Martin Raifer

Hm, that's actually implemented like that by design: #23 What's your use-case that requires multiple osm objects (with the same id) to result in multiple GeoJSON features?

Can you share an example of what data is returned by `/search/pass/` in this code (i.e. what's inside the `data` varaible when calling `osmtogeojson(data)`)? Are you sure it is a...

ooh, it must be that I forgot to update the browserified&minified version before pushing the 3.0.0-beta(s?) to github/npm (it works as expected in the non pre-built code) and must have...

Try additionally: `sudo apt-get install nodejs-legacy` (did you try [googling](http://www.lmgtfy.com/?q=%2Fusr%2Fbin%2Fenv%3A+%E2%80%98node%E2%80%99%3A+No+such+file+or+directory) the error message :wink: ?) The problem is that this depends from system to system. For example, if one works...

fixed in #136 (released as version `3.0.0-beta.5` on npm)

> decide if this needs to be integrated more closely into the library The question here is, if tiny-osmpbf should be fully integrated into osmtogeojson (i.e. `osmtogeojson(pbf_data)` would work), or...

Hmm. Wouldn't it be better to perform this kind of analysis on the OSM data before converting it to geojson? Then, you could simply have separate information like _"door with...

Ok, that does indeed sound like a useful information to have!

since now (almost) all old-style "simple" multipolygons are [gone](https://lists.openstreetmap.org/pipermail/talk/2017-May/077938.html), it could make sense to disable this option by default.

Yeah… But the question is if the respective OSM relations actually do represent some "geometric" feature on their own or not. For example in the case of this turn restriction:...