[webkit-reviews] review denied: [Bug 26342] Absolutely positioned HTML element within foreignObject of absolutely positioned SVG crashes Safari : [Attachment 39636] resubmit patch with test cases, and reverse un-related changes for ASSERT

[bugzilla-daemon at webkit.org]

Eric Seidel <eric at webkit.org> has denied Charles Wei
<charles.wei at torchmobile.com.cn>'s request for review:
Bug 26342: Absolutely positioned HTML element within foreignObject of
absolutely positioned SVG crashes Safari
https://bugs.webkit.org/show_bug.cgi?id=26342

Attachment 39636: resubmit patch with test cases, and reverse un-related
changes for ASSERT
https://bugs.webkit.org/attachment.cgi?id=39636&action=review

------- Additional Comments from Eric Seidel <eric at webkit.org>
Yes, this is the wrong fix. :)

ForeignObjects are a RenderBlock, but they should know how to find their
containing SVGRenderRoot, which is also a RenderBlock.	For SVG this is
slightly tricky becuase SVG doesn't use the concept of "containingBlock" like
the rest of the rendering tree does.  SVG cares about "nearestViewportElement"
instead.   There is some missing logic in SVG here.  Whatever it is that thinks
it wants the foreignObject's containing block might not actually.  Or if it
does, it wants the containing SVGRenderRoot (which it should be able to
find...).  The "return 0" that you're returning should really be:
ASSERT_NOT_REACHED();
return 0;