On Wed, Jul 14, 2021 at 4:08 PM Thomas Munro <thomas.munro@gmail.com> wrote:
> On Wed, Jul 14, 2021 at 6:17 AM Tom Lane <tgl@sss.pgh.pa.us> wrote:
> > (I suppose a hacky solution might be to never define USE_SIGWAIT
> > on Solaris.)
>
> Yeah I was thinking about that. I've just got a shell on an illumos
> VM and will try a couple of ideas out if I can remember how to drive
> this thing... more soon.
I decided to try to make it work properly on that OS instead of the
hacky solution now that I have access. I took your last patch, moved
-D_POSIX_PTHREAD_SEMANTICS into CFLAGS in src/template/solaris,
removed the special flags and libs from the new configure check where
you had them, and then used HAVE_POSIX_SIGWAIT directly without the
c.h change. I can't find any evidence on the 'net that any other OS
cares about _POSIX_PTHREAD_SEMANTICS (every reference says it's a
Solaris thing) so this should just remove some useless clutter, and
it's not like we want POSIX draft 6 behaviour anywhere. (It's a bit
weird to define a macro that has PTHREAD in the name when building
things that aren't multi-threaded, but that was their choice).
Do you see any problems with this approach? The main thing I can
think of is that people who are using a custom CFLAGS on that OS will
need to update the value they're using, but that seems OK (they'll
likely see the new error and be alerted to the change). Another thing
I noticed is that config/ax_pthread.m4 (something from autoconf that
we don't want to modify) will add another copy of
-D_POSIX_PTHREAD_SEMANTICS, but that's already the case.