Tolerance silently ignored in ST_RemoveRepeatedPoint for multipoint

[strk]

According to ​http://postgis.net/docs/ST_RemoveRepeatedPoints.html ST_RemoveRepeatedPoints considers vertices within the tolerance of one another as identical in case a tolerance is given, but this does NOT happen for MULTIPOINT.

Also, I suggest to specify that for lines the distance between points is only computed for *consecutive* points.

[pramsey]

I don't necessarily agree with the premise of the ticket: for multi-point the operation of the function doesn't make much sense. Maybe use unaryunion for duplicate point removal in multipoint?

[pramsey]

See lwmpoint_remove_repeated_points and add a tolerance there. It's expensive, but hey!

[pramsey]

Doco is correct, current code removes exact repeated points, and shouldn't be too hard to make it respect tolerance, though algorithm in there is O(n2). Forgot to re-open ticket, am actually going to add the tolerance support.

[robe]

pramsey this one seems safe enough to do even after I release beta. You can even do this in 2.3.4 if you are so energized after 6 days of basking in the sun.

[pramsey]

OK, so made the code match the doco, seemed reasonable for this corner case even if the code isn't so efficient. Done in trunk at r15637