The following layout test is failing on Windows and Linux fast/images/color-jpeg-with-color-profile.html There are green sky and purple hill. Probable cause: Unknown
http://test-results.appspot.com/dashboards/flakiness_dashboard.html#showExpectations=true&showLargeExpectations=true&tests=color-jpeg-with
This image requires the interpretation of an ICC v2 profile. This used to work on Chromium Mac, was unintentionally broken around Chrome 12, and has recently been fixed (working in Canary). It also works in the Skia on Mac configuration and on Safari. On Linux and Windows, there's no support for color profiles, although a suitable library has been checked in at third_party/qcms. The steps to connect the two is documented here: http://code.google.com/p/chromium/issues/detail?id=37028#c26 (the "- enable profiles" step is now complete)
This test is failing/flakey on Lion too. Please fix.
(In reply to comment #1) > http://test-results.appspot.com/dashboards/flakiness_dashboard.html#showExpectations=true&showLargeExpectations=true&tests=color-jpeg-with Looking at this flakiness dashboard link, I don't see anything that would make a bot red. The test is marked as "FAIL MISSING" or "PASS FAIL" on all platforms shown there, so what's the problem? In other words, can you please point us at a bot run that is actually red (or even flaky) as a result of this test?
See http://build.webkit.org/TestFailures/#/Lion%20Intel%20Debug%20(Tests) I guess the actual issue is that it has no expected result. Why not?
I added results in http://trac.webkit.org/changeset/97401
(In reply to comment #4) > (In reply to comment #1) > > http://test-results.appspot.com/dashboards/flakiness_dashboard.html#showExpectations=true&showLargeExpectations=true&tests=color-jpeg-with > > Looking at this flakiness dashboard link, I don't see anything that would make a bot red. The test is marked as "FAIL MISSING" or "PASS FAIL" on all platforms shown there, so what's the problem? The expectation should be PASS. We marked it FAIL MISSING or PASS FAIL for now to make bots green. > In other words, can you please point us at a bot run that is actually red (or even flaky) as a result of this test? On the dashboard page, Colors of bars in the flakiness column should be green, not blue or brown. You can see wrong images in the page for "ACTUAL RESULTS: Webkit Win", "ACTUAL RESULTS: Webkit Linux", and so on.
Marking test failures as WontFix. Bug is still accessible and recording in TestExpectations.