[webkit-reviews] review denied: [Bug 67007] r93794 results in ruby tests causing assertion failure in RenderBlock.cpp : [Attachment 108826] Patch

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Thu Apr 26 19:21:24 PDT 2012


Abhishek Arya <inferno at chromium.org> has denied Roland Steiner
<rolandsteiner at chromium.org>'s request for review:
Bug 67007: r93794 results in ruby tests causing assertion failure in
RenderBlock.cpp
https://bugs.webkit.org/show_bug.cgi?id=67007

Attachment 108826: Patch
https://bugs.webkit.org/attachment.cgi?id=108826&action=review

------- Additional Comments from Abhishek Arya <inferno at chromium.org>
We don't want ruby to have its own code for managing generated content. This
will open a can of security worms. The bug looks to be in
RenderBlock::addChildIgnoringAnonymousColumnBlocks because it cannot passing
the child addition to the correct anonymous parent and then hits that assert.
We recently had similar bug fixed for fullscreen. See
http://trac.webkit.org/changeset/111277. I don't think you need to change to
isAnonymousBlock (as you said in c#5), but you do need to teach
addChildIgnoringAnonymousColumnBlocks to handle ruby wrappers. Also see my c#7
and please first try to see if the test causes incorrect rendering, like does
any testcase result in the wrong location of :before, :after content. Once we
have a testcase showing wrong rendering, then it will be easier to see the
problem.


More information about the webkit-reviews mailing list