[PEAK] Major update to generic functions in CVS

[Phillip J. Eby]

At 09:57 AM 4/25/05 -0400, Jay Parlar wrote:

>I think that I was seeing this error before the fixes you put in over
>the weekend.

I rather doubt that, since the problem is a direct result of the 
refactoring I did.  Specifically, the tree building algorithm expected a 
given index to always produce casemaps with the same number of keys, 
regardless of which cases it was producing the map for.  This is no longer 
true with the optimized InequalityIndex, so I had to refactor to accomodate 
that.

A fix is now in CVS; thanks for the report.  Sorry I couldn't get to this 
sooner; I've pretty much been sick since Sunday and I still don't feel too 
well now, but this is the first I've had enough brain cells since then to 
be able to even think about programming.  :(