On Thu, Feb 4, 2010 at 1:10 PM, Greg Smith <greg@2ndquadrant.com> wrote:
> Yes, but that doesn't make the user very happy--just better informed about
> why they're dead in the water. I'd hate to see the next Long-Term support
> version of Debian and Ubuntu, both likely to hang around a while, have
> versions of pgAdmin that are fundamentally incompatible with the next PG
> version, due in only a few months. Seems bad from an advocacy perspective.
> And it looks like that's exactly how it's going to play out unless
> something gets done very soon here, which I just realized myself today.
pgAdmin 1.12 will be released before, or with PostgreSQL 9.0, so there
won't be a version of Ubuntu shipping with an incompatible PostgreSQL
and pgAdmin combination. That's how we've done it for the last 10
years or so.
> Obviously backpatching past 1.10 is silly. There seem to be three
> potentially useful ideas that would improve the state of this part of the
> world:
>
> 1) The documentation update I already suggested, bringing the Ubuntu page
> back to current and then reasonable to link on the main download section
> again.
> 2) Backpatch basic 9.0 support into a V1.10.2 release. That could work like
> this:
>
> a) Release 1.10.2 with this fix ASAP--if that can't happen within a week or
> less, the rest of what I'm suggesting is probably dead.
There is zero chance I'll have time to produce a release in a week. I
leave for Brussels tomorrow as do most of the rest of the active
developers, and have a full calendar next week.
Besides, there are far more changes required than just fixing one
query - more so than we'd ever consider back porting, even if you did
ignore the 9.0 feature enhancements. Off-the-top-of-the-head examples
include: role/database GUC storage has completely changed and
pg_trigger.tgiscontrainst is replaced by pg_trigger.tgconstraint
(which has a slightly different meaning).
> Doing just (3) is probably sufficient to make my problems go away, and now
> that I know nobody else has that on their radar I'll start looking into it.
I'm perfectly happy with that option.
> Now that I realize the scope of the problem here, it would be nice to
> consider a broader plan too. Anybody know Gerfried well enough to ping for
> his opinion here?
Don't know him at all.
--
Dave Page
EnterpriseDB UK: http://www.enterprisedb.com