1
0
mirror of https://gitlab.gnome.org/GNOME/libxml2.git synced 2024-10-26 03:55:04 +03:00

warning patches from Peter Breitenlohner serious DTD parsing speedups,

* testdso.c xmlregexp.c: warning patches from Peter Breitenlohner
* include/libxml/valid.h valid.c parser.c: serious DTD parsing
  speedups, start to deprecate 3 ElementDef related entry point
  and replace them with better ones.
Daniel
This commit is contained in:
Daniel Veillard 2005-01-25 00:22:52 +00:00
parent b9e2f26ed6
commit cee2b3a5f1
6 changed files with 293 additions and 95 deletions

View File

@ -1,3 +1,10 @@
Tue Jan 25 01:20:11 CET 2005 Daniel Veillard <daniel@veillard.com>
* testdso.c xmlregexp.c: warning patches from Peter Breitenlohner
* include/libxml/valid.h valid.c parser.c: serious DTD parsing
speedups, start to deprecate 3 ElementDef related entry point
and replace them with better ones.
Mon Jan 24 00:47:41 CET 2005 Daniel Veillard <daniel@veillard.com>
* xmlschemas.c: more hash dictionary interning changes

View File

@ -172,6 +172,7 @@ XMLPUBFUN void XMLCALL
#endif /* LIBXML_OUTPUT_ENABLED */
/* Element Content */
/* the non Doc version are being deprecated */
XMLPUBFUN xmlElementContentPtr XMLCALL
xmlNewElementContent (const xmlChar *name,
xmlElementContentType type);
@ -179,6 +180,17 @@ XMLPUBFUN xmlElementContentPtr XMLCALL
xmlCopyElementContent (xmlElementContentPtr content);
XMLPUBFUN void XMLCALL
xmlFreeElementContent (xmlElementContentPtr cur);
/* the new versions with doc argument */
XMLPUBFUN xmlElementContentPtr XMLCALL
xmlNewDocElementContent (xmlDocPtr doc,
const xmlChar *name,
xmlElementContentType type);
XMLPUBFUN xmlElementContentPtr XMLCALL
xmlCopyDocElementContent(xmlDocPtr doc,
xmlElementContentPtr content);
XMLPUBFUN void XMLCALL
xmlFreeDocElementContent(xmlDocPtr doc,
xmlElementContentPtr cur);
XMLPUBFUN void XMLCALL
xmlSnprintfElementContent(char *buf,
int size,

View File

@ -4739,7 +4739,7 @@ xmlParseElementMixedContentDecl(xmlParserCtxtPtr ctxt, int inputchk) {
NULL);
}
NEXT;
ret = xmlNewElementContent(NULL, XML_ELEMENT_CONTENT_PCDATA);
ret = xmlNewDocElementContent(ctxt->myDoc, NULL, XML_ELEMENT_CONTENT_PCDATA);
if (RAW == '*') {
ret->ocur = XML_ELEMENT_CONTENT_MULT;
NEXT;
@ -4747,22 +4747,22 @@ xmlParseElementMixedContentDecl(xmlParserCtxtPtr ctxt, int inputchk) {
return(ret);
}
if ((RAW == '(') || (RAW == '|')) {
ret = cur = xmlNewElementContent(NULL, XML_ELEMENT_CONTENT_PCDATA);
ret = cur = xmlNewDocElementContent(ctxt->myDoc, NULL, XML_ELEMENT_CONTENT_PCDATA);
if (ret == NULL) return(NULL);
}
while (RAW == '|') {
NEXT;
if (elem == NULL) {
ret = xmlNewElementContent(NULL, XML_ELEMENT_CONTENT_OR);
ret = xmlNewDocElementContent(ctxt->myDoc, NULL, XML_ELEMENT_CONTENT_OR);
if (ret == NULL) return(NULL);
ret->c1 = cur;
if (cur != NULL)
cur->parent = ret;
cur = ret;
} else {
n = xmlNewElementContent(NULL, XML_ELEMENT_CONTENT_OR);
n = xmlNewDocElementContent(ctxt->myDoc, NULL, XML_ELEMENT_CONTENT_OR);
if (n == NULL) return(NULL);
n->c1 = xmlNewElementContent(elem, XML_ELEMENT_CONTENT_ELEMENT);
n->c1 = xmlNewDocElementContent(ctxt->myDoc, elem, XML_ELEMENT_CONTENT_ELEMENT);
if (n->c1 != NULL)
n->c1->parent = n;
cur->c2 = n;
@ -4775,7 +4775,7 @@ xmlParseElementMixedContentDecl(xmlParserCtxtPtr ctxt, int inputchk) {
if (elem == NULL) {
xmlFatalErrMsg(ctxt, XML_ERR_NAME_REQUIRED,
"xmlParseElementMixedContentDecl : Name expected\n");
xmlFreeElementContent(cur);
xmlFreeDocElementContent(ctxt->myDoc, cur);
return(NULL);
}
SKIP_BLANKS;
@ -4783,7 +4783,7 @@ xmlParseElementMixedContentDecl(xmlParserCtxtPtr ctxt, int inputchk) {
}
if ((RAW == ')') && (NXT(1) == '*')) {
if (elem != NULL) {
cur->c2 = xmlNewElementContent(elem,
cur->c2 = xmlNewDocElementContent(ctxt->myDoc, elem,
XML_ELEMENT_CONTENT_ELEMENT);
if (cur->c2 != NULL)
cur->c2->parent = cur;
@ -4796,7 +4796,7 @@ xmlParseElementMixedContentDecl(xmlParserCtxtPtr ctxt, int inputchk) {
}
SKIP(2);
} else {
xmlFreeElementContent(ret);
xmlFreeDocElementContent(ctxt->myDoc, ret);
xmlFatalErr(ctxt, XML_ERR_MIXED_NOT_STARTED, NULL);
return(NULL);
}
@ -4861,7 +4861,7 @@ xmlParseElementChildrenContentDecl (xmlParserCtxtPtr ctxt, int inputchk) {
xmlFatalErr(ctxt, XML_ERR_ELEMCONTENT_NOT_STARTED, NULL);
return(NULL);
}
cur = ret = xmlNewElementContent(elem, XML_ELEMENT_CONTENT_ELEMENT);
cur = ret = xmlNewDocElementContent(ctxt->myDoc, elem, XML_ELEMENT_CONTENT_ELEMENT);
if (cur == NULL) {
xmlErrMemory(ctxt, NULL);
return(NULL);
@ -4898,18 +4898,18 @@ xmlParseElementChildrenContentDecl (xmlParserCtxtPtr ctxt, int inputchk) {
"xmlParseElementChildrenContentDecl : '%c' expected\n",
type);
if ((last != NULL) && (last != ret))
xmlFreeElementContent(last);
xmlFreeDocElementContent(ctxt->myDoc, last);
if (ret != NULL)
xmlFreeElementContent(ret);
xmlFreeDocElementContent(ctxt->myDoc, ret);
return(NULL);
}
NEXT;
op = xmlNewElementContent(NULL, XML_ELEMENT_CONTENT_SEQ);
op = xmlNewDocElementContent(ctxt->myDoc, NULL, XML_ELEMENT_CONTENT_SEQ);
if (op == NULL) {
if ((last != NULL) && (last != ret))
xmlFreeElementContent(last);
xmlFreeElementContent(ret);
xmlFreeDocElementContent(ctxt->myDoc, last);
xmlFreeDocElementContent(ctxt->myDoc, ret);
return(NULL);
}
if (last == NULL) {
@ -4938,19 +4938,19 @@ xmlParseElementChildrenContentDecl (xmlParserCtxtPtr ctxt, int inputchk) {
"xmlParseElementChildrenContentDecl : '%c' expected\n",
type);
if ((last != NULL) && (last != ret))
xmlFreeElementContent(last);
xmlFreeDocElementContent(ctxt->myDoc, last);
if (ret != NULL)
xmlFreeElementContent(ret);
xmlFreeDocElementContent(ctxt->myDoc, ret);
return(NULL);
}
NEXT;
op = xmlNewElementContent(NULL, XML_ELEMENT_CONTENT_OR);
op = xmlNewDocElementContent(ctxt->myDoc, NULL, XML_ELEMENT_CONTENT_OR);
if (op == NULL) {
if ((last != NULL) && (last != ret))
xmlFreeElementContent(last);
xmlFreeDocElementContent(ctxt->myDoc, last);
if (ret != NULL)
xmlFreeElementContent(ret);
xmlFreeDocElementContent(ctxt->myDoc, ret);
return(NULL);
}
if (last == NULL) {
@ -4971,7 +4971,7 @@ xmlParseElementChildrenContentDecl (xmlParserCtxtPtr ctxt, int inputchk) {
} else {
xmlFatalErr(ctxt, XML_ERR_ELEMCONTENT_NOT_FINISHED, NULL);
if (ret != NULL)
xmlFreeElementContent(ret);
xmlFreeDocElementContent(ctxt->myDoc, ret);
return(NULL);
}
GROW;
@ -4989,10 +4989,10 @@ xmlParseElementChildrenContentDecl (xmlParserCtxtPtr ctxt, int inputchk) {
if (elem == NULL) {
xmlFatalErr(ctxt, XML_ERR_ELEMCONTENT_NOT_STARTED, NULL);
if (ret != NULL)
xmlFreeElementContent(ret);
xmlFreeDocElementContent(ctxt->myDoc, ret);
return(NULL);
}
last = xmlNewElementContent(elem, XML_ELEMENT_CONTENT_ELEMENT);
last = xmlNewDocElementContent(ctxt->myDoc, elem, XML_ELEMENT_CONTENT_ELEMENT);
if (RAW == '?') {
last->ocur = XML_ELEMENT_CONTENT_OPT;
NEXT;
@ -5213,6 +5213,9 @@ xmlParseElementDecl(xmlParserCtxtPtr ctxt) {
if (RAW != '>') {
xmlFatalErr(ctxt, XML_ERR_GT_REQUIRED, NULL);
if (content != NULL) {
xmlFreeDocElementContent(ctxt->myDoc, content);
}
} else {
if (input != ctxt->input) {
xmlFatalErrMsg(ctxt, XML_ERR_ENTITY_BOUNDARY,
@ -5221,12 +5224,23 @@ xmlParseElementDecl(xmlParserCtxtPtr ctxt) {
NEXT;
if ((ctxt->sax != NULL) && (!ctxt->disableSAX) &&
(ctxt->sax->elementDecl != NULL))
(ctxt->sax->elementDecl != NULL)) {
if (content != NULL)
content->parent = NULL;
ctxt->sax->elementDecl(ctxt->userData, name, ret,
content);
if ((content != NULL) && (content->parent == NULL)) {
/*
* this is a trick: if xmlAddElementDecl is called,
* instead of copying the full tree it is plugged directly
* if called from the parser. Avoid duplicating the
* interfaces or change the API/ABI
*/
xmlFreeDocElementContent(ctxt->myDoc, content);
}
} else if (content != NULL) {
xmlFreeDocElementContent(ctxt->myDoc, content);
}
if (content != NULL) {
xmlFreeElementContent(content);
}
}
return(ret);

View File

@ -3,7 +3,9 @@
#define IN_LIBXML
#include "libxml/xmlexports.h"
XMLPUBFUN int hello_world(void)
XMLPUBFUN int hello_world(void);
int hello_world(void)
{
printf("Success!\n");
return 0;

245
valid.c
View File

@ -895,17 +895,23 @@ xmlFreeValidCtxt(xmlValidCtxtPtr cur) {
#endif /* LIBXML_VALID_ENABLED */
/**
* xmlNewElementContent:
* xmlNewDocElementContent:
* @doc: the document
* @name: the subelement name or NULL
* @type: the type of element content decl
*
* Allocate an element content structure.
* Allocate an element content structure for the document.
*
* Returns NULL if not, otherwise the new element content structure
*/
xmlElementContentPtr
xmlNewElementContent(const xmlChar *name, xmlElementContentType type) {
xmlNewDocElementContent(xmlDocPtr doc, const xmlChar *name,
xmlElementContentType type) {
xmlElementContentPtr ret;
xmlDictPtr dict = NULL;
if (doc != NULL)
dict = doc->dict;
switch(type) {
case XML_ELEMENT_CONTENT_ELEMENT:
@ -939,16 +945,121 @@ xmlNewElementContent(const xmlChar *name, xmlElementContentType type) {
ret->type = type;
ret->ocur = XML_ELEMENT_CONTENT_ONCE;
if (name != NULL) {
xmlChar *prefix = NULL;
ret->name = xmlSplitQName2(name, &prefix);
if (ret->name == NULL)
int l;
const xmlChar *tmp;
tmp = xmlSplitQName3(name, &l);
if (tmp == NULL) {
if (dict == NULL)
ret->name = xmlStrdup(name);
ret->prefix = prefix;
else
ret->name = xmlDictLookup(dict, name, -1);
} else {
ret->name = NULL;
ret->prefix = NULL;
if (dict == NULL) {
ret->prefix = xmlStrndup(name, l);
ret->name = xmlStrdup(tmp);
} else {
ret->prefix = xmlDictLookup(dict, name, l);
ret->name = xmlDictLookup(dict, tmp, -1);
}
}
}
return(ret);
}
/**
* xmlNewElementContent:
* @name: the subelement name or NULL
* @type: the type of element content decl
*
* Allocate an element content structure.
* Deprecated in favor of xmlNewDocElementContent
*
* Returns NULL if not, otherwise the new element content structure
*/
xmlElementContentPtr
xmlNewElementContent(const xmlChar *name, xmlElementContentType type) {
return(xmlNewDocElementContent(NULL, name, type));
}
/**
* xmlCopyDocElementContent:
* @doc: the document owning the element declaration
* @cur: An element content pointer.
*
* Build a copy of an element content description.
*
* Returns the new xmlElementContentPtr or NULL in case of error.
*/
xmlElementContentPtr
xmlCopyDocElementContent(xmlDocPtr doc, xmlElementContentPtr cur) {
xmlElementContentPtr ret = NULL, prev = NULL, tmp;
xmlDictPtr dict = NULL;
if (cur == NULL) return(NULL);
if (doc != NULL)
dict = doc->dict;
ret = (xmlElementContentPtr) xmlMalloc(sizeof(xmlElementContent));
if (ret == NULL) {
xmlVErrMemory(NULL, "malloc failed");
return(NULL);
}
memset(ret, 0, sizeof(xmlElementContent));
ret->type = cur->type;
ret->ocur = cur->ocur;
if (cur->name != NULL) {
if (dict)
ret->name = xmlDictLookup(dict, cur->name, -1);
else
ret->name = xmlStrdup(cur->name);
}
if (cur->prefix != NULL) {
if (dict)
ret->prefix = xmlDictLookup(dict, cur->prefix, -1);
else
ret->prefix = xmlStrdup(cur->prefix);
}
if (cur->c1 != NULL)
ret->c1 = xmlCopyDocElementContent(doc, cur->c1);
if (ret->c1 != NULL)
ret->c1->parent = ret;
if (cur->c2 != NULL) {
prev = ret;
cur = cur->c2;
while (cur != NULL) {
tmp = (xmlElementContentPtr) xmlMalloc(sizeof(xmlElementContent));
if (tmp == NULL) {
xmlVErrMemory(NULL, "malloc failed");
return(ret);
}
memset(tmp, 0, sizeof(xmlElementContent));
tmp->type = cur->type;
tmp->ocur = cur->ocur;
prev->c2 = tmp;
if (cur->name != NULL) {
if (dict)
tmp->name = xmlDictLookup(dict, cur->name, -1);
else
tmp->name = xmlStrdup(cur->name);
}
if (cur->prefix != NULL) {
if (dict)
tmp->prefix = xmlDictLookup(dict, cur->prefix, -1);
else
tmp->prefix = xmlStrdup(cur->prefix);
}
if (cur->c1 != NULL)
tmp->c1 = xmlCopyDocElementContent(doc,cur->c1);
if (tmp->c1 != NULL)
tmp->c1->parent = ret;
prev = tmp;
cur = cur->c2;
}
}
ret->c1 = ret->c2 = ret->parent = NULL;
return(ret);
}
@ -957,40 +1068,32 @@ xmlNewElementContent(const xmlChar *name, xmlElementContentType type) {
* @cur: An element content pointer.
*
* Build a copy of an element content description.
* Deprecated, use xmlCopyDocElementContent instead
*
* Returns the new xmlElementContentPtr or NULL in case of error.
*/
xmlElementContentPtr
xmlCopyElementContent(xmlElementContentPtr cur) {
xmlElementContentPtr ret;
if (cur == NULL) return(NULL);
ret = xmlNewElementContent((xmlChar *) cur->name, cur->type);
if (ret == NULL) {
xmlVErrMemory(NULL, "malloc failed");
return(NULL);
}
if (cur->prefix != NULL)
ret->prefix = xmlStrdup(cur->prefix);
ret->ocur = cur->ocur;
if (cur->c1 != NULL) ret->c1 = xmlCopyElementContent(cur->c1);
if (ret->c1 != NULL)
ret->c1->parent = ret;
if (cur->c2 != NULL) ret->c2 = xmlCopyElementContent(cur->c2);
if (ret->c2 != NULL)
ret->c2->parent = ret;
return(ret);
return(xmlCopyDocElementContent(NULL, cur));
}
/**
* xmlFreeElementContent:
* xmlFreeDocElementContent:
* @doc: the document owning the element declaration
* @cur: the element content tree to free
*
* Free an element content structure. This is a recursive call !
* Free an element content structure. The whole subtree is removed.
*/
void
xmlFreeElementContent(xmlElementContentPtr cur) {
if (cur == NULL) return;
xmlFreeDocElementContent(xmlDocPtr doc, xmlElementContentPtr cur) {
xmlElementContentPtr next;
xmlDictPtr dict = NULL;
if (doc != NULL)
dict = doc->dict;
while (cur != NULL) {
next = cur->c2;
switch (cur->type) {
case XML_ELEMENT_CONTENT_PCDATA:
case XML_ELEMENT_CONTENT_ELEMENT:
@ -1003,11 +1106,31 @@ xmlFreeElementContent(xmlElementContentPtr cur) {
NULL);
return;
}
if (cur->c1 != NULL) xmlFreeElementContent(cur->c1);
if (cur->c2 != NULL) xmlFreeElementContent(cur->c2);
if (cur->c1 != NULL) xmlFreeDocElementContent(doc, cur->c1);
if (dict) {
if ((cur->name != NULL) && (!xmlDictOwns(dict, cur->name)))
xmlFree((xmlChar *) cur->name);
if ((cur->prefix != NULL) && (!xmlDictOwns(dict, cur->prefix)))
xmlFree((xmlChar *) cur->prefix);
} else {
if (cur->name != NULL) xmlFree((xmlChar *) cur->name);
if (cur->prefix != NULL) xmlFree((xmlChar *) cur->prefix);
}
xmlFree(cur);
cur = next;
}
}
/**
* xmlFreeElementContent:
* @cur: the element content tree to free
*
* Free an element content structure. The whole subtree is removed.
* Deprecated, use xmlFreeDocElementContent instead
*/
void
xmlFreeElementContent(xmlElementContentPtr cur) {
xmlFreeDocElementContent(NULL, cur);
}
#ifdef LIBXML_OUTPUT_ENABLED
@ -1216,7 +1339,7 @@ static void
xmlFreeElement(xmlElementPtr elem) {
if (elem == NULL) return;
xmlUnlinkNode((xmlNodePtr) elem);
xmlFreeElementContent(elem->content);
xmlFreeDocElementContent(elem->doc, elem->content);
if (elem->name != NULL)
xmlFree((xmlChar *) elem->name);
if (elem->prefix != NULL)
@ -1419,7 +1542,19 @@ xmlAddElementDecl(xmlValidCtxtPtr ctxt,
* Finish to fill the structure.
*/
ret->etype = type;
ret->content = xmlCopyElementContent(content);
/*
* Avoid a stupid copy when called by the parser
* and flag it by setting a special parent value
* so the parser doesn't unallocate it.
*/
if ((ctxt->finishDtd == XML_CTXT_FINISH_DTD_0) ||
(ctxt->finishDtd == XML_CTXT_FINISH_DTD_1)) {
ret->content = content;
if (content != NULL)
content->parent = (xmlElementContentPtr) 1;
} else {
ret->content = xmlCopyDocElementContent(dtd->doc, content);
}
/*
* Link it to the DTD
@ -1763,10 +1898,27 @@ xmlScanIDAttributeDecl(xmlValidCtxtPtr ctxt, xmlElementPtr elem) {
*/
static void
xmlFreeAttribute(xmlAttributePtr attr) {
xmlDictPtr dict;
if (attr == NULL) return;
if (attr->doc != NULL)
dict = attr->doc->dict;
else
dict = NULL;
xmlUnlinkNode((xmlNodePtr) attr);
if (attr->tree != NULL)
xmlFreeEnumeration(attr->tree);
if (dict) {
if ((attr->elem != NULL) && (!xmlDictOwns(dict, attr->elem)))
xmlFree((xmlChar *) attr->elem);
if ((attr->name != NULL) && (!xmlDictOwns(dict, attr->name)))
xmlFree((xmlChar *) attr->name);
if ((attr->prefix != NULL) && (!xmlDictOwns(dict, attr->prefix)))
xmlFree((xmlChar *) attr->prefix);
if ((attr->defaultValue != NULL) &&
(!xmlDictOwns(dict, attr->defaultValue)))
xmlFree((xmlChar *) attr->defaultValue);
} else {
if (attr->elem != NULL)
xmlFree((xmlChar *) attr->elem);
if (attr->name != NULL)
@ -1775,6 +1927,7 @@ xmlFreeAttribute(xmlAttributePtr attr) {
xmlFree((xmlChar *) attr->defaultValue);
if (attr->prefix != NULL)
xmlFree((xmlChar *) attr->prefix);
}
xmlFree(attr);
}
@ -1805,6 +1958,7 @@ xmlAddAttributeDecl(xmlValidCtxtPtr ctxt,
xmlAttributePtr ret;
xmlAttributeTablePtr table;
xmlElementPtr elemDef;
xmlDictPtr dict = NULL;
if (dtd == NULL) {
xmlFreeEnumeration(tree);
@ -1818,6 +1972,8 @@ xmlAddAttributeDecl(xmlValidCtxtPtr ctxt,
xmlFreeEnumeration(tree);
return(NULL);
}
if (dtd->doc != NULL)
dict = dtd->doc->dict;
#ifdef LIBXML_VALID_ENABLED
/*
@ -1879,9 +2035,6 @@ xmlAddAttributeDecl(xmlValidCtxtPtr ctxt,
*/
table = (xmlAttributeTablePtr) dtd->attributes;
if (table == NULL) {
xmlDictPtr dict = NULL;
if (dtd->doc != NULL)
dict = dtd->doc->dict;
table = xmlHashCreateDict(0, dict);
dtd->attributes = (void *) table;
}
@ -1904,19 +2057,29 @@ xmlAddAttributeDecl(xmlValidCtxtPtr ctxt,
* fill the structure.
*/
ret->atype = type;
if (dict) {
ret->name = xmlDictLookup(dict, name, -1);
ret->prefix = xmlDictLookup(dict, ns, -1);
ret->elem = xmlDictLookup(dict, elem, -1);
} else {
ret->name = xmlStrdup(name);
ret->prefix = xmlStrdup(ns);
ret->elem = xmlStrdup(elem);
}
ret->def = def;
ret->tree = tree;
if (defaultValue != NULL)
if (defaultValue != NULL) {
if (dict)
ret->defaultValue = xmlDictLookup(dict, defaultValue, -1);
else
ret->defaultValue = xmlStrdup(defaultValue);
}
/*
* Validity Check:
* Search the DTD for previous declarations of the ATTLIST
*/
if (xmlHashAddEntry3(table, name, ns, elem, ret) < 0) {
if (xmlHashAddEntry3(table, ret->name, ret->prefix, ret->elem, ret) < 0) {
#ifdef LIBXML_VALID_ENABLED
/*
* The attribute is already defined in this DTD.

View File

@ -19,7 +19,7 @@
#ifdef LIBXML_REGEXP_ENABLED
#define DEBUG_ERR
/* #define DEBUG_ERR */
#include <stdio.h>
#include <string.h>
@ -3343,7 +3343,7 @@ xmlRegExecErrInfo(xmlRegExecCtxtPtr exec, const xmlChar **string,
#ifdef DEBUG_ERR
static void testerr(xmlRegExecCtxtPtr exec) {
const xmlChar *string;
const xmlChar *values[5];
xmlChar *values[5];
int nb = 5;
int nbneg;
int terminal;