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

HTML parsing, output is now correct, added HTMLtests target and testcases, Daniel

Browse Source
This commit is contained in:
Daniel Veillard
1999-07-07 07:32:15 +00:00
parent 167b5092b8
commit 82150d8a99
15 changed files with 380 additions and 81 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
ChangeLog
View File

@ -1,3 +1,9 @@
Wed Jul 7 09:28:43 CEST 1999 Daniel Veillard <Daniel.Veillard@w3.org>
* HTMLparser.[ch], HTMLtree.[ch]: more work for HTML parsing and
output.
* Makefile.am, test/HTML/*, result/HTML/*: added HTMLtests targetestHTMLt
Wed Jul 7 00:25:42 CEST 1999 Daniel Veillard <Daniel.Veillard@w3.org> Wed Jul 7 00:25:42 CEST 1999 Daniel Veillard <Daniel.Veillard@w3.org>
* parser.h : Oops removed the binary compatibility problem * parser.h : Oops removed the binary compatibility problem

8
HTMLparser.c
View File

@ -34,7 +34,7 @@
#include "valid.h" #include "valid.h"
#include "parserInternals.h" #include "parserInternals.h"
#define DEBUG */ /* #define DEBUG */
/************************************************************************ /************************************************************************
* * * *
@ -351,7 +351,6 @@ htmlInitAutoClose(void) {
htmlElemDescPtr htmlElemDescPtr
htmlTagLookup(const CHAR *tag) { htmlTagLookup(const CHAR *tag) {
int i = 0; int i = 0;
int cnt;
for (i = 0; i < (sizeof(html40ElementTable) / for (i = 0; i < (sizeof(html40ElementTable) /
sizeof(html40ElementTable[0]));i++) { sizeof(html40ElementTable[0]));i++) {
@ -408,7 +407,6 @@ htmlCheckAutoClose(const CHAR *new, const CHAR *old) {
*/ */
void void
htmlAutoClose(htmlParserCtxtPtr ctxt, const CHAR *new) { htmlAutoClose(htmlParserCtxtPtr ctxt, const CHAR *new) {
const CHAR *old;
while ((ctxt->node != NULL) && while ((ctxt->node != NULL) &&
(htmlCheckAutoClose(new, ctxt->node->name))) { (htmlCheckAutoClose(new, ctxt->node->name))) {
@ -1933,7 +1931,7 @@ htmlParseStartTag(htmlParserCtxtPtr ctxt) {
atts = (const CHAR **) malloc(maxatts * sizeof(CHAR *)); atts = (const CHAR **) malloc(maxatts * sizeof(CHAR *));
if (atts == NULL) { if (atts == NULL) {
fprintf(stderr, "malloc of %ld byte failed\n", fprintf(stderr, "malloc of %ld byte failed\n",
maxatts * sizeof(CHAR *)); maxatts * (long)sizeof(CHAR *));
return(NULL); return(NULL);
} }
} else if (nbatts + 2 < maxatts) { } else if (nbatts + 2 < maxatts) {
@ -1941,7 +1939,7 @@ htmlParseStartTag(htmlParserCtxtPtr ctxt) {
atts = (const CHAR **) realloc(atts, maxatts * sizeof(CHAR *)); atts = (const CHAR **) realloc(atts, maxatts * sizeof(CHAR *));
if (atts == NULL) { if (atts == NULL) {
fprintf(stderr, "realloc of %ld byte failed\n", fprintf(stderr, "realloc of %ld byte failed\n",
maxatts * sizeof(CHAR *)); maxatts * (long)sizeof(CHAR *));
return(NULL); return(NULL);
} }
} }

6
HTMLparser.h
View File

@ -48,8 +48,10 @@ typedef struct htmlEntityDesc {
/* /*
* There is only few public functions. * There is only few public functions.
*/ */
htmlEntityDescPtr htmlElemDescPtr htmlTagLookup(const CHAR *tag);
htmlParseEntityRef(htmlParserCtxtPtr ctxt, CHAR **str); htmlEntityDescPtr htmlEntityLookup(const CHAR *name);
htmlEntityDescPtr htmlParseEntityRef(htmlParserCtxtPtr ctxt, CHAR **str);
int htmlParseCharRef(htmlParserCtxtPtr ctxt); int htmlParseCharRef(htmlParserCtxtPtr ctxt);
void htmlParseElement(htmlParserCtxtPtr ctxt); void htmlParseElement(htmlParserCtxtPtr ctxt);

120
HTMLtree.c
View File

@ -12,14 +12,11 @@
#include <stdlib.h> #include <stdlib.h>
#include <string.h> /* for memset() only ! */ #include <string.h> /* for memset() only ! */
#include "tree.h" #include "HTMLparser.h"
#include "HTMLtree.h"
#include "entities.h" #include "entities.h"
#include "valid.h" #include "valid.h"
#define HTML_TEXT_NODE XML_TEXT_NODE
#define HTML_ENTITY_REF_NODE XML_ENTITY_REF_NODE
#define HTML_COMMENT_NODE XML_COMMENT_NODE
/** /**
* htmlDtdDump: * htmlDtdDump:
* @buf: the HTML buffer output * @buf: the HTML buffer output
@ -46,23 +43,6 @@ htmlDtdDump(xmlBufferPtr buf, xmlDocPtr doc) {
xmlBufferWriteChar(buf, " SYSTEM "); xmlBufferWriteChar(buf, " SYSTEM ");
xmlBufferWriteQuotedString(buf, cur->SystemID); xmlBufferWriteQuotedString(buf, cur->SystemID);
} }
if ((cur->entities == NULL) && (cur->elements == NULL) &&
(cur->attributes == NULL) && (cur->notations == NULL)) {
xmlBufferWriteChar(buf, ">\n");
return;
}
xmlBufferWriteChar(buf, " [\n");
if (cur->entities != NULL)
xmlDumpEntitiesTable(buf, (xmlEntitiesTablePtr) cur->entities);
if (cur->notations != NULL)
xmlDumpNotationTable(buf, (xmlNotationTablePtr) cur->notations);
if (cur->elements != NULL)
xmlDumpElementTable(buf, (xmlElementTablePtr) cur->elements);
if (cur->attributes != NULL)
xmlDumpAttributeTable(buf, (xmlAttributeTablePtr) cur->attributes);
xmlBufferWriteChar(buf, "]");
/* TODO !!! a lot more things to dump ... */
xmlBufferWriteChar(buf, ">\n"); xmlBufferWriteChar(buf, ">\n");
} }
@ -116,30 +96,23 @@ htmlAttrListDump(xmlBufferPtr buf, xmlDocPtr doc, xmlAttrPtr cur) {
static void static void
htmlNodeDump(xmlBufferPtr buf, xmlDocPtr doc, xmlNodePtr cur, int level); htmlNodeDump(xmlBufferPtr buf, xmlDocPtr doc, xmlNodePtr cur);
/** /**
* htmlNodeListDump: * htmlNodeListDump:
* @buf: the HTML buffer output * @buf: the HTML buffer output
* @doc: the document * @doc: the document
* @cur: the first node * @cur: the first node
* @level: the imbrication level for indenting
* *
* Dump an HTML node list, recursive behaviour,children are printed too. * Dump an HTML node list, recursive behaviour,children are printed too.
*/ */
static void static void
htmlNodeListDump(xmlBufferPtr buf, xmlDocPtr doc, xmlNodePtr cur, int level) { htmlNodeListDump(xmlBufferPtr buf, xmlDocPtr doc, xmlNodePtr cur) {
int i;
if (cur == NULL) { if (cur == NULL) {
fprintf(stderr, "htmlNodeListDump : node == NULL\n"); fprintf(stderr, "htmlNodeListDump : node == NULL\n");
return; return;
} }
while (cur != NULL) { while (cur != NULL) {
if ((cur->type != HTML_TEXT_NODE) && htmlNodeDump(buf, doc, cur);
(cur->type != HTML_ENTITY_REF_NODE)) {
xmlBufferWriteChar(buf, "\n");
}
htmlNodeDump(buf, doc, cur, level);
cur = cur->next; cur = cur->next;
} }
} }
@ -149,22 +122,26 @@ htmlNodeListDump(xmlBufferPtr buf, xmlDocPtr doc, xmlNodePtr cur, int level) {
* @buf: the HTML buffer output * @buf: the HTML buffer output
* @doc: the document * @doc: the document
* @cur: the current node * @cur: the current node
* @level: the imbrication level for indenting
* *
* Dump an HTML node, recursive behaviour,children are printed too. * Dump an HTML node, recursive behaviour,children are printed too.
*/ */
static void static void
htmlNodeDump(xmlBufferPtr buf, xmlDocPtr doc, xmlNodePtr cur, int level) { htmlNodeDump(xmlBufferPtr buf, xmlDocPtr doc, xmlNodePtr cur) {
int i; int i;
htmlElemDescPtr info;
if (cur == NULL) { if (cur == NULL) {
fprintf(stderr, "htmlNodeDump : node == NULL\n"); fprintf(stderr, "htmlNodeDump : node == NULL\n");
return; return;
} }
/*
* Special cases.
*/
if (cur->type == HTML_TEXT_NODE) { if (cur->type == HTML_TEXT_NODE) {
if (cur->content != NULL) { if (cur->content != NULL) {
CHAR *buffer; CHAR *buffer;
/* uses the HTML encoding routine !!!!!!!!!! */
buffer = xmlEncodeEntitiesReentrant(doc, cur->content); buffer = xmlEncodeEntitiesReentrant(doc, cur->content);
if (buffer != NULL) { if (buffer != NULL) {
xmlBufferWriteCHAR(buf, buffer); xmlBufferWriteCHAR(buf, buffer);
@ -188,20 +165,38 @@ htmlNodeDump(xmlBufferPtr buf, xmlDocPtr doc, xmlNodePtr cur, int level) {
return; return;
} }
xmlBufferWriteChar(buf, "<"); /*
if ((cur->ns != NULL) && (cur->ns->prefix != NULL)) { * Get specific HTmL info for taht node.
xmlBufferWriteCHAR(buf, cur->ns->prefix); */
xmlBufferWriteChar(buf, ":"); info = htmlTagLookup(cur->name);
}
xmlBufferWriteChar(buf, "<");
xmlBufferWriteCHAR(buf, cur->name); xmlBufferWriteCHAR(buf, cur->name);
if (cur->nsDef)
xmlNsListDump(buf, cur->nsDef);
if (cur->properties != NULL) if (cur->properties != NULL)
htmlAttrListDump(buf, doc, cur->properties); htmlAttrListDump(buf, doc, cur->properties);
if (info->empty) {
xmlBufferWriteChar(buf, ">");
if (cur->next != NULL) {
if ((cur->next->type != HTML_TEXT_NODE) &&
(cur->next->type != HTML_ENTITY_REF_NODE))
xmlBufferWriteChar(buf, "\n");
}
return;
}
if ((cur->content == NULL) && (cur->childs == NULL)) { if ((cur->content == NULL) && (cur->childs == NULL)) {
xmlBufferWriteChar(buf, "/>\n"); if (info->endTag != 0)
xmlBufferWriteChar(buf, ">");
else {
xmlBufferWriteChar(buf, "></");
xmlBufferWriteCHAR(buf, cur->name);
xmlBufferWriteChar(buf, ">");
}
if (cur->next != NULL) {
if ((cur->next->type != HTML_TEXT_NODE) &&
(cur->next->type != HTML_ENTITY_REF_NODE))
xmlBufferWriteChar(buf, "\n");
}
return; return;
} }
xmlBufferWriteChar(buf, ">"); xmlBufferWriteChar(buf, ">");
@ -215,16 +210,22 @@ htmlNodeDump(xmlBufferPtr buf, xmlDocPtr doc, xmlNodePtr cur, int level) {
} }
} }
if (cur->childs != NULL) { if (cur->childs != NULL) {
htmlNodeListDump(buf, doc, cur->childs, level + 1); if ((cur->childs->type != HTML_TEXT_NODE) &&
(cur->childs->type != HTML_ENTITY_REF_NODE))
xmlBufferWriteChar(buf, "\n");
htmlNodeListDump(buf, doc, cur->childs);
if ((cur->last->type != HTML_TEXT_NODE) &&
(cur->last->type != HTML_ENTITY_REF_NODE))
xmlBufferWriteChar(buf, "\n");
} }
xmlBufferWriteChar(buf, "</"); xmlBufferWriteChar(buf, "</");
if ((cur->ns != NULL) && (cur->ns->prefix != NULL)) {
xmlBufferWriteCHAR(buf, cur->ns->prefix);
xmlBufferWriteChar(buf, ":");
}
xmlBufferWriteCHAR(buf, cur->name); xmlBufferWriteCHAR(buf, cur->name);
xmlBufferWriteChar(buf, ">\n"); xmlBufferWriteChar(buf, ">");
if (cur->next != NULL) {
if ((cur->next->type != HTML_TEXT_NODE) &&
(cur->next->type != HTML_ENTITY_REF_NODE))
xmlBufferWriteChar(buf, "\n");
}
} }
/** /**
@ -236,29 +237,12 @@ htmlNodeDump(xmlBufferPtr buf, xmlDocPtr doc, xmlNodePtr cur, int level) {
*/ */
static void static void
htmlDocContentDump(xmlBufferPtr buf, xmlDocPtr cur) { htmlDocContentDump(xmlBufferPtr buf, xmlDocPtr cur) {
xmlBufferWriteChar(buf, "<?xml version=");
if (cur->version != NULL)
xmlBufferWriteQuotedString(buf, cur->version);
else
xmlBufferWriteChar(buf, "\"1.0\"");
if (cur->encoding != NULL) {
xmlBufferWriteChar(buf, " encoding=");
xmlBufferWriteQuotedString(buf, cur->encoding);
}
switch (cur->standalone) {
case 0:
xmlBufferWriteChar(buf, " standalone=\"no\"");
break;
case 1:
xmlBufferWriteChar(buf, " standalone=\"yes\"");
break;
}
xmlBufferWriteChar(buf, "?>\n");
if (cur->intSubset != NULL) if (cur->intSubset != NULL)
htmlDtdDump(buf, cur); htmlDtdDump(buf, cur);
if (cur->root != NULL) { if (cur->root != NULL) {
htmlNodeDump(buf, cur, cur->root, 0); htmlNodeDump(buf, cur, cur->root);
} }
xmlBufferWriteChar(buf, "\n");
} }
/** /**

34
HTMLtree.h Normal file
View File

@ -0,0 +1,34 @@
/*
* tree.h : describes the structures found in an tree resulting
* from an XML parsing.
*
* See Copyright for the status of this software.
*
* Daniel.Veillard@w3.org
*/
#ifndef __HTML_TREE_H__
#define __HTML_TREE_H__
#ifdef __cplusplus
extern "C" {
#endif
#include <stdio.h>
#include "tree.h"
#define HTML_TEXT_NODE XML_TEXT_NODE
#define HTML_ENTITY_REF_NODE XML_ENTITY_REF_NODE
#define HTML_COMMENT_NODE XML_COMMENT_NODE
void htmlDocDumpMemory(xmlDocPtr cur, CHAR**mem, int *size);
void htmlDocDump(FILE *f, xmlDocPtr cur);
int htmlSaveFile(const char *filename, xmlDocPtr cur);
#ifdef __cplusplus
}
#endif
#endif /* __HTML_TREE_H__ */

19
Makefile.am
View File

@ -61,7 +61,24 @@ check-local: tests
testall : tests SVGtests SAXtests testall : tests SVGtests SAXtests
tests : tester tests: HTMLtests XMLtests
HTMLtests : testHTML
@(DIR=`pwd`; cd $(srcdir) ; \
for i in test/HTML/* ; do \
if [ ! -d $$i ] ; then \
if [ ! -f result/HTML/`basename $$i` ] ; then \
echo New test file `basename $$i` ; \
$$DIR/testHTML $$i > result/HTML/`basename $$i` ; \
else \
echo Testing `basename $$i` ; \
$$DIR/testHTML $$i > result.`basename $$i` ; \
diff result/HTML/`basename $$i` result.`basename $$i` ; \
$$DIR/testHTML result.`basename $$i` > result2.`basename $$i` ; \
diff result.`basename $$i` result2.`basename $$i` ; \
rm result.`basename $$i` result2.`basename $$i` ; \
fi ; fi ; done)
XMLtests : tester
@(DIR=`pwd`; cd $(srcdir) ; \ @(DIR=`pwd`; cd $(srcdir) ; \
for i in test/* ; do \ for i in test/* ; do \
if [ ! -d $$i ] ; then \ if [ ! -d $$i ] ; then \

6
include/libxml/HTMLparser.h
View File

@ -48,8 +48,10 @@ typedef struct htmlEntityDesc {
/* /*
* There is only few public functions. * There is only few public functions.
*/ */
htmlEntityDescPtr htmlElemDescPtr htmlTagLookup(const CHAR *tag);
htmlParseEntityRef(htmlParserCtxtPtr ctxt, CHAR **str); htmlEntityDescPtr htmlEntityLookup(const CHAR *name);
htmlEntityDescPtr htmlParseEntityRef(htmlParserCtxtPtr ctxt, CHAR **str);
int htmlParseCharRef(htmlParserCtxtPtr ctxt); int htmlParseCharRef(htmlParserCtxtPtr ctxt);
void htmlParseElement(htmlParserCtxtPtr ctxt); void htmlParseElement(htmlParserCtxtPtr ctxt);

34
include/libxml/HTMLtree.h Normal file
View File

@ -0,0 +1,34 @@
/*
* tree.h : describes the structures found in an tree resulting
* from an XML parsing.
*
* See Copyright for the status of this software.
*
* Daniel.Veillard@w3.org
*/
#ifndef __HTML_TREE_H__
#define __HTML_TREE_H__
#ifdef __cplusplus
extern "C" {
#endif
#include <stdio.h>
#include "tree.h"
#define HTML_TEXT_NODE XML_TEXT_NODE
#define HTML_ENTITY_REF_NODE XML_ENTITY_REF_NODE
#define HTML_COMMENT_NODE XML_COMMENT_NODE
void htmlDocDumpMemory(xmlDocPtr cur, CHAR**mem, int *size);
void htmlDocDump(FILE *f, xmlDocPtr cur);
int htmlSaveFile(const char *filename, xmlDocPtr cur);
#ifdef __cplusplus
}
#endif
#endif /* __HTML_TREE_H__ */

12
result/HTML/Down.html Normal file
View File

@ -0,0 +1,12 @@
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN" "http://www.w3.org/TR/REC-html40/loose.dtd">
<HTML>
<HEAD>
<TITLE>This service is temporary down</TITLE>
</HEAD>
<BODY bgcolor="#FFFFFF">
<H1 align="center">Sorry, this service is temporary down</H1>
We are doing our best to get it back on-line,
<P>The W3C system administrators</P>
</BODY>
</HTML>

41
result/HTML/test2.html Normal file
View File

@ -0,0 +1,41 @@
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN" "http://www.w3.org/TR/REC-html40/loose.dtd">
<HTML>
<HEAD>
<TITLE>Linux Today</TITLE>
</HEAD>
<BODY bgcolor="White" link="Blue" text="Black" VLINK="Black" ALINK="Red">
<CENTER>
<TABLE BORDER="0" WIDTH="100%" CELLSPACING="0" CELLPADDING="0">
<TR BGCOLOR="#FFFFFF">
<TD HEIGHT="90">
<A href="http://linuxtoday.com/cgi-bin/click.pl?adnum=49">
<IMG src="/pics/door_linux.gif" border="0" width="468" height="60" alt="Atipa Linux solutions. Your reliable cluster, server, and workstation solution. Win a Free Celeron Linux Workstation!">
</A>
</TD>
<TD>
<IMG SRC="/pics/lt.gif" VSPACE="5" alt="Linux Today Logo">
<BR>
<FONT size="-1">
<A href="http://linux.com">linux.com</A> partner</FONT>
<P>
</TD>
</TR>
</TABLE>
<FONT size="2" face="Helvetica">
[ <A href="http://linuxtoday.com/">headlines</A> |
<A href="http://features.linuxtoday.com/">features</A> |
<A href="http://commercial.linuxtoday.com/">commercial</A> |
<A href="http://security.linuxtoday.com/">security</A> |
<A href="http://jobs.linuxtoday.com/">jobs</A> |
<A href="http://linuxtoday.com/volt/">volt</A> |
<A href="http://linuxtoday.com/contrib.pl">contribute/submit</A> |
<A href="http://linuxtoday.com/advertise/">advertise</A> |
<A href="http://linuxtoday.com/search.html">search</A> |
<A href="http://linuxtoday.com/digests/">site digests</A> |
<A href="http://linuxtoday.com/mail-lists">mailing lists</A> |
<A href="http://linuxtoday.com/about/">about us</A> |
<A href="http://linuxtoday.com/linkus.html">link us</A> ]</FONT>
</CENTER>
<P>
</BODY>
</HTML>

88
result/HTML/test3.html Normal file
View File

@ -0,0 +1,88 @@
<!DOCTYPE HTML>
<HTML>
<HEAD>
<BASE target="contents">
</HEAD>
<A name="ProblemDomain.Package">
<H2>Component Package diagram ProblemDomain</H2>
</A>
<P>
<HR>
<DL>
<DT>
<B>Stereotype </B>problem domain</DT>
<DT>
<B>Alias </B>Problem Domain</DT>
<DT>
<B>Note </B>
<DD>The Problem Domain package is the model behind the Human
<DD>Interface, thats stores and manipulates the Family Tree.
</DD>
</DD>
</DT>
</DL>
<P>
<HR>
<DL>
<DT>
<H4>Class <A href="HumanInterface.FamilyFrame.html#HumanInterface.FamilyFrame">HumanInterface.FamilyFrame</A>
</H4>
</DT>
<DT>
<H4>Class <A href="ProblemDomain.Birth.html#ProblemDomain.Birth">ProblemDomain.Birth</A>
</H4>
</DT>
<DT>
<H4>Class <A href="ProblemDomain.Death.html#ProblemDomain.Death">ProblemDomain.Death</A>
</H4>
</DT>
<DT>
<H4>Class <A href="ProblemDomain.Divorce.html#ProblemDomain.Divorce">ProblemDomain.Divorce</A>
</H4>
</DT>
<DT>
<H4>Class <A href="ProblemDomain.Family.html#ProblemDomain.Family">ProblemDomain.Family</A>
</H4>
</DT>
<DT>
<H4>Class <A href="ProblemDomain.Individual.html#ProblemDomain.Individual">ProblemDomain.Individual</A>
</H4>
</DT>
<DT>
<H4>Class <A href="ProblemDomain.LifeEvent.html#ProblemDomain.LifeEvent">ProblemDomain.LifeEvent</A>
</H4>
</DT>
<DT>
<H4>Class <A href="ProblemDomain.Marriage.html#ProblemDomain.Marriage">ProblemDomain.Marriage</A>
</H4>
</DT>
<DT>
<H4>Class <A href="ProblemDomain.Note.html#ProblemDomain.Note">ProblemDomain.Note</A>
</H4>
</DT>
</DL>
<H4>
<B>Links</B>
</H4>
<UL>
<LI>
<B>Link to </B>
<A href="HumanInterface.Package.html#HumanInterface.Package">HumanInterface</A>
</LI>
</UL>
<DIR></DIR>
<UL>
<LI>
<B>Link to </B>
<A href="DataManagement.FlatFile.Package.html#DataManagement.FlatFile.Package">DataManagement.FlatFile</A>
</LI>
</UL>
<DIR></DIR>
<UL>
<LI>
<B>Link to </B>
<A href="DataManagement.Package.html#DataManagement.Package">DataManagement</A>
</LI>
</UL>
<DIR></DIR>
</HTML>

14
test/HTML/Down.html Normal file
View File

@ -0,0 +1,14 @@
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN"
"http://www.w3.org/TR/REC-html40/loose.dtd">
<html>
<head>
<title>This service is temporary down</title>
</head>
<body bgcolor="#FFFFFF">
<h1 align="center">Sorry, this service is temporary down</h1>
We are doing our best to get it back on-line,
<p>The W3C system administrators</p>
</body>
</html>

33
test/HTML/test2.html Normal file
View File

@ -0,0 +1,33 @@
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN" "http://www.w3.org/TR/REC-html40/loose.dtd">
<HTML><HEAD> <TITLE>Linux Today</TITLE></HEAD>
<body bgcolor="White" link="Blue" text="Black" VLINK="Black" ALINK="Red">
<center>
<TABLE BORDER="0" WIDTH="100%" CELLSPACING="0" CELLPADDING="0">
<TR BGCOLOR="#FFFFFF">
<TD HEIGHT="90">
<a href="http://linuxtoday.com/cgi-bin/click.pl?adnum=49"><img src="/pics/door_linux.gif" border="0" width="468" height="60" alt="Atipa Linux solutions. Your reliable cluster, server, and workstation solution. Win a Free Celeron Linux Workstation!"></A>
</TD>
<TD><IMG SRC="/pics/lt.gif" VSPACE=5 alt="Linux Today Logo"><br><font size="-1"><a href="http://linux.com">linux.com</a> partner</font><p></TD>
</TR>
</TABLE>
<font size="2" face="Helvetica">
[ <a href="http://linuxtoday.com/">headlines</A> |
<a href="http://features.linuxtoday.com/">features</A> |
<a href="http://commercial.linuxtoday.com/">commercial</a> |
<a href="http://security.linuxtoday.com/">security</a> |
<a href="http://jobs.linuxtoday.com/">jobs</a> |
<a href="http://linuxtoday.com/volt/">volt</a> |
<a href="http://linuxtoday.com/contrib.pl">contribute/submit</a> |
<a href="http://linuxtoday.com/advertise/">advertise</A> |
<a href="http://linuxtoday.com/search.html">search</A> |
<a href="http://linuxtoday.com/digests/">site digests</A> |
<a href="http://linuxtoday.com/mail-lists">mailing lists</A> |
<a href="http://linuxtoday.com/about/">about us</a> |
<a href="http://linuxtoday.com/linkus.html">link us</A> ]</font>
</center>
<P>
</body>
</html>

34
test/HTML/test3.html Normal file
View File

@ -0,0 +1,34 @@
<html>
<head>
<base target=contents>
</head>
<a name=ProblemDomain.Package><h2>Component Package diagram ProblemDomain</h2>
</a><p><hr></p>
<dl>
<dt><b>Stereotype </b>problem domain</dt>
<dt><b>Alias </b>Problem Domain</dt>
<dt><b>Note </b><dd>The Problem Domain package is the model behind the Human
<dd>Interface, thats stores and manipulates the Family Tree.
</dl>
<p><hr></p>
<dl>
<dt><h4>Class <a href=HumanInterface.FamilyFrame.html#HumanInterface.FamilyFrame>HumanInterface.FamilyFrame</a></h4></dt>
<dt><h4>Class <a href=ProblemDomain.Birth.html#ProblemDomain.Birth>ProblemDomain.Birth</a></h4></dt>
<dt><h4>Class <a href=ProblemDomain.Death.html#ProblemDomain.Death>ProblemDomain.Death</a></h4></dt>
<dt><h4>Class <a href=ProblemDomain.Divorce.html#ProblemDomain.Divorce>ProblemDomain.Divorce</a></h4></dt>
<dt><h4>Class <a href=ProblemDomain.Family.html#ProblemDomain.Family>ProblemDomain.Family</a></h4></dt>
<dt><h4>Class <a href=ProblemDomain.Individual.html#ProblemDomain.Individual>ProblemDomain.Individual</a></h4></dt>
<dt><h4>Class <a href=ProblemDomain.LifeEvent.html#ProblemDomain.LifeEvent>ProblemDomain.LifeEvent</a></h4></dt>
<dt><h4>Class <a href=ProblemDomain.Marriage.html#ProblemDomain.Marriage>ProblemDomain.Marriage</a></h4></dt>
<dt><h4>Class <a href=ProblemDomain.Note.html#ProblemDomain.Note>ProblemDomain.Note</a></h4></dt>
</dl>
<h4><b>Links</h4></b>
<ul><li><b>Link to </b><a href=HumanInterface.Package.html#HumanInterface.Package>HumanInterface</a></li></ul>
<dir></dir>
<ul><li><b>Link to </b><a href=DataManagement.FlatFile.Package.html#DataManagement.FlatFile.Package>DataManagement.FlatFile</a></li></ul>
<dir></dir>
<ul><li><b>Link to </b><a href=DataManagement.Package.html#DataManagement.Package>DataManagement</a></li></ul>
<dir></dir>
</html>

6
testHTML.c
View File

@ -27,7 +27,7 @@
#include <stdlib.h> #include <stdlib.h>
#include "HTMLparser.h" #include "HTMLparser.h"
#include "tree.h" #include "HTMLtree.h"
#include "debugXML.h" #include "debugXML.h"
static int debug = 0; static int debug = 0;
@ -80,7 +80,7 @@ void parseAndPrintFile(char *filename) {
* print it. * print it.
*/ */
if (!debug) if (!debug)
xmlDocDump(stdout, doc); htmlDocDump(stdout, doc);
else else
xmlDebugDumpDocument(stdout, doc); xmlDebugDumpDocument(stdout, doc);
@ -111,7 +111,7 @@ void parseAndPrintBuffer(CHAR *buf) {
* print it. * print it.
*/ */
if (!debug) if (!debug)
xmlDocDump(stdout, doc); htmlDocDump(stdout, doc);
else else
xmlDebugDumpDocument(stdout, doc); xmlDebugDumpDocument(stdout, doc);