Eric filed a bug with HTML5: "in body" insertion mode, "any other end tag", step 2.2 seems wrong. https://www.w3.org/Bugs/Public/show_bug.cgi?id=10080 Since then the HTML5 spec is updated to change the 'end tag' processing to not imply its own end tag, since that makes no sense. http://www.whatwg.org/specs/web-apps/current-work/multipage/tree-construction.html#parsing-main-inbody 2.1 Generate implied end tags. -> 2.1 Generate implied end tags, except for elements with the same tag name as the token.
Created attachment 152710 [details] Patch
Comment on attachment 152710 [details] Patch OK.
Comment on attachment 152710 [details] Patch Clearing flags on attachment: 152710 Committed r122831: <http://trac.webkit.org/changeset/122831>
All reviewed patches have been landed. Closing bug.
Is there any observable change in behavior with this change? If so, we should have added a test.
(In reply to comment #5) > Is there any observable change in behavior with this change? If so, we should have added a test. As I mentioned briefly in the change log, there is no observable change in behavior. This patch reduces parse errors, but that's not observable because HTMLTreeBuilder::parseError(AtomicToken&) is just an empty marker.
Thanks!