1023553 - WebGL exposes some Draft extensions without having them behind the draft flag

Status: RESOLVED FIXED

(Core :: Graphics: CanvasWebGL, defect)

Description

[Kelsey Gilbert [:jgilbert]]

This is mostly my fault for not realizing this as we put together our patches for these extensions.

The following extensions are Draft extensions, and should be hidden behind a pref:
EXT_color_buffer_half_float
EXT_sRGB
WEBGL_color_buffer_float
WEBGL_compressed_texture_atc
WEBGL_compressed_texture_etc1
WEBGL_compressed_texture_pvrtc

We should try to promote these, and failing that (for a convincing reason), probably try to put them back under the flag. People *shouldn't* be using these widely, since no one else exposes them unconditionally. The only thing we need to watch out for is Apps on our platforms, which might have chosen to take advantage of one or more of these. Push comes to shove, we may be able to expose these only to Apps, but no longer to general webpages.

That said, we should not forever be gated on there being two+ implementing UAs as a criteria for promotion from Draft.

Comment 1

[Kelsey Gilbert [:jgilbert]]

Attachment: draft-exts

Comment 2

[Kelsey Gilbert [:jgilbert]]

A bunch of extensions were promoted past draft, so the only draft extensions we need to worry about right now are: 
EXT_color_buffer_half_float
WEBGL_color_buffer_float

Comment 3

[u480271]

Comment on attachment 8449827 [details] [diff] [review]
draft-exts

Review of attachment 8449827 [details] [diff] [review]:
-----------------------------------------------------------------

Do we need to audit all the extensions?

Comment 4

[Kelsey Gilbert [:jgilbert]]

(In reply to Dan Glastonbury :djg :kamidphish from comment #3)
> Comment on attachment 8449827 [details] [diff] [review]
> draft-exts
> 
> Review of attachment 8449827 [details] [diff] [review]:
> -----------------------------------------------------------------
> 
> Do we need to audit all the extensions?

Not sure what you mean by 'audit'. What we need to do is propose a new pair of extensions which expose similar functionality, but are unobjectionable to the rest of the working group. (Largely, this means implementability on ANGLE)

Comment 5

[Kelsey Gilbert [:jgilbert]]

https://hg.mozilla.org/integration/mozilla-inbound/rev/7535d57be721

Comment 6

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/mozilla-central/rev/7535d57be721

Comment 7

[u480271]

Jeff,
I meant audit in the sense of making sure all our extensions are currently exposed properly according to draft/community/etc status?

Comment 8

[Kelsey Gilbert [:jgilbert]]

(In reply to Dan Glastonbury :djg :kamidphish from comment #7)
> Jeff,
> I meant audit in the sense of making sure all our extensions are currently
> exposed properly according to draft/community/etc status?

Ah, I checked after the most recent round of extensions were promoted, so we should be good now.