Intent to Ship: motion path module level 1

[Boris Chiou]

Hi, All

As of Firefox 72, I intend to turn the preference of motion-path,
layout.css.motion-path.enabled, on by default on all platforms. Blink has
shipped it already but Webkit doesn't support it yet. There are some
properties defined in the spec, and I would like to ship part of them, to
match the behaviors in Blink:

1. offset-path: none | path()

2. offset-distance

3. offset-rotate

4. offset-anchor

Note: We have implemented ray() for offset-path
<https://www.w3.org/TR/motion-1/#valdef-offsetpath-ray>, but there are
still some critical spec issues not resolved, so I will add a new
preference to disable it on beta and release channels.


*Bug to turn on by default*:
https://bugzilla.mozilla.org/show_bug.cgi?id=1582554

*Spec*: https://www.w3.org/TR/motion-1/ (or
https://drafts.fxtf.org/motion-1/)

*DevTools*: We don't support DevTools for this motion-path now.

*WPT*:
https://searchfox.org/mozilla-central/source/testing/web-platform/tests/css/motion

Thanks

[Dirk Schulze]

Hi Boris,

Blink doesn't support offset-anchor either it seems. Are you keeping it behind a feature flag as well?

The WebKit bug is here: https://bugs.webkit.org/show_bug.cgi?id=203847
At least initially I do not plan to implement offset-anchor either.

The spec currently is a working draft and not a candidate recommendation (CR). Is that going to get considered? Or did the CSS WG agree to ship?

For the spec it might make sense to split into 2 levels: Level 1 could be the 3 longhand properties offset-path (with the limitations you mentioned), offset-distance and offset-rotate and the shorthand offset. Level 2 would include the 2 remaining properties and the ray() and basic shape functions. If 2 implementations support the 3 longhand properties consistently it brings the spec of level 1 to CR faster.

Greetings,
Dirk

________________________________________
From: dev-platform <dev-platfo...@lists.mozilla.org> on behalf of Boris Chiou <bch...@mozilla.com>
Sent: Thursday, November 14, 2019 11:44 PM
To: dev-platform
Subject: Intent to Ship: motion path module level 1

Hi, All

1. offset-path: none | path()

2. offset-distance

3. offset-rotate

4. offset-anchor

*WPT*:
https://searchfox.org/mozilla-central/source/testing/web-platform/tests/css/motion

Thanks
_______________________________________________
dev-platform mailing list
dev-pl...@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-platform

[Dirk Schulze]

Hi Boris,

On 18. Nov 2019, at 23:31, Boris Chiou <bch...@mozilla.com<mailto:bch...@mozilla.com>> wrote:

Hi Dirk,

Thanks for providing the link to the issue in Webkit. I reply to the questions below:

> Blink doesn't support offset-anchor either it seems. Are you keeping it behind a feature flag as well?

No, I am not planning to add a feature flag for it. It seems Blink has almost finished this feature. Eric Willigers told me Blink has a bug on `offset-anchor` (on an edge case, e.g. 0% 0%), so they haven't shipped it. However I didn't see any new spec issues on it, so I think shipping it should be fine. I'm ok to add an extra feature pref if you have concerns about shipping it.

> The spec currently is a working draft and not a candidate recommendation (CR). Is that going to get considered? Or did the CSS WG agree to ship?

There are still some other spec issues, at least for ray() function, so the spec may not be ready for CR. However, since Blink has shipped it (partially), and `offset-path:none|path()`, `offset-distance`, and `offset-rotate` are stable, so it should be ok to ship them without asking for permission.

> For the spec it might make sense to split into 2 levels: ... If 2 implementations support the 3 longhand properties consistently it brings the spec of level 1 to CR faster.
Agree. it's a great suggestion. Should I file a spec issue for this, or you would like to update the spec as 2 levels?

Yes, that would be great! Thanks.

Dirk


Regards,
Boris

On Sun, Nov 17, 2019 at 9:43 PM Dirk Schulze <dsch...@adobe.com<mailto:dsch...@adobe.com>> wrote:
Hi Boris,

Blink doesn't support offset-anchor either it seems. Are you keeping it behind a feature flag as well?

The WebKit bug is here: https://bugs.webkit.org/show_bug.cgi?id=203847
At least initially I do not plan to implement offset-anchor either.

The spec currently is a working draft and not a candidate recommendation (CR). Is that going to get considered? Or did the CSS WG agree to ship?

For the spec it might make sense to split into 2 levels: Level 1 could be the 3 longhand properties offset-path (with the limitations you mentioned), offset-distance and offset-rotate and the shorthand offset. Level 2 would include the 2 remaining properties and the ray() and basic shape functions. If 2 implementations support the 3 longhand properties consistently it brings the spec of level 1 to CR faster.

Greetings,
Dirk

________________________________________

From: dev-platform <dev-platfo...@lists.mozilla.org<mailto:dev-platfo...@lists.mozilla.org>> on behalf of Boris Chiou <bch...@mozilla.com<mailto:bch...@mozilla.com>>

dev-pl...@lists.mozilla.org<mailto:dev-pl...@lists.mozilla.org>
https://lists.mozilla.org/listinfo/dev-platform

[Boris Chiou]

Hi Dirk,

Thanks for providing the link to the issue in Webkit. I reply to the
questions below:

> Blink doesn't support offset-anchor either it seems. Are you keeping it
behind a feature flag as well?
No, I am not planning to add a feature flag for it. It seems Blink has
almost finished this feature. Eric Willigers told me Blink has a bug on
`offset-anchor` (on an edge case, e.g. 0% 0%), so they haven't shipped it.
However I didn't see any new spec issues on it, so I think shipping it
should be fine. I'm ok to add an extra feature pref if you have concerns
about shipping it.

> The spec currently is a working draft and not a candidate recommendation
(CR). Is that going to get considered? Or did the CSS WG agree to ship?
There are still some other spec issues, at least for ray() function, so the
spec may not be ready for CR. However, since Blink has shipped it
(partially), and `offset-path:none|path()`, `offset-distance`, and
`offset-rotate` are stable, so it should be ok to ship them without asking
for permission.

> For the spec it might make sense to split into 2 levels: ... If 2
implementations support the 3 longhand properties consistently it brings
the spec of level 1 to CR faster.
Agree. it's a great suggestion. Should I file a spec issue for this, or you
would like to update the spec as 2 levels?

Regards,
Boris

On Sun, Nov 17, 2019 at 9:43 PM Dirk Schulze <dsch...@adobe.com> wrote:

> Hi Boris,
>
> Blink doesn't support offset-anchor either it seems. Are you keeping it
> behind a feature flag as well?
>
> The WebKit bug is here: https://bugs.webkit.org/show_bug.cgi?id=203847
> At least initially I do not plan to implement offset-anchor either.
>
> The spec currently is a working draft and not a candidate recommendation
> (CR). Is that going to get considered? Or did the CSS WG agree to ship?
>
> For the spec it might make sense to split into 2 levels: Level 1 could be
> the 3 longhand properties offset-path (with the limitations you mentioned),
> offset-distance and offset-rotate and the shorthand offset. Level 2 would
> include the 2 remaining properties and the ray() and basic shape functions.
> If 2 implementations support the 3 longhand properties consistently it
> brings the spec of level 1 to CR faster.
>
> Greetings,
> Dirk
>
> ________________________________________

> From: dev-platform <dev-platfo...@lists.mozilla.org> on behalf of
> Boris Chiou <bch...@mozilla.com>

> https://lists.mozilla.org/listinfo/dev-platform
>