Bug 240814

Summary:	REGRESSION (r293885): [ macOS wk2 ] webrtc/canvas-to-peer-connection.html is a consistent failure
Product:	WebKit	Reporter:	Karl Rackler <rackler>
Component:	WebRTC	Assignee:	Karl Rackler <rackler>
Status:	RESOLVED FIXED
Severity:	Normal	CC:	gsnedders, webkit-bot-watchers-bugzilla, webkit-bug-importer, youennf
Priority:	P2	Keywords:	InRadar
Version:	WebKit Nightly Build
Hardware:	Unspecified
OS:	Unspecified

Description Karl Rackler 2022-05-23 10:21:31 PDT

Description:
webrtc/canvas-to-peer-connection.html

The first failure I saw on the dashboard was on 5/6/2022 at r293885.  r293885 removed usage of PixelBufferConformer in RealtimeOutgoingVideoSourceCocoa so that possibly introduced the consistent failure on macOS X86_64 wk2. 

REPRODUCTION STEPS
I can reproduce this on r293885.

Command: 
run-webkit-tests --root 293885 --exit-after-n-failures 1 --exit-after-n-crashes-or-timeouts 1 --iterations 50 -f --no-retry webrtc/canvas-to-peer-connection.html

Result: 
Regressions: Unexpected text-only failures (1)
  webrtc/canvas-to-peer-connection.html [ Failure ]

History:
https://results.webkit.org/?suite=layout-tests&test=webrtc%2Fcanvas-to-peer-connection.html&platform=mac&limit=50000&architecture=x86_64

Diff:
--- /Volumes/Data/worker/Apple-BigSur-Debug-WK2-Tests/build/layout-test-results/webrtc/canvas-to-peer-connection-expected.txt
+++ /Volumes/Data/worker/Apple-BigSur-Debug-WK2-Tests/build/layout-test-results/webrtc/canvas-to-peer-connection-actual.txt
@@ -1,3 +1,3 @@
 
-PASS draw: withWebGL
+FAIL draw: withWebGL assert_array_approx_equals: rect 0,255,255,255 at 0.9,0.9 not found iteration: 0 property 1, expected 255 +/- 25, expected 255 but got 229

I have marked this test as a consistent failure while this issue is investigated.

Comment 1 Radar WebKit Bug Importer 2022-05-23 10:21:47 PDT

<rdar://problem/93768909>

Comment 2 Karl Rackler 2022-05-23 10:30:19 PDT

Pull request: https://github.com/WebKit/WebKit/pull/925

Comment 3 EWS 2022-05-23 10:32:58 PDT

Test gardening commit r294652 (250876@main): <https://commits.webkit.org/250876@main>

Reviewed commits have been landed. Closing PR #925 and removing active labels.

Comment 4 Alexey Proskuryakov 2022-05-23 19:35:18 PDT

Bots seem to be very certain on r293885 being the culprit, yes.

Comment 5 EWS 2022-05-30 07:06:11 PDT

Committed r295026 (251121@main): <https://commits.webkit.org/251121@main>

Reviewed commits have been landed. Closing PR #1005 and removing active labels.

Comment 6 youenn fablet 2022-05-31 08:10:58 PDT

*** Bug 241006 has been marked as a duplicate of this bug. ***