On 5 June 2012 23:55, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> Simon Riggs <simon@2ndQuadrant.com> writes:
>> Can't we have a trial branch where quarantined patches can be placed
>> on trial for inclusion in main release?
>
> [ shrug... ] You're welcome to publish a personal repo somewhere with
> such things. But even if we did that in the master repo, it would have
> approximately nothing to do with released versions. You might as well
> just figure on submitting the patch into 9.3.
As you know, many of us have personal repos on github and
git.postgresql.org already. Technically, the official repo is nothing
more than just another clone too, owing to git's distributed
architecture, and yet we don't give out commit bits to just anyone.
Simon is proposing precisely that such a branch *would* have something
to do with released versions, as the linux-next branch does, for
example, and that our processes would change to merge upstream patches
from this branch, to better match the processes of other large open
source projects, in particular, those of the Linux kernel.
One way to make the process of patch acceptance scale without
compromising on quality is to pipeline it, which is precisely what the
Linux guys did, and, I believe, is a major reason why Linus Torvalds
took the time to develop his own source control system.
You may disagree with the view that we should do this, and that's
obviously a perfectly valid view. However, process matters, and
ceremony matters, and the suggestion that an officially blessed repo
is essentially equivalent to something somebody privately produces is
simply not accurate. If it was, why would you even care?
--
Peter Geoghegan http://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Training and Services