Get/set n-th bit of BigUint and BigInt

[janmarthedal]

This PR implements bit and set_bit for BigUint and BigInt.

The method names have been chosen to match those of Ramp (https://docs.rs/ramp/0.5.9/ramp/int/struct.Int.html#method.bit).

For BigInt the implementation uses the number's two's complement representation when the number is negative. This matches what libraries like Ramp or languages like Python do.

Resolves #172

[cuviper]

Thanks! The two's complement stuff is a bit hairy, as expected, but I don't see any problem with it.

[janmarthedal]

@cuviper I have now updated the PR with a more efficient implementation of BigInt::set_bit. Basically, I have made sure that only the relevant digits will be iterated over instead of all of them, as you also pointed out. In a special benchmark it saved about 70% in runtime.

I have also implemented a special case for bit < trailing_zeros && value which saves some runtime compared to the method used in the case bit == trailing_zeros && !value, but only about 5-10%. The two cases can be merged to make set_bit a bit shorter.

[mkaczanowski]

@cuviper it's been a while since the last review :) Is this PR all good? (asking, since I am waiting for this feature)

[cuviper]

I added a couple tweaks to the edge cases of u64 to usize, but otherwise it looks good!

bors r+

[bors]

Build failed:

• Test (1.34.0)

[cuviper]

bors r+

[bors]

Build succeeded:

• Format
• No Std
• Test (1.31.0)
• Test (1.32.0)
• Test (1.34.0)
• Test (1.36.0)
• Test (beta)
• Test (i686)
• Test (nightly)
• Test (stable)