[FFmpeg-devel] [RFC] wma encoder regtest

[Mike Melanson]

Baptiste Coudurier wrote:
> On 2/17/2009 8:04 AM, Michael Niedermayer wrote:
>> [...]
>>
>> telling that the md5 differs is not usefull alone...
> 
> Maybe we could think of a mechanism that helps us spotting what _change_
> appeared and _where_ in a more efficient way ?
> 
> Md5 seems not to be efficient in facilitating debugging.

FATE already does this, stipulating that the test in question ever 
worked in the first place. Which is not the case for these particular 
regressions. :)

-- 
     -Mike Melanson