Change MultiPoint WKT to use parentheses in sub-members

[mwtoews]

This is a potentially breaking change that switches the WKT output of MultiPoint from:

MULTIPOINT (2 8, 3 9)

to

MULTIPOINT ((2 8), (3 9))

This aligns with JTS' outputs of MultiPoint and with ISO/OGC specs.

See also postgis/postgis#682 which was implemented for PostGIS 3.3

[pramsey]

I wonder how many downstream projects have tests against the WKT output of MULTIPOINT.

[mwtoews]

Possibly a few, for instance I'll try shapely later. Another strategy is to modify WKTWriter to enable an option, keeping the default to use old behaviour.

[dr-jts]

I wonder how many downstream projects have tests against the WKT output of MULTIPOINT.

I think we have to note this as a breaking change and move on. It's more important to be standards-compliant.

[dr-jts]

WKTReader can still handle both forms, right?

It's good to update the unit tests that aren't specifically checking this behaviour. But are there still tests for WKTReader that check both formats?

[mwtoews]

WKTReader can still handle both forms, right?

Yes, and I intentionally left a few test cases with the previous form. The only exception that I'm aware of #902 which will be closed as "not planned".