Difference of gpolygons giving wrong result

[gastald88]

I have two polygons, p1 (orange) and p2 (blue) as in figure.
The difference of p1 and p2 returns a (wrong) empty result.

This can be replicated in the following godbolt link:
https://godbolt.org/z/GT7aWaWzW

[barendgehrels]

Thanks for the report. I can reproduce it.

issue_1349<d> a: 2354350.1 b: 230426.9 union: 2355360.2 intersection: 229416.76
diff(a-b): 0 diff(b-a): 1010.1469 sym_diff: 1010.1469 
checks: 1.1641532e-10 2124933.3 -2.6557245e-10 0 OK WRONG OK OK

[barendgehrels]

Detailed note (mainly for myself).

It's caused by the fix of #869 , a condition for difference added by me in 2021.

I have to revise that part of the code. Skipping the reverseMeaningInTurn is fine for all test cases, except issue_869 which has correct areas, but is considered as invalid.

In the traversal_switch_detector.hpp it should not be necessary change behavior for difference.

[barendgehrels]

PR fixing this issue is published