1371388 - Stylo: animation removal mutation not fired on class change

Status: RESOLVED WORKSFORME

(Core :: CSS Parsing and Computation, defect, P2)

Description

[J. Ryan Stinnett [:jryans] (Use needinfo, replies may be slow)]

DevTools expects to receive an animation mutation (via MutationObserver) when an animation is removed because of a class change, but this appears to not happen with Stylo.

This leads to timeouts in tests like:

* devtools/server/tests/browser/browser_animation_emitMutations.js[1]

[1]: https://treeherder.mozilla.org/logviewer.html#?job_id=105301401&repo=try&lineNumber=3470

Comment 1

[Hiroyuki Ikezoe (:hiro)]

Attachment: An example

Interesting. Initially I thought we don't cancel the animation when the corresponding class name is removed, but as far as I can see, the animation is cancelled.

Comment 2

[Brian Birtles (:birtles, Feb 26~ parental leave)]

(In reply to Hiroyuki Ikezoe (:hiro) from comment #1)
> Interesting. Initially I thought we don't cancel the animation when the
> corresponding class name is removed, but as far as I can see, the animation
> is cancelled.

There are definitely some cases for transitions where we don't cancel them, but for animations we should.

Comment 3

[Brian Birtles (:birtles, Feb 26~ parental leave)]

This may be the cause of the failures in dom/animation/test/chrome/test_animation_observers_a?sync.html too (a test I plan to look into later today).

Comment 4

[J. Ryan Stinnett [:jryans] (Use needinfo, replies may be slow)]

The test page used by this DevTools test is:

http://searchfox.org/mozilla-central/rev/a798ee4fc323f9387b7576dbed177859d29d09b7/devtools/server/tests/browser/animation.html

Specifically, it appears to change the node with the class "not-animated" to "multiple-animations" (which seems to work) and then later back to "not-animated" (which is the step where removed animations are missed).

The test page seems to contain a mix of both animations and transitions together.

Comment 5

[Cameron McCormack (:heycam)]

This test passes for me, locally.  (Despite a few assertions, which are bug 1189015.)  Ryan, can you confirm that this test is passing on try in your next try run?

Comment 6

[Cameron McCormack (:heycam)]

(In reply to Brian Birtles (:birtles) from comment #3)
> This may be the cause of the failures in
> dom/animation/test/chrome/test_animation_observers_a?sync.html too (a test I
> plan to look into later today).

The sync version of that test is passing for me too.  The async version is failing, and I think that's covered by bug 1376248.

Comment 7

[J. Ryan Stinnett [:jryans] (Use needinfo, replies may be slow)]

(In reply to Cameron McCormack (:heycam) from comment #5)
> This test passes for me, locally.  (Despite a few assertions, which are bug
> 1189015.)  Ryan, can you confirm that this test is passing on try in your
> next try run?

Looks like it is now passing on try and locally for me as well! :)