding_utf16_le, encoding_utf16_le, encoding_utf16_le, + encoding_utf32_be, encoding_utf32_be, encoding_utf32_be, + encoding_utf32_le, encoding_utf32_le, encoding_utf32_le, + encoding_utf8, encoding_utf8, encoding_utf8}; for (unsigned int i = 0; i < sizeof(files) / sizeof(files[0]); ++i) { @@ -967,32 +985,30 @@ TEST(document_load_file_convert_auto) TEST(document_load_file_convert_specific) { const char* files[] = - { - "tests/data/utftest_utf16_be.xml", - "tests/data/utftest_utf16_be_bom.xml", - "tests/data/utftest_utf16_be_nodecl.xml", - "tests/data/utftest_utf16_le.xml", - "tests/data/utftest_utf16_le_bom.xml", - "tests/data/utftest_utf16_le_nodecl.xml", - "tests/data/utftest_utf32_be.xml", - "tests/data/utftest_utf32_be_bom.xml", - "tests/data/utftest_utf32_be_nodecl.xml", - "tests/data/utftest_utf32_le.xml", - "tests/data/utftest_utf32_le_bom.xml", - "tests/data/utftest_utf32_le_nodecl.xml", - "tests/data/utftest_utf8.xml", - "tests/data/utftest_utf8_bom.xml", - "tests/data/utftest_utf8_nodecl.xml" - }; + { + "tests/data/utftest_utf16_be.xml", + "tests/data/utftest_utf16_be_bom.xml", + "tests/data/utftest_utf16_be_nodecl.xml", + "tests/data/utftest_utf16_le.xml", + "tests/data/utftest_utf16_le_bom.xml", + "tests/data/utftest_utf16_le_nodecl.xml", + "tests/data/utftest_utf32_be.xml", + "tests/data/utftest_utf32_be_bom.xml", + "tests/data/utftest_utf32_be_nodecl.xml", + "tests/data/utftest_utf32_le.xml", + "tests/data/utftest_utf32_le_bom.xml", + "tests/data/utftest_utf32_le_nodecl.xml", + "tests/data/utftest_utf8.xml", + "tests/data/utftest_utf8_bom.xml", + "tests/data/utftest_utf8_nodecl.xml"}; xml_encoding encodings[] = - { - encoding_utf16_be, encoding_utf16_be, encoding_utf16_be, - encoding_utf16_le, encoding_utf16_le, encoding_utf16_le, - encoding_utf32_be, encoding_utf32_be, encoding_utf32_be, - encoding_utf32_le, encoding_utf32_le, encoding_utf32_le, - encoding_utf8, encoding_utf8, encoding_utf8 - }; + { + encoding_utf16_be, encoding_utf16_be, encoding_utf16_be, + encoding_utf16_le, encoding_utf16_le, encoding_utf16_le, + encoding_utf32_be, encoding_utf32_be, encoding_utf32_be, + encoding_utf32_le, encoding_utf32_le, encoding_utf32_le, + encoding_utf8, encoding_utf8, encoding_utf8}; for (unsigned int i = 0; i < sizeof(files) / sizeof(files[0]); ++i) { @@ -1021,30 +1037,28 @@ TEST(document_load_file_convert_specific) TEST(document_load_file_convert_native_endianness) { const char* files[2][6] = - { - { - "tests/data/utftest_utf16_be.xml", - "tests/data/utftest_utf16_be_bom.xml", - "tests/data/utftest_utf16_be_nodecl.xml", - "tests/data/utftest_utf32_be.xml", - "tests/data/utftest_utf32_be_bom.xml", - "tests/data/utftest_utf32_be_nodecl.xml", - }, - { - "tests/data/utftest_utf16_le.xml", - "tests/data/utftest_utf16_le_bom.xml", - "tests/data/utftest_utf16_le_nodecl.xml", - "tests/data/utftest_utf32_le.xml", - "tests/data/utftest_utf32_le_bom.xml", - "tests/data/utftest_utf32_le_nodecl.xml", - } - }; + { + { + "tests/data/utftest_utf16_be.xml", + "tests/data/utftest_utf16_be_bom.xml", + "tests/data/utftest_utf16_be_nodecl.xml", + "tests/data/utftest_utf32_be.xml", + "tests/data/utftest_utf32_be_bom.xml", + "tests/data/utftest_utf32_be_nodecl.xml", + }, + { + "tests/data/utftest_utf16_le.xml", + "tests/data/utftest_utf16_le_bom.xml", + "tests/data/utftest_utf16_le_nodecl.xml", + "tests/data/utftest_utf32_le.xml", + "tests/data/utftest_utf32_le_bom.xml", + "tests/data/utftest_utf32_le_nodecl.xml", + }}; xml_encoding encodings[] = - { - encoding_utf16, encoding_utf16, encoding_utf16, - encoding_utf32, encoding_utf32, encoding_utf32 - }; + { + encoding_utf16, encoding_utf16, encoding_utf16, + encoding_utf32, encoding_utf32, encoding_utf32}; for (unsigned int i = 0; i < sizeof(files[0]) / sizeof(files[0][0]); ++i) { @@ -1084,24 +1098,22 @@ TEST(document_load_file_convert_native_endianness) struct file_data_t { - const char* path; - xml_encoding encoding; + const char* path; + xml_encoding encoding; - char* data; - size_t size; + char* data; + size_t size; }; - TEST(document_contents_preserve) { file_data_t files[] = - { - {"tests/data/utftest_utf16_be_clean.xml", encoding_utf16_be, 0, 0}, - {"tests/data/utftest_utf16_le_clean.xml", encoding_utf16_le, 0, 0}, - {"tests/data/utftest_utf32_be_clean.xml", encoding_utf32_be, 0, 0}, - {"tests/data/utftest_utf32_le_clean.xml", encoding_utf32_le, 0, 0}, - {"tests/data/utftest_utf8_clean.xml", encoding_utf8, 0, 0} - }; + { + {"tests/data/utftest_utf16_be_clean.xml", encoding_utf16_be, 0, 0}, + {"tests/data/utftest_utf16_le_clean.xml", encoding_utf16_le, 0, 0}, + {"tests/data/utftest_utf32_be_clean.xml", encoding_utf32_be, 0, 0}, + {"tests/data/utftest_utf32_le_clean.xml", encoding_utf32_le, 0, 0}, + {"tests/data/utftest_utf8_clean.xml", encoding_utf8, 0, 0}}; // load files in memory for (unsigned int i = 0; i < sizeof(files) / sizeof(files[0]); ++i) @@ -1133,10 +1145,9 @@ TEST(document_contents_preserve) TEST(document_contents_preserve_latin1) { file_data_t files[] = - { - {"tests/data/latintest_utf8.xml", encoding_utf8, 0, 0}, - {"tests/data/latintest_latin1.xml", encoding_latin1, 0, 0} - }; + { + {"tests/data/latintest_utf8.xml", encoding_utf8, 0, 0}, + {"tests/data/latintest_latin1.xml", encoding_latin1, 0, 0}}; // load files in memory for (unsigned int i = 0; i < sizeof(files) / sizeof(files[0]); ++i) @@ -1216,18 +1227,17 @@ TEST(document_convert_invalid_utf16) TEST(document_load_buffer_empty) { xml_encoding encodings[] = - { - encoding_auto, - encoding_utf8, - encoding_utf16_le, - encoding_utf16_be, - encoding_utf16, - encoding_utf32_le, - encoding_utf32_be, - encoding_utf32, - encoding_wchar, - encoding_latin1 - }; + { + encoding_auto, + encoding_utf8, + encoding_utf16_le, + encoding_utf16_be, + encoding_utf16, + encoding_utf32_le, + encoding_utf32_be, + encoding_utf32, + encoding_wchar, + encoding_latin1}; char buffer[1]; @@ -1252,18 +1262,17 @@ TEST(document_load_buffer_empty) TEST(document_load_buffer_empty_fragment) { xml_encoding encodings[] = - { - encoding_auto, - encoding_utf8, - encoding_utf16_le, - encoding_utf16_be, - encoding_utf16, - encoding_utf32_le, - encoding_utf32_be, - encoding_utf32, - encoding_wchar, - encoding_latin1 - }; + { + encoding_auto, + encoding_utf8, + encoding_utf16_le, + encoding_utf16_be, + encoding_utf16, + encoding_utf32_le, + encoding_utf32_be, + encoding_utf32, + encoding_wchar, + encoding_latin1}; char buffer[1]; @@ -1390,91 +1399,92 @@ TEST(document_load_buffer_inplace_short) #ifndef PUGIXML_NO_EXCEPTIONS TEST(document_load_exceptions) { - bool thrown = false; + bool thrown = false; - try - { - xml_document doc; - if (!doc.load_string(STR("", parse_default | parse_declaration | parse_comments) { - CHECK(doc.document_element() == doc.child(STR("node"))); + CHECK(doc.document_element() == doc.child(STR("node"))); } TEST_XML_FLAGS(document_element_absent, "", parse_comments | parse_fragment) { - CHECK(doc.document_element() == xml_node()); + CHECK(doc.document_element() == xml_node()); } TEST_XML(document_reset, "") { - CHECK(doc.first_child()); + CHECK(doc.first_child()); - doc.reset(); - CHECK(!doc.first_child()); - CHECK_NODE(doc, STR("")); + doc.reset(); + CHECK(!doc.first_child()); + CHECK_NODE(doc, STR("")); - doc.reset(); - CHECK(!doc.first_child()); - CHECK_NODE(doc, STR("")); + doc.reset(); + CHECK(!doc.first_child()); + CHECK_NODE(doc, STR("")); - CHECK(doc.load_string(STR(""))); - CHECK(doc.first_child()); - CHECK_NODE(doc, STR("")); + CHECK(doc.load_string(STR(""))); + CHECK(doc.first_child()); + CHECK_NODE(doc, STR("")); - doc.reset(); - CHECK(!doc.first_child()); - CHECK_NODE(doc, STR("")); + doc.reset(); + CHECK(!doc.first_child()); + CHECK_NODE(doc, STR("")); } TEST(document_reset_empty) { - xml_document doc; + xml_document doc; - doc.reset(); - CHECK(!doc.first_child()); - CHECK_NODE(doc, STR("")); + doc.reset(); + CHECK(!doc.first_child()); + CHECK_NODE(doc, STR("")); } TEST_XML(document_reset_copy, "") { - xml_document doc2; + xml_document doc2; - CHECK_NODE(doc2, STR("")); + CHECK_NODE(doc2, STR("")); - doc2.reset(doc); + doc2.reset(doc); - CHECK_NODE(doc2, STR("")); - CHECK(doc.first_child() != doc2.first_child()); + CHECK_NODE(doc2, STR("")); + CHECK(doc.first_child() != doc2.first_child()); - doc.reset(doc2); + doc.reset(doc2); - CHECK_NODE(doc, STR("")); - CHECK(doc.first_child() != doc2.first_child()); + CHECK_NODE(doc, STR("")); + CHECK(doc.first_child() != doc2.first_child()); - CHECK(doc.first_child().offset_debug() == -1); + CHECK(doc.first_child().offset_debug() == -1); } TEST_XML(document_reset_copy_self, "") { - CHECK_NODE(doc, STR("")); + CHECK_NODE(doc, STR("")); - doc.reset(doc); + doc.reset(doc); - CHECK(!doc.first_child()); - CHECK_NODE(doc, STR("")); + CHECK(!doc.first_child()); + CHECK_NODE(doc, STR("")); } TEST(document_load_buffer_utf_truncated) @@ -1487,20 +1497,20 @@ TEST(document_load_buffer_utf_truncated) struct document_data_t { - xml_encoding encoding; + xml_encoding encoding; - const unsigned char* data; - size_t size; + const unsigned char* data; + size_t size; }; const document_data_t data[] = - { - { encoding_utf8, utf8, sizeof(utf8) }, - { encoding_utf16_be, utf16_be, sizeof(utf16_be) }, - { encoding_utf16_le, utf16_le, sizeof(utf16_le) }, - { encoding_utf32_be, utf32_be, sizeof(utf32_be) }, - { encoding_utf32_le, utf32_le, sizeof(utf32_le) }, - }; + { + {encoding_utf8, utf8, sizeof(utf8)}, + {encoding_utf16_be, utf16_be, sizeof(utf16_be)}, + {encoding_utf16_le, utf16_le, sizeof(utf16_le)}, + {encoding_utf32_be, utf32_be, sizeof(utf32_be)}, + {encoding_utf32_le, utf32_le, sizeof(utf32_le)}, + }; for (size_t i = 0; i < sizeof(data) / sizeof(data[0]); ++i) { @@ -1520,11 +1530,11 @@ TEST(document_load_buffer_utf_truncated) const char_t* name = doc.first_child().name(); - #ifdef PUGIXML_WCHAR_MODE +#ifdef PUGIXML_WCHAR_MODE CHECK(name[0] == 0x20ac && name[1] == 0); - #else +#else CHECK_STRING(name, "\xe2\x82\xac"); - #endif +#endif } else { @@ -1565,11 +1575,11 @@ TEST(document_load_stream_truncated) { const char_t* name = doc.first_child().name(); - #ifdef PUGIXML_WCHAR_MODE +#ifdef PUGIXML_WCHAR_MODE CHECK(name[0] == 0x20ac && name[1] == 0); - #else +#else CHECK_STRING(name, "\xe2\x82\xac"); - #endif +#endif } } } @@ -1594,14 +1604,13 @@ TEST(document_alignment) TEST(document_convert_out_of_memory) { file_data_t files[] = - { - {"tests/data/utftest_utf16_be_clean.xml", encoding_utf16_be, 0, 0}, - {"tests/data/utftest_utf16_le_clean.xml", encoding_utf16_le, 0, 0}, - {"tests/data/utftest_utf32_be_clean.xml", encoding_utf32_be, 0, 0}, - {"tests/data/utftest_utf32_le_clean.xml", encoding_utf32_le, 0, 0}, - {"tests/data/utftest_utf8_clean.xml", encoding_utf8, 0, 0}, - {"tests/data/latintest_latin1.xml", encoding_latin1, 0, 0} - }; + { + {"tests/data/utftest_utf16_be_clean.xml", encoding_utf16_be, 0, 0}, + {"tests/data/utftest_utf16_le_clean.xml", encoding_utf16_le, 0, 0}, + {"tests/data/utftest_utf32_be_clean.xml", encoding_utf32_be, 0, 0}, + {"tests/data/utftest_utf32_le_clean.xml", encoding_utf32_le, 0, 0}, + {"tests/data/utftest_utf8_clean.xml", encoding_utf8, 0, 0}, + {"tests/data/latintest_latin1.xml", encoding_latin1, 0, 0}}; // load files in memory for (unsigned int i = 0; i < sizeof(files) / sizeof(files[0]); ++i) @@ -1761,7 +1770,7 @@ TEST(document_move_empty_zero_alloc) test_runner::_memory_fail_threshold = 1; for (int i = 1; i < 32; ++i) - docs[i] = std::move(docs[i-1]); + docs[i] = std::move(docs[i - 1]); delete[] docs; } @@ -1776,7 +1785,7 @@ TEST(document_move_repeated_zero_alloc) test_runner::_memory_fail_threshold = 1; for (int i = 1; i < 32; ++i) - docs[i] = std::move(docs[i-1]); + docs[i] = std::move(docs[i - 1]); for (int i = 0; i < 31; ++i) CHECK(!docs[i].first_child()); @@ -1797,15 +1806,15 @@ TEST(document_move_compact_fail) int safe_count = 21; for (int i = 1; i <= safe_count; ++i) - docs[i] = std::move(docs[i-1]); + docs[i] = std::move(docs[i - 1]); - CHECK_ALLOC_FAIL(docs[safe_count+1] = std::move(docs[safe_count])); + CHECK_ALLOC_FAIL(docs[safe_count + 1] = std::move(docs[safe_count])); for (int i = 0; i < safe_count; ++i) CHECK(!docs[i].first_child()); CHECK_NODE(docs[safe_count], STR("")); - CHECK(!docs[safe_count+1].first_child()); + CHECK(!docs[safe_count + 1].first_child()); } #endif #endif diff --git a/tests/test_dom_modify.cpp b/tests/test_dom_modify.cpp index 9696827..79672d8 100644 --- a/tests/test_dom_modify.cpp +++ b/tests/test_dom_modify.cpp @@ -3,9 +3,9 @@ #include #include +#include #include #include -#include using namespace pugi; @@ -174,7 +174,7 @@ TEST_XML(dom_attr_assign_large_number_float, "") node.attribute(STR("attr")) = std::numeric_limits::max(); CHECK(test_node(node, STR(""), STR(""), format_raw) || - test_node(node, STR(""), STR(""), format_raw)); + test_node(node, STR(""), STR(""), format_raw)); } TEST_XML(dom_attr_assign_large_number_double, "") @@ -880,12 +880,12 @@ TEST(dom_node_declaration_name) TEST(dom_node_declaration_attributes) { - xml_document doc; - xml_node node = doc.append_child(node_declaration); - node.append_attribute(STR("version")) = STR("1.0"); - node.append_attribute(STR("encoding")) = STR("utf-8"); + xml_document doc; + xml_node node = doc.append_child(node_declaration); + node.append_attribute(STR("version")) = STR("1.0"); + node.append_attribute(STR("encoding")) = STR("utf-8"); - CHECK_NODE(doc, STR("")); + CHECK_NODE(doc, STR("")); } TEST(dom_node_declaration_top_level) @@ -926,7 +926,8 @@ TEST(dom_string_out_of_memory) const unsigned int length = 65536; static char_t string[length + 1]; - for (unsigned int i = 0; i < length; ++i) string[i] = 'a'; + for (unsigned int i = 0; i < length; ++i) + string[i] = 'a'; string[length] = 0; xml_document doc; @@ -976,8 +977,8 @@ TEST(dom_node_out_of_memory) xml_attribute a = n.append_attribute(STR("a")); CHECK(a); - CHECK_ALLOC_FAIL(while (n.append_child(node_comment)) { /* nop */ }); - CHECK_ALLOC_FAIL(while (n.append_attribute(STR("b"))) { /* nop */ }); + CHECK_ALLOC_FAIL(while (n.append_child(node_comment)){/* nop */}); + CHECK_ALLOC_FAIL(while (n.append_attribute(STR("b"))){/* nop */}); // verify all node modification operations CHECK_ALLOC_FAIL(CHECK(!n.append_child())); @@ -1005,7 +1006,8 @@ TEST(dom_node_memory_limit) const unsigned int length = 65536; static char_t string[length + 1]; - for (unsigned int i = 0; i < length; ++i) string[i] = 'a'; + for (unsigned int i = 0; i < length; ++i) + string[i] = 'a'; string[length] = 0; test_runner::_memory_fail_threshold = 32768 * 2 + sizeof(string); @@ -1024,7 +1026,8 @@ TEST(dom_node_memory_limit_pi) const unsigned int length = 65536; static char_t string[length + 1]; - for (unsigned int i = 0; i < length; ++i) string[i] = 'a'; + for (unsigned int i = 0; i < length; ++i) + string[i] = 'a'; string[length] = 0; test_runner::_memory_fail_threshold = 32768 * 2 + sizeof(string); @@ -1073,15 +1076,15 @@ TEST(dom_node_doctype_copy) TEST(dom_node_doctype_value) { - xml_document doc; - xml_node node = doc.append_child(node_doctype); + xml_document doc; + xml_node node = doc.append_child(node_doctype); - CHECK(node.type() == node_doctype); - CHECK_STRING(node.value(), STR("")); - CHECK_NODE(node, STR("")); + CHECK(node.type() == node_doctype); + CHECK_STRING(node.value(), STR("")); + CHECK_NODE(node, STR("")); - CHECK(node.set_value(STR("id [ foo ]"))); - CHECK_NODE(node, STR("")); + CHECK(node.set_value(STR("id [ foo ]"))); + CHECK_NODE(node, STR("")); } TEST_XML(dom_node_append_buffer_native, "test") @@ -1112,7 +1115,6 @@ TEST_XML(dom_node_append_buffer_convert, "test") CHECK_NODE(doc, STR("test")); } - TEST_XML(dom_node_append_buffer_remove, "test") { xml_node node = doc.child(STR("node")); @@ -1127,15 +1129,21 @@ TEST_XML(dom_node_append_buffer_remove, "test") CHECK_NODE(doc, STR("testtexttext")); - while (node.remove_child(STR("child2"))) {} + while (node.remove_child(STR("child2"))) + { + } CHECK_NODE(doc, STR("test")); - while (node.remove_child(STR("child1"))) {} + while (node.remove_child(STR("child1"))) + { + } CHECK_NODE(doc, STR("test")); - while (node.remove_child(STR("child3"))) {} + while (node.remove_child(STR("child3"))) + { + } CHECK_NODE(doc, STR("test")); @@ -1665,7 +1673,8 @@ TEST(dom_node_copy_declaration_empty_name) CHECK_STRING(decl2.name(), STR("")); } -template bool fp_equal(T lhs, T rhs) +template +bool fp_equal(T lhs, T rhs) { // Several compilers compare float/double values on x87 stack without proper rounding // This causes roundtrip tests to fail, although they correctly preserve the data. @@ -1703,13 +1712,13 @@ TEST(dom_fp_roundtrip_min_max_double) } const double fp_roundtrip_base[] = -{ - 0.31830988618379067154, - 0.43429448190325182765, - 0.57721566490153286061, - 0.69314718055994530942, - 0.70710678118654752440, - 0.78539816339744830962, + { + 0.31830988618379067154, + 0.43429448190325182765, + 0.57721566490153286061, + 0.69314718055994530942, + 0.70710678118654752440, + 0.78539816339744830962, }; TEST(dom_fp_roundtrip_float) @@ -1738,17 +1747,17 @@ TEST(dom_fp_roundtrip_double) { for (size_t i = 0; i < sizeof(fp_roundtrip_base) / sizeof(fp_roundtrip_base[0]); ++i) { - #if (defined(_MSC_VER) && _MSC_VER < 1400) || defined(__MWERKS__) +#if (defined(_MSC_VER) && _MSC_VER < 1400) || defined(__MWERKS__) // Not all runtime libraries guarantee roundtripping for denormals if (e == -1021 && fp_roundtrip_base[i] < 0.5) continue; - #endif +#endif - #ifdef __DMC__ +#ifdef __DMC__ // Digital Mars C does not roundtrip on exactly one combination if (e == -12 && i == 1) continue; - #endif +#endif double value = ldexp(fp_roundtrip_base[i], e); diff --git a/tests/test_dom_text.cpp b/tests/test_dom_text.cpp index 23b41e5..48a09a7 100644 --- a/tests/test_dom_text.cpp +++ b/tests/test_dom_text.cpp @@ -8,57 +8,57 @@ using namespace pugi; TEST_XML_FLAGS(dom_text_empty, "foo", parse_default | parse_pi) { - xml_node node = doc.child(STR("node")); + xml_node node = doc.child(STR("node")); - CHECK(node.child(STR("a")).text()); - CHECK(node.child(STR("b")).text()); - CHECK(!node.child(STR("c")).text()); - CHECK(!node.child(STR("d")).text()); - CHECK(!xml_node().text()); - CHECK(!xml_text()); + CHECK(node.child(STR("a")).text()); + CHECK(node.child(STR("b")).text()); + CHECK(!node.child(STR("c")).text()); + CHECK(!node.child(STR("d")).text()); + CHECK(!xml_node().text()); + CHECK(!xml_text()); - generic_empty_test(node.child(STR("a")).text()); + generic_empty_test(node.child(STR("a")).text()); } TEST_XML(dom_text_bool_ops, "foo") { - generic_bool_ops_test(doc.child(STR("node")).text()); + generic_bool_ops_test(doc.child(STR("node")).text()); } TEST_XML_FLAGS(dom_text_get, "foo", parse_default | parse_pi) { - xml_node node = doc.child(STR("node")); + xml_node node = doc.child(STR("node")); - CHECK_STRING(node.child(STR("a")).text().get(), STR("foo")); - CHECK_STRING(node.child(STR("a")).first_child().text().get(), STR("foo")); + CHECK_STRING(node.child(STR("a")).text().get(), STR("foo")); + CHECK_STRING(node.child(STR("a")).first_child().text().get(), STR("foo")); - CHECK_STRING(node.child(STR("b")).text().get(), STR("bar")); - CHECK_STRING(node.child(STR("b")).last_child().text().get(), STR("bar")); + CHECK_STRING(node.child(STR("b")).text().get(), STR("bar")); + CHECK_STRING(node.child(STR("b")).last_child().text().get(), STR("bar")); - CHECK_STRING(node.child(STR("c")).text().get(), STR("")); - CHECK_STRING(node.child(STR("c")).first_child().text().get(), STR("")); + CHECK_STRING(node.child(STR("c")).text().get(), STR("")); + CHECK_STRING(node.child(STR("c")).first_child().text().get(), STR("")); - CHECK_STRING(node.child(STR("d")).text().get(), STR("")); + CHECK_STRING(node.child(STR("d")).text().get(), STR("")); - CHECK_STRING(xml_node().text().get(), STR("")); + CHECK_STRING(xml_node().text().get(), STR("")); } TEST_XML_FLAGS(dom_text_as_string, "foo", parse_default | parse_pi) { - xml_node node = doc.child(STR("node")); + xml_node node = doc.child(STR("node")); - CHECK_STRING(node.child(STR("a")).text().as_string(), STR("foo")); - CHECK_STRING(node.child(STR("a")).first_child().text().as_string(), STR("foo")); + CHECK_STRING(node.child(STR("a")).text().as_string(), STR("foo")); + CHECK_STRING(node.child(STR("a")).first_child().text().as_string(), STR("foo")); - CHECK_STRING(node.child(STR("b")).text().as_string(), STR("bar")); - CHECK_STRING(node.child(STR("b")).last_child().text().as_string(), STR("bar")); + CHECK_STRING(node.child(STR("b")).text().as_string(), STR("bar")); + CHECK_STRING(node.child(STR("b")).last_child().text().as_string(), STR("bar")); - CHECK_STRING(node.child(STR("c")).text().as_string(), STR("")); - CHECK_STRING(node.child(STR("c")).first_child().text().as_string(), STR("")); + CHECK_STRING(node.child(STR("c")).text().as_string(), STR("")); + CHECK_STRING(node.child(STR("c")).first_child().text().as_string(), STR("")); - CHECK_STRING(node.child(STR("d")).text().as_string(), STR("")); + CHECK_STRING(node.child(STR("d")).text().as_string(), STR("")); - CHECK_STRING(xml_node().text().as_string(), STR("")); + CHECK_STRING(xml_node().text().as_string(), STR("")); } TEST_XML(dom_text_as_int, "1-1-214748364821474836470") @@ -70,19 +70,19 @@ TEST_XML(dom_text_as_int, "1-1-214748 CHECK(node.child(STR("text2")).text().as_int() == -1); CHECK(node.child(STR("text3")).text().as_int() == -2147483647 - 1); CHECK(node.child(STR("text4")).text().as_int() == 2147483647); - CHECK(node.child(STR("text5")).text().as_int() == 0); + CHECK(node.child(STR("text5")).text().as_int() == 0); } TEST_XML(dom_text_as_int_hex, "07770x5ab0XFf-0x20-0x800000000x") { - xml_node node = doc.child(STR("node")); + xml_node node = doc.child(STR("node")); - CHECK(node.child(STR("text1")).text().as_int() == 777); // no octal support! intentional - CHECK(node.child(STR("text2")).text().as_int() == 1451); - CHECK(node.child(STR("text3")).text().as_int() == 255); - CHECK(node.child(STR("text4")).text().as_int() == -32); - CHECK(node.child(STR("text5")).text().as_int() == -2147483647 - 1); - CHECK(node.child(STR("text6")).text().as_int() == 0); + CHECK(node.child(STR("text1")).text().as_int() == 777); // no octal support! intentional + CHECK(node.child(STR("text2")).text().as_int() == 1451); + CHECK(node.child(STR("text3")).text().as_int() == 255); + CHECK(node.child(STR("text4")).text().as_int() == -32); + CHECK(node.child(STR("text5")).text().as_int() == -2147483647 - 1); + CHECK(node.child(STR("text6")).text().as_int() == 0); } TEST_XML(dom_text_as_uint, "01214748364742949672950") @@ -94,29 +94,29 @@ TEST_XML(dom_text_as_uint, "012147483 CHECK(node.child(STR("text2")).text().as_uint() == 1); CHECK(node.child(STR("text3")).text().as_uint() == 2147483647); CHECK(node.child(STR("text4")).text().as_uint() == 4294967295u); - CHECK(node.child(STR("text5")).text().as_uint() == 0); + CHECK(node.child(STR("text5")).text().as_uint() == 0); } TEST_XML(dom_text_as_uint_hex, "07770x5ab0XFf0x200xFFFFFFFF0x") { - xml_node node = doc.child(STR("node")); + xml_node node = doc.child(STR("node")); - CHECK(node.child(STR("text1")).text().as_uint() == 777); // no octal support! intentional - CHECK(node.child(STR("text2")).text().as_uint() == 1451); - CHECK(node.child(STR("text3")).text().as_uint() == 255); - CHECK(node.child(STR("text4")).text().as_uint() == 32); - CHECK(node.child(STR("text5")).text().as_uint() == 4294967295u); - CHECK(node.child(STR("text6")).text().as_uint() == 0); + CHECK(node.child(STR("text1")).text().as_uint() == 777); // no octal support! intentional + CHECK(node.child(STR("text2")).text().as_uint() == 1451); + CHECK(node.child(STR("text3")).text().as_uint() == 255); + CHECK(node.child(STR("text4")).text().as_uint() == 32); + CHECK(node.child(STR("text5")).text().as_uint() == 4294967295u); + CHECK(node.child(STR("text6")).text().as_uint() == 0); } TEST_XML(dom_text_as_integer_space, " \t\n1234\n\t 0x123- 16- 0x10") { - xml_node node = doc.child(STR("node")); + xml_node node = doc.child(STR("node")); - CHECK(node.child(STR("text1")).text().as_int() == 1234); - CHECK(node.child(STR("text2")).text().as_int() == 291); - CHECK(node.child(STR("text3")).text().as_int() == 0); - CHECK(node.child(STR("text4")).text().as_int() == 0); + CHECK(node.child(STR("text1")).text().as_int() == 1234); + CHECK(node.child(STR("text2")).text().as_int() == 291); + CHECK(node.child(STR("text3")).text().as_int() == 0); + CHECK(node.child(STR("text4")).text().as_int() == 0); } TEST_XML(dom_text_as_float, "010.12-5.13e-43.14159265358979323846") @@ -211,42 +211,42 @@ TEST_XML(dom_text_as_ullong_hex, "07770x5ab< TEST_XML(dom_text_get_no_state, "") { - xml_node node = doc.child(STR("node")); - xml_text t = node.text(); + xml_node node = doc.child(STR("node")); + xml_text t = node.text(); - CHECK(!t); - CHECK(t.get() && *t.get() == 0); - CHECK(!node.first_child()); + CHECK(!t); + CHECK(t.get() && *t.get() == 0); + CHECK(!node.first_child()); - node.append_child(node_pcdata); + node.append_child(node_pcdata); - CHECK(t); - CHECK_STRING(t.get(), STR("")); + CHECK(t); + CHECK_STRING(t.get(), STR("")); - node.first_child().set_value(STR("test")); + node.first_child().set_value(STR("test")); - CHECK(t); - CHECK_STRING(t.get(), STR("test")); + CHECK(t); + CHECK_STRING(t.get(), STR("test")); } TEST_XML(dom_text_set, "") { - xml_node node = doc.child(STR("node")); - xml_text t = node.text(); + xml_node node = doc.child(STR("node")); + xml_text t = node.text(); - t.set(STR("")); - CHECK(node.first_child().type() == node_pcdata); - CHECK_NODE(node, STR("")); + t.set(STR("")); + CHECK(node.first_child().type() == node_pcdata); + CHECK_NODE(node, STR("")); - t.set(STR("boo")); - CHECK(node.first_child().type() == node_pcdata); - CHECK(node.first_child() == node.last_child()); - CHECK_NODE(node, STR("boo")); + t.set(STR("boo")); + CHECK(node.first_child().type() == node_pcdata); + CHECK(node.first_child() == node.last_child()); + CHECK_NODE(node, STR("boo")); - t.set(STR("foobarfoobar")); - CHECK(node.first_child().type() == node_pcdata); - CHECK(node.first_child() == node.last_child()); - CHECK_NODE(node, STR("foobarfoobar")); + t.set(STR("foobarfoobar")); + CHECK(node.first_child().type() == node_pcdata); + CHECK(node.first_child() == node.last_child()); + CHECK_NODE(node, STR("foobarfoobar")); } TEST_XML(dom_text_assign, "") @@ -399,49 +399,49 @@ TEST_XML(dom_text_set_value_llong, "") TEST_XML(dom_text_middle, "notthisonetext") { - xml_node node = doc.child(STR("node")); - xml_text t = node.text(); + xml_node node = doc.child(STR("node")); + xml_text t = node.text(); - CHECK_STRING(t.get(), STR("text")); - t.set(STR("notext")); + CHECK_STRING(t.get(), STR("text")); + t.set(STR("notext")); - CHECK_NODE(node, STR("notthisonenotext")); - CHECK(node.remove_child(t.data())); + CHECK_NODE(node, STR("notthisonenotext")); + CHECK(node.remove_child(t.data())); - CHECK(!t); - CHECK_NODE(node, STR("notthisone")); + CHECK(!t); + CHECK_NODE(node, STR("notthisone")); - t.set(STR("yestext")); + t.set(STR("yestext")); - CHECK(t); - CHECK_NODE(node, STR("notthisoneyestext")); - CHECK(t.data() == node.last_child()); + CHECK(t); + CHECK_NODE(node, STR("notthisoneyestext")); + CHECK(t.data() == node.last_child()); } TEST_XML_FLAGS(dom_text_data, "foo", parse_default | parse_pi) { - xml_node node = doc.child(STR("node")); + xml_node node = doc.child(STR("node")); - CHECK(node.child(STR("a")).text().data() == node.child(STR("a")).first_child()); - CHECK(node.child(STR("b")).text().data() == node.child(STR("b")).first_child()); - CHECK(!node.child(STR("c")).text().data()); - CHECK(!node.child(STR("d")).text().data()); - CHECK(!xml_text().data()); + CHECK(node.child(STR("a")).text().data() == node.child(STR("a")).first_child()); + CHECK(node.child(STR("b")).text().data() == node.child(STR("b")).first_child()); + CHECK(!node.child(STR("c")).text().data()); + CHECK(!node.child(STR("d")).text().data()); + CHECK(!xml_text().data()); } TEST(dom_text_defaults) { - xml_text text; + xml_text text; - CHECK_STRING(text.as_string(STR("foo")), STR("foo")); - CHECK(text.as_int(42) == 42); - CHECK(text.as_uint(42) == 42); - CHECK(text.as_double(42) == 42); - CHECK(text.as_float(42) == 42); - CHECK(text.as_bool(true) == true); + CHECK_STRING(text.as_string(STR("foo")), STR("foo")); + CHECK(text.as_int(42) == 42); + CHECK(text.as_uint(42) == 42); + CHECK(text.as_double(42) == 42); + CHECK(text.as_float(42) == 42); + CHECK(text.as_bool(true) == true); #ifdef PUGIXML_HAS_LONG_LONG - CHECK(text.as_llong(42) == 42); - CHECK(text.as_ullong(42) == 42); + CHECK(text.as_llong(42) == 42); + CHECK(text.as_ullong(42) == 42); #endif } diff --git a/tests/test_dom_traverse.cpp b/tests/test_dom_traverse.cpp index b135674..52ec8eb 100644 --- a/tests/test_dom_traverse.cpp +++ b/tests/test_dom_traverse.cpp @@ -4,28 +4,34 @@ #include "test.hpp" -#include #include +#include #include -#include -#include #include #include +#include +#include #include "helpers.hpp" using namespace pugi; #ifdef PUGIXML_NO_STL -template static I move_iter(I base, int n) +template +static I move_iter(I base, int n) { - if (n > 0) while (n--) ++base; - else while (n++) --base; + if (n > 0) + while (n--) + ++base; + else + while (n++) + --base; return base; } #else -template static I move_iter(I base, int n) +template +static I move_iter(I base, int n) { std::advance(base, n); return base; @@ -237,18 +243,18 @@ TEST_XML(dom_attr_as_ullong_hex, "") { - xml_node node = doc.child(STR("node")); + xml_node node = doc.child(STR("node")); - const xml_attribute_iterator i1 = node.attributes_begin(); - const xml_attribute_iterator i2 = ++xml_attribute_iterator(i1); - const xml_attribute_iterator i3 = ++xml_attribute_iterator(i2); + const xml_attribute_iterator i1 = node.attributes_begin(); + const xml_attribute_iterator i2 = ++xml_attribute_iterator(i1); + const xml_attribute_iterator i3 = ++xml_attribute_iterator(i2); - CHECK(*i1 == node.attribute(STR("attr1"))); - CHECK(*i2 == node.attribute(STR("attr2"))); - CHECK(i3 == node.attributes_end()); + CHECK(*i1 == node.attribute(STR("attr1"))); + CHECK(*i2 == node.attribute(STR("attr2"))); + CHECK(i3 == node.attributes_end()); - CHECK_STRING(i1->name(), STR("attr1")); - CHECK_STRING(i2->name(), STR("attr2")); + CHECK_STRING(i1->name(), STR("attr1")); + CHECK_STRING(i2->name(), STR("attr2")); } TEST_XML(dom_node_bool_ops, "") @@ -461,18 +467,18 @@ TEST_XML(dom_node_iterator_invalidate, "") { - xml_node node = doc.child(STR("node")); + xml_node node = doc.child(STR("node")); - const xml_node_iterator i1 = node.begin(); - const xml_node_iterator i2 = ++xml_node_iterator(i1); - const xml_node_iterator i3 = ++xml_node_iterator(i2); + const xml_node_iterator i1 = node.begin(); + const xml_node_iterator i2 = ++xml_node_iterator(i1); + const xml_node_iterator i3 = ++xml_node_iterator(i2); - CHECK(*i1 == node.child(STR("child1"))); - CHECK(*i2 == node.child(STR("child2"))); - CHECK(i3 == node.end()); + CHECK(*i1 == node.child(STR("child1"))); + CHECK(*i2 == node.child(STR("child2"))); + CHECK(i3 == node.end()); - CHECK_STRING(i1->name(), STR("child1")); - CHECK_STRING(i2->name(), STR("child2")); + CHECK_STRING(i1->name(), STR("child1")); + CHECK_STRING(i2->name(), STR("child2")); } TEST_XML(dom_node_parent, "") @@ -659,11 +665,13 @@ struct find_predicate_const { bool result; - find_predicate_const(bool result_): result(result_) + find_predicate_const(bool result_) + : result(result_) { } - template bool operator()(const T&) const + template + bool operator()(const T&) const { return result; } @@ -673,18 +681,20 @@ struct find_predicate_prefix { const char_t* prefix; - find_predicate_prefix(const char_t* prefix_): prefix(prefix_) + find_predicate_prefix(const char_t* prefix_) + : prefix(prefix_) { } - template bool operator()(const T& obj) const + template + bool operator()(const T& obj) const { - #ifdef PUGIXML_WCHAR_MODE +#ifdef PUGIXML_WCHAR_MODE // can't use wcsncmp here because of a bug in DMC return std::basic_string(obj.name()).compare(0, wcslen(prefix), prefix) == 0; - #else +#else return strncmp(obj.name(), prefix, strlen(prefix)) == 0; - #endif +#endif } }; @@ -782,13 +792,15 @@ TEST_XML(dom_node_first_element_by_path, "text log; unsigned int call_count; unsigned int stop_count; - test_walker(unsigned int stop_count_ = 0): call_count(0), stop_count(stop_count_) + test_walker(unsigned int stop_count_ = 0) + : call_count(0) + , stop_count(stop_count_) { } @@ -797,14 +809,14 @@ struct test_walker: xml_tree_walker char buf[32]; sprintf(buf, "%d", depth()); - #ifdef PUGIXML_WCHAR_MODE +#ifdef PUGIXML_WCHAR_MODE wchar_t wbuf[32]; std::copy(buf, buf + strlen(buf) + 1, &wbuf[0]); return std::basic_string(wbuf); - #else +#else return std::basic_string(buf); - #endif +#endif } virtual bool begin(xml_node& node) PUGIXML_OVERRIDE @@ -935,7 +947,7 @@ TEST_XML_FLAGS(dom_offset_debug, "pcd TEST(dom_offset_debug_encoding) { - char buf[] = { 0, '<', 0, 'n', 0, '/', 0, '>' }; + char buf[] = {0, '<', 0, 'n', 0, '/', 0, '>'}; xml_document doc; CHECK(doc.load_buffer(buf, sizeof(buf))); @@ -979,17 +991,17 @@ TEST_XML(dom_internal_object, "value") CHECK(xml_node().internal_object() == 0); CHECK(xml_attribute().internal_object() == 0); - CHECK(node.internal_object() != 0); - CHECK(value.internal_object() != 0); - CHECK(node.internal_object() != value.internal_object()); + CHECK(node.internal_object() != 0); + CHECK(value.internal_object() != 0); + CHECK(node.internal_object() != value.internal_object()); - CHECK(attr.internal_object() != 0); + CHECK(attr.internal_object() != 0); - xml_node node_copy = node; - CHECK(node_copy.internal_object() == node.internal_object()); + xml_node node_copy = node; + CHECK(node_copy.internal_object() == node.internal_object()); - xml_attribute attr_copy = attr; - CHECK(attr_copy.internal_object() == attr.internal_object()); + xml_attribute attr_copy = attr; + CHECK(attr_copy.internal_object() == attr.internal_object()); } TEST_XML(dom_hash_value, "value") @@ -1001,17 +1013,17 @@ TEST_XML(dom_hash_value, "value") CHECK(xml_node().hash_value() == 0); CHECK(xml_attribute().hash_value() == 0); - CHECK(node.hash_value() != 0); - CHECK(value.hash_value() != 0); - CHECK(node.hash_value() != value.hash_value()); + CHECK(node.hash_value() != 0); + CHECK(value.hash_value() != 0); + CHECK(node.hash_value() != value.hash_value()); - CHECK(attr.hash_value() != 0); + CHECK(attr.hash_value() != 0); - xml_node node_copy = node; - CHECK(node_copy.hash_value() == node.hash_value()); + xml_node node_copy = node; + CHECK(node_copy.hash_value() == node.hash_value()); - xml_attribute attr_copy = attr; - CHECK(attr_copy.hash_value() == attr.hash_value()); + xml_attribute attr_copy = attr; + CHECK(attr_copy.hash_value() == attr.hash_value()); } TEST_XML(dom_node_named_iterator, "") @@ -1121,15 +1133,15 @@ TEST_XML(dom_ranged_for, "35 { int index = 1; - for (xml_node n: doc.children()) + for (xml_node n : doc.children()) { - for (xml_attribute a: n.attributes()) + for (xml_attribute a : n.attributes()) { CHECK(a.as_int() == index); index++; } - for (xml_node c: n.children(STR("test"))) + for (xml_node c : n.children(STR("test"))) { CHECK(c.text().as_int() == index); index++; diff --git a/tests/test_header_guard.cpp b/tests/test_header_guard.cpp index 2f65928..1ca6592 100644 --- a/tests/test_header_guard.cpp +++ b/tests/test_header_guard.cpp @@ -1,3 +1,2 @@ // Tests header guards #include "../src/pugixml.hpp" -#include "../src/pugixml.hpp" diff --git a/tests/test_header_iosfwd_2.cpp b/tests/test_header_iosfwd_2.cpp index 865d0d8..0ed528a 100644 --- a/tests/test_header_iosfwd_2.cpp +++ b/tests/test_header_iosfwd_2.cpp @@ -1,3 +1,3 @@ // Tests compatibility with iosfwd -#include #include "../src/pugixml.hpp" +#include diff --git a/tests/test_header_iostream_2.cpp b/tests/test_header_iostream_2.cpp index c0be50b..a836d4f 100644 --- a/tests/test_header_iostream_2.cpp +++ b/tests/test_header_iostream_2.cpp @@ -1,3 +1,3 @@ // Tests compatibility with iostream -#include #include "../src/pugixml.hpp" +#include diff --git a/tests/test_header_only_1.cpp b/tests/test_header_only_1.cpp index 2e90b33..c708e31 100644 --- a/tests/test_header_only_1.cpp +++ b/tests/test_header_only_1.cpp @@ -5,7 +5,6 @@ // Check header guards #include "../src/pugixml.hpp" -#include "../src/pugixml.hpp" using namespace pugi; diff --git a/tests/test_header_only_2.cpp b/tests/test_header_only_2.cpp index e4e876d..8a52d8c 100644 --- a/tests/test_header_only_2.cpp +++ b/tests/test_header_only_2.cpp @@ -5,7 +5,6 @@ // Check header guards #include "../src/pugixml.hpp" -#include "../src/pugixml.hpp" using namespace pugi; diff --git a/tests/test_header_string_2.cpp b/tests/test_header_string_2.cpp index 2813fc9..07d1263 100644 --- a/tests/test_header_string_2.cpp +++ b/tests/test_header_string_2.cpp @@ -1,3 +1,3 @@ // Tests compatibility with string -#include #include "../src/pugixml.hpp" +#include diff --git a/tests/test_header_string_iostream.cpp b/tests/test_header_string_iostream.cpp index 3210080..5160ccd 100644 --- a/tests/test_header_string_iostream.cpp +++ b/tests/test_header_string_iostream.cpp @@ -1,5 +1,5 @@ // Tests compatibility with string/iostream -#include #include "../src/pugixml.hpp" #include #include +#include diff --git a/tests/test_memory.cpp b/tests/test_memory.cpp index 3258736..85ae4df 100644 --- a/tests/test_memory.cpp +++ b/tests/test_memory.cpp @@ -1,7 +1,7 @@ #include "test.hpp" -#include "writer_string.hpp" #include "allocator.hpp" +#include "writer_string.hpp" #include #include @@ -10,28 +10,28 @@ using namespace pugi; namespace { - int page_allocs = 0; - int page_deallocs = 0; +int page_allocs = 0; +int page_deallocs = 0; - bool is_page(size_t size) - { - return size >= 16384; - } - - void* allocate(size_t size) - { - void* ptr = memory_allocate(size); - page_allocs += is_page(memory_size(ptr)); - return ptr; - } - - void deallocate(void* ptr) - { - page_deallocs += is_page(memory_size(ptr)); - memory_deallocate(ptr); - } +bool is_page(size_t size) +{ + return size >= 16384; } +void* allocate(size_t size) +{ + void* ptr = memory_allocate(size); + page_allocs += is_page(memory_size(ptr)); + return ptr; +} + +void deallocate(void* ptr) +{ + page_deallocs += is_page(memory_size(ptr)); + memory_deallocate(ptr); +} +} // namespace + TEST(memory_custom_memory_management) { page_allocs = page_deallocs = 0; @@ -116,7 +116,7 @@ TEST(memory_large_allocations) } // prune - for (node = doc.first_child(); node; ) + for (node = doc.first_child(); node;) { xml_node next = node.next_sibling().next_sibling(); @@ -229,7 +229,8 @@ TEST(memory_string_allocate_decreasing) std::basic_string s = STR("ab"); - for (int i = 0; i < 17; ++i) s += s; + for (int i = 0; i < 17; ++i) + s += s; for (int j = 0; j < 17; ++j) { @@ -286,7 +287,8 @@ TEST(memory_string_allocate_decreasing_inplace) std::basic_string s = STR("ab"); - for (int i = 0; i < 17; ++i) s += s; + for (int i = 0; i < 17; ++i) + s += s; for (int j = 0; j < 17; ++j) { diff --git a/tests/test_parse.cpp b/tests/test_parse.cpp index 180c70a..639b628 100644 --- a/tests/test_parse.cpp +++ b/tests/test_parse.cpp @@ -279,132 +279,132 @@ TEST(parse_ws_pcdata_parse) static int get_tree_node_count(xml_node n) { - int result = 1; + int result = 1; - for (xml_node c = n.first_child(); c; c = c.next_sibling()) - result += get_tree_node_count(c); + for (xml_node c = n.first_child(); c; c = c.next_sibling()) + result += get_tree_node_count(c); - return result; + return result; } TEST(parse_ws_pcdata_permutations) { - struct test_data_t - { - unsigned int mask; // 1 = default flags, 2 = parse_ws_pcdata, 4 = parse_ws_pcdata_single - const char_t* source; - const char_t* result; - int nodes; // negative if parsing should fail - }; + struct test_data_t + { + unsigned int mask; // 1 = default flags, 2 = parse_ws_pcdata, 4 = parse_ws_pcdata_single + const char_t* source; + const char_t* result; + int nodes; // negative if parsing should fail + }; - test_data_t test_data[] = - { - // external pcdata should be discarded (whitespace or not) - {7, STR("ext1"), STR(""), 2}, - {7, STR("ext1ext2"), STR(""), 2}, - {7, STR(" "), STR(""), 2}, - {7, STR(" "), STR(""), 2}, - {7, STR(" "), STR(""), 2}, - // inner pcdata should be preserved - {7, STR("inner"), STR("inner"), 3}, - {7, STR("inner1inner2"), STR("inner1inner2"), 5}, - {7, STR("inner1deepinner2"), STR("inner1deepinner2"), 6}, - // empty pcdata nodes should never be created - {7, STR("inner1inner2"), STR("inner1inner2"), 5}, - {7, STR("inner2"), STR("inner2"), 4}, - {7, STR("inner1"), STR("inner1"), 4}, - {7, STR(""), STR(""), 3}, - // comments, pi or other nodes should not cause pcdata creation either - {7, STR(""), STR(""), 4}, - // leading/trailing pcdata whitespace should be preserved (note: this will change if parse_ws_pcdata_trim is introduced) - {7, STR("\t \tinner1 deep \t\ninner2\n\t"), STR("\t \tinner1 deep \t\ninner2\n\t"), 6}, - // whitespace-only pcdata preservation depends on the parsing mode - {1, STR("\n\t \n\t \n\t\n\t"), STR(""), 5}, - {2, STR("\n\t \n\t \n\t\n\t"), STR("\n\t \n\t \n\t\n\t"), 13}, - {4, STR("\n\t \n\t \n\t\n\t"), STR(" "), 7}, - // current implementation of parse_ws_pcdata_single has an unfortunate bug; reproduce it here - {4, STR("\t\t\n\n"), STR("\n\n"), 3}, - // error case: terminate PCDATA in the middle - {7, STR("abcdef"), STR("abcdef"), -3}, - {5, STR(" "), STR(""), -2}, - {2, STR(" "), STR(" "), -3}, - // error case: terminate PCDATA as early as possible - {7, STR(""), STR(""), -2}, - {7, STR("a"), STR("a"), -3}, - {5, STR(" "), STR(""), -2}, - {2, STR(" "), STR(" "), -3}, - }; + test_data_t test_data[] = + { + // external pcdata should be discarded (whitespace or not) + {7, STR("ext1"), STR(""), 2}, + {7, STR("ext1ext2"), STR(""), 2}, + {7, STR(" "), STR(""), 2}, + {7, STR(" "), STR(""), 2}, + {7, STR(" "), STR(""), 2}, + // inner pcdata should be preserved + {7, STR("inner"), STR("inner"), 3}, + {7, STR("inner1inner2"), STR("inner1inner2"), 5}, + {7, STR("inner1deepinner2"), STR("inner1deepinner2"), 6}, + // empty pcdata nodes should never be created + {7, STR("inner1inner2"), STR("inner1inner2"), 5}, + {7, STR("inner2"), STR("inner2"), 4}, + {7, STR("inner1"), STR("inner1"), 4}, + {7, STR(""), STR(""), 3}, + // comments, pi or other nodes should not cause pcdata creation either + {7, STR(""), STR(""), 4}, + // leading/trailing pcdata whitespace should be preserved (note: this will change if parse_ws_pcdata_trim is introduced) + {7, STR("\t \tinner1 deep \t\ninner2\n\t"), STR("\t \tinner1 deep \t\ninner2\n\t"), 6}, + // whitespace-only pcdata preservation depends on the parsing mode + {1, STR("\n\t \n\t \n\t\n\t"), STR(""), 5}, + {2, STR("\n\t \n\t \n\t\n\t"), STR("\n\t \n\t \n\t\n\t"), 13}, + {4, STR("\n\t \n\t \n\t\n\t"), STR(" "), 7}, + // current implementation of parse_ws_pcdata_single has an unfortunate bug; reproduce it here + {4, STR("\t\t\n\n"), STR("\n\n"), 3}, + // error case: terminate PCDATA in the middle + {7, STR("abcdef"), STR("abcdef"), -3}, + {5, STR(" "), STR(""), -2}, + {2, STR(" "), STR(" "), -3}, + // error case: terminate PCDATA as early as possible + {7, STR(""), STR(""), -2}, + {7, STR("a"), STR("a"), -3}, + {5, STR(" "), STR(""), -2}, + {2, STR(" "), STR(" "), -3}, + }; - for (size_t i = 0; i < sizeof(test_data) / sizeof(test_data[0]); ++i) - { - const test_data_t& td = test_data[i]; + for (size_t i = 0; i < sizeof(test_data) / sizeof(test_data[0]); ++i) + { + const test_data_t& td = test_data[i]; - for (int flag = 0; flag < 3; ++flag) - { - if (td.mask & (1 << flag)) - { - unsigned int flags[] = {parse_default, parse_default | parse_ws_pcdata, parse_default | parse_ws_pcdata_single}; + for (int flag = 0; flag < 3; ++flag) + { + if (td.mask & (1 << flag)) + { + unsigned int flags[] = {parse_default, parse_default | parse_ws_pcdata, parse_default | parse_ws_pcdata_single}; - xml_document doc; - CHECK((td.nodes > 0) == doc.load_string(td.source, flags[flag])); - CHECK_NODE(doc, td.result); + xml_document doc; + CHECK((td.nodes > 0) == doc.load_string(td.source, flags[flag])); + CHECK_NODE(doc, td.result); - int nodes = get_tree_node_count(doc); - CHECK((td.nodes < 0 ? -td.nodes : td.nodes) == nodes); - } - } - } + int nodes = get_tree_node_count(doc); + CHECK((td.nodes < 0 ? -td.nodes : td.nodes) == nodes); + } + } + } } TEST(parse_ws_pcdata_fragment_permutations) { - struct test_data_t - { - unsigned int mask; // 1 = default flags, 2 = parse_ws_pcdata, 4 = parse_ws_pcdata_single - const char_t* source; - const char_t* result; - int nodes; // negative if parsing should fail - }; + struct test_data_t + { + unsigned int mask; // 1 = default flags, 2 = parse_ws_pcdata, 4 = parse_ws_pcdata_single + const char_t* source; + const char_t* result; + int nodes; // negative if parsing should fail + }; - test_data_t test_data[] = - { - // external pcdata should be preserved - {7, STR("ext1"), STR("ext1"), 2}, - {5, STR(" "), STR(""), 1}, - {2, STR(" "), STR(" "), 2}, - {7, STR("ext1"), STR("ext1"), 3}, - {7, STR("ext2"), STR("ext2"), 3}, - {7, STR("ext1ext2"), STR("ext1ext2"), 4}, - {7, STR("ext1ext2ext3"), STR("ext1ext2ext3"), 6}, - {5, STR(" "), STR(""), 2}, - {2, STR(" "), STR(" "), 3}, - {5, STR(" "), STR(""), 2}, - {2, STR(" "), STR(" "), 3}, - {5, STR(" "), STR(""), 2}, - {2, STR(" "), STR(" "), 4}, - {5, STR(" "), STR(""), 3}, - {2, STR(" "), STR(" "), 6}, - }; + test_data_t test_data[] = + { + // external pcdata should be preserved + {7, STR("ext1"), STR("ext1"), 2}, + {5, STR(" "), STR(""), 1}, + {2, STR(" "), STR(" "), 2}, + {7, STR("ext1"), STR("ext1"), 3}, + {7, STR("ext2"), STR("ext2"), 3}, + {7, STR("ext1ext2"), STR("ext1ext2"), 4}, + {7, STR("ext1ext2ext3"), STR("ext1ext2ext3"), 6}, + {5, STR(" "), STR(""), 2}, + {2, STR(" "), STR(" "), 3}, + {5, STR(" "), STR(""), 2}, + {2, STR(" "), STR(" "), 3}, + {5, STR(" "), STR(""), 2}, + {2, STR(" "), STR(" "), 4}, + {5, STR(" "), STR(""), 3}, + {2, STR(" "), STR(" "), 6}, + }; - for (size_t i = 0; i < sizeof(test_data) / sizeof(test_data[0]); ++i) - { - const test_data_t& td = test_data[i]; + for (size_t i = 0; i < sizeof(test_data) / sizeof(test_data[0]); ++i) + { + const test_data_t& td = test_data[i]; - for (int flag = 0; flag < 3; ++flag) - { - if (td.mask & (1 << flag)) - { - unsigned int flags[] = {parse_default, parse_default | parse_ws_pcdata, parse_default | parse_ws_pcdata_single}; + for (int flag = 0; flag < 3; ++flag) + { + if (td.mask & (1 << flag)) + { + unsigned int flags[] = {parse_default, parse_default | parse_ws_pcdata, parse_default | parse_ws_pcdata_single}; - xml_document doc; - CHECK((td.nodes > 0) == doc.load_string(td.source, flags[flag] | parse_fragment)); - CHECK_NODE(doc, td.result); + xml_document doc; + CHECK((td.nodes > 0) == doc.load_string(td.source, flags[flag] | parse_fragment)); + CHECK_NODE(doc, td.result); - int nodes = get_tree_node_count(doc); - CHECK((td.nodes < 0 ? -td.nodes : td.nodes) == nodes); - } - } - } + int nodes = get_tree_node_count(doc); + CHECK((td.nodes < 0 ? -td.nodes : td.nodes) == nodes); + } + } + } } TEST(parse_pcdata_no_eol) @@ -439,49 +439,48 @@ TEST(parse_pcdata_error) TEST(parse_pcdata_trim) { - struct test_data_t - { - const char_t* source; - const char_t* result; - unsigned int flags; - }; + struct test_data_t + { + const char_t* source; + const char_t* result; + unsigned int flags; + }; - test_data_t test_data[] = - { - { STR(" text"), STR("text"), 0 }, - { STR("\t\n text"), STR("text"), 0 }, - { STR("text "), STR("text"), 0 }, - { STR("text \t\n"), STR("text"), 0 }, - { STR("\r\n\t text \t\n\r"), STR("text"), 0 }, - { STR(" text"), STR("text"), parse_fragment }, - { STR("\t\n text"), STR("text"), parse_fragment }, - { STR("text "), STR("text"), parse_fragment }, - { STR("text \t\n"), STR("text"), parse_fragment }, - { STR("\r\n\t text \t\n\r"), STR("text"), parse_fragment }, - { STR("\r\n\t text \t\n\r more \r\n\t"), STR("text \t\n\r more"), 0 }, - { STR("\r\n\t text \t\n\r more \r\n\t"), STR("text \t\n\n more"), parse_eol }, - { STR("\r\n\t text \r\n\r\n\r\n\r\n\r\n\r\n\r\n more \r\n\t"), STR("text \n\n\n\n\n\n\n more"), parse_eol }, - { STR(" test&&&&&&& "), STR("test&&&&&&&"), 0 }, - { STR(" test&&&&&&& "), STR("test&&&&&&&"), parse_escapes }, - { STR(" test&&&&&&& "), STR("test&&&&&&&"), parse_fragment | parse_escapes }, - { STR("\r\n\t text \t\n\r m&&e \r\n\t"), STR("text \t\n\n m&&e"), parse_eol | parse_escapes } - }; + test_data_t test_data[] = + { + {STR(" text"), STR("text"), 0}, + {STR("\t\n text"), STR("text"), 0}, + {STR("text "), STR("text"), 0}, + {STR("text \t\n"), STR("text"), 0}, + {STR("\r\n\t text \t\n\r"), STR("text"), 0}, + {STR(" text"), STR("text"), parse_fragment}, + {STR("\t\n text"), STR("text"), parse_fragment}, + {STR("text "), STR("text"), parse_fragment}, + {STR("text \t\n"), STR("text"), parse_fragment}, + {STR("\r\n\t text \t\n\r"), STR("text"), parse_fragment}, + {STR("\r\n\t text \t\n\r more \r\n\t"), STR("text \t\n\r more"), 0}, + {STR("\r\n\t text \t\n\r more \r\n\t"), STR("text \t\n\n more"), parse_eol}, + {STR("\r\n\t text \r\n\r\n\r\n\r\n\r\n\r\n\r\n more \r\n\t"), STR("text \n\n\n\n\n\n\n more"), parse_eol}, + {STR(" test&&&&&&& "), STR("test&&&&&&&"), 0}, + {STR(" test&&&&&&& "), STR("test&&&&&&&"), parse_escapes}, + {STR(" test&&&&&&& "), STR("test&&&&&&&"), parse_fragment | parse_escapes}, + {STR("\r\n\t text \t\n\r m&&e \r\n\t"), STR("text \t\n\n m&&e"), parse_eol | parse_escapes}}; - for (size_t i = 0; i < sizeof(test_data) / sizeof(test_data[0]); ++i) - { - const test_data_t& td = test_data[i]; + for (size_t i = 0; i < sizeof(test_data) / sizeof(test_data[0]); ++i) + { + const test_data_t& td = test_data[i]; - xml_document doc; - CHECK(doc.load_string(td.source, td.flags | parse_trim_pcdata)); + xml_document doc; + CHECK(doc.load_string(td.source, td.flags | parse_trim_pcdata)); - const char_t* value = doc.child(STR("node")) ? doc.child_value(STR("node")) : doc.text().get(); - CHECK_STRING(value, td.result); - } + const char_t* value = doc.child(STR("node")) ? doc.child_value(STR("node")) : doc.text().get(); + CHECK_STRING(value, td.result); + } } TEST(parse_pcdata_trim_empty) { - unsigned int flags[] = { 0, parse_ws_pcdata, parse_ws_pcdata_single, parse_ws_pcdata | parse_ws_pcdata_single }; + unsigned int flags[] = {0, parse_ws_pcdata, parse_ws_pcdata_single, parse_ws_pcdata | parse_ws_pcdata_single}; for (size_t i = 0; i < sizeof(flags) / sizeof(flags[0]); ++i) { @@ -696,7 +695,6 @@ TEST(parse_attribute_variations) } } - TEST(parse_attribute_error) { xml_document doc; @@ -926,10 +924,10 @@ TEST(parse_out_of_memory_halfway_node) for (unsigned int i = 0; i < count; ++i) { - text[4*i + 0] = '<'; - text[4*i + 1] = 'n'; - text[4*i + 2] = '/'; - text[4*i + 3] = '>'; + text[4 * i + 0] = '<'; + text[4 * i + 1] = 'n'; + text[4 * i + 2] = '/'; + text[4 * i + 3] = '>'; } test_runner::_memory_fail_threshold = 65536; @@ -949,11 +947,11 @@ TEST(parse_out_of_memory_halfway_attr) for (unsigned int i = 0; i < count; ++i) { - text[5*i + 2] = ' '; - text[5*i + 3] = 'a'; - text[5*i + 4] = '='; - text[5*i + 5] = '"'; - text[5*i + 6] = '"'; + text[5 * i + 2] = ' '; + text[5 * i + 3] = 'a'; + text[5 * i + 4] = '='; + text[5 * i + 5] = '"'; + text[5 * i + 6] = '"'; } text[5 * count + 2] = '/'; @@ -984,10 +982,10 @@ TEST(parse_out_of_memory_allocator_state_sync) for (unsigned int i = 0; i < count; ++i) { - text[4*i + 0] = '<'; - text[4*i + 1] = 'n'; - text[4*i + 2] = '/'; - text[4*i + 3] = '>'; + text[4 * i + 0] = '<'; + text[4 * i + 1] = 'n'; + text[4 * i + 2] = '/'; + text[4 * i + 3] = '>'; } test_runner::_memory_fail_threshold = 65536; @@ -1072,18 +1070,18 @@ TEST(parse_bom_fragment) }; const test_data_t data[] = - { - { encoding_utf8, "\xef\xbb\xbf", 3, STR("") }, - { encoding_utf8, "\xef\xbb\xbftest", 7, STR("test") }, - { encoding_utf16_be, "\xfe\xff", 2, STR("") }, - { encoding_utf16_be, "\xfe\xff\x00t\x00o\x00s\x00t", 10, STR("tost") }, - { encoding_utf16_le, "\xff\xfe", 2, STR("") }, - { encoding_utf16_le, "\xff\xfet\x00o\x00s\x00t\x00", 10, STR("tost") }, - { encoding_utf32_be, "\x00\x00\xfe\xff", 4, STR("") }, - { encoding_utf32_be, "\x00\x00\xfe\xff\x00\x00\x00t\x00\x00\x00o\x00\x00\x00s\x00\x00\x00t", 20, STR("tost") }, - { encoding_utf32_le, "\xff\xfe\x00\x00", 4, STR("") }, - { encoding_utf32_le, "\xff\xfe\x00\x00t\x00\x00\x00o\x00\x00\x00s\x00\x00\x00t\x00\x00\x00", 20, STR("tost") }, - }; + { + {encoding_utf8, "\xef\xbb\xbf", 3, STR("")}, + {encoding_utf8, "\xef\xbb\xbftest", 7, STR("test")}, + {encoding_utf16_be, "\xfe\xff", 2, STR("")}, + {encoding_utf16_be, "\xfe\xff\x00t\x00o\x00s\x00t", 10, STR("tost")}, + {encoding_utf16_le, "\xff\xfe", 2, STR("")}, + {encoding_utf16_le, "\xff\xfet\x00o\x00s\x00t\x00", 10, STR("tost")}, + {encoding_utf32_be, "\x00\x00\xfe\xff", 4, STR("")}, + {encoding_utf32_be, "\x00\x00\xfe\xff\x00\x00\x00t\x00\x00\x00o\x00\x00\x00s\x00\x00\x00t", 20, STR("tost")}, + {encoding_utf32_le, "\xff\xfe\x00\x00", 4, STR("")}, + {encoding_utf32_le, "\xff\xfe\x00\x00t\x00\x00\x00o\x00\x00\x00s\x00\x00\x00t\x00\x00\x00", 20, STR("tost")}, + }; for (size_t i = 0; i < sizeof(data) / sizeof(data[0]); ++i) { @@ -1171,12 +1169,11 @@ TEST(parse_close_tag_eof) TEST(parse_fuzz_doctype) { unsigned char data[] = - { - 0x3b, 0x3c, 0x21, 0x44, 0x4f, 0x43, 0x54, 0x59, 0x50, 0x45, 0xef, 0xbb, 0xbf, 0x3c, 0x3f, 0x78, - 0x6d, 0x6c, 0x20, 0x76, 0x65, 0x72, 0x73, 0x69, 0x6f, 0x6e, 0x3d, 0x22, 0x31, 0x2e, 0x30, 0x22, - 0x3f, 0x3e, 0x3c, 0x21, 0x2d, 0x2d, 0x20, 0xe9, 0x80, 0xb1, 0xe5, 0xa0, 0xb1, 0xe3, 0x82, 0xb4, - 0xe3, 0x83, 0xb3, 0x20, 0xef, 0x83, 0x97, 0xe3, 0xa9, 0x2a, 0x20, 0x2d, 0x2d, 0x3e - }; + { + 0x3b, 0x3c, 0x21, 0x44, 0x4f, 0x43, 0x54, 0x59, 0x50, 0x45, 0xef, 0xbb, 0xbf, 0x3c, 0x3f, 0x78, + 0x6d, 0x6c, 0x20, 0x76, 0x65, 0x72, 0x73, 0x69, 0x6f, 0x6e, 0x3d, 0x22, 0x31, 0x2e, 0x30, 0x22, + 0x3f, 0x3e, 0x3c, 0x21, 0x2d, 0x2d, 0x20, 0xe9, 0x80, 0xb1, 0xe5, 0xa0, 0xb1, 0xe3, 0x82, 0xb4, + 0xe3, 0x83, 0xb3, 0x20, 0xef, 0x83, 0x97, 0xe3, 0xa9, 0x2a, 0x20, 0x2d, 0x2d, 0x3e}; xml_document doc; CHECK(doc.load_buffer(data, sizeof(data)).status == status_bad_doctype); @@ -1215,9 +1212,9 @@ TEST(parse_embed_pcdata) CHECK_STRING(child.child_value(), STR("outer")); CHECK_STRING(child.child_value(STR("inner2")), STR("value2")); - #ifndef PUGIXML_NO_XPATH +#ifndef PUGIXML_NO_XPATH CHECK_XPATH_NUMBER(doc, STR("count(node/child/*[starts-with(., 'value')])"), 2); - #endif +#endif CHECK_NODE(doc, STR("valuevalue1value2outertext")); CHECK_NODE_EX(doc, STR("\nvalue\n\nvalue1\nvalue2outer\ntext\n\n\n"), STR("\t"), 0); @@ -1286,23 +1283,23 @@ TEST(parse_encoding_detect_auto) }; const data_t data[] = - { - // BOM - { "\x00\x00\xfe\xff", 4, encoding_utf32_be }, - { "\xff\xfe\x00\x00", 4, encoding_utf32_le }, - { "\xfe\xff ", 4, encoding_utf16_be }, - { "\xff\xfe ", 4, encoding_utf16_le }, - { "\xef\xbb\xbf ", 4, encoding_utf8 }, - // automatic tag detection for < or ", 16, encoding_utf32_be }, - { "<\x00\x00\x00n\x00\x00\x00/\x00\x00\x00>\x00\x00\x00", 16, encoding_utf32_le }, - { "\x00<\x00?\x00n\x00?\x00>", 10, encoding_utf16_be }, - { "<\x00?\x00n\x00?\x00>\x00", 10, encoding_utf16_le }, - { "\x00<\x00n\x00/\x00>", 8, encoding_utf16_be }, - { "<\x00n\x00/\x00>\x00", 8, encoding_utf16_le }, - // ", 25, encoding_latin1 }, - }; + { + // BOM + {"\x00\x00\xfe\xff", 4, encoding_utf32_be}, + {"\xff\xfe\x00\x00", 4, encoding_utf32_le}, + {"\xfe\xff ", 4, encoding_utf16_be}, + {"\xff\xfe ", 4, encoding_utf16_le}, + {"\xef\xbb\xbf ", 4, encoding_utf8}, + // automatic tag detection for < or ", 16, encoding_utf32_be}, + {"<\x00\x00\x00n\x00\x00\x00/\x00\x00\x00>\x00\x00\x00", 16, encoding_utf32_le}, + {"\x00<\x00?\x00n\x00?\x00>", 10, encoding_utf16_be}, + {"<\x00?\x00n\x00?\x00>\x00", 10, encoding_utf16_le}, + {"\x00<\x00n\x00/\x00>", 8, encoding_utf16_be}, + {"<\x00n\x00/\x00>\x00", 8, encoding_utf16_le}, + // ", 25, encoding_latin1}, + }; for (size_t i = 0; i < sizeof(data) / sizeof(data[0]); ++i) { @@ -1324,27 +1321,27 @@ TEST(parse_encoding_detect_auto_incomplete) }; const data_t data[] = - { - // BOM - { "\x00\x00\xfe ", 4, encoding_utf8 }, - { "\x00\x00 ", 4, encoding_utf8 }, - { "\xff\xfe\x00 ", 4, encoding_utf16_le }, - { "\xfe ", 4, encoding_utf8 }, - { "\xff ", 4, encoding_utf8 }, - { "\xef\xbb ", 4, encoding_utf8 }, - { "\xef ", 4, encoding_utf8 }, - // automatic tag detection for < or \x00", 8, encoding_utf16_le }, - { "\x00", 8, encoding_utf16_be }, - { "<\x00?n/\x00>\x00", 8, encoding_utf16_le }, - { "\x00 ", 2, encoding_utf8 }, - // ", 25, encoding_utf8 }, - { "", 25, encoding_utf8 }, - { "", 25, encoding_utf8 }, - { "<_ABC encoding='latin1'/>", 25, encoding_utf8 }, - }; + { + // BOM + {"\x00\x00\xfe ", 4, encoding_utf8}, + {"\x00\x00 ", 4, encoding_utf8}, + {"\xff\xfe\x00 ", 4, encoding_utf16_le}, + {"\xfe ", 4, encoding_utf8}, + {"\xff ", 4, encoding_utf8}, + {"\xef\xbb ", 4, encoding_utf8}, + {"\xef ", 4, encoding_utf8}, + // automatic tag detection for < or \x00", 8, encoding_utf16_le}, + {"\x00", 8, encoding_utf16_be}, + {"<\x00?n/\x00>\x00", 8, encoding_utf16_le}, + {"\x00 ", 2, encoding_utf8}, + // ", 25, encoding_utf8}, + {"", 25, encoding_utf8}, + {"", 25, encoding_utf8}, + {"<_ABC encoding='latin1'/>", 25, encoding_utf8}, + }; for (size_t i = 0; i < sizeof(data) / sizeof(data[0]); ++i) { diff --git a/tests/test_parse_doctype.cpp b/tests/test_parse_doctype.cpp index 861e772..c55c2f0 100644 --- a/tests/test_parse_doctype.cpp +++ b/tests/test_parse_doctype.cpp @@ -3,8 +3,8 @@ #include "test.hpp" #include -#include #include +#include using namespace pugi; @@ -30,21 +30,28 @@ static bool test_doctype_wf(const char_t* decl) xml_document doc; // standalone - if (!load_concat(doc, decl) || !doc.first_child().empty()) return false; + if (!load_concat(doc, decl) || !doc.first_child().empty()) + return false; // pcdata pre/postfix - if (!load_concat(doc, STR("a"), decl) || !test_node(doc, STR("a"), STR(""), format_raw)) return false; - if (!load_concat(doc, decl, STR("b")) || !test_node(doc, STR("b"), STR(""), format_raw)) return false; - if (!load_concat(doc, STR("a"), decl, STR("b")) || !test_node(doc, STR("ab"), STR(""), format_raw)) return false; + if (!load_concat(doc, STR("a"), decl) || !test_node(doc, STR("a"), STR(""), format_raw)) + return false; + if (!load_concat(doc, decl, STR("b")) || !test_node(doc, STR("b"), STR(""), format_raw)) + return false; + if (!load_concat(doc, STR("a"), decl, STR("b")) || !test_node(doc, STR("ab"), STR(""), format_raw)) + return false; // node pre/postfix - if (!load_concat(doc, STR(""), decl) || !test_node(doc, STR(""), STR(""), format_raw)) return false; - if (!load_concat(doc, decl, STR("")) || !test_node(doc, STR(""), STR(""), format_raw)) return false; - if (!load_concat(doc, STR(""), decl, STR("")) || !test_node(doc, STR(""), STR(""), format_raw)) return false; + if (!load_concat(doc, STR(""), decl) || !test_node(doc, STR(""), STR(""), format_raw)) + return false; + if (!load_concat(doc, decl, STR("")) || !test_node(doc, STR(""), STR(""), format_raw)) + return false; + if (!load_concat(doc, STR(""), decl, STR("")) || !test_node(doc, STR(""), STR(""), format_raw)) + return false; - // check load-store contents preservation - CHECK(doc.load_string(decl, parse_doctype | parse_fragment)); - CHECK_NODE(doc, decl); + // check load-store contents preservation + CHECK(doc.load_string(decl, parse_doctype | parse_fragment)); + CHECK_NODE(doc, decl); return true; } @@ -54,13 +61,16 @@ static bool test_doctype_nwf(const char_t* decl) xml_document doc; // standalone - if (load_concat(doc, decl).status != status_bad_doctype) return false; + if (load_concat(doc, decl).status != status_bad_doctype) + return false; // pcdata postfix - if (load_concat(doc, decl, STR("b")).status != status_bad_doctype) return false; + if (load_concat(doc, decl, STR("b")).status != status_bad_doctype) + return false; // node postfix - if (load_concat(doc, decl, STR("")).status != status_bad_doctype) return false; + if (load_concat(doc, decl, STR("")).status != status_bad_doctype) + return false; return true; } @@ -281,9 +291,9 @@ TEST(parse_doctype_xmlconf_oasis_1) TEST_DOCTYPE_WF(" ]>"); TEST_DOCTYPE_WF("]>"); - // not actually a doctype :) - xml_document doc; - CHECK(doc.load_string(STR(" "), parse_full | parse_fragment) && doc.first_child().type() == node_comment && doc.last_child().type() == node_comment && doc.first_child().next_sibling() == doc.last_child()); + // not actually a doctype :) + xml_document doc; + CHECK(doc.load_string(STR(" "), parse_full | parse_fragment) && doc.first_child().type() == node_comment && doc.last_child().type() == node_comment && doc.first_child().next_sibling() == doc.last_child()); CHECK(doc.load_string(STR(" &a%b&#c?>"), parse_full | parse_fragment) && doc.first_child().type() == node_pi && doc.first_child() == doc.last_child()); } @@ -303,22 +313,22 @@ TEST(parse_doctype_xmlconf_xmltest_1) TEST_XML_FLAGS(parse_doctype_value, " \"> ]>", parse_fragment | parse_doctype) { - xml_node n = doc.first_child(); + xml_node n = doc.first_child(); - CHECK(n.type() == node_doctype); - CHECK_STRING(n.value(), STR("doc [ \"> ]")); + CHECK(n.type() == node_doctype); + CHECK_STRING(n.value(), STR("doc [ \"> ]")); } TEST(parse_doctype_error_toplevel) { - xml_document doc; - CHECK(doc.load_string(STR("")).status == status_bad_doctype); - CHECK(doc.load_string(STR(""), parse_doctype).status == status_bad_doctype); + xml_document doc; + CHECK(doc.load_string(STR("")).status == status_bad_doctype); + CHECK(doc.load_string(STR(""), parse_doctype).status == status_bad_doctype); } TEST(parse_doctype_error_ignore) { - xml_document doc; + xml_document doc; CHECK(doc.load_string(STR(" s = L"abcd"; + std::basic_string s = L"abcd"; - CHECK(as_utf8(s) == "abcd"); + CHECK(as_utf8(s) == "abcd"); } #endif diff --git a/tests/test_write.cpp b/tests/test_write.cpp index 95cc566..b563bf8 100644 --- a/tests/test_write.cpp +++ b/tests/test_write.cpp @@ -2,9 +2,9 @@ #include "writer_string.hpp" -#include #include #include +#include using namespace pugi; @@ -199,11 +199,11 @@ TEST_XML(write_escape, "text") TEST_XML(write_escape_unicode, "") { #ifdef PUGIXML_WCHAR_MODE - #ifdef U_LITERALS - CHECK_NODE(doc, STR("")); - #else - CHECK_NODE(doc, STR("")); - #endif +#ifdef U_LITERALS + CHECK_NODE(doc, STR("")); +#else + CHECK_NODE(doc, STR("")); +#endif #else CHECK_NODE(doc, STR("")); #endif @@ -217,7 +217,7 @@ TEST_XML(write_no_escapes, "text") CHECK_NODE_EX(doc, STR("'\"&\x04\r\n\t\"><>'\"&\x04\r\n\t"), STR(""), format_raw | format_no_escapes); } -struct test_writer: xml_writer +struct test_writer : xml_writer { std::basic_string contents; @@ -302,7 +302,7 @@ TEST(write_encodings) CHECK(v.size() == 10 && v[0] == '<' && v[1] == 0x54 && v[2] == 0xA2 && v[3] == 0x20AC && v[4] == wchar_cast(0xd852) && v[5] == wchar_cast(0xdf62) && v[6] == ' ' && v[7] == '/' && v[8] == '>' && v[9] == '\n'); } - CHECK(test_write_narrow(doc, format_default, encoding_latin1, "<\x54\xA2?? />\n", 9)); + CHECK(test_write_narrow(doc, format_default, encoding_latin1, "<\x54\xA2?? />\n", 9)); } #ifdef PUGIXML_WCHAR_MODE @@ -313,7 +313,8 @@ TEST(write_encoding_huge) // make a large utf16 name consisting of 6-byte char pairs (6 does not divide internal buffer size, so will need split correction) std::string s_utf16 = std::string("\x00<", 2); - for (unsigned int i = 0; i < N; ++i) s_utf16 += "\x20\xAC\xd8\x52\xdf\x62"; + for (unsigned int i = 0; i < N; ++i) + s_utf16 += "\x20\xAC\xd8\x52\xdf\x62"; s_utf16 += std::string("\x00/\x00>", 4); @@ -322,7 +323,8 @@ TEST(write_encoding_huge) std::string s_utf8 = "<"; - for (unsigned int j = 0; j < N; ++j) s_utf8 += "\xE2\x82\xAC\xF0\xA4\xAD\xA2"; + for (unsigned int j = 0; j < N; ++j) + s_utf8 += "\xE2\x82\xAC\xF0\xA4\xAD\xA2"; s_utf8 += " />\n"; @@ -340,7 +342,8 @@ TEST(write_encoding_huge_invalid) // make a large utf16 name consisting of leading surrogate chars std::basic_string s_utf16; - for (unsigned int i = 0; i < N; ++i) s_utf16 += static_cast(0xd852); + for (unsigned int i = 0; i < N; ++i) + s_utf16 += static_cast(0xd852); xml_document doc; doc.append_child().set_name(s_utf16.c_str()); @@ -356,7 +359,8 @@ TEST(write_encoding_huge) // make a large utf8 name consisting of 3-byte chars (3 does not divide internal buffer size, so will need split correction) std::string s_utf8 = "<"; - for (unsigned int i = 0; i < N; ++i) s_utf8 += "\xE2\x82\xAC"; + for (unsigned int i = 0; i < N; ++i) + s_utf8 += "\xE2\x82\xAC"; s_utf8 += "/>"; @@ -365,7 +369,8 @@ TEST(write_encoding_huge) std::string s_utf16 = std::string("\x00<", 2); - for (unsigned int j = 0; j < N; ++j) s_utf16 += "\x20\xAC"; + for (unsigned int j = 0; j < N; ++j) + s_utf16 += "\x20\xAC"; s_utf16 += std::string("\x00 \x00/\x00>\x00\n", 8); @@ -379,7 +384,8 @@ TEST(write_encoding_huge_invalid) // make a large utf8 name consisting of non-leading chars std::string s_utf8; - for (unsigned int i = 0; i < N; ++i) s_utf8 += "\x82"; + for (unsigned int i = 0; i < N; ++i) + s_utf8 += "\x82"; xml_document doc; doc.append_child().set_name(s_utf8.c_str()); @@ -416,22 +422,22 @@ TEST(write_unicode_invalid_utf16) if (wcharsize == 2) { // check non-terminated degenerate handling - #ifdef U_LITERALS +#ifdef U_LITERALS CHECK(test_write_unicode_invalid(L"a\uda1d", "a")); CHECK(test_write_unicode_invalid(L"a\uda1d_", "a_")); - #else +#else CHECK(test_write_unicode_invalid(L"a\xda1d", "a")); CHECK(test_write_unicode_invalid(L"a\xda1d_", "a_")); - #endif +#endif // check incorrect leading code - #ifdef U_LITERALS +#ifdef U_LITERALS CHECK(test_write_unicode_invalid(L"a\ude24", "a")); CHECK(test_write_unicode_invalid(L"a\ude24_", "a_")); - #else +#else CHECK(test_write_unicode_invalid(L"a\xde24", "a")); CHECK(test_write_unicode_invalid(L"a\xde24_", "a_")); - #endif +#endif } } #else @@ -569,19 +575,17 @@ TEST(write_pcdata_whitespace_fixedpoint) const char_t* data = STR(" test \n \n \n"); static const unsigned int flags_parse[] = - { - 0, - parse_ws_pcdata, - parse_ws_pcdata_single, - parse_trim_pcdata - }; + { + 0, + parse_ws_pcdata, + parse_ws_pcdata_single, + parse_trim_pcdata}; static const unsigned int flags_format[] = - { - 0, - format_raw, - format_indent - }; + { + 0, + format_raw, + format_indent}; for (unsigned int i = 0; i < sizeof(flags_parse) / sizeof(flags_parse[0]); ++i) { @@ -618,7 +622,7 @@ TEST_XML(write_no_empty_element_tags, "textpremid<text&escapepostfin", parse_full) { - const unsigned int flagset[] = { format_indent, format_raw, format_no_declaration, format_indent_attributes, format_no_empty_element_tags }; + const unsigned int flagset[] = {format_indent, format_raw, format_no_declaration, format_indent_attributes, format_no_empty_element_tags}; size_t flagcount = sizeof(flagset) / sizeof(flagset[0]); for (size_t i = 0; i < (size_t(1) << flagcount); ++i) @@ -669,7 +673,7 @@ TEST(write_flush_coverage) } #ifndef PUGIXML_NO_EXCEPTIONS -struct throwing_writer: xml_writer +struct throwing_writer : xml_writer { virtual void write(const void*, size_t) PUGIXML_OVERRIDE { diff --git a/tests/test_xpath.cpp b/tests/test_xpath.cpp index 1a2c477..35921ed 100644 --- a/tests/test_xpath.cpp +++ b/tests/test_xpath.cpp @@ -5,10 +5,10 @@ #include #include -#include -#include #include #include +#include +#include using namespace pugi; @@ -24,7 +24,8 @@ TEST(xpath_allocator_many_pages) { std::basic_string query = STR("0"); - for (int i = 0; i < 128; ++i) query += STR("+string-length('abcdefgh')"); + for (int i = 0; i < 128; ++i) + query += STR("+string-length('abcdefgh')"); CHECK_XPATH_NUMBER(xml_node(), query.c_str(), 1024); } @@ -33,7 +34,8 @@ TEST(xpath_allocator_large_page) { std::basic_string query; - for (int i = 0; i < 1024; ++i) query += STR("abcdefgh"); + for (int i = 0; i < 1024; ++i) + query += STR("abcdefgh"); CHECK_XPATH_NUMBER(xml_node(), (STR("string-length('") + query + STR("')")).c_str(), 8192); } @@ -147,10 +149,10 @@ TEST(xpath_sort_random_medium) { xml_document doc; load_document_copy(doc, STR("") - STR("testtest") - STR("testtest") - STR("testtest") - STR("")); + STR("testtest") + STR("testtest") + STR("testtest") + STR("")); xpath_node_set ns = doc.select_nodes(STR("//node() | //@*")); @@ -162,24 +164,25 @@ TEST(xpath_sort_random_medium) xpath_node_set_tester tester(copy, "sorted order failed"); - for (unsigned int i = 2; i < 39; ++i) tester % i; + for (unsigned int i = 2; i < 39; ++i) + tester % i; } TEST(xpath_sort_random_large) { xml_document doc; load_document_copy(doc, STR("") - STR("testtest") - STR("testtest") - STR("testtest") - STR("testtest") - STR("testtest") - STR("testtest") - STR("testtest") - STR("testtest") - STR("testtest") - STR("testtest") - STR("test")); + STR("testtest") + STR("testtest") + STR("testtest") + STR("testtest") + STR("testtest") + STR("testtest") + STR("testtest") + STR("testtest") + STR("testtest") + STR("testtest") + STR("test")); xpath_node_set ns = doc.select_nodes(STR("//node() | //@*")); @@ -191,7 +194,8 @@ TEST(xpath_sort_random_large) xpath_node_set_tester tester(copy, "sorted order failed"); - for (unsigned int i = 2; i < 129; ++i) tester % i; + for (unsigned int i = 2; i < 129; ++i) + tester % i; } TEST(xpath_long_numbers_parse) @@ -247,7 +251,8 @@ TEST(xpath_denorm_numbers) // 10^-318 - double denormal for (int i = 0; i < 106; ++i) { - if (i != 0) query += STR(" * "); + if (i != 0) + query += STR(" * "); query += STR("0.001"); } @@ -282,7 +287,7 @@ TEST_XML(xpath_rexml_2, "zzz< TEST_XML(xpath_rexml_3, "
free flowing text.
free flowing text.
free flowing text.
") { CHECK_XPATH_NODESET(doc, STR("//section[../self::section[@role=\"division\"]]")) % 10 % 15; - CHECK_XPATH_NODESET(doc, STR("//section[@role=\"subdivision\" and not(../self::section[@role=\"division\"])]")) % 3; + CHECK_XPATH_NODESET(doc, STR("//section[@role=\"subdivision\" and not(../self::section[@role=\"division\"])]")) % 3; CHECK_XPATH_NODESET(doc, STR("//section[@role=\"subdivision\"][not(../self::section[@role=\"division\"])]")) % 3; } @@ -716,7 +721,7 @@ TEST(xpath_sort_crossdoc_different_depth) TEST_XML(xpath_sort_empty_node, "") { xml_node n = doc.child(STR("node")); - xpath_node nodes[] = { n.child(STR("child2")), xml_node(), n.child(STR("child1")), xml_node() }; + xpath_node nodes[] = {n.child(STR("child2")), xml_node(), n.child(STR("child1")), xml_node()}; xpath_node_set ns(nodes, nodes + sizeof(nodes) / sizeof(nodes[0])); ns.sort(); @@ -728,9 +733,10 @@ TEST(xpath_allocate_string_out_of_memory) { std::basic_string query; - for (int i = 0; i < 1024; ++i) query += STR("abcdefgh"); + for (int i = 0; i < 1024; ++i) + query += STR("abcdefgh"); - test_runner::_memory_fail_threshold = 8*1024; + test_runner::_memory_fail_threshold = 8 * 1024; #ifndef __DMC__ // DigitalMars exception handling crashes instead of catching the exception... CHECK_ALLOC_FAIL(CHECK(!xpath_query(query.c_str()))); diff --git a/tests/test_xpath_api.cpp b/tests/test_xpath_api.cpp index 7abf7d9..9926008 100644 --- a/tests/test_xpath_api.cpp +++ b/tests/test_xpath_api.cpp @@ -141,10 +141,10 @@ TEST_XML(xpath_api_nodeset_copy, "") TEST(xpath_api_nodeset_copy_empty) { - xpath_node_set set; - xpath_node_set set2 = set; - xpath_node_set set3; - set3 = set; + xpath_node_set set; + xpath_node_set set2 = set; + xpath_node_set set3; + set3 = set; } TEST_XML(xpath_api_evaluate, "") @@ -430,7 +430,6 @@ TEST_XML(xpath_api_nodeset_move_ctor, "") CHECK(move[1] == doc.first_child().first_child()); } - TEST_XML(xpath_api_nodeset_move_ctor_single, "") { xpath_node_set set = doc.select_nodes(STR("node/bar")); diff --git a/tests/test_xpath_functions.cpp b/tests/test_xpath_functions.cpp index 0d1d630..3dcead9 100644 --- a/tests/test_xpath_functions.cpp +++ b/tests/test_xpath_functions.cpp @@ -57,7 +57,7 @@ TEST_XML(xpath_number_sum, "123789") CHECK_XPATH_NUMBER(n, STR("sum(.)"), 123789); // 123 .. 789 CHECK_XPATH_NUMBER(n, STR("sum(./descendant-or-self::node())"), 125490); // node + 123 + child + 789 = 123789 + 123 + 789 + 789 = 125490 - CHECK_XPATH_NUMBER(n, STR("sum(.//node())"), 1701); // 123 + child + 789 = 123 + 789 + 789 + CHECK_XPATH_NUMBER(n, STR("sum(.//node())"), 1701); // 123 + child + 789 = 123 + 789 + 789 CHECK_XPATH_NUMBER_NAN(doc.last_child(), STR("sum(.)")); // sum with 2 arguments @@ -423,9 +423,9 @@ TEST(xpath_string_substring_after) TEST_XML(xpath_string_substring_after_heap, "foobar") { - CHECK_XPATH_STRING(doc, STR("substring-after(node, 'fo')"), STR("obar")); - CHECK_XPATH_STRING(doc, STR("substring-after(node, 'fooba')"), STR("r")); - CHECK_XPATH_STRING(doc, STR("substring-after(node, 'foobar')"), STR("")); + CHECK_XPATH_STRING(doc, STR("substring-after(node, 'fo')"), STR("obar")); + CHECK_XPATH_STRING(doc, STR("substring-after(node, 'fooba')"), STR("r")); + CHECK_XPATH_STRING(doc, STR("substring-after(node, 'foobar')"), STR("")); } TEST(xpath_string_substring) @@ -452,7 +452,7 @@ TEST(xpath_string_substring) CHECK_XPATH_STRING(c, STR("substring('abcd', 0 div 0)"), STR("")); CHECK_XPATH_STRING(c, STR("substring('', 1)"), STR("")); CHECK_XPATH_STRING(c, STR("substring('', 0)"), STR("")); - CHECK_XPATH_STRING(c, STR("substring(substring('internalexternalcorrect substring',9),9)"), STR("correct substring")); + CHECK_XPATH_STRING(c, STR("substring(substring('internalexternalcorrect substring',9),9)"), STR("correct substring")); // substring with 3 arguments CHECK_XPATH_STRING(c, STR("substring('abcd', 2, 1)"), STR("b")); @@ -490,9 +490,9 @@ TEST(xpath_string_substring) TEST_XML(xpath_string_substring_heap, "foobar") { - CHECK_XPATH_STRING(doc, STR("substring(node, 3)"), STR("obar")); - CHECK_XPATH_STRING(doc, STR("substring(node, 6)"), STR("r")); - CHECK_XPATH_STRING(doc, STR("substring(node, 7)"), STR("")); + CHECK_XPATH_STRING(doc, STR("substring(node, 3)"), STR("obar")); + CHECK_XPATH_STRING(doc, STR("substring(node, 6)"), STR("r")); + CHECK_XPATH_STRING(doc, STR("substring(node, 7)"), STR("")); } TEST_XML(xpath_string_string_length, "123") @@ -590,7 +590,7 @@ TEST_XML(xpath_nodeset_last, "") // last with 0 arguments CHECK_XPATH_NUMBER(n, STR("last()"), 1); CHECK_XPATH_NODESET(n, STR("c1[last() = 1]")); - CHECK_XPATH_NODESET(n, STR("c1[last() = 2]")) % 3 % 4; // c1, c1 + CHECK_XPATH_NODESET(n, STR("c1[last() = 2]")) % 3 % 4; // c1, c1 CHECK_XPATH_NODESET(n, STR("c2/preceding-sibling::node()[last() = 2]")) % 4 % 3; // c1, c1 // last with 1 argument @@ -829,7 +829,8 @@ TEST(xpath_string_translate_table_out_of_memory) for (size_t i = 0; i < count; ++i) { - if (i != 0) query += STR(","); + if (i != 0) + query += STR(","); query += STR("translate('a','a','A')"); } diff --git a/tests/test_xpath_parse.cpp b/tests/test_xpath_parse.cpp index daa12bf..0e241d7 100644 --- a/tests/test_xpath_parse.cpp +++ b/tests/test_xpath_parse.cpp @@ -99,82 +99,80 @@ TEST(xpath_semantics_posinv) // coverage for contains() TEST(xpath_parse_paths_valid) { - const char_t* paths[] = - { - // From Jaxen tests - STR("foo[.='bar']"), STR("foo[.!='bar']"), STR("/"), STR("*"), STR("//foo"), STR("/*"), STR("/."), STR("/foo[/bar[/baz]]"), - STR("/foo/bar/baz[(1 or 2) + 3 * 4 + 8 and 9]"), STR("/foo/bar/baz"), STR("(.)[1]"), STR("self::node()"), STR("."), STR("count(/)"), - STR("foo[1]"), STR("/baz[(1 or 2) + 3 * 4 + 8 and 9]"), STR("foo/bar[/baz[(1 or 2) - 3 mod 4 + 8 and 9 div 8]]"), - STR("foo/bar/yeah:baz[a/b/c and toast]"), STR("/foo/bar[../x='123']"), STR("/foo[@bar='1234']"), STR("foo|bar"), - STR("/foo|/bar[@id='1234']"), STR("count(//author/attribute::*)"), STR("/child::node()/child::node()[@id='_13563275']"), - STR("10 + (count(descendant::author) * 5)"), STR("10 + count(descendant::author) * 5"), STR("2 + (2 * 5)"), STR("//foo:bar"), - STR("count(//author)+5"), STR("count(//author)+count(//author/attribute::*)"), STR("/foo/bar[@a='1' and @c!='2']"), - STR("12 + (count(//author)+count(//author/attribute::*)) div 2"), STR("text()[.='foo']"), STR("/*/*[@id='123']") - STR("/foo/bar[@a='1' and @b='2']"), STR("/foo/bar[@a='1' and @b!='2']"), STR("//attribute::*[.!='crunchy']"), - STR("'//*[contains(string(text()),\"yada yada\")]'"), + const char_t* paths[] = + { + // From Jaxen tests + STR("foo[.='bar']"), STR("foo[.!='bar']"), STR("/"), STR("*"), STR("//foo"), STR("/*"), STR("/."), STR("/foo[/bar[/baz]]"), + STR("/foo/bar/baz[(1 or 2) + 3 * 4 + 8 and 9]"), STR("/foo/bar/baz"), STR("(.)[1]"), STR("self::node()"), STR("."), STR("count(/)"), + STR("foo[1]"), STR("/baz[(1 or 2) + 3 * 4 + 8 and 9]"), STR("foo/bar[/baz[(1 or 2) - 3 mod 4 + 8 and 9 div 8]]"), + STR("foo/bar/yeah:baz[a/b/c and toast]"), STR("/foo/bar[../x='123']"), STR("/foo[@bar='1234']"), STR("foo|bar"), + STR("/foo|/bar[@id='1234']"), STR("count(//author/attribute::*)"), STR("/child::node()/child::node()[@id='_13563275']"), + STR("10 + (count(descendant::author) * 5)"), STR("10 + count(descendant::author) * 5"), STR("2 + (2 * 5)"), STR("//foo:bar"), + STR("count(//author)+5"), STR("count(//author)+count(//author/attribute::*)"), STR("/foo/bar[@a='1' and @c!='2']"), + STR("12 + (count(//author)+count(//author/attribute::*)) div 2"), STR("text()[.='foo']"), STR("/*/*[@id='123']") STR("/foo/bar[@a='1' and @b='2']"), STR("/foo/bar[@a='1' and @b!='2']"), STR("//attribute::*[.!='crunchy']"), + STR("'//*[contains(string(text()),\"yada yada\")]'"), - // From ajaxslt tests - STR("@*"), STR("@*|node()"), STR("/descendant-or-self::div"), STR("/div"), STR("//div"), STR("/descendant-or-self::node()/child::para"), - STR("substring('12345', 0, 3)"), STR("//title | //link"), STR("x//title"), STR("x/title"), STR("id('a')//title"), STR("//*[@about]"), - STR("count(descendant::*)"), STR("count(descendant::*) + count(ancestor::*)"), STR("@*|text()"), STR("*|/"), STR("source|destination"), - STR("page != 'to' and page != 'from'"), STR("substring-after(icon/@image, '/mapfiles/marker')"), STR("substring-before(str, c)"), STR("page = 'from'"), - STR("segments/@time"), STR("child::para"), STR("child::*"), STR("child::text()"), STR("child::node()"), STR("attribute::name"), STR("attribute::*"), - STR("descendant::para"), STR("ancestor::div"), STR("ancestor-or-self::div"), STR("descendant-or-self::para"), STR("self::para"), STR("child::*/child::para"), - STR("concat(substring-before(@image,'marker'),'icon',substring-after(@image,'marker'))"), STR("/"), STR("/descendant::para"), STR("/descendant::olist/child::item"), - STR("child::para[position()=1]"), STR("child::para[position()=last()]"), STR("child::para[position()=last()-1]"), STR("child::para[position()>1]"), - STR("following-sibling::chapter[position()=1]"), STR("preceding-sibling::chapter[position()=1]"), STR("/descendant::figure[position()=42]"), - STR("/child::doc/child::chapter[position()=5]/child::section[position()=2]"), STR("child::chapter/descendant::para"), STR("child::para[attribute::type='warning']"), - STR("child::para[attribute::type='warning'][position()=5]"), STR("child::para[position()=5][attribute::type='warning']"), STR("child::chapter[child::title='Introduction']"), - STR("child::chapter[child::title]"), STR("child::*[self::chapter or self::appendix]"), STR("child::*[self::chapter or self::appendix][position()=last()]"), - STR("count(//*[id='u1']|//*[id='u2'])"), STR("count(//*[id='u1']|//*[class='u'])"), STR("count(//*[class='u']|//*[class='u'])"), STR("count(//*[class='u']|//*[id='u1'])"), - STR("count(//*[@id='self']/ancestor-or-self::*)"), STR("count(//*[@id='self']/ancestor::*)"), STR("count(//*[@id='self']/attribute::*)"), STR("count(//*[@id='self']/child::*)"), - STR("count(//*[@id='self']/descendant-or-self::*)"), STR("count(//*[@id='self']/descendant::*)"), STR("count(//*[@id='self']/following-sibling::*)"), - STR("count(//*[@id='self']/following::*)"), STR("//*[@id='self']/parent::*/@id"), STR("count(//*[@id='self']/preceding-sibling::*)"), - STR("count(//*[@id='self']/preceding::*)"), STR("//*[@id='self']/self::*/@id"), STR("id('nested1')/div[1]//input[2]"), STR("id('foo')//div[contains(@id, 'useful')]//input"), - STR("(//table[@class='stylee'])//th[text()='theHeaderText']/../td"), STR("address"), STR("address=string(/page/user/defaultlocation)"), STR("count-of-snippet-of-url = 0"), - STR("daddr"), STR("form"), STR("form = 'from'"), STR("form = 'to'"), STR("form='near'"), STR("home"), STR("i"), STR("i > page and i < page + range"), - STR("i < page and i >= page - range"), STR("i < @max"), STR("i <= page"), STR("i + 1"), STR("i = page"), STR("i = 1"), STR("info = position() or (not(info) and position() = 1)"), - STR("is-first-order"), STR("is-first-order and snippets-exist"), STR("more"), STR("more > 0"), STR("near-point"), STR("page"), STR("page != 'from'"), STR("page != 'to'"), - STR("page != 'to' and page != 'from'"), STR("page > 1"), STR("page = 'basics'"), STR("page = 'details'"), STR("page = 'from'"), STR("page = 'to'"), STR("page='from'"), - STR("page='to'"), STR("r >= 0.5"), STR("r >= 1"), STR("r - 0"), STR("r - 1"), STR("r - 2"), STR("r - 3"), STR("r - 4"), STR("saddr"), STR("sources"), STR("sources[position() < details]"), - STR("src"), STR("str"), STR("\"'\""), STR("(//location[string(info/references/reference[1]/url)=string(current-url)]/info/references/reference[1])[1]"), - STR("(not(count-of-snippet-of-url = 0) and (position() = 1) or not(current-url = //locations/location[position() = last-pos]//reference[1]/url))"), - STR("(not(info) and position() = 1) or info = position()"), STR("."), STR("../@arg0"), STR("../@filterpng"), STR("/page/@filterpng"), STR("4"), STR("@attribution"), - STR("@id"), STR("@max > @num"), STR("@meters > 16093"), STR("@name"), STR("@start div @num + 1"), STR("@url"), STR("ad"), STR("address/line"), STR("adsmessage"), - STR("attr"), STR("boolean(location[@id='near'][icon/@image])"), STR("bubble/node()"), STR("calltoaction/node()"), STR("category"), STR("contains(str, c)"), - STR("count(//location[string(info/references/reference[1]/url)=string(current-url)]//snippet)"), STR("count(//snippet)"), STR("count(attr)"), STR("count(location)"), - STR("count(structured/source) > 1"), STR("description/node()"), STR("destination"), STR("destinationAddress"), STR("domain"), STR("false()"), STR("icon/@class != 'noicon'"), - STR("icon/@image"), STR("info"), STR("info/address/line"), STR("info/distance"), STR("info/distance and near-point"), STR("info/distance and info/phone and near-point"), - STR("info/distance or info/phone"), STR("info/panel/node()"), STR("info/phone"), STR("info/references/reference[1]"), STR("info/references/reference[1]/snippet"), - STR("info/references/reference[1]/url"), STR("info/title"), STR("info/title/node()"), STR("line"), STR("location"), STR("location[@id!='near']"), STR("location[@id='near'][icon/@image]"), - STR("location[position() > umlocations div 2]"), STR("location[position() <= numlocations div 2]"), STR("locations"), STR("locations/location"), STR("near"), STR("node()"), - STR("not(count-of-snippets = 0)"), STR("not(form = 'from')"), STR("not(form = 'near')"), STR("not(form = 'to')"), STR("not(../@page)"), STR("not(structured/source)"), STR("notice"), - STR("number(../@info)"), STR("number(../@items)"), STR("number(/page/@linewidth)"), STR("page/ads"), STR("page/directions"), STR("page/error"), STR("page/overlay"), - STR("page/overlay/locations/location"), STR("page/refinements"), STR("page/request/canonicalnear"), STR("page/request/near"), STR("page/request/query"), STR("page/spelling/suggestion"), - STR("page/user/defaultlocation"), STR("phone"), STR("position()"), STR("position() != 1"), STR("position() != last()"), STR("position() > 1"), STR("position() < details"), - STR("position()-1"), STR("query"), STR("references/@total"), STR("references/reference"), STR("references/reference/domain"), STR("references/reference/url"), - STR("reviews/@positive div (reviews/@positive + reviews/@negative) * 5"), STR("reviews/@positive div (reviews/@positive + reviews/@negative) * (5)"), STR("reviews/@total"), - STR("reviews/@total > 1"), STR("reviews/@total > 5"), STR("reviews/@total = 1"), STR("segments/@distance"), STR("segments/@time"), STR("segments/segment"), STR("shorttitle/node()"), - STR("snippet"), STR("snippet/node()"), STR("source"), STR("sourceAddress"), STR("sourceAddress and destinationAddress"), STR("string(../@daddr)"), STR("string(../@form)"), - STR("string(../@page)"), STR("string(../@saddr)"), STR("string(info/title)"), STR("string(page/request/canonicalnear) != ''"), STR("string(page/request/near) != ''"), - STR("string-length(address) > linewidth"), STR("structured/@total - details"), STR("structured/source"), STR("structured/source[@name]"), STR("substring(address, 1, linewidth - 3)"), - STR("substring-after(str, c)"), STR("substring-after(icon/@image, '/mapfiles/marker')"), STR("substring-before(str, c)"), STR("tagline/node()"), STR("targetedlocation"), - STR("title"), STR("title/node()"), STR("true()"), STR("url"), STR("visibleurl"), STR("id(\"level10\")/ancestor::SPAN"), STR("id(\"level10\")/ancestor-or-self::SPAN"), STR("//attribute::*"), - STR("child::HTML/child::BODY/child::H1"), STR("descendant::node()"), STR("descendant-or-self::SPAN"), STR("id(\"first\")/following::text()"), STR("id(\"first\")/following-sibling::node()"), - STR("id(\"level10\")/parent::node()"), STR("id(\"last\")/preceding::text()"), STR("id(\"last\")/preceding-sibling::node()"), STR("/HTML/BODY/H1/self::node()"), STR("//*[@name]"), - STR("id(\"pet\")/SELECT[@name=\"species\"]/OPTION[@selected]/@value"), STR("descendant::INPUT[@name=\"name\"]/@value"), STR("id(\"pet\")/INPUT[@name=\"gender\" and @checked]/@value"), - STR("//TEXTAREA[@name=\"description\"]/text()"), STR("id(\"div1\")|id(\"div2\")|id(\"div3 div4 div5\")"), STR("//LI[1]"), STR("//LI[last()]/text()"), STR("//LI[position() mod 2]/@class"), - STR("//text()[.=\"foo\"]"), STR("descendant-or-self::SPAN[position() > 2]"), STR("descendant::*[contains(@class,\" fruit \")]"), + // From ajaxslt tests + STR("@*"), STR("@*|node()"), STR("/descendant-or-self::div"), STR("/div"), STR("//div"), STR("/descendant-or-self::node()/child::para"), + STR("substring('12345', 0, 3)"), STR("//title | //link"), STR("x//title"), STR("x/title"), STR("id('a')//title"), STR("//*[@about]"), + STR("count(descendant::*)"), STR("count(descendant::*) + count(ancestor::*)"), STR("@*|text()"), STR("*|/"), STR("source|destination"), + STR("page != 'to' and page != 'from'"), STR("substring-after(icon/@image, '/mapfiles/marker')"), STR("substring-before(str, c)"), STR("page = 'from'"), + STR("segments/@time"), STR("child::para"), STR("child::*"), STR("child::text()"), STR("child::node()"), STR("attribute::name"), STR("attribute::*"), + STR("descendant::para"), STR("ancestor::div"), STR("ancestor-or-self::div"), STR("descendant-or-self::para"), STR("self::para"), STR("child::*/child::para"), + STR("concat(substring-before(@image,'marker'),'icon',substring-after(@image,'marker'))"), STR("/"), STR("/descendant::para"), STR("/descendant::olist/child::item"), + STR("child::para[position()=1]"), STR("child::para[position()=last()]"), STR("child::para[position()=last()-1]"), STR("child::para[position()>1]"), + STR("following-sibling::chapter[position()=1]"), STR("preceding-sibling::chapter[position()=1]"), STR("/descendant::figure[position()=42]"), + STR("/child::doc/child::chapter[position()=5]/child::section[position()=2]"), STR("child::chapter/descendant::para"), STR("child::para[attribute::type='warning']"), + STR("child::para[attribute::type='warning'][position()=5]"), STR("child::para[position()=5][attribute::type='warning']"), STR("child::chapter[child::title='Introduction']"), + STR("child::chapter[child::title]"), STR("child::*[self::chapter or self::appendix]"), STR("child::*[self::chapter or self::appendix][position()=last()]"), + STR("count(//*[id='u1']|//*[id='u2'])"), STR("count(//*[id='u1']|//*[class='u'])"), STR("count(//*[class='u']|//*[class='u'])"), STR("count(//*[class='u']|//*[id='u1'])"), + STR("count(//*[@id='self']/ancestor-or-self::*)"), STR("count(//*[@id='self']/ancestor::*)"), STR("count(//*[@id='self']/attribute::*)"), STR("count(//*[@id='self']/child::*)"), + STR("count(//*[@id='self']/descendant-or-self::*)"), STR("count(//*[@id='self']/descendant::*)"), STR("count(//*[@id='self']/following-sibling::*)"), + STR("count(//*[@id='self']/following::*)"), STR("//*[@id='self']/parent::*/@id"), STR("count(//*[@id='self']/preceding-sibling::*)"), + STR("count(//*[@id='self']/preceding::*)"), STR("//*[@id='self']/self::*/@id"), STR("id('nested1')/div[1]//input[2]"), STR("id('foo')//div[contains(@id, 'useful')]//input"), + STR("(//table[@class='stylee'])//th[text()='theHeaderText']/../td"), STR("address"), STR("address=string(/page/user/defaultlocation)"), STR("count-of-snippet-of-url = 0"), + STR("daddr"), STR("form"), STR("form = 'from'"), STR("form = 'to'"), STR("form='near'"), STR("home"), STR("i"), STR("i > page and i < page + range"), + STR("i < page and i >= page - range"), STR("i < @max"), STR("i <= page"), STR("i + 1"), STR("i = page"), STR("i = 1"), STR("info = position() or (not(info) and position() = 1)"), + STR("is-first-order"), STR("is-first-order and snippets-exist"), STR("more"), STR("more > 0"), STR("near-point"), STR("page"), STR("page != 'from'"), STR("page != 'to'"), + STR("page != 'to' and page != 'from'"), STR("page > 1"), STR("page = 'basics'"), STR("page = 'details'"), STR("page = 'from'"), STR("page = 'to'"), STR("page='from'"), + STR("page='to'"), STR("r >= 0.5"), STR("r >= 1"), STR("r - 0"), STR("r - 1"), STR("r - 2"), STR("r - 3"), STR("r - 4"), STR("saddr"), STR("sources"), STR("sources[position() < details]"), + STR("src"), STR("str"), STR("\"'\""), STR("(//location[string(info/references/reference[1]/url)=string(current-url)]/info/references/reference[1])[1]"), + STR("(not(count-of-snippet-of-url = 0) and (position() = 1) or not(current-url = //locations/location[position() = last-pos]//reference[1]/url))"), + STR("(not(info) and position() = 1) or info = position()"), STR("."), STR("../@arg0"), STR("../@filterpng"), STR("/page/@filterpng"), STR("4"), STR("@attribution"), + STR("@id"), STR("@max > @num"), STR("@meters > 16093"), STR("@name"), STR("@start div @num + 1"), STR("@url"), STR("ad"), STR("address/line"), STR("adsmessage"), + STR("attr"), STR("boolean(location[@id='near'][icon/@image])"), STR("bubble/node()"), STR("calltoaction/node()"), STR("category"), STR("contains(str, c)"), + STR("count(//location[string(info/references/reference[1]/url)=string(current-url)]//snippet)"), STR("count(//snippet)"), STR("count(attr)"), STR("count(location)"), + STR("count(structured/source) > 1"), STR("description/node()"), STR("destination"), STR("destinationAddress"), STR("domain"), STR("false()"), STR("icon/@class != 'noicon'"), + STR("icon/@image"), STR("info"), STR("info/address/line"), STR("info/distance"), STR("info/distance and near-point"), STR("info/distance and info/phone and near-point"), + STR("info/distance or info/phone"), STR("info/panel/node()"), STR("info/phone"), STR("info/references/reference[1]"), STR("info/references/reference[1]/snippet"), + STR("info/references/reference[1]/url"), STR("info/title"), STR("info/title/node()"), STR("line"), STR("location"), STR("location[@id!='near']"), STR("location[@id='near'][icon/@image]"), + STR("location[position() > umlocations div 2]"), STR("location[position() <= numlocations div 2]"), STR("locations"), STR("locations/location"), STR("near"), STR("node()"), + STR("not(count-of-snippets = 0)"), STR("not(form = 'from')"), STR("not(form = 'near')"), STR("not(form = 'to')"), STR("not(../@page)"), STR("not(structured/source)"), STR("notice"), + STR("number(../@info)"), STR("number(../@items)"), STR("number(/page/@linewidth)"), STR("page/ads"), STR("page/directions"), STR("page/error"), STR("page/overlay"), + STR("page/overlay/locations/location"), STR("page/refinements"), STR("page/request/canonicalnear"), STR("page/request/near"), STR("page/request/query"), STR("page/spelling/suggestion"), + STR("page/user/defaultlocation"), STR("phone"), STR("position()"), STR("position() != 1"), STR("position() != last()"), STR("position() > 1"), STR("position() < details"), + STR("position()-1"), STR("query"), STR("references/@total"), STR("references/reference"), STR("references/reference/domain"), STR("references/reference/url"), + STR("reviews/@positive div (reviews/@positive + reviews/@negative) * 5"), STR("reviews/@positive div (reviews/@positive + reviews/@negative) * (5)"), STR("reviews/@total"), + STR("reviews/@total > 1"), STR("reviews/@total > 5"), STR("reviews/@total = 1"), STR("segments/@distance"), STR("segments/@time"), STR("segments/segment"), STR("shorttitle/node()"), + STR("snippet"), STR("snippet/node()"), STR("source"), STR("sourceAddress"), STR("sourceAddress and destinationAddress"), STR("string(../@daddr)"), STR("string(../@form)"), + STR("string(../@page)"), STR("string(../@saddr)"), STR("string(info/title)"), STR("string(page/request/canonicalnear) != ''"), STR("string(page/request/near) != ''"), + STR("string-length(address) > linewidth"), STR("structured/@total - details"), STR("structured/source"), STR("structured/source[@name]"), STR("substring(address, 1, linewidth - 3)"), + STR("substring-after(str, c)"), STR("substring-after(icon/@image, '/mapfiles/marker')"), STR("substring-before(str, c)"), STR("tagline/node()"), STR("targetedlocation"), + STR("title"), STR("title/node()"), STR("true()"), STR("url"), STR("visibleurl"), STR("id(\"level10\")/ancestor::SPAN"), STR("id(\"level10\")/ancestor-or-self::SPAN"), STR("//attribute::*"), + STR("child::HTML/child::BODY/child::H1"), STR("descendant::node()"), STR("descendant-or-self::SPAN"), STR("id(\"first\")/following::text()"), STR("id(\"first\")/following-sibling::node()"), + STR("id(\"level10\")/parent::node()"), STR("id(\"last\")/preceding::text()"), STR("id(\"last\")/preceding-sibling::node()"), STR("/HTML/BODY/H1/self::node()"), STR("//*[@name]"), + STR("id(\"pet\")/SELECT[@name=\"species\"]/OPTION[@selected]/@value"), STR("descendant::INPUT[@name=\"name\"]/@value"), STR("id(\"pet\")/INPUT[@name=\"gender\" and @checked]/@value"), + STR("//TEXTAREA[@name=\"description\"]/text()"), STR("id(\"div1\")|id(\"div2\")|id(\"div3 div4 div5\")"), STR("//LI[1]"), STR("//LI[last()]/text()"), STR("//LI[position() mod 2]/@class"), + STR("//text()[.=\"foo\"]"), STR("descendant-or-self::SPAN[position() > 2]"), STR("descendant::*[contains(@class,\" fruit \")]"), - // ajaxslt considers this path invalid, however I believe it's valid as per spec - STR("***"), + // ajaxslt considers this path invalid, however I believe it's valid as per spec + STR("***"), - // Oasis MSFT considers this path invalid, however I believe it's valid as per spec - STR("**..**"), + // Oasis MSFT considers this path invalid, however I believe it's valid as per spec + STR("**..**"), - // Miscellaneous - STR("..***..***.***.***..***..***..") - }; + // Miscellaneous + STR("..***..***.***.***..***..***..")}; for (size_t i = 0; i < sizeof(paths) / sizeof(paths[0]); ++i) { @@ -185,73 +183,125 @@ TEST(xpath_parse_paths_valid) #if defined(PUGIXML_WCHAR_MODE) || !defined(PUGIXML_NO_STL) TEST(xpath_parse_paths_valid_unicode) { - // From ajaxslt + // From ajaxslt const wchar_t* paths[] = - { - #ifdef U_LITERALS - L"/descendant-or-self::\u90e8\u5206", L"//\u90e8\u5206", L"substring('\uff11\uff12\uff13\uff14\uff15', 0, 3)", L"//\u30bf\u30a4\u30c8\u30eb | //\u30ea\u30f3\u30af", - L"\u8b0e//\u30bf\u30a4\u30c8\u30eb", L"//*[@\u30c7\u30b9\u30c6\u30a3\u30cd\u30a4\u30b7\u30e7\u30f3]", L"\u30da\u30fc\u30b8 = '\u304b\u3089'", - L"concat(substring-before(@\u30a4\u30e1\u30fc\u30b8,'\u76ee\u5370'),'\u30a2\u30a4\u30b3\u30f3',substring-after(@\u30a4\u30e1\u30fc\u30b8,'\u76ee\u5370'))", - L"\u30bd\u30fc\u30b9|\u30c7\u30b9\u30c6\u30a3\u30cd\u30a4\u30b7\u30e7\u30f3", L"\u30da\u30fc\u30b8 != '\u307e\u3067' and \u30da\u30fc\u30b8 != '\u304b\u3089'", - L"substring-after(\u30a2\u30a4\u30b3\u30f3/@\u30a4\u30e1\u30fc\u30b8, '/\u5730\u56f3\u30d5\u30a1\u30a4\u30eb/\u76ee\u5370')", L"child::\u6bb5\u843d", - L"substring-before(\u6587\u5b57\u5217, \u6587\u5b57)", L"\u30bb\u30b0\u30e1\u30f3\u30c8/@\u6642\u523b", L"attribute::\u540d\u524d", L"descendant::\u6bb5\u843d", - L"ancestor::\u90e8\u5206", L"ancestor-or-self::\u90e8\u5206", L"descendant-or-self::\u6bb5\u843d", L"self::\u6bb5\u843d", L"child::\u7ae0/descendant::\u6bb5\u843d", - L"child::*/child::\u6bb5\u843d", L"/descendant::\u6bb5\u843d", L"/descendant::\u9806\u5e8f\u30ea\u30b9\u30c8/child::\u9805\u76ee", L"child::\u6bb5\u843d[position()=1]", - L"child::\u6bb5\u843d[position()=last()]", L"child::\u6bb5\u843d[position()=last()-1]", L"child::\u6bb5\u843d[position()>1]", L"following-sibling::\u7ae0[position()=1]", - L"preceding-sibling::\u7ae0[position()=1]", L"/descendant::\u56f3\u8868[position()=42]", L"/child::\u6587\u66f8/child::\u7ae0[position()=5]/child::\u7bc0[position()=2]", - L"child::\u6bb5\u843d[attribute::\u30bf\u30a4\u30d7='\u8b66\u544a']", L"child::\u6bb5\u843d[attribute::\u30bf\u30a4\u30d7='\u8b66\u544a'][position()=5]", - L"child::\u6bb5\u843d[position()=5][attribute::\u30bf\u30a4\u30d7='\u8b66\u544a']", L"child::\u7ae0[child::\u30bf\u30a4\u30c8\u30eb='\u306f\u3058\u3081\u306b']", - L"child::\u7ae0[child::\u30bf\u30a4\u30c8\u30eb]", L"child::*[self::\u7ae0 or self::\u4ed8\u9332]", L"child::*[self::\u7ae0 or self::\u4ed8\u9332][position()=last()]", - #else - L"/descendant-or-self::\x90e8\x5206", L"//\x90e8\x5206", L"substring('\xff11\xff12\xff13\xff14\xff15', 0, 3)", L"//\x30bf\x30a4\x30c8\x30eb | //\x30ea\x30f3\x30af", - L"\x8b0e//\x30bf\x30a4\x30c8\x30eb", L"//*[@\x30c7\x30b9\x30c6\x30a3\x30cd\x30a4\x30b7\x30e7\x30f3]", L"\x30da\x30fc\x30b8 = '\x304b\x3089'", - L"concat(substring-before(@\x30a4\x30e1\x30fc\x30b8,'\x76ee\x5370'),'\x30a2\x30a4\x30b3\x30f3',substring-after(@\x30a4\x30e1\x30fc\x30b8,'\x76ee\x5370'))", - L"\x30bd\x30fc\x30b9|\x30c7\x30b9\x30c6\x30a3\x30cd\x30a4\x30b7\x30e7\x30f3", L"\x30da\x30fc\x30b8 != '\x307e\x3067' and \x30da\x30fc\x30b8 != '\x304b\x3089'", - L"substring-after(\x30a2\x30a4\x30b3\x30f3/@\x30a4\x30e1\x30fc\x30b8, '/\x5730\x56f3\x30d5\x30a1\x30a4\x30eb/\x76ee\x5370')", L"child::\x6bb5\x843d", - L"substring-before(\x6587\x5b57\x5217, \x6587\x5b57)", L"\x30bb\x30b0\x30e1\x30f3\x30c8/@\x6642\x523b", L"attribute::\x540d\x524d", L"descendant::\x6bb5\x843d", - L"ancestor::\x90e8\x5206", L"ancestor-or-self::\x90e8\x5206", L"descendant-or-self::\x6bb5\x843d", L"self::\x6bb5\x843d", L"child::\x7ae0/descendant::\x6bb5\x843d", - L"child::*/child::\x6bb5\x843d", L"/descendant::\x6bb5\x843d", L"/descendant::\x9806\x5e8f\x30ea\x30b9\x30c8/child::\x9805\x76ee", L"child::\x6bb5\x843d[position()=1]", - L"child::\x6bb5\x843d[position()=last()]", L"child::\x6bb5\x843d[position()=last()-1]", L"child::\x6bb5\x843d[position()>1]", L"following-sibling::\x7ae0[position()=1]", - L"preceding-sibling::\x7ae0[position()=1]", L"/descendant::\x56f3\x8868[position()=42]", L"/child::\x6587\x66f8/child::\x7ae0[position()=5]/child::\x7bc0[position()=2]", - L"child::\x6bb5\x843d[attribute::\x30bf\x30a4\x30d7='\x8b66\x544a']", L"child::\x6bb5\x843d[attribute::\x30bf\x30a4\x30d7='\x8b66\x544a'][position()=5]", - L"child::\x6bb5\x843d[position()=5][attribute::\x30bf\x30a4\x30d7='\x8b66\x544a']", L"child::\x7ae0[child::\x30bf\x30a4\x30c8\x30eb='\x306f\x3058\x3081\x306b']", - L"child::\x7ae0[child::\x30bf\x30a4\x30c8\x30eb]", L"child::*[self::\x7ae0 or self::\x4ed8\x9332]", L"child::*[self::\x7ae0 or self::\x4ed8\x9332][position()=last()]", - #endif - }; + { +#ifdef U_LITERALS + L"/descendant-or-self::\u90e8\u5206", + L"//\u90e8\u5206", + L"substring('\uff11\uff12\uff13\uff14\uff15', 0, 3)", + L"//\u30bf\u30a4\u30c8\u30eb | //\u30ea\u30f3\u30af", + L"\u8b0e//\u30bf\u30a4\u30c8\u30eb", + L"//*[@\u30c7\u30b9\u30c6\u30a3\u30cd\u30a4\u30b7\u30e7\u30f3]", + L"\u30da\u30fc\u30b8 = '\u304b\u3089'", + L"concat(substring-before(@\u30a4\u30e1\u30fc\u30b8,'\u76ee\u5370'),'\u30a2\u30a4\u30b3\u30f3',substring-after(@\u30a4\u30e1\u30fc\u30b8,'\u76ee\u5370'))", + L"\u30bd\u30fc\u30b9|\u30c7\u30b9\u30c6\u30a3\u30cd\u30a4\u30b7\u30e7\u30f3", + L"\u30da\u30fc\u30b8 != '\u307e\u3067' and \u30da\u30fc\u30b8 != '\u304b\u3089'", + L"substring-after(\u30a2\u30a4\u30b3\u30f3/@\u30a4\u30e1\u30fc\u30b8, '/\u5730\u56f3\u30d5\u30a1\u30a4\u30eb/\u76ee\u5370')", + L"child::\u6bb5\u843d", + L"substring-before(\u6587\u5b57\u5217, \u6587\u5b57)", + L"\u30bb\u30b0\u30e1\u30f3\u30c8/@\u6642\u523b", + L"attribute::\u540d\u524d", + L"descendant::\u6bb5\u843d", + L"ancestor::\u90e8\u5206", + L"ancestor-or-self::\u90e8\u5206", + L"descendant-or-self::\u6bb5\u843d", + L"self::\u6bb5\u843d", + L"child::\u7ae0/descendant::\u6bb5\u843d", + L"child::*/child::\u6bb5\u843d", + L"/descendant::\u6bb5\u843d", + L"/descendant::\u9806\u5e8f\u30ea\u30b9\u30c8/child::\u9805\u76ee", + L"child::\u6bb5\u843d[position()=1]", + L"child::\u6bb5\u843d[position()=last()]", + L"child::\u6bb5\u843d[position()=last()-1]", + L"child::\u6bb5\u843d[position()>1]", + L"following-sibling::\u7ae0[position()=1]", + L"preceding-sibling::\u7ae0[position()=1]", + L"/descendant::\u56f3\u8868[position()=42]", + L"/child::\u6587\u66f8/child::\u7ae0[position()=5]/child::\u7bc0[position()=2]", + L"child::\u6bb5\u843d[attribute::\u30bf\u30a4\u30d7='\u8b66\u544a']", + L"child::\u6bb5\u843d[attribute::\u30bf\u30a4\u30d7='\u8b66\u544a'][position()=5]", + L"child::\u6bb5\u843d[position()=5][attribute::\u30bf\u30a4\u30d7='\u8b66\u544a']", + L"child::\u7ae0[child::\u30bf\u30a4\u30c8\u30eb='\u306f\u3058\u3081\u306b']", + L"child::\u7ae0[child::\u30bf\u30a4\u30c8\u30eb]", + L"child::*[self::\u7ae0 or self::\u4ed8\u9332]", + L"child::*[self::\u7ae0 or self::\u4ed8\u9332][position()=last()]", +#else + L"/descendant-or-self::\x90e8\x5206", + L"//\x90e8\x5206", + L"substring('\xff11\xff12\xff13\xff14\xff15', 0, 3)", + L"//\x30bf\x30a4\x30c8\x30eb | //\x30ea\x30f3\x30af", + L"\x8b0e//\x30bf\x30a4\x30c8\x30eb", + L"//*[@\x30c7\x30b9\x30c6\x30a3\x30cd\x30a4\x30b7\x30e7\x30f3]", + L"\x30da\x30fc\x30b8 = '\x304b\x3089'", + L"concat(substring-before(@\x30a4\x30e1\x30fc\x30b8,'\x76ee\x5370'),'\x30a2\x30a4\x30b3\x30f3',substring-after(@\x30a4\x30e1\x30fc\x30b8,'\x76ee\x5370'))", + L"\x30bd\x30fc\x30b9|\x30c7\x30b9\x30c6\x30a3\x30cd\x30a4\x30b7\x30e7\x30f3", + L"\x30da\x30fc\x30b8 != '\x307e\x3067' and \x30da\x30fc\x30b8 != '\x304b\x3089'", + L"substring-after(\x30a2\x30a4\x30b3\x30f3/@\x30a4\x30e1\x30fc\x30b8, '/\x5730\x56f3\x30d5\x30a1\x30a4\x30eb/\x76ee\x5370')", + L"child::\x6bb5\x843d", + L"substring-before(\x6587\x5b57\x5217, \x6587\x5b57)", + L"\x30bb\x30b0\x30e1\x30f3\x30c8/@\x6642\x523b", + L"attribute::\x540d\x524d", + L"descendant::\x6bb5\x843d", + L"ancestor::\x90e8\x5206", + L"ancestor-or-self::\x90e8\x5206", + L"descendant-or-self::\x6bb5\x843d", + L"self::\x6bb5\x843d", + L"child::\x7ae0/descendant::\x6bb5\x843d", + L"child::*/child::\x6bb5\x843d", + L"/descendant::\x6bb5\x843d", + L"/descendant::\x9806\x5e8f\x30ea\x30b9\x30c8/child::\x9805\x76ee", + L"child::\x6bb5\x843d[position()=1]", + L"child::\x6bb5\x843d[position()=last()]", + L"child::\x6bb5\x843d[position()=last()-1]", + L"child::\x6bb5\x843d[position()>1]", + L"following-sibling::\x7ae0[position()=1]", + L"preceding-sibling::\x7ae0[position()=1]", + L"/descendant::\x56f3\x8868[position()=42]", + L"/child::\x6587\x66f8/child::\x7ae0[position()=5]/child::\x7bc0[position()=2]", + L"child::\x6bb5\x843d[attribute::\x30bf\x30a4\x30d7='\x8b66\x544a']", + L"child::\x6bb5\x843d[attribute::\x30bf\x30a4\x30d7='\x8b66\x544a'][position()=5]", + L"child::\x6bb5\x843d[position()=5][attribute::\x30bf\x30a4\x30d7='\x8b66\x544a']", + L"child::\x7ae0[child::\x30bf\x30a4\x30c8\x30eb='\x306f\x3058\x3081\x306b']", + L"child::\x7ae0[child::\x30bf\x30a4\x30c8\x30eb]", + L"child::*[self::\x7ae0 or self::\x4ed8\x9332]", + L"child::*[self::\x7ae0 or self::\x4ed8\x9332][position()=last()]", +#endif + }; for (size_t i = 0; i < sizeof(paths) / sizeof(paths[0]); ++i) { - #if defined(PUGIXML_WCHAR_MODE) +#if defined(PUGIXML_WCHAR_MODE) xpath_query q(paths[i]); - #elif !defined(PUGIXML_NO_STL) +#elif !defined(PUGIXML_NO_STL) std::basic_string path_utf8 = as_utf8(paths[i]); xpath_query q(path_utf8.c_str()); - #endif +#endif } } #endif TEST(xpath_parse_invalid) { - const char_t* paths[] = - { - // From Jaxen tests - STR("//:p"), STR("/foo/bar/"), STR("12 + (count(//author)+count(//author/attribute::*)) / 2"), STR("id()/2"), STR("+"), - STR("///triple slash"), STR("/numbers numbers"), STR("/a/b[c > d]efg"), STR("/inv/child::"), STR("/invoice/@test[abcd"), - STR("/invoice/@test[abcd > x"), STR("string-length('a"), STR("/descendant::()"), STR("(1 + 1"), STR("!false()"), - STR("$author"), STR("10 + $foo"), STR("$foo:bar"), STR("$varname[@a='1']"), STR("foo/$variable/foo"), - STR(".[1]"), STR("chyld::foo"), STR("foo/tacos()"), STR("foo/tacos()"), STR("/foo/bar[baz"), STR("//"), STR("*:foo"), - STR("/cracker/cheese[(mold > 1) and (sense/taste"), + const char_t* paths[] = + { + // From Jaxen tests + STR("//:p"), STR("/foo/bar/"), STR("12 + (count(//author)+count(//author/attribute::*)) / 2"), STR("id()/2"), STR("+"), + STR("///triple slash"), STR("/numbers numbers"), STR("/a/b[c > d]efg"), STR("/inv/child::"), STR("/invoice/@test[abcd"), + STR("/invoice/@test[abcd > x"), STR("string-length('a"), STR("/descendant::()"), STR("(1 + 1"), STR("!false()"), + STR("$author"), STR("10 + $foo"), STR("$foo:bar"), STR("$varname[@a='1']"), STR("foo/$variable/foo"), + STR(".[1]"), STR("chyld::foo"), STR("foo/tacos()"), STR("foo/tacos()"), STR("/foo/bar[baz"), STR("//"), STR("*:foo"), + STR("/cracker/cheese[(mold > 1) and (sense/taste"), - // From xpath-as3 tests - STR("a b"), STR("//self::node())"), STR("/x/y[contains(self::node())"), STR("/x/y[contains(self::node()]"), STR("///"), STR("text::a"), + // From xpath-as3 tests + STR("a b"), STR("//self::node())"), STR("/x/y[contains(self::node())"), STR("/x/y[contains(self::node()]"), STR("///"), STR("text::a"), - // From haXe-xpath tests - STR("|/gjs"), STR("+3"), STR("/html/body/p != ---'div'/a"), STR(""), STR("@"), STR("#akf"), STR(",") + // From haXe-xpath tests + STR("|/gjs"), STR("+3"), STR("/html/body/p != ---'div'/a"), STR(""), STR("@"), STR("#akf"), STR(",") - // Miscellaneous - STR("..."), STR("...."), STR("**"), STR("****"), STR("******"), STR("..***..***.***.***..***..***..*"), STR("/[1]") - }; + // Miscellaneous + STR("..."), + STR("...."), STR("**"), STR("****"), STR("******"), STR("..***..***.***.***..***..***..*"), STR("/[1]")}; for (size_t i = 0; i < sizeof(paths) / sizeof(paths[0]); ++i) { @@ -376,7 +426,7 @@ TEST(xpath_parse_oom_propagation) { std::basic_string literal(i, 'a'); std::basic_string query = STR("processing-instruction('") + literal + STR("') | ") + query_base; - + CHECK_ALLOC_FAIL(CHECK_XPATH_FAIL(query.c_str())); } } diff --git a/tests/test_xpath_paths.cpp b/tests/test_xpath_paths.cpp index 54a47ae..b751754 100644 --- a/tests/test_xpath_paths.cpp +++ b/tests/test_xpath_paths.cpp @@ -27,9 +27,9 @@ TEST_XML(xpath_paths_axes_descendant, " CHECK_XPATH_NODESET(n.child(STR("last")), STR("following-sibling:: node()")); CHECK_XPATH_NODESET(n, STR("@attr1/following-sibling:: node()")); // attributes are not siblings - CHECK_XPATH_NODESET(na, STR("following-sibling:: node()")); // attributes are not siblings + CHECK_XPATH_NODESET(na, STR("following-sibling:: node()")); // attributes are not siblings } TEST_XML(xpath_paths_axes_preceding_sibling, "") @@ -97,7 +97,7 @@ TEST_XML(xpath_paths_axes_preceding_sibling, " CHECK_XPATH_NODESET(n.child(STR("last")), STR("preceding-sibling:: node()")) % 8 % 5; // another, child CHECK_XPATH_NODESET(n, STR("@attr2/following-sibling:: node()")); // attributes are not siblings - CHECK_XPATH_NODESET(na, STR("following-sibling:: node()")); // attributes are not siblings + CHECK_XPATH_NODESET(na, STR("following-sibling:: node()")); // attributes are not siblings } TEST_XML(xpath_paths_axes_following, "") @@ -108,14 +108,14 @@ TEST_XML(xpath_paths_axes_following, "") @@ -126,14 +126,14 @@ TEST_XML(xpath_paths_axes_preceding, "") @@ -146,9 +146,9 @@ TEST_XML(xpath_paths_axes_attribute, "") @@ -189,13 +189,13 @@ TEST_XML(xpath_paths_axes_descendant_or_self, "") @@ -207,13 +207,13 @@ TEST_XML(xpath_paths_axes_ancestor_or_self, "") @@ -561,7 +561,7 @@ TEST_XML(xpath_paths_descendant_double_slash_w3c, "") { - CHECK_XPATH_NODESET(doc, STR("(node/child/subchild)[2]")) % 7; + CHECK_XPATH_NODESET(doc, STR("(node/child/subchild)[2]")) % 7; } TEST_XML(xpath_paths_descendant_filters, "") @@ -657,53 +657,53 @@ TEST_XML(xpath_paths_optimize_compare_attribute, " TEST_XML(xpath_paths_optimize_step_once, "") { - CHECK_XPATH_BOOLEAN(doc, STR("node//para2/following::*"), true); - CHECK_XPATH_BOOLEAN(doc, STR("node//para6/following::*"), false); + CHECK_XPATH_BOOLEAN(doc, STR("node//para2/following::*"), true); + CHECK_XPATH_BOOLEAN(doc, STR("node//para6/following::*"), false); - CHECK_XPATH_STRING(doc, STR("name(node//para2/following::*)"), STR("para3")); - CHECK_XPATH_STRING(doc, STR("name(node//para6/following::*)"), STR("")); + CHECK_XPATH_STRING(doc, STR("name(node//para2/following::*)"), STR("para3")); + CHECK_XPATH_STRING(doc, STR("name(node//para6/following::*)"), STR("")); - CHECK_XPATH_BOOLEAN(doc, STR("node//para1/preceding::*"), false); - CHECK_XPATH_BOOLEAN(doc, STR("node//para6/preceding::*"), true); + CHECK_XPATH_BOOLEAN(doc, STR("node//para1/preceding::*"), false); + CHECK_XPATH_BOOLEAN(doc, STR("node//para6/preceding::*"), true); - CHECK_XPATH_STRING(doc, STR("name(node//para1/preceding::*)"), STR("")); - CHECK_XPATH_STRING(doc, STR("name(node//para6/preceding::*)"), STR("para1")); + CHECK_XPATH_STRING(doc, STR("name(node//para1/preceding::*)"), STR("")); + CHECK_XPATH_STRING(doc, STR("name(node//para6/preceding::*)"), STR("para1")); - CHECK_XPATH_BOOLEAN(doc, STR("node//para6/preceding::para4"), true); + CHECK_XPATH_BOOLEAN(doc, STR("node//para6/preceding::para4"), true); - CHECK_XPATH_BOOLEAN(doc, STR("//@attr5/ancestor-or-self::*"), true); - CHECK_XPATH_BOOLEAN(doc, STR("//@attr5/ancestor::*"), true); + CHECK_XPATH_BOOLEAN(doc, STR("//@attr5/ancestor-or-self::*"), true); + CHECK_XPATH_BOOLEAN(doc, STR("//@attr5/ancestor::*"), true); - CHECK_XPATH_BOOLEAN(doc, STR("//@attr5/following::para6"), true); - CHECK_XPATH_STRING(doc, STR("name(//@attr5/following::para6)"), STR("para6")); + CHECK_XPATH_BOOLEAN(doc, STR("//@attr5/following::para6"), true); + CHECK_XPATH_STRING(doc, STR("name(//@attr5/following::para6)"), STR("para6")); - CHECK_XPATH_BOOLEAN(doc, STR("//para5/ancestor-or-self::*"), true); - CHECK_XPATH_BOOLEAN(doc, STR("//para5/ancestor::*"), true); + CHECK_XPATH_BOOLEAN(doc, STR("//para5/ancestor-or-self::*"), true); + CHECK_XPATH_BOOLEAN(doc, STR("//para5/ancestor::*"), true); - CHECK_XPATH_BOOLEAN(doc, STR("//@attr5/ancestor-or-self::node()"), true); + CHECK_XPATH_BOOLEAN(doc, STR("//@attr5/ancestor-or-self::node()"), true); } TEST_XML(xpath_paths_null_nodeset_entries, "") { - xpath_node nodes[] = - { - xpath_node(doc.first_child()), - xpath_node(xml_node()), - xpath_node(doc.first_child().first_attribute(), doc.first_child()), - xpath_node(xml_attribute(), doc.first_child()), - xpath_node(xml_attribute(), xml_node()), - }; + xpath_node nodes[] = + { + xpath_node(doc.first_child()), + xpath_node(xml_node()), + xpath_node(doc.first_child().first_attribute(), doc.first_child()), + xpath_node(xml_attribute(), doc.first_child()), + xpath_node(xml_attribute(), xml_node()), + }; - xpath_node_set ns(nodes, nodes + sizeof(nodes) / sizeof(nodes[0])); + xpath_node_set ns(nodes, nodes + sizeof(nodes) / sizeof(nodes[0])); - xpath_variable_set vars; - vars.set(STR("x"), ns); + xpath_variable_set vars; + vars.set(STR("x"), ns); - xpath_node_set rs = xpath_query(STR("$x/."), &vars).evaluate_node_set(xml_node()); + xpath_node_set rs = xpath_query(STR("$x/."), &vars).evaluate_node_set(xml_node()); - CHECK(rs.size() == 2); - CHECK(rs[0] == nodes[0]); - CHECK(rs[1] == nodes[2]); + CHECK(rs.size() == 2); + CHECK(rs[0] == nodes[0]); + CHECK(rs[1] == nodes[2]); } TEST_XML(xpath_paths_step_leaf_coverage, "") diff --git a/tests/test_xpath_variables.cpp b/tests/test_xpath_variables.cpp index 8a65e3c..830b48a 100644 --- a/tests/test_xpath_variables.cpp +++ b/tests/test_xpath_variables.cpp @@ -404,11 +404,11 @@ TEST(xpath_variables_name_case) TEST(xpath_variables_name_unicode) { #ifdef PUGIXML_WCHAR_MODE - #ifdef U_LITERALS - const char_t* name = L"\u0400\u203D"; - #else - const char_t* name = L"\x0400\x203D"; - #endif +#ifdef U_LITERALS + const char_t* name = L"\u0400\u203D"; +#else + const char_t* name = L"\x0400\x203D"; +#endif #else const char_t* name = "\xd0\x80\xe2\x80\xbd"; #endif diff --git a/tests/test_xpath_xalan_1.cpp b/tests/test_xpath_xalan_1.cpp index 0127ddd..7c6856a 100644 --- a/tests/test_xpath_xalan_1.cpp +++ b/tests/test_xpath_xalan_1.cpp @@ -322,16 +322,16 @@ TEST(xpath_xalan_math_5) CHECK_XPATH_NUMBER(c, STR("10+5+25+20+15+50+35+40"), 200); CHECK_XPATH_NUMBER(c, STR("100-9-7-4-17-18-5"), 40); CHECK_XPATH_NUMBER(c, STR("3*2+5*4-4*2-1"), 17); - CHECK_XPATH_NUMBER(c, STR("6*5-8*2+5*2"), 24); - CHECK_XPATH_NUMBER(c, STR("10*5-4*2+6*1 -3*3"), 39); + CHECK_XPATH_NUMBER(c, STR("6*5-8*2+5*2"), 24); + CHECK_XPATH_NUMBER(c, STR("10*5-4*2+6*1 -3*3"), 39); - CHECK_XPATH_NUMBER(c, STR("(24 div 3 +2) div (40 div 8 -3)"), 5); - CHECK_XPATH_NUMBER(c, STR("80 div 2 + 12 div 2 - 4 div 2"), 44); - CHECK_XPATH_NUMBER(c, STR("70 div 10 - 18 div 6 + 10 div 2"), 9); + CHECK_XPATH_NUMBER(c, STR("(24 div 3 +2) div (40 div 8 -3)"), 5); + CHECK_XPATH_NUMBER(c, STR("80 div 2 + 12 div 2 - 4 div 2"), 44); + CHECK_XPATH_NUMBER(c, STR("70 div 10 - 18 div 6 + 10 div 2"), 9); - CHECK_XPATH_NUMBER(c, STR("48 mod 17 - 2 mod 9 + 13 mod 5"), 15); - CHECK_XPATH_NUMBER(c, STR("56 mod round(5*2+1.444) - 6 mod 4 + 7 mod 4"), 2); - CHECK_XPATH_NUMBER(c, STR("(77 mod 10 + 5 mod 8) mod 10"), 2); + CHECK_XPATH_NUMBER(c, STR("48 mod 17 - 2 mod 9 + 13 mod 5"), 15); + CHECK_XPATH_NUMBER(c, STR("56 mod round(5*2+1.444) - 6 mod 4 + 7 mod 4"), 2); + CHECK_XPATH_NUMBER(c, STR("(77 mod 10 + 5 mod 8) mod 10"), 2); } TEST_XML(xpath_xalan_math_6, "37x") diff --git a/tests/test_xpath_xalan_2.cpp b/tests/test_xpath_xalan_2.cpp index 4e8c10e..60f689b 100644 --- a/tests/test_xpath_xalan_2.cpp +++ b/tests/test_xpath_xalan_2.cpp @@ -4,8 +4,8 @@ #include "test.hpp" -#include #include +#include using namespace pugi; diff --git a/tests/test_xpath_xalan_3.cpp b/tests/test_xpath_xalan_3.cpp index 5228cde..b26db1e 100644 --- a/tests/test_xpath_xalan_3.cpp +++ b/tests/test_xpath_xalan_3.cpp @@ -193,7 +193,7 @@ TEST_XML(xpath_xalan_axes_9, "
1< CHECK_XPATH_NODESET(c, STR("processing-instruction()[2]")) % 34; CHECK_XPATH_NODESET(c, STR("processing-instruction('pi')[2]")) % 34; CHECK_XPATH_NODESET(c, STR("comment()[2]")) % 36; - CHECK_XPATH_NODESET(c, STR("a/*[last()]")) % 5 % 8 % 11 % 15 % 18 % 21 % 25 % 29; - CHECK_XPATH_NODESET(c, STR("a/child::*[last()]")) % 5 % 8 % 11 % 15 % 18 % 21 % 25 % 29; - CHECK_XPATH_NODESET(c, STR("a/descendant::*[last()]")) % 5 % 8 % 11 % 15 % 18 % 21 % 25 % 31; - CHECK_XPATH_NODESET(c, STR("a/child::node()[last()]")) % 5 % 8 % 11 % 15 % 18 % 21 % 25 % 29; - CHECK_XPATH_NODESET(c, STR("a/descendant::text()[last()]")) % 6 % 9 % 12 % 16 % 19 % 22 % 26 % 32; - CHECK_XPATH_NODESET(c, STR("child::comment()[last()]")) % 36; + CHECK_XPATH_NODESET(c, STR("a/*[last()]")) % 5 % 8 % 11 % 15 % 18 % 21 % 25 % 29; + CHECK_XPATH_NODESET(c, STR("a/child::*[last()]")) % 5 % 8 % 11 % 15 % 18 % 21 % 25 % 29; + CHECK_XPATH_NODESET(c, STR("a/descendant::*[last()]")) % 5 % 8 % 11 % 15 % 18 % 21 % 25 % 31; + CHECK_XPATH_NODESET(c, STR("a/child::node()[last()]")) % 5 % 8 % 11 % 15 % 18 % 21 % 25 % 29; + CHECK_XPATH_NODESET(c, STR("a/descendant::text()[last()]")) % 6 % 9 % 12 % 16 % 19 % 22 % 26 % 32; + CHECK_XPATH_NODESET(c, STR("child::comment()[last()]")) % 36; } TEST_XML(xpath_xalan_position_3, "
AAA
BBBAbout this article
CCCThis is the section titled 'ZZZ'.DDDDon't worry.
EEEThis is the deep subsection.
") @@ -111,7 +111,6 @@ TEST_XML(xpath_xalan_match_1, "< CHECK_XPATH_NODESET(c, STR("x[(((((2*10)-4)+9) div 5) mod 3)]")) % 6; } - TEST_XML(xpath_xalan_match_2, "doc-l1-v2doc-l1-x2doc-l1-l2-v3doc-l1-l2-w3doc-l1-l2-x3doc-l1-l2-y3doc-l1-l2-l3-v4doc-l1-l2-l3-x4") { CHECK_XPATH_STRING(doc, STR("doc/l1/v2"), STR("doc-l1-v2")); diff --git a/tests/test_xpath_xalan_5.cpp b/tests/test_xpath_xalan_5.cpp index 6daa234..9e35a56 100644 --- a/tests/test_xpath_xalan_5.cpp +++ b/tests/test_xpath_xalan_5.cpp @@ -111,7 +111,7 @@ TEST_XML(xpath_xalan_select_11, "descendant number 1< TEST_XML(xpath_xalan_select_12, "child number 1grandchild number 1child number 2grandchild number 2child number 3grandchild number 3child number 4grandchild number 4") { - CHECK_XPATH_NODESET(doc, STR("//child/ancestor-or-self::sub | //child/ancestor-or-self::sub-sub")) % 3 % 7 % 15 % 19 % 31; + CHECK_XPATH_NODESET(doc, STR("//child/ancestor-or-self::sub | //child/ancestor-or-self::sub-sub")) % 3 % 7 % 15 % 19 % 31; } TEST_XML(xpath_xalan_select_13, "Carmelo MontanezNineDavid MarstonSevenMary BradyTenLynne RosenthalFive") @@ -154,17 +154,17 @@ TEST_XML(xpath_xalan_select_17, "HelloThereWorld") { - CHECK_XPATH_NODESET(doc, STR("/para/font[@color='green']")) % 6; - CHECK_XPATH_NODESET(doc.child(STR("para")), STR("/para/font[@color='green']")) % 6; - CHECK_XPATH_NODESET(doc.child(STR("para")).last_child(), STR("/para/font[@color='green']")) % 6; + CHECK_XPATH_NODESET(doc, STR("/para/font[@color='green']")) % 6; + CHECK_XPATH_NODESET(doc.child(STR("para")), STR("/para/font[@color='green']")) % 6; + CHECK_XPATH_NODESET(doc.child(STR("para")).last_child(), STR("/para/font[@color='green']")) % 6; } TEST_XML_FLAGS(xpath_xalan_select_19, "1in-a2345678in-c9", parse_default | parse_comments | parse_pi) diff --git a/tests/writer_string.cpp b/tests/writer_string.cpp index 26bca8d..0b215fd 100644 --- a/tests/writer_string.cpp +++ b/tests/writer_string.cpp @@ -5,10 +5,12 @@ static bool test_narrow(const std::string& result, const char* expected, size_t length) { // check result - if (result != std::string(expected, expected + length)) return false; + if (result != std::string(expected, expected + length)) + return false; // check comparison operator (incorrect implementation can theoretically early-out on zero terminators...) - if (length > 0 && result == std::string(expected, expected + length - 1) + "?") return false; + if (length > 0 && result == std::string(expected, expected + length - 1) + "?") + return false; return true; } @@ -27,7 +29,7 @@ std::basic_string xml_writer_string::as_wide() const { CHECK(contents.size() % sizeof(wchar_t) == 0); - // round-trip pointer through void* to avoid pointer alignment warnings; contents data should be heap allocated => safe to cast + // round-trip pointer through void* to avoid pointer alignment warnings; contents data should be heap allocated => safe to cast return std::basic_string(static_cast(static_cast(contents.data())), contents.size() / sizeof(wchar_t)); } @@ -37,7 +39,7 @@ std::basic_string xml_writer_string::as_string() const CHECK(contents.size() % sizeof(pugi::char_t) == 0); #endif - // round-trip pointer through void* to avoid pointer alignment warnings; contents data should be heap allocated => safe to cast + // round-trip pointer through void* to avoid pointer alignment warnings; contents data should be heap allocated => safe to cast return std::basic_string(static_cast(static_cast(contents.data())), contents.size() / sizeof(pugi::char_t)); } diff --git a/tests/writer_string.hpp b/tests/writer_string.hpp index 539377c..3562c20 100644 --- a/tests/writer_string.hpp +++ b/tests/writer_string.hpp @@ -5,7 +5,7 @@ #include -struct xml_writer_string: public pugi::xml_writer +struct xml_writer_string : public pugi::xml_writer { std::string contents;