docs: update VartimeRistrettoPrecomputation documentation

[AaronFeickert]

It is the case that performing variable-time Ristretto multiscalar multiplication evaluation using VartimeRistrettoPrecomputation may be less efficient than the corresponding VartimeMultiscalarMul implementation, depending on the number of points used.

This PR updates the documentation to make this clear, since the decision of whether or not to use precomputation may depend highly on the use case.

[rozbb]

Nice! I can't quite tell what the formatting error is here

[AaronFeickert]

Looks like the formatter doesn't like it when an empty comment line includes a trailing space. Fixed.

[rozbb]

Thank you!