[Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.9.2a1pre) Gecko/20090715 SeaMonkey/2.1a1pre] (home, optim default) (W2Ksp4) (http://hg.mozilla.org/mozilla-central/rev/d190d9b6ccd1 +http://hg.mozilla.org/comm-central/rev/773809454cf2 + bug 503724 patch) { ... ... | Page2: Ensure we could scrol. - Didn't expect 0, but got it. ... ... | Page2Again: Ensure we can still scroll. - Didn't expect 0, but got it. } NB: While there, add missing 'l' to 'scrol'. (I don't know if this is anyhow related to bug 465096.)

[Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.9.2a1pre) Gecko/20090715 SeaMonkey/2.0b1pre] (experimental/_m-c_, home, optim default) (W2Ksp4) (http://hg.mozilla.org/mozilla-central/rev/1057ca8f2a91 +http://hg.mozilla.org/comm-central/rev/d1a9b79cf5c2) Passes. (http://hg.mozilla.org/mozilla-central/rev/cabb8925dcd3 +http://hg.mozilla.org/comm-central/rev/d1a9b79cf5c2) Fails. Regression timeframe: https://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=1057ca8f2a91&tochange=cabb8925dcd3 2 changesets: bug 488846 is Firefox only, then this is a "Bug 178324, refactor focus" fallout. This is probably a case of bug 497839.

What's the issue here? Is test_bug430723 failing when you add some patch to seamonkey branch?

(In reply to comment #2) > What's the issue here? The test used to pass even without being focused. > Is test_bug430723 failing when you add some patch to seamonkey branch? No sure what you mean.

(In reply to comment #3) > (In reply to comment #2) > The test used to pass even without being focused. Maybe a regression from the great focus rewrite then. Where is the test failing and under what conditions? > > Is test_bug430723 failing when you add some patch to seamonkey branch? > > No sure what you mean. Your bug report is unclear. I was trying to guess what you meant based on the you comments provided.

(In reply to comment #4) > Maybe a regression from the great focus rewrite then. Comment 1 says it is... > Where is the test failing and under what conditions? Comment 0 says 'Page2' and 'Page2Again' checks... Summary says 'when test window does not have focus'... (As in: run the test command from shell, (quickly) switch to another window, let the test load and run in background, ...) > Your bug report is unclear. Well, I thought it couldn't be clearer: has condition, error report, regression timeframe, ...

(In reply to comment #5) > Summary says 'when test window does not have focus'... > (As in: run the test command from shell, (quickly) switch to another window, > let the test load and run in background, ...) You should always leave the focus on the mochitest window. It's known to have problems when it loses focus. Many tests fail if the window loses focus.

(In reply to comment #6) > It's known to have problems when it loses focus. It is ... and this bug is about the regression affecting this very test...

From what I understand from it, eventutils.sendkeyevent can't be expected to work when the window doesn't have focus. That this apparently used to work is a bug, which is now fixed by bug 178324.

Linux mozilla-central test mochitests on 2009/07/22 19:35:28 http://tinderbox.mozilla.org/showlog.cgi?log=Firefox-Unittest/1248316528.1248318609.21339.gz

WINNT 5.2 mozilla-central test mochitests on 2009/07/28 10:26:40 http://tinderbox.mozilla.org/showlog.cgi?log=Firefox-Unittest/1248802000.1248804768.21084.gz

29085 ERROR TEST-UNEXPECTED-FAIL | /tests/docshell/test/navigation/test_bug430723.html | Page2: Ensure we could scrol. - Didn't expect 0, but got it. http://tinderbox.mozilla.org/showlog.cgi?log=Firefox-Unittest/1249925504.1249927242.28886.gz

Linux mozilla-central unit test on 2009/08/20 17:46:20 http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1250815580.1250822663.11076.gz 31093 ERROR TEST-UNEXPECTED-FAIL | /tests/docshell/test/navigation/test_bug430723.html | Page2: Ensure we could scrol. - Didn't expect 0, but got it. etc.

Linux mozilla-central test mochitests on 2009/09/01 19:45:24 http://tinderbox.mozilla.org/showlog.cgi?log=Firefox-Unittest/1251859524.1251861874.13258.gz

http://tinderbox.mozilla.org/showlog.cgi?log=Firefox-Unittest/1253114345.1253116296.23138.gz Linux mozilla-central test mochitests on 2009/09/16 08:19:05 ... | Page2: Ensure we could scrol. - Didn't expect 0, but got it.

http://tinderbox.mozilla.org/showlog.cgi?log=Firefox3.6-Unittest/1258457641.1258461442.15374.gz WINNT 5.2 mozilla-1.9.2 test mochitests on 2009/11/17 03:34:01

http://tinderbox.mozilla.org/showlog.cgi?log=Firefox3.6-Unittest/1258602637.1258605859.22542.gz Linux mozilla-1.9.2 test mochitests on 2009/11/18 19:50:37

The cause isn't a focus problem. The scroll by arrow keys are asynchronous scrolling. So, the scroll might not finish by next executeSoon. So, the tests should listen to scroll events before sending key events. try: https://tbpl.mozilla.org/?tree=Try&usebuildbot=1&rev=f670c36557ec

Comment on attachment 580056 [details] [diff] [review] Patch oops, we shouldn't send key events directly from event handlers. I'll post a new patch.