drop false dead-ends

[anastassiavybornova]

after first simplification round, drop edges (which have NOT been introduced by the algorithm but were already there in the orig network) and which now have dead-ends (nodes of degree 1 at their end). (example: nonplanar highway intersection Liège)

[martinfleis]

Do these actually exist?

[jGaboardi]

(example: nonplanar highway intersection Liège)

@anastassiavybornova Can you point the the area/bbox where this is happening and I can attempt to recreate an example...

Do these actually exist?

... if it exists?