"Theo de Raadt" <dera...@openbsd.org> wrote:
> Nope, that is not correct.
>
> errno is not being cleared. It just happens to be zero. Future
> code changes could insert another operation above which would set
> errno, and then this would print a report about that error.
Although I was being sarcastic with """Everything is alright""", yes,
correct. Point taken.
> No, your diff is still wrong.
>
> errno is only updated when a system call returns -1.
>
> So your diff is looking at an old, unrelated, errno.
How? This is now correctly looking at errno only when {,p}read returns
-1, and is using warnx in the other cases.
