Johnny,

I think that your comment is wrong in one detail: 2.1 will be a
patch-release to 2.0.  It will be to 2.0 what 1.6.1 was to 1.6.
So there should be no new functionality in 2.1 relative to 2.0
(hence relative to 2.99.<x> for pre-3.0 -current).

So a 2.1 release having functionality not in a 2.99.<something>
-current is about as likely as 1.6.1 having functionality that
is absent in some post-1.6 -current versions.

That is my understanding.


I never saw the old scheme as broken, so I was not anxious to see
it fixed, but at least the problem that you are describing does
not seem to be introduced by the new scheme.  (I also think that
linearizing a branching tree is not the best of ideas, even if
in practice the NetBSD tree is close to a line and all anomalies
can be avoided.)


-- 
  "I probably don't know what I'm talking about."  http://www.olib.org/~rkr/