[FFmpeg-devel] [PATCH 1/8] png: Clear up the calculation of max packet size

[wm4]

On Sun, 29 Mar 2015 11:05:40 +0000
Donny Yang <work at kota.moe> wrote:

> Signed-off-by: Donny Yang <work at kota.moe>
> ---
>  libavcodec/pngenc.c | 19 +++++++++++++++----
>  1 file changed, 15 insertions(+), 4 deletions(-)
> 
> diff --git a/libavcodec/pngenc.c b/libavcodec/pngenc.c
> index d6233d0..3697dbb 100644
> --- a/libavcodec/pngenc.c
> +++ b/libavcodec/pngenc.c
> @@ -360,12 +360,23 @@ static int encode_frame(AVCodecContext *avctx, AVPacket *pkt,
>          return -1;
>  
>      enc_row_size    = deflateBound(&s->zstream, row_size);
> -    max_packet_size = avctx->height * (int64_t)(enc_row_size +
> -                                       ((enc_row_size + IOBUF_SIZE - 1) / IOBUF_SIZE) * 12)
> -                      + FF_MIN_BUFFER_SIZE;
> +    max_packet_size =
> +        8 +             // PNGSIG
> +        13 + 12 +       // IHDR
> +        9 + 12 +        // pHYs
> +        1 + 12 +        // sRGB
> +        32 + 12 +       // cHRM
> +        4 + 12 +        // gAMA
> +        256 * 3 + 12 +  // PLTE
> +        256 + 12 +      // tRNS
> +        avctx->height * (
> +            enc_row_size +
> +            12 * (((int64_t)enc_row_size + IOBUF_SIZE - 1) / IOBUF_SIZE) // 12 * ceil(enc_row_size / IOBUF_SIZE)
> +        );

This looks extremely painful and unmaintainable. (I know, the previous
code is not much better.)

>      if (max_packet_size > INT_MAX)
>          return AVERROR(ENOMEM);
> -    if ((ret = ff_alloc_packet2(avctx, pkt, max_packet_size)) < 0)
> +    ret = ff_alloc_packet2(avctx, pkt, max_packet_size < FF_MIN_BUFFER_SIZE ? FF_MIN_BUFFER_SIZE : max_packet_size);
> +    if (ret)
>          return ret;
>  
>      s->bytestream_start =