External/pugixml
8
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Add parse_trim_pcdata parse option.

Browse Source 
git-svn-id: https://pugixml.googlecode.com/svn/trunk@987 99668b35-9821-0410-8761-19e4c4f06640
This commit is contained in:
Arseny Kapoulkine 2014-02-25 03:41:54 +00:00
parent cbd8131d09
commit 0a747e6c1a
2 changed files with 36 additions and 14 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

45
src/pugixml.cpp
View File

@ -1875,19 +1875,27 @@ PUGI__NS_BEGIN
typedef char_t* (*strconv_pcdata_t)(char_t*); typedef char_t* (*strconv_pcdata_t)(char_t*);
template <typename opt_eol, typename opt_escape> struct strconv_pcdata_impl template <typename opt_trim, typename opt_eol, typename opt_escape> struct strconv_pcdata_impl
{ {
static char_t* parse(char_t* s) static char_t* parse(char_t* s)
{ {
gap g; gap g;
char_t* begin = s;
while (true) while (true)
{ {
while (!PUGI__IS_CHARTYPE(*s, ct_parse_pcdata)) ++s; while (!PUGI__IS_CHARTYPE(*s, ct_parse_pcdata)) ++s;
if (*s == '<') // PCDATA ends here if (*s == '<') // PCDATA ends here
{ {
*g.flush(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 + 1; return s + 1;
} }
@ -1903,8 +1911,14 @@ PUGI__NS_BEGIN
} }
else if (*s == 0) else if (*s == 0)
{ {
*g.flush(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; return s;
} }
else ++s; else ++s;
@ -1914,14 +1928,18 @@ PUGI__NS_BEGIN
PUGI__FN strconv_pcdata_t get_strconv_pcdata(unsigned int optmask) PUGI__FN strconv_pcdata_t get_strconv_pcdata(unsigned int optmask)
{ {
PUGI__STATIC_ASSERT(parse_escapes == 0x10 && parse_eol == 0x20); PUGI__STATIC_ASSERT(parse_escapes == 0x10 && parse_eol == 0x20 && parse_trim_pcdata == 0x0800);
switch ((optmask >> 4) & 3) // get bitmask for flags (eol escapes) switch (((optmask >> 4) & 3) | ((optmask >> 9) & 4)) // get bitmask for flags (eol escapes trim)
{ {
case 0: return strconv_pcdata_impl<opt_false, opt_false>::parse; case 0: return strconv_pcdata_impl<opt_false, opt_false, opt_false>::parse;
case 1: return strconv_pcdata_impl<opt_false, opt_true>::parse; case 1: return strconv_pcdata_impl<opt_false, opt_false, opt_true>::parse;
case 2: return strconv_pcdata_impl<opt_true, opt_false>::parse; case 2: return strconv_pcdata_impl<opt_false, opt_true, opt_false>::parse;
case 3: return strconv_pcdata_impl<opt_true, opt_true>::parse; case 3: return strconv_pcdata_impl<opt_false, opt_true, opt_true>::parse;
case 4: return strconv_pcdata_impl<opt_true, opt_false, opt_false>::parse;
case 5: return strconv_pcdata_impl<opt_true, opt_false, opt_true>::parse;
case 6: return strconv_pcdata_impl<opt_true, opt_true, opt_false>::parse;
case 7: return strconv_pcdata_impl<opt_true, opt_true, opt_true>::parse;
default: assert(false); return 0; // should not get here default: assert(false); return 0; // should not get here
} }
} }
@ -2636,7 +2654,7 @@ PUGI__NS_BEGIN
// We skipped some whitespace characters because otherwise we would take the tag branch instead of PCDATA one // We skipped some whitespace characters because otherwise we would take the tag branch instead of PCDATA one
assert(mark != s); assert(mark != s);
if (!PUGI__OPTSET(parse_ws_pcdata | parse_ws_pcdata_single)) if (!PUGI__OPTSET(parse_ws_pcdata | parse_ws_pcdata_single) || PUGI__OPTSET(parse_trim_pcdata))
{ {
continue; continue;
} }
@ -2646,7 +2664,8 @@ PUGI__NS_BEGIN
} }
} }
s = mark; if (!PUGI__OPTSET(parse_trim_pcdata))
s = mark;
if (cursor->parent || PUGI__OPTSET(parse_fragment)) if (cursor->parent || PUGI__OPTSET(parse_fragment))
{ {

5
src/pugixml.hpp
View File

@ -151,9 +151,12 @@ namespace pugi
// This flag is off by default; turning it on may result in slower parsing and more memory consumption. // 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; 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 that does not have a parent node is added to the DOM tree, and if an empty document // 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. // is a valid document. This flag is off by default.
const unsigned int parse_fragment = 0x0800; const unsigned int parse_fragment = 0x1000;
// The default parsing mode. // The default parsing mode.
// Elements, PCDATA and CDATA sections are added to the DOM tree, character/reference entities are expanded, // Elements, PCDATA and CDATA sections are added to the DOM tree, character/reference entities are expanded,