On Sat, May 17, 2008 at 1:57 AM, Frederik Ramm <[EMAIL PROTECTED]> wrote: > Hi, > >> I ran a test here with similar inputs (code appended). > > Thanks for the effort! > > [...] > >> Anyway, I got the images: >> http://beerwarmer.randomjunk.co.uk/progress-images/karlsruhe-ll.png >> http://beerwarmer.randomjunk.co.uk/progress-images/karlsruhe-sph.png >> >> which is what you'd expect. > > Yes. > >> If you send me your exact python and osm.xml I can take a closer look >> to see if I can spot the problem > > Just to be sure, I have used exactly the scripts you provided, > together with my style files, and had the same wrong result as before > - the ll variant looks ok, the spherical broken. > > My style files are at > > http://www.remote.org/frederik/tmp/osm-ll.xml > http://www.remote.org/frederik/tmp/osm-sphmerc.xml > > - they are pretty much OSM bog standard I should think.
You have the postgis inputs with the srs set to latlong, but you also have the estimate_extent off, and extent hardcoded to the (spherical) mercator coordinates... this may be interfering with mapnik's attempt to determine whether the layer contains any relevant data. This kind of makes sense of your images, which seem to be the right projection compared to mine. just missing a lot of data (if you switch between my sphmerc image, and your "wrong" image the forest and river matches exactly). Try setting the estimate_extent parameter to true, and removing the extent parameter (you won't have any probs with this in germany... just southern hemisphere and far east will get missed because postgis' estimation is rubbish). Alternatively just set the extent parameter in latlong coords (but I can't remember the lat, long order... probably -179,-89,179,89 will work). If it still doesn't work I'll see if I can replicate the situation more precisely my end. Dave _______________________________________________ Mapnik-users mailing list [email protected] https://lists.berlios.de/mailman/listinfo/mapnik-users
