Revision r192138 <http://trac.webkit.org/r192138> has caused the following 6 new failures: Regressions: Unexpected image-only failures (6) imported/mozilla/svg/filter-scaled-02.html [ ImageOnlyFailure ] imported/mozilla/svg/image/image-filter-01.svg [ ImageOnlyFailure ] svg/filters/color-space-conversion.svg [ ImageOnlyFailure ] svg/filters/container-with-filters.svg [ ImageOnlyFailure ] svg/filters/feComposite-background-rect-control-operators.svg [ ImageOnlyFailure ] svg/filters/feDisplacementMap-filterUnits.svg [ ImageOnlyFailure ] I have tested to revert locally r192138 on top of r194637 and the above 6 tests pass again. The diffs are the following: https://build.webkit.org/results/GTK%20Linux%2064-bit%20Release%20%28Tests%29/r194637%20%2812965%29/imported/mozilla/svg/filter-scaled-02-diffs.html https://build.webkit.org/results/GTK%20Linux%2064-bit%20Release%20%28Tests%29/r194637%20%2812965%29/imported/mozilla/svg/image/image-filter-01-diffs.html https://build.webkit.org/results/GTK%20Linux%2064-bit%20Release%20%28Tests%29/r194637%20%2812965%29/svg/filters/color-space-conversion-diffs.html https://build.webkit.org/results/GTK%20Linux%2064-bit%20Release%20%28Tests%29/r194637%20%2812965%29/svg/filters/container-with-filters-diffs.html https://build.webkit.org/results/GTK%20Linux%2064-bit%20Release%20%28Tests%29/r194637%20%2812965%29/svg/filters/feComposite-background-rect-control-operators-diffs.html https://build.webkit.org/results/GTK%20Linux%2064-bit%20Release%20%28Tests%29/r194637%20%2812965%29/svg/filters/feDisplacementMap-filterUnits-diffs.html Note: imported/mozilla/svg/image/image-filter-01.svg that is marked as failing on the general TestExpectations because of bug 139282 was passing on the GTK platform before r192138 (it was marked as passing on the GTK TestExpectations).
Why did GTK EWS go green on the original patch? Does it not run tests?
(In reply to comment #1) > Why did GTK EWS go green on the original patch? Does it not run tests? It does not run tests, the GTK EWS only tests if the build passes or not.
Simon fixed all of these tests (and *many* more) in r225797 "Remove ColorSpaceDeviceRGB and most users of the obsolete deviceRGB colorspace". Exception: imported/mozilla/svg/filter-scaled-02.html is still broken, so we cannot close this bug yet.
Nice.
The test(s) filed under this bug have been consistently passing for the last 4000 revisions. Marking bug as fixed. Committed r263254: <https://trac.webkit.org/changeset/263254>
<rdar://problem/64522506>