some weirdness in diff output?

[Christopher Li]

On Wed, May 25, 2005 at 07:12:45AM -0600, Jake Edge wrote:
> mercurial/mdiff.py calls difflib.unified_diff so I thought that its 
> output should conform to some standard (if one exists) for unified diff 
> output ... perhaps this is just a gray area and different 
> implementations do different things ... in the end, we probably mostly 
> care what patch will handle for input ...

Actually, you are right. I have the wrong impression that we convert the
binary diff to unidiff.  The difflib.unidiff should handle the no newline
at ending.

Chris