[Webkit-unassigned] [Bug 153991] Refactor layout functions to avoid using flexbox in MathML

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=153991

--- Comment #3 from Frédéric Wang (:fredw) <fred.wang at free.fr> ---
Comment on attachment 270869
  --> https://bugs.webkit.org/attachment.cgi?id=270869
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=270869&action=review

> LayoutTests/TestExpectations:860
> +mathml/presentation/foreign-mi-dynamic.html [ Skip ]

So I verified that one. In the test, you have some <mi>x</mi> and try to append some <span> inside it. MathMLTextElement::childrenChanged is called, but the renderer for the <span> does not seem to be attached yet when RenderMathMLToken::updateStyle is called and so WebKit still thinks it has to use italic. Moreover, MathMLTextElement::didAttachRenderers is not called and so italic style is not removed either ; maybe this is because the children of the token element are wrapped inside an anonymous RenderBlock. 

I've fixed that on the MathMLLayout-bis branch by delaying the determination of when mathvariant=italic applies as well as the selection of the italic glyph to use. I'm not sure that can easily be done without the refactoring of RenderMathMLToken to remove anonymous node/style, so indeed it seems sensible to skip it at this point.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.webkit.org/pipermail/webkit-unassigned/attachments/20160211/45076f6c/attachment.html>