with the IR compiler we actually compute the whole thing as a 32-bit array and then try to copy it over. In addition to having the wrong values, this can overwrite other things.
I think it is a type inference issue? Marking one of them as u16 is sufficient to fix it.
If we do something like
with the IR compiler we actually compute the whole thing as a 32-bit array and then try to copy it over. In addition to having the wrong values, this can overwrite other things.
I think it is a type inference issue? Marking one of them as
u16
is sufficient to fix it.