At Thu, 15 Feb 2018 10:01:57 -0800, Andres Freund <andres@anarazel.de> wrote in
<20180215180157.7q55aytrhif7vvzw@alap3.anarazel.de>
> On 2018-02-15 20:11:07 +0900, Kyotaro HORIGUCHI wrote:
> > As in another mail just before, spin.c seems a bit strange
> > (without acutual harm).
> >
> > spin.h doesn't include pg_sema.h when HAVE_SPINLOCKS is defined,
> > but spin.c always includes it even in the case. The file is
> > included only to use sizeof(PGSemaphore) to calcualte
> > SpinlockSemaSize as 0.
> >
> > The codes that requires PGSempaphore is inactivated when the
> > symbol is defined in all other places so it seems to me that we
> > ought to refrain from using it there, too. The attched patch does
>
> IDK, I don't quite see the point of the change here...
No actual gain, but I just feel it uneasy that utterly-unused
symbol is used to yield zero. From other point of view, it seems
to be inconsistency that a header file is disabled in a file, but
not disabled in another on one configuration.
regards.
--
Kyotaro Horiguchi
NTT Open Source Software Center