Hi WanMil, it's a combination of errors. OSM data is incorrect because it contains two admin-level 4 boundaries that intersect (r1561776 and r349048) The program is incorrect because it doesn't tag this intersection with admin-level 2 information. The trunk version doesn't care about intersections of equal levels and produces better results. I try to find a fix for mkgmap. Gerd WanMil wrote

Hi,

I have commited a new tool BoundaryCoverageUtil to the performance branch. Thanks to Gerd who ported it to the new boundary format.

I have tested it with freshly compiled boundaries of africa and found some mysteries. http://files.mkgmap.org.uk/detail/60 contains a zip file with the coverage of africa with admin_level=2 (as GPX). If you look at the Canaries you see that the sea boundary is ok, but there is a hole along the coastline.

I haven't started to track down the problem but maybe you have an idea, Gerd? It is possible that the boundary data is wrong, or the Preparer has a problem or that the coverage util fails.

WanMil _______________________________________________ mkgmap-dev mailing list mkgmap-dev@.org http://www.mkgmap.org.uk/mailman/listinfo/mkgmap-dev

-- View this message in context: http://gis.19327.n5.nabble.com/BoundaryCoverageUtil-tp5559258p5560568.html Sent from the Mkgmap Development mailing list archive at Nabble.com.