Closed HDembinski closed 1 year ago
I noticed that this particular test trips if the output is reference histogram is generated on an M1 Mac. For a previous PR I regenerated it on the thread ripper and then it passed on both. I'll try it and if the tests pass merge before the impy->chromo PR since most changes here are for binaries.
No, don't bother. It is not practical if I cannot generate the reference myself. I think we should rather skip that test.
Besides, I think we should generally skip test_generators on Windows, where it takes much longer than on Macos and Linux.
Ok, agree.
Regarding windows, it seems to be related to hepmc. These tests take much longer than on *nix.
I reduced the binning of the histograms used by test_generators in an attempt to make the test more robust against the issue we observed in #112
Unfortunately, it does not work. We should just skip the affected UrQMD test.