Attachment 268999 Details for Bug 471067

IRC log discussing the problem

irc.log (text/plain), 3.22 KB, created by Brian Merrell on 2009-01-30 19:06:30 UTC

(hide)

Description:

Filename:

MIME Type:

Creator: Brian Merrell

Created: 2009-01-30 19:06:30 UTC

Size: 3.22 KB

patch

obsolete

>10:49 <@bgmerrell> i have kind of a crappy situation, that seems to be getting worse.  When i try to quit a winforms application with Strongwind (usually using Alt+F4), the application doesn't seeem to exit all of the way.. the window closes, but Strongwind just waits and eventually times out, which causes the test to fail
>10:50 <@bgmerrell> we had a bug about the applications not exiting all of the way and it was fixed
>10:50 <@bgmerrell> and i can't ever duplicate this problem manually
>10:51 <@bgmerrell> even though Strongwind just issues a alt+f4 key combo to the desktop
>10:52 <@bgmerrell> but out of 4 smoke test runs, 3 of them have failed because of this problem
>10:52 <@bgmerrell> (doesn't ever happen with Gtk that I can tell)
>10:57 <@knocte> bgmerrell: do you see this problem only with strongwind but not when you test it manually ?
>10:57 <@bgmerrell> yes, i haven't tried testing it just with at-spi though.
>10:58 <@bgmerrell> by manually, i mean just physically pressing alt+f4 on my keyboard
>10:59 <@knocte> bgmerrell: can you give a higher timeout to strongwind ?
>10:59 <@mario> bgmerrell: so, pressing alt+f4 hangs the winform app?
>10:59 <@mario> bgmerrell: what app are you testing?
>11:00 <@bgmerrell> knocte: it's already really high
>11:00 <@knocte> :o
>11:00 <@bgmerrell> mario: several, it happens on different ones each time
>11:00 <@bgmerrell> it tries for probably 60 seconds
>11:00 <@bgmerrell> if you run harness/local_run.py -s, chances are that you'll see it
>11:01 <@mario> ok, running.
>11:01 <@bgmerrell> Action: Press <Alt>F4.
>11:01 <@bgmerrell> Expected result: The StatusBar controls window disappears.
>11:01 <@bgmerrell> Watchdog timeout reached; exiting
>11:02 <@bgmerrell> will looking something like that.
>11:02 <@bgmerrell> the watchdog line is Strongwind's
>11:04 <@mario> exiting normally means everything is ok, I guess.
>11:04 <@mario> so, everything is ok.
>11:04 <@mario> rerunning!
>11:04 <@bgmerrell> welcome to my day ;)
>11:04 <@brad> bgmerrell: what is the test environment? openSUSE 11.0?
>11:04 <@bgmerrell> yes
>11:06 <@bgmerrell> the watchdog timeout is 180 seconds.. so Strongwind waits 180 seconds and if nothing happens it kills the test
>11:06 <@mario> I can't reproduce it. the window is closed
>11:07 <@bgmerrell> yeah, it might take several times.. it seems totally random
>11:07 <@bgmerrell> it's been happening for a long time, and i've just been ignoring it
>11:07 <@mario> http://paste2.org/p/138704
>11:07 <@bgmerrell> but now that we're actually automated again, it's important
>11:09 <@bgmerrell> mario: i think that label_smoke_test should be fixed if you svn up
>11:09 <@bgmerrell> (line 75)
>11:10 <@bgmerrell> The window actually does close when this happens
>11:11 <@bgmerrell> but the process doesn't seem to... let me check what hangs around
>11:13 <@bgmerrell> yeah.. i duplicated it on the first try
>11:15 <@bgmerrell> http://rafb.net/p/W2l0he54.html
>11:15 <@bgmerrell> probably most interesting at the very bottom
>11:16 <@bgmerrell> so the window has closed, but the winforms app process is still there (line 114)
>11:18 <@bgmerrell> http://rafb.net/p/WKvysb24.html
>11:18 <@bgmerrell> Strongwind test at the top, ps ax at the bottom
>11:20 <@mario> ok, I reproduced it
>11:20 <@mario> is random
>11:21 <@mario> http://paste2.org/p/138710

Actions: View

Attachments on bug 471067: 268995 | 268996 | 268999