On 04/29/2010 01:14 AM, Paolo Bonzini wrote:
On 04/29/2010 05:41 AM, Eric Blake wrote:
> Gnulib guarantees that pthread.h exists, but for now, it is a dummy
> header with no support for most pthread_* functions. Modify our
> use of pthread to use function checks, rather than header checks,
> to determine how much pthread support is present.
Shouldn't this be 1/5 for the sake of bisectability?
1/5 passes 'make check' independently on Linux, but I see your point at
how it will fail on mingw (link errors) if this is not squashed in. So
I'll go ahead and squash 1 and 2 into a single patch before pushing (I'm
also trying to get a mingw build going before I push).
> -#ifdef HAVE_PTHREAD_H
> +#ifdef HAVE_PTHREAD_SIGMASK
> sigset_t oldmask, blockedsigs;
> #endif
If there is no pthread_sigmask but there is sigprocmask, it is correct
to use that too. But this should be done in gnulib though, so I suppose
the patch is fine as is.
Ultimately, it would be nice if gnulib guaranteed pthread_sigmask, but
we're not there yet. You are right, though, that gnulib guarantees
sigprocmask, even form mingw, so it will probably be worth investigating
in a future patch. At any rate, incremental steps are better than
nothing :)
--
Eric Blake eblake(a)redhat.com +1-801-349-2682
Libvirt virtualization library
http://libvirt.org