Update Fuzz Tests

[Chillee]

Update Fuzz Tests

[simonlindholm]

fast-delaunay.cpp (Not sure what's going on here... Is looks like # is test failures? That seems bad? It also floods my screen for output and takes forever. @simonlindholm ?).

# indicates that the test case was skipped due to duplicate points. There used to be , and ? markers for skipping tests for concyclicity/colinearity reasons as well, but those were commented-out with the latest Delaunay implementation which handles that. The points of the markers are to indicate a) that the test hasn't hung, b) approximate performance, c) that the test doesn't skip all cases, in a simple-to-implement way. They could probably be removed/commented out. Real test failures terminate the test.

As for the performance, well, more tests is good, right? Delaunay triangulation probably has a lot of special cases. But a fast/slow mode for that test would likely be a good thing.

2sat.cpp (Running main1() causes a segfault. Don't know whether it's because it's an outdated test or 2Sat has problems).

You're just hitting the stack limit.

[Chillee]

Got it. Updated the issue in response.

2sat.cpp has some other issues. Namely, it's unclear to me what the tests are meant to output. For main1, 0 and 1 respectively. For main2, I can reason about the correct output. For main3, it looks like it should be outputting something ~95/100?

[simonlindholm]

Right. You can comment out the printouts, those were mainly just for some intuition about how random 2-sat behaves. It's still asserting correctness.