search

pnorman / ogr2osm

pnorman's version of UVM's Rewrite of ogr2osm
Other
78 stars 46 forks source link

Proposal for issue#30: Splitting long ways #34

Closed flire closed 7 years ago

pnorman commented 7 years ago

Is there anything in particular that needs scrutiny? Otherwise it looks good and I'll merge it untested.

flire commented 7 years ago

I have second thoughts about the following case: when a way doesn't belong to a relation, a new relation will be created, and all the parts of the way will get empty role (""). I've considered to assign "outer" role, but in that case a splitted way should be closed to form a multipolygon, which isn't always true. Other than that everything should be fine.

pnorman commented 7 years ago

If it's a way from a LINESTRING it shouldn't result in a relation at all

flire commented 7 years ago

So should it just be splitted without any additional info that a bunch of ways was a single linestring? And considering the fact I split ways after the parsing, I should store some kind of info what type a feature had before splitting. It's not a big deal though, I'll implement that and update the pull request.

flire commented 7 years ago

Updated.

pnorman commented 7 years ago

So should it just be splitted without any additional info that a bunch of ways was a single linestring?

Yes

flire commented 7 years ago

So should it just be splitted without any additional info that a bunch of ways was a single linestring? Yes

The current version should do exactly this: parts of a polygon are joined within a relation, parts of a line become separate ways.

pnorman commented 7 years ago

merged, and granted commit access.

As you might have noticed, I'm not very active with ogr2osm these days, having other priorities at present.