[postgis-devel] Possible Bug in isvalid()

[strk at refractions.net]

On Thu, Apr 07, 2005 at 11:42:28AM +0200, Markus Schaber wrote:
> Hi,
> 
> I just tried to apply my isvalid constraint to our raw data tables, and
> got the following error message on one of them:
> 
> NOTICE:  AssertionFailedException: Unable to find a hole point not a
> vertex of the shell
> ERROR:  GEOS isvalid() threw an error!
> 
> This sounds rather strange to me. IMHO, it should not be possible to
> make isvalid() throw an error by feeding it an invalid geometry.

I agree, sounds as a bug in JTS, would be the second of this kind.
(we have had a similar problem in the past, where an 'inner' class
 used by validOp threw exceptions not handled by it)

If you provide the exact geometry I'll check against GEOS-2.1.1 and
a recent JTS.

--strk;

> 
> I'll try to track down which geometries cause this error message...
> 
> Markus



> _______________________________________________
> postgis-devel mailing list
> postgis-devel at postgis.refractions.net
> http://postgis.refractions.net/mailman/listinfo/postgis-devel