Somewhere in this regression range: http://hg.mozilla.org/mozilla-central/pushloghtml?tochange=fb81c9a433e4&fromchange=fe7a433de8ff we've started failing on Plugin tests (Linux & Win, Mac doesn't run these): SUMMARY-UNEXPECTED-FAIL | e:\buildbot\comm-central-win32-opt-unittest-mozmill\build\mozmill\content-tabs\test-plugin-crashing.js | test-plugin-crashing.js::test_crashed_plugin_notification_bar EXCEPTION: Timeout waiting for alert at: utils.js line 429 TimeoutError("Timeout waiting for alert") utils.js 429 waitFor((function () this.alerted),"Timeout waiting for alert",10000,100,[object Object]) utils.js 467 controller.js 648 test-content-tab-helpers.js 115 test_crashed_plugin_notification_bar() test-plugin-crashing.js 227 frame.js 557 frame.js 626 frame.js 669 frame.js 497 frame.js 675 server.js 179 server.js 183 SUMMARY-PASS | test-plugin-crashing.js::setupTest SUMMARY-PASS | test-plugin-crashing.js::teardownTest SUMMARY-UNEXPECTED-FAIL | e:\buildbot\comm-central-win32-opt-unittest-mozmill\build\mozmill\content-tabs\test-plugin-crashing.js | test-plugin-crashing.js::test_crashed_plugin_notification_inline EXCEPTION: Expected the status to be "noReport" or "please" at: test-folder-display-helpers.js line 2842 assert_true(false,"Expected the status to be \"noReport\" or \"please\"") test-folder-display-helpers.js 2842 test_crashed_plugin_notification_inline() test-plugin-crashing.js 265 frame.js 557 frame.js 626 frame.js 669 frame.js 497 frame.js 675 server.js 179 server.js 183

For help, an ARPL log link: http://clicky.visophyte.org/tools/arbpl-mozmill-standalone/?log=http://tinderbox.mozilla.org/showlog.cgi?log=ThunderbirdTrunk/1328832897.1328833568.23102.gz&fulltext=1

Csets in that range explicitly mentioning plugins and thus good culprits to investigate: http://hg.mozilla.org/mozilla-central/rev/606a6ab0d0a6 Bug 626813 - Hide plugins in background tabs to avoid trackpad drivers wanting to scroll them. r=jmathies http://hg.mozilla.org/mozilla-central/rev/874062cc7df7 Fix for nsPluginNativeWindowWin.cpp: 'return' : truncation from 'nsresult' to 'bool'. r=josh http://hg.mozilla.org/mozilla-central/rev/1a345b043b47 Bug 724886: Push images from plugin code instead of pulling from nsObjectFrame. r=roc

I'd already looked at those, and the first two wouldn't explain the linux failure, and the third one doesn't look like it would cause this as we're not comparing what the plugin shows. So this probably needs a bisect running on it to find the actual failure spot.

Sigh, I should have spotted this earlier, but thanks to a bad backout comment, I didn't: http://hg.mozilla.org/mozilla-central/rev/2087f423a1a4 I've therefore landed the same backouts as I've been doing on aurora for it: http://hg.mozilla.org/comm-central/rev/eda26cb3231e