On 08/05/09 12:46, Paragon Corporation wrote: > Mark, > > To add I think Simon's dataset was too big to go thru 1.3 at a decent time > (as I recall think its about 3000 polygons he's dealing with). I think I > have run into some of these issues myself in my torture tests but ignored > them since they passed regress (I mean all versions failed) and I didn't > absolutely verify they weren't invalid, but I'll try to dig some of these up > to verify they were invalid. > > I suspect its a 1.3 issue as well, but I expect we'll see more of these > kinds of things popping up as people will be more able to handle larger > geometries and thus will be stress testing PostGIS more.
I run into this kind of problem every now and then. I've created a ticket (http://trac.osgeo.org/geos/ticket/275) with a test case. The problem still exists with geos-3.1.1. That specific test case works when unioned in PostGIS rather than geos though. When I encounter this kind of problem I usually simplify and/or snap to grid my geometries and with a bit of luck it works again :). -- Maxime _______________________________________________ postgis-users mailing list postgis-users@postgis.refractions.net http://postgis.refractions.net/mailman/listinfo/postgis-users
