Nathan Goldbaum

I’m not sure where python-zstandard is coming in on the dependency tree. The build issue is fixed in `main`: https://github.com/indygreg/python-zstandard/issues/231. I’ll ping about doing a release. 3.14 also ships with...

@bschoenmaeckers can you try again? Also cryptography has a cp314t wheel now, IIRC that was showing up here

This PR also needs something like this for `srsly` to become installable: ``` diff --git a/setup.cfg b/setup.cfg index 350d943..fe5f34a 100644 --- a/setup.cfg +++ b/setup.cfg @@ -27,7 +27,7 @@ classifiers =...

Duplicate of https://github.com/numpy/numpy/issues/26850. Is there any particular you're problem with getting a scalar here? If it's a duck typing issue we should fix the issue in the scalars like in...

Maybe we could add a `__dlpack__` implementation (in C) to the scalars that converts to a 0-D array and delegates to the ndarray `__dlpack__`?

> skip_file_prefixes may sometimes be useful, and sometimes just a more cumbersome and verbose way of achieving the same thing - it depends on code structure. From my point-of-view it's...

> I am sure there are other cases where stacklevel is equivalent or better, because a tuple of absolute path to file names constructed with __file__ is a pretty poor...

I opened https://github.com/numpy/numpy/issues/30358 and minimized the comments about that topic in this issue.

It occurs to me: should we keep code for pypy support in-tree? There's a lot of `IS_PYPY` in the tests, a few places in the implementation we check `sys.implementation.name ==...

The issue with only compiling part of your stack with sanitizers is you will have both false positives and false negatives at library boundaries. IMO it's worth the internal consistency...