Guilherme Leobas

Sure. Take a look at the file [`numba/core/typing/mathdecl.py`](https://github.com/numba/numba/blob/main/numba/core/typing/mathdecl.py). I think `math.nextafter` typing rule should look similar to `math.hypot`: https://github.com/numba/numba/blob/3edb458c69daeaaffbfc938145c367eda1cf043f/numba/core/typing/mathdecl.py#L79-L86

@gmarkall, it seems that the gpuCI failure is not related to this PR?

@gmarkall, CI seems to be all green. Can you confirm on GPU CI?

I've submitted https://github.com/numba/numba/pull/9416 which partially fixes this issue for the CPU case. GPU case should be easy to address as well.

CUDA support for np.log2 is here: https://github.com/numba/numba/pull/9417

Thanks for the contribution @i7878. Could you also include a [release notes](https://numba.readthedocs.io/en/stable/developer/contributing.html#release-notes)?

@gmarkall, I've labeled it as ready to merge as the CI failure is not related to this PR.

The cpu and parallel implementations of guvectorize do not share the same logic under the hood. I've submitted https://github.com/numba/numba/pull/9377 which should fix this.

@jaredjeya, could you include the tests from NumPy for the set of functions added in this PR? https://github.com/numpy/numpy/blob/b0371ef240560e78b651a5d7c9407ae3212a3d56/numpy/lib/tests/test_arraysetops.py#L15

> According to the [release notes](https://numpy.org/doc/stable/release/1.24.0-notes.html#faster-version-of-np-isin-and-np-in1d-for-integer-arrays), pre-1.24, the kind="sort" behaviour was the default. Newer versions automatically pick the optimal algorithm. I think we can have an if-else statement to support...