TST: add one more python release to xfail list by bsipocz · Pull Request #266 · scientific-python/pytest-doctestplus
Yes, I was thinking about making it more robust, but it doesn't fail with all 3.11 releases (see my comment about not being able to reproduce locally, I was using 3.11.7).
So I would postpone a version wildcard hackary until it pops up again. The numpy ufunc test already rang a bell when I first saw the tracelog so it won't be too much of a rabbit hole to go into next time either).