Owen Kirby <[email protected]> writes:
> Actually, that's not entirely correct. I have seen some architectures
> that will turn a shift into a nop when the shift width is greater than
> the word size.
Shifts greater than the word size can give different types of undefined
behaviour. Ignoring higher bits of the shift count is also common, I
think, depending on compiler and instruction set.
But in this case, I think it can't cause any real harm, because these
shifts are inside an
if (authlen >= (0x01ULL << 32)) {
which should be always false if authlen (of type size_t) is 32-bits.
> I have some ARM boards that I can cross compile for, so
> I'll take a look at cleaning up the warnings when I get some time this
> weekend.
I think you get the same warnings if you configure with CC='gcc -m32'
CXX='g++ -m32' on a plain x86_64 gnu/linux box or similar.
Regards,
/Niels
--
Niels Möller. PGP-encrypted email is preferred. Keyid C0B98E26.
Internet email is subject to wholesale government surveillance.
_______________________________________________
nettle-bugs mailing list
[email protected]
http://lists.lysator.liu.se/mailman/listinfo/nettle-bugs
