[geos-devel] CVS GEOS union results

[strk at refractions.net]

Could you try to compute the steps again but with a GEOS version
*not* inlined ?

I'm working right now on yet another case where inlined/outlined
GEOS versions give different results (the outlined one is correct
for my case).

We really need a way to stop this FP madness...

--strk;

On Tue, Jun 13, 2006 at 09:33:59AM -0500, Howard Butler wrote:
> The geometry is my own, and it is not from the JTS test suite.  Using 
> Python Cartographic Library with a GEOS 2.2.2 produces a result that 
> is more in line with the expected one (a single polygon rather than a 
> concatenated multipolygon).
> 
> 
> 
> At 04:52 PM 6/12/2006, strk at refractions.net wrote:
> >>From which file from JTS did you extract the test ?
> >I haven't seen the geometries visually, did you ?
> >The default precision model is FLOATING (double), right,
> >but for XML tests the precision model can be specified
> >with the <precisionModel> tag.
> >
> >--strk;