splitter broken with multiple files

I updated this morning and ran the splitter with mass/conn/ri extracts (3 files). I got a map that had only connecticut. SO I wonder if the bounds code in r102 is using the last bound, rather than unioning them. here's my log. search down for connecticut bounds, and the bounds chosen for writeout. cache= description= geonames-file= legacy-mode=false mapid=63240001 max-areas=255 max-nodes=800000 mixed=false overlap=2000 resolution=13 split-file= status-freq=120 write-kml= Elapsed time: 0s Memory: Current 80MB (2MB used, 78MB free) Max 1994MB Time started: Thu Dec 24 09:07:50 EST 2009 Map is being split for resolution 13: - area boundaries are aligned to 0x800 map units - areas are multiples of 0x1000 map units wide and high The input osm file(s) will be re-parsed during the split (slower) because no --cache parameter was specified Processing massachusetts.osm.bz2 A <bounds/> tag was found. Area covered is (41.23593807220459,-73.50664615631104) to (42.888779640197754,-69.93184089660645) 2,500,000 nodes processed... 5,000,000 nodes processed... Elapsed time: 2m 0s Memory: Current 80MB (27MB used, 53MB free) Max 1994MB 7,500,000 nodes processed... Processing rhode_island.osm.bz2 A <bounds/> tag was found. Area covered is (41.149635314941406,-71.89511060714722) to (42.0211386680603,-71.12078905105591) Processing connecticut.osm.bz2 A <bounds/> tag was found. Area covered is (40.99842309951782,-73.72570753097534) to (42.0546555519104,-71.78900241851807) A total of 9,772,793 nodes, 0 ways and 0 relations were processed in 3 files Min node ID = 196471 Max node ID = 591033429 Time: Thu Dec 24 09:10:22 EST 2009 Exact map coverage is (40.99842309951782,-73.72570753097534) to (42.0546555519104,-71.78900241851807) Trimmed and rounded map coverage is (40.95703125,-73.7841796875) to (42.099609375,-71.7626953125) Splitting nodes into areas containing a maximum of 800,000 nodes each... 3 areas: Area 63240001 covers (0x1d2000,0xffcb8800) to (0x1df000,0xffcc3800) Area 63240002 covers (0x1d2000,0xffcc3800) to (0x1da000,0xffccf800) Area 63240003 covers (0x1da000,0xffcc3800) to (0x1df000,0xffccf800) Writing out split osm files Thu Dec 24 09:10:22 EST 2009 Processing 3 areas in a single pass Starting pass 1 of 1, processing 3 areas (63240001 to 63240003) Processing massachusetts.osm.bz2 A <bounds/> tag was found. Area covered is (41.23593807220459,-73.50664615631104) to (42.888779640197754,-69.93184089660645) 2,500,000 nodes processed... Elapsed time: 4m 0s Memory: Current 80MB (49MB used, 31MB free) Max 1994MB 5,000,000 nodes processed... 7,500,000 nodes processed... Writing ways Thu Dec 24 09:12:34 EST 2009 500,000 ways processed... Writing relations Thu Dec 24 09:13:19 EST 2009 Processing rhode_island.osm.bz2 A <bounds/> tag was found. Area covered is (41.149635314941406,-71.89511060714722) to (42.0211386680603,-71.12078905105591) Processing connecticut.osm.bz2 A <bounds/> tag was found. Area covered is (40.99842309951782,-73.72570753097534) to (42.0546555519104,-71.78900241851807) Elapsed time: 6m 0s Memory: Current 80MB (58MB used, 22MB free) Max 1994MB Wrote 9,772,793 nodes, 799,729 ways, 1,537 relations Time finished: Thu Dec 24 09:14:03 EST 2009 Total time taken: 372s

Greg Troxel <gdt@ir.bbn.com> writes:
I updated this morning and ran the splitter with mass/conn/ri extracts (3 files). I got a map that had only connecticut. SO I wonder if the bounds code in r102 is using the last bound, rather than unioning them.
I updated to -r101, and that produced a map of all three states. So it seems that the combination of bounds code and multiple input files is trouble.

Yes you're right, unfortunately the latest changes I made (to allow recursive splitting) didn't take into account splitting mutliple files at once. It shouldn't be too hard for me to fix this however I doubt I'm going to get a chance before the New Year. Sorry for the inconvenience, I'm glad to hear r101 is working OK for you in the meantime though. Merry Christmas to everyone in the mkgmap community! Chris GT> --=-=-= GT> GT> Greg Troxel <gdt@ir.bbn.com> writes: GT>
I updated this morning and ran the splitter with mass/conn/ri extracts (3 files). I got a map that had only connecticut. SO I wonder if the bounds code in r102 is using the last bound, rather than unioning them.
GT> I updated to -r101, and that produced a map of all three states. GT> GT> So it seems that the combination of bounds code and multiple input GT> files is trouble. GT> GT> --=-=-= GT> Content-Type: application/pgp-signature GT> -----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.10 (NetBSD)
iEYEARECAAYFAksztIUACgkQ+vesoDJhHiU5ygCgmXC9Lk/cXuQS4yNB0prsAUPh lpgAnAy8/KBmSYMSwYSvREzU4uyFJmHB =/Pl2 GT> -----END PGP SIGNATURE----- --=-=-=--
participants (2)
-
Chris Miller
-
Greg Troxel