That's correct Jody, the resulting geometry looks pretty much like the ascii art in your email. The top point of the "bow tie" that crosses the edge is the point that causes the ring consistency check to fail.
Graham. Jody Garnett wrote: > Perhaps it is time to call for another IRC meeting of the "Geometry > Fan Club"? > > If I understand you have found one talented bug - our construction of > Rings is not quite up to snuff. > +------------+--------+ > | / \ | > | + + | > | \ / | > | + | > | / \ | > | / \ | > | +-----+ | > +---------------------+ > Kind of like the classic "bow tie" test case, but this time with holes > touching the outer edge. > > Nasty. > > Anyone have any ideas? Or are we off to look at the JTS codebase .... > > Cheers, > Jody > > Graham Davis wrote: > >> I've been working on getting the JTS XML tests to work with the >> geometry module so we can have better confidence that it is working >> correctly. There have been many errors that I'm working through, but >> I've come across one that seems to be a fairly big problem with how >> the module is constructing polygons with the polygonbuilder. >> >> Right now, if you run the test GeometryConformanceTest.java in >> org.geotools.geometry.xml it will run all the xml file tests in >> src/test/resources. The property files in there have some special >> cases to get some tests to run, or to skip problematic ones for now >> while I work through the problems. If you look at >> TestFunctionAA.properties and stop it from skipping the first 6 tests >> (comment them out) and run the test, you can see the errors I am >> talking about. >> >> All the errors are similar, but if you look at the first one in >> particular, you will see it is trying to UNION 2 polygons together. >> The resulting geometry is being built up as a MaximalRingEdge, which >> implements RingImpl. The resulting geometry is a polygon with 2 >> holes in it, which should be constructed as 3 rings I believe, but >> instead it is being built as 1 large ring which isn't a simple >> geometry. Since RingImpl has built in consistency checks, it crashes >> saying there is a point that intersects but isn't an end or start >> point (basically it's saying this ring is not simple and is incorrect). >> >> The other 5 tests that I am skipping are similar problems. They are >> doing UNIONs or SYMDIFFERENCEs on polygons, and the resulting >> geometry is being built up as one big ring when it should be multiple >> rings. I've been looking at this problem and trying to figure out >> the best way to resolve this, but I'm not sure if the implementation >> of this area needs a whole rework or perhaps there is a smaller tweak >> that can be made to fix this. >> >> Perhaps the original coders of this module have some insight and >> might be able to help? Now that there are some strong tests to run >> against this module, it is good that these sorts of problems are >> popping up. I just hope that the original coders, or someone, can >> help with some of these larger design/implementation problems. Is >> anyone brave enough to offer some insight?? :) Thanks! >> >> (I can create some images to better display the problem in the first >> exampe with the 2 polygons and how they union if that will help >> anyone, just let me know). >> >> > > -- Graham Davis Refractions Research Inc. [EMAIL PROTECTED] ------------------------------------------------------------------------- This SF.net email is sponsored by DB2 Express Download DB2 Express C - the FREE version of DB2 express and take control of your XML. No limits. Just data. Click to get it now. http://sourceforge.net/powerbar/db2/ _______________________________________________ Geotools-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/geotools-devel
