inline btoi to reduce compile times

[benmkw]

This PR inlines the calls to btoi into gix-utils to reduce compile times (removes num-traits as well).
Compared to the previous diff I extracted basic tests from the doc comments and created a macro to make it easier to add new types/ reduce noise.
If needed I can squash the commits.
I have https://github.com/tamasfe/taplo installed and it reformatted the touched Cargo.toml files, I can revert that as well if wanted.

ref discussion in #729 (comment)

[Byron]

Thanks a lot for your tremendous help with this!

I will merge once CI is green after my refactor round.

[fbstj]

Suggested change

	/// minimal subset of traits used by btoi_radix and btou_radix
	/// minimal subset of traits used by [`to_signed_with_radix `] and [`to_signed_with_radix`]

or however it is that doclinks (or just generally improve this doc?)

[Byron]

That's a great catch, thanks for the suggestion.
I didn't apply the fix because it uses to_signed_with_radix twice, but applied the intended fix locally and force-pushed.

[fbstj]

have you considered changing this to

Suggested change

	///
	fn zero() -> Self;
	/// the 0 value of this type
	const ZERO: Self;

[fbstj]

ahhh apologies for doing this post-merge, the browser hadn't refreshed

[Byron]

Oh, I see! Some of these could probably be const, maybe const-fn in traits is allowed too?

In any case, if you would like to submit a PR, that would be very welcome.

[benmkw]

I just took this from num-traits with minor changes (changed &x parameters of checked_* to pass by value cause its only used for small copy types here). Making it an associated const seems reasonable to me as well.

[fbstj]

I realized later in the evening that it was likely to do with MSRV that such consts were not used, but maybe that is 'too conservative'?

I'll see about a PR if that's ok with benmkw :)