Bug 48396 – Siblings of floated elements should be cleared below the float if they are too wide to fit in the containing block.

• Component: Layout and Rendering
• Status: RESOLVED FIXED

[Andy Estes]

Siblings of floated elements should be cleared below the float if they are too wide to fit in the containing block.

[Andy Estes]

<rdar://problem/7699849>

[Andy Estes]

Created attachment 71980 [details]
Reduction

[Andy Estes]

CSS 2.1 says in section 9.5:

"The border box of a table, a block-level replaced element, or an element in the normal flow that establishes a new block formatting context (such as an element with 'overflow' other than 'visible') must not overlap any floats in the same block formatting context as the element itself. If necessary, implementations should clear the said element by placing it below any preceding floats, but may place it adjacent to such floats if there is sufficient space. They may even make the border box of said element narrower than defined by section 10.3.3. CSS2 does not define when a UA may put said element next to the float or by how much said element may become narrower."

WebKit does not clear elements adjacent to floats if they are wider than the containing block, instead keeping them on the same line and allowing them to spill out the right side of the containing block. We should match the spec in this regard and clear the element below the float even if it is too wide for the containing block. This seems to also match IE's and Firefox's behavior.

[Andy Estes]

Created attachment 71982 [details]
Patch

[Andy Estes]

Created attachment 71984 [details]
Patch

[Simon Fraser (smfr)]

Comment on attachment 71984 [details]
Patch

It would be good to run through some CSS 2.1 testcases to see if we pass any more of them with this change.

[Andy Estes]

(In reply to comment #6)
> (From update of attachment 71984 [details])
> It would be good to run through some CSS 2.1 testcases to see if we pass any more of them with this change.

I'd be happy to.

[Andy Estes]

Comment on attachment 71984 [details]
Patch

View in context: https://bugs.webkit.org/attachment.cgi?id=71984&action=review

> WebCore/rendering/RenderBlock.cpp:4046
> +            if (widthAtY == availableLogicalWidth())

This was a bad idea. availableLogicalWidth() only needs to be called once in this function.

[Andy Estes]

Comment on attachment 71982 [details]
Patch

The first patch should be the one reviewed after all.

[Dave Hyatt]

Comment on attachment 71982 [details]
Patch

r=me, but should probably remove the "Jeffrey Zeldman" stuff from the title element in the test case.

[Andy Estes]

Committed r70619: <http://trac.webkit.org/changeset/70619>

[Andy Estes]

(In reply to comment #10)
> (From update of attachment 71982 [details])
> r=me, but should probably remove the "Jeffrey Zeldman" stuff from the title element in the test case.

Thanks Dave! I had actually made that change to the test, but the patch looks weird when you svn mv a file then make changes to it.