Repository navigation
np.place segfaults for quaddtype #7
Description
import numpy as np
from numpy_quaddtype import QuadPrecDType
a = np.zeros((18, 1), dtype=QuadPrecDType())
m = np.zeros((18, 1), dtype=np.bool)
m[3:-3] = True
v = np.ones(12, dtype=QuadPrecDType())
np.place(a, m, v)
print(a, m, v)never reaches the print in numpy-quaddtype==0.2.0.
@SwayamInSync has this been fixed since?
Activity
SwayamInSync commentedon Dec 1, 2025
This requires the implementation of
copyswapotherwise if urgent then a similar operationnp.copyto(arr, vals, where=mask)worksSwayamInSync commentedon Dec 1, 2025
@seberg just a heads up on another function that relies on
copyswapand inside
dtypemeta.cit assigns NULL to that function pointer leading to segfaultjuntyr commentedon Dec 1, 2025
np.copytois unfortunately not the right workaround sincenp.placetakes a flat differently sizedvalsarray. However,np.put(arr, np.flatnonzero(mask), vals)seems to workjuntyr commentedon Dec 1, 2025
@SwayamInSync This should definitely be fixed, but since there is a workaround it's not urgent (can be included in the upcoming big release)
seberg commentedon Dec 1, 2025
Yeah, should at least raise a nice error. For most use-cases, I think
a[m] = vis nicer (yes,placehas some funny semantics).juntyr commentedon Jan 16, 2026
@SwayamInSync This is still not fixed in the current
mainbranch versionSwayamInSync commentedon Jan 16, 2026
Yes this and other issues makred at #bugs are depending on the same changes to be made on numpy + adding support to different endian casting within quaddtype
So this might take a while and not that critical as a workaround exists, we will cover this in future releases
juntyr commentedon Jan 16, 2026
Ok, thanks for the update! I pulled some wheels from CI to test v1.0 with my downstream project and removed all bug-workarounds from earlier; this is the only one I still need but everything else seems to be working so far