[FFmpeg-devel] Ogg regression broken on PowerPC

[Måns Rullgård]

Mike Melanson <mike at multimedia.cx> writes:

> Hi,
>
> Compiling SVN 12599 wsing gcc 4.1.2 on a PowerPC. Traditionally, the WMA
> encoder breaks when running 'make test'.

I fixed that some time ago.

> However, this is the current problem:
>
> --- /home/melanson/ffmpeg/ffmpeg-main/tests/libav.regression.ref
> 2008-03-20 23:04:04.000000000 -0700
> +++ tests/data/lavf.regression  2008-03-26 13:50:39.000000000 -0700
> @@ -84,8 +84,8 @@
>  8d117c49d6b210abe783d1b0b897cec7 *./tests/data/b-libav.voc
>    32768 ./tests/data/b-libav.voc
>  ./tests/data/b-libav.voc CRC=0x49972c8c
> -d7e6e9091a600b7208b29500be94aa89 *./tests/data/b-libav.ogg
> -23624 ./tests/data/b-libav.ogg
> +6a0489ee6cf21ada85c45ac7eb8c0c26 *./tests/data/b-libav.ogg
> +23618 ./tests/data/b-libav.ogg
>  ./tests/data/b-libav.ogg CRC=0x93baa056
>  ce356ce2708cb6033ab5d762da93cfd4 *./tests/data/b-libav-yuv420p.yuv
>   304128 ./tests/data/b-libav-yuv420p.yuv

The problem is in the flac encoder.  There's another floating-point
oddity there.  The compression is still lossless, just different.

-- 
M?ns Rullg?rd
mans at mansr.com