Hi all, I've committed r239 now. - r237 and r238 fixed o5m read and write issues r239: - don't calculate bbox for all ways of a multipolygon rel, instead try to find closed polygons and calculate the bbox for each polygon. - don't treat type=boundariy relation like multipolygon Relation - add --output=simulate for testing, this allows to run the split process without writing the OSM output files. - make output better readable I think the --keep-complete is now working good enough, please let me know if anything is wrong or missing. Next on my todo list: - add/complete javadoc - update the doc file intro.txt - add o5m support to mkgmap - make sea generator compatible with new osmwriter interface - try to use thread(s) for the OSM reader - try to calculate reasonable value for max-areas based on available heap, number of tiles and file size -- View this message in context: http://gis.19327.n5.nabble.com/splitter-r239-tp5736433.html Sent from the Mkgmap Development mailing list archive at Nabble.com.