1149906 - Play-pending animations should still be considered to be playing

Status: RESOLVED FIXED

(Core :: CSS Parsing and Computation, defect)

Description

[Brian Birtles (:birtles)]

Patch 18 from bug 1109390 changed the logic used when fetching animations eligible to be run on the compositor such that play-pending animations are no longer included. That's bad since it means we'll paint such animations' first frame, record their start time, set it on the next refresh driver tick and only *then* stick them in the next layer transaction for the compositor.

This is causing at least test_deferred_start.html to fail on some platforms when we turn OMTA on.

See also bug 980770 comment 17.

Comment 1

[Brian Birtles (:birtles)]

Attachment: Consider play-pending animations to be playing as well

Comment 2

[Brian Birtles (:birtles)]

Try run with OMTA switched on:
https://treeherder.mozilla.org/#/jobs?repo=try&revision=e19dc0b5016b

Comment 3

[Brian Birtles (:birtles)]

And with OMTA off but the full set of mochitests:
https://treeherder.mozilla.org/#/jobs?repo=try&revision=16a482f3b96a

Comment 4

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

It looks like this adds new failures in test_animations_omta.html, though.

Comment 5

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

When I enabled OMTA last night, the failures on Linux debug mochitest-e10s I got in:
https://treeherder.mozilla.org/#/jobs?repo=mozilla-inbound&revision=a765547ac232
were:

 174 INFO TEST-UNEXPECTED-FAIL | dom/animation/test/mozilla/test_deferred_start.html | Starting an animation with a delay starts from the correct point - Starting an animation with a delay starts from the correct point: assert_equals: Got a valid transform matrix on the compositor (got: "") expected 6 but got 0 
 654 INFO TEST-UNEXPECTED-FAIL | layout/style/test/test_animations_omta.html | animation-iteration-count test 2 at 3.1s - got 0.121983, expected 1 - expected PASS 
 655 INFO TEST-UNEXPECTED-FAIL | layout/style/test/test_animations_omta.html | animation-iteration-count test 2 at 3.1s: OMTA style and computed style should be equal - OMTA 0.121983, computed 1 - expected PASS 

This patch fixes the first failure, but adds a new one, so the set of failures is now:
 650 INFO TEST-UNEXPECTED-FAIL | layout/style/test/test_animations_omta.html | anim1 + anim2 + anim3, translate at 10.1s: should NOT be animating on compositor - expected PASS 
 849 INFO TEST-UNEXPECTED-FAIL | layout/style/test/test_animations_omta.html | animation-iteration-count test 2 at 3.1s - got 0.121983, expected 1 - expected PASS 
 850 INFO TEST-UNEXPECTED-FAIL | layout/style/test/test_animations_omta.html | animation-iteration-count test 2 at 3.1s: OMTA style and computed style should be equal - OMTA 0.121983, computed 1 - expected PASS 

The same addition is present on OS X 10.6 opt, although added to the beginning of a longer list of test_animations_omta.html failures.

Comment 6

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

https://hg.mozilla.org/integration/mozilla-inbound/rev/b0758f9d5f6c

Comment 7

[Carsten Book [:Tomcat]]

https://hg.mozilla.org/mozilla-central/rev/b0758f9d5f6c