[selectors] Update :focus-visible tests from WPT
Created attachment 426097 [details] Patch
Created attachment 426101 [details] Patch
Created attachment 426103 [details] Patch
Comment on attachment 426103 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=426103&action=review > LayoutTests/imported/w3c/web-platform-tests/css/selectors/focus-visible-002.html:33 > - This test checks that <code>:focus-visible</code> always matches on <code><input></code> elements which take text input, regardless of focus mechanism. > + This test checks that <code>:focus-visible</code> always matches on <code><input></code> elements which take text input, regardless of focus mechanism. Do we know who made this type of change in WPT and why?
Created attachment 426122 [details] Patch
(In reply to Darin Adler from comment #4) > Comment on attachment 426103 [details] > Patch > > View in context: > https://bugs.webkit.org/attachment.cgi?id=426103&action=review > > > LayoutTests/imported/w3c/web-platform-tests/css/selectors/focus-visible-002.html:33 > > - This test checks that <code>:focus-visible</code> always matches on <code><input></code> elements which take text input, regardless of focus mechanism. > > + This test checks that <code>:focus-visible</code> always matches on <code><input></code> elements which take text input, regardless of focus mechanism. > > Do we know who made this type of change in WPT and why? The test in WPT hasn't changed, so maybe it's the import script doing the change now.
(In reply to Manuel Rego Casasnovas from comment #6) > The test in WPT hasn't changed, so maybe it's the import script doing the > change now. Can you research this? Like changes have there been to the import script since we imported this correctly? I’m really concerned about damaging the tests like this while importing them. For now it might be harmless? I would like to help fix this.
(In reply to Darin Adler from comment #7) > (In reply to Manuel Rego Casasnovas from comment #6) > > The test in WPT hasn't changed, so maybe it's the import script doing the > > change now. > > Can you research this? Like changes have there been to the import script > since we imported this correctly? I’m really concerned about damaging the > tests like this while importing them. For now it might be harmless? > > I would like to help fix this. Yeah, I'll investigate it, and report my findings on a separated bug. Maybe the first time I copied the tests manually instead of using the import (cannot remember, though I usually go with the import script), or maybe there were some recent change on the import script.
Created attachment 426182 [details] Patch
Created attachment 426184 [details] Patch
Created attachment 426193 [details] Patch for landing
(In reply to Manuel Rego Casasnovas from comment #8) > (In reply to Darin Adler from comment #7) > > (In reply to Manuel Rego Casasnovas from comment #6) > > > The test in WPT hasn't changed, so maybe it's the import script doing the > > > change now. > > > > Can you research this? Like changes have there been to the import script > > since we imported this correctly? I’m really concerned about damaging the > > tests like this while importing them. For now it might be harmless? > > > > I would like to help fix this. > > Yeah, I'll investigate it, and report my findings on a separated bug. > > Maybe the first time I copied the tests manually instead of using the import > (cannot remember, though I usually go with the import script), or maybe > there were some recent change on the import script. This was an issue in Python 3.5 or bigger, reported it with a fix: https://bugs.webkit.org/show_bug.cgi?id=224658 I've used python2 on the last patch, so we avoid those modifications. Nice catch Darin.
Committed r276127 (236622@main): <https://commits.webkit.org/236622@main> All reviewed patches have been landed. Closing bug and clearing flags on attachment 426193 [details].
Looks to me like the thing we landed does still have the changes due to the bug in the expected files!
I’m concerned that some of the files now have incorrect expectations in them, but are also skipped. Not a great state to be in.
One example: web-platform-tests/css/selectors/focus-visible-002-expected.txt I don’t think this should have been changed at all as part of this import.
Reopening to attach new patch.
Created attachment 426394 [details] Patch
<rdar://problem/76827947>
Created attachment 426395 [details] Patch
Committed r276237 (236719@main): <https://commits.webkit.org/236719@main> All reviewed patches have been landed. Closing bug and clearing flags on attachment 426395 [details].