#180 highlights that GHC still has a lot of existing baggage from how it's maintained upstream.
In addition, it's very difficult to actually "package" GHC without introducing a few dozen hackage packages, and that complexity is tackled in https://github.com/ekala-project/haskell-pkgs.
I say all this because I'm starting to be more open to just moving all haskell related stuff; including the ghc binary dists, into the haskell overlay. Nothing in the repo is a haskell package because of the bootstrapping issue. Other than shellcheck, there's not really a haskell utilty which is central to development, building software, or deployments.
#180 highlights that GHC still has a lot of existing baggage from how it's maintained upstream.
In addition, it's very difficult to actually "package" GHC without introducing a few dozen hackage packages, and that complexity is tackled in https://github.com/ekala-project/haskell-pkgs.
I say all this because I'm starting to be more open to just moving all haskell related stuff; including the ghc binary dists, into the haskell overlay. Nothing in the repo is a haskell package because of the bootstrapping issue. Other than
shellcheck, there's not really a haskell utilty which is central to development, building software, or deployments.