RESOLVED FIXED 40716
[Chromium] fast/js/missing-title-end-js.html is failing on Chrome stating from r61234 
https://bugs.webkit.org/show_bug.cgi?id=40716
Summary [Chromium] fast/js/missing-title-end-js.html is failing on Chrome stating fro...
Andrei Popescu
Reported 2010-06-16 10:23:03 PDT
Starting from http://trac.webkit.org/changeset/61234, fast/js/missing-title-end-js.html started failing on Chromium. It seems that the test throws an exception and dumpAsText() is not called.
Attachments
Add attachment proposed patch, testcase, etc.
Adam Barth
Comment 1 2010-06-16 11:03:48 PDT
That's correct. The new behavior is for an unclosed <title> to swallow the whole document.
Tony Gentilcore
Comment 2 2010-06-16 11:04:45 PDT
(In reply to comment #0) > Starting from http://trac.webkit.org/changeset/61234, fast/js/missing-title-end-js.html started failing on Chromium. It seems that the test throws an exception and dumpAsText() is not called. When I run this from chromium with run_webkit_tests.bat, it looks like the HTML5 parser just isn't enabled there. If it were, the test would pass. The platform specific expectations in platform/chromium-win/fast/tokenizer look correct for the HTML5 parser (empty image with entire document swallowed). Bulach has a patch for this: http://codereview.chromium.org/2809011 Landing that patch should fix this.
Andrei Popescu
Comment 3 2010-06-16 11:18:19 PDT
(In reply to comment #2) > (In reply to comment #0) > > Starting from http://trac.webkit.org/changeset/61234, fast/js/missing-title-end-js.html started failing on Chromium. It seems that the test throws an exception and dumpAsText() is not called. > > When I run this from chromium with run_webkit_tests.bat, it looks like the HTML5 parser just isn't enabled there. If it were, the test would pass. The platform specific expectations in platform/chromium-win/fast/tokenizer look correct for the HTML5 parser (empty image with entire document swallowed). > > Bulach has a patch for this: http://codereview.chromium.org/2809011 > > Landing that patch should fix this. Actually, we did land that patch: http://codereview.chromium.org/2843005/show As far as I can tell, the problem seems to be that, with the HTML5 lexer enabled, the test throws on Chrome and does not call dumpAsText(). The bot then thinks it must've been a pixel test and looks for a png output, doesn't find it and marks the test as failed with "MISSING" output.
Note You need to log in before you can comment on or make changes to this bug.