Refine childShouldCreateRenderer for MathML elements

[Frédéric Wang (:fredw)]

Follow-up of bug 124128.

[Frédéric Wang (:fredw)]

Created attachment 227290 [details]
Patch

This adds more restriction and more tests.

[Frédéric Wang (:fredw)]

BTW, I've reported https://www.w3.org/Bugs/Public/show_bug.cgi?id=25104

[chris fleizach]

Comment on attachment 227290 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=227290&action=review

> LayoutTests/mathml/presentation/annotation-children.html:9
> +    <p><math><semantics><csymbol>Content MathML</csymbol><annotation>PA<mtext>ERROR</mtext>SS</annotation></semantics></math></p>

can you added a comment line here to indicate what should be happening

> LayoutTests/mathml/presentation/foreign-element-in-token.html:127
> +  <p>var: <math><mtext><var>mtext</var></mtext></math></p>

bad indendation

> LayoutTests/mathml/presentation/foreign-element-in-token.html:133
> +    <p>h1: <math><mi><h1>mi</h1></mi><mtext><h1>mi</h1></mtext><mn><h1>mn</h1></mn></math></p>

do we need to test all the h tags

> Source/WebCore/mathml/MathMLElement.cpp:286
> +            return htmlElement.hasTagName(HTMLNames::htmlTag) || (isFlowContent(htmlElement) && StyledElement::childShouldCreateRenderer(child));

it looks like isFlowContent is already doing the StyledElement::childShouldCreateRenderer check

[Frédéric Wang (:fredw)]

Created attachment 227430 [details]
Patch

[chris fleizach]

(In reply to comment #4)
> Created an attachment (id=227430) [details]
> Patch

Should this be r?

[Frédéric Wang (:fredw)]

Comment on attachment 227430 [details]
Patch

Yes, I forgot that one.

[Frédéric Wang (:fredw)]

Committed r166065: <http://trac.webkit.org/changeset/166065>