Re: [xml] HTMLparser: comments in <style> element

[Michael Day <mikeday yeslogic com>]

Hi Daniel,

Here is the patch to stop htmlParseScript() interpreting <!-- as the 
start of a comment.

Best regards,

Michael

Index: HTMLparser.c
===================================================================
--- HTMLparser.c        (revision 3597)
+++ HTMLparser.c        (working copy)
@@ -2678,23 +2678,7 @@
     SHRINK;
     cur = CUR_CHAR(l);
     while (IS_CHAR_CH(cur)) {
-       if ((cur == '<') && (NXT(1) == '!') && (NXT(2) == '-') &&
-           (NXT(3) == '-')) {
-           if ((nbchar != 0) && (ctxt->sax != NULL) && (!ctxt->disableSAX)) {
-               if (ctxt->sax->cdataBlock!= NULL) {
-                   /*
-                    * Insert as CDATA, which is the same as HTML_PRESERVE_NODE
-                    */
-                   ctxt->sax->cdataBlock(ctxt->userData, buf, nbchar);
-               } else if (ctxt->sax->characters != NULL) {
-                   ctxt->sax->characters(ctxt->userData, buf, nbchar);
-               }
-           }
-           nbchar = 0;
-           htmlParseComment(ctxt);
-           cur = CUR_CHAR(l);
-           continue;
-       } else if ((cur == '<') && (NXT(1) == '/')) {
+       if ((cur == '<') && (NXT(1) == '/')) {
             /*
              * One should break here, the specification is clear:
              * Authors should therefore escape "</" within the content.

Index: HTMLparser.c
===================================================================
--- HTMLparser.c        (revision 3597)
+++ HTMLparser.c        (working copy)
@@ -2678,23 +2678,7 @@
     SHRINK;
     cur = CUR_CHAR(l);
     while (IS_CHAR_CH(cur)) {
-       if ((cur == '<') && (NXT(1) == '!') && (NXT(2) == '-') &&
-           (NXT(3) == '-')) {
-           if ((nbchar != 0) && (ctxt->sax != NULL) && (!ctxt->disableSAX)) {
-               if (ctxt->sax->cdataBlock!= NULL) {
-                   /*
-                    * Insert as CDATA, which is the same as HTML_PRESERVE_NODE
-                    */
-                   ctxt->sax->cdataBlock(ctxt->userData, buf, nbchar);
-               } else if (ctxt->sax->characters != NULL) {
-                   ctxt->sax->characters(ctxt->userData, buf, nbchar);
-               }
-           }
-           nbchar = 0;
-           htmlParseComment(ctxt);
-           cur = CUR_CHAR(l);
-           continue;
-       } else if ((cur == '<') && (NXT(1) == '/')) {
+       if ((cur == '<') && (NXT(1) == '/')) {
             /*
              * One should break here, the specification is clear:
              * Authors should therefore escape "</" within the content.