Bug 42168 – REGRESSION (r63192): Pixel test failure in contain-and-cover.html

• Component: Layout and Rendering
• Status: RESOLVED FIXED

[Pavel Podivilov]

contain-and-cover.html start failing after http://trac.webkit.org/changeset/63192, http://build.chromium.org/buildbot/waterfall.fyi/builders/Webkit%20Linux%20(webkit.org)/builds/32026
Could you please have a look?

[Jeremy Orlow]

Why am I cc'ed on this?  Also you really need to provide WAY more information to non-Chromium contributors.  There's basically no context here.  And no link to where he can see the actual failing test/results.  (Hopefully we'll be able to get pixel tests working on build.webkit.org soon!)

[Pavel Podivilov]

Created attachment 61383 [details]
add contain-and-cover.html to test_expectations

[Jeremy Orlow]

Is this a real regression?  If so (and you can't get ahold of Andreas) then you should revert it and not paper over the regression.  If it's not a real regression then you should probably go ahead and check in a new baseline.

[Pavel Podivilov]

(In reply to comment #3)
> Is this a real regression?  If so (and you can't get ahold of Andreas) then you should revert it and not paper over the regression.  If it's not a real regression then you should probably go ahead and check in a new baseline.

I've cc'ed you just because you were in https://bugs.webkit.org/show_bug.cgi?id=38108 which seemed very similar to this problem. 

I'm not sure if it's a regression or not, that's why I've created this bug.

[Andreas Kling]

Just a quick note to let you know I'm looking at this right now.

[Pavel Podivilov]

(In reply to comment #5)
> Just a quick note to let you know I'm looking at this right now.

Andreas, it's my bad, this is actually the same problem as https://bugs.webkit.org/show_bug.cgi?id=38108, and absolutely not connected to r63192. For some reason image diff was first detected by buildbots just after your change.

[Pavel Podivilov]

*** This bug has been marked as a duplicate of bug 38108 ***

[Andreas Kling]

Not so fast, this actually did regress with my change (r63192). Didn't notice it since we don't have pixel tests for Qt. ;(
Working on a fix right now.

[Pavel Podivilov]

If so, could you please revert it?

*** This bug has been marked as a duplicate of bug 38108 ***

[Pavel Podivilov]

Duplicate -> Unconfirmed

[Andreas Kling]

Reverted with <http://trac.webkit.org/changeset/63213>