Re: fixing old_snapshot_threshold's time->xid mapping

On Thu, Sep 17, 2020 at 1:47 AM Michael Paquier <michael(at)paquier(dot)xyz> wrote:
> On Sat, Aug 15, 2020 at 10:09:15AM +1200, Thomas Munro wrote:
> > And ... now that this has a commitfest entry, cfbot told me about a
> > small problem in a makefile. Third time lucky?
>
> Still lucky since then, and the CF bot does not complain. So... The
> meat of the patch is in 0003 which is fixing an actual bug. Robert,
> Thomas, anything specific you are waiting for here? As this is a bug
> fix, perhaps it would be better to just move on with some portions of
> the set?

Yeah, I plan to push forward with 0001 through 0003 soon, but 0001
needs to be revised with a PGDLLIMPORT marking, I think, and 0002
needs documentation. Not sure whether there's going to be adequate
support for back-patching given that it's adding a new contrib module
for observability and not just fixing a bug, so my tentative plan is
to just push into master. If there is a great clamor for back-patching
then I can, but I'm not very excited about pushing the bug fix into
the back-branches without the observability stuff, because then if
somebody claims that it's not working properly, it'll be almost
impossible to understand why.

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company