[MPlayer-dev-eng] [PATCH] initial DragonFly BSD support

[Joey Parrish]

On Tue, Aug 31, 2004 at 08:05:24PM +0200, Bernd Ernesti wrote:
> On Tue, Aug 31, 2004 at 07:20:48PM +0200, Pawel Biernacki wrote:
> [..]
> > Ok, is this patch is good enough to be commited?
> 
> No.
> 
> [..]
> > --- /dev/null   2004-08-31 20:19:47.000000000 +0200
> > +++ osdep/lrintf.c      2004-08-31 21:01:39.000000000 +0200
> > @@ -0,0 +1,7 @@
> > +#include "../config.h"
> > +
> > +#ifndef lrint
> > +#include <math.h>
> > +#define        lrintf(x) rintf(x)
> > +#endif
> > +
> 
> Why are you checking for lrint?
> 
> Thats seems for me plain wrong.
> Hint: HAVE_LRINTF
> 
> Did you even test that on a system which has lrintf?
> I would say no.

I agree.  Use HAVE_LRINTF.

Also, this lrintf.c will not do anything, because that #definition will
only have the scope of that one file.  You must write an inline c
function.  Also, you might try this:

#ifndef HAVE_LRINTF
inline long int lrintf(float blah) { ... }
#endif

And then unconditionally compile that file.  For systems with lrintf,
the object is empty, but this is done for many demuxers already.  It's
not a new thing in the MPlayer source.

--Joey

-- 
"There's no blood in your brain?  You're one of the lucky ones." --Chad