I thought some more about the approach in bug 1390384 and I think it's better if we actually just treat all shorthands as non-discrete. Obviously their longhand components may be discrete, and will be animated as such, we simply won't fallback to the special SMIL keyTimes discrete behavior for such properties unless the author specifies calc="discrete". The reasoning is: * If a shorthand grows an extra longhand that is non-discrete, or if one of the existing longhand components becomes non-discrete, then presumably we should no longer animate the shorthand discretely. However, doing that could presumably break any content that relied on the previous behavior. (In practice it never will because there's not much SMIL content and other browsers don't support shorthands for SMIL anyway.) * Authors can (and should) specify calcMode="linear" if they want to ensure a shorthand and all its longhands use discrete animation (and therefore apply the special keyTimes behavior).

(In reply to Brian Birtles (:birtles) from comment #0) > * Authors can (and should) specify calcMode="linear" if they want to ensure > a shorthand and all its longhands use discrete animation (and therefore > apply > the special keyTimes behavior). You mean calcMode="discrete"?

(In reply to Hiroyuki Ikezoe (:hiro) from comment #2) > (In reply to Brian Birtles (:birtles) from comment #0) > > * Authors can (and should) specify calcMode="linear" if they want to ensure > > a shorthand and all its longhands use discrete animation (and therefore > > apply > > the special keyTimes behavior). > > You mean calcMode="discrete"? Yes, sorry!

(In reply to Brian Birtles (:birtles) from comment #3) > (In reply to Hiroyuki Ikezoe (:hiro) from comment #2) > > (In reply to Brian Birtles (:birtles) from comment #0) > > > * Authors can (and should) specify calcMode="linear" if they want to ensure > > > a shorthand and all its longhands use discrete animation (and therefore > > > apply > > > the special keyTimes behavior). > > > > You mean calcMode="discrete"? > > Yes, sorry! Good to know. :)

Comment on attachment 8897752 [details] Bug 1390766 - Drop complex handling of shorthand properties when determining if they are discrete or not; https://reviewboard.mozilla.org/r/169054/#review174370

Pushed by bbirtles@mozilla.com: https://hg.mozilla.org/integration/autoland/rev/cc542de6ec53 Drop complex handling of shorthand properties when determining if they are discrete or not; r=hiro