[Ffmpeg-devel] [PATCH] Fix some gcc warnings

[Aurelien Jacobs]

On Wed, 21 Sep 2005 23:59:43 +0200
Michael Niedermayer <michaelni at gmx.at> wrote:

> Hi
> 
> On Wed, Sep 21, 2005 at 11:40:24PM +0200, Aurelien Jacobs wrote:
> > On Wed, 21 Sep 2005 22:12:18 +0100
> > M?ns Rullg?rd <mru at inprovide.com> wrote:
> > 
> > > Aurelien Jacobs <aurel at gnuage.org> writes:
> > > 
> > > > On Wed, 21 Sep 2005 10:28:00 +0100 (BST)
> > > > M?ns Rullg?rd <mru at inprovide.com> wrote:
> > > >
> > > >> 
> > > >> Colin Ward said:
> > > >> >
> > > >> > Michel Bardiaux wrote:
> > > >> >>
> > > >> >
> > > >> >    [Snip]
> > > >> >
> > > >> >    Yay!
> > > >> >
> > > >> >    I noticed that FFMPEG can be a little noisy to compile. 
> > > >Lots of > > "unused variable" type warnings.  I was wondering if
> > > >people would be > > interested in patches for these.  Should I
> > > >submit some or are people > > not really worried?
> > > >> 
> > > >> I thought I had cleaned up most of those.
> > > >
> > > > That's true. But there are still some of them left, especially
> > > > in snow.c. Attached a patch which fixes all the remaining
> > > > "unused variable" I had with my config.
> > > >
> > > > Also note that the patch which initiated this thread don't fix
> > > > "unused variable". It in fact fixes "assignment makes pointer
> > > > from integer without a cast" in bitstream.h which is included in
> > > > a lot of files, and so which causes warnings all along the
> > > > build.
> > > 
> > > Are you using gcc 4, or some extra pedantic flags?  I don't see
> > > those warnings.
> > 
> > Oh, sure, I should have mentioned that I use gcc 4 !
> > And that I also get tons of warnings about pointers which differ
> > in signedness. I may try to workaround those too...
> 
> iam not sure how you want to "workaround" those but
> iam against making the code less readable due to silly signedness
> warnings

I'm aware of this.
But there are lots of cases were (un)signed types are mixed when
they should really not. It's often solved by simply adding or
removing an unsigned somewhere.
Unfortunately there are other cases which are not as simple
and which may have no clean solution.

> [...]
> > > > Index: libavcodec/snow.c
> > > > ===============================================================
> > > > ==== RCS file: /cvsroot/ffmpeg/ffmpeg/libavcodec/snow.c,v
> > > > retrieving revision 1.62
> > > > diff -u -r1.62 snow.c
> > > > --- libavcodec/snow.c	21 Aug 2005 22:17:41 -0000	1.62
> > > > +++ libavcodec/snow.c	21 Sep 2005 20:34:56 -0000
> > > 
> > > I didn't touch snow last time around, since it is fairly actively
> > > worked on, and I didn't want to disrupt any work in progress.
> > 
> > Hum... somewhat true. But there wasn't lots of commits in snow.c
> > since quite some time. Moreover, removed variables should be
> > really harmless (mostly some 'i').
> > Michael, are you ok with this ?
> 
> ok

Done.

Aurel