Hi all.

I wasn't around for the discussions alluded to below...

What would be the problem with the following?

Begin with RELEASE 1.5.0
Modifications result in "beta" versions such as 1.5.0A, 1.5.0B,
etc.  Eventually, some to all of the changes in the beta versions are
rolled into a new RELEASE which in this case would be 1.5.1.
Features that are not included in 1.5.1 but are still being tested
would be rolled into 1.5.1A.

People needing stable, production quality code just need to know
to stay away from versions with letters, as they are considered
"beta" releases.  People that want the latest-and-greatest features
just need to know to look for the highest release number with the
"highest" character.

If there is a need to specifically differentiate a "release candidate"
for a version, maybe use something like 1.5.1-RC.0, and the -RC.0 part
can be incremented for  multiple release candidates.

Lets Go Canes!

--- David Burgess <burgess@mitre.org> wrote:
> There are people that can explain this much more succinctly than I,
> but
> here's a shot....
> 
> 1.5W and 1.5.1 are both upgrades from 1.5.  While they may share some
> common upgrades/fixes, they are definitely not the same.  
> 
> The difference is the quality of the release.  1.5.1 is available for
> 
> people like me (that run businesses using NetBSD) and need
> 'commercial'
> quality while 1.5W is the 'research' quality release.
> 
> We spent quite a bit of time arguing about the release numbering
> scheme 
> a couple of years ago and there isn't really any good alternative.  
> Unless we want to reopen that can of worms, this kind of problem will
> 
> continue to occur.
> 
> The problem is that the 1.5[A-ZZ] releases are actually precursors to
> 
> 1.6, where the 1.5.[1-9] releases are extensions to the past.  How we
> 
> number that in a clear way just isn't an easy nut to crack.  
> 
> Unless we change the scheme so that the numbered releases identify
> the
> goal (which would make -current 1.6W), I don't see any way around
> this
> problem.  I'm not advocating a change, there are good reasons to go 
> with the current method.  Even then, we still need to make it clear
> that
> 1.5.1 is not an upgrade compatible with 1.6W.  Of course, we then get
> 
> into the problem of what happens when we want to release the 1.6
> Beta?
> 
> Sigh :-(
> 
> Dave
> 
> Ken Wellsch wrote:
> > 
> > Steven Sartorius wrote:
> > >
> > > I'm currently running 1.5W and decided to upgrade to 1.5.1 by
> installing
> > > from source.   This morning I grabbed the tarballs from the 1.5.1
> release
> > > directory and built a new (GENERIC) kernel with no problems. 
> When I go to
> > > boot the new kernel, though, things start off normally (the
> kernel starts to
> > > load) but before I see any boot messages the machine
> spontaneously reboots.
> > > Never seen this before in NetBSD.  Anyone have any ideas?
> > 
> > 1.5W is a release of -current.  While 1.5.1 is a upgrade from 1.5.
> > So you are running an "old" kernel via 1.5.1 with a new userland
> (-current).
> > 
> > Trust me, 1.5.1 is *not* an upgrade to -current (e.g. 1.5W).
> > 
> > Apologies this is confusing... the lettered 1.5 versions are
> branches
> > of -current as I understand the naming scheme.
> > 
> > -- Ken


=====
--------------
Lets Go Canes!

__________________________________________________
Do You Yahoo!?
Get personalized email addresses from Yahoo! Mail
http://personal.mail.yahoo.com/