Alvaro Herrera wrote:
> Bruce Momjian wrote:
> > Alvaro Herrera wrote:
> > > Bruce Momjian wrote:
> > >
> > > > The hold queue has patches that still need discussion, or ideas for
> > > > patches, so it is more than just patches ready for application, and
> > > > moving the whole thing at once would overwhelm patch reviewers.
> > >
> > > So why aren't all patches that are posted to the -patches list in the
> > > hold queue?
> >
> > Because I haven't looked them over yet, and wasn't putting things in the
> > queue while we were waiting on 8.2.1.
>
> No, I mean in principle, not in this particular case. If we have two
> queues, and there's a barrier to moving patches from the "hold" queue to
> the other queue, why aren't patches posted in pgsql-patches put right
> away in the "hold" queue?
They could be, but remember, my queues are only for patches that no one
else has delt with, so auto-add doesn't make lots of sense, plus many
patches aren't sent to patches, or are discussions in the patches list,
or are ideas that have to be made into patches.
> After all, there's already a barrier to applying a patch in the non-hold
> queue, which is that someone reviews and approves it. Does it make
> sense to have three barriers to the patch managing process? ISTM two is
> good enough (first when moving a patch from the hold queue to the main
> queue, and then when applying a patch from the main queue).
>
> I hope I'm making sense here :-)
Yea. We could just throw things in the hold queue if we were sure we
would get only good patches/ideas from all lists. Right now the hold
queue is only used during this transition period between releases.
I am afraid that to capture everything, you would basically just
duplicate the archives.
-- Bruce Momjian bruce@momjian.us EnterpriseDB http://www.enterprisedb.com
+ If your life is a hard drive, Christ can be your backup. +