[Mplayer-cvslog] Re:

[Gábor Lénárt]

On Sat, Nov 23, 2002 at 06:35:33PM +0100, Michael Niedermayer wrote:
> Hi
> 
> On Saturday 23 November 2002 18:25, Michael Niedermayer wrote:
> > On Saturday 23 November 2002 18:15, Alex Beregszaszi wrote:
> > [...]
> >
> > > > > >  	    case MOV_FOURCC('s','t','t','s'): {
> > > > > > -		int temp=stream_read_dword(demuxer->stream);
> > > > > > +		stream_read_dword(demuxer->stream);
> > > > >
> > > > > 1000L !
> > > > > reverse this file !
> > > >
> > > > ok, but could u explain why it is bad? i cant see where temp is used
> > > > ...
> > >
> > > it's not used but at least my gcc 2.95.4 didn't support declaring
> > > variables after some func calls. so:
> > > {
> > > stream_read_dword(demuxer->stream);
> > > int len = stream_read_dword(demuxer->stream);
> > > }
> > > is bad.
> >
> > ooooooops, thats really 1000l
> > only strange is that gcc 3.2 did compile it
> just checked gcc 3.0 compiles it too, only 2.95.4 complained, i dont have any 
> other gcc version installed so dunno about them
> 
> is this a gcc bug? or a feature ;)

Really strange! I thought it's a strictly C rule that you cannot declare variables
elsewhere than at begining of a { ... } block (not sounting global variables).
Sure, maybe this rule is constructed by myself only when learning C ;-)

- Gábor (larta'H)