On Thu, Dec 19, 2013 at 04:34:52PM +0200, Martin Storsjö wrote:
> On Thu, 19 Dec 2013, Diego Biurrun wrote:
> >On Wed, Dec 18, 2013 at 09:21:28PM +0200, Martin Storsjö wrote:
> >>On Tue, 17 Dec 2013, Diego Biurrun wrote:
> >>
> >>>Also unify the related ifdeffery in the atomics implementation.
> >>>---
> >>>
> >>>This fixes compilation on systems w/o any atomics implementation available,
> >>>e.g. MinGW32 < 4.0.
> >>
> >>.. by disabling threading silently in those configurations - right?
> >>E.g., configurations where w32threads would be enabled implicitly
> >>(but isn't due to the lack of proper support for atomics), you will
> >>silently get a slightly crippled build, with no threading?
> >>
> >>In general this might work smoother (no failed builds), but is it
> >>better? What do others think? I'm a bit undecided.
> >
> >I'm also slightly undecided, but in any case we should fail early during
> >configure and not late in the compile process.  Other than that, I don't
> >have much of an opinion myself.
> 
> Yes, the earlier we can fail, the better of course.
> 
> I tested this in a setup where atomics aren't supported, and it
> doesn't look all too understandable to the user - configure still
> says "threading support    w32threads" and is built with
> HAVE_W32THREADS=1 - only the main HAVE_THREADS is set to 0. So it's
> not all too easy for the user to realize that threading actually
> wasn't enabled even though all indicators show that it is.

OK, I'll see if there is a way to make all of this more transparent.

Diego
_______________________________________________
libav-devel mailing list
[email protected]
https://lists.libav.org/mailman/listinfo/libav-devel

Reply via email to