On Wed, 11 Oct 2006, Daniel de Kok wrote:
>> 2. Offering sdiff(1) has regression tests and passed them all. While
>> our GNU sdiff(1) does not pass more than half of them (from 35 tests).
>
> I suppose half of the sdiff regression tests? Since you propose to
> replace GNU diff with sdiff, it would be far more interesting to see
> how sdiff performs with GNU diff regressions tests.

Yup - nice that the 'new' sdiff has a regression test, but if 50% of the 
tests fail with the current (GNU) sdiff, I wonder what the point of the 
regression test is - I always thought it was to make sure things are 
compatible...


  - Hubert