Open cBournhonesque opened 2 months ago
Thanks, The negative floats problem was already identified in #11 , as for infinity your idea is great.
Something to consider here is that wgsl does not support infinity as a literal, at all. You can't even bypass it by doing something like let x = 1.0 / 0.0;
, the compiler will catch it.
F32::new(f32::INFINITY)
does not work, it expands asinstead of something like
Similar problem with negative floats:
F32::new(-10.0)
does not work