[FFmpeg-devel] [PATCH] VP8 de/encode via libvpx

[Diego Biurrun]

On Wed, May 19, 2010 at 07:57:13PM -0400, James Zern wrote:
> 
> --- libavcodec/libvpxdec.c	(revision 0)
> +++ libavcodec/libvpxdec.c	(revision 0)
> @@ -0,0 +1,127 @@
> +
> +typedef struct VP8DecoderContext {
> +    vpx_codec_ctx_t decoder;

Apparently libvpx invades the _t POSIX namespace left and right.
Why don't you go ahead and change this quickly before anything
is set in stone?  The library is new, you can now easily make
incompatible changes...

> +    vpx_codec_dec_cfg_t deccfg = { /*token partitions+1 would be a decent choice */

space after *

> +    /*end of stream, vp8 does no buffering */

ditto

> +    if (avpkt->size == 0)

(!avpkt->size)


> +        const char *error = vpx_codec_error(&ctx->decoder);
> +        const char *detail = vpx_codec_error_detail(&ctx->decoder);

align the =

> +AVCodec libvpx_vp8_decoder = {
> +    "libvpx_vp8",

Sometimes you call this libvpx, sometimes libvpx_vp8.  Stick to one
version.  You said..

> >> --- configure (revision 23157)
> >> +++ configure (working copy)
> >> @@ -183,6 +183,7 @@ External library support:
> >> + ?--enable-libvpx-vp8 ? ? ?enable VP8 support via libvpx [no]
> >
> > Does the name of the lib really contain "vp8"? ?If not, then drop the
> > silly "-vp8" postfix.
> >
> Removed.

.. so I guess the -vp8 suffix is completely made up.
Drop it everywhere then.

> --- libavcodec/allcodecs.c	(revision 23201)
> +++ libavcodec/allcodecs.c	(working copy)
> @@ -349,6 +349,7 @@
>      REGISTER_ENCODER (LIBVORBIS, libvorbis);
> +    REGISTER_DECODER (LIBVPX_VP8, libvpx_vp8);

ditto

> --- libavcodec/Makefile	(revision 23201)
> +++ libavcodec/Makefile	(working copy)
> @@ -536,6 +536,7 @@
>  OBJS-$(CONFIG_LIBVORBIS_ENCODER)          += libvorbis.o
> +OBJS-$(CONFIG_LIBVPX_VP8_DECODER)         += libvpxdec.o

same

Diego