dedicated handling of touching inner rings in multipoygons

[tyrasd]

The previous workaround of using buffer(0) to let JTS join any touching inner rings can be very slow, up to the point of potentially bringing a query to an almost complete halt when it has to be called on very large multipolygons (e.g. relation/1205151).

This adds dedicated handling of touching inner rings, and merges them before the geometry object is created.

A few test cases were adjusted, because the new handling is following more strictly the multipolygon validity rules given in the osm wiki. (related: #123, #129)

Also includes some minor code cleanup (e.g. adding of a private construtor to the static geometry builder utility class to hide the implicit public constructor).

Checklist

• My code follows the code-style rules and I have checked on the static analyses and benchmark (if applicable) results
• I have commented my code
• I have written javadoc (required for public methods)
• I have added sufficient unit tests
• I have made corresponding changes to the documentation
• I have updated the CHANGELOG.md
• I have adjusted the examples and benchmark if necessary

[joker234]

Some open points need to be considered, mostly questions or suggestions.

[joker234]

@tyrasd I checked all finished points of your checklist above. Can you think about whether we need to do the remaining ones (benchmarks, changelog, examples, documentation)

[tyrasd]

@tyrasd I checked all finished points of your checklist above. Can you think about whether we need to do the remaining ones (benchmarks, changelog, examples, documentation)

I would definitely add this to the changelog. additional benchmarks would be nice, but let's see if we find time for that. documentation/examples are probably not needed, since the API stayed the same.