[css-transitions-1] Make discrete properties transition at 50%; exclude them from 'all'. #4441

[fantasai]

No description provided.

[dbaron]

I think this is mostly good, although there are a number of things that are awkward about doing this as a delta spec and it's not completely clear how diff-like to make some of the edits.

I have a few comments below.

In addition to those comments, I think you should specifically patch all occurences of "transitionable" in the "Starting of transitions" section, which does need to be patched to make this part of the detailed algorithm. I think that may need a term (I have mixed feelings about whether to re-use "transitionable") that considers whether the relevant part of transition-property was all or not.

[fantasai]

I think you should specifically patch all occurences of "transitionable" in the "Starting of transitions" section, which does need to be patched to make this part of the detailed algorithm

I'm patching the definition of "transitionable" here; I don't see that, with that change, anything in "Starting of transitions" needs edits?

[dbaron]

I'm patching the definition of "transitionable" here; I don't see that, with that change, anything in "Starting of transitions" needs edits?

With your patch to "transitionable", the detailed algorithm in "Starting of transitions" will start transitions for discretely-animatable properties when all is specified for transition-property, which is incorrect.

[fantasai]

With your patch to "transitionable", the detailed algorithm in "Starting of transitions" will start transitions for discretely-animatable properties when all is specified for transition-property, which is incorrect.

I don't see where it says that. It doesn't mention the all keyword in that section at all, afaict. And the section we're editing here now says very clearly that discretely-animatable properties are not included in all. So I don't see how you're arriving at this conclusion.

[dbaron]

I don't see where it says that. It doesn't mention the all keyword in that section at all, afaict. And the section we're editing here now says very clearly that discretely-animatable properties are not included in all. So I don't see how you're arriving at this conclusion.

OK, after rereading, I agree.

[dbaron]

OK, looks good. Thanks.

[dholbert]

I have an editorial nit with the new text here which I filed as #8761

[dholbert]

I also have one question about various tests that were updated to test this change but I think are mistaken; see web-platform-tests/wpt#39871

(Alternately, maybe the tests are right and the spec-text just needs clarification?)