Re: [PATCH] 1(3) hurd+glibc: Support for file record locking

[Svante Signell]

On Mon, 2019-10-28 at 01:40 +0100, Samuel Thibault wrote:
> There was an issue with rlock-tweak.c, revealed by the tdb testsuite:
> 
> Index: hurd-debian/libfshelp/rlock-tweak.c
> ===================================================================
> --- hurd-debian.orig/libfshelp/rlock-tweak.c
> +++ hurd-debian/libfshelp/rlock-tweak.c
> 
> +         else if (l->start < start && start <= l->start + l->len
> +                  && (len == 0 || start + len > l->start + l->len))
> /* Our start falls within the locked region or exactly one byte after
> it and our end falls beyond it.  We know that we cannot consume the
> entire region.  */
> 
> Apart from that it seemed to work just fine. I'll let the glibc
> testsuite run this night to be sure before committing.

Nice catch. I never tried running the tdb testsuite explicitly during
the years, only when trying to build tdb itself. How did the glibc
testsuite run go?

Thanks!