Release notes for 1.2.0.0 release mid September

[jodygarnett]

These release notes match up with: https://github.com/locationtech/jts/milestone/8?closed=1

This is the release review page: https://projects.eclipse.org/projects/locationtech.jts/releases/1.20.0

For background reading on RelateNG:

• https://lin-ear-th-inking.blogspot.com/2024/05/jts-topological-relationships-next.html
• https://lin-ear-th-inking.blogspot.com/2024/05/relateng-performance.html

[jiayuasu]

@jodygarnett Could you help @dr-jts with this PR? #734

We are pushing the Apache Parquet and Apache Iceberg community to use JTS WKB reader / writer (ISO variant) for encoding geometries with XYZM coordinates. I really want to fix this issue in this release otherwise this will lead to unexpected outcome in many cloud db engines.

[jodygarnett]

I am a little out of the loop; I see your questions on the PR - I think we should assume lack of resources for the project.
I know I have been away for personal reasons and doing my best to make a release to help out.

I am trying to think of developers who would be comfortable reviewing such work... maybe Andrea or Gabe from the GeoTools project?

aside: can you rebase and resolve conflicts please.

[jodygarnett]

Getting some test failures downstream in GeoTools:

[ERROR] Errors: 
[ERROR]   ClippedFeatureCollectionTest.testClipGeometry:362 » IllegalArgument Coordinate...
[ERROR]   ClippedFeatureCollectionTest.testClipPoly3DNewVertices:134 » IllegalArgument C...
[ERROR]   CoordinateSequencesTest.testEqualityND:163 » IllegalArgument CoordinateXY dime...

The failures all look legitimate so far.

Resolved with: #1071

[jodygarnett]

Next change (unclear if regression or just different) comes from MaximumInscribedCircle.getCenter(multiPolygon, precision).

    Polygon polygon = (Polygon) read("Polygon ((0 5, 5 10, 10 6, 15 10, 20 5, 15 0, 10 4, 5 0, 0 5),(5.4267578125 6.68164062499999822, 3.7451171875 5.30761718749999822, 5.365234375 3.21582031249999822, 8.3388671875 5.08203124999999822, 5.4267578125 6.68164062499999822))");
    Point point = MaximumInscribedCircle.getCenter(polygon, 1);
    Point expected = (Point) read("Point (15 5)");

    Assert.assertEquals(expected, point);

Internally this seems to be a difference with this call:

Cell farthestCell = createCentroidCell(inputGeom);

Returning:

• JTS 1.19.0: POINT (15,5)
• JTS 1.20.0: POINT (14.116248497596153 5.041015624999999)

Okay they uses different starting locations to start their grid:

• JTS 1.19.0 uses: Point p = geom.getInteriorPoint();
• JTS 1.20.0 uses: Point p = geom.getCentroid();

[jodygarnett]

Next change/regression appears to be a fragile test checking CentreLine and Skeletons coming out of VoronoiDiagramBuilder processing. Cannot really tell if this is an intensional change or a side effect as test lacks sufficient detail.

[ERROR]   CenterLineTest.testGetCentreLine:69 equalsNorm
[ERROR]   CenterLineTest.testGetSkeleton:78 expectedSkel

Here is the PR if you wish to see the different out: geotools/geotools#4884