bleeding-edge CI is red again, example log
This time, the "culprit" seems to be sympy 1.12.1a1
However, I'm hesitant to report this upstream: the one failing test was originally added because of a similar "regression" where newest sympy broke unpickling old pickle files. The reason why we have this test is that yt's CI relies on pickle files to store known-good test results. However, pickles were never intended to be a reliable long term storage solution, so I don't think it's fair to put the burden of maintaining this system onto sympy maintainers, and unyt is only a middle-man with no actual responsibility or power over this, so I think our best shot is to actually kill the test.
That said, I will still warn other yt maintainers about the upcoming expiration of some test answers.
bleeding-edge CI is red again, example log This time, the "culprit" seems to be sympy 1.12.1a1
However, I'm hesitant to report this upstream: the one failing test was originally added because of a similar "regression" where newest sympy broke unpickling old pickle files. The reason why we have this test is that
yt
's CI relies on pickle files to store known-good test results. However, pickles were never intended to be a reliable long term storage solution, so I don't think it's fair to put the burden of maintaining this system onto sympy maintainers, and unyt is only a middle-man with no actual responsibility or power over this, so I think our best shot is to actually kill the test.That said, I will still warn other yt maintainers about the upcoming expiration of some test answers.