I find it weird that for C floats, -0<+0 is not true. Had to write a little bit of extra code to enforce this.

-0 could represent an incrementally small number below zero but greater than the next lower quantisation level.

@mysth +0 is just 0. -0 (0x80000000) is the only extra thing that we have to deal with.

< is orderable and implies not > and not ==
> is orderable and implies not ==
== is an equivalence relation
< U > U == includes all pairs of numeric types

A language that doesn't respect set theory is a bad language.

@inawhile Mathematically yes. That much is obvious. But in C, -0 does not equal +0. Also, 0 is still 0 nobody is disputing it. It’s just about where we put -0 (0x80000000) on the number line for the purposes of the comparison operator.

@r20408e122449d If -0 doesn't equal +0 in C then C is breaking the floating point specification.