Hi Michael, The JOSM format differs from the Osmosis format in a number of ways, not just version information. In order to open a JOSM file properly you really need to handle other information including the extra action attribute.
I try to keep Osmosis "pure" wherever possible to avoid lots of special case coding, avoid weird corner cases, and keep the behaviour simple and predictable. For that reason I'm hesitant to add JOSM specific logic to the parser. I'd prefer to treat the JOSM file format separately and create a new task capable of reading JOSM (eg. --read-josm-xml). Brett On Sun, Aug 22, 2010 at 7:27 PM, Michael Zilske <michael.zil...@gmail.com>wrote: > Hi, > > I'm using Osmosis for extracting road networks and public transit > routes from OSM and converting them to our own file format with my own > Sink. (We do transportation research.) > > We work on an API 0.6 planet extract .osm file which we sometimes open > with JOSM to manually add some tags of our own, e.g. as conversion > hints for our application. We're not checking this in, because it's of > no use to anyone else. > > Now my problem is that the Osmosis FastXMLParser chokes on the .osm > files saved by JOSM, because newly added entities get neither a > "version" nor a "changeset" attribute in JOSM's interpretation of the > .osm format. > > I helped myself by patching the Parser to simply ignore the fact that > these mandatory attributes aren't there. Is there a better solution to > this? Would it be an option to integrate this into Osmosis, or don't > you want to parse these "improper" files? > > Cheers, > Michael > > _______________________________________________ > osmosis-dev mailing list > osmosis-dev@openstreetmap.org > http://lists.openstreetmap.org/listinfo/osmosis-dev >
_______________________________________________ osmosis-dev mailing list osmosis-dev@openstreetmap.org http://lists.openstreetmap.org/listinfo/osmosis-dev