On Mon, Feb 13, 2012 at 09:53:33AM -0500, Justin Ruggles wrote:
> On 02/13/2012 09:07 AM, Diego Biurrun wrote:
> 
> > On Mon, Feb 13, 2012 at 08:18:09AM -0500, Justin Ruggles wrote:
> >> On 02/13/2012 04:15 AM, Diego Biurrun wrote:
> >>> --- a/libavcodec/ac3enc.c
> >>> +++ b/libavcodec/ac3enc.c
> >>> @@ -1386,8 +1386,7 @@ static void 
> >>> ac3_output_frame_header(AC3EncodeContext *s)
> >>>  static void output_audio_block(AC3EncodeContext *s, int blk)
> >>>  {
> >>> -    int ch, i, baie, bnd, got_cpl;
> >>> -    int av_uninit(ch0);
> >>> +    int ch, i, baie, bnd, got_cpl, ch0;
> >>>      AC3Block *block = &s->blocks[blk];
> >>>  
> >>> @@ -2236,8 +2235,7 @@ static av_cold int 
> >>> validate_options(AC3EncodeContext *s)
> >>>  static av_cold void set_bandwidth(AC3EncodeContext *s)
> >>>  {
> >>> -    int blk, ch;
> >>> -    int av_uninit(cpl_start);
> >>> +    int blk, ch, cpl_start;
> >>>  
> >>> --- a/libavcodec/ac3enc_template.c
> >>> +++ b/libavcodec/ac3enc_template.c
> >>> @@ -335,7 +335,7 @@ static void 
> >>> compute_rematrixing_strategy(AC3EncodeContext *s)
> >>>      int nb_coefs;
> >>>      int blk, bnd, i;
> >>> -    AC3Block *block, *av_uninit(block0);
> >>> +    AC3Block *block, *block0;
> >>
> >> I'm not against removing these, but I want to point out that removing
> >> them causes (invalid) gcc warnings when making seemingly unrelated
> >> changes to the code. I did this in the first place to cut out the
> >> unnecessary noise, but if people really want these removed, that's fine.
> >> I've been told clang gives warnings when we use av_uninit() when it
> >> otherwise would not.
> > 
> > What warnings do you get?  I don't get any new warnings after removing
> > these attributes.
> 
> They may not be there currently, but when I was actively working on the
> code, making unrelated changes would sometimes trigger the "may be used
> uninitialized" warning.

Wouldn't that simply be because you were using too old a gcc version
back then?

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

Reply via email to