[vtk-developers] Mysterious test failures...

[Bill Lorensen]

Dave,

Try running the test manually on one of those machines. Maybe some
message is not being captured by the test driver?

Bill

2008/2/27 David Cole <david.cole at kitware.com>:
> Last Friday, we started getting 78 tests failing on krondor and midworld,
> two old Macs we have here at Kitware running OSX 10.3.
>
> Failing test list for krondor:
> http://www.vtk.org/Testing/Sites/krondor.kitware/Darwin-c++-carbon/20080222-0300-Nightly/Test.html
>
> These are the files that were updated for this build:
> http://www.vtk.org/Testing/Sites/krondor.kitware/Darwin-c++-carbon/20080222-0300-Nightly/Update.html
>
> For the life of me, I do not see which of these changes might have caused
> these tests to start failing on that day....
>
>
> For the test "expCos" (and many of the others have similar output) the
> following output is emitted when a vtkRenderWindowInteractor is created:
> ERROR: In /Users/kitware/Dashboards/My
> Tests/VTK/Filtering/vtkDemandDrivenPipeline.cxx, line
> 686
> vtkStreamingDemandDrivenPipeline (0x2b05c60): Algorithm
> vtkOutlineSource(0x2b05700) did not create output for port 0 when asked by
> REQUEST_DATA_OBJECT and does not specify any DATA_TYPE_NAME.
> (...same output repeated 6 more times...)
> And then the test crashes on line 65 of Graphics/Testing/Cxx/expCos.cxx
> because "input" is NULL...
> Any clues why this might have started happening on this day? To the best of
> my knowledge, nothing changed on the machines other than updating the source
> tree every night...
>
>
>
> Thanks,
> David Cole
>
>
>
> _______________________________________________
> vtk-developers mailing list
> vtk-developers at vtk.org
> http://www.vtk.org/mailman/listinfo/vtk-developers
>
>