Pending work for transitions/animations

[emilio]

Assuming #11766 lands, this would be the pending work left to do for animations and transitions. I'm planning in doing it, though if someone wants to help that would be appreciated.
Pending work for animations/transitions:

• Support more animated properties: These are marked with a FIXME in
  style/properties/longhands/. These are easy, it's just a matter of
  following the transitions spec and implement the interpolation
  appropriately.
• Support the animation shorthand. This should be easy.
• Scan for all the animatable properties of all the keyframes, not just the
  first one. This is not specified, but all other browsers seem to do this.
• Apply all the previous keyframes' declarations when computed the style. We
  only apply the previous keyframe, and interpolate it with the target one.
  We should do all the chain, maybe storing the current animation's computed
  style in the state to be a bit more efficient? We should be able to do that
  while the animation is ticking, but not on style because we don't have
  write access to the hashmap (well, we could, but it could become really
  slow).
• Don't remove animations from the layout hashmap when they get expired. This
  might cause them to get retriggered on the next restyle, which is wrong.
• Implement animation-fill-mode. This will allow us to test the animation
  code.
• Fix the parsing of the steps() function, which was wrong.
• Figure out how to test all this stuff (#12120).
• Split the SharedLayoutContext in two different pieces, one cloneable per
  thread, and one inmutable as it's today. This will allow us to take rid of
  the Mutex in new_animations_sender.
• File a spec bug about why Gecko doesn't seem to parse
  @keyframes "bar" {}, or fix it in gecko directly.
• Probably a few more things.

[emilio]

cc @SimonSapin

[emilio]

@SimonSapin: The string part for the animation name seems to have been added lately, so it's just no-one else has implemented it yet :)

[mrobinson]

@emilio Can you confirm whether or not parsing of steps() has been fixed? I think if it has, we can close this issue in favor of #26328.

[emilio]

Yeah, I believe steps parsing should be fine now. Seems fine to close as a dupe. Thanks!