F32.from_nat wraps values at 2^32 #1609

F32.from_nat narrows through U32.from_nat, so it silently wraps at 2^32 even though these values fit the supported Nat range.

import Base

def main() -> IO(Unit):
  IO.print(F32.show(F32.from_nat(Nat.double(2147483648n))))

On 7e120d5 this prints 0. The input is 4294967296, an exactly representable F32; it should not become zero. The literal is built with Nat.double because Nat literals themselves stop at U32_MAX. Current main has the same definition.

The fix removes the narrowing conversion, with a direct Nat-to-F32 intrinsic on compiled lanes. I checked the 2^32 boundary, rounding just above half an ulp, a tie to even, and the maximum supported Nat on Bun and native C; the existing number kit still passes. AI-assisted investigation and fix.