Make handling of FORCE_COLOR environment variable match the emerging standard

[AlexWaygood]

See the conversation in https://github.com/python/mypy/pull/13814/files#r990652344: the emerging standard for this is to treat the presence of the FORCE_COLOR environment variable as indicating that color is desired, even if the variable is set to "0" or "False". This PR makes mypy's behaviour here match the behaviour of pytest, nox, rich and pypa/build in how they all handle this environment variable.

(It doesn't match the behaviour of tox, but tox seems to be on its own here.)

[github-actions]

According to mypy_primer, this change has no effect on the checked open source code. 🤖🎉

[hauntsaninja]

Unless there's actually some thought out standard and not just tired maintainers merging quick PRs, let's treat FORCE_COLOR=0 as not forcing colour. Setting env vars to 0 or 1 is about as standard as it gets and it's weird that it's inconsistent with MYPY_FORCE_COLOR. Also looks like it's not just tox, it's node too.

[AlexWaygood]

it's weird that it's inconsistent with MYPY_FORCE_COLOR

Yeah, that's fair. I don't really have a strong opinion one way or another :)

@hugovk might have some thoughts!

[hugovk]

If starting from scratch I think it would make sense to have 0 disable it, but it's been quite widely implemented so 0 also enables it. 🤷

There's some value to following other implementations, but I agree it's pretty weird to set FORCE_COLOR=0 and expect colour enabled! And if I saw it, I would definitely recommend against using FORCE_COLOR=0 to enable.

Personally, as long as I can set FORCE_COLOR=1 on GitHub Actions and get coloured mypy output I'll be happy :)