Indefinite prime sieve

[tfpf]

Serves to document my first attempt at implementing an indefinite prime sieve. Inspired by this Stack Overflow question.

Benchmarks indicate that my sieve of Atkin implementation is way faster than this implementation when both are executed up to a definite upper bound.

     Running benches/benches.rs (target/release/deps/benches-4a4f39c75b2fd568)
sieve_of_atkin          time:   [732.33 µs 733.87 µs 735.51 µs]
primes                  time:   [16.117 ms 16.137 ms 16.160 ms]