Edgar Andrés Margffoy Tuay

Hi @poppyschmo, thanks for your input on this one, right now we're stuck until @ccordoba12 obtains write permissions on this repo

Hi @delta003, this PR brakes compatibility with the python-language-server, since we don't have a PR ready there. I don't know what @ccordoba12 thinks about creating a release as the last...

# NumPy mismatch After comparing each signature (across all namespaces) between NumPy and CuPy by using this script https://gist.github.com/andfoy/5ad7e4c10fb82e851b5e359ae44d6eba, it is possible to conclude that the following functions' signatures present...

# SciPy comparison The SciPy comparison is analog to the NumPy one, where functions and modules were compared using this script: https://gist.github.com/andfoy/a5acb14a60d7b46f0dd20fd6e0074d7d #### Notes 1. The following functions cannot be...

@kmaehashi @asi1024, do you know which distributions are still not supporting `size=None` properly?

After a little bit of inspection using the following [script](https://gist.github.com/andfoy/2b52426b4e0bddcb4c264b1f1e9eb4b7), the status regarding the broadcasting support when `size=None`. Checked means that the generator supports array broadcasting on its parameters, otherwise...

@kmaehashi @asi1024 I had a question regarding `randint`, the current implementation makes use of `_interval`, which states that `generates multiple integers independently sampled uniformly from [0, mx]`. What would be...

Nevermind, `uniform` only supports floating types

Thanks for the pointers! I was trying to fiddle with `core.pyx` but I didn't know the functions had to check for `__cupy_get_ndarray__`, I'll give it a go

@kmaehashi @leofang I've just added a test, please let me know if the changes are correct