[FFmpeg-devel] [PATCH] avcodec: improve avcodec_receive_frame() doxy

[Andreas Rheinhardt]

Anton Khirnov:
> Make it clear that the returned reference is owned by the caller.
> ---
>  libavcodec/avcodec.h | 8 ++++----
>  1 file changed, 4 insertions(+), 4 deletions(-)
> 
> diff --git a/libavcodec/avcodec.h b/libavcodec/avcodec.h
> index c9baf859ac..cadd213301 100644
> --- a/libavcodec/avcodec.h
> +++ b/libavcodec/avcodec.h
> @@ -3728,10 +3728,10 @@ int avcodec_send_packet(AVCodecContext *avctx, const AVPacket *avpkt);
>   * Return decoded output data from a decoder.
>   *
>   * @param avctx codec context
> - * @param frame This will be set to a reference-counted video or audio
> - *              frame (depending on the decoder type) allocated by the
> - *              decoder. Note that the function will always call
> - *              av_frame_unref(frame) before doing anything else.
> + * @param frame This parameter should point to a "clean" (unallocated) frame,
> + *              into which the decoder will write a reference to the decoded
> + *              frame. The returned reference is owned by the caller and must be
> + *              freed with av_frame_unref().
>   *
>   * @return
>   *      0:                 success, a frame was returned
> 
You are adding a requirement for the frame to be clean initially. This
doesn't seem to be intended, as the commit message doesn't mention it.

- Andreas