[GHC] #16209: GHC behaves differently when binding has separate or inlined type declaration

#16209: GHC behaves differently when binding has separate or inlined type declaration -------------------------------------+------------------------------------- Reporter: fghibellini | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.10.3 Keywords: | Operating System: Unknown/Multiple Architecture: | Type of failure: GHC rejects Unknown/Multiple | valid program Test Case: | Blocked By: Blocking: | Related Tickets: Differential Rev(s): | Wiki Page: -------------------------------------+------------------------------------- {{{#!hs {-# LANGUAGE RankNTypes #-} {-# LANGUAGE ScopedTypeVariables #-} main = do let a :: forall t. Functor t => t () = undefined let g :: forall u. Functor u => u () = a putStrLn "success" }}} fails with: {{{ source_file.hs:6:44: Couldn't match expected type ‘forall (u :: * -> *). Functor u => u ()’ with actual type ‘t0 ()’ When checking that: t0 () is more polymorphic than: forall (u :: * -> *). Functor u => u () In the expression: a }}} but the following compiles successfully {{{#!hs {-# LANGUAGE RankNTypes #-} {-# LANGUAGE ScopedTypeVariables #-} main = do let a :: forall t. Functor t => t () = undefined let g :: forall u. Functor u => u () g = a putStrLn "success" }}} On GHC 7.10 it works as described above. On newer ones it fails even on the "functioning code". {{{ a.hs:5:44: error: • Cannot instantiate unification variable ‘a0’ with a type involving foralls: forall (t :: * -> *). Functor t => t () GHC doesn't yet support impredicative polymorphism • In the expression: undefined In a pattern binding: a :: forall (t :: * -> *). Functor t => t () = undefined In the expression: do let a :: forall t. Functor t => t () = undefined let g :: forall u. Functor u => u () g = a putStrLn "success" }}} -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/16209 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

#16209: GHC behaves differently when binding has separate or inlined type declaration -------------------------------------+------------------------------------- Reporter: fghibellini | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.10.3 Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: Type of failure: GHC rejects | Unknown/Multiple valid program | Test Case: Blocked By: | Blocking: Related Tickets: | Differential Rev(s): Wiki Page: | -------------------------------------+------------------------------------- Description changed by fghibellini: Old description:
{{{#!hs {-# LANGUAGE RankNTypes #-} {-# LANGUAGE ScopedTypeVariables #-}
main = do let a :: forall t. Functor t => t () = undefined let g :: forall u. Functor u => u () = a putStrLn "success" }}}
fails with:
{{{ source_file.hs:6:44: Couldn't match expected type ‘forall (u :: * -> *). Functor u => u ()’ with actual type ‘t0 ()’ When checking that: t0 () is more polymorphic than: forall (u :: * -> *). Functor u => u () In the expression: a }}}
but the following compiles successfully
{{{#!hs {-# LANGUAGE RankNTypes #-} {-# LANGUAGE ScopedTypeVariables #-}
main = do let a :: forall t. Functor t => t () = undefined let g :: forall u. Functor u => u () g = a putStrLn "success" }}}
On GHC 7.10 it works as described above. On newer ones it fails even on the "functioning code".
{{{ a.hs:5:44: error: • Cannot instantiate unification variable ‘a0’ with a type involving foralls: forall (t :: * -> *). Functor t => t () GHC doesn't yet support impredicative polymorphism • In the expression: undefined In a pattern binding: a :: forall (t :: * -> *). Functor t => t () = undefined In the expression: do let a :: forall t. Functor t => t () = undefined let g :: forall u. Functor u => u () g = a putStrLn "success" }}}
New description: {{{#!hs {-# LANGUAGE RankNTypes #-} {-# LANGUAGE ScopedTypeVariables #-} main = do let a :: forall t. Functor t => t () = undefined let g :: forall u. Functor u => u () = a putStrLn "success" }}} fails with: {{{ source_file.hs:6:44: Couldn't match expected type ‘forall (u :: * -> *). Functor u => u ()’ with actual type ‘t0 ()’ When checking that: t0 () is more polymorphic than: forall (u :: * -> *). Functor u => u () In the expression: a }}} but the following compiles successfully {{{#!hs {-# LANGUAGE RankNTypes #-} {-# LANGUAGE ScopedTypeVariables #-} main = do let a :: forall t. Functor t => t () = undefined let g :: forall u. Functor u => u () g = a putStrLn "success" }}} On GHC 7.10 it works as described above. On 8.4.4 it fails even on the "functioning code". {{{ a.hs:5:44: error: • Cannot instantiate unification variable ‘a0’ with a type involving foralls: forall (t :: * -> *). Functor t => t () GHC doesn't yet support impredicative polymorphism • In the expression: undefined In a pattern binding: a :: forall (t :: * -> *). Functor t => t () = undefined In the expression: do let a :: forall t. Functor t => t () = undefined let g :: forall u. Functor u => u () g = a putStrLn "success" }}} -- -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/16209#comment:1 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler

#16209: GHC behaves differently when binding has separate or inlined type declaration -------------------------------------+------------------------------------- Reporter: fghibellini | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.10.3 Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: Type of failure: GHC rejects | Unknown/Multiple valid program | Test Case: Blocked By: | Blocking: Related Tickets: | Differential Rev(s): Wiki Page: | -------------------------------------+------------------------------------- Comment (by dmwit): See further discussion on [https://stackoverflow.com/q/54293485/791604 StackOverflow]. Executive summary of my thoughts there: 7.10.3 is correct to disallow the first program, and buggy because it allows the second program; meanwhile, modern GHC correctly disallows both. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/16209#comment:2 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler
participants (1)
-
GHC