Bugfix: ExactEquals not correct when both geoms are empty points

[peterstace]

Description

Fix bug in `ExactEquals` for empty Points

The `ExactEquals` method didn't properly take into consideration
coordinates type (XY vs. XYZ vs. XYM vs. XYZM) when both Point arguments
were empty.

The actual fix is only in the `pointsEq` method. In the previous code,
`ok1` and `ok2` would both evaluate to `false` (because the `Points` are
both empty). This would then skip the call to `c.eq(c1, c2)`, which was
where the coordinates type comparison occurred.

The fix is to explicitly check for the case where both `Point`s are
empty, and handle the coordinates type check there. The `c.eq` method
still needs to check for coordinate type, to ensure that inputs such as
`POINT (1 2)` and `POINT Z (1 2 3)` continue to compared as non-equal.

Check List

Have you:

• Added unit tests? Yes.

• Add cmprefimpl tests? (if appropriate?) N/A

• Updated release notes? (if appropriate?) N/A

Related Issue

• N/A

Benchmark Results

N/A

[peterstace]

This is the actual bugfix. The other changes are just additions to the unit test cases.

[albertteoh]

Can two non-empty points with different coordinate types ever be considered equal? Or perhaps this is handled by c.eq(c1, c2).

[peterstace]

Yes, that case is handled by the call to c.eq(c1, c2).