media/video-playing-and-pause.html test is flakey: http://build.webkit.org/results/Apple%20MountainLion%20Debug%20WK2%20(Tests)/r146137%20(7888)/results.html Not sure if this is a flexbox or subpixel issue.
We still have this platform expectation. Blink has similar name test, which uses 'testharness', should we try to import it, if it test something similar?