just found out that round(-.2) in glsl gives you -0 and round(.2) gives you +0

Follow

@SuricrasiaOnline having positive and negative zero is one of the most annoying "features" with IEEE floating point

@ChlorideCull @SuricrasiaOnline if you want edge cases to give you +infinity or -infinity when doing computations, you need to know from which sign you got zero.

@Feufochmar @SuricrasiaOnline It's also pretty logical from a spec standpoint, as the sign is a separate bit, so not allowing both positive and negative zero would be a special case unlike just allowing it to be a thing

Sign in to participate in the conversation

Instance run by a non-profit association, with a mission to encourage an open internet, welcoming to everyone.