RE: sander's diff library

On Sun, 2003-02-09 at 10:19, Mats Nilsson wrote:
> Compare for instance how Microsoft developed their Excel recalculation
> engine. They had two versions of the engine in operation in the debug
> builds. One optimized assembly version and one slow but understandable
> version. Both would run and if their output differed that would trigger an
> assertion.

Doesn't work out in this case. There are multiple valid diffs between a
given source and target. Consider:

   int a() int a()
   { {
   } +}
  + +
  +int b() +int b()
  +{ +{
  +} }

(I bet that looks really bad in a proportional font.)

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@subversion.tigris.org
For additional commands, e-mail: dev-help@subversion.tigris.org
Received on Sun Feb 9 16:28:54 2003