Based on the error message you've shown, I'm pretty sure there isn't a problem with the XML parser itself (it handles <changeset/> vs </changeset> automatically). The end of the input stream is being reached, so it looks like either the planet file that you're trying to parse is corrupt, or the bz2 code is having trouble uncompressing the file and thinks it has reached the end earlier than it should. I've seen the latter happen before, unfortunately the bz2 unpacker we're using isn't all that great... If you're sure the planet file is OK and suspect the bz2 code is at fault, you could uncompress the planet file (assuming you have a spare 140GB or so!) and try again. I realise this is hardly ideal. Rather than try to fix the bz2 code, I'll probably add a parameter that allows the input to come from stdin. That way it will be possible to pipe the osm file from an external uncompression tool, which should also greatly help performance. RK> Every time I try to split the planet OSM file I get this: RK> RK> A <bounds/> tag was found. Area covered is (-90.0,-180.0) to RK> (90.0,180.0) RK> Error opening or reading file java.io.EOFException: no more data RK> available - RK> expected end tags </changeset></osm> to close start tag <changeset> RK> from RK> line 4312 and start tag <osm> from line 2, parser stopped on TEXT RK> seen RK> ...006-04-21T20:39:35Z" closed_at="2006-04-21T23:28:58Z" RK> open="false"... RK> @4312:103 RK> java.io.EOFException: no more data available - expected end tags RK> </changeset></osm> to close start tag <changeset> from line 4312 and RK> start RK> tag <osm> from line 2, parser stopped on TEXT seen RK> ...006-04-21T20:39:35Z" RK> closed_at="2006-04-21T23:28:58Z" open="false"... @4312:103 RK> at org.xmlpull.mxp1.MXParser.fillBuf(MXParser.java:3035) RK> at org.xmlpull.mxp1.MXParser.more(MXParser.java:3046) RK> at RK> org.xmlpull.mxp1.MXParser.parseStartTag(MXParser.java:1800) RK> at org.xmlpull.mxp1.MXParser.nextImpl(MXParser.java:1127) RK> at org.xmlpull.mxp1.MXParser.next(MXParser.java:1093) RK> at RK> uk.me.parabola.splitter.AbstractXppParser.parse(AbstractXppParser.ja RK> va:62) RK> at RK> uk.me.parabola.splitter.Main.processOsmFiles(Main.java:388) RK> at uk.me.parabola.splitter.Main.processMap(Main.java:377) RK> at RK> uk.me.parabola.splitter.Main.calculateAreas(Main.java:267) RK> at uk.me.parabola.splitter.Main.split(Main.java:158) RK> at uk.me.parabola.splitter.Main.start(Main.java:106) RK> at uk.me.parabola.splitter.Main.main(Main.java:95) RK> Time finished: Thu Jan 14 16:47:45 CET 2010 RK> Total time taken: 0s RK> RK> Looks like the splitter expects a <changeset></changeset> pair but RK> gets <changeset [...] /> instead. Is this a bug? RK>