- From: Randy Waki <rwaki@flipdog.com>
- Date: Wed, 11 Oct 2000 20:41:26 -0600
- To: <dsr@w3.org>, <html-tidy@w3.org>
If a frame element appears after the end of a frameset, 4-Aug-2000 Tidy
enters an infinite loop. Tidy mistakenly tries to wrap the frame in a
noframes element, but since frame is not allowed in noframes, Tidy then
ends the noframes element which puts it right back where it started. IE
and Netscape simply ignore the illegal frame element.
This is the same bug previously reported by Johny Jugianto:
http://lists.w3.org/Archives/Public/html-tidy/2000JulSep/0336.html
The following change to ParseHTML() in parser.c should fix this.
(The equivalent change in Java Tidy is working for me.)
*** \dev\tidy\2000-08-04-C\c\parser.c Fri Aug 04 16:32:04 2000
--- parser.c Wed Oct 11 19:19:32 2000
*************** void ParseHTML(Lexer *lexer, Node *html,
*** 3187,3192 ****
--- 3187,3199 ----
MoveToHead(lexer, html, node);
continue;
}
+
+ /* discard illegal frame element following a frameset */
+ if (frameset != null && node->tag == tag_frame)
+ {
+ ReportWarning(lexer, html, node, DISCARDING_UNEXPECTED);
+ continue;
+ }
}
UngetToken(lexer);
------------------------ Example HTML document -------------------------
<html>
<head><title></title></head>
<frameset>
<frame name="name" src="one.html">
</frameset>
<frame name="ignored-by-browsers" src="two.html">
This text belongs in a noframes element.
</html>
------------------------------------------------------------------------
Randy
Received on Wednesday, 11 October 2000 22:42:46 UTC