Hi all, I got no feedback on the do_not_split_0x4a_polygon.patch which I provided to fix the problems reported by Thomas and Andrzej, see http://gis.19327.n5.nabble.com/How-do-I-transform-the-name-of-all-Ways-tp587... I checked older versions of the source and found out that older versions of the code did contain a special treatment for the 0x4a polygons, I removed those in the overview2 branch and I don't exactly remember why, so I've reverted that part of the change. I have the feeling that the code should be changed somewere else, the tests in PolygonSubdivSizeSplitterFilter.java were introduced by WanMil long before we changed the split algo in MapSplitter. Both sources use diverse hard coded limits, and I have no idea why these are not relevant for 0x4a polygons. I assume that there is a maximum tile size (at least we have it in splitter) so maybe mkgmap should stop with an error message when one tries to create a map with larger tiles? Gerd