Incorrect 'N' for scrypt.hash

[chrisveness]

It's taken me far longer for the penny to drop than it probably should have, but If I understand well, scrypt.hash is misinterpreting the N parameter.

In the params for scrypt.kdf, the N is actually logN – which while the naming is a little confusing, is much easier to work with than N itself, and I like.

However, in the params for scrypt.hash, the N parameter is actually N.

Hence scrypt.hash treats { N: 16, r: 8, p: 1} as if it were { N: 4, r: 8, p: 1}, resulting in a work factor of 16 instead of 65536.

[BrandonZacharie]

This is handled by #119

[chrisveness]

Good stuff, thanks.

I'll close once the fix hits npm (so it doesn't get re-reported if anyone else hits it).

Excellent library, thanks for the work.