Use Element for checking Settings in CSSComputedStyleDeclaration. Right now renderer is used but this will not work for non-rendered elements.
Created attachment 424338 [details]
Created attachment 424343 [details]
Comment on attachment 424343 [details]
8 changes, but only 2 tests; I’d expect we’d need 8 different tests to cover them all. Also, did these tests fail before making the change?
(In reply to Darin Adler from comment #4)
> Comment on attachment 424343 [details]
> 8 changes, but only 2 tests; I’d expect we’d need 8 different tests to cover
> them all. Also, did these tests fail before making the change?
Yes, I verified for aspect-ratio and contain. I missed that scroll-behaviour has a similar test. Should I adapt that one and add a similar if-disabled test for cssIndividualTransformProperties before landing?
Yes, I suggest we make sure we have a test for every change. I don’t feel strongly about how to structure those tests; one test, many, use existing tests that happen to have the correct set-up.
Created attachment 424795 [details]
Looks like some of the new tests are failing on Windows. Maybe DumpRenderTree doesn’t have the necessary support for the CSSIndividualTransformPropertiesEnabled=false on Windows?
Created attachment 424876 [details]
Committed r275351: <https://commits.webkit.org/r275351>
All reviewed patches have been landed. Closing bug and clearing flags on attachment 424876 [details].
(In reply to Darin Adler from comment #8)
> Looks like some of the new tests are failing on Windows. Maybe
> DumpRenderTree doesn’t have the necessary support for the
> CSSIndividualTransformPropertiesEnabled=false on Windows?
Yes, this happened before with the aspect-ratio test, it was marked as failure. Looks like the bug on Windows is that experimental features that are enabled by default can't be disabled for testing. I will leave this to the windows experts.