Oleg Grenrus
Oleg Grenrus
I conjecture that mentioning Nix only adds confusion. Almost all people I know use `v2-build` (well actually they use `new-build`, but maybe `v2-build` is a bit better). E.g. - One...
This is wrong, please don't.  - https://hackage.haskell.org/package/puresat - https://hackage.haskell.org/package/haskell-ci - https://hackage.haskell.org/package/attoparsec
`KeyMap` (object) is strict values. I don't see a reason why array values couldn't as well. This will be a major breaking change though, comparable to `aeson-2.0.0.0` breakage.
It depends on `Stream`, which specifically may become a liability ``` [1 of 1] Compiling Data.Stream ( Data/Stream.hs, dist/build/Data/Stream.o, dist/build/Data/Stream.dyn_o ) Data/Stream.hs:105:3: warning: [GHC-22705] [-Wnoncanonical-monad-instances] Noncanonical ‘return’ definition detected in...
With https://github.com/well-typed/large-records/pull/185 `ghc-tcplugin-api`'s `splitTyConApp_upTo` gives us list of coercions used to elaborate applications. We should thread those through and pass to `mkPluginUnivCo` & `mkPluginUnivEvTerm`. This might be easier to do...
Currently the parsing code in `large-anon` (i.e. code which elaborates constraints/givens etc.) uses `Maybe` as it's main monad. This means that we either get `Just result` (good) or machine says...
I understand what `splitTyConApp_maybe :: HasDebugCallStack => Type -> Maybe (TyCon, [Type])` does. It's a reverse of `mkTyConApp :: TyCon -> [Type] -> Type`. I don't really understand what `splitTyConApp_upTo`...
But maybe some others too.
It's not uncommon when people make new tests by copying old ones and forget to update names, resulting in test-suites with duplicate names. This makes `-p` not being able to...