Re: Removing pg_migrator limitations

On Sun, Dec 27, 2009 at 5:15 PM, Bruce Momjian <bruce(at)momjian(dot)us> wrote:
>> Is that something we're comfortable adding in the final commitfest?
>
> Uh, no idea.  It would be nice, of course.

Do we know if there's active development in progress on page CRCs? If
so, when can we expect to see a working patch?

With respect to adding it in the final CommitFest, I am concerned that
this might be a big enough feature that it would be destabilizing, and
I'm pretty disinclined to do anything that will destabilize the tree
at this point in the release cycle. Part of my concern is that we
just committed a really big feature (Hot Standby) that has already had
a few bug reports and also has a known issue with VACUUM FULL that
needs to be addressed. It seems likely there is going to be a good
deal more work that has to be done there. I'm worried that adding
more big features in the relatively small amount of time that remains
before we are ostensibly going to beta is going to result in a very
buggy beta and/or a buggy release (I am also concerned about Streaming
Replication in this regard).

On the other hand, since we have yet to see an actual patch to
implement page CRCs, it may be premature to draw conclusions about how
destabilizing and/or invasive it will be. If it's not that bad, maybe
it's OK. Another option, if it turns out that we have several major
patches that we don't feel comfortable committing for 8.5, is to
branch the tree prior to the release - for example, at the start of
beta. Then we could commit those features for 8.6 and just apply any
remaining changes for 8.5 to both branches. This is a little more
work for the committers, but it has the advantage of getting the big
patches into our tree early, versus leaving them elsewhere where they
may bitrot or fall through the cracks, so I think it might be worth
it.

...Robert