[MPlayer-cvslog] r29123 - trunk/libmpdemux/parse_mp4.c

[Uoti Urpala]

On Thu, 2009-04-02 at 00:20 +0200, Diego Biurrun wrote:
> On Thu, Apr 02, 2009 at 01:15:22AM +0300, Uoti Urpala wrote:
> > On Thu, 2009-04-02 at 00:07 +0200, Diego Biurrun wrote:
> > > On Wed, Apr 01, 2009 at 10:41:47PM +0300, Uoti Urpala wrote:
> > > > On Wed, 2009-04-01 at 19:09 +0200, Diego Biurrun wrote:
> > > > > On Wed, Apr 01, 2009 at 07:48:32PM +0300, Uoti Urpala wrote:
> > > > > > On Wed, 2009-04-01 at 18:28 +0200, diego wrote:
> > > > > > > 
> > > > > > > Log:
> > > > > > > Add config.h #include, necessary for HAVE_MALLOC_H check.
> > > > > > 
> > > > > > You should first verify whether the check is correct and needed - as far
> > > > > > as I can tell it isn't.
> > > > > 
> > > > > How can the test be incorrect and unneeded?  malloc.h is not available
> > > > > everywhere..
> > > > 
> > > > In the other direction - what would fail without HAVE_MALLOC_H?
> > > 
> > > Some systems lack malloc.h, compilation would fail on those systems.
> > 
> > I'm not sure just how you interpreted "without HAVE_MALLOC_H",
> 
> Without the line
> 
>   #if HAVE_MALLOC_H
> 
> present in the file, seems like the natural interpretation to me.
> 
> > but I meant "in case HAVE_MALLOC_H was not defined". In other words I
> > see no reason why that check would ever need to _succeed_.
> 
> If it's undefined we get a warning, if it's defined to 0, the header is
> skipped, if it's defined to 1, the header is included.  We have the
> desired behavior now.

And the advantage over deleting the whole #ifdef stuff is? Is there
anything that actually needs it?