[OSM-dev] planet diff and excerpts

Brett Henderson brett at bretth.com
Thu Oct 11 14:14:26 BST 2007

I should mention that if anybody comes across bugs in osmosis changeset 
handling, please let me know and I'll fix it before anything else.  I'm 
keen to see this working properly.

Of course, if anybody is prepared to fix bugs in osmosis for me that 
would also be most welcome :-)

The issues with the polygon code are not as high a priority with the 
perl script being available.  I'd just like it to be available in 
osmosis because it's pretty powerful to have all that functionality in a 
single tool.

Brett Henderson wrote:
> Frederik Ramm wrote:
>> Hi,
>>     probably a question to Brett H but I think others may be interested.
>> Say I have a planet extract containing Germany, and I want to keep  
>> that up to date using the planet diffs.
> I've been wondering about this myself ...
>> Then I'll just apply the normal (full planet) diffs to my extract 
>> and  run a new polygon extraction on the result which will give me 
>> the  updated excerpt, right?
> Yeah, I guess that should work.  I must say I hadn't thought of doing 
> this though.  It will rely on the polygon tool being able to copy with 
> ways and relations referring to non-existent nodes and ways.  I don't 
> think this will be a problem though.
>> And afterwards I could diff my old data and my new data and thus  
>> create a diff that is limited to the excerpted area.
> Yes, that should be possible.
>> I don't suppose it would be possible to make something like a 
>> polygon  excerpt from a diff, because the diff doesn't necessarily 
>> contain the  location where a change occurred (if the change is e.g. 
>> in a way).
> Yes, that's the tricky bit.
> I've just posted some questions to the "relations without members" 
> thread.  I'm curious exactly how the extract-polygon perl script 
> works.  Osmosis polygon extraction is broken at the moment (I forgot 
> to finish it during the 0.5 changes ...), I'm hoping to get it working 
> again but if it's become horribly complicated I may have to pull the 
> tasks out again.  If relations can contain other relations I don't 
> think osmosis will be able to support it because it always works in a 
> streamy fashion, working around that will be difficult.
> Cheers,
> Brett

More information about the dev mailing list