On Fri, 20 Feb 1998 09:55:13 -0500 
 "Perry E. Metzger" <perry@piermont.com> wrote:

 > My problem is that I actually run my snapshots for a while before
 > releasing them to make sure they don't totally destroy people's
 > machines, and right now I'm running UVM. Backing out would be painful
 > because the userlands differ. (Indeed, you need a different userland
 > for UVM *and* PMAP_NEW, too!)
 > 
 > Given this, I'm only really in a position to produce UVM snapshots. I
 > was going to do a new one in about two weeks (once the rubble finished
 > bouncing).

What are you talking about?  Build with a different mk.conf and
BSDOBJDIR, then!  The fact that _you're_ running UVM should have
no effect on your ability to make regular snapshots.

Binary snapshots should not be built w/ UVM, period.  They're meant to
reflect the state of the source tree with standard defaults when the
snapshot is generated.

 > 
 > > Yeah, sure, we want to get UVM tested, but like it or not people have
 > > to run -current on "stable-ish" machines, and they shouldn't _have_ to
 > > switch to UVM until it's known to be pretty stable.
 > 
 > That is one reason I intend to wait a little while longer.
 > 
 > Perry

Jason R. Thorpe                                       thorpej@nas.nasa.gov
NASA Ames Research Center                            Home: +1 408 866 1912
NAS: M/S 258-5                                       Work: +1 650 604 0935
Moffett Field, CA 94035                             Pager: +1 415 428 6939