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

Parsing flags refactoring (removed trim flags, eol flags merged together, escapes flags merged together, removed wnorm_pcdata flag, changed wnorm_attribute flag (it's space normalization + trimming now), fixed default flags, changed documentation accordingly

Browse Source 
git-svn-id: http://pugixml.googlecode.com/svn/trunk@26 99668b35-9821-0410-8761-19e4c4f06640
This commit is contained in:
arseny.kapoulkine 2007-01-05 20:05:10 +00:00
parent 784235c5e6
commit 2777da9faa
3 changed files with 50 additions and 142 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

39
docs/index.html
View File

@ -275,30 +275,13 @@ So, these are the processing flags:
</p> </p>
<ul> <ul>
<li>If <b>parse_trim_pcdata</b> is on, then the trimming of leading/trailing space-like characters <li>If <b>parse_escapes</b> is on, then the character reference expansion is done for PCDATA content
is performed for PCDATA content and for attribute values (replacing &lt;lt; with &lt;, &amp;#4c; with L, etc.).
<br>Default value: on
<br>In W3C mode: off</li>
<li>If <b>parse_trim_attribute</b> is on, then the trimming of leading/trailing space-like characters
is performed for attribute values (this is non-standard behavior and is here only for compatibility
reasons (PugXML had this flag).
<br>Default value: off
<br>In W3C mode: off</li>
<li>If <b>parse_escapes_pcdata</b> is on, then the character reference expansion is done for PCDATA
content (replacing &lt;lt; with &lt;, &amp;#4c; with L, etc.).
<br>Default value: on <br>Default value: on
<br>In W3C mode: on</li> <br>In W3C mode: on</li>
<li>If <b>parse_escapes_attribute</b> is on, then the character reference expansion is done for
attribute values (replacing &lt;lt; with &lt;, &amp;#4c; with L, etc.).
<br>Default value: on
<br>In W3C mode: on</li>
<li>If <b>parse_wnorm_pcdata</b> is on, then the whitespace normalisation is done for PCDATA content
(this includes replacing any space-like character by a space character and converting sequences of
spaces into a single space)
<br>Default value: on
<br>In W3C mode: off</li>
<li>If <b>parse_wnorm_attribute</b> is on, then the whitespace normalisation is done for attribute <li>If <b>parse_wnorm_attribute</b> is on, then the whitespace normalisation is done for attribute
values values (this includes replacing any space-like character by a space character, converting sequences of
spaces into a single space and trimming of leading/trailing spaces)
<br>Default value: on <br>Default value: on
<br>In W3C mode: off</li> <br>In W3C mode: off</li>
<li>If <b>parse_wconv_attribute</b> is on, then the whitespace conversion is done for attribute <li>If <b>parse_wconv_attribute</b> is on, then the whitespace conversion is done for attribute
@ -306,15 +289,9 @@ values (this is a subset of whitespace normalization, and includes only replacin
with spaces). If <b>parse_wnorm_attribute</b> is on, this flag has no effect. with spaces). If <b>parse_wnorm_attribute</b> is on, this flag has no effect.
<br>Default value: on <br>Default value: on
<br>In W3C mode: on</li> <br>In W3C mode: on</li>
<li>If <b>parse_eol_pcdata</b> is on, then the end-of-line handling is done for PCDATA content (this <li>If <b>parse_eol</b> is on, then the end-of-line handling is done for PCDATA/CDATA content and for
includes converting any pair of 0x0d 0x0a characters to a single 0x0a and converting any standalone attribute values (this includes converting any pair of 0x0d 0x0a characters to a single 0x0a and
0x0d to 0x0a). converting any standalone 0x0d to 0x0a).
<br>Default value: on
<br>In W3C mode: on</li>
<li>If <b>parse_eol_attribute</b> is on, then the end-of-line handling is done for attribute values.
<br>Default value: on
<br>In W3C mode: on</li>
<li>If <b>parse_eol_cdata</b> is on, then the end-of-line handling is done for CDATA content.
<br>Default value: on <br>Default value: on
<br>In W3C mode: on</li> <br>In W3C mode: on</li>
</ul> </ul>
@ -329,7 +306,7 @@ correctly). This is controlled by <b>parse_match_end_tags</b>, which is on by de
<li>just treat the tag as a closing tag for the node (so that <b>&lt;foo&gt; ... &lt;/bar&gt;</b> will <li>just treat the tag as a closing tag for the node (so that <b>&lt;foo&gt; ... &lt;/bar&gt;</b> will
be parsed as <b>&lt;foo&gt; ... &lt;/foo&gt;</b>). This is the fastest way, and this is what <i>pugxml</i> be parsed as <b>&lt;foo&gt; ... &lt;/foo&gt;</b>). This is the fastest way, and this is what <i>pugxml</i>
is doing, but it can corrupt your DOM tree. This way is chosen if both <b>parse_check_end_tags</b> and is doing, but it can corrupt your DOM tree. This way is chosen if both <b>parse_check_end_tags</b> and
<b>parsse_match_end_tags</b> are off. <b>parse_match_end_tags</b> are off.
</ul> </ul>
Note, that these 2 flags are mutually exclusive. Note, that these 2 flags are mutually exclusive.
</p> </p>

124
src/pugixml.cpp
View File

@ -252,14 +252,14 @@ namespace pugi
static bool chartype_lbracket(char c) { return c == '['; } static bool chartype_lbracket(char c) { return c == '['; }
static bool chartype_rbracket(char c) { return c == ']'; } static bool chartype_rbracket(char c) { return c == ']'; }
template <bool opt_trim, bool opt_escape, bool opt_wnorm, bool opt_wconv, bool opt_eol> static void strconv_t(char** s) template <bool opt_escape, bool opt_wnorm, bool opt_wconv, bool opt_eol> static void strconv_t(char** s)
{ {
if (!s || !*s) return; if (!s || !*s) return;
if (!opt_trim && !opt_escape && !opt_wnorm && !opt_wconv && !opt_eol) return; if (!opt_escape && !opt_wnorm && !opt_wconv && !opt_eol) return;
// Trim whitespaces // Trim whitespaces
if (opt_trim) while (chartype_space(**s)) ++(*s); if (opt_wnorm) while (chartype_space(**s)) ++(*s);
char* str = *s; char* str = *s;
@ -270,6 +270,7 @@ namespace pugi
{ {
if (opt_escape && *str == '&') break; if (opt_escape && *str == '&') break;
if ((opt_wnorm || opt_wconv || opt_eol) && chartype_space(*str)) break; if ((opt_wnorm || opt_wconv || opt_eol) && chartype_space(*str)) break;
++str; ++str;
} }
} }
@ -406,7 +407,7 @@ namespace pugi
*lastpos++ = *str++; *lastpos++ = *str++;
} }
if (opt_trim) if (opt_wnorm)
{ {
do *lastpos-- = 0; do *lastpos-- = 0;
while (chartype_space(*lastpos)); while (chartype_space(*lastpos));
@ -414,66 +415,34 @@ namespace pugi
else *lastpos = 0; else *lastpos = 0;
} }
static void strconv_setup(void (*&func)(char**), unsigned int opt_trim, unsigned int opt_escape, unsigned int opt_wnorm, unsigned int opt_wconv, unsigned int opt_eol) static void strconv_setup(void (*&func)(char**), unsigned int opt_escape, unsigned int opt_wnorm, unsigned int opt_wconv, unsigned int opt_eol)
{ {
if (opt_eol) if (opt_eol)
{ {
if (opt_wconv) if (opt_wconv)
{ {
if (opt_trim) if (opt_escape)
{ {
if (opt_escape) if (opt_wnorm) func = &strconv_t<true, true, true, true>;
{ else func = &strconv_t<true, false, true, true>;
if (opt_wnorm) func = &strconv_t<true, true, true, true, true>;
else func = &strconv_t<true, true, false, true, true>;
}
else
{
if (opt_wnorm) func = &strconv_t<true, false, true, true, true>;
else func = &strconv_t<true, false, false, true, true>;
}
} }
else else
{ {
if (opt_escape) if (opt_wnorm) func = &strconv_t<false, true, true, true>;
{ else func = &strconv_t<false, false, true, true>;
if (opt_wnorm) func = &strconv_t<false, true, true, true, true>;
else func = &strconv_t<false, true, false, true, true>;
}
else
{
if (opt_wnorm) func = &strconv_t<false, false, true, true, true>;
else func = &strconv_t<false, false, false, true, true>;
}
} }
} }
else else
{ {
if (opt_trim) if (opt_escape)
{ {
if (opt_escape) if (opt_wnorm) func = &strconv_t<true, true, false, true>;
{ else func = &strconv_t<true, false, false, true>;
if (opt_wnorm) func = &strconv_t<true, true, true, false, true>;
else func = &strconv_t<true, true, false, false, true>;
}
else
{
if (opt_wnorm) func = &strconv_t<true, false, true, false, true>;
else func = &strconv_t<true, false, false, false, true>;
}
} }
else else
{ {
if (opt_escape) if (opt_wnorm) func = &strconv_t<false, true, false, true>;
{ else func = &strconv_t<false, false, false, true>;
if (opt_wnorm) func = &strconv_t<false, true, true, false, true>;
else func = &strconv_t<false, true, false, false, true>;
}
else
{
if (opt_wnorm) func = &strconv_t<false, false, true, false, true>;
else func = &strconv_t<false, false, false, false, true>;
}
} }
} }
} }
@ -481,64 +450,33 @@ namespace pugi
{ {
if (opt_wconv) if (opt_wconv)
{ {
if (opt_trim) if (opt_escape)
{ {
if (opt_escape) if (opt_wnorm) func = &strconv_t<true, true, true, false>;
{ else func = &strconv_t<true, false, true, false>;
if (opt_wnorm) func = &strconv_t<true, true, true, true, false>;
else func = &strconv_t<true, true, false, true, false>;
}
else
{
if (opt_wnorm) func = &strconv_t<true, false, true, true, false>;
else func = &strconv_t<true, false, false, true, false>;
}
} }
else else
{ {
if (opt_escape) if (opt_wnorm) func = &strconv_t<false, true, true, false>;
{ else func = &strconv_t<false, false, true, false>;
if (opt_wnorm) func = &strconv_t<false, true, true, true, false>;
else func = &strconv_t<false, true, false, true, false>;
}
else
{
if (opt_wnorm) func = &strconv_t<false, false, true, true, false>;
else func = &strconv_t<false, false, false, true, false>;
}
} }
} }
else else
{ {
if (opt_trim) if (opt_escape)
{ {
if (opt_escape) if (opt_wnorm) func = &strconv_t<true, true, false, false>;
{ else func = &strconv_t<true, false, false, false>;
if (opt_wnorm) func = &strconv_t<true, true, true, false, false>;
else func = &strconv_t<true, true, false, false, false>;
}
else
{
if (opt_wnorm) func = &strconv_t<true, false, true, false, false>;
else func = &strconv_t<true, false, false, false, false>;
}
} }
else else
{ {
if (opt_escape) if (opt_wnorm) func = &strconv_t<false, true, false, false>;
{ else func = &strconv_t<false, false, false, false>;
if (opt_wnorm) func = &strconv_t<false, true, true, false, false>;
else func = &strconv_t<false, true, false, false, false>;
}
else
{
if (opt_wnorm) func = &strconv_t<false, false, true, false, false>;
else func = &strconv_t<false, false, false, false, false>;
}
} }
} }
} }
} }
// Allocate & append a new xml_node_struct onto the given parent. // Allocate & append a new xml_node_struct onto the given parent.
// \param parent - pointer to parent node. // \param parent - pointer to parent node.
// \param type - desired node type. // \param type - desired node type.
@ -608,8 +546,8 @@ namespace pugi
void (*strconv_pcdata)(char**); void (*strconv_pcdata)(char**);
void (*strconv_attribute)(char**); void (*strconv_attribute)(char**);
strconv_setup(strconv_attribute, OPTSET(parse_trim_attribute), OPTSET(parse_escapes_attribute), OPTSET(parse_wnorm_attribute), OPTSET(parse_wconv_attribute), OPTSET(parse_eol_attribute)); strconv_setup(strconv_attribute, OPTSET(parse_escapes), OPTSET(parse_wnorm_attribute), OPTSET(parse_wconv_attribute), OPTSET(parse_eol));
strconv_setup(strconv_pcdata, OPTSET(parse_trim_pcdata), OPTSET(parse_escapes_pcdata), OPTSET(parse_wnorm_pcdata), false, OPTSET(parse_eol_pcdata)); strconv_setup(strconv_pcdata, OPTSET(parse_escapes), false, false, OPTSET(parse_eol));
char ch = 0; // Current char, in cases where we must null-terminate before we test. char ch = 0; // Current char, in cases where we must null-terminate before we test.
xml_node_struct* cursor = xmldoc; // Tree node cursor. xml_node_struct* cursor = xmldoc; // Tree node cursor.
@ -702,9 +640,9 @@ namespace pugi
SCANFOR(chartype_rbracket(*s) && chartype_rbracket(*(s+1)) && chartype_leave(*(s+2))); SCANFOR(chartype_rbracket(*s) && chartype_rbracket(*(s+1)) && chartype_leave(*(s+2)));
ENDSEG(); // Zero-terminate this segment. ENDSEG(); // Zero-terminate this segment.
if (OPTSET(parse_eol_cdata)) if (OPTSET(parse_eol))
{ {
strconv_t<false, false, false, false, true>(&cursor->value); strconv_t<false, false, false, true>(&cursor->value);
} }
POPNODE(); // Pop since this is a standalone. POPNODE(); // Pop since this is a standalone.

29
src/pugixml.hpp
View File

@ -48,28 +48,21 @@ namespace pugi
const unsigned int parse_pi = 0x00000001; ///< Parse '<?...?>' const unsigned int parse_pi = 0x00000001; ///< Parse '<?...?>'
const unsigned int parse_comments = 0x00000002; ///< Parse '<!--...-->' const unsigned int parse_comments = 0x00000002; ///< Parse '<!--...-->'
const unsigned int parse_cdata = 0x00000004; ///< Parse '<![CDATA[...]]>' const unsigned int parse_cdata = 0x00000004; ///< Parse '<![CDATA[...]]>'
const unsigned int parse_ws_pcdata = 0x00000008; ///< Skip PCDATA that consists only of whitespaces const unsigned int parse_ws_pcdata = 0x00000008; ///< Do not skip PCDATA that consists only of whitespaces
const unsigned int parse_ext_pcdata = 0x00000010; ///< Skip PCDATA that is outside all tags (i.e. root) const unsigned int parse_ext_pcdata = 0x00000010; ///< Do not skip PCDATA that is outside all tags (i.e. root)
const unsigned int parse_trim_pcdata = 0x00000020; ///< Trim '>...<' const unsigned int parse_escapes = 0x00000020; ///< Parse &lt;, &gt;, &amp;, &quot;, &apos;, &#.. sequences
const unsigned int parse_trim_attribute = 0x00000040; ///< Trim 'foo="..."'. const unsigned int parse_wnorm_attribute = 0x00000080; ///< Normalize spaces in attributes (convert space-like characters to spaces + merge adjacent spaces + trim leading/trailing spaces)
const unsigned int parse_escapes_pcdata = 0x00000080; ///< Parse &lt;, &gt;, &amp;, &quot;, &apos;, &#.. sequences const unsigned int parse_wconv_attribute = 0x00000100; ///< Convert space-like characters to spaces in attributes (only if wnorm is not set)
const unsigned int parse_escapes_attribute = 0x00000100; ///< Parse &lt;, &gt;, &amp;, &quot;, &apos;, &#.. sequences const unsigned int parse_eol = 0x00000200; ///< Perform EOL handling
const unsigned int parse_wnorm_pcdata = 0x00000200; ///< Normalize spaces in pcdata const unsigned int parse_check_end_tags = 0x00000400; ///< Check start and end tag names and return error if names mismatch
const unsigned int parse_wnorm_attribute = 0x00000400; ///< Normalize spaces in attributes const unsigned int parse_match_end_tags = 0x00000800; ///< Try to find corresponding start tag for an end tag
const unsigned int parse_wconv_attribute = 0x00000800; ///< Convert space-like characters to spaces in attributes (only if wnorm is not set)
const unsigned int parse_eol_pcdata = 0x00001000; ///< Perform EOL handling in pcdata
const unsigned int parse_eol_attribute = 0x00002000; ///< Perform EOL handling in attrobites
const unsigned int parse_eol_cdata = 0x00004000; ///< Perform EOL handling in CDATA sections
const unsigned int parse_check_end_tags = 0x00010000; ///< Check start and end tag names and return error if names mismatch
const unsigned int parse_match_end_tags = 0x00020000; ///< Try to find corresponding start tag for an end tag
///< Set all flags, except parse_ws_pcdata, parse_trim_attribute, parse_pi and parse_comments ///< Set all flags, except parse_ws_pcdata, parse_trim_attribute, parse_pi and parse_comments
const unsigned int parse_default = 0x00FFFFFF & ~parse_ws_pcdata & ~parse_trim_attribute & ~parse_pi & ~parse_comments; const unsigned int parse_default = parse_cdata | parse_ext_pcdata | parse_escapes | parse_wconv_attribute | parse_eol | parse_check_end_tags;
const unsigned int parse_noset = 0x80000000; ///< Parse with flags in xml_parser const unsigned int parse_noset = 0x80000000; ///< Parse with flags in xml_parser
const unsigned int parse_w3c = parse_pi | parse_comments | parse_cdata | const unsigned int parse_w3c = parse_pi | parse_comments | parse_cdata |
parse_escapes_pcdata | parse_escapes_attribute | parse_escapes | parse_wconv_attribute |
parse_wconv_attribute | parse_check_end_tags | parse_check_end_tags | parse_ws_pcdata | parse_eol;
parse_ws_pcdata | parse_eol_cdata;
/// Forward declarations /// Forward declarations
struct xml_attribute_struct; struct xml_attribute_struct;