added a parser XML_PARSE_COMPACT option to allocate small text nodes (less
* 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
diff --git a/ChangeLog b/ChangeLog
index c928c97..a32e9f7 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -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>
* configure.in: patch from Andrew W. Nosenko, use se $GCC = 'yes'
diff --git a/HTMLparser.c b/HTMLparser.c
index fe36c89..6b8b562 100644
--- a/HTMLparser.c
+++ b/HTMLparser.c
@@ -5820,6 +5820,10 @@
ctxt->recovery = 1;
} else
ctxt->recovery = 0;
+ if (options & HTML_PARSE_COMPACT) {
+ ctxt->options |= HTML_PARSE_COMPACT;
+ options -= HTML_PARSE_COMPACT;
+ }
ctxt->dictNames = 0;
return (options);
}
diff --git a/SAX2.c b/SAX2.c
index 241e90d..a73fa1f 100644
--- a/SAX2.c
+++ b/SAX2.c
@@ -1777,6 +1777,7 @@
xmlErrMemory(ctxt, "xmlSAX2Characters");
return(NULL);
}
+ memset(ret, 0, sizeof(xmlNode));
/*
* intern the formatting blanks found between tags, or the
* very short strings
@@ -1784,7 +1785,14 @@
if (ctxt->dictNames) {
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] != '!')))) {
intern = xmlDictLookup(ctxt->dict, str, len);
} else if (IS_BLANK_CH(*str) && (len < 60) && (cur == '<') &&
@@ -1798,7 +1806,6 @@
}
}
skip:
- memset(ret, 0, sizeof(xmlNode));
ret->type = XML_TEXT_NODE;
ret->name = xmlStringText;
@@ -2407,8 +2414,11 @@
* We try to minimaze realloc() uses and avoid copying
* and recomputing length over and over.
*/
- if ((ctxt->nodemem == ctxt->nodelen + 1) &&
- (xmlDictOwns(ctxt->dict, lastChild->content))) {
+ if (lastChild->content == (xmlChar *)&(lastChild->properties)) {
+ 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);
}
if (ctxt->nodelen + len >= ctxt->nodemem) {
diff --git a/debugXML.c b/debugXML.c
index 3cb2848..24c7c4d 100644
--- a/debugXML.c
+++ b/debugXML.c
@@ -902,9 +902,15 @@
if (!ctxt->check) {
xmlCtxtDumpSpaces(ctxt);
if (node->name == (const xmlChar *) xmlStringTextNoenc)
- fprintf(ctxt->output, "TEXT no enc\n");
+ fprintf(ctxt->output, "TEXT no enc");
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;
case XML_CDATA_SECTION_NODE:
@@ -1005,9 +1011,9 @@
fprintf(ctxt->output, "PBM: doc == NULL !!!\n");
}
ctxt->depth++;
- if (node->nsDef != NULL)
+ if ((node->type == XML_ELEMENT_NODE) && (node->nsDef != NULL))
xmlCtxtDumpNamespaceList(ctxt, node->nsDef);
- if (node->properties != NULL)
+ if ((node->type == XML_ELEMENT_NODE) && (node->properties != NULL))
xmlCtxtDumpAttrList(ctxt, node->properties);
if (node->type != XML_ENTITY_REF_NODE) {
if ((node->type != XML_ELEMENT_NODE) && (node->content != NULL)) {
diff --git a/include/libxml/HTMLparser.h b/include/libxml/HTMLparser.h
index 2c1e8d2..8477efb 100644
--- a/include/libxml/HTMLparser.h
+++ b/include/libxml/HTMLparser.h
@@ -178,7 +178,8 @@
HTML_PARSE_NOWARNING= 1<<6, /* suppress warning reports */
HTML_PARSE_PEDANTIC = 1<<7, /* pedantic error reporting */
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;
XMLPUBFUN void XMLCALL
diff --git a/include/libxml/parser.h b/include/libxml/parser.h
index 6e750e7..6d6ea37 100644
--- a/include/libxml/parser.h
+++ b/include/libxml/parser.h
@@ -1088,7 +1088,8 @@
XML_PARSE_NODICT = 1<<12,/* Do not reuse the context dictionnary */
XML_PARSE_NSCLEAN = 1<<13,/* remove redundant namespaces declarations */
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;
XMLPUBFUN void XMLCALL
diff --git a/parser.c b/parser.c
index 44e6180..6a905a0 100644
--- a/parser.c
+++ b/parser.c
@@ -684,7 +684,7 @@
}
/*
- * 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.
*/
name = xmlSplitQName3(fullattr, &len);
@@ -11319,8 +11319,9 @@
*lst = cur;
while (cur != NULL) {
#ifdef LIBXML_VALID_ENABLED
- if (oldctxt->validate && oldctxt->wellFormed &&
- oldctxt->myDoc && oldctxt->myDoc->intSubset) {
+ if ((oldctxt->validate) && (oldctxt->wellFormed) &&
+ (oldctxt->myDoc) && (oldctxt->myDoc->intSubset) &&
+ (cur->type == XML_ELEMENT_NODE)) {
oldctxt->valid &= xmlValidateElement(&oldctxt->vctxt,
oldctxt->myDoc, cur);
}
@@ -12843,6 +12844,10 @@
ctxt->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;
return (options);
}
diff --git a/result/XPath/tests/chaptersbase b/result/XPath/tests/chaptersbase
index 03c6567..10636c3 100644
--- a/result/XPath/tests/chaptersbase
+++ b/result/XPath/tests/chaptersbase
@@ -72,10 +72,10 @@
Expression: /child::EXAMPLE/child::head/node()
Object is a Node Set :
Set contains 3 nodes:
-1 TEXT
+1 TEXT compact
content=
2 ELEMENT title
-3 TEXT
+3 TEXT compact
content=
========================
diff --git a/result/XPath/tests/idsimple b/result/XPath/tests/idsimple
index 891b52e..76d9286 100644
--- a/result/XPath/tests/idsimple
+++ b/result/XPath/tests/idsimple
@@ -5,7 +5,7 @@
Set contains 1 nodes:
1 ELEMENT EXAMPLE
ATTRIBUTE id
- TEXT
+ TEXT compact
content=root
ATTRIBUTE prop1
TEXT
diff --git a/result/XPath/tests/langsimple b/result/XPath/tests/langsimple
index d8d7afd..0254462 100644
--- a/result/XPath/tests/langsimple
+++ b/result/XPath/tests/langsimple
@@ -5,30 +5,30 @@
Set contains 9 nodes:
1 ELEMENT b
ATTRIBUTE lang
- TEXT
+ TEXT compact
content=en
2 ELEMENT x
3 ELEMENT x
4 ELEMENT para
ATTRIBUTE lang
- TEXT
+ TEXT compact
content=en
5 ELEMENT div
ATTRIBUTE lang
- TEXT
+ TEXT compact
content=en
6 ELEMENT para
7 ELEMENT para
ATTRIBUTE lang
- TEXT
+ TEXT compact
content=EN
8 ELEMENT para
ATTRIBUTE lang
- TEXT
+ TEXT compact
content=en-us
9 ELEMENT para
ATTRIBUTE lang
- TEXT
+ TEXT compact
content=EN-US
========================
@@ -37,11 +37,11 @@
Set contains 2 nodes:
1 ELEMENT para
ATTRIBUTE lang
- TEXT
+ TEXT compact
content=en-us
2 ELEMENT para
ATTRIBUTE lang
- TEXT
+ TEXT compact
content=EN-US
========================
diff --git a/result/XPath/tests/mixedpat b/result/XPath/tests/mixedpat
index 3be0c49..f3d746a 100644
--- a/result/XPath/tests/mixedpat
+++ b/result/XPath/tests/mixedpat
@@ -5,7 +5,7 @@
Set contains 1 nodes:
1 ELEMENT s
ATTRIBUTE p
- TEXT
+ TEXT compact
content=root
========================
@@ -14,11 +14,11 @@
Set contains 2 nodes:
1 ELEMENT s
ATTRIBUTE p
- TEXT
+ TEXT compact
content=root
2 ELEMENT s
ATTRIBUTE p
- TEXT
+ TEXT compact
content=p1
========================
@@ -27,11 +27,11 @@
Set contains 2 nodes:
1 ELEMENT s
ATTRIBUTE p
- TEXT
+ TEXT compact
content=root
2 ELEMENT s
ATTRIBUTE p
- TEXT
+ TEXT compact
content=p1
========================
@@ -40,11 +40,11 @@
Set contains 2 nodes:
1 ELEMENT s
ATTRIBUTE p
- TEXT
+ TEXT compact
content=root
2 ELEMENT s
ATTRIBUTE p
- TEXT
+ TEXT compact
content=p1
========================
@@ -53,15 +53,15 @@
Set contains 3 nodes:
1 ELEMENT s
ATTRIBUTE p
- TEXT
+ TEXT compact
content=root
2 ELEMENT s
ATTRIBUTE p
- TEXT
+ TEXT compact
content=p1
3 ELEMENT s
ATTRIBUTE p
- TEXT
+ TEXT compact
content=p2
========================
@@ -70,16 +70,16 @@
Set contains 4 nodes:
1 ELEMENT s
ATTRIBUTE p
- TEXT
+ TEXT compact
content=root
2 ELEMENT p1
3 ELEMENT s
ATTRIBUTE p
- TEXT
+ TEXT compact
content=p1
4 ELEMENT s
ATTRIBUTE p
- TEXT
+ TEXT compact
content=p2
========================
@@ -88,14 +88,14 @@
Set contains 4 nodes:
1 ELEMENT s
ATTRIBUTE p
- TEXT
+ TEXT compact
content=root
2 ELEMENT p1
3 ELEMENT s
ATTRIBUTE p
- TEXT
+ TEXT compact
content=p1
4 ELEMENT s
ATTRIBUTE p
- TEXT
+ TEXT compact
content=p2
diff --git a/result/XPath/tests/simpleabbr b/result/XPath/tests/simpleabbr
index 85da41a..6137812 100644
--- a/result/XPath/tests/simpleabbr
+++ b/result/XPath/tests/simpleabbr
@@ -45,7 +45,7 @@
Set contains 2 nodes:
1 TEXT
content=bla bla bla ...
-2 TEXT
+2 TEXT compact
content=...
========================
@@ -54,7 +54,7 @@
Set contains 2 nodes:
1 TEXT
content=bla bla bla ...
-2 TEXT
+2 TEXT compact
content=...
========================
@@ -63,7 +63,7 @@
Set contains 2 nodes:
1 TEXT
content=bla bla bla ...
-2 TEXT
+2 TEXT compact
content=...
========================
@@ -77,5 +77,5 @@
Expression: (//p/text())[position()=last()]
Object is a Node Set :
Set contains 1 nodes:
-1 TEXT
+1 TEXT compact
content=...
diff --git a/result/XPath/tests/simplebase b/result/XPath/tests/simplebase
index 38fa144..3ba1749 100644
--- a/result/XPath/tests/simplebase
+++ b/result/XPath/tests/simplebase
@@ -53,10 +53,10 @@
Expression: /child::EXAMPLE/child::head/node()
Object is a Node Set :
Set contains 3 nodes:
-1 TEXT
+1 TEXT compact
content=
2 ELEMENT title
-3 TEXT
+3 TEXT compact
content=
========================
diff --git a/result/XPath/tests/usr1check b/result/XPath/tests/usr1check
index 0308991..f284198 100644
--- a/result/XPath/tests/usr1check
+++ b/result/XPath/tests/usr1check
@@ -5,8 +5,8 @@
Set contains 1 nodes:
1 ELEMENT ITEM
ATTRIBUTE monto
- TEXT
+ TEXT compact
content=50.12
ATTRIBUTE divisa
- TEXT
+ TEXT compact
content=DOL
diff --git a/result/XPath/xptr/strrange b/result/XPath/xptr/strrange
index c14ed91..9c0096d 100644
--- a/result/XPath/xptr/strrange
+++ b/result/XPath/xptr/strrange
@@ -44,7 +44,7 @@
Object is a Location Set:
1 : Object is a range :
From index 3 in node
- TEXT
+ TEXT compact
content=a diff
To index 4 in node
TEXT
@@ -56,10 +56,10 @@
Object is a Location Set:
1 : Object is a range :
From index 3 in node
- TEXT
+ TEXT compact
content=a span
To index 3 in node
- TEXT
+ TEXT compact
content=ing one
@@ -80,17 +80,17 @@
Object is a Location Set:
1 : Object is a collapsed range :
index 1 in node
- TEXT
+ TEXT compact
content=123
2 : Object is a collapsed range :
index 2 in node
- TEXT
+ TEXT compact
content=123
3 : Object is a collapsed range :
index 3 in node
- TEXT
+ TEXT compact
content=123
4 : Object is a collapsed range :
index 4 in node
- TEXT
+ TEXT compact
content=123
diff --git a/result/XPath/xptr/strrange2 b/result/XPath/xptr/strrange2
index ea6ee45..fd6c0ea 100644
--- a/result/XPath/xptr/strrange2
+++ b/result/XPath/xptr/strrange2
@@ -60,5 +60,5 @@
Object is a Location Set:
1 : Object is a collapsed range :
index 3 in node
- TEXT
+ TEXT compact
content=a diff
diff --git a/result/xmlid/id_err2.xml b/result/xmlid/id_err2.xml
index 33ee896..c7956dc 100644
--- a/result/xmlid/id_err2.xml
+++ b/result/xmlid/id_err2.xml
@@ -2,5 +2,5 @@
Set contains 1 nodes:
1 ELEMENT foo
ATTRIBUTE id
- TEXT
+ TEXT compact
content=bar
diff --git a/result/xmlid/id_tst1.xml b/result/xmlid/id_tst1.xml
index 33ee896..c7956dc 100644
--- a/result/xmlid/id_tst1.xml
+++ b/result/xmlid/id_tst1.xml
@@ -2,5 +2,5 @@
Set contains 1 nodes:
1 ELEMENT foo
ATTRIBUTE id
- TEXT
+ TEXT compact
content=bar
diff --git a/result/xmlid/id_tst2.xml b/result/xmlid/id_tst2.xml
index 33ee896..c7956dc 100644
--- a/result/xmlid/id_tst2.xml
+++ b/result/xmlid/id_tst2.xml
@@ -2,5 +2,5 @@
Set contains 1 nodes:
1 ELEMENT foo
ATTRIBUTE id
- TEXT
+ TEXT compact
content=bar
diff --git a/result/xmlid/id_tst3.xml b/result/xmlid/id_tst3.xml
index e2f8228..bf020eb 100644
--- a/result/xmlid/id_tst3.xml
+++ b/result/xmlid/id_tst3.xml
@@ -2,5 +2,5 @@
Set contains 1 nodes:
1 ELEMENT o:o
ATTRIBUTE id
- TEXT
+ TEXT compact
content=bar
diff --git a/result/xmlid/id_tst4.xml b/result/xmlid/id_tst4.xml
index 33ee896..c7956dc 100644
--- a/result/xmlid/id_tst4.xml
+++ b/result/xmlid/id_tst4.xml
@@ -2,5 +2,5 @@
Set contains 1 nodes:
1 ELEMENT foo
ATTRIBUTE id
- TEXT
+ TEXT compact
content=bar
diff --git a/runtest.c b/runtest.c
index d03322d..897cb9b 100644
--- a/runtest.c
+++ b/runtest.c
@@ -4306,7 +4306,7 @@
testErrorsSize = 0;
testErrors[0] = 0;
res = tst->func(globbuf.gl_pathv[i], result, error,
- tst->options);
+ tst->options | XML_PARSE_COMPACT);
xmlResetLastError();
if (res != 0) {
fprintf(stderr, "File %s generated an error\n",
diff --git a/testXPath.c b/testXPath.c
index 8086a81..677419a 100644
--- a/testXPath.c
+++ b/testXPath.c
@@ -178,9 +178,9 @@
}
if (document == NULL) {
if (filename == NULL)
- document = xmlReadDoc(buffer,NULL,NULL,0);
+ document = xmlReadDoc(buffer,NULL,NULL,XML_PARSE_COMPACT);
else
- document = xmlReadFile(filename,NULL,0);
+ document = xmlReadFile(filename,NULL,XML_PARSE_COMPACT);
}
for (i = 1; i < argc ; i++) {
if ((!strcmp(argv[i], "-i")) || (!strcmp(argv[i], "--input"))) {
diff --git a/tree.c b/tree.c
index 3e5378f..1da8400 100644
--- a/tree.c
+++ b/tree.c
@@ -3287,7 +3287,8 @@
if ((cur->type != XML_ELEMENT_NODE) &&
(cur->type != XML_XINCLUDE_START) &&
(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)
}
if (((cur->type == XML_ELEMENT_NODE) ||
@@ -3356,7 +3357,8 @@
(cur->content != NULL) &&
(cur->type != XML_ENTITY_REF_NODE) &&
(cur->type != XML_XINCLUDE_END) &&
- (cur->type != XML_XINCLUDE_START)) {
+ (cur->type != XML_XINCLUDE_START) &&
+ (cur->content != (xmlChar *) &(cur->properties))) {
DICT_FREE(cur->content)
}
@@ -3811,7 +3813,7 @@
if (!extended)
goto out;
- if (node->nsDef != NULL)
+ if ((node->type == XML_ELEMENT_NODE) && (node->nsDef != NULL))
ret->nsDef = xmlCopyNamespaceList(node->nsDef);
if (node->ns != NULL) {
@@ -3838,7 +3840,7 @@
ret->ns = ns;
}
}
- if (node->properties != NULL)
+ if ((node->type == XML_ELEMENT_NODE) && (node->properties != NULL))
ret->properties = xmlCopyPropList(ret, node->properties);
if (node->type == XML_ENTITY_REF_NODE) {
if ((doc == NULL) || (node->doc != doc)) {
@@ -5106,7 +5108,8 @@
case XML_ENTITY_NODE:
case XML_PI_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) &&
(xmlDictOwns(cur->doc->dict, cur->content))))
xmlFree(cur->content);
@@ -5117,6 +5120,8 @@
cur->content = xmlStrdup(content);
} else
cur->content = NULL;
+ cur->properties = NULL;
+ cur->nsDef = NULL;
break;
case XML_DOCUMENT_NODE:
case XML_HTML_DOCUMENT_NODE:
@@ -5178,8 +5183,11 @@
case XML_PI_NODE:
case XML_COMMENT_NODE:
case XML_NOTATION_NODE:
- if (cur->content != NULL) {
- xmlFree(cur->content);
+ if ((cur->content != NULL) &&
+ (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);
cur->children = cur->last = NULL;
@@ -5187,6 +5195,8 @@
cur->content = xmlStrndup(content, len);
} else
cur->content = NULL;
+ cur->properties = NULL;
+ cur->nsDef = NULL;
break;
case XML_DOCUMENT_NODE:
case XML_DTD_NODE:
@@ -5257,10 +5267,12 @@
case XML_COMMENT_NODE:
case XML_NOTATION_NODE:
if (content != NULL) {
- if ((cur->doc != NULL) && (cur->doc->dict != NULL) &&
- xmlDictOwns(cur->doc->dict, cur->content)) {
- cur->content =
- xmlStrncatNew(cur->content, content, len);
+ if ((cur->content == (xmlChar *) &(cur->properties)) ||
+ ((cur->doc != NULL) && (cur->doc->dict != NULL) &&
+ xmlDictOwns(cur->doc->dict, cur->content))) {
+ cur->content = xmlStrncatNew(cur->content, content, len);
+ cur->properties = NULL;
+ cur->nsDef = NULL;
break;
}
cur->content = xmlStrncat(cur->content, content, len);
@@ -5903,7 +5915,8 @@
xmlAttrPtr prop;
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
*/
@@ -5959,7 +5972,7 @@
xmlDocPtr doc;
#endif /* LIBXML_TREE_ENABLED */
- if (node == NULL)
+ if ((node == NULL) || (node->type != XML_ELEMENT_NODE) || (name == NULL))
return(NULL);
prop = node->properties;
@@ -6057,7 +6070,9 @@
xmlAttrPtr prop;
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
*/
@@ -6114,7 +6129,8 @@
xmlAttrPtr prop;
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
*/
@@ -6172,7 +6188,7 @@
xmlDocPtr doc;
xmlNsPtr ns;
- if (node == NULL)
+ if ((node == NULL) || (node->type != XML_ELEMENT_NODE))
return(NULL);
prop = node->properties;
@@ -6236,7 +6252,7 @@
xmlUnsetProp(xmlNodePtr node, const xmlChar *name) {
xmlAttrPtr prop, prev = NULL;;
- if ((node == NULL) || (name == NULL))
+ if ((node == NULL) || (node->type != XML_ELEMENT_NODE) || (name == NULL))
return(-1);
prop = node->properties;
while (prop != NULL) {
@@ -6265,7 +6281,7 @@
xmlUnsetNsProp(xmlNodePtr node, xmlNsPtr ns, const xmlChar *name) {
xmlAttrPtr prop, prev = NULL;;
- if ((node == NULL) || (name == NULL))
+ if ((node == NULL) || (node->type != XML_ELEMENT_NODE) || (name == NULL))
return(-1);
prop = node->properties;
if (ns == NULL)
@@ -6471,12 +6487,14 @@
return(-1);
}
/* need to check if content is currently in the dictionary */
- if ((node->doc != NULL) && (node->doc->dict != NULL) &&
- xmlDictOwns(node->doc->dict, node->content)) {
+ if ((node->content == (xmlChar *) &(node->properties)) ||
+ ((node->doc != NULL) && (node->doc->dict != NULL) &&
+ xmlDictOwns(node->doc->dict, node->content))) {
node->content = xmlStrncatNew(node->content, content, len);
} else {
node->content = xmlStrncat(node->content, content, len);
}
+ node->properties = NULL;
if (node->content == NULL)
return(-1);
return(0);
diff --git a/valid.c b/valid.c
index 35c21ae..c0141ff 100644
--- a/valid.c
+++ b/valid.c
@@ -5860,24 +5860,12 @@
NULL,NULL,NULL);
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) {
xmlErrValidNode(ctxt, elem, XML_ERR_INTERNAL_ERROR,
"Text element has namespace !\n",
NULL,NULL,NULL);
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) {
xmlErrValidNode(ctxt, elem, XML_ERR_INTERNAL_ERROR,
"Text element has no content !\n",
@@ -6302,23 +6290,25 @@
}
ret &= xmlValidateOneElement(ctxt, doc, elem);
- attr = elem->properties;
- while (attr != NULL) {
- value = xmlNodeListGetString(doc, attr->children, 0);
- ret &= xmlValidateOneAttribute(ctxt, doc, elem, attr, value);
- if (value != NULL)
- xmlFree((char *)value);
- attr= attr->next;
- }
- ns = elem->nsDef;
- while (ns != NULL) {
- if (elem->ns == NULL)
- ret &= xmlValidateOneNamespace(ctxt, doc, elem, NULL,
- ns, ns->href);
- else
- ret &= xmlValidateOneNamespace(ctxt, doc, elem, elem->ns->prefix,
- ns, ns->href);
- ns = ns->next;
+ if (elem->type == XML_ELEMENT_NODE) {
+ attr = elem->properties;
+ while (attr != NULL) {
+ value = xmlNodeListGetString(doc, attr->children, 0);
+ ret &= xmlValidateOneAttribute(ctxt, doc, elem, attr, value);
+ if (value != NULL)
+ xmlFree((char *)value);
+ attr= attr->next;
+ }
+ ns = elem->nsDef;
+ while (ns != NULL) {
+ if (elem->ns == NULL)
+ ret &= xmlValidateOneNamespace(ctxt, doc, elem, NULL,
+ ns, ns->href);
+ else
+ ret &= xmlValidateOneNamespace(ctxt, doc, elem,
+ elem->ns->prefix, ns, ns->href);
+ ns = ns->next;
+ }
}
child = elem->children;
while (child != NULL) {
diff --git a/xmllint.c b/xmllint.c
index 65e17b3..607cea3 100644
--- a/xmllint.c
+++ b/xmllint.c
@@ -196,7 +196,7 @@
static xmlPatternPtr patternc = NULL;
static xmlStreamCtxtPtr patstream = NULL;
#endif
-static int options = 0;
+static int options = XML_PARSE_COMPACT;
static int sax = 0;
/************************************************************************
@@ -2793,6 +2793,7 @@
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--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--nowrap : do not put HTML doc wrapper\n");
#ifdef LIBXML_VALID_ENABLED
@@ -3175,6 +3176,9 @@
} else if ((!strcmp(argv[i], "-nonet")) ||
(!strcmp(argv[i], "--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")) ||
(!strcmp(argv[i], "--load-trace"))) {
load_trace++;
diff --git a/xmlreader.c b/xmlreader.c
index 5fdf3a5..7efa273 100644
--- a/xmlreader.c
+++ b/xmlreader.c
@@ -348,7 +348,8 @@
(cur->type == XML_XINCLUDE_END)) &&
(cur->properties != NULL))
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_END) &&
(cur->type != XML_ENTITY_REF_NODE)) {
@@ -422,7 +423,8 @@
(cur->type == XML_XINCLUDE_END)) &&
(cur->properties != NULL))
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_END) &&
(cur->type != XML_ENTITY_REF_NODE)) {
@@ -2810,7 +2812,9 @@
reader->faketext = xmlNewDocText(reader->node->doc,
ns->href);
} else {
- if (reader->faketext->content != NULL)
+ if ((reader->faketext->content != NULL) &&
+ (reader->faketext->content !=
+ (xmlChar *) &(reader->faketext->properties)))
xmlFree(reader->faketext->content);
reader->faketext->content = xmlStrdup(ns->href);
}
@@ -4776,6 +4780,12 @@
if (reader == NULL)
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->entNr = 0;
reader->parserFlags = options;