@olebole - I think this fixes the segmentation faults in our code that occurred in https://github.com/astrofrog/fast-histogram/issues/60 but I'm now running into segmentation faults in Numpy itself during the computation of the reference values. That will take a bit longer to investigate. These are basically popping up because of new test cases Hypothesis is doing, but I need to figure out what is triggering the segfaults on the Numpy side.
@olebole - I think this fixes the segmentation faults in our code that occurred in https://github.com/astrofrog/fast-histogram/issues/60 but I'm now running into segmentation faults in Numpy itself during the computation of the reference values. That will take a bit longer to investigate. These are basically popping up because of new test cases Hypothesis is doing, but I need to figure out what is triggering the segfaults on the Numpy side.