Nathan Goldbaum
Nathan Goldbaum
Out of curiosity, how much slower does this make the free-threaded CI run?
This is an issue with the current version of spin. I think @stefanv is working on a fix. For now I think installing spin 0.8 should let you get past...
Can you turn actions on for your fork to make sure CI passes?
Did you try running e.g. the example use-case from the readme on the GIL-enabled and free-threaded build?
Looking at [the reverse dependencies](https://www.wheelodex.org/projects/cymem/rdepends/), I only recognize thinc. You could look at how that library is using cymem. The original comment was just towards making sure this is getting...
I think it would probably be ok to just universally return False for equality comparisons with types that aren’t instances of `unyt_array`. Seems sensible to me.
Of course we’d need to treat ndarrays differently, I meant any random python type like string or other unrelated object.
Probably just needs a call to `np.asanyarray` in `__eq__` to handle all array-likes, then anything that ends up with a string or object dtype gets rejected.
Right, that's why I said, > then anything that ends up with a string or object dtype gets rejected I guess also datetimes too. `U5` is a string dtype.
Running this script under py-spy with `--native`: ```python import numpy as np import random options = ["a", "bb", "cc", "ddd"] lst = random.choices(options, k=1000) arr_s = np.array(lst, dtype="T") for _...