st_segmentize changes start/end point coordinates of edges for lat/long data about sfnetworks

[luukvdmeer]

RIght now, the way of working was to only write an sfnetwork method for geometric unary operations that never break the restrictions - i.e. 1) nodes being points and edges, if spatially explicit, being lines and 2) endpoints of edges, if spatially explicit, are exactly equal to the coordinates of its corresponding nodes.

Because of this rule, the methods didn't have to pass through the checks of st_geometry<-.sfnetwork when replacing the geometries. It seems st_segmentize breaks restriction 2 in some cases, but not always. There are two options:

• Keep the workflow as it is, and thus, don't have a method for st_segmentize.
• Always pass through the geometry checks when applying a geometric unary operation, and thus allowing methods for operations that sometimes break the restrictions, but not always.

(Either way I think applying st_segmentize to the edges does not have a real practical use case, but that is another story)

[luukvdmeer]

I opted for solution 1 for now