lib/libxml2
1
0
Fork 0
mirror of https://gitlab.gnome.org/GNOME/libxml2.git synced 2025-07-29 11:41:22 +03:00
Code Activity

added a parser XML_PARSE_COMPACT option to allocate small text nodes (less

Browse Source 
* HTMLparser.c parser.c SAX2.c debugXML.c tree.c valid.c xmlreader.c
  xmllint.c include/libxml/HTMLparser.h include/libxml/parser.h:
  added a parser XML_PARSE_COMPACT option to allocate small
  text nodes (less than 8 bytes on 32bits, less than 16bytes on 64bits)
  directly within the node, various changes to cope with this.
* result/XPath/tests/* result/XPath/xptr/* result/xmlid/*: this
  slightly change the output
Daniel
This commit is contained in:
Daniel Veillard
2005-08-25 13:19:21 +00:00
parent 3854c57a99
commit 8874b94cd2
27 changed files with 176 additions and 117 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
ChangeLog
View File

@ -1,3 +1,13 @@
Thu Aug 25 15:14:56 CEST 2005 Daniel Veillard <daniel@veillard.com>
* HTMLparser.c parser.c SAX2.c debugXML.c tree.c valid.c xmlreader.c
xmllint.c include/libxml/HTMLparser.h include/libxml/parser.h:
added a parser XML_PARSE_COMPACT option to allocate small
text nodes (less than 8 bytes on 32bits, less than 16bytes on 64bits)
directly within the node, various changes to cope with this.
* result/XPath/tests/* result/XPath/xptr/* result/xmlid/*: this
slightly change the output
Thu Aug 25 12:16:26 CEST 2005 Daniel Veillard <daniel@veillard.com> Thu Aug 25 12:16:26 CEST 2005 Daniel Veillard <daniel@veillard.com>
* configure.in: patch from Andrew W. Nosenko, use se $GCC = 'yes' * configure.in: patch from Andrew W. Nosenko, use se $GCC = 'yes'

4
HTMLparser.c
View File

@ -5820,6 +5820,10 @@ htmlCtxtUseOptions(htmlParserCtxtPtr ctxt, int options)
ctxt->recovery = 1; ctxt->recovery = 1;
} else } else
ctxt->recovery = 0; ctxt->recovery = 0;
if (options & HTML_PARSE_COMPACT) {
ctxt->options |= HTML_PARSE_COMPACT;
options -= HTML_PARSE_COMPACT;
}
ctxt->dictNames = 0; ctxt->dictNames = 0;
return (options); return (options);
} }

18
SAX2.c
View File

@ -1777,6 +1777,7 @@ xmlSAX2TextNode(xmlParserCtxtPtr ctxt, const xmlChar *str, int len) {
xmlErrMemory(ctxt, "xmlSAX2Characters"); xmlErrMemory(ctxt, "xmlSAX2Characters");
return(NULL); return(NULL);
} }
memset(ret, 0, sizeof(xmlNode));
/* /*
* intern the formatting blanks found between tags, or the * intern the formatting blanks found between tags, or the
* very short strings * very short strings
@ -1784,7 +1785,14 @@ xmlSAX2TextNode(xmlParserCtxtPtr ctxt, const xmlChar *str, int len) {
if (ctxt->dictNames) { if (ctxt->dictNames) {
xmlChar cur = str[len]; xmlChar cur = str[len];
if ((len <= 3) && ((cur == '"') || (cur == '\'') || if ((len < (int) (2 * sizeof(void *))) &&
(ctxt->options & XML_PARSE_COMPACT)) {
/* store the string in the node overrithing properties and nsDef */
xmlChar *tmp = (xmlChar *) &(ret->properties);
memcpy(tmp, str, len);
tmp[len] = 0;
intern = tmp;
} else if ((len <= 3) && ((cur == '"') || (cur == '\'') ||
((cur == '<') && (str[len + 1] != '!')))) { ((cur == '<') && (str[len + 1] != '!')))) {
intern = xmlDictLookup(ctxt->dict, str, len); intern = xmlDictLookup(ctxt->dict, str, len);
} else if (IS_BLANK_CH(*str) && (len < 60) && (cur == '<') && } else if (IS_BLANK_CH(*str) && (len < 60) && (cur == '<') &&
@ -1798,7 +1806,6 @@ xmlSAX2TextNode(xmlParserCtxtPtr ctxt, const xmlChar *str, int len) {
} }
} }
skip: skip:
memset(ret, 0, sizeof(xmlNode));
ret->type = XML_TEXT_NODE; ret->type = XML_TEXT_NODE;
ret->name = xmlStringText; ret->name = xmlStringText;
@ -2407,8 +2414,11 @@ xmlSAX2Characters(void *ctx, const xmlChar *ch, int len)
* We try to minimaze realloc() uses and avoid copying * We try to minimaze realloc() uses and avoid copying
* and recomputing length over and over. * and recomputing length over and over.
*/ */
if ((ctxt->nodemem == ctxt->nodelen + 1) && if (lastChild->content == (xmlChar *)&(lastChild->properties)) {
(xmlDictOwns(ctxt->dict, lastChild->content))) { lastChild->content = xmlStrdup(lastChild->content);
lastChild->properties = NULL;
} else if ((ctxt->nodemem == ctxt->nodelen + 1) &&
(xmlDictOwns(ctxt->dict, lastChild->content))) {
lastChild->content = xmlStrdup(lastChild->content); lastChild->content = xmlStrdup(lastChild->content);
} }
if (ctxt->nodelen + len >= ctxt->nodemem) { if (ctxt->nodelen + len >= ctxt->nodemem) {

14
debugXML.c
View File

@ -902,9 +902,15 @@ xmlCtxtDumpOneNode(xmlDebugCtxtPtr ctxt, xmlNodePtr node)
if (!ctxt->check) { if (!ctxt->check) {
xmlCtxtDumpSpaces(ctxt); xmlCtxtDumpSpaces(ctxt);
if (node->name == (const xmlChar *) xmlStringTextNoenc) if (node->name == (const xmlChar *) xmlStringTextNoenc)
fprintf(ctxt->output, "TEXT no enc\n"); fprintf(ctxt->output, "TEXT no enc");
else else
fprintf(ctxt->output, "TEXT\n"); fprintf(ctxt->output, "TEXT");
if (node->content == (xmlChar *) &(node->properties))
fprintf(ctxt->output, " compact\n");
else if (xmlDictOwns(ctxt->dict, node->content) == 1)
fprintf(ctxt->output, " interned\n");
else
fprintf(ctxt->output, "\n");
} }
break; break;
case XML_CDATA_SECTION_NODE: case XML_CDATA_SECTION_NODE:
@ -1005,9 +1011,9 @@ xmlCtxtDumpOneNode(xmlDebugCtxtPtr ctxt, xmlNodePtr node)
fprintf(ctxt->output, "PBM: doc == NULL !!!\n"); fprintf(ctxt->output, "PBM: doc == NULL !!!\n");
} }
ctxt->depth++; ctxt->depth++;
if (node->nsDef != NULL) if ((node->type == XML_ELEMENT_NODE) && (node->nsDef != NULL))
xmlCtxtDumpNamespaceList(ctxt, node->nsDef); xmlCtxtDumpNamespaceList(ctxt, node->nsDef);
if (node->properties != NULL) if ((node->type == XML_ELEMENT_NODE) && (node->properties != NULL))
xmlCtxtDumpAttrList(ctxt, node->properties); xmlCtxtDumpAttrList(ctxt, node->properties);
if (node->type != XML_ENTITY_REF_NODE) { if (node->type != XML_ENTITY_REF_NODE) {
if ((node->type != XML_ELEMENT_NODE) && (node->content != NULL)) { if ((node->type != XML_ELEMENT_NODE) && (node->content != NULL)) {

3
include/libxml/HTMLparser.h
View File

@ -178,7 +178,8 @@ typedef enum {
HTML_PARSE_NOWARNING= 1<<6, /* suppress warning reports */ HTML_PARSE_NOWARNING= 1<<6, /* suppress warning reports */
HTML_PARSE_PEDANTIC = 1<<7, /* pedantic error reporting */ HTML_PARSE_PEDANTIC = 1<<7, /* pedantic error reporting */
HTML_PARSE_NOBLANKS = 1<<8, /* remove blank nodes */ HTML_PARSE_NOBLANKS = 1<<8, /* remove blank nodes */
HTML_PARSE_NONET = 1<<11 /* Forbid network access */ HTML_PARSE_NONET = 1<<11,/* Forbid network access */
HTML_PARSE_COMPACT = 1<<16 /* compact small text nodes */
} htmlParserOption; } htmlParserOption;
XMLPUBFUN void XMLCALL XMLPUBFUN void XMLCALL

3
include/libxml/parser.h
View File

@ -1088,7 +1088,8 @@ typedef enum {
XML_PARSE_NODICT = 1<<12,/* Do not reuse the context dictionnary */ XML_PARSE_NODICT = 1<<12,/* Do not reuse the context dictionnary */
XML_PARSE_NSCLEAN = 1<<13,/* remove redundant namespaces declarations */ XML_PARSE_NSCLEAN = 1<<13,/* remove redundant namespaces declarations */
XML_PARSE_NOCDATA = 1<<14,/* merge CDATA as text nodes */ XML_PARSE_NOCDATA = 1<<14,/* merge CDATA as text nodes */
XML_PARSE_NOXINCNODE= 1<<15 /* do not generate XINCLUDE START/END nodes */ XML_PARSE_NOXINCNODE= 1<<15,/* do not generate XINCLUDE START/END nodes */
XML_PARSE_COMPACT = 1<<16 /* compact small text nodes */
} xmlParserOption; } xmlParserOption;
XMLPUBFUN void XMLCALL XMLPUBFUN void XMLCALL

11
parser.c
View File

@ -684,7 +684,7 @@ xmlAddDefAttrs(xmlParserCtxtPtr ctxt,
} }
/* /*
* plit the element name into prefix:localname , the string found * Split the element name into prefix:localname , the string found
* are within the DTD and hen not associated to namespace names. * are within the DTD and hen not associated to namespace names.
*/ */
name = xmlSplitQName3(fullattr, &len); name = xmlSplitQName3(fullattr, &len);
@ -11319,8 +11319,9 @@ xmlParseBalancedChunkMemoryInternal(xmlParserCtxtPtr oldctxt,
*lst = cur; *lst = cur;
while (cur != NULL) { while (cur != NULL) {
#ifdef LIBXML_VALID_ENABLED #ifdef LIBXML_VALID_ENABLED
if (oldctxt->validate && oldctxt->wellFormed && if ((oldctxt->validate) && (oldctxt->wellFormed) &&
oldctxt->myDoc && oldctxt->myDoc->intSubset) { (oldctxt->myDoc) && (oldctxt->myDoc->intSubset) &&
(cur->type == XML_ELEMENT_NODE)) {
oldctxt->valid &= xmlValidateElement(&oldctxt->vctxt, oldctxt->valid &= xmlValidateElement(&oldctxt->vctxt,
oldctxt->myDoc, cur); oldctxt->myDoc, cur);
} }
@ -12843,6 +12844,10 @@ xmlCtxtUseOptions(xmlParserCtxtPtr ctxt, int options)
ctxt->options |= XML_PARSE_NONET; ctxt->options |= XML_PARSE_NONET;
options -= XML_PARSE_NONET; options -= XML_PARSE_NONET;
} }
if (options & XML_PARSE_COMPACT) {
ctxt->options |= XML_PARSE_COMPACT;
options -= XML_PARSE_COMPACT;
}
ctxt->linenumbers = 1; ctxt->linenumbers = 1;
return (options); return (options);
} }

4
result/XPath/tests/chaptersbase
View File

@ -72,10 +72,10 @@ Set contains 1 nodes:
Expression: /child::EXAMPLE/child::head/node() Expression: /child::EXAMPLE/child::head/node()
Object is a Node Set : Object is a Node Set :
Set contains 3 nodes: Set contains 3 nodes:
1 TEXT 1 TEXT compact
content= content=
2 ELEMENT title 2 ELEMENT title
3 TEXT 3 TEXT compact
content= content=
======================== ========================

2
result/XPath/tests/idsimple
View File

@ -5,7 +5,7 @@ Object is a Node Set :
Set contains 1 nodes: Set contains 1 nodes:
1 ELEMENT EXAMPLE 1 ELEMENT EXAMPLE
ATTRIBUTE id ATTRIBUTE id
TEXT TEXT compact
content=root content=root
ATTRIBUTE prop1 ATTRIBUTE prop1
TEXT TEXT

16
result/XPath/tests/langsimple
View File

@ -5,30 +5,30 @@ Object is a Node Set :
Set contains 9 nodes: Set contains 9 nodes:
1 ELEMENT b 1 ELEMENT b
ATTRIBUTE lang ATTRIBUTE lang
TEXT TEXT compact
content=en content=en
2 ELEMENT x 2 ELEMENT x
3 ELEMENT x 3 ELEMENT x
4 ELEMENT para 4 ELEMENT para
ATTRIBUTE lang ATTRIBUTE lang
TEXT TEXT compact
content=en content=en
5 ELEMENT div 5 ELEMENT div
ATTRIBUTE lang ATTRIBUTE lang
TEXT TEXT compact
content=en content=en
6 ELEMENT para 6 ELEMENT para
7 ELEMENT para 7 ELEMENT para
ATTRIBUTE lang ATTRIBUTE lang
TEXT TEXT compact
content=EN content=EN
8 ELEMENT para 8 ELEMENT para
ATTRIBUTE lang ATTRIBUTE lang
TEXT TEXT compact
content=en-us content=en-us
9 ELEMENT para 9 ELEMENT para
ATTRIBUTE lang ATTRIBUTE lang
TEXT TEXT compact
content=EN-US content=EN-US
======================== ========================
@ -37,11 +37,11 @@ Object is a Node Set :
Set contains 2 nodes: Set contains 2 nodes:
1 ELEMENT para 1 ELEMENT para
ATTRIBUTE lang ATTRIBUTE lang
TEXT TEXT compact
content=en-us content=en-us
2 ELEMENT para 2 ELEMENT para
ATTRIBUTE lang ATTRIBUTE lang
TEXT TEXT compact
content=EN-US content=EN-US
======================== ========================

32
result/XPath/tests/mixedpat
View File

@ -5,7 +5,7 @@ Object is a Node Set :
Set contains 1 nodes: Set contains 1 nodes:
1 ELEMENT s 1 ELEMENT s
ATTRIBUTE p ATTRIBUTE p
TEXT TEXT compact
content=root content=root
======================== ========================
@ -14,11 +14,11 @@ Object is a Node Set :
Set contains 2 nodes: Set contains 2 nodes:
1 ELEMENT s 1 ELEMENT s
ATTRIBUTE p ATTRIBUTE p
TEXT TEXT compact
content=root content=root
2 ELEMENT s 2 ELEMENT s
ATTRIBUTE p ATTRIBUTE p
TEXT TEXT compact
content=p1 content=p1
======================== ========================
@ -27,11 +27,11 @@ Object is a Node Set :
Set contains 2 nodes: Set contains 2 nodes:
1 ELEMENT s 1 ELEMENT s
ATTRIBUTE p ATTRIBUTE p
TEXT TEXT compact
content=root content=root
2 ELEMENT s 2 ELEMENT s
ATTRIBUTE p ATTRIBUTE p
TEXT TEXT compact
content=p1 content=p1
======================== ========================
@ -40,11 +40,11 @@ Object is a Node Set :
Set contains 2 nodes: Set contains 2 nodes:
1 ELEMENT s 1 ELEMENT s
ATTRIBUTE p ATTRIBUTE p
TEXT TEXT compact
content=root content=root
2 ELEMENT s 2 ELEMENT s
ATTRIBUTE p ATTRIBUTE p
TEXT TEXT compact
content=p1 content=p1
======================== ========================
@ -53,15 +53,15 @@ Object is a Node Set :
Set contains 3 nodes: Set contains 3 nodes:
1 ELEMENT s 1 ELEMENT s
ATTRIBUTE p ATTRIBUTE p
TEXT TEXT compact
content=root content=root
2 ELEMENT s 2 ELEMENT s
ATTRIBUTE p ATTRIBUTE p
TEXT TEXT compact
content=p1 content=p1
3 ELEMENT s 3 ELEMENT s
ATTRIBUTE p ATTRIBUTE p
TEXT TEXT compact
content=p2 content=p2
======================== ========================
@ -70,16 +70,16 @@ Object is a Node Set :
Set contains 4 nodes: Set contains 4 nodes:
1 ELEMENT s 1 ELEMENT s
ATTRIBUTE p ATTRIBUTE p
TEXT TEXT compact
content=root content=root
2 ELEMENT p1 2 ELEMENT p1
3 ELEMENT s 3 ELEMENT s
ATTRIBUTE p ATTRIBUTE p
TEXT TEXT compact
content=p1 content=p1
4 ELEMENT s 4 ELEMENT s
ATTRIBUTE p ATTRIBUTE p
TEXT TEXT compact
content=p2 content=p2
======================== ========================
@ -88,14 +88,14 @@ Object is a Node Set :
Set contains 4 nodes: Set contains 4 nodes:
1 ELEMENT s 1 ELEMENT s
ATTRIBUTE p ATTRIBUTE p
TEXT TEXT compact
content=root content=root
2 ELEMENT p1 2 ELEMENT p1
3 ELEMENT s 3 ELEMENT s
ATTRIBUTE p ATTRIBUTE p
TEXT TEXT compact
content=p1 content=p1
4 ELEMENT s 4 ELEMENT s
ATTRIBUTE p ATTRIBUTE p
TEXT TEXT compact
content=p2 content=p2

8
result/XPath/tests/simpleabbr
View File

@ -45,7 +45,7 @@ Object is a Node Set :
Set contains 2 nodes: Set contains 2 nodes:
1 TEXT 1 TEXT
content=bla bla bla ... content=bla bla bla ...
2 TEXT 2 TEXT compact
content=... content=...
======================== ========================
@ -54,7 +54,7 @@ Object is a Node Set :
Set contains 2 nodes: Set contains 2 nodes:
1 TEXT 1 TEXT
content=bla bla bla ... content=bla bla bla ...
2 TEXT 2 TEXT compact
content=... content=...
======================== ========================
@ -63,7 +63,7 @@ Object is a Node Set :
Set contains 2 nodes: Set contains 2 nodes:
1 TEXT 1 TEXT
content=bla bla bla ... content=bla bla bla ...
2 TEXT 2 TEXT compact
content=... content=...
======================== ========================
@ -77,5 +77,5 @@ Set contains 1 nodes:
Expression: (//p/text())[position()=last()] Expression: (//p/text())[position()=last()]
Object is a Node Set : Object is a Node Set :
Set contains 1 nodes: Set contains 1 nodes:
1 TEXT 1 TEXT compact
content=... content=...

4
result/XPath/tests/simplebase
View File

@ -53,10 +53,10 @@ Set contains 1 nodes:
Expression: /child::EXAMPLE/child::head/node() Expression: /child::EXAMPLE/child::head/node()
Object is a Node Set : Object is a Node Set :
Set contains 3 nodes: Set contains 3 nodes:
1 TEXT 1 TEXT compact
content= content=
2 ELEMENT title 2 ELEMENT title
3 TEXT 3 TEXT compact
content= content=
======================== ========================

4
result/XPath/tests/usr1check
View File

@ -5,8 +5,8 @@ Object is a Node Set :
Set contains 1 nodes: Set contains 1 nodes:
1 ELEMENT ITEM 1 ELEMENT ITEM
ATTRIBUTE monto ATTRIBUTE monto
TEXT TEXT compact
content=50.12 content=50.12
ATTRIBUTE divisa ATTRIBUTE divisa
TEXT TEXT compact
content=DOL content=DOL

14
result/XPath/xptr/strrange
View File

@ -44,7 +44,7 @@ Expression: xpointer(string-range(//p, 'difficult'))
Object is a Location Set: Object is a Location Set:
1 : Object is a range : 1 : Object is a range :
From index 3 in node From index 3 in node
TEXT TEXT compact
content=a diff content=a diff
To index 4 in node To index 4 in node
TEXT TEXT
@ -56,10 +56,10 @@ Expression: xpointer(string-range(//p, 'spanning'))
Object is a Location Set: Object is a Location Set:
1 : Object is a range : 1 : Object is a range :
From index 3 in node From index 3 in node
TEXT TEXT compact
content=a span content=a span
To index 3 in node To index 3 in node
TEXT TEXT compact
content=ing one content=ing one
@ -80,17 +80,17 @@ Expression: xpointer(string-range(//seq, ''))
Object is a Location Set: Object is a Location Set:
1 : Object is a collapsed range : 1 : Object is a collapsed range :
index 1 in node index 1 in node
TEXT TEXT compact
content=123 content=123
2 : Object is a collapsed range : 2 : Object is a collapsed range :
index 2 in node index 2 in node
TEXT TEXT compact
content=123 content=123
3 : Object is a collapsed range : 3 : Object is a collapsed range :
index 3 in node index 3 in node
TEXT TEXT compact
content=123 content=123
4 : Object is a collapsed range : 4 : Object is a collapsed range :
index 4 in node index 4 in node
TEXT TEXT compact
content=123 content=123

2
result/XPath/xptr/strrange2
View File

@ -60,5 +60,5 @@ Expression: xpointer(string-range(//p, 'difficult', 1, 0))
Object is a Location Set: Object is a Location Set:
1 : Object is a collapsed range : 1 : Object is a collapsed range :
index 3 in node index 3 in node
TEXT TEXT compact
content=a diff content=a diff

2
result/xmlid/id_err2.xml
View File

@ -2,5 +2,5 @@ Object is a Node Set :
Set contains 1 nodes: Set contains 1 nodes:
1 ELEMENT foo 1 ELEMENT foo
ATTRIBUTE id ATTRIBUTE id
TEXT TEXT compact
content=bar content=bar

2
result/xmlid/id_tst1.xml
View File

@ -2,5 +2,5 @@ Object is a Node Set :
Set contains 1 nodes: Set contains 1 nodes:
1 ELEMENT foo 1 ELEMENT foo
ATTRIBUTE id ATTRIBUTE id
TEXT TEXT compact
content=bar content=bar

2
result/xmlid/id_tst2.xml
View File

@ -2,5 +2,5 @@ Object is a Node Set :
Set contains 1 nodes: Set contains 1 nodes:
1 ELEMENT foo 1 ELEMENT foo
ATTRIBUTE id ATTRIBUTE id
TEXT TEXT compact
content=bar content=bar

2
result/xmlid/id_tst3.xml
View File

@ -2,5 +2,5 @@ Object is a Node Set :
Set contains 1 nodes: Set contains 1 nodes:
1 ELEMENT o:o 1 ELEMENT o:o
ATTRIBUTE id ATTRIBUTE id
TEXT TEXT compact
content=bar content=bar

2
result/xmlid/id_tst4.xml
View File

@ -2,5 +2,5 @@ Object is a Node Set :
Set contains 1 nodes: Set contains 1 nodes:
1 ELEMENT foo 1 ELEMENT foo
ATTRIBUTE id ATTRIBUTE id
TEXT TEXT compact
content=bar content=bar

2
runtest.c
View File

@ -4306,7 +4306,7 @@ launchTests(testDescPtr tst) {
testErrorsSize = 0; testErrorsSize = 0;
testErrors[0] = 0; testErrors[0] = 0;
res = tst->func(globbuf.gl_pathv[i], result, error, res = tst->func(globbuf.gl_pathv[i], result, error,
tst->options); tst->options | XML_PARSE_COMPACT);
xmlResetLastError(); xmlResetLastError();
if (res != 0) { if (res != 0) {
fprintf(stderr, "File %s generated an error\n", fprintf(stderr, "File %s generated an error\n",

4
testXPath.c
View File

@ -178,9 +178,9 @@ int main(int argc, char **argv) {
} }
if (document == NULL) { if (document == NULL) {
if (filename == NULL) if (filename == NULL)
document = xmlReadDoc(buffer,NULL,NULL,0); document = xmlReadDoc(buffer,NULL,NULL,XML_PARSE_COMPACT);
else else
document = xmlReadFile(filename,NULL,0); document = xmlReadFile(filename,NULL,XML_PARSE_COMPACT);
} }
for (i = 1; i < argc ; i++) { for (i = 1; i < argc ; i++) {
if ((!strcmp(argv[i], "-i")) || (!strcmp(argv[i], "--input"))) { if ((!strcmp(argv[i], "-i")) || (!strcmp(argv[i], "--input"))) {

58
tree.c
View File

@ -3287,7 +3287,8 @@ xmlFreeNodeList(xmlNodePtr cur) {
if ((cur->type != XML_ELEMENT_NODE) && if ((cur->type != XML_ELEMENT_NODE) &&
(cur->type != XML_XINCLUDE_START) && (cur->type != XML_XINCLUDE_START) &&
(cur->type != XML_XINCLUDE_END) && (cur->type != XML_XINCLUDE_END) &&
(cur->type != XML_ENTITY_REF_NODE)) { (cur->type != XML_ENTITY_REF_NODE) &&
(cur->content != (xmlChar *) &(cur->properties))) {
DICT_FREE(cur->content) DICT_FREE(cur->content)
} }
if (((cur->type == XML_ELEMENT_NODE) || if (((cur->type == XML_ELEMENT_NODE) ||
@ -3356,7 +3357,8 @@ xmlFreeNode(xmlNodePtr cur) {
(cur->content != NULL) && (cur->content != NULL) &&
(cur->type != XML_ENTITY_REF_NODE) && (cur->type != XML_ENTITY_REF_NODE) &&
(cur->type != XML_XINCLUDE_END) && (cur->type != XML_XINCLUDE_END) &&
(cur->type != XML_XINCLUDE_START)) { (cur->type != XML_XINCLUDE_START) &&
(cur->content != (xmlChar *) &(cur->properties))) {
DICT_FREE(cur->content) DICT_FREE(cur->content)
} }
@ -3811,7 +3813,7 @@ xmlStaticCopyNode(const xmlNodePtr node, xmlDocPtr doc, xmlNodePtr parent,
if (!extended) if (!extended)
goto out; goto out;
if (node->nsDef != NULL) if ((node->type == XML_ELEMENT_NODE) && (node->nsDef != NULL))
ret->nsDef = xmlCopyNamespaceList(node->nsDef); ret->nsDef = xmlCopyNamespaceList(node->nsDef);
if (node->ns != NULL) { if (node->ns != NULL) {
@ -3838,7 +3840,7 @@ xmlStaticCopyNode(const xmlNodePtr node, xmlDocPtr doc, xmlNodePtr parent,
ret->ns = ns; ret->ns = ns;
} }
} }
if (node->properties != NULL) if ((node->type == XML_ELEMENT_NODE) && (node->properties != NULL))
ret->properties = xmlCopyPropList(ret, node->properties); ret->properties = xmlCopyPropList(ret, node->properties);
if (node->type == XML_ENTITY_REF_NODE) { if (node->type == XML_ENTITY_REF_NODE) {
if ((doc == NULL) || (node->doc != doc)) { if ((doc == NULL) || (node->doc != doc)) {
@ -5106,7 +5108,8 @@ xmlNodeSetContent(xmlNodePtr cur, const xmlChar *content) {
case XML_ENTITY_NODE: case XML_ENTITY_NODE:
case XML_PI_NODE: case XML_PI_NODE:
case XML_COMMENT_NODE: case XML_COMMENT_NODE:
if (cur->content != NULL) { if ((cur->content != NULL) &&
(cur->content != (xmlChar *) &(cur->properties))) {
if (!((cur->doc != NULL) && (cur->doc->dict != NULL) && if (!((cur->doc != NULL) && (cur->doc->dict != NULL) &&
(xmlDictOwns(cur->doc->dict, cur->content)))) (xmlDictOwns(cur->doc->dict, cur->content))))
xmlFree(cur->content); xmlFree(cur->content);
@ -5117,6 +5120,8 @@ xmlNodeSetContent(xmlNodePtr cur, const xmlChar *content) {
cur->content = xmlStrdup(content); cur->content = xmlStrdup(content);
} else } else
cur->content = NULL; cur->content = NULL;
cur->properties = NULL;
cur->nsDef = NULL;
break; break;
case XML_DOCUMENT_NODE: case XML_DOCUMENT_NODE:
case XML_HTML_DOCUMENT_NODE: case XML_HTML_DOCUMENT_NODE:
@ -5178,8 +5183,11 @@ xmlNodeSetContentLen(xmlNodePtr cur, const xmlChar *content, int len) {
case XML_PI_NODE: case XML_PI_NODE:
case XML_COMMENT_NODE: case XML_COMMENT_NODE:
case XML_NOTATION_NODE: case XML_NOTATION_NODE:
if (cur->content != NULL) { if ((cur->content != NULL) &&
xmlFree(cur->content); (cur->content != (xmlChar *) &(cur->properties))) {
if (!((cur->doc != NULL) && (cur->doc->dict != NULL) &&
(xmlDictOwns(cur->doc->dict, cur->content))))
xmlFree(cur->content);
} }
if (cur->children != NULL) xmlFreeNodeList(cur->children); if (cur->children != NULL) xmlFreeNodeList(cur->children);
cur->children = cur->last = NULL; cur->children = cur->last = NULL;
@ -5187,6 +5195,8 @@ xmlNodeSetContentLen(xmlNodePtr cur, const xmlChar *content, int len) {
cur->content = xmlStrndup(content, len); cur->content = xmlStrndup(content, len);
} else } else
cur->content = NULL; cur->content = NULL;
cur->properties = NULL;
cur->nsDef = NULL;
break; break;
case XML_DOCUMENT_NODE: case XML_DOCUMENT_NODE:
case XML_DTD_NODE: case XML_DTD_NODE:
@ -5257,10 +5267,12 @@ xmlNodeAddContentLen(xmlNodePtr cur, const xmlChar *content, int len) {
case XML_COMMENT_NODE: case XML_COMMENT_NODE:
case XML_NOTATION_NODE: case XML_NOTATION_NODE:
if (content != NULL) { if (content != NULL) {
if ((cur->doc != NULL) && (cur->doc->dict != NULL) && if ((cur->content == (xmlChar *) &(cur->properties)) ||
xmlDictOwns(cur->doc->dict, cur->content)) { ((cur->doc != NULL) && (cur->doc->dict != NULL) &&
cur->content = xmlDictOwns(cur->doc->dict, cur->content))) {
xmlStrncatNew(cur->content, content, len); cur->content = xmlStrncatNew(cur->content, content, len);
cur->properties = NULL;
cur->nsDef = NULL;
break; break;
} }
cur->content = xmlStrncat(cur->content, content, len); cur->content = xmlStrncat(cur->content, content, len);
@ -5903,7 +5915,8 @@ xmlHasProp(xmlNodePtr node, const xmlChar *name) {
xmlAttrPtr prop; xmlAttrPtr prop;
xmlDocPtr doc; xmlDocPtr doc;
if ((node == NULL) || (name == NULL)) return(NULL); if ((node == NULL) || (node->type != XML_ELEMENT_NODE) || (name == NULL))
return(NULL);
/* /*
* Check on the properties attached to the node * Check on the properties attached to the node
*/ */
@ -5959,7 +5972,7 @@ xmlHasNsProp(xmlNodePtr node, const xmlChar *name, const xmlChar *nameSpace) {
xmlDocPtr doc; xmlDocPtr doc;
#endif /* LIBXML_TREE_ENABLED */ #endif /* LIBXML_TREE_ENABLED */
if (node == NULL) if ((node == NULL) || (node->type != XML_ELEMENT_NODE) || (name == NULL))
return(NULL); return(NULL);
prop = node->properties; prop = node->properties;
@ -6057,7 +6070,9 @@ xmlGetProp(xmlNodePtr node, const xmlChar *name) {
xmlAttrPtr prop; xmlAttrPtr prop;
xmlDocPtr doc; xmlDocPtr doc;
if ((node == NULL) || (name == NULL)) return(NULL); if ((node == NULL) || (node->type != XML_ELEMENT_NODE) || (name == NULL))
return(NULL);
/* /*
* Check on the properties attached to the node * Check on the properties attached to the node
*/ */
@ -6114,7 +6129,8 @@ xmlGetNoNsProp(xmlNodePtr node, const xmlChar *name) {
xmlAttrPtr prop; xmlAttrPtr prop;
xmlDocPtr doc; xmlDocPtr doc;
if ((node == NULL) || (name == NULL)) return(NULL); if ((node == NULL) || (node->type != XML_ELEMENT_NODE) || (name == NULL))
return(NULL);
/* /*
* Check on the properties attached to the node * Check on the properties attached to the node
*/ */
@ -6172,7 +6188,7 @@ xmlGetNsProp(xmlNodePtr node, const xmlChar *name, const xmlChar *nameSpace) {
xmlDocPtr doc; xmlDocPtr doc;
xmlNsPtr ns; xmlNsPtr ns;
if (node == NULL) if ((node == NULL) || (node->type != XML_ELEMENT_NODE))
return(NULL); return(NULL);
prop = node->properties; prop = node->properties;
@ -6236,7 +6252,7 @@ int
xmlUnsetProp(xmlNodePtr node, const xmlChar *name) { xmlUnsetProp(xmlNodePtr node, const xmlChar *name) {
xmlAttrPtr prop, prev = NULL;; xmlAttrPtr prop, prev = NULL;;
if ((node == NULL) || (name == NULL)) if ((node == NULL) || (node->type != XML_ELEMENT_NODE) || (name == NULL))
return(-1); return(-1);
prop = node->properties; prop = node->properties;
while (prop != NULL) { while (prop != NULL) {
@ -6265,7 +6281,7 @@ int
xmlUnsetNsProp(xmlNodePtr node, xmlNsPtr ns, const xmlChar *name) { xmlUnsetNsProp(xmlNodePtr node, xmlNsPtr ns, const xmlChar *name) {
xmlAttrPtr prop, prev = NULL;; xmlAttrPtr prop, prev = NULL;;
if ((node == NULL) || (name == NULL)) if ((node == NULL) || (node->type != XML_ELEMENT_NODE) || (name == NULL))
return(-1); return(-1);
prop = node->properties; prop = node->properties;
if (ns == NULL) if (ns == NULL)
@ -6471,12 +6487,14 @@ xmlTextConcat(xmlNodePtr node, const xmlChar *content, int len) {
return(-1); return(-1);
} }
/* need to check if content is currently in the dictionary */ /* need to check if content is currently in the dictionary */
if ((node->doc != NULL) && (node->doc->dict != NULL) && if ((node->content == (xmlChar *) &(node->properties)) ||
xmlDictOwns(node->doc->dict, node->content)) { ((node->doc != NULL) && (node->doc->dict != NULL) &&
xmlDictOwns(node->doc->dict, node->content))) {
node->content = xmlStrncatNew(node->content, content, len); node->content = xmlStrncatNew(node->content, content, len);
} else { } else {
node->content = xmlStrncat(node->content, content, len); node->content = xmlStrncat(node->content, content, len);
} }
node->properties = NULL;
if (node->content == NULL) if (node->content == NULL)
return(-1); return(-1);
return(0); return(0);

48
valid.c
View File

@ -5860,24 +5860,12 @@ xmlValidateOneElement(xmlValidCtxtPtr ctxt, xmlDocPtr doc,
NULL,NULL,NULL); NULL,NULL,NULL);
return(0); return(0);
} }
if (elem->properties != NULL) {
xmlErrValidNode(ctxt, elem, XML_ERR_INTERNAL_ERROR,
"Text element has attribute !\n",
NULL,NULL,NULL);
return(0);
}
if (elem->ns != NULL) { if (elem->ns != NULL) {
xmlErrValidNode(ctxt, elem, XML_ERR_INTERNAL_ERROR, xmlErrValidNode(ctxt, elem, XML_ERR_INTERNAL_ERROR,
"Text element has namespace !\n", "Text element has namespace !\n",
NULL,NULL,NULL); NULL,NULL,NULL);
return(0); return(0);
} }
if (elem->nsDef != NULL) {
xmlErrValidNode(ctxt, elem, XML_ERR_INTERNAL_ERROR,
"Text element has namespace !\n",
NULL,NULL,NULL);
return(0);
}
if (elem->content == NULL) { if (elem->content == NULL) {
xmlErrValidNode(ctxt, elem, XML_ERR_INTERNAL_ERROR, xmlErrValidNode(ctxt, elem, XML_ERR_INTERNAL_ERROR,
"Text element has no content !\n", "Text element has no content !\n",
@ -6302,23 +6290,25 @@ xmlValidateElement(xmlValidCtxtPtr ctxt, xmlDocPtr doc, xmlNodePtr elem) {
} }
ret &= xmlValidateOneElement(ctxt, doc, elem); ret &= xmlValidateOneElement(ctxt, doc, elem);
attr = elem->properties; if (elem->type == XML_ELEMENT_NODE) {
while (attr != NULL) { attr = elem->properties;
value = xmlNodeListGetString(doc, attr->children, 0); while (attr != NULL) {
ret &= xmlValidateOneAttribute(ctxt, doc, elem, attr, value); value = xmlNodeListGetString(doc, attr->children, 0);
if (value != NULL) ret &= xmlValidateOneAttribute(ctxt, doc, elem, attr, value);
xmlFree((char *)value); if (value != NULL)
attr= attr->next; xmlFree((char *)value);
} attr= attr->next;
ns = elem->nsDef; }
while (ns != NULL) { ns = elem->nsDef;
if (elem->ns == NULL) while (ns != NULL) {
ret &= xmlValidateOneNamespace(ctxt, doc, elem, NULL, if (elem->ns == NULL)
ns, ns->href); ret &= xmlValidateOneNamespace(ctxt, doc, elem, NULL,
else ns, ns->href);
ret &= xmlValidateOneNamespace(ctxt, doc, elem, elem->ns->prefix, else
ns, ns->href); ret &= xmlValidateOneNamespace(ctxt, doc, elem,
ns = ns->next; elem->ns->prefix, ns, ns->href);
ns = ns->next;
}
} }
child = elem->children; child = elem->children;
while (child != NULL) { while (child != NULL) {

6
xmllint.c
View File

@ -196,7 +196,7 @@ static const char *pattern = NULL;
static xmlPatternPtr patternc = NULL; static xmlPatternPtr patternc = NULL;
static xmlStreamCtxtPtr patstream = NULL; static xmlStreamCtxtPtr patstream = NULL;
#endif #endif
static int options = 0; static int options = XML_PARSE_COMPACT;
static int sax = 0; static int sax = 0;
/************************************************************************ /************************************************************************
@ -2793,6 +2793,7 @@ static void usage(const char *name) {
printf("\t--path 'paths': provide a set of paths for resources\n"); printf("\t--path 'paths': provide a set of paths for resources\n");
printf("\t--load-trace : print trace of all external entites loaded\n"); printf("\t--load-trace : print trace of all external entites loaded\n");
printf("\t--nonet : refuse to fetch DTDs or entities over network\n"); printf("\t--nonet : refuse to fetch DTDs or entities over network\n");
printf("\t--nocompact : do not generate compact text nodes\n");
printf("\t--htmlout : output results as HTML\n"); printf("\t--htmlout : output results as HTML\n");
printf("\t--nowrap : do not put HTML doc wrapper\n"); printf("\t--nowrap : do not put HTML doc wrapper\n");
#ifdef LIBXML_VALID_ENABLED #ifdef LIBXML_VALID_ENABLED
@ -3175,6 +3176,9 @@ main(int argc, char **argv) {
} else if ((!strcmp(argv[i], "-nonet")) || } else if ((!strcmp(argv[i], "-nonet")) ||
(!strcmp(argv[i], "--nonet"))) { (!strcmp(argv[i], "--nonet"))) {
options |= XML_PARSE_NONET; options |= XML_PARSE_NONET;
} else if ((!strcmp(argv[i], "-nocompact")) ||
(!strcmp(argv[i], "--nocompact"))) {
options &= ~XML_PARSE_COMPACT;
} else if ((!strcmp(argv[i], "-load-trace")) || } else if ((!strcmp(argv[i], "-load-trace")) ||
(!strcmp(argv[i], "--load-trace"))) { (!strcmp(argv[i], "--load-trace"))) {
load_trace++; load_trace++;

16
xmlreader.c
View File

@ -348,7 +348,8 @@ xmlTextReaderFreeNodeList(xmlTextReaderPtr reader, xmlNodePtr cur) {
(cur->type == XML_XINCLUDE_END)) && (cur->type == XML_XINCLUDE_END)) &&
(cur->properties != NULL)) (cur->properties != NULL))
xmlTextReaderFreePropList(reader, cur->properties); xmlTextReaderFreePropList(reader, cur->properties);
if ((cur->type != XML_ELEMENT_NODE) && if ((cur->content != (xmlChar *) &(cur->properties)) &&
(cur->type != XML_ELEMENT_NODE) &&
(cur->type != XML_XINCLUDE_START) && (cur->type != XML_XINCLUDE_START) &&
(cur->type != XML_XINCLUDE_END) && (cur->type != XML_XINCLUDE_END) &&
(cur->type != XML_ENTITY_REF_NODE)) { (cur->type != XML_ENTITY_REF_NODE)) {
@ -422,7 +423,8 @@ xmlTextReaderFreeNode(xmlTextReaderPtr reader, xmlNodePtr cur) {
(cur->type == XML_XINCLUDE_END)) && (cur->type == XML_XINCLUDE_END)) &&
(cur->properties != NULL)) (cur->properties != NULL))
xmlTextReaderFreePropList(reader, cur->properties); xmlTextReaderFreePropList(reader, cur->properties);
if ((cur->type != XML_ELEMENT_NODE) && if ((cur->content != (xmlChar *) &(cur->properties)) &&
(cur->type != XML_ELEMENT_NODE) &&
(cur->type != XML_XINCLUDE_START) && (cur->type != XML_XINCLUDE_START) &&
(cur->type != XML_XINCLUDE_END) && (cur->type != XML_XINCLUDE_END) &&
(cur->type != XML_ENTITY_REF_NODE)) { (cur->type != XML_ENTITY_REF_NODE)) {
@ -2810,7 +2812,9 @@ xmlTextReaderReadAttributeValue(xmlTextReaderPtr reader) {
reader->faketext = xmlNewDocText(reader->node->doc, reader->faketext = xmlNewDocText(reader->node->doc,
ns->href); ns->href);
} else { } else {
if (reader->faketext->content != NULL) if ((reader->faketext->content != NULL) &&
(reader->faketext->content !=
(xmlChar *) &(reader->faketext->properties)))
xmlFree(reader->faketext->content); xmlFree(reader->faketext->content);
reader->faketext->content = xmlStrdup(ns->href); reader->faketext->content = xmlStrdup(ns->href);
} }
@ -4776,6 +4780,12 @@ xmlTextReaderSetup(xmlTextReaderPtr reader,
if (reader == NULL) if (reader == NULL)
return (-1); return (-1);
/*
* we force the generation of compact text nodes on the reader
* since usr applications should never modify the tree
*/
options |= XML_PARSE_COMPACT;
reader->doc = NULL; reader->doc = NULL;
reader->entNr = 0; reader->entNr = 0;
reader->parserFlags = options; reader->parserFlags = options;