AES* generate_key argument name differs in actual class vs .pyi

[jborean93]

The definition for the AESGCM.generate_key class method defines the first argument as bit_length: usize whereas the AESGCM.generate_key pyi definition is key_size. This causes issues with mypy validation if calling this method with an explicit kwarg like

AESGCM.generate_key(bit_length=256)

There are other AES* classes that also have a bit_length argument in the source but the .pyi defines it as key_size. These should align and to avoid backwards compatibility problems the .pyi files should be changed to bit_length. If you are happy with this change I can send through a PR with those changes.

[alex]

Sigh, good catch.

Yes, we'd be happy to take a patch, and you're correct that bit_length is the right name.

[jborean93]

Have opened #12170. I'm unsure if you want a test for one or all classes to call this method with the kwarg but just left it to the .pyi changes for now.