Andrea Aime

Tested manually, everything seems to work correctly. Merging.

Reading the code, I see a more complex situation: * The value of the variable is [first tried out as an absolute path](https://github.com/geoserver/geoserver/blob/b1a7f63984c47e0033b1fa7c2ebb98cb0bfaafd1/src/platform/src/main/java/org/geoserver/platform/GeoServerEnvironment.java#L117), if it exist, it's going to be...

@dr-jts, I believe that having such setter would work. I found this while working on curves support in SQL Server. After this report I switched the data store to using...

Just stumbled into this by doing some PMD related cleanups in GeoTools. One of the PMD rules for tests is not to use assertTrue when assertEquals is a logical equivalent,...

I'm in agreement with any move that resolves the inconsistencies, the current situation is hard explain and reason about. The test that failed after migration from assertTrue to assertEquals really...

Cause I already have spent enough time on this. Those upgrading JTS in GeoTools will have a reason to change the tests (they are not many, I'm not setting a...

Ideally, it should not, any code trying to compare against anything but a Geometry would receive a NPE. But it would cause a binary incompatibility for sure, since compareTo switches...

Well, no, if there are subclasses implementing compareTo, outside of JTS, they will have to switch the argument from Object to Geometry. In GeoTools we [have a few of those](https://github.com/geotools/geotools/blob/f66689b5305eaf381e21dc2feec8011138d4cb24/modules/library/main/src/main/java/org/geotools/geometry/jts/CompoundRing.java#L518)....

Indeed not interested in either cases for the moment. We could switch regardless to the JTS version if it has the same or better performance characteristics (calculations, allocations), one less...

By the way, both @jodygarnett and @jnh5y were made aware of the GeoTools implementation, you did not talk to the JTS folks before starting this work?