I think the problem is in the Java2DConverter.areaToShapes() method which rounds the float values to int. This is not a problem for the boundaries that are created with OSM points, but with areas created with Area.intersect() or Area.substract() we see some points with fractions. In Way we are using these rounded values to calculate the area size and determine whether the way is clockwise or counter-clockwise. Sometimes the rounding error is too big (bigger than the area size), so the clockwise() method returns wrong results. I have coded a quick hack with coords stored in floats, and this seems to fix the problem, but I find many other places where this problem might show up, so I'll continue searching for a real solution. Gerd
Date: Tue, 24 Jan 2012 22:50:07 +0100 From: wmgcnfg@web.de To: mkgmap-dev@lists.mkgmap.org.uk Subject: Re: [mkgmap-dev] mkgmap r2179: Still NullPointerException in BoundaryUtil
Rather the same method is implemented in Multipolygon.calcAreaSize(List<Coord> points). Maybe the two methods can be merged to use the same area calculation. The Multipolygon variant uses doubles. Maybe the fixes it.
I don't mind moving the calculation from the Multipolygon class to a better util class.
WanMil
Hi WanMil,
I verified the result of the area calculation with a slightly different algorithm that I found in the net, that produced always exactly the same result. Maybe it is not a good idea to use integer/long values for this. I'll continue investigation with this tomorrow.
ciao, Gerd
Date: Tue, 24 Jan 2012 22:30:50 +0100 From: wmgcnfg@web.de To: mkgmap-dev@lists.mkgmap.org.uk Subject: Re: [mkgmap-dev] mkgmap r2179: Still NullPointerException in BoundaryUtil
Hi,
in case your are right Gerd (I have no time today to check that in deep), it is a bug of the clockwise method. So this method needs a bugfix - not the parts that use the method clockwise method.
WanMil
Hi Thorsten,
I can confirm this problem. I can repdroduce it with a "boundary" which contains these 4 points: [2412980/371559, 2412981/371558, 2412992/371544, 2412980/371559] The Way.clockwise() method says it is counter-clockwise. Besides rounding errors, it is a line, not a shape, and I guess that causes the problem. The clockwise() method calculates an area size of 3, which is almost zero.
It is quite likely that the intersection of two areas produces these "nearly empty" (parts of) areas, what we need is a filter that throws them away.
I am not sure if WanMil is already working on this?
Gerd
Thorsten Kukuk wrote
Hi,
I compiled mkgmap r2179 and tried to create the bounds files with it:
osmconvert data/planet/planet-120122.osm.pbf --out-o5m osmfilter data/boundaries/world-20120122.o5m --keep-nodes= '--keep-ways-relations=boundary=administrative =postal_code
postal_code='
mkgmap --createboundsfile=data/boundaries/boundaries-20120122.osm.gz --bounds=data/boundaries/20120122 Schwerwiegend (BoundarySaver): Calculate bbox to (-90.12222290039062,-180.24444580078125) to (82.6120376586914,179.17156219482422) Exception in thread "main" java.lang.NullPointerException at
uk.me.parabola.mkgmap.reader.osm.boundary.BoundaryUtil.loadBoundaryFile(BoundaryUtil.java:169)
at
uk.me.parabola.mkgmap.reader.osm.boundary.BoundaryPreparer.workoutBoundaryRelations(BoundaryPreparer.java:144)
at
uk.me.parabola.mkgmap.reader.osm.boundary.BoundaryPreparer.run(BoundaryPreparer.java:110)
at uk.me.parabola.mkgmap.main.Main.endOptions(Main.java:333) at
uk.me.parabola.mkgmap.CommandArgsReader.readArgs(CommandArgsReader.java:126)
at uk.me.parabola.mkgmap.main.Main.main(Main.java:112) Command exited with non-zero status 1
Thorsten
-- Thorsten Kukuk, Project Manager/Release Manager SLES SUSE LINUX Products GmbH, Maxfeldstr. 5, D-90409 Nuernberg GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer, HRB 16746 (AG Nürnberg) _______________________________________________ mkgmap-dev mailing list mkgmap-dev@.org http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
-- View this message in context: http://gis.638310.n2.nabble.com/mkgmap-r2179-Still-NullPointerException-in-B... Sent from the Mkgmap Development mailing list archive at Nabble.com. _______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
_______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
_______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev
_______________________________________________ mkgmap-dev mailing list mkgmap-dev@lists.mkgmap.org.uk http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev