Ian, Made the changes to a 2.3.3 stack install version running on Ubuntu Desktop ver 8.xx. Edited the .js file by hand and then restarted the zenoss using "zenoss restart" as the zenoss user.
Got the same problems as before; when using lat/lon for address the portlet moves the point to the nearest road (most of the time) or off into a nearby location (some of the time). Some of the addresses are in another post in this thread. Let me know if I can somehow get you more information. Rod ian wrote: > Well, you've exposed an excellent bug. Here's what was happening: the > lat/long would be sent to Google for geocoding. Google would try to /reverse/ > geocode it to an address, then /forward/ geocode the result, obtaining a set > of coordinates that in no way necessarily correspond to the originals. I have > no idea why it would do this, but that's my best guess as to what happened > based on the geocoding responses. > > Luckily, the fix is simple. There's no reason at all to send lat/lng > coordinates for geocoding in the first place, so if the address of the > Location is coordinates we can just skip that step. I've put the fix in the > trunk: > > http://dev.zenoss.org/trac/changeset/12202 > > Unfortunately the Google Maps code is quite a bit different in the trunk so > you won't be able to apply the patch directly to an older version. You can > manually apply it to the ZenGeoMap.geocode function, though. You'll need to > make the change in geomap.js /and/ geomap-2.1.js. OR you can just wait for > Blue Crab. > > Let me know how it works out. > > --Ian -------------------- m2f -------------------- Read this topic online here: http://forums.zenoss.com/viewtopic.php?p=32562#32562 -------------------- m2f -------------------- _______________________________________________ zenoss-users mailing list [email protected] http://lists.zenoss.org/mailman/listinfo/zenoss-users
