http://bugs.winehq.org/show_bug.cgi?id=32003
Bug #: 32003 Summary: Properly account for WineTestBot errors Product: Wine-Testbot Version: unspecified Platform: x86 OS/Version: Linux Status: NEW Severity: normal Priority: P2 Component: unknown AssignedTo: wine-bugs@winehq.org ReportedBy: fgouget@codeweavers.com Classification: Unclassified
Sometimes a test cannot be run on a Windows VM for a reason internal to WineTestBot rather than due to the patch author. This can be due to a WineTestBot bug or due to an administrator error. For instance: * A Linux VM that got declared as a Windows one. * TestAgentd was not started on the VM. * The VM is not reachable because it has the wrong IP address or some such.
In those cases WineTestBot entirely reports something like this in the log: * Can't copy exe to VM (Bad IP address) * Child process died unexpectedly (Linux VM used as a Windows one)
And yet on the summary page it shows 'completed - 0 failures'. Instead the summary page should either show something like 'completed with errors - 0 failures' or 'completed - 1 failures' so one knows something is wrong.
http://bugs.winehq.org/show_bug.cgi?id=32003
François Gouget fgouget@codeweavers.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|NEW |ASSIGNED AssignedTo|wine-bugs@winehq.org |fgouget@codeweavers.com
--- Comment #1 from François Gouget fgouget@codeweavers.com 2012-10-24 17:24:32 CDT --- The logic for UpdateStatus() and Check(), where it is duplicated, is broken: if a task failed its step is marked as failed but the job won't be marked as failed unless the step was a build step. Presumably this was done this way to avoid having test failures cause a job to be marked as failed. However a task status only reflects whether the task ran as expected. So if there are test failures the TestFailures field will be non-zero and the Status 'completed'.
I will send a patch to fix this and deduplicate the code.
http://bugs.winehq.org/show_bug.cgi?id=32003
--- Comment #2 from François Gouget fgouget@codeweavers.com 2013-01-30 10:38:20 CST --- This is mostly done:
commit 4c266ab9e904222186b0fae0c28b190b80d22ce1 Author: Francois Gouget fgouget@codeweavers.com Date: Fri Nov 2 00:11:00 2012 +0100
testbot: Fix the Job and Step status update.
A job that has a non-build step fail should still be marked as failed. That's because when there are test failures the TestFailures field gets set to a non-zero value but the task's Status is still set to completed. So a failed Step really reflects a bot error. Also simplify the code and avoid duplicating the UpdateStatus() logic in the Check() method.
However now if a patch fails to apply or compile the job is marked as failed which may not be appropriate.
http://bugs.winehq.org/show_bug.cgi?id=32003
François Gouget fgouget@codeweavers.com changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|ASSIGNED |RESOLVED Resolution| |FIXED
--- Comment #3 from François Gouget fgouget@codeweavers.com 2013-04-12 10:45:42 CDT --- The remaining issue got fixed by the following commit:
commit fab277d195f56acd2e5e6078045077df215fdfa5 Author: Francois Gouget fgouget@codeweavers.com Date: Fri Mar 29 14:02:16 2013 +0100
testbot: Add new status values to distinguish bad patches, build errors and TestBot errors.
Also rename the 'failed' status to 'boterror' to make its meaning clearer and avoid confusion with test failures.
http://bugs.winehq.org/show_bug.cgi?id=32003
André H. nerv@dawncrow.de changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED CC| |nerv@dawncrow.de
--- Comment #4 from André H. nerv@dawncrow.de 2013-04-13 16:33:12 CDT --- closing