Devil Take the Hindmost?

classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

Devil Take the Hindmost?

Paul Ramsey
Would it be too terrible to drop support for 7.4?

http://code.google.com/p/postgis/issues/detail?id=6

So many things got better with 8.X...

P.

Reply | Threaded
Open this post in threaded view
|

Re: Devil Take the Hindmost?

Mark Leslie-2-2
Paul Ramsey wrote:

> Would it be too terrible to drop support for 7.4?
>
> http://code.google.com/p/postgis/issues/detail?id=6
>
> So many things got better with 8.X...
>
> P.
> _______________________________________________
> postgis-devel mailing list
> [hidden email]
> http://postgis.refractions.net/mailman/listinfo/postgis-devel
>

Does anyone know offhand how much longer PostgreSQL is committed to
maintaining 7.4?

It would seem to me that if they are maintaining 7.4 for a short period
of time, call it < x, then we're better off supporting it until then,
just for the look of it.  For > x, we should cut support as soon as the
support becomes significant (today perhaps?).  From my perspective, x =
1 year feels right, but it depends very much on how much development is
expected on PostGIS in that year.
I haven't done the benchmarking, but I would expect running against 7.4
would be less performant, and thus provide a lesser impression of
PostGIS.  At the same time, telling everyone to upgrade their databases
won't be received with a ticker-tape parade either.

My suggestion would be pick a date to end support.  The same time
PostgreSQL does would be ideal, but if it's too far, give it a few
months and cut a minor release at that time which explicitly prohibits
compiling against 7.4.

--
Mark Leslie
Geospatial Software Architect
LISAsoft Pty Ltd
+61 (0)2 8570 5050

Reply | Threaded
Open this post in threaded view
|

Re: Devil Take the Hindmost?

Ron Mayer
In reply to this post by Paul Ramsey
Paul Ramsey wrote:
> Would it be too terrible to drop support for 7.4?
>
> http://code.google.com/p/postgis/issues/detail?id=6
>
> So many things got better with 8.X...

Which was the postgresql release that fixed vacuuming
GiST indexes taking a lock blocking most other queries?

Our PostGIS based app was so painful on those releases[1]
I'd vot for dropping support for those since it was
a rather crippling and hard to work around issue; and
left many systems (those using autovacuum or any similar
vacuum scripts) with unpredictably slow behavior (especially
when using the vacuum_cost_delay).


[1] http://archives.postgresql.org/pgsql-general/2005-02/msg01510.php


Reply | Threaded
Open this post in threaded view
|

Re: Devil Take the Hindmost?

Tom Lane
In reply to this post by Mark Leslie-2-2
Mark Leslie <[hidden email]> writes:
> Does anyone know offhand how much longer PostgreSQL is committed to
> maintaining 7.4?

There's no specific plan.

We just obsoleted 7.3, and 7.4 is about a year newer, so consistency
might suggest obsoleting 7.4 in about a year.  Whether it will actually
fall out that way is hard to say.

What's reasonable for PostGIS to support is not necessarily the same.
You guys are heavily dependent on some features like GIST indexes that
are not seen as critical by the core project.  Ron M notes nearby that
the 8.1 fixes to enable concurrent use of GIST indexes made a huge
difference to him --- so if you were to drop support for everything
older than 8.1, I wouldn't argue with you.  You should certainly be
trying to light a fire under pre-8.1 PostGIS users to upgrade.

                        regards, tom lane