Bug 54573 – Root element should establish a new block formatting context

• Status: RESOLVED
• Resolution: FIXED
• Component: Layout and Rendering

[Simon Fraser (smfr)]

The root element should establish a new block formatting context. See http://wiki.csswg.org/spec/css2.1#issue-209

[Dave Hyatt]

I think it effectively already does in every case that matters except for expanding to enclose overhanging floats.  We just need to patch that case.

[Levi Weintraub]

Created attachment 83730 [details]
Patch

[Levi Weintraub]

Comment on attachment 83730 [details]
Patch

Removing the r?. If this is accurate (it renders correctly), there are a lot of associated expected result changes I'll have to include as there are many cases where the root's size doesn't include overhanging floats.

[Levi Weintraub]

Created attachment 90605 [details]
Patch

[Levi Weintraub]

(In reply to comment #4)
> Created an attachment (id=90605) [details]
> Patch

This is an easy fix. Can I get a review before the layout test updates atrophy?

[Eric Seidel (no email)]

Comment on attachment 90605 [details]
Patch

OK.

[WebKit Commit Bot]

Comment on attachment 90605 [details]
Patch

Clearing flags on attachment: 90605

Committed r85011: <http://trac.webkit.org/changeset/85011>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.

[WebKit Review Bot]

http://trac.webkit.org/changeset/85029 might have broken Chromium Win Release

[Jessie Berlin]

Note: you only updated the Mac results, ignoring other platform results such as Windows and causing ~10 failures:

http://build.webkit.org/builders/Windows%207%20Release%20%28Tests%29/builds/12162
http://build.webkit.org/results/Windows%207%20Release%20(Tests)/r85011%20(12163)/results.html

I am debating rolling out this patch as a result, since it broke so many tests.

[Jessie Berlin]

I updated the Windows results in http://trac.webkit.org/changeset/85052 instead of rolling out the patch, but it would be better in the future to stick around after you commit if you think there might be any rebaselining to be done on other platforms.

[Levi Weintraub]

(In reply to comment #11)
> I updated the Windows results in http://trac.webkit.org/changeset/85052 instead of rolling out the patch, but it would be better in the future to stick around after you commit if you think there might be any rebaselining to be done on other platforms.

As I stated on https://bugs.webkit.org/show_bug.cgi?id=54573 this was landed in my absence via the commit-queue. I hadn't listed it as cq? so I expected to be the one landing, so I could then be around to prevent this backlash.

Thanks for updating the expected results. I'd have been there to help if I'd known :)

[Levi Weintraub]

Erm, in https://bugs.webkit.org/show_bug.cgi?id=59581 rather.

[Eric Seidel (no email)]

We're working on teaching the EWSes to run tests (currently one cr-linux-ews does!) which may help prevent this sort of trouble in the future.

Committers are ultimately responsible for their patches of course.  But the commit-queue tries to make things as safe as it knows how.  Right now, that means it should never break mac. :)

Thanks for the fix Jessie.