[web-animations-1] The definition of a current effect should reflect the playback rate

[birtles]

This has come up in a couple of contexts for me recently, but I'm starting to think that the definition of a current effect should reflect the sign of the playback rate.

If you have an animation with a negative playback rate that is in the end delay phase, it should probably be returned by getAnimations() even if it is not filling.

[birtles]

Discussed and the suggestion is instead to flip the phases.

[birtles]

I looked into doing this and flipping the phases doesn't make sense. They are used for a variety of calculations such as applying fill modes etc. and changing their meaning would render all those calculations wrong. If someone feels uncomfortable using the terms "before" and "after" then we should just rename them to "left phase" and "right phase" or something similar.

[birtles]

I've made up a spec patch for this but I'm waiting to see if this change causes any unexpected breakage to any of the Firefox automated tests before putting it up for review.

[birtles]

When reviewing this, @hiikezoe raised a good question about the behavior when playbackRate == 0.

One the one hand, in many places we treat playbackRate == 0 as a positive playback rate, but in this case we probably don't need to.

Do you think it would make sense to treat an animation in the before phase with playbackRate == 0 as NOT current?

One effect of this is that if an author does anim.playbackRate = 0 to pause an animation in its delay phase it will mean that animation no longer appears in the result of anim.effect.target.getAnimations(). I think that might be ok though?

[birtles]

Specifically this is the change I'm considering:
birtles@964469a