Bug 46439 – Unskip remaining canvas layout tests so that they can track progressions/regressions

• Component: Canvas
• Status: RESOLVED FIXED

[Matthew Delaney]

There are a handful of *mostly* failing canvas layout tests that are still skipped. Some are now passing. These all should be unskipped and their expectations reset for them to be currently useful.

[Matthew Delaney]

Created attachment 68632 [details]
Patch

[Csaba Osztrogonác]

Are there opened bugs for fixing these failing?

[Matthew Delaney]

Yes and no. There is one master bug opened to "pass them all" and then at least a few tracking individual failing tests. We should definitely open bugs for each though.

[Adele Peterson]

Comment on attachment 68632 [details]
Patch

I agree this makes sense to unskip these.  Please file individual bugs for the tests that have failing expected results, so we don't lose track of them.

[Matthew Delaney]

I created this master bug (https://bugs.webkit.org/show_bug.cgi?id=46506) that will depend on all of the individual bugs filed per test. In the process of filing all those.

[WebKit Commit Bot]

Comment on attachment 68632 [details]
Patch

Clearing flags on attachment: 68632

Committed r68568: <http://trac.webkit.org/changeset/68568>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.

[WebKit Review Bot]

http://trac.webkit.org/changeset/68568 might have broken Qt Linux Release
The following changes are on the blame list:
http://trac.webkit.org/changeset/68568
http://trac.webkit.org/changeset/68566
http://trac.webkit.org/changeset/68567

[Csaba Osztrogonác]

(In reply to comment #8)
> http://trac.webkit.org/changeset/68568 might have broken Qt Linux Release
> The following changes are on the blame list:
> http://trac.webkit.org/changeset/68568
> http://trac.webkit.org/changeset/68566
> http://trac.webkit.org/changeset/68567

Qt buildbot don't like this patch, because the following tests pass:

canvas/philip/tests/2d.gradient.radial.cone.front.html
canvas/philip/tests/2d.gradient.radial.inside2.html
canvas/philip/tests/2d.gradient.radial.inside3.html
canvas/philip/tests/2d.gradient.radial.outside1.html
canvas/philip/tests/2d.line.cap.closed.html
canvas/philip/tests/2d.line.join.parallel.html
canvas/philip/tests/2d.shadow.enable.blur.html
canvas/philip/tests/2d.shadow.enable.x.html
canvas/philip/tests/2d.shadow.enable.y.html
canvas/philip/tests/2d.shadow.gradient.alpha.html
canvas/philip/tests/2d.shadow.gradient.basic.html
canvas/philip/tests/2d.shadow.gradient.transparent.2.html
canvas/philip/tests/2d.strokeRect.zero.4.html
canvas/philip/tests/toDataURL.jpeg.alpha.html

[Csaba Osztrogonác]

Qt specific expected files landed for passing tests in http://trac.webkit.org/changeset/68633

[Csaba Osztrogonác]

(In reply to comment #8)
> http://trac.webkit.org/changeset/68568 might have broken Qt Linux Release
> The following changes are on the blame list:
> http://trac.webkit.org/changeset/68568
> http://trac.webkit.org/changeset/68566
> http://trac.webkit.org/changeset/68567

Dear Matthew and Adele, please don't ignore this kind of notices.
We usually fix what we break and don't leave the tree burning.

[Eric Seidel (no email)]

It looks like this broke canvas/philip/tests/toDataURL.jpeg.alpha.html on Snow Leopard.

[Eric Seidel (no email)]

This was a very Mac-centric patch.  Why not check in mac-specific results for these?  I expect a number of these tests are related to canvas on CG systems.

[Eric Seidel (no email)]

We could just roll this out, since the landing was actually done by the commit-queue.

I think the failure in this patch was not considering all platforms.  Canvas is sadly rather graphics-library specific.

[Eric Seidel (no email)]

canvas/philip/tests/2d.imageData.get.source.outside.html is failing on Leopard too.

[Eric Seidel (no email)]

If the commit-bot landed this, then all the tests must have passed for it on either leopard or snow leopard (we have nodes of both OSes and its' random which one you get).  So one of these failures must be flaky.

[Eric Seidel (no email)]

It was landed by a Snow Leopard bot:
https://webkit-commit-queue.appspot.com/results/4210003
(you can tell by the python version).

https://webkit-commit-queue.appspot.com/patch/68632

also shows the the tests flaked once before the land completed correctly.  So it seems this added a flaky test too.

I'm going to roll this out, especially since I feel partially responsible as the cq-admin.  We can roll this back in when someone is around to deal with the tree fires.

[Eric Seidel (no email)]

Reverted r68568 for reason:

Failing tests on both Leopard and Snow Leopard

Committed r68737: <http://trac.webkit.org/changeset/68737>

[WebKit Review Bot]

http://trac.webkit.org/changeset/68737 might have broken SnowLeopard Intel Release (Tests)

[Matthew Delaney]

Currently preparing a patch to add the expected results to the proper locations and unskip the appropriate tests - so as to not mess up the other ports. Sorry about that!

[Matthew Delaney]

*** Bug 46830 has been marked as a duplicate of this bug. ***

[Matthew Delaney]

Created attachment 69395 [details]
Unskip canvas layout tests / update platform expectations

Unskips all but two of the remaining philip canvas layout tests and adds in the mac platform specific results.

[Matthew Delaney]

Created attachment 69400 [details]
Unskip canvas layout tests / update platform expectations

Unskips all but two of the remaining philip canvas layout tests and adds in the mac platform specific results.

[Geoffrey Garen]

Comment on attachment 69400 [details]
Unskip canvas layout tests / update platform expectations

Let's give the commit queue another go at this.

[WebKit Commit Bot]

Comment on attachment 69400 [details]
Unskip canvas layout tests / update platform expectations

Rejecting patch 69400 from commit-queue.

Failed to run "['./WebKitTools/Scripts/webkit-patch', '--status-host=queues.webkit.org', '--bot-id=abarth-cq-sl', 'apply-attachment', '--force-clean', '--non-interactive', '--quiet', 69400]" exit_code: 2
Cleaning working directory
Updating working directory
Logging in as [e-mail]...
Fetching: https://bugs.webkit.org/attachment.cgi?id=69400&action=edit
Fetching: https://bugs.webkit.org/show_bug.cgi?id=46439&ctype=xml
Processing 1 patch from 1 bug.
Processing patch 69400 from bug 46439.
Failed to run "[u'/Users/abarth/git/webkit-queue/WebKitTools/Scripts/svn-apply', u'--reviewer', u'Geoffrey Garen', u'--force']" exit_code: 1

Full output: http://queues.webkit.org/results/4439036

[Eric Seidel (no email)]

The patch doesn't apply.  Sorry the output is kinda lame.

[Matthew Delaney]

Created attachment 72221 [details]
Patch

[WebKit Commit Bot]

Comment on attachment 72221 [details]
Patch

Clearing flags on attachment: 72221

Committed r70803: <http://trac.webkit.org/changeset/70803>

[WebKit Commit Bot]

All reviewed patches have been landed.  Closing bug.