1286716 - Implement plane splitting (for OpenGL backend)

Status: RESOLVED FIXED

(Core :: Graphics, defect, P3)

Description

[Miko Mynttinen]

This bug is about integrating the changes from bug 1286412 and bug 1274673 to implement plane splitting for OpenGL backend.

Comment 1

[Miko Mynttinen]

Attachment: Bug 1286716 - Implement plane splitting (for OpenGL backend)

Review commit: https://reviewboard.mozilla.org/r/98598/diff/#index_header
See other reviews: https://reviewboard.mozilla.org/r/98598/

Comment 2

[Miko Mynttinen]

This patch enables plane splitting flag for OpenGL compositor backend.

Comment 3

[Matt Woodrow (:mattwoodrow)]

Comment on attachment 8819028 [details]
Bug 1286716 - Implement plane splitting (for OpenGL backend)

https://reviewboard.mozilla.org/r/98598/#review99598

Comment 4

[Miko Mynttinen]

https://treeherder.mozilla.org/#/jobs?repo=try&revision=c4aedc5d93e0d8cb5c4cc89096f40784de3711a7

Comment 5

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/f5c0b7682e2f
Implement plane splitting (for OpenGL backend) r=mattwoodrow

Comment 6

[Phil Ringnalda (:philor)]

https://hg.mozilla.org/mozilla-central/rev/f5c0b7682e2f

Comment 7

[Thomas Wisniewski]

Just FYI, Mozregression is telling me that this is the cause for a webcompat issue where a CSS transition is finishing and then the content is vanishing (https://github.com/webcompat/web-bugs/issues/4284).

Comment 8

[Miko Mynttinen]

(In reply to Thomas Wisniewski from comment #7)
> Just FYI, Mozregression is telling me that this is the cause for a webcompat
> issue where a CSS transition is finishing and then the content is vanishing
> (https://github.com/webcompat/web-bugs/issues/4284).

Thank you so much for this. I am currently debugging a similar regression I noticed at http://www.simulationcurriculum.com, but this test case will make things a lot easier.

I will report back when I have more information.

Comment 9

[Thomas Wisniewski]

Sorry to dredge this up, but it looks like this patchset broke Spotify (according to mozregression) on OSX and Linux, and I'm not 100% sure if it's something we were aware of (or still in the middle of fixing).

>Last good revision: a4ddb0dd052046a056271628a570d86152361245
>First bad revision: f5c0b7682e2ff43f9bbe7400dc0ec7d397f256b3
>Pushlog:
>https://hg.mozilla.org/integration/autoland/pushloghtml?fromchange=a4ddb0dd052046a056271628a570d86152361245&tochange=f5c0b7682e2ff43f9bbe7400dc0ec7d397f256b3
>Looks like the following bug has the changes which introduced the regression:
>https://bugzilla.mozilla.org/show_bug.cgi?id=1286716