On Tue, Dec 21, 2010 at 10:28 AM, Jochen Topf <[email protected]> wrote: > On Tue, Dec 21, 2010 at 10:08:41AM -0600, Scott Crosby wrote: >> On Mon, Dec 20, 2010 at 6:01 PM, Scott Crosby <[email protected]> wrote: >> > Yes. That is an osmosis bug. That header block is required. >> >> I would like to clarify this slightly. There must be an OSMHeader >> block before the first OSMData block, however, the OSMHeader block >> does not have to be the first block in the file. Also, parsers should >> ignore fileblocks types that they do not understand. > > Hm. I see I have some more coding to do then... :-)
Sorry. I could have been more explicit. This is intended for forward compatibility; anyone can define and add their own application-specific block types to the file. It should be fairly easy to skip those blocks; have a front-end filter that skips any block thats not 'OSMHeader' or 'OSMData'. Then the first unskipped block must contain an OSMHeader block. > > I am pretty happy at the moment though. My code that generates all the > statistics for Taginfo (http://taginfo.openstreetmap.de) runs in about 2.5 > hours for the whole planet, thats less than half the time it needed when > it was using the XML file format. PBF rocks! Cool! Scott _______________________________________________ dev mailing list [email protected] http://lists.openstreetmap.org/listinfo/dev
