Avoid unsafe casts from float to unsigned int

[QuLogic]

Note that this adds .view to duck_array_ops; I'm not certain if this is new user API or not.

This works across all non-x86_64 architectures in Fedora. I did not test RISC-V, as that is not a primary architecture, but it is likely the same issue and maybe @U2FsdGVkX1 can give it a try.

• Fixes Test failure on RISC-V platform #9815
• Tests added
• User visible changes (including notable bug fixes) are documented in whats-new.rst
• New functions/methods are listed in api.rst

[Illviljan]

Which package supports this? I can't find xp.view in the standard https://data-apis.org/array-api/draft/API_specification/index.html

[QuLogic]

Oh, I didn't realize this was implementing a standard when @kmuehlbauer brought it up in #9815. As in #9815, it works for NumPy and Dask. I don't know about any other packages other than tests don't seem to fail about it specifically.

[kmuehlbauer]

I didn't check this fully, but to be honest, I also tried it the same way locally.

@Illviljan can you suggest a way forward here, so that it is usable for numpy and dask?

[QuLogic]

I can also try view and fall back to astype(copy=False) if that's more portable. And also not expose view so publicly if desired.

[kmuehlbauer]

@dcherian Could you have a look here, too? This would unblock @QuLogic for some Fedora builds. Is the implementation of view the right way?

[Illviljan]

Using hasattr(__array_namespace__) implies the method is included in the array api standard. But the standard does not include a .view. That's why I simply don't think the __array_namespace__ should be used in this case.

But using .astype(copy=False) does indeed seem like a nice way forward, no?