Bug 242194

Summary:	REGRESSION (251723@main): [ iOS ] compositing/images/positioned-image-content-rect.html is a flaky image failure
Product:	WebKit	Reporter:	Karl Rackler <rackler>
Component:	Compositing	Assignee:	Nobody <webkit-unassigned>
Status:	NEW ---
Severity:	Normal	CC:	ap, mattwoodrow, simon.fraser, webkit-bot-watchers-bugzilla, webkit-bug-importer
Priority:	P2	Keywords:	InRadar
Version:	WebKit Nightly Build
Hardware:	Unspecified
OS:	Unspecified

Description Karl Rackler 2022-06-30 10:42:36 PDT

Description:
compositing/images/positioned-image-content-rect.html

The first failure that I saw on the dashboard was on 6/21/2022 at 251723@main.

History:
https://results.webkit.org/?suite=layout-tests&test=compositing%2Fimages%2Fpositioned-image-content-rect.html&platform=ios&limit=50000

Image Diff:
https://build.webkit.org/results/Apple-iOS-15-Simulator-Release-arm64-WK2-Tests/251723@main%20(94)/compositing/images/positioned-image-content-rect-diffs.html

Comment 1 Radar WebKit Bug Importer 2022-06-30 10:43:09 PDT

<rdar://problem/96230716>

Comment 2 Karl Rackler 2022-06-30 11:02:21 PDT

REPRODUCTION STEPS
I can reproduce this on 251723@main, but unable to reproduce it on 251722@main or earlier.

Command: 
run-webkit-tests --ios-simulator --debug --exit-after-n-failures 1 --exit-after-n-crashes-or-timeouts 1 --iterations 50 --no-retry compositing/images/positioned-image-content-rect.html

Result: 
Unexpected flakiness: image-only failures (1)
  compositing/images/positioned-image-content-rect.html [ ImageOnlyFailure Pass ]

Comment 3 Karl Rackler 2022-06-30 11:04:58 PDT

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

Comment 4 EWS 2022-06-30 11:12:02 PDT

Test gardening commit 252003@main (769fc294900a): <https://commits.webkit.org/252003@main>

Reviewed commits have been landed. Closing PR #1953 and removing active labels.

Comment 5 Alexey Proskuryakov 2022-07-04 12:03:51 PDT

Looks like an antialiasing or upscaling issue.

Would there be any problem with reverting 251723@main?

Comment 6 Simon Fraser (smfr) 2022-07-05 10:56:54 PDT

Looks like one is a non-Retina image. Possibly affected by an earlier test in the sequence.

Comment 7 Alexey Proskuryakov 2022-07-05 11:54:09 PDT

> Would there be any problem with reverting 251723@main?

I didn't try, but it seems unlikely that it can be reverted cleanly, because of 251965@main.