boost instability

From: nbecker_at_[hidden]
Date: 2001-12-04 09:00:59

As I mentioned, the recent change from 1.25.1 to 1.26.0 causes
different results for random number generators.  So far, I found the
results for uniform_real and uniform_int have changed.

Nothing in the release notes suggested this.  If possible, please let
us know if a change to the library is likely to cause the results to
differ.  I'm not complaining.  It would just make my life easier.

I'm using boost for (among other things) generating test vectors for a
VLSI design.  All the vectors just changed unexpectedly.