// shouldn't be used with doubles, but we get too many of them and
// removing these operators is probably not a good idea
//
- // Maybe we should alway compare doubles up to some "epsilon" precision
+ // Maybe we should always compare doubles up to some "epsilon" precision
#pragma warning(push)
// floating-point equality and inequality comparisons are unreliable