Regression: fast/hidpi/image-srcset-relative-svg-canvas-2x.html is consistently failing on iOS EWS.
This test was passing earlier on iOS, since yesterday (Jan 29) it started failing consistently on iOS EWS.
First failure in:
- Build: https://ews-build.webkit.org/#/builders/24/builds/9513 (the ToT used in this build was r255337)
- Result: https://ews-build.webkit.org/results/iOS-13-Simulator-WK2-Tests-EWS/r389098-9513-clean-tree/results.html
Seems like a regression from r255331 (Bug 206254: REGRESSION: [ Mac ] fast/hidpi/image-srcset-relative-svg-canvas-2x.html is a flaky failure)
This issue doesn't seems to be happening on build.webkit.org bots https://results.webkit.org/?suite=layout-tests&test=fast%2Fhidpi%2Fimage-srcset-relative-svg-canvas-2x.html
This is an antialiasing diff on circle border.
This sometimes means that a preceding test changes antialiasing settings, which affects subsequent tests. May be reproducible by running same tests as this single process does on EWS, in the same order.
Marked as failing in https://trac.webkit.org/changeset/255462/webkit in an attempt to speed up EWS.
(In reply to Ryan Haddad from comment #5)
> Marked as failing in https://trac.webkit.org/changeset/255462/webkit in an
> attempt to speed up EWS.
I think this caused another test to fail https://bugs.webkit.org/show_bug.cgi?id=207038
We'll need to repro this locally.
Test Expectations for this test were updated again so as to run this in EWS in https://trac.webkit.org/changeset/257091/webkit (See https://bugs.webkit.org/show_bug.cgi?id=207038#c11)
This has again started failing on EWS. e.g.: