bl-bldxp01 is failing to test with output Running MozillaAliveTest test ... Timeout = 45 seconds. Begin: Sat Jun 9 11:50:04 2007 cmd = c:/builds/tinderbox/Fx-Trunk/WINNT_5.1_Depend/firefox/firefox/firefox.exe -P default End: Sat Jun 9 11:50:12 2007 ----------- Output from MozillaAliveTest ------------- ----------- End Output from MozillaAliveTest --------- Error: firefox.exe: exited with status 5 MozillaAliveTest: test failed This is after bug 383742 eating the storage for the tinderbox VM, and rhelmer disabling BreadkPad for bug 383861. The build does pass MozillaAliveTest on fx-win32-tbox. When I try to run the 2007060904 build on Win XP then the crash reporter comes up but fails to submit a report. If I disable BreakPad using MOZ_CRASHREPORTER_DISABLE=1 then the Windows crash reporter appears but not Talkback. However, there are several reports from nightly testers that the build is fine on Vista, and single reports it's ok on XP & 2000. http://forums.mozillazine.org/viewtopic.php?t=556878 I tried the previous build, http://hourly-archive.localgho.st/win32/20070608_0322_firefox-3.0a6pre.en-US.win32.zip and that starts up fine. Which gives a regression window of http://hourly-archive.localgho.st/win32/20070609_0420_checkins.html We (wildly??) speculating on IRC that this might be similar to bug 331404.

Tested with a new XP Sp2 Profile and latest trunk. When i start Minefield the first time, its hanging around in the taskmgr with 13 MB use of Ram, but its not starting. When i close the Firefox.exe Process and restart minefield, i get the crash reporter and so i can confirm comment #0 from Nick. Vista and Linux is fine here.

I can't reproduce this on XP SP2, it starts fine for me.

(In reply to comment #0) > We (wildly??) speculating on IRC that this might be similar to bug 331404. > Does that mean the checkin for bug 350616 is suspected as triggering this issue?

(In reply to comment #3) Yes. I ran findold.exe from bug 331404 but it reported no files with old dates.

Looks like the same situation on Thunderbird trunk. 08-Jun-2007 03:22 partial update was OK Next partial defaulted to a full, and crashed on startup. No talkback. Full houry zip downloads since then, still crash on startup. This is with winxp sp2 on a p3-s

I've tried backing out the patch for bug 350616, the first cycle still crashed. I then tried triggering a clobber, I'm waiting for those results.

The 2007061004 nightly starts up fine, so the backout would appear to have fixed the problem. The performance test boxes are not cycling because of bug 383919.

(In reply to comment #7) > The 2007061004 nightly starts up fine, so the backout would appear to have > fixed the problem. This is testing on the same machine as I had the problem with the 20070609 build.

I was able to reproduce on a VM, and I submitted: https://crash-reports.mozilla.com/reports/report/index/27a59de6-1758-11dc-a54e-001a4bd43ed6 but it hasn't processed yet.

Ok, bl-bldxp01 is green and the numbers are consistent with previous values over 3 cycles.

Are you sure it isn't regressed again ? I got the error again in the 2007061206 build from http://ftp.mozilla.org/pub/mozilla.org/firefox/nightly/2007-06-12-06-trunk/ (currently latest-trunk), on Windows XP SP2. But 20070611 is fine. Again, Breakpad isn't able to submit the crash-report.

(In reply to comment #11) > Are you sure it isn't regressed again ? I got the error again in the 2007061206 > build from > http://ftp.mozilla.org/pub/mozilla.org/firefox/nightly/2007-06-12-06-trunk/ > (currently latest-trunk), on Windows XP SP2. But 20070611 is fine. Again, > Breakpad isn't able to submit the crash-report. > Sorry to reply on myself - it's now handled in bug 384179