[Webkit-unassigned] [Bug 77541] SVG pixel tests can't catch regressions anymore

bugzilla-daemon at webkit.org bugzilla-daemon at webkit.org
Fri Feb 3 03:37:24 PST 2012


--- Comment #1 from Nikolas Zimmermann <zimmermann at kde.org>  2012-02-03 03:37:23 PST ---
I went back until Oct 15, where its still broken. Can't go back much further, as Safari depends on a symbol thats not present in older WebKit2 libraries.
Anyhow, I've compared DRT, the WebKit2-testrunner to Safari, and can only reproduce the "old behavior" with crude hacks forcing repaints inside DRT. Still this is not reliable.

I've decided to stop wasting my time and instead fix the SVG tests. I'll grep for all occurrences of setTimeout and/or layoutTestController.waitUntilDone(), and will have to fix all testcases that don't use dumpAsText, to correctly enforce repainting through layoutTestController.display() and forcing a layout before, by eg. document.rootElement.offsetLeft.

As I initially said, I share James opinion that the tests have to be fixed, but I was confident it was a recent change that changed the timing. What recently changed on my side is that I got a new fast machine now. It turns out the repainting trouble I ran into happens _much_ less frequently on my older MBP from 2006 that I used as main machine until December. Now with my shiny iMac, it's easy to see that there are obvious races between dumping & painting, that I won't be able to fix using any setTimeout tricks.

/me goes fixing the SVG tests once for all.

Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

More information about the webkit-unassigned mailing list