[FATE] fax-g4 on bsd

[Vitor Sessak]

On 08/28/2010 09:31 PM, Michael Kostylev wrote:
> On Sat Aug 28 20:17:18 2010
> Michael Kostylev wrote:
>
>    
>> [...]
>> Program received signal SIGSEGV, Segmentation fault.
>> [Switching to Thread 0x8013020b0 (LWP 100071)]
>> decode_group3_2d_line (avctx=0x80130b800, gb=0x7fffffffcbe0, width=2464, runs=0x80134804c, runend=0x80134a684, ref=0x801500000)
>>      at /usr/home/mik/src/fate/src/libavcodec/faxcompr.c:232
>> 232                 run_off += *ref++;
>>      
> Mostly clear:
>
> ==22584== Conditional jump or move depends on uninitialised value(s)
> ==22584==    at 0x7C6B48: decode_group3_2d_line (faxcompr.c:231)
> ==22584==    by 0x7C7465: ff_ccitt_unpack (faxcompr.c:291)
> ==22584==    by 0x68904A: decode_frame (tiff.c:160)
> ==22584==    by 0x697715: avcodec_decode_video2 (utils.c:610)
> ==22584==    by 0x47E87D: av_find_stream_info (utils.c:2049)
> ==22584==    by 0x4079F8: opt_input_file (ffmpeg.c:3205)
> ==22584==    by 0x40ED9A: parse_options (cmdutils.c:179)
> ==22584==    by 0x409B39: main (ffmpeg.c:4320)
> ==22584==  Uninitialised value was created by a heap allocation
>    

Thanks. I think this is a decoder bug and I've removed the test (at 
least until it is fixed).

-Vitor