shaba/libxml2
1
0
Fork 0
mirror of https://gitlab.gnome.org/GNOME/libxml2.git synced 2024-12-25 23:21:26 +03:00
Code

applied the same kind of refactoring to the HTML saving code. slight API

Browse Source 
* HTMLtree.c include/libxml/HTMLtree.h: applied the same kind
  of refactoring to the HTML saving code.
* doc/libxml2-*.xml doc/API*.html: slight API changes got reflected
  in the doc.
Daniel
This commit is contained in:
Daniel Veillard 2002-11-27 19:39:27 +00:00
parent ebc4ca9456
commit 8db67d2704
7 changed files with 114 additions and 307 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

7
ChangeLog
View File

@ -1,3 +1,10 @@
Wed Nov 27 20:36:08 CET 2002 Daniel Veillard <daniel@veillard.com>
* HTMLtree.c include/libxml/HTMLtree.h: applied the same kind
of refactoring to the HTML saving code.
* doc/libxml2-*.xml doc/API*.html: slight API changes got reflected
in the doc.
Wed Nov 27 12:40:16 CET 2002 Daniel Veillard <daniel@veillard.com>
* tree.c include/libxml/tree.h: refactored the XML dump of a node

339
HTMLtree.c
View File

@ -11,6 +11,8 @@
#include "libxml.h"
#ifdef LIBXML_HTML_ENABLED
#include <string.h> /* for memset() only ! */
#ifdef HAVE_CTYPE_H
#include <ctype.h>
#endif
@ -319,146 +321,10 @@ htmlIsBooleanAttr(const xmlChar *name)
* *
************************************************************************/
static void
htmlDocContentDump(xmlBufferPtr buf, xmlDocPtr cur, int format);
static void
static int
htmlNodeDumpFormat(xmlBufferPtr buf, xmlDocPtr doc, xmlNodePtr cur,
int format);
/**
* htmlDtdDump:
* @buf: the HTML buffer output
* @doc: the document
*
* Dump the HTML document DTD, if any.
*/
static void
htmlDtdDump(xmlBufferPtr buf, xmlDocPtr doc) {
xmlDtdPtr cur = doc->intSubset;
if (cur == NULL) {
xmlGenericError(xmlGenericErrorContext,
"htmlDtdDump : no internal subset\n");
return;
}
xmlBufferWriteChar(buf, "<!DOCTYPE ");
xmlBufferWriteCHAR(buf, cur->name);
if (cur->ExternalID != NULL) {
xmlBufferWriteChar(buf, " PUBLIC ");
xmlBufferWriteQuotedString(buf, cur->ExternalID);
if (cur->SystemID != NULL) {
xmlBufferWriteChar(buf, " ");
xmlBufferWriteQuotedString(buf, cur->SystemID);
}
} else if (cur->SystemID != NULL) {
xmlBufferWriteChar(buf, " SYSTEM ");
xmlBufferWriteQuotedString(buf, cur->SystemID);
}
xmlBufferWriteChar(buf, ">\n");
}
/**
* htmlAttrDump:
* @buf: the HTML buffer output
* @doc: the document
* @cur: the attribute pointer
*
* Dump an HTML attribute
*/
static void
htmlAttrDump(xmlBufferPtr buf, xmlDocPtr doc, xmlAttrPtr cur) {
xmlChar *value;
/*
* TODO: The html output method should not escape a & character
* occurring in an attribute value immediately followed by
* a { character (see Section B.7.1 of the HTML 4.0 Recommendation).
*/
if (cur == NULL) {
xmlGenericError(xmlGenericErrorContext,
"htmlAttrDump : property == NULL\n");
return;
}
xmlBufferWriteChar(buf, " ");
xmlBufferWriteCHAR(buf, cur->name);
if ((cur->children != NULL) && (!htmlIsBooleanAttr(cur->name))) {
value = xmlNodeListGetString(doc, cur->children, 0);
if (value) {
xmlBufferWriteChar(buf, "=");
if ((!xmlStrcasecmp(cur->name, BAD_CAST "href")) ||
(!xmlStrcasecmp(cur->name, BAD_CAST "src"))) {
xmlChar *escaped;
xmlChar *tmp = value;
while (IS_BLANK(*tmp)) tmp++;
escaped = xmlURIEscapeStr(tmp, BAD_CAST"@/:=?;#%&");
if (escaped != NULL) {
xmlBufferWriteQuotedString(buf, escaped);
xmlFree(escaped);
} else {
xmlBufferWriteQuotedString(buf, value);
}
} else {
xmlBufferWriteQuotedString(buf, value);
}
xmlFree(value);
} else {
xmlBufferWriteChar(buf, "=\"\"");
}
}
}
/**
* htmlAttrListDump:
* @buf: the HTML buffer output
* @doc: the document
* @cur: the first attribute pointer
*
* Dump a list of HTML attributes
*/
static void
htmlAttrListDump(xmlBufferPtr buf, xmlDocPtr doc, xmlAttrPtr cur, int format) {
int i = 0;
if (cur == NULL) {
xmlGenericError(xmlGenericErrorContext,
"htmlAttrListDump : property == NULL\n");
return;
}
while (cur != NULL) {
i++;
if ((format) && (i >= 5)) {
i = 0;
xmlBufferWriteChar(buf, "\n");
}
htmlAttrDump(buf, doc, cur);
cur = cur->next;
}
}
/**
* htmlNodeListDump:
* @buf: the HTML buffer output
* @doc: the document
* @cur: the first node
*
* Dump an HTML node list, recursive behaviour,children are printed too.
*/
static void
htmlNodeListDump(xmlBufferPtr buf, xmlDocPtr doc, xmlNodePtr cur, int format) {
if (cur == NULL) {
xmlGenericError(xmlGenericErrorContext,
"htmlNodeListDump : node == NULL\n");
return;
}
while (cur != NULL) {
htmlNodeDumpFormat(buf, doc, cur, format);
cur = cur->next;
}
}
/**
* htmlNodeDumpFormat:
* @buf: the HTML buffer output
@ -467,150 +333,41 @@ htmlNodeListDump(xmlBufferPtr buf, xmlDocPtr doc, xmlNodePtr cur, int format) {
* @format: should formatting spaces been added
*
* Dump an HTML node, recursive behaviour,children are printed too.
*
* Returns the number of byte written or -1 in case of error
*/
static void
static int
htmlNodeDumpFormat(xmlBufferPtr buf, xmlDocPtr doc, xmlNodePtr cur,
int format) {
const htmlElemDesc * info;
unsigned int use;
int ret;
xmlOutputBufferPtr outbuf;
if (cur == NULL) {
xmlGenericError(xmlGenericErrorContext,
"htmlNodeDumpFormat : node == NULL\n");
return;
return (-1);
}
/*
* Special cases.
*/
if (cur->type == XML_DTD_NODE)
return;
if (cur->type == XML_HTML_DOCUMENT_NODE) {
htmlDocContentDump(buf, (xmlDocPtr) cur, format);
return;
if (buf == NULL) {
return (-1);
}
if (cur->type == HTML_TEXT_NODE) {
if (cur->content != NULL) {
if (((cur->name == (const xmlChar *)xmlStringText) ||
(cur->name != (const xmlChar *)xmlStringTextNoenc)) &&
((cur->parent == NULL) ||
((xmlStrcasecmp(cur->parent->name, BAD_CAST "script")) &&
(xmlStrcasecmp(cur->parent->name, BAD_CAST "style"))))) {
xmlChar *buffer;
outbuf = (xmlOutputBufferPtr) xmlMalloc(sizeof(xmlOutputBuffer));
if (outbuf == NULL) {
xmlGenericError(xmlGenericErrorContext,
"htmlNodeDumpFormat: out of memory!\n");
return (-1);
}
memset(outbuf, 0, (size_t) sizeof(xmlOutputBuffer));
outbuf->buffer = buf;
outbuf->encoder = NULL;
outbuf->writecallback = NULL;
outbuf->closecallback = NULL;
outbuf->context = NULL;
outbuf->written = 0;
buffer = xmlEncodeEntitiesReentrant(doc, cur->content);
if (buffer != NULL) {
xmlBufferWriteCHAR(buf, buffer);
xmlFree(buffer);
}
} else {
xmlBufferWriteCHAR(buf, cur->content);
}
}
return;
}
if (cur->type == HTML_COMMENT_NODE) {
if (cur->content != NULL) {
xmlBufferWriteChar(buf, "<!--");
xmlBufferWriteCHAR(buf, cur->content);
xmlBufferWriteChar(buf, "-->");
}
return;
}
if (cur->type == HTML_PI_NODE) {
if (cur->name == NULL)
return;
xmlBufferWriteChar(buf, "<?");
xmlBufferWriteCHAR(buf, cur->name);
if (cur->content != NULL) {
xmlBufferWriteChar(buf, " ");
xmlBufferWriteCHAR(buf, cur->content);
}
xmlBufferWriteChar(buf, ">");
return;
}
if (cur->type == HTML_ENTITY_REF_NODE) {
xmlBufferWriteChar(buf, "&");
xmlBufferWriteCHAR(buf, cur->name);
xmlBufferWriteChar(buf, ";");
return;
}
if (cur->type == HTML_PRESERVE_NODE) {
if (cur->content != NULL) {
xmlBufferWriteCHAR(buf, cur->content);
}
return;
}
/*
* Get specific HTML info for that node.
*/
info = htmlTagLookup(cur->name);
xmlBufferWriteChar(buf, "<");
xmlBufferWriteCHAR(buf, cur->name);
if (cur->properties != NULL)
htmlAttrListDump(buf, doc, cur->properties, format);
if ((info != NULL) && (info->empty)) {
xmlBufferWriteChar(buf, ">");
if ((format) && (info != NULL) && (!info->isinline) &&
(cur->next != NULL)) {
if ((cur->next->type != HTML_TEXT_NODE) &&
(cur->next->type != HTML_ENTITY_REF_NODE))
xmlBufferWriteChar(buf, "\n");
}
return;
}
if (((cur->type == XML_ELEMENT_NODE) || (cur->content == NULL)) &&
(cur->children == NULL)) {
if ((info != NULL) && (info->saveEndTag != 0) &&
(xmlStrcmp(BAD_CAST info->name, BAD_CAST "html")) &&
(xmlStrcmp(BAD_CAST info->name, BAD_CAST "body"))) {
xmlBufferWriteChar(buf, ">");
} else {
xmlBufferWriteChar(buf, "></");
xmlBufferWriteCHAR(buf, cur->name);
xmlBufferWriteChar(buf, ">");
}
if ((format) && (info != NULL) && (!info->isinline) &&
(cur->next != NULL)) {
if ((cur->next->type != HTML_TEXT_NODE) &&
(cur->next->type != HTML_ENTITY_REF_NODE))
xmlBufferWriteChar(buf, "\n");
}
return;
}
xmlBufferWriteChar(buf, ">");
if ((cur->type != XML_ELEMENT_NODE) && (cur->content != NULL)) {
xmlChar *buffer;
buffer = xmlEncodeEntitiesReentrant(doc, cur->content);
if (buffer != NULL) {
xmlBufferWriteCHAR(buf, buffer);
xmlFree(buffer);
}
}
if (cur->children != NULL) {
if ((format) && (info != NULL) && (!info->isinline) &&
(cur->children->type != HTML_TEXT_NODE) &&
(cur->children->type != HTML_ENTITY_REF_NODE) &&
(cur->children != cur->last))
xmlBufferWriteChar(buf, "\n");
htmlNodeListDump(buf, doc, cur->children, format);
if ((format) && (info != NULL) && (!info->isinline) &&
(cur->last->type != HTML_TEXT_NODE) &&
(cur->last->type != HTML_ENTITY_REF_NODE) &&
(cur->children != cur->last))
xmlBufferWriteChar(buf, "\n");
}
xmlBufferWriteChar(buf, "</");
xmlBufferWriteCHAR(buf, cur->name);
xmlBufferWriteChar(buf, ">");
if ((format) && (info != NULL) && (!info->isinline) &&
(cur->next != NULL)) {
if ((cur->next->type != HTML_TEXT_NODE) &&
(cur->next->type != HTML_ENTITY_REF_NODE))
xmlBufferWriteChar(buf, "\n");
}
use = buf->use;
htmlNodeDumpFormatOutput(outbuf, doc, cur, NULL, format);
xmlFree(outbuf);
ret = buf->use - use;
return (ret);
}
/**
@ -621,10 +378,12 @@ htmlNodeDumpFormat(xmlBufferPtr buf, xmlDocPtr doc, xmlNodePtr cur,
*
* Dump an HTML node, recursive behaviour,children are printed too,
* and formatting returns are added.
*
* Returns the number of byte written or -1 in case of error
*/
void
int
htmlNodeDump(xmlBufferPtr buf, xmlDocPtr doc, xmlNodePtr cur) {
htmlNodeDumpFormat(buf, doc, cur, 1);
return(htmlNodeDumpFormat(buf, doc, cur, 1));
}
/**
@ -693,36 +452,6 @@ htmlNodeDumpFile(FILE *out, xmlDocPtr doc, xmlNodePtr cur) {
htmlNodeDumpFileFormat(out, doc, cur, NULL, 1);
}
/**
* htmlDocContentDump:
* @buf: the HTML buffer output
* @cur: the document
*
* Dump an HTML document.
*/
static void
htmlDocContentDump(xmlBufferPtr buf, xmlDocPtr cur, int format) {
int type;
/*
* force to output the stuff as HTML, especially for entities
*/
type = cur->type;
cur->type = XML_HTML_DOCUMENT_NODE;
if (cur->intSubset != NULL)
htmlDtdDump(buf, cur);
else {
/* Default to HTML-4.0 transitional @@@@ */
xmlBufferWriteChar(buf, "<!DOCTYPE HTML PUBLIC \"-//W3C//DTD HTML 4.0 Transitional//EN\" \"http://www.w3.org/TR/REC-html40/loose.dtd\">");
}
if (cur->children != NULL) {
htmlNodeListDump(buf, cur, cur->children, format);
}
xmlBufferWriteChar(buf, "\n");
cur->type = (xmlElementType) type;
}
/**
* htmlDocDumpMemory:
* @cur: the document

31
doc/APIchunk19.html
View File

@ -813,6 +813,37 @@ A:link, A:visited, A:active { text-decoration: underline }
<a href="html/libxml-encoding.html#XMLUTF8STRLOC">xmlUTF8Strloc</a>
<a href="html/libxml-encoding.html#XMLUTF8STRPOS">xmlUTF8Strpos</a>
</dd>
<dt>provided</dt>
<dd>
<a href="html/libxml-docbparser.html#DOCBCREATEFILEPARSERCTXT">docbCreateFileParserCtxt</a>
<a href="html/libxml-docbparser.html#DOCBPARSEFILE">docbParseFile</a>
<a href="html/libxml-docbparser.html#DOCBSAXPARSEDOC">docbSAXParseDoc</a>
<a href="html/libxml-docbparser.html#DOCBSAXPARSEFILE">docbSAXParseFile</a>
<a href="html/libxml-nanoftp.html#FTPDATACALLBACK">ftpDataCallback</a>
<a href="html/libxml-nanoftp.html#FTPLISTCALLBACK">ftpListCallback</a>
<a href="html/libxml-parserinternals.html#HTMLCREATEFILEPARSERCTXT">htmlCreateFileParserCtxt</a>
<a href="html/libxml-htmltree.html#HTMLNEWDOCNODTD">htmlNewDocNoDtD</a>
<a href="html/libxml-htmlparser.html#HTMLPARSEFILE">htmlParseFile</a>
<a href="html/libxml-htmlparser.html#HTMLSAXPARSEDOC">htmlSAXParseDoc</a>
<a href="html/libxml-htmlparser.html#HTMLSAXPARSEFILE">htmlSAXParseFile</a>
<a href="html/libxml-xmlautomata.html#XMLAUTOMATANEWCOUNTEDTRANS">xmlAutomataNewCountedTrans</a>
<a href="html/libxml-c14n.html#XMLC14NDOCSAVETO">xmlC14NDocSaveTo</a>
<a href="html/libxml-c14n.html#XMLC14NEXECUTE">xmlC14NExecute</a>
<a href="html/libxml-parserinternals.html#XMLCREATEENTITYPARSERCTXT">xmlCreateEntityParserCtxt</a>
<a href="html/libxml-parserinternals.html#XMLCREATEFILEPARSERCTXT">xmlCreateFileParserCtxt</a>
<a href="html/libxml-hash.html#XMLHASHFREE">xmlHashFree</a>
<a href="html/libxml-parser.html#XMLKEEPBLANKSDEFAULT">xmlKeepBlanksDefault</a>
<a href="html/libxml-list.html#XMLLISTWALKER">xmlListWalker</a>
<a href="html/libxml-nanohttp.html#XMLNANOHTTPFETCH">xmlNanoHTTPFetch</a>
<a href="html/libxml-xmlio.html#XMLOUTPUTBUFFERCREATEFILENAME">xmlOutputBufferCreateFilename</a>
<a href="html/libxml-parser.html#XMLPARSEFILE">xmlParseFile</a>
<a href="html/libxml-xmlio.html#XMLPARSERINPUTBUFFERCREATEFILENAME">xmlParserInputBufferCreateFilename</a>
<a href="html/libxml-parser.html#XMLRECOVERFILE">xmlRecoverFile</a>
<a href="html/libxml-parser.html#XMLSAXPARSEFILE">xmlSAXParseFile</a>
<a href="html/libxml-parser.html#XMLSAXPARSEFILEWITHDATA">xmlSAXParseFileWithData</a>
<a href="html/libxml-parser.html#XMLSAXUSERPARSEFILE">xmlSAXUserParseFile</a>
<a href="html/libxml-debugxml.html#XMLSHELLVALIDATE">xmlShellValidate</a>
</dd>
<dt>providing</dt>
<dd>
<a href="html/libxml-parserinternals.html#INPUT-CHUNK-CAPS">INPUT_CHUNK</a>

2
doc/APIfiles.html
View File

@ -1117,6 +1117,8 @@ A:link, A:visited, A:active { text-decoration: underline }
<a href="html/libxml-wincecompat.html#CLOSE">close</a>
<a href="html/libxml-wincecompat.html#GETENV">getenv</a>
<a href="html/libxml-wincecompat.html#STRERROR">strerror</a>
<a href="">snprintf</a>
<a href="">vsnprintf</a>
<a href="html/libxml-wincecompat.html#PERROR">perror</a>
</p>
<h2>

8
doc/libxml2-api.xml
View File

@ -330,6 +330,8 @@
<exports symbol='close'/>
<exports symbol='getenv'/>
<exports symbol='strerror'/>
<exports symbol='snprintf'/>
<exports symbol='vsnprintf'/>
<exports symbol='perror'/>
</file>
<file name='parser'>
@ -3040,6 +3042,8 @@
<arg name='ctx' type='void *' info='the user data (XML parser context) '/>
<arg name='name' type='const xmlChar *' info='the namespace prefix '/>
</function>
<macro name='snprintf' file='wincecompat'>
</macro>
<function name='startDocument' file='SAX'>
<info>called when the document start being processed.</info>
<return type='void'/>
@ -3097,6 +3101,8 @@
<arg name='ctxt' type='xmlXPathParserContextPtr' info='an XPath evaluation context '/>
<arg name='value' type='xmlXPathObjectPtr' info='the XPath object '/>
</function>
<macro name='vsnprintf' file='wincecompat'>
</macro>
<functype name='warningSAXFunc' file='parser'>
<info>Display and format a warning messages, callback.</info>
<return type='void'/>
@ -5742,7 +5748,7 @@
</function>
<function name='xmlNodeDump' file='tree'>
<info>Dump an XML node, recursive behaviour,children are printed too. Note that format = 1 provide node indenting only if xmlIndentTreeOutput = 1 or xmlKeepBlanksDefault(0) was called</info>
<return type='void'/>
<return type='int' info='the number of bytes written to the buffer or -1 in case of error '/>
<arg name='buf' type='xmlBufferPtr' info='the XML buffer output '/>
<arg name='doc' type='xmlDocPtr' info='the document '/>
<arg name='cur' type='xmlNodePtr' info='the current node '/>

32
doc/libxml2-refs.xml
View File

@ -8402,6 +8402,8 @@
<ref name='close'/>
<ref name='getenv'/>
<ref name='strerror'/>
<ref name='snprintf'/>
<ref name='vsnprintf'/>
<ref name='perror'/>
</file>
<file name='wsockcompat'>
@ -18799,6 +18801,36 @@
<ref name='xmlUTF8Strloc'/>
<ref name='xmlUTF8Strpos'/>
</word>
<word name='provided'>
<ref name='docbCreateFileParserCtxt'/>
<ref name='docbParseFile'/>
<ref name='docbSAXParseDoc'/>
<ref name='docbSAXParseFile'/>
<ref name='ftpDataCallback'/>
<ref name='ftpListCallback'/>
<ref name='htmlCreateFileParserCtxt'/>
<ref name='htmlNewDocNoDtD'/>
<ref name='htmlParseFile'/>
<ref name='htmlSAXParseDoc'/>
<ref name='htmlSAXParseFile'/>
<ref name='xmlAutomataNewCountedTrans'/>
<ref name='xmlC14NDocSaveTo'/>
<ref name='xmlC14NExecute'/>
<ref name='xmlCreateEntityParserCtxt'/>
<ref name='xmlCreateFileParserCtxt'/>
<ref name='xmlHashFree'/>
<ref name='xmlKeepBlanksDefault'/>
<ref name='xmlListWalker'/>
<ref name='xmlNanoHTTPFetch'/>
<ref name='xmlOutputBufferCreateFilename'/>
<ref name='xmlParseFile'/>
<ref name='xmlParserInputBufferCreateFilename'/>
<ref name='xmlRecoverFile'/>
<ref name='xmlSAXParseFile'/>
<ref name='xmlSAXParseFileWithData'/>
<ref name='xmlSAXUserParseFile'/>
<ref name='xmlShellValidate'/>
</word>
<word name='providing'>
<ref name='INPUT_CHUNK'/>
</word>

2
include/libxml/HTMLtree.h
View File

@ -70,7 +70,7 @@ int htmlDocDump (FILE *f,
xmlDocPtr cur);
int htmlSaveFile (const char *filename,
xmlDocPtr cur);
void htmlNodeDump (xmlBufferPtr buf,
int htmlNodeDump (xmlBufferPtr buf,
xmlDocPtr doc,
xmlNodePtr cur);
void htmlNodeDumpFile (FILE *out,