WIP: Add test for GEOS reported Orientation#isCCW robustness issue

[strk]

See libgeos/geos#1425

[strk]

For the record, tested locally the test fails:

OrientationIsCCWTest.testGeosTicketGH1425:113->checkCCW:118 Coordinate array isCCW: expected:false but was:true

[strk]

@dr-jts by reading the code I was under the impression that similar case should be correctly handled but the Orientation#isCCW method, am I wrong ?

[strk]

The case at hand is this (in HEXWKB form):

0103000000010000000A000000890205230D6C31408177583619475140FD613B270D6C314031AA5436194751405D58985776943140195EE965DE47514085FFA6647B9F31408FC1C62B394851405D58985776943140195EE965DE475140FD613B270D6C314031AA543619475140A4613B270D6C31402FAA543619475140890205230D6C31408177583619475140320205230D6C31407E77583619475140890205230D6C31408177583619475140

[strk]

Simpler case:

POLYGON ((0 0,0 2, 2 1, 2 0, 0 2, 0 0))

that is, the next figure going clockwise starting from the bottom-left vertex is reported to be counterclockwise instead

[dr-jts]

Both of these polygons compute as invalid. The CCW algorithm isn't certified to work correctly on invalid geometry.

(The invalidity of the second polygon is obvious; the first one is hard to see).