Strange routing result

Any idea what is causing this weird routing result ?

The first problem here is that there is a no-right-turn restriction that is circumvented by driving this odd loop (which is not forbidden according to the current OSM data, even though in reality it is illegal/not possible). The second problem is that the route doesn’t simply go straight from the beginning, which is probably due to the weighting of the different highway types.

This way should probably rather be trunk_link than trunk, just like this way

Maybe some minor turn costs would have prevented this? configure turn cost calculation by karussell · Pull Request #2957 · graphhopper/graphhopper · GitHub

And find another strange result by the GH router.

The osrm result is without that unexpected unnessary diversion.

This is a bug that we’ll have to fix.