1582554 - [motion-1] Ship motion-path

Status: RESOLVED FIXED

(Core :: CSS Transitions and Animations, enhancement, P2)

Description

[Boris Chiou [:boris]]

This includes the properties and values we have implemented:

1. offset-path: none|path()
2. offset-distance
3. offset-rotate
4. offset-anchor
5. offset shorthand

Note: offset-position is not implemented yet. We intentionally partially ship these properties to match Blink.

Updated: drop ray() because its spec issues haven't been resolved.

Comment 1

[Boris Chiou [:boris]]

Note for me: we should add a pref for ray function and don't ship it for now because there are still some critical spec issues about ray().

Comment 2

[Boris Chiou [:boris]]

We are ready to ship. I will disable ray() and turn the pref on for release/beta.

Comment 3

[Boris Chiou [:boris]]

Attachment: Bug 1582554 - Add a preference for offset-path:ray().

Comment 4

[Boris Chiou [:boris]]

Attachment: Bug 1582554 - Ship motion-path without offset-path:ray().

Comment 5

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

I think it's worth having a bug blocking motion-1 on file to ship ray(). (Maybe also a bug for the other parts of offset-path that we don't implement, and a bug for offset-position.

Are there bugs on file for the cases in https://wpt.fyi/results/css/motion where we fail tests that Chromium passes?

Comment 6

[David Baron :dbaron: (⌚️UTC-4, no longer working on Mozilla)]

(In reply to David Baron :dbaron: 🏴󠁵󠁳󠁣󠁡󠁿 ⌚UTC-8 from comment #5)

Are there bugs on file for the cases in https://wpt.fyi/results/css/motion where we fail tests that Chromium passes?

I think the only one that's concerning is http://w3c-test.org/css/motion/offset-path-serialization.html .

(I think the Edge running there is more representative than the Chrome, since I think the Chrome there has experimental web platform features enabled.)

Comment 7

[Boris Chiou [:boris]]

(In reply to David Baron :dbaron: 🏴󠁵󠁳󠁣󠁡󠁿 ⌚UTC-8 from comment #6)

(In reply to David Baron :dbaron: 🏴󠁵󠁳󠁣󠁡󠁿 ⌚UTC-8 from comment #5)

Are there bugs on file for the cases in https://wpt.fyi/results/css/motion where we fail tests that Chromium passes?

I think the only one that's concerning is http://w3c-test.org/css/motion/offset-path-serialization.html .

About the serialization of offset-path:path(), there is a spec issue for this: https://github.com/w3c/svgwg/issues/321:
We normalize the SVG path at animation time, instead of computed time. Its gecko bug is Bug 1489392. That's why we didn't pass the serialization of the computed value. It seems there is no conclusion in the spec issue, so we still use "as specified value" for the computed value of offset-path.

Comment 8

[Boris Chiou [:boris]]

(In reply to David Baron :dbaron: 🏴󠁵󠁳󠁣󠁡󠁿 ⌚UTC-8 from comment #5)

I think it's worth having a bug blocking motion-1 on file to ship ray(). (Maybe also a bug for the other parts of offset-path that we don't implement, and a bug for offset-position.

I filed some bugs blocking motion-1 for these. They should all of the properties & values and the shipment. Thanks for the reminder.

Comment 9

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/683152ca004c
Add a preference for offset-path:ray(). r=emilio
https://hg.mozilla.org/integration/autoland/rev/43ce89f91487
Ship motion-path without offset-path:ray(). r=hiro

Comment 10

[Andreea Pavel [:apavel]]

https://hg.mozilla.org/mozilla-central/rev/683152ca004c
https://hg.mozilla.org/mozilla-central/rev/43ce89f91487

Comment 11

[Martin Balfanz [:mbalfanz]]

[Why is this notable]: I think this deserves developer facing release notes as it's a new CSS feature and has been implemented already by Chrome.
[Suggested wording]: Firefox 72 implements 'Motion paths', which allow authors to animate elements along an author-specified path.
[Links (documentation, blog post, etc)]: spec https://drafts.fxtf.org/motion-1/

Comment 12

[Julien Cristau [:jcristau]]

AFAIK dev-doc-needed covers the developer-facing release notes on MDN.

Comment 13

[Sebastian Zartner [:sebo]]

(In reply to Julien Cristau [:jcristau] from comment #12)

AFAIK dev-doc-needed covers the developer-facing release notes on MDN.

That's correct, though it may also be mentioned in the "Developer" section of the release notes.

Sebastian