Quoting Diego Biurrun (2016-09-30 15:08:20)
> On Fri, Sep 30, 2016 at 07:28:44AM +0200, Anton Khirnov wrote:
> > Quoting Diego Biurrun (2016-09-29 18:19:13)
> > > On Wed, Sep 28, 2016 at 02:12:53PM +0200, Anton Khirnov wrote:
> > > > Quoting Diego Biurrun (2016-09-08 12:34:14)
> > > > > On Wed, Sep 07, 2016 at 02:59:38PM +0200, Anton Khirnov wrote:
> > > > > > --- a/configure
> > > > > > +++ b/configure
> > > > > > @@ -4533,6 +4555,12 @@ check_header 
> > > > > > VideoDecodeAcceleration/VDADecoder.h
> > > > > >  
> > > > > > +# it seems there are versions of clang in some distros that try to 
> > > > > > use the
> > > > > > +# gcc headers, which explodes for stdatomic
> > > > > > +# so we also check that atomics actually work here
> > > > > > +check_header stdatomic.h &&
> > > > > > +    check_code cc "stdatomic.h" "atomic_int foo; 
> > > > > > atomic_store(&foo, 0)" || disable stdatomic_h
> > > > > 
> > > > > check_func_headers()
> > > > 
> > > > Does not do what I want here.
> > > 
> > > You need to know not only that the header is present, but that the
> > > functions it exposes are usable. So just check for a function, here in
> > > configure, and in the C code. It's what we do in other such cases.
> > 
> > Those things are not really proper functions, they will be typically
> > compiler builtins.
> 
> But you can check for them as if they were functions, can you not?

No, you cannot use its address, since it's not a function but something
expanded by the compiler at compile time. So that test would just always
fail.

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

Reply via email to