Re: [gpsd-dev] Updated docs on NTP segment management

[Eric S. Raymond]

Harlan Stenn <address@hidden>:
> "Eric S. Raymond" writes:
> > Harlan Stenn <address@hidden>:
> > > So aside from your patch, I need to have configure look for stdatomic.h,
> > > and iff we find it we #include stdatomic.h in refclock_shm.c, right?
> > > Once that's done the rest should be OK?
> > 
> > I believe so, yes. Of course I could have a bug in the way shm_query()
> > is mated to shm_timer(), but I think the logging will make that pretty
> > obvious to diagnose if I do.
> 
> OK, after more looking I'm curious about the STD_ATOMIC_H guard - that's
> something that needs to be set too, right?

Yes, I think I mentioned that in my original email.

> If so, I'm wondering if it would be better to use something configure
> will detect for us, either HAVE_STDATOMIC_H (for the header file) or
> HAVE_ATOMIC_THREAD_FENCE (for the function we want to use).

I assumed you would do something like that.

For the record, I hate autotools. I deal with it when I have to; I'm 
even good at beating it into submission.  But I hate it.  Intensely.
-- 
                <a href="http://www.catb.org/~esr/";>Eric S. Raymond</a>