Closed rahulporuri closed 2 years ago
This looks redundant with #1050
This looks redundant with #1050
I don't know which of these is the "right"/"better" solution. @prabhuramachandran comments?
Over on #1096 I'm investigating a related issue having to do with hangs during the test suite run. It appears that with this PR, there are still hangs related to vtkDataEncoder
in the test suite. If we're definitely not wrapping vtkDataEncoder
, it's not clear to me how we're getting hold of an instance of that class in the tests. Is there some other class that uses it that we also need to exclude from the wrapping?
Here's a transcript of an example shell session (zsh, inside VS Code, on macOS 10.15.7 / Intel, with Python 3.9.7 from MacPorts).
Here's the LLDB backtrace from the main thread at the point where the hang occurs:
If we're definitely not wrapping
vtkDataEncoder
, it's not clear to me how we're getting hold of an instance of that class in the tests.
Ah, sorry; this is a test that digs into vtk
directly. We probably need to exclude vtkDataEncoder
explicitly in that test, too, as something that's known to cause issues.
closing given that we've gone with #1050
This PR ignores two VTK classes -
vtkDataEncoder
andvtkWebApplication
to prevent the package build/install from hanging with the latest version of vtk - 9.0.3. We should investigate why these classes are causing the build to hang but for now, we need to ensure that the build doesnt hang. See #1097fixes https://github.com/enthought/mayavi/issues/1073