[geos-devel] non precision related geometric issue

strk at refractions.net strk at refractions.net
Tue Jun 21 04:34:11 EDT 2005


On Mon, Jun 20, 2005 at 08:02:49PM -0400, Carl Anderson wrote:
> 
> Based on recent improvements in the geos library, I have started in 
> depth geometric testing.
> Thanks to all for the improvements.
> 
> attached is a test file for XMLTester for a failure case.
> 
> The B object has an edge intersection with itself, violating assertions 
> 2.1.12.1 and 2.1.12.2 of the SF SQL spec.  GEOS reports it as valid but 
> upon intersection it abends.  JTS/GEOS is not conditioned to test for 
> those assertions but Martin may have had those assertions in mind when 
> writing the overlayOp bts.

That's an error in XMLTester, swap A and B geometries and you'll
read two 'false' from isvalid. 'arg1' is not honourated !
I'll take a look at it.

--strk;

> 
> This is not a precision issue as can be clearly be seen from the test file.
> 
> There is another variation of this issue that I cannot, as of yet, 
> reliably express with a simple test case.
> The variation invloves shared edges between the two test objects.  I 
> will keep digging.
> 
> C.
> 
> 
> 


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




More information about the geos-devel mailing list