Samuel Freilich

Well, full unified diff might be excessive, but you have splitlines() and could probably manage _some_ sort of multi-line diff.

This seems to be fixed at head.

Ah, wait, nevermind, that was with TZ=UTC. I'm still confused why the test expects datetime.date.

Interesting, the thread in https://github.com/python/cpython/issues/59648 suggests patches that add a `fromnanoseconds` method (though I think if going with that design it could use `tonanoseconds` as well). The thing that got...

@patrickmeehan Probably not going to poke at this one myself, but I want to point out that I added tests to this library recently, and it would be pretty easy...

I don't remember the context exactly, but I think I was trying to optimize some low-latency graphics code that tries hard to avoid unnecessary allocations and wound up taking a...

I'll focus on that first. Skipping zero-length edges seems like it would make sense, but is there some strong expectation that every input vertex is used?

Removing that one zero-length edge doesn't seem to fix the crash / assert failure here, though. I was able to cut the repro down to: ``` {0.f, -1.f}, {1.f, -3.f},...

I'm trying to better understand the computing of winding numbers in AddRightEdges and getting a bit stuck: https://github.com/memononen/libtess2/blob/9a450cc9e5b4b79c36b89648f8b92fe65b6dd407/Source/sweep.c#L328 ``` { // Initialization ... int firstTime = TRUE; // Loop until...

Another interesting thing about this example is that it has origin coordinates with different signs for the 0. If the 0, 0 coordinate is removed (leaving the pair of (-0,...