Dashboard cleanup

ShapeLabelMapFixture.3D_T2x2x2_Spacing test has been failing for a while now. Can we fix it? @blowekamp If I remember correctly, you were the last fiddling with it. Is this still on your to-do list?

@jhlegarreta Do you know if the three failing dynamic analysis tests are false positives, or is there some real problem with them? For example,
itkSimpleFastMutexLockTest does not seem to touch file descriptors at all. If they are false positive, can we prevent them from being reported on the dashboard?

The other two are itkDiscreteHessianGaussianImageFunctionTest and itkTimeVaryingBSplineVelocityFieldImageRegistrationTest.

Here a simple patch for it:

Different issue, but still dashboard cleanup: PythonLazyLoadingImage is failing, but this patch should solve the issue.

Not an expert on dynamic analysis, but gave it a try.

I may be more time to look into detail, but at first sight, using an up-to-date master (Win 7; 64-bit; MSVC 2010):

  • itkSimpleFastMutexLockTest is passing locally.

  • itkTimeVaryingBSplineVelocityFieldImageRegistrationTest is supposedly passing (there is no comparison to a baseline).

  • I run into the same issues for the itkDiscreteHessianGaussianImageFunctionTest. I need to have a closer look.

OK, so my guess is that it may be related to the timeout. Running the test manually works, and the test finishes successfully (excluding the comparison to the baseline which I haven’t tested).

A timeout error presence depends on the machine the test is run, so I’ll set a value that I consider is fair; but we’ll need to check again once the patch set gets hopefully merged.

Change-Id 22799 proposes the TIMEOUT workaround for itkDiscreteHessianGaussianImageFunctionTest.