Matthías Páll Gissurarson

I removed the unneccesary ones, per @ClearlyClaire. I still think this is the way to go, to render the domain when it's rendered. But I can change it so that...

Alright, much simpler now! (sorry for the all the pushes, my git got into some weird state there for a second).

The merge request has been approved, so now we only need to wait for 9.0.2 and it's release! https://gitlab.haskell.org/ghc/ghc/-/merge_requests/5784

Well, we compile code with the same version we target. But it works with 8.8 and 8.10 I think, and then 9.0.2, but not 9.0.1 haha

True, but if I see a "1 success" now and `stop_on_results` is false, it keeps going. If I then do Ctrl-C, there's no way for me to see what that...

Finding fixes for these disjoint sets could also be done by a different type of search, maybe exhaustive or random. Then we could merge those solutions together with genetic search.

No, we can infer the disjoint sets ourselves, so the user doesn't do anything. But the search itself would be doing a multi-objective search which is (hopefully) an easier problem...

Seems to work in the CI @lapplislazuli, are you doing something different?

@lapplislazuli note: the version of `check-helpers` was bumped, so you'll need to upgrade.

That's weird, since the bug you show is exactly what would happen if `check-helpers` was outdated. Did you delete: + the cabal store `~/.cabal` + the ghc env `~/.ghc` +...