Questions regarding the right strategy for upgrading to ghc-9.8

[lucasdicioccio]

Hi there, I'm attempting to increase this repository to ghc-9.8.2 with an aim for a Hackage release (so that my blog-engine can be on Hackage as well) at some point before this summer (personal reasons for the deadline). I work on a fork at https://github.com/lucasdicioccio/dhall-haskell .

My patch is not ready yet but an early attempt shows:

• a few bounds could be/would have to be relaxed
• some deeper failures will need bound-relaxing/uploads as well (hnix for dhall-lsp-server, dhall-nix and dhall-nixpkgs)
• some CPP will be needed to support (without warning) old imports (liftA2 being part of the class Control.Applicative)
• some rework around (safe? -- i haven't convinced myself) calls to Prelude.head which turns a deprecation warning
• my auto-formatting vim autocommand to fourmolu-format on save trapped me
• some more CPP needed as megaparsec now has an additional packages

I don't mind continuing, however I'm interested to know whether:

• are you aware of someone already on a further track than myself?
• is there some target-compatibility level with older GHCs/megaparsec? (to cut on CPP)
• is there some plan to fourmolu-format the code (in which case I'd rather see that done in a lockstep before xor after porting to 9.8.2 to reduce diff noise/rebase conflicts)?