wkt do not duplicate first point at the end for "non closed" ring about geometry

[vissarion]

If you define a ring as non closed and then you omit the repeated point it is a valid geometry. If you define a ring as closed and then you omit the repeated point the geometry is non valid. You can run correct algorithm and then the geometry will be corrected to have the repeated point. So I think everything is as expected here.

Interestingly though, if you define a geometry to be open (as in your case) and you include the repeated point in the creation then the geometry is valid. @barendgehrels is this intended?

[GS6902]

If you define a ring as non closed and then you omit the repeated point it is a valid geometry. If you define a ring as closed and then you omit the repeated point the geometry is non valid. You can run correct algorithm and then the geometry will be corrected to have the repeated point. So I think everything is as expected here.

Interestingly though, if you define a geometry to be open (as in your case) and you include the repeated point in the creation then the geometry is valid. @barendgehrels is this intended?

Thanks a lot for your answer,
My point is that a non closed ring is an oxymoron.
If I follow you a non closed ring (a ring with template parameter set as non closed) is just non closed and a synonym of a linestring ?
Or it is a ring. So it is closed but by convinience the last point can be omitted in case of a "non closed" ring (by setting the template parameter)

In that case I suppose the result of valid ring (closed and non closed) shouldn't change because it model the same concept

[vissarion]

The correct case is the second case you mentioned. A non closed ring is not a linestring.