Closed olexandr-konovalov closed 5 years ago
I have suppressed this output for now. It is hard to believe that this can be triggered by merely a compiler change (NaturalHomomorphismByNormalSubgroup now expresses its result with a different, redundant set of generators).
When all packages are loaded, there are diffs in the tests: see e.g. https://travis-ci.org/gap-system/gap-docker-pkg-tests-stable-4.9/jobs/392042838
If the output is mathematically correct, and the diff is due to some randomness involved, perhaps it's possible to suppress the output and check some properties of the result instead?