On Tue, Jun 03, 2014 at 03:56:09PM -0400, Andrew Stone wrote:
> On Tue, Jun 3, 2014 at 3:21 PM, Diego Biurrun <[email protected]> wrote:
> >> --- a/libavutil/dict.c
> >> +++ b/libavutil/dict.c
> >> @@ -186,7 +186,7 @@ void av_dict_free(AVDictionary **pm)
> >>
> >> -void av_dict_copy(AVDictionary **dst, AVDictionary *src, int flags)
> >> +void av_dict_copy(AVDictionary **dst, const AVDictionary *src, int flags)
> >>  {
> >>      AVDictionaryEntry *t = NULL;
> >>
> >> --- a/libavutil/dict.h
> >> +++ b/libavutil/dict.h
> >> @@ -141,7 +141,7 @@ int av_dict_parse_string(AVDictionary **pm, const char 
> >> *str,
> >>   * @param flags flags to use when setting entries in *dst
> >>   * @note metadata is read using the AV_DICT_IGNORE_SUFFIX flag
> >>   */
> >> -void av_dict_copy(AVDictionary **dst, AVDictionary *src, int flags);
> >> +void av_dict_copy(AVDictionary **dst, const AVDictionary *src, int flags);
> >
> > This looks unrelated.
> 
> For the various av_opt_set_* functions that take pointers, all values
> are const. To keep things consistent, av_opt_set_dict_val also uses a
> const AVDictionary* as the val, and then passes it to av_dict_copy. To
> allow that, though, it was necessary to change av_dict_copy to accept
> a const pointer, or just suffer through a compiler warning or cast.
> Changing to a const* seemed best.

My point was that this change could be done as a separate patch, which I
have just sent.

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

Reply via email to