On Mon Jul 31 16:34:03 2023 +0000, Fabian Maurer wrote:
I mean, we could also keep it around, although I personally don't see the point in a test that doesn't test what was made to test.
Nothing wrong the reasoning, but there is more information now. I agree that the AddPathArc version of the test may be removed. It appears to have no value given the more precise test. And as for the cause of the broken test this MR is addressing, that is better handled by the new test for zero-width/zero-height.