keyword expansion and post_commit hook in 2.0 (was Re: DRAFT 2.0.0 ANNOUNCEMENT)

[Stephen J. Turnbull]

Ian Clatworthy writes:

 > I still don't understand exactly what point you're arguing. We added a
 > "feature" after 2.0rc1? If so, 2.0rc1 was incorrectly named - it was
 > meant to be called 2.0beta1. After it was released and announced, it was
 > too late to change it.

 > FWIW, as RM, poolie refused to take the risk that patch carried the day
 > or two before 2.0.0 went gold. That was much to my disappointment but a
 > perfectly reasonable decision. It will *hopefully* be approved and land
 > in 2.1b1 and 2.0.1.

The point I'm arguing is that under this policy, I see no way for
anybody who is not closely following bzr development to distinguish
between 2.1b1 and 2.0.1.  Especially since you imply that you think it
would be reasonable had the patch been accepted a day or two before
2.0.0 went gold.  If that's anything close to an accurate assessment
of policy, you're risking a situation where people are saying "don't
bother with 2.0.X, it's as beta as 2.1bY."  That's a disservice to the
users and a waste of development effort.

IMO, your betas just are not that risky that it's worth splitting the
difference between "risky as a bzr beta" and "read my lips: No New
Features" in 2.0.X releases.  The latter policy simplifies the RM's
job, too.