[FFmpeg-devel] [PATCH] Merge (vp3|theora)_calculate_pixel_addresses

[Michael Niedermayer]

On Tue, Sep 30, 2008 at 11:47:06PM -0400, David Conrad wrote:
> On Sep 30, 2008, at 9:13 PM, David Conrad wrote:
>
>> On Sep 30, 2008, at 3:32 PM, matthieu castet wrote:
>>
>>> David Conrad wrote:
>>>> Hi,
>>>>
>>>> As noted in the FIXME.
>>>>
>>>> On a side note, does anyone have any Theora files from before the
>>>> bitstream finalization?
>>> You've got : http://samples.mplayerhq.hu/ogg/Theora/theora.ogg
>>
>> Looks like I broke this about a year and a half ago in r8506, then it was 
>> further broken in r8928.
>> Attached fixes the second break.
>
> I need to pay more attention to warnings. Fixed version of patch 1 
> attached. Also attached is a fix for the first issue.

[...]
> diff --git a/libavcodec/xiph.c b/libavcodec/xiph.c
> index 6554264..11caed4 100644
> --- a/libavcodec/xiph.c
> +++ b/libavcodec/xiph.c
> @@ -18,6 +18,7 @@
>   * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
>   */
>  
> +#include "avcodec.h"
>  #include "xiph.h"
>  
>  int ff_split_xiph_headers(uint8_t *extradata, int extradata_size,
> @@ -25,6 +26,10 @@ int ff_split_xiph_headers(uint8_t *extradata, int extradata_size,
>                            int header_len[3])
>  {
>      int i;
> +    for (i = 0; i < 3; i++) {
> +        header_start[i] = extradata;
> +        header_len[i] = 0;
> +    }
>  
>      if (extradata_size >= 6 && AV_RB16(extradata) == first_header_size) {
>          int overall_len = 6;
> @@ -33,8 +38,10 @@ int ff_split_xiph_headers(uint8_t *extradata, int extradata_size,
>              extradata += 2;
>              header_start[i] = extradata;
>              extradata += header_len[i];
> -            if (overall_len > extradata_size - header_len[i])
> -                return -1;
> +            if (overall_len > extradata_size - header_len[i]) {
> +                av_log(NULL, AV_LOG_WARNING, "Only found %d Xiph headers, possibly old Theora\n", i+1);
> +                return 0;
> +            }
>              overall_len += header_len[i];
>          }

I dont see how this can work correctly, 
if header_len isset to a too large value then instead of failing it returns
success, this isnt correct.

[...]
-- 
Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

It is not what we do, but why we do it that matters.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
URL: <http://lists.mplayerhq.hu/pipermail/ffmpeg-devel/attachments/20081005/6a932940/attachment.pgp>