[postgis-users] Ready for 1.0.0 ?

[Fernão Lopes]

Hello!

I have the 0.9.1 installed and working. I'm using Postgresql 8.0.1 in
a debian with 2.6.8 686 kernel.

I've tried to install the 1.0 RC3, but experienced some problems.
I can read old postgis data and publish it with mapserver; but I can't
enable new databases using the lwpostgis. I received the following
error:

DETAIL:  Creating a shell type definition.
psql:lwpostgis.sql:36: ERROR:  could not load library
"/usr/local/pgsql/lib/liblwgeom.so.1.0":
/usr/local/pgsql/lib/liblwgeom.so.1.0: undefined symbol:
_ZN4geos11Polygonizer11getPolygonsEv

The file is there.

This message is received when I installed it in a system that already
had postgis 0.91 installed AND in a new one, with a postgresql 8.0.1
with no data, from the scratch.

Any ideas?

Thanks,

Fernão Lopes


On Fri, 04 Mar 2005 12:00:07 +0100, Markus Schaber <schabi at logix-tt.com> wrote:
> Hi, Mark,
> 
> Mark Cave-Ayland schrieb:
> 
> >>BTW, I successfully compiled postgis using JAVAC="jikes-sun
> >>-target 1.1"  for java 1.1 compatibility. :-)
> >
> > Yes, the PostGIS JDBC driver compiles fine for me with -target 1.1 under the
> > Sun JDK. I think we should make this the default :)
> 
> I think this is a good idea.
> 
> However, I'm afraid not all of the java compilers around do support
> -target, so I'm hesitating to put it in the makefile as default, as this
> could annoy Alex Bodnaru :-)
> 
> >>You can use "make alltests" to run them, they test both
> >>client and some server functionality, but you will have to
> >>set up the database connection info via environment or
> >>makefile. I think this will be the first Windows tests for
> >>postgis jdbc :-)
> > The tests appear to run successfully here for me. I just wanted to check
> > with you as there seem to be multiple entries along the lines of "Finished,
> > 0 tests failed" as opposed to a single grand total at the end. I'll email
> > the JDBC results to you separately as they are quite long, but it's looking
> > promising :)
> 
> Yes. The reason for this is that two of the test applications are only
> slightly adopted versions of Pauls original tests, while the others were
> derived from our in-house work. Another problem was that, originally,
> some of the test cases were "allowed" to fail against older PostGIS
> versions, needing manual re-checking of the output. It was not until
> recently that I added a check for the postgis version so the test
> programs know which tests to run, and which are allowed to fail.
> 
> On my todo list for those tests is to have them output only a small
> summary, and write everything else into a test log file. Another idea
> may be using junit or another test framework, but I don't like the idea
> of adding another dependency.
> 
> Markus
> --
> Markus Schaber | Dipl. Informatiker | Software Development GIS
> 
> Fight against software patents in EU! http://ffii.org/
>                                       http://nosoftwarepatents.org/
> _______________________________________________
> postgis-users mailing list
> postgis-users at postgis.refractions.net
> http://postgis.refractions.net/mailman/listinfo/postgis-users
>