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

576368 – htmlChunkParser with special attributes

Browse Source 
* HTMLparser.c: htmlChunkParsing failed when the chunk ends inside
  element after some attribute which  has a '>' char in its value.
This commit is contained in:
Jiri Netolicky
2009-08-07 17:05:36 +02:00
committed by Daniel Veillard
parent 933e5de96c
commit 446e126de5
1 changed files with 36 additions and 17 deletions
Download Patch File Download Diff File Expand all files Collapse all files

53
HTMLparser.c
View File

@ -4521,11 +4521,14 @@ htmlCreateDocParserCtxt(const xmlChar *cur, const char *encoding) {
*/ */
static int static int
htmlParseLookupSequence(htmlParserCtxtPtr ctxt, xmlChar first, htmlParseLookupSequence(htmlParserCtxtPtr ctxt, xmlChar first,
xmlChar next, xmlChar third, int iscomment) { xmlChar next, xmlChar third, int iscomment,
int ignoreattrval) {
int base, len; int base, len;
htmlParserInputPtr in; htmlParserInputPtr in;
const xmlChar *buf; const xmlChar *buf;
int incomment = 0; int incomment = 0;
int invalue = 0;
char valdellim = 0x0;
in = ctxt->input; in = ctxt->input;
if (in == NULL) return(-1); if (in == NULL) return(-1);
@ -4552,6 +4555,22 @@ htmlParseLookupSequence(htmlParserCtxtPtr ctxt, xmlChar first,
base += 2; base += 2;
} }
} }
if (ignoreattrval) {
if (buf[base] == '"' || buf[base] == '\'') {
if (invalue) {
if (buf[base] == valdellim) {
invalue = 0;
continue;
}
} else {
valdellim = buf[base];
invalue = 1;
continue;
}
} else if (invalue) {
continue;
}
}
if (incomment) { if (incomment) {
if (base + 3 > len) if (base + 3 > len)
return(-1); return(-1);
@ -4731,7 +4750,7 @@ htmlParseTryOrFinish(htmlParserCtxtPtr ctxt, int terminate) {
(UPP(6) == 'Y') && (UPP(7) == 'P') && (UPP(6) == 'Y') && (UPP(7) == 'P') &&
(UPP(8) == 'E')) { (UPP(8) == 'E')) {
if ((!terminate) && if ((!terminate) &&
(htmlParseLookupSequence(ctxt, '>', 0, 0, 0) < 0)) (htmlParseLookupSequence(ctxt, '>', 0, 0, 0, 1) < 0))
goto done; goto done;
#ifdef DEBUG_PUSH #ifdef DEBUG_PUSH
xmlGenericError(xmlGenericErrorContext, xmlGenericError(xmlGenericErrorContext,
@ -4764,7 +4783,7 @@ htmlParseTryOrFinish(htmlParserCtxtPtr ctxt, int terminate) {
if ((cur == '<') && (next == '!') && if ((cur == '<') && (next == '!') &&
(in->cur[2] == '-') && (in->cur[3] == '-')) { (in->cur[2] == '-') && (in->cur[3] == '-')) {
if ((!terminate) && if ((!terminate) &&
(htmlParseLookupSequence(ctxt, '-', '-', '>', 1) < 0)) (htmlParseLookupSequence(ctxt, '-', '-', '>', 1, 1) < 0))
goto done; goto done;
#ifdef DEBUG_PUSH #ifdef DEBUG_PUSH
xmlGenericError(xmlGenericErrorContext, xmlGenericError(xmlGenericErrorContext,
@ -4774,7 +4793,7 @@ htmlParseTryOrFinish(htmlParserCtxtPtr ctxt, int terminate) {
ctxt->instate = XML_PARSER_MISC; ctxt->instate = XML_PARSER_MISC;
} else if ((cur == '<') && (next == '?')) { } else if ((cur == '<') && (next == '?')) {
if ((!terminate) && if ((!terminate) &&
(htmlParseLookupSequence(ctxt, '>', 0, 0, 0) < 0)) (htmlParseLookupSequence(ctxt, '>', 0, 0, 0, 1) < 0))
goto done; goto done;
#ifdef DEBUG_PUSH #ifdef DEBUG_PUSH
xmlGenericError(xmlGenericErrorContext, xmlGenericError(xmlGenericErrorContext,
@ -4788,7 +4807,7 @@ htmlParseTryOrFinish(htmlParserCtxtPtr ctxt, int terminate) {
(UPP(6) == 'Y') && (UPP(7) == 'P') && (UPP(6) == 'Y') && (UPP(7) == 'P') &&
(UPP(8) == 'E')) { (UPP(8) == 'E')) {
if ((!terminate) && if ((!terminate) &&
(htmlParseLookupSequence(ctxt, '>', 0, 0, 0) < 0)) (htmlParseLookupSequence(ctxt, '>', 0, 0, 0, 1) < 0))
goto done; goto done;
#ifdef DEBUG_PUSH #ifdef DEBUG_PUSH
xmlGenericError(xmlGenericErrorContext, xmlGenericError(xmlGenericErrorContext,
@ -4824,7 +4843,7 @@ htmlParseTryOrFinish(htmlParserCtxtPtr ctxt, int terminate) {
if ((cur == '<') && (next == '!') && if ((cur == '<') && (next == '!') &&
(in->cur[2] == '-') && (in->cur[3] == '-')) { (in->cur[2] == '-') && (in->cur[3] == '-')) {
if ((!terminate) && if ((!terminate) &&
(htmlParseLookupSequence(ctxt, '-', '-', '>', 1) < 0)) (htmlParseLookupSequence(ctxt, '-', '-', '>', 1, 1) < 0))
goto done; goto done;
#ifdef DEBUG_PUSH #ifdef DEBUG_PUSH
xmlGenericError(xmlGenericErrorContext, xmlGenericError(xmlGenericErrorContext,
@ -4834,7 +4853,7 @@ htmlParseTryOrFinish(htmlParserCtxtPtr ctxt, int terminate) {
ctxt->instate = XML_PARSER_PROLOG; ctxt->instate = XML_PARSER_PROLOG;
} else if ((cur == '<') && (next == '?')) { } else if ((cur == '<') && (next == '?')) {
if ((!terminate) && if ((!terminate) &&
(htmlParseLookupSequence(ctxt, '>', 0, 0, 0) < 0)) (htmlParseLookupSequence(ctxt, '>', 0, 0, 0, 1) < 0))
goto done; goto done;
#ifdef DEBUG_PUSH #ifdef DEBUG_PUSH
xmlGenericError(xmlGenericErrorContext, xmlGenericError(xmlGenericErrorContext,
@ -4871,7 +4890,7 @@ htmlParseTryOrFinish(htmlParserCtxtPtr ctxt, int terminate) {
if ((cur == '<') && (next == '!') && if ((cur == '<') && (next == '!') &&
(in->cur[2] == '-') && (in->cur[3] == '-')) { (in->cur[2] == '-') && (in->cur[3] == '-')) {
if ((!terminate) && if ((!terminate) &&
(htmlParseLookupSequence(ctxt, '-', '-', '>', 1) < 0)) (htmlParseLookupSequence(ctxt, '-', '-', '>', 1, 1) < 0))
goto done; goto done;
#ifdef DEBUG_PUSH #ifdef DEBUG_PUSH
xmlGenericError(xmlGenericErrorContext, xmlGenericError(xmlGenericErrorContext,
@ -4881,7 +4900,7 @@ htmlParseTryOrFinish(htmlParserCtxtPtr ctxt, int terminate) {
ctxt->instate = XML_PARSER_EPILOG; ctxt->instate = XML_PARSER_EPILOG;
} else if ((cur == '<') && (next == '?')) { } else if ((cur == '<') && (next == '?')) {
if ((!terminate) && if ((!terminate) &&
(htmlParseLookupSequence(ctxt, '>', 0, 0, 0) < 0)) (htmlParseLookupSequence(ctxt, '>', 0, 0, 0, 1) < 0))
goto done; goto done;
#ifdef DEBUG_PUSH #ifdef DEBUG_PUSH
xmlGenericError(xmlGenericErrorContext, xmlGenericError(xmlGenericErrorContext,
@ -4931,7 +4950,7 @@ htmlParseTryOrFinish(htmlParserCtxtPtr ctxt, int terminate) {
break; break;
} }
if ((!terminate) && if ((!terminate) &&
(htmlParseLookupSequence(ctxt, '>', 0, 0, 0) < 0)) (htmlParseLookupSequence(ctxt, '>', 0, 0, 0, 1) < 0))
goto done; goto done;
failed = htmlParseStartTag(ctxt); failed = htmlParseStartTag(ctxt);
@ -5056,7 +5075,7 @@ htmlParseTryOrFinish(htmlParserCtxtPtr ctxt, int terminate) {
int idx; int idx;
xmlChar val; xmlChar val;
idx = htmlParseLookupSequence(ctxt, '<', '/', 0, 0); idx = htmlParseLookupSequence(ctxt, '<', '/', 0, 0, 1);
if (idx < 0) if (idx < 0)
goto done; goto done;
val = in->cur[idx + 2]; val = in->cur[idx + 2];
@ -5083,7 +5102,7 @@ htmlParseTryOrFinish(htmlParserCtxtPtr ctxt, int terminate) {
(UPP(6) == 'Y') && (UPP(7) == 'P') && (UPP(6) == 'Y') && (UPP(7) == 'P') &&
(UPP(8) == 'E')) { (UPP(8) == 'E')) {
if ((!terminate) && if ((!terminate) &&
(htmlParseLookupSequence(ctxt, '>', 0, 0, 0) < 0)) (htmlParseLookupSequence(ctxt, '>', 0, 0, 0, 1) < 0))
goto done; goto done;
htmlParseErr(ctxt, XML_HTML_STRUCURE_ERROR, htmlParseErr(ctxt, XML_HTML_STRUCURE_ERROR,
"Misplaced DOCTYPE declaration\n", "Misplaced DOCTYPE declaration\n",
@ -5093,7 +5112,7 @@ htmlParseTryOrFinish(htmlParserCtxtPtr ctxt, int terminate) {
(in->cur[2] == '-') && (in->cur[3] == '-')) { (in->cur[2] == '-') && (in->cur[3] == '-')) {
if ((!terminate) && if ((!terminate) &&
(htmlParseLookupSequence( (htmlParseLookupSequence(
ctxt, '-', '-', '>', 1) < 0)) ctxt, '-', '-', '>', 1, 1) < 0))
goto done; goto done;
#ifdef DEBUG_PUSH #ifdef DEBUG_PUSH
xmlGenericError(xmlGenericErrorContext, xmlGenericError(xmlGenericErrorContext,
@ -5103,7 +5122,7 @@ htmlParseTryOrFinish(htmlParserCtxtPtr ctxt, int terminate) {
ctxt->instate = XML_PARSER_CONTENT; ctxt->instate = XML_PARSER_CONTENT;
} else if ((cur == '<') && (next == '?')) { } else if ((cur == '<') && (next == '?')) {
if ((!terminate) && if ((!terminate) &&
(htmlParseLookupSequence(ctxt, '>', 0, 0, 0) < 0)) (htmlParseLookupSequence(ctxt, '>', 0, 0, 0, 1) < 0))
goto done; goto done;
#ifdef DEBUG_PUSH #ifdef DEBUG_PUSH
xmlGenericError(xmlGenericErrorContext, xmlGenericError(xmlGenericErrorContext,
@ -5131,7 +5150,7 @@ htmlParseTryOrFinish(htmlParserCtxtPtr ctxt, int terminate) {
break; break;
} else if (cur == '&') { } else if (cur == '&') {
if ((!terminate) && if ((!terminate) &&
(htmlParseLookupSequence(ctxt, ';', 0, 0, 0) < 0)) (htmlParseLookupSequence(ctxt, ';', 0, 0, 0, 1) < 0))
goto done; goto done;
#ifdef DEBUG_PUSH #ifdef DEBUG_PUSH
xmlGenericError(xmlGenericErrorContext, xmlGenericError(xmlGenericErrorContext,
@ -5147,7 +5166,7 @@ htmlParseTryOrFinish(htmlParserCtxtPtr ctxt, int terminate) {
* data detection. * data detection.
*/ */
if ((!terminate) && if ((!terminate) &&
(htmlParseLookupSequence(ctxt, '<', 0, 0, 0) < 0)) (htmlParseLookupSequence(ctxt, '<', 0, 0, 0, 1) < 0))
goto done; goto done;
ctxt->checkIndex = 0; ctxt->checkIndex = 0;
#ifdef DEBUG_PUSH #ifdef DEBUG_PUSH
@ -5173,7 +5192,7 @@ htmlParseTryOrFinish(htmlParserCtxtPtr ctxt, int terminate) {
if (avail < 2) if (avail < 2)
goto done; goto done;
if ((!terminate) && if ((!terminate) &&
(htmlParseLookupSequence(ctxt, '>', 0, 0, 0) < 0)) (htmlParseLookupSequence(ctxt, '>', 0, 0, 0, 1) < 0))
goto done; goto done;
htmlParseEndTag(ctxt); htmlParseEndTag(ctxt);
if (ctxt->nameNr == 0) { if (ctxt->nameNr == 0) {