Open musicinmybrain opened 1 month ago
Does https://github.com/numpy/numpy/pull/27579 fix this?
Does #27579 fix this?
Well – it was supposed to, and I could have sworn I validated it, but I can still reproduce this.
@musicinmybrain this is a simple test failure. The test uses self.assertRaises(TypeError, max, [0, -1, 2, -3])
, but whatever the underlying function does, it converts the -1
to a uint8, etc.
Just replacing it with 0, 1, 2, 3
(removing the negatives) should be OK, I think... Or use some other random set of positive numbers.
The maybe bigger thing is: apparently we have tests for this swig code (I didn't even know!) and they are not run by our CI...
Describe the issue:
Working on Fedora 40, with swig 4.2.1, several errors of the form
ERROR: testMaxWrongDim (__main__.ucharTestCase.testMaxWrongDim)
occur when attempting to run the SWIG tests.Reproduce the code example:
Error message:
Python and NumPy Versions:
Runtime Environment:
Context for the issue:
I noticed these errors while attempting to investigate a more serious incompatibility with SWIG 4.3. It’s not immediately obvious to me what the correct fix would be.