Re: fossil build failure on riscv64
> On Apr 16, 2025, at 5:24 PM, Barak A. Pearlmutter <barak@pearlmutter.net> wrote:
>
> I don't see why that would be relevant: the issue is that uVal%10 was
> coming out negative, where uVal is an unsigned int.
That’s correct. But‘s cast to „char“ which is signed on x86_64 but unsigned on riscv64 meaning that the resulting cast is differently interpreted on both architectures.
It‘s not an unusual for hidden bugs to show only on certain architectures due to the different behavior.
Adrian
Reply to: