ST_OrderingEquals returning true for mis-ordered multipoints

[pramsey]

This query seems wrong on the face:

select '69a',st_orderingequals('MULTIPOINT(2 2, 1 1)'::GEOMETRY,'MULTIPOINT(1 1,2 2)'::GEOMETRY) as bool;

Returns true but should return false? This was/is actually part of the old regression suite for ~= operator (now changed to _st_orderingequals()) and expected the "wrong" answer (true) there too.

[pramsey]

Updated behavior of lwgeom_same to match "orderingequals" semantics in trunk r4789