Paweł Paprota
Paweł Paprota
OOM killer kicked in again tonight... Some logs: ``` D, [2013-01-04T02:44:57.257553 #7989] DEBUG -- : W 152813734 (2) D, [2013-01-04T02:44:57.280255 #7989] DEBUG -- : tile_count = 2147504146 D, [2013-01-04T02:55:37.107015 #7989]...
Right now the feed generated by OWL does not have too much interesting content, just changeset id and user name. There is much more information available so let's make use...
Right now they are derived from tiles at zoom level 16 which makes them tiny and also makes a lot of them at high zoom levels - this makes browsers...
At this point any changes made to relations are simply ignored. I'm not sure how to handle it. Geometry changes to relation members is reflected in node/way changes so nothing...
Now it's just x/y/zoom. Need to implement something like quad tree for indexing tiles.
`-L` option is missing in `find` invocation in that script.
Look at this bad boy: http://www.openstreetmap.org/browse/changeset/13277695 OSC file is nearly 4MB! I got HTTP timeout from Changemonger on this one. Perhaps there should be some limit of how many elements...
Right now some descriptions are like "XXX edited unnamed fence" :-) I think just "XXX edited a fence" would be enough. In fact I can't see for what features "unnamed"...
Test instance reports HTTP 500. My own instance says "Nicky edited twenty-four buildings" while it's pretty clear that it was much less than 24: http://www.openstreetmap.org/browse/changeset/13308929
Calling encode('utf-8') is a Bad Idea (tm) - there should be unicode objects flowing from the beginning of time. So for example in changemonger.py: `xml = et.XML(data.encode('utf-8'))` should be `xml...