Build failure with GHC 9.4.1

[mitchellwrosen]

src/Language/Haskell/GHC/ExactPrint/Orphans.hs:55:18: error:
    Not in scope: type constructor or class ‘AnnsLet’
   |
55 | instance Default AnnsLet where
   |                  ^^^^^^^

src/Language/Haskell/GHC/ExactPrint/Lookup.hs:19:3: warning: [-Wincomplete-patterns]
    Pattern match(es) are non-exhaustive
    In a case alternative:
        Patterns of type ‘AnnKeywordId’ not matched: AnnCases
   |
19 |   case kw of
   |   ^^^^^^^^^^...

[ndmitchell]

This is also blocking HLint from moving forward with GHC 9.4.1 - tracked at https://github.com/ndmitchell/hlint/issues/1413

[shayne-fletcher]

also affects haskell-language-server (hlint plugin, quickfixes) it seems https://github.com/haskell/haskell-language-server/issues/3146

[milloni]

Hi, I've been trying to port ghc-exactpoint to GHC 9.4.1 and I was wondering if anyone is working on this already? I don't want to duplicate the effort

[alanz]

Let me take a look

[milloni]

@alanz I posted an early draft of the work so far here, if you want to take a look

https://github.com/milloni/ghc-exactprint/tree/milloni/ghc-9.4

[alanz]

I just realised I actually did this work already in my https://github.com/alanz/ghc-exactprint/tree/ghc-9.4 branch.

I just spent some time getting the repo back into a state of having multiple source trees with common areas, in https://github.com/alanz/ghc-exactprint/tree/ghc-9.4.1

But my current ghc-9.4 branch can easily be compared to the GHC source tree utils/check-exact directory, and that comparison gets hard if there is a lot of CPP floating around.

So I will switch back to polishing my ghc-9.4 branch, and when it is published to hackage it will support only GHC 9.4.x, much lke apply-refact does now.

[andreasabel]

Issue fixed? Close?