Re: Possible GCC optimization bug

Segher Boessenkool <[email protected]> Thu, 12 Feb 2026 16:11:38 -0600

Newsgroups	gmane.comp.gcc.help
Message-ID	<aY5QGmJGPIbegL0g@gate>

On Thu, Feb 12, 2026 at 04:35:38PM +0000, Anubis 1101 via Gcc-help wrote:
> That does appear to "fix" it, but it doesn't tell me what was actually wrong.

"unsigned" in C (and C++, it inherited it) does not mean "not signed".
Instead it means that your numbers are not part of (a submodule of) ℤ
but instead of ℤ/Mℤ (with M some power of two typically, usually 2**32
or maybe even 2**64).  x >0 in some "unsigned integer" thing means
exactly the same as "x != 0".


Segher