[Bazel] Fix USB dependencies, __init__.py creation

[RyanDraves]

The 2.3.0 release seems to depend on nonexistent target for the USB reset library. A la raspberrypi/picotool#351, new rules_py warnings spam about the py_binary usage in the repo; these changes silence the warning.

I didn't make an issue, sorry for the hygiene.

[armandomontanez]

Sorry for the silence on this. I think we should definitely fix the dependency issue.

The Python warning is newly introduced in the latest version of rules Python. I think there's a path to rolling that out in a way where rules_python doesn't require this to be set on every py_test and py_bianry. Warnings like this are very noisy, I'll try to spin up a proposed rules_python rollout mechanism today.

If I can't get that out today, I'll approve this as a short-erm workaround. In the interim, I recommend early-adopting --incompatible_default_to_explicit_init_py.

[RyanDraves]

Thanks. I didn't have success with --incompatible_default_to_explicit_init_py with pico-sdk as a dependency, but I'll try and give it another go today as well to see if that silences it. Ironically - very ironically - I also got a ton of warning spam from rules_py's py_binary's involved in the build process. Ugh.

[RyanDraves]

Thanks. I didn't have success with --incompatible_default_to_explicit_init_py with pico-sdk as a dependency, but I'll try and give it another go today as well to see if that silences it. Ironically - very ironically - I also got a ton of warning spam from rules_py's py_binary's involved in the build process. Ugh.

I was mistaken; common --incompatible_default_to_explicit_init_py does indeed silence the warnings. Totally fair to not make the spam your problem if you just want the dependency fix.

[kilograham]

I'm going to merge this as is since i'm not sure what common --incompatible_default_to_explicit_init_py means ;-)

feel free to add another PR to fix up

[lurch]

The Bazel "buildifier" check is also causing the CI to fail on this PR...

[RyanDraves]

The Bazel "buildifier" check is also causing the CI to fail on this PR...

My bad. Could I interest y'all in a bazel_env setup in another PR?

[armandomontanez]

Sent out bazel-contrib/rules_python#3997 to hopefully pave a better path for the broader Bazel Python world.