shell.nix improvements, and fix problems on Darwin

[purcell]

Improvements to shell.nix were made in #8910 by @thorstenweber83, but the result was that the nix shell no longer worked on Darwin.

I have tidied up and improved some of the Nix expression, fixing this issue along the way. If someone would be kind enough to test this on Linux too, that would be appreciated, though I don't expect issues.

Description

• Removed unnecessary overrides
• Provided a more complete Python environment, adequate for actually running/developing bin/qmk
• Unset NIX_TARGET_CFLAGS_COMPILE in the Nix shell, to avoid incompatible host CC flags getting picked up by avr-gcc.

Types of Changes

• Core
• Bugfix
• New feature
• Enhancement/optimization
• Keyboard (addition or update)
• Keymap/layout/userspace (addition or update)
• Documentation

Issues Fixed or Closed by This PR

• None

Checklist

• My code follows the code style of this project.
• My change requires a change to the documentation.
• I have updated the documentation accordingly.
• I have read the CONTRIBUTING document.
• I have added tests to cover my changes.
• I have tested the changes and verified that they work and don't break anything (as well as I can manage).

[thorstenweber83]

LGTM.

make all:default successfully built 997 of 1084 keyboards on my NixOS machine.

Edit: actually that does not look that good considering there must be nearly 100 boards failing.
I think i'm a bit confused about how many keyboards there are and how many i should be able to build.
Looking at the travis log i can see 980 lines of "Making xyz with keymap default".
I counted my firmware files and got to 997 ❓ .

Did you test make all:default?

Also running the tests with make test:all gives me an error in one of the serial_link tests.
I'll add some more details later.

[purcell]

I think i'm a bit confused about how many keyboards there are and how many i should be able to build.

I guess the key question is whether you get more or fewer failures/successes with or without this PR. I can imagine blanking out NIX_TARGET_CFLAGS_COMPILE might have negative effects in some cases, but it was actively necessary for me to do that with avr-gcc.

[thorstenweber83]

I guess the key question is whether you get more or fewer failures/successes with or without this PR.

I had failures before, too and it was about the same number. So it does not look like your PR makes things worse on linux.

[purcell]

So it does not look like your PR makes things worse on linux.

Cool, thanks for confirming - this should be unproblematic to merge, then.

[purcell]

Thanks @Erovia and @fauxpark