Sloppy_Math

The smallest value of the numeric type T: the lowest integer, so -128 for s8 and 0 for any unsigned type.

assert(min_of_type(s8) == -128);
assert(min_of_type(s32) == -2147483648);
assert(min_of_type(u16) == 0);
assert(min_of_type(float32) > 0);

The largest finite value of the numeric type T: 255 for u8, Math.FLOAT32_MAX for float32. Any non-numeric type fails an assert.

#import "Math";
assert(max_of_type(u8) == 255);
assert(max_of_type(s64) == 9223372036854775807);
assert(max_of_type(u64) == 0xFFFF_FFFF_FFFF_FFFF);
assert(max_of_type(float32) == FLOAT32_MAX);

True when abs(f) is at most epsilon, which is 0.0001 unless given. The bound is inclusive, and a NaN is never approximately zero.

assert(is_approximately_zero(0.00005));
assert(is_approximately_zero(-0.00005));
assert(!is_approximately_zero(0.01));
assert(is_approximately_zero(0.01, 0.1));

True when f0 and f1 differ by at most epsilon (0.0001 unless given). The test is absolute, not relative, and the bound is inclusive.

#import "Math";
assert(values_are_close(1.0, 1.00005));
assert(!values_are_close(1.0, 1.01));
assert(values_are_close(1.0, 1.01, 0.1));
assert(values_are_close(xyz(1, 2, 3), xyz(1, 2, 3.00001)));
assert(!values_are_close(xy(1, 2), xy(1, 3)));