Re: commit fests (was Re: primary key error message)

Robert Haas <robertmhaas(at)gmail(dot)com> writes:
> I'm not sure whether you're stating a position that's been agreed to
> by -core or some other group, or just expressing your own opinion, but
> I think feature freeze should be the beginning of the last CommitFest,
> not the end.

I think traditionally we understood "feature freeze" to be the point at
which we stopped *committing* new features, not the point at which it
was too late to *submit* them. So by that definition feature freeze
starts at the end of the last CF.

I agree with Peter that things are a bit different in the CF process.
Rather than a binary frozen-or-not state, we now have a gradual
congealing (if you will), where the size of an acceptable new feature
gets smaller as we get towards the end of the development cycle.

regards, tom lane