On Mon, 23 Feb 2015 17:19:16 +0100
Michael Niedermayer <michae...@gmx.at> wrote:
> This should allow simplifying many encoders
>
> Signed-off-by: Michael Niedermayer <michae...@gmx.at>
> ---
> libavcodec/utils.c | 12 ++++++++++++
> 1 file changed, 12 insertions(+)
>
> diff --git a/libavcodec/utils.c b/libavcodec/utils.c
> index c5e6300..c6d588a 100644
> --- a/libavcodec/utils.c
> +++ b/libavcodec/utils.c
> @@ -1695,6 +1695,16 @@ int attribute_align_arg avcodec_open2(AVCodecContext
> *avctx, const AVCodec *code
> av_assert0(*(const AVClass **)avctx->priv_data == codec->priv_class);
> }
>
> + if (av_codec_is_encoder(avctx->codec) &&
> + avctx->codec_type == AVMEDIA_TYPE_VIDEO &&
> + !avctx->coded_frame) {
> + avctx->coded_frame = av_frame_alloc();
> + if (avctx->coded_frame) {
> + av_assert0(avctx->coded_frame->key_frame);
> + avctx->coded_frame->pict_type = AV_PICTURE_TYPE_I;
> + }
> + }
> +
> end:
> ff_unlock_avcodec();
> if (options) {
> @@ -2852,6 +2862,8 @@ av_cold int avcodec_close(AVCodecContext *avctx)
> av_freep(&avctx->internal->hwaccel_priv_data);
>
> av_freep(&avctx->internal);
> + if (av_codec_is_encoder(avctx->codec))
> + av_frame_free(&avctx->coded_frame);
This isn't the same condition. You might free the frame, even if you
didn't allocate it.
What's the value in allocating the frame anyway? Seems like it
increases the memory management pain levels by 200%.
> }
>
> if (avctx->priv_data && avctx->codec && avctx->codec->priv_class)
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
