[postgis-users] Ready for 1.0.0 ?

[Mike Leahy]

Hey folks,

I'm going to suggest that RC3 is not ready...I just tried out the Mark's
installer and ran into trouble.  There are still issues with MapServer
connectivity, which are effectively the opposite of the problem that I had with
RC2.  To get a mapserver to draw a layer from a postgis points table using RC2,
I had to use the following DATA string:

DATA "the_geom from (select oid, place_id, entry, name, collect(the_geom) as
the_geom from gis_culturepoints group by oid, place_id, entry, name) as foo
using srid=32619 using uniqe oid"

It's not ideal, but I guess that the 'collect(the_geom)' solved a problem with
force_collection() not working, am I right?  At any rate, after installing RC3
(with PostgreSQL 8.0.1/WinXP, etc), the same DATA string causes postgres to crash.

I changed the DATA string for the layer back to what it normally would be, and
had no problems with mapserver:

DATA "the_geom from gis_culturepoints using srid=32619 using uniqe oid"

Just to be sure, I tried the collect() function in psql from a command window,
and it crashed as well, so I think this is a more significant bug than just
MapServer connectivity:

testdb=# select collect(the_geom) from gis_culturepoints;
server closed the connection unexpectedly
        This probably means the server terminated abnormally
        before or while processing the request.
The connection to the server was lost. Attempting reset: Failed.


So...my first guess is that perhaps whatever fix solved the force_collection()
problem also broke the collect() function, though maybe something else
altogether has happened.

FYI: When I uninstalled RC2, I verified that that the postgis-related dlls and
sql files (that I know of) were successfully removed.  I backed-up all
non-spatial tables, and dropped my database.  After installing RC3, I recreated
my database, reloaded the spatial table from scripts previously generated by
shp2pgsql, then reloaded the other non-spatial tables.  Is there anything I
might have missed that could be causing an incompatibility with my current
PosgreSQL installation and RC3?

Does anyone know of anything else I should mention, or any tests I should perform?

Kind regards,
Mike

Quoting Mark Cave-Ayland <m.cave-ayland at webbased.co.uk>:

> 
> > -----Original Message-----
> > From: postgis-users-bounces at postgis.refractions.net 
> > [mailto:postgis-users-bounces at postgis.refractions.net] On 
> > Behalf Of Markus Schaber
> > Sent: 03 March 2005 11:14
> > To: PostGIS Users Discussion
> > Subject: Re: [postgis-users] Ready for 1.0.0 ?
> 
> (cut) 
> 
> > Do you think an RC4 would be the correct solution to make the 
> > people test the latest fixes? Strk could then release 1.0.0 
> > unchanged if RC4 is fine.
> 
> Yes, I think especially as strk has fixed one crashing bug since then. I'm
> just not convinced that RC3 has been out long enough for people to test to
> make sure that the Mapserver connector issues have been solved etc.
> 
> > Fine. What is your experience building the jdbc2 stuff on 
> > windows? I'm pretty shure some additional weird problems 
> > might pop up there. You should check out the latest CVS, as 
> > there are some Makefile changes compared to RC3 that may ease 
> > building in your environment.
> 
> The only change I've had to make so far is to change RUNCP and BUILDCP to
> use semi-colons instead of colons - otherwise the building under Windows
> fails. I don't know if it's possible to build some of this functionality
> into the Makefile or to just simply make a note of it in the comments
> somewhere.
> 
> Another thing: I've compiled the JARs as standard using a 1.4 JDK. However
> the notes on jdbc.postgresql.org aren't particularly clear as to which
> versions to use. In particular, it notes that JDBC1 is obsolete with
> PostgreSQL 8.0. What do you think would be the best way to compile the jdbc2
> driver? I'm thinking of compiling with a -target 1.3 option, and including
> the resulting standard/debug JARs in the Win32 installer?
> 
> > I think we should make as much testing as possible.
> 
> Agreed. I must admit I haven't looked at the regression tests in a long time
> to see what coverage they give us, but I'm hoping that passing the
> regression tests should be a good enough indication that PostGIS works on a
> particular platform.
> 
> 
> Kind regards,
> 
> Mark.
> 
> ------------------------
> WebBased Ltd
> South West Technology Centre
> Tamar Science Park
> Plymouth
> PL6 8BT 
> 
> T: +44 (0)1752 791021
> F: +44 (0)1752 791023
> W: http://www.webbased.co.uk
> 
> 
> _______________________________________________
> postgis-users mailing list
> postgis-users at postgis.refractions.net
> http://postgis.refractions.net/mailman/listinfo/postgis-users
> 




----------------------------------------
This mail sent through www.mywaterloo.ca