You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently, we cannot use memoryview in SageMath because of what appears to be a cython bug: cython/cython#6442
As recommended in cython documentation, typed memoryview is preferred to the numpy.ndarray[...] syntax.
Proposed Solution
We disable fast_getattr as suggested in the cython bug linked above.
Side note, the commit that introduces fast_getattr is 7f33b86 which is #10493
--- however I can't find any discussion on why the fast_getattr is added?
Alternatives Considered
Additional Information
No response
Is there an existing issue for this?
I have searched the existing issues for a bug report that matches the one I want to file, without success.
The text was updated successfully, but these errors were encountered:
My working theory is that the directive isn't actually doing much - it prevents Cython from creating a __getattr__ C method in the class. However that just affects what happens if you do Cls.__getattr__ or Cls().__getattr__. I don't think it has any impact at all on the speed of attribute lookup.
So it'd be good to know if I'm missing something and there's actually a good reason for this.
Problem Description
Currently, we cannot use memoryview in SageMath because of what appears to be a cython bug: cython/cython#6442
As recommended in cython documentation, typed memoryview is preferred to the
numpy.ndarray[...]
syntax.Proposed Solution
We disable
fast_getattr
as suggested in the cython bug linked above.Side note, the commit that introduces
fast_getattr
is 7f33b86 which is #10493--- however I can't find any discussion on why the
fast_getattr
is added?Alternatives Considered
Additional Information
No response
Is there an existing issue for this?
The text was updated successfully, but these errors were encountered: