[postgis-devel] PSC Vote to officially drop support for PostgreSQL 8.3 in PostGIS 2.0
lr at pcorp.us
Thu Apr 21 17:57:12 PDT 2011
Agree with Paul -- we did say all PSC should at least feel comfortable with
our position and be able to defend it. Though probably something we need to
clarify in our voting rules.
I've cc'd the regular users group since I feel they would be most affected
by this decision and would like to hear their opinions on it.
First let's keep things in perspective. We are talking about not having
support for PostgreSQL 8.3 for PostGIS 2.0. We will still do our duty and
support PostgreSQL 8.3 on PostGIS 1.3-1.5 and if we don't have to worry
about also supporting it on 2.0, we'll have many more cycles to support
issues that arise in 1.3-1.5.
More food for thought -
>From all the signals I have seen, I just feel trying to support PostgreSQL
8.3 on PostGIS 2.0 is a really bad idea.
I will add this. It's not just the testing, its the fact that requiring
our 2.0 code work on PostgreSQL 8.3 is going to slow our release as all
PostGIS developers will need to limit their feature set to work on 8.3 and
avoid new features that will make programming easier and more efficient. We
have much more plpgsql code in PostGIS 2.0, than we have ever had in prior
versions, which makes the task much more difficult.
>From what I can gather most distros package just one version of PostGIS with
each version of PostgreSQL if they package PostGIS at all. I just helped a
client port their database to an ubuntu server on a different host and the
stable on Ubuntu 10 is 8.4 with PostGIS 1.4. In fact even the backports
that have PostgreSQL 9.0, I can't find 1.5 so had to compile ourselves to
get 1.5. This is not something most users new to PostgreSQL or PostGIS will
be willing to do. So the reality is if they want to stay stable they'll be
using 8.2 with 1.4. Similar story with centos. Yum rpms packages just one
version of PostGIS with 8.4 and 9.0. For 9.0 it's 1.5.
If we don't make release before the PostgreSQL 9.1 cut (which I really
fear given all we need to test and finish), we are going to have a lot of
new users starting off their PostGIS experience on 1.5 and it's going to be
next to impossible to get them to upgrade.
If you are at all concerned about new users, you need to take this into
consideration. The ratio of new users is exponential so that I suspect in
a year's time if it is not the case already we will have a lot more new
users with less than 1 years experience with PostGIS / PostgreSQL than we
have users with more than 1 years experience.
From: postgis-devel-bounces at postgis.refractions.net
[mailto:postgis-devel-bounces at postgis.refractions.net] On Behalf Of Paul
Sent: Thursday, April 21, 2011 6:54 PM
To: PostGIS Development Discussion
Subject: Re: [postgis-devel] PSC Vote to officially drop support for
PostgreSQL 8.3 in PostGIS 2.0
-1 means "veto and I'll do what it takes to make my position stick".
So you're OK being 8.3 compatibility tester and fixer?
On Thu, Apr 21, 2011 at 12:27 PM, Mark Cave-Ayland
<mark.cave-ayland at siriusit.co.uk> wrote:
> On 14/04/11 12:02, Paragon Corporation wrote:
>> I hereby call a PSC Vote to officially drop support for PostgreSQL
>> 8.3 in PostGIS 2.0.
>> I think enough has been said on the topic that it is clear to me
>> 8.3 is becoming a bit of a burden to maintain for both regression
>> testing and as well as having to hold back on features introduced in
>> newer versions of PostgreSQL.
>> With this official drop we will then be available to take advantage
>> 1) Window functions, CTES, variadic functions, RETURN QUERY EXECUTE,
>> CASE in pl/pgsql, EXECUTE using, user-defined exceptions,
>> and a slew of other features itemized in PostgreSQL feature matrix
>> 2) Not have to regress test against 8.3 any longer
>> 3) Get rid of the stupid hack we have in place for pgxs
>> 4) Get rid of that conditional logic we have in place for aggregation
>> to handle versions that don't support windowing
>> This is just one step, but my more aggressive requirement which I
>> shall put in as a second PSC Vote, is to not support more than
>> 3 versions of PostgreSQL on any version of PostGIS unless there are
>> extenuating circumstances. More on that later.
>> All PSC voters, please give your vote.
> I'm probably in the minority, but -1 from me.
> Mark Cave-Ayland - Senior Technical Architect PostgreSQL - PostGIS
> Sirius Corporation plc - control through freedom
> t: +44 870 608 0063
> Sirius Labs: http://www.siriusit.co.uk/labs
> postgis-devel mailing list
> postgis-devel at postgis.refractions.net
postgis-devel mailing list
postgis-devel at postgis.refractions.net
The substance of this message, including any attachments, may be
confidential, legally privileged and/or exempt from disclosure pursuant to
Massachusetts law. It is intended solely for the addressee. If you received
this in error, please contact the sender and delete the material from any
More information about the postgis-devel