✔️ Linting Passed

All linting checks passed. Your pull request is in excellent shape! ☀️

_{Generated for commit: 569ab2b. Link to the linter CI: here}

@betatim I think the CUDA CI is currently failing because the array-api-strict is not updated to the latest version in the CUDA CI. Either we need to update the version or change how we handle maximum as its available in the latest version but not in prior ones.

@betatim I think the CUDA CI is currently failing because the array-api-strict is not updated to the latest version in the CUDA CI. Either we need to update the version or change how we handle maximum as its available in the latest version but not in prior ones.

I thought that updating the lockfile (as I did in this PR) would not fix it, but it seems like it did?

I think supporting more than the latest version of array-api-strict is not wort the effort. The reason for that is that it isn't meant to be used by humans (like Numpy or cupy) but only for testing how compliant your use of the array API is. So I think if there are no errors in the latest version, we are good. What do you think?

The unsovled mystery for me is why the array-api-strict related error went away in this PR? What change did I make that lead to a newer version of it being installed?

I think supporting more than the latest version of array-api-strict is not wort the effort. The reason for that is that it isn't meant to be used by humans (like Numpy or cupy) but only for testing how compliant your use of the array API is. So I think if there are no errors in the latest version, we are good. What do you think?

The unsovled mystery for me is why the array-api-strict related error went away in this PR? What change did I make that lead to a newer version of it being installed?

I agree with you and I think we went with the implementation based on considering the latest version of the api. I think the error is now fixed because the lock file is updated and the supported version in the lock file is now == 2.0.1. Previously array-api-strict was == 1.1.1 in the lock file and that version does not support the maximum function.

Ergh. I looked at the lockfile diff twice and couldn't spot array-api-strict. Thanks for finding it for me.

@lesteve should we merge this or #29373 (comment) and then this. Or something else completely?

Let's merge this and close #29373.