Updates for numpy=2.

[erykoff]

No description provided.

[beckermr]

We need to somehow get numpy2 into the CI to test the builds. Maybe we wait to merge until numpy2 is official on conda-forge.

[esheldon]

what does copy=None mean in the new numpy?

[erykoff]

I'll restore the string_ for old numpy on the PR. The new behavior in numpy 2 is that copy is always False as I understand it, and I think they left in copy=None as a way to allow backwards compatibility.

[erykoff]

Trying to add py312 and numpy2 to the build matrix. Let's see what happens!

[esheldon]

OK, I guess their release notes are out of date

[erykoff]

This isn't working:

if numpy.lib.NumpyVersion(numpy.__version__) >= "2.0.0":
    _stypes += (numpy.str_,)
else:
    _stypes += (numpy.string_, numpy.str_,)

[erykoff]

It's a dumb rc thing ...

[beckermr]

Yeah copy=None is wrong here. Here are the docs: https://numpy.org/devdocs/release/2.0.0-notes.html#new-copy-keyword-meaning-for-array-and-asarray-constructors

[beckermr]

link is busted

New copy keyword meaning for array and asarray constructors

Now [numpy.array](https://numpy.org/devdocs/reference/generated/numpy.array.html#numpy.array) and [numpy.asarray](https://numpy.org/devdocs/reference/generated/numpy.asarray.html#numpy.asarray) support three values for copy parameter:

None - A copy will only be made if it is necessary.

True - Always make a copy.

False - Never make a copy. If a copy is required a ValueError is raised.

The meaning of False changed as it now raises an exception if a copy is needed.```

[beckermr]

We specify copies or no copies on purpose so properly pass by reference in places iirc.

[erykoff]

So the old numpy1x behavior was copy=False would try to do no copy, but would if it needed? And that is now copy=None I think.

[beckermr]

Ahhhhhhh yes that's right.

doc string in 1.26:

If true (default), then the object is copied. Otherwise, a copy will only be made if __array__ returns a copy, if obj is a nested sequence, or if a copy is needed to satisfy any of the other requirements (dtype, order, etc.).

[erykoff]

I wonder if all our pushes made the system mad and now won't allocate any nodes for the macos builds ... thoughts?

[beckermr]

We probably had too many jobs in the queue and just had to wait for the old ones. We should be able to add bit to the GHA to cancel previous jobs.

[erykoff]

All tests have passed and I think this is ready.