Closed Bug 1376430 Opened 7 years ago Closed 1 month ago

Intermittent /2dcontext/path-objects/2d.path.stroke.prune.arc.html | Zero-length line segments from arcTo and arc are removed before stroking - assert_equals: Red channel of the pixel at (50, 25) expected 0 but got 255

Categories

(Core :: Graphics: Canvas2D, defect, P3)

defect

Tracking

()

RESOLVED INCOMPLETE

People

(Reporter: intermittent-bug-filer, Unassigned)

References

Details

(Keywords: intermittent-failure, Whiteboard: [gfx-noted] [stockwell disabled])

Attachments

(1 file)

Whiteboard: [gfx-noted]
See Also: → 1376273
This bug (and bug 1376273) happen only on Windows 10, where wpt is tier 2, and seems to only run on mozilla-central. On central, this started with https://treeherder.mozilla.org/#/jobs?repo=mozilla-central&revision=42593af5ec7ec9d5a07c9c0e899ac99244d15ae6&filter-searchStr=windows+web-platform-tests. Perhaps a consequence of bug 1363428?
:milan - This bug and bug 1376273 are on track for 100+ failures / week. Can you find someone to look into it, or can we disable this test on Windows?
Flags: needinfo?(milan)
This test is confusing me, at a first glance it seems to be asking for the wrong value.
(In reply to Geoff Brown [:gbrown] from comment #3) > :milan - This bug and bug 1376273 are on track for 100+ failures / week. Can > you find someone to look into it, or can we disable this test on Windows? Does it ever succeed?
I'm sorry :milan, I think I see where this has gone wrong now. This test always fails (at least on windows and linux), as you say. The test is marked as failing for many platforms: https://dxr.mozilla.org/mozilla-central/rev/b07db5d650b7056c78ba0dbc409d060ec4e922cd/testing/web-platform/meta/2dcontext/path-objects/2d.path.stroke.prune.arc.html.ini#1 and that explains why those test jobs are green for other platforms. I think the annotation for Windows 10 is wrong, and that's why this test is perma-fail on Windows. :jgraham -- What's the best way to mark this as failing on all Windows?
Flags: needinfo?(milan) → needinfo?(james)
Well, obviously that would be os == 'win'. I guess I'm wondering, why are we testing for all these windows versions, and can we collapse those to os == 'win'?
I am still not sure what was wrong with the existing annotations, but "collapsing" them to (os == "win") seems to work well: https://treeherder.mozilla.org/#/jobs?repo=try&revision=937dc2aedb7881b0c65f5b733a633edba565e80e
Flags: needinfo?(james)
Attachment #8885986 - Flags: review?(james)
Whiteboard: [gfx-noted] → [gfx-noted] [stockwell needswork]
Comment on attachment 8885986 [details] [diff] [review] simplify some windows test expectations Review of attachment 8885986 [details] [diff] [review]: ----------------------------------------------------------------- so much simpler!
Attachment #8885986 - Flags: review?(james) → review+
Pushed by gbrown@mozilla.com: https://hg.mozilla.org/integration/mozilla-inbound/rev/91f429bd8dc0 Modify test expectations for some wpt 2dcontext/path-objects tests, on Windows; r=jmaher
Keywords: leave-open
Whiteboard: [gfx-noted] [stockwell needswork] → [gfx-noted] [stockwell disabled]
FYI the reason they aren't collapsed by default is just that the script for updating the metadata isn't clever enough to work out it can do that (partially because it's just not that good and partially because it doesn't know what the full range of possible values is). I will try to work on that as part of the wpt sync improvements this quater.
Severity: normal → S3
See Also: → 1820608
Status: NEW → RESOLVED
Closed: 1 month ago
Resolution: --- → INCOMPLETE
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Created:
Updated:
Size: