Closed Bug 795627 Opened 12 years ago Closed 8 years ago

[Skia] Allow the Skia canvas backend to be used on accelerated Windows

Categories

(Core :: Graphics, defect)

x86_64
Windows 7
defect
Not set
normal

Tracking

()

RESOLVED FIXED
mozilla53
Tracking Status
firefox53 --- fixed

People

(Reporter: gw280, Assigned: milan)

Details

Attachments

(1 file, 1 obsolete file)

Currently, BACKEND_SKIA is only added to the allowable backends bitmask if direct2d is disabled. I think we should allow Skia to be usable on machines that are accelerated as well.
Attached patch Add BACKEND_SKIA to the bitmask (obsolete) — Splinter Review
Attachment #666234 - Flags: review?(bas.schouten)
Oh, hrm, text doesn't currently work in Skia/Canvas if Direct2D is enabled.
Attachment #666234 - Flags: review?(bas.schouten)
Attachment #666234 - Attachment is obsolete: true
Attachment #8818734 - Flags: review?(gwright) → review+
Assignee: nobody → milan
(In reply to Milan Sreckovic [:milan] from comment #5) > https://treeherder.mozilla.org/#/ > jobs?repo=try&revision=790ec47c7196f041a8461f43b4763355ab23068e A couple of intermittent oranges, I don't imagine would get more frequent with this patch.
Pushed by msreckovic@mozilla.com: https://hg.mozilla.org/integration/autoland/rev/11e2a07abdaa Allow Skia canvas with accelerated Windows. r=gw280
Status: NEW → RESOLVED
Closed: 8 years ago
Resolution: --- → FIXED
Target Milestone: --- → mozilla53
Milan, does this fix have automated coverage? Would it benefit from manual testing as well?
Flags: qe-verify?
Flags: needinfo?(milan)
There are automated tests, but I don't think we run this configuration by default, so doing manual testing may be useful. Kats, can you do a try push with the preference disabling D2D content/canvas, just to see if something unexpected is happening? Still with the accelerated compositor.
Flags: needinfo?(milan) → needinfo?(bugmail)
Try results are in. There's gpu job failures, but that's expected (it's asserting D2D is on but it's not). And then there's a handful of WPT tests that are passing instead of failing, so those jobs are failing with TEST-UNEXPECTED-PASS. There's one test which changed from a pass to a fail. I haven't looked at the tests so I'm not sure what they're doing, but overall it doesn't seem like there's a lot of unexpected bustage.
Flags: needinfo?(bugmail)
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: