http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1312945686.1312946575.15509.gz Rev3 Fedora 12x64 mozilla-central opt test mochitests-2/5 on 2011/08/09 20:08:06 9638 INFO TEST-PASS | /tests/dom/tests/mochitest/bugs/test_resize_move_windows.html | Window height should have changed - 8 should not equal 100 9639 ERROR TEST-UNEXPECTED-FAIL | /tests/dom/tests/mochitest/bugs/test_resize_move_windows.html | Test timed out. XScreenSaver state: Disabled User input has been idle for 906 seconds args: ['/home/cltbld/talos-slave/test/build/bin/screentopng'] 9640 INFO TEST-END | /tests/dom/tests/mochitest/bugs/test_resize_move_windows.html | finished in 326172ms

This orange became much more frequent after bug 769254. I don't know why but I saw an existing leak turn permaorange with that bug; the leak was triggered (in part) by this test. So although I don't know why bug 769254 tickles this randomorange, this bug is now blocking the most important browser-api bug we have. We either need to fix this test, fix whatever code happens to be broken, or disable the test. And we need to do this now. Mounir, you wrote this test; can you try to fix it? I doubt that the increased orange frequency here is due to a bug in my patches, and I can't reproduce the orange on my Windows VM, with or without bug 769254.

I'm not sure whether it's significant, but bug 769254 caused timeouts in several tests besides this one, which suggests a different failure mode than the ones above: https://tbpl.mozilla.org/php/getParsedLog.php?id=13789809&tree=Mozilla-Inbound 6112 ERROR TEST-UNEXPECTED-FAIL | /tests/dom/tests/mochitest/bugs/test_resize_move_windows.html | Test timed out. 6113 ERROR TEST-UNEXPECTED-FAIL | /tests/dom/tests/mochitest/bugs/test_resize_move_windows.html | [SimpleTest.finish()] waitForFocus() was called a different number of times from the number of callbacks run. Maybe the test terminated prematurely -- be sure to use SimpleTest.waitForExplicitFinish(). - got 1, expected 0 14254 ERROR TEST-UNEXPECTED-FAIL | /tests/dom/tests/mochitest/general/test_497898.html | Test timed out. 14470 ERROR TEST-UNEXPECTED-FAIL | /tests/dom/tests/mochitest/general/test_clipboard_events.html | Test timed out. 14538 ERROR TEST-UNEXPECTED-FAIL | /tests/dom/tests/mochitest/general/test_focus_legend_noparent.html | Test timed out. 14539 ERROR TEST-UNEXPECTED-FAIL | (SimpleTest/TestRunner.js) | 4 test timeouts, giving up. 14540 ERROR TEST-UNEXPECTED-FAIL | (SimpleTest/TestRunner.js) | Skipping 67 remaining tests. 14541 ERROR TEST-UNEXPECTED-FAIL | (SimpleTest/TestRunner.js) | [SimpleTest.finish()] waitForFocus() was called a different number of times from the number of callbacks run. Maybe the test terminated prematurely -- be sure to use SimpleTest.waitForExplicitFinish(). - got 1, expected 0 14542 INFO TEST-UNEXPECTED-FAIL | (SimpleTest/TestRunner.js) | finished in a non-clean fashion (in /tests/dom/tests/mochitest/general/test_focus_legend_noparent.html) buildbot.slave.commands.TimeoutError: command timed out: 5400 seconds elapsed, attempting to kill

Something made that bug happening a dozen time a day while it was happening once every few days :(

IME, randomoranges sometimes get tickled into being more or less random by entirely innocent code. I think someone needs to actually fix this test.