-
Notifications
You must be signed in to change notification settings - Fork 521
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
h5py install fails on M1 #1934
Comments
I'm also having this problem but with python 3.8 on an m1 cpu. I believe this is the cause/fix: numpy/numpy#18160. Numpy's devs appear to be working out a patch that removes the A workaround might be disabling pep517, but I didn't have much luck with the --no-use-pep517 (using pipenv myself) but I did successfully install h5py via git cloning this repo and changing the referenced numpy version in
I'm sure that's pinned for a reason; in my case I'm not using data stored in hdf5 format with numpy, so no harm. |
@NAThompson , @hagan correctly identified that as a numpy installation issue. @hagan Those are the minimum versions of numpy we support, it is expected that h5py will work with any newer version. When you access data from a dataset we will return a numpy array and use numpy heavily internally to handle the buffers that libhdf5 returns so you are definitely reliant on numpy working correctly for h5py to work correctly! |
I'm going to close this because there is nothing we can do on the h5py side as this is a numpy build / installation issue. |
@tacaswell I just talked to the NumPy devs, and they said they would not be making another release of NumPy 1.19. They recommended using oldest-supported-numpy, which has a stanza for Darwin + arm64. See numpy/numpy#18160 (comment) and the README of oldest-supported-numpy for some additional detail. With this information, should this issue be reopened? |
@tacaswell I've opened #1993, which I believe should fix this. Could you take a look and let me know what you think? |
Background info:
To reproduce:
The text was updated successfully, but these errors were encountered: