fast/frames/iframe-window-focus.html This test is consistently failing on most macOS wk2 configurations. HISTORY: https://results.webkit.org/?suite=layout-tests&test=fast%2Fframes%2Fiframe-window-focus.html&flavor=wk2&platform=mac TEXT DIFF: This tests that focusing an editable iframe's window works correctly and allows text and newlines to be entered. -TEST +EST PASS DIFF URL: https://build.webkit.org/results/Apple-Sonoma-Release-WK2-Tests/270636%40main%20(623)/fast/frames/iframe-window-focus-pretty-diff.html REPRODUCIBILITY: I was able to reproduce this issue on macOS Sonoma (14.1, 23B74/arm64) ToT using the following command: run-webkit-tests --clobber-old-results --verbose --iterations 100 --exit-after-n-failures 1 fast/frames/iframe-window-focus.html REGRESSION: I am unable to bisect an exact regression point, as there were a string of build failures spanning the range in which we would expect it to be (268626@main - 268640@main). My best guess is that 268632@main caused the failure, as changes seem like they could be related.
<rdar://problem/118340493>
Test gardening commit 270644@main (34368464f184): <https://commits.webkit.org/270644@main> Reviewed commits have been landed. Closing PR #20424 and removing active labels.
Thank you Ben for the initial report and bisection. I was able to fix the compilation issue locally, and further bisect the regression. The regression point is actually https://commits.webkit.org/268639@main aka https://bugs.webkit.org/show_bug.cgi?id=262286 rdar://116167166 The failing test sends keyDown events, and the 268639@main patch is about key events, that seems to confirm my bisection. I'll ping the author to investigate further...
Thanks Gerald! I modified the title to reflect your findings.
Pull request: https://github.com/WebKit/WebKit/pull/20496
Committed 270731@main (7326eec80566): <https://commits.webkit.org/270731@main> Reviewed commits have been landed. Closing PR #20496 and removing active labels.