GHC issueshttps://gitlab.haskell.org/ghc/ghc//issues20200123T19:27:40Zhttps://gitlab.haskell.org/ghc/ghc//issues/14319Stuck type families can lead to lousy error messages20200123T19:27:40ZDavid FeuerStuck type families can lead to lousy error messagesI first noticed this problem at the type level:
```hs
{# language TypeFamilies, TypeInType, ScopedTypeVariables #}
module ArityError where
import Data.Kind
import GHC.TypeLits
import Data.Proxy
type family F (s :: Symbol) :: Type
type family G (s :: Symbol) :: F s
type instance G "Hi" = Maybe
```
This produces the error message
```hs
ArityError.hs:10:24: error:
• Expecting one more argument to ‘Maybe’
Expected kind ‘F "Hi"’, but ‘Maybe’ has kind ‘* > *’
• In the type ‘Maybe’
In the type instance declaration for ‘G’

10  type instance G "Hi" = Maybe
 ^^^^^
```
This looks utterly bogus: `F "Hi"` is stuck, so we have no idea what arity it indicates.

I just realized we have a similar problem at the term level:
```hs
f :: forall (s :: Symbol). Proxy s > F s
f _ _ = undefined
```
produces
```hs
ArityError.hs:14:1: error:
• Couldn't match expected type ‘F s’ with actual type ‘p0 > a0’
The type variables ‘p0’, ‘a0’ are ambiguous
• The equation(s) for ‘f’ have two arguments,
but its type ‘Proxy s > F s’ has only one
• Relevant bindings include
f :: Proxy s > F s (bound at ArityError.hs:14:1)

14  f _ _ = undefined
 ^^^^^^^^^^^^^^^^^
```
The claim that `Proxy s > F s` has only one argument is bogus; we only know that it has *at least* one argument. The fix (I imagine) is to refrain from reporting arity errors when we don't know enough about the relevant arities.I first noticed this problem at the type level:
```hs
{# language TypeFamilies, TypeInType, ScopedTypeVariables #}
module ArityError where
import Data.Kind
import GHC.TypeLits
import Data.Proxy
type family F (s :: Symbol) :: Type
type family G (s :: Symbol) :: F s
type instance G "Hi" = Maybe
```
This produces the error message
```hs
ArityError.hs:10:24: error:
• Expecting one more argument to ‘Maybe’
Expected kind ‘F "Hi"’, but ‘Maybe’ has kind ‘* > *’
• In the type ‘Maybe’
In the type instance declaration for ‘G’

10  type instance G "Hi" = Maybe
 ^^^^^
```
This looks utterly bogus: `F "Hi"` is stuck, so we have no idea what arity it indicates.

I just realized we have a similar problem at the term level:
```hs
f :: forall (s :: Symbol). Proxy s > F s
f _ _ = undefined
```
produces
```hs
ArityError.hs:14:1: error:
• Couldn't match expected type ‘F s’ with actual type ‘p0 > a0’
The type variables ‘p0’, ‘a0’ are ambiguous
• The equation(s) for ‘f’ have two arguments,
but its type ‘Proxy s > F s’ has only one
• Relevant bindings include
f :: Proxy s > F s (bound at ArityError.hs:14:1)

14  f _ _ = undefined
 ^^^^^^^^^^^^^^^^^
```
The claim that `Proxy s > F s` has only one argument is bogus; we only know that it has *at least* one argument. The fix (I imagine) is to refrain from reporting arity errors when we don't know enough about the relevant arities.https://gitlab.haskell.org/ghc/ghc//issues/14198Inconsistent treatment of implicitly bound kind variables as freefloating20191211T15:53:21ZRyan ScottInconsistent treatment of implicitly bound kind variables as freefloating(Spun off from the discussion starting at https://phabricator.haskell.org/D3872#109927.)
This program is accepted:
```hs
{# LANGUAGE ExistentialQuantification #}
{# LANGUAGE PolyKinds #}
{# LANGUAGE RankNTypes #}
import Data.Proxy
data Foo = MkFoo (forall a. Proxy a)
```
There's something interesting going on here, however. Because `PolyKinds` is enabled, the kind of `a` is generalized to `k`. But where does `k` get quantified? It turns out that it's implicitly quantified as an existential type variable to `MkFoo`:
```
λ> :i Foo
data Foo = forall k. MkFoo (forall (a :: k). Proxy a)
```
This was brought up some time ago in #7873, where the conclusion was to keep this behavior. But it's strange becuase the `k` is existential, so the definition is probably unusable.
But to make things stranger, it you write out the kind of `a` explicitly:
```hs
{# LANGUAGE ExistentialQuantification #}
{# LANGUAGE PolyKinds #}
{# LANGUAGE RankNTypes #}
import Data.Proxy
data Foo2 = MkFoo2 (forall (a :: k). Proxy a)
```
Then GHC will reject it:
```
Bug.hs:7:1: error:
Kind variable ‘k’ is implicitly bound in data type
‘Foo2’, but does not appear as the kind of any
of its type variables. Perhaps you meant
to bind it (with TypeInType) explicitly somewhere?

7  data Foo2 = MkFoo2 (forall (a :: k). Proxy a)
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
```
So GHC's treatment is inconsistent. What should GHC do? We could:
1. Have both be an error.
2. Have both be accepted, and implicitly quantify `k` as an existential type variable
3. Have both be accepted, and implicitly quantify `k` in the `forall` itself. That is:
```hs
MkFoo :: (forall {k} (a :: k). Proxy k a) > Foo
```
4. Something else. When you try a similar trick with type synonyms:
```hs
{# LANGUAGE ExistentialQuantification #}
{# LANGUAGE PolyKinds #}
{# LANGUAGE RankNTypes #}
import Data.Proxy
type Foo3 = forall a. Proxy a
```
Instead of generalizing the kind of `a`, its kind will default to `Any`:
```
λ> :i Foo3
type Foo3 = forall (a :: GHC.Types.Any). Proxy a
```
Would that be an appropriate trick for data types as well?(Spun off from the discussion starting at https://phabricator.haskell.org/D3872#109927.)
This program is accepted:
```hs
{# LANGUAGE ExistentialQuantification #}
{# LANGUAGE PolyKinds #}
{# LANGUAGE RankNTypes #}
import Data.Proxy
data Foo = MkFoo (forall a. Proxy a)
```
There's something interesting going on here, however. Because `PolyKinds` is enabled, the kind of `a` is generalized to `k`. But where does `k` get quantified? It turns out that it's implicitly quantified as an existential type variable to `MkFoo`:
```
λ> :i Foo
data Foo = forall k. MkFoo (forall (a :: k). Proxy a)
```
This was brought up some time ago in #7873, where the conclusion was to keep this behavior. But it's strange becuase the `k` is existential, so the definition is probably unusable.
But to make things stranger, it you write out the kind of `a` explicitly:
```hs
{# LANGUAGE ExistentialQuantification #}
{# LANGUAGE PolyKinds #}
{# LANGUAGE RankNTypes #}
import Data.Proxy
data Foo2 = MkFoo2 (forall (a :: k). Proxy a)
```
Then GHC will reject it:
```
Bug.hs:7:1: error:
Kind variable ‘k’ is implicitly bound in data type
‘Foo2’, but does not appear as the kind of any
of its type variables. Perhaps you meant
to bind it (with TypeInType) explicitly somewhere?

7  data Foo2 = MkFoo2 (forall (a :: k). Proxy a)
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
```
So GHC's treatment is inconsistent. What should GHC do? We could:
1. Have both be an error.
2. Have both be accepted, and implicitly quantify `k` as an existential type variable
3. Have both be accepted, and implicitly quantify `k` in the `forall` itself. That is:
```hs
MkFoo :: (forall {k} (a :: k). Proxy k a) > Foo
```
4. Something else. When you try a similar trick with type synonyms:
```hs
{# LANGUAGE ExistentialQuantification #}
{# LANGUAGE PolyKinds #}
{# LANGUAGE RankNTypes #}
import Data.Proxy
type Foo3 = forall a. Proxy a
```
Instead of generalizing the kind of `a`, its kind will default to `Any`:
```
λ> :i Foo3
type Foo3 = forall (a :: GHC.Types.Any). Proxy a
```
Would that be an appropriate trick for data types as well?https://gitlab.haskell.org/ghc/ghc//issues/14180Permit levity polymorphism in kinds20200916T18:18:18ZDavid FeuerPermit levity polymorphism in kinds```hs
{# language TypeInType, TypeFamilies, MagicHash #}
import GHC.Exts
type family MatchInt (f :: Int) :: () where
MatchInt ('I# x) = '()
```
produces
```
<interactive>:2:59: error:
• Couldn't match a lifted type with an unlifted type
When matching kinds
k0 :: *
Int# :: TYPE 'IntRep
Expected kind ‘Int#’, but ‘x’ has kind ‘k0’
• In the first argument of ‘ 'I#’, namely ‘x’
In the first argument of ‘MatchInt’, namely ‘( 'I# x)’
In the type family declaration for ‘MatchInt’
```
If, however, I replace `x` in the pattern with `_`, the type checker is satisfied. If I give it a pattern signature,
```hs
MatchInt ('I# (x :: Int#)) = '()
```
I get a different (and equally unhelpful) error message,
```
• Expecting a lifted type, but ‘Int#’ is unlifted
• In the kind ‘Int#’
In the first argument of ‘ 'I#’, namely ‘(x :: Int#)’
In the first argument of ‘MatchInt’, namely ‘( 'I# (x :: Int#))’
```
<details><summary>Trac metadata</summary>
 Trac field  Value 
    
 Version  8.3 
 Type  Bug 
 TypeOfFailure  OtherFailure 
 Priority  normal 
 Resolution  Unresolved 
 Component  Compiler (Type checker) 
 Test case  
 Differential revisions  
 BlockedBy  
 Related  
 Blocking  
 CC  
 Operating system  
 Architecture  
</details>
<! {"blocked_by":[],"summary":"Strange/bad error message binding unboxed type variable","status":"New","operating_system":"","component":"Compiler (Type checker)","related":[],"milestone":"8.2.2","resolution":"Unresolved","owner":{"tag":"Unowned"},"version":"8.3","keywords":["TypeInType"],"differentials":[],"test_case":"","architecture":"","cc":[""],"type":"Bug","description":"{{{#!hs\r\n{# language TypeInType, TypeFamilies, MagicHash #}\r\n\r\nimport GHC.Exts\r\n\r\ntype family MatchInt (f :: Int) :: () where\r\n MatchInt ('I# x) = '()\r\n}}}\r\n\r\nproduces\r\n\r\n{{{\r\n<interactive>:2:59: error:\r\n • Couldn't match a lifted type with an unlifted type\r\n When matching kinds\r\n k0 :: *\r\n Int# :: TYPE 'IntRep\r\n Expected kind ‘Int#’, but ‘x’ has kind ‘k0’\r\n • In the first argument of ‘ 'I#’, namely ‘x’\r\n In the first argument of ‘MatchInt’, namely ‘( 'I# x)’\r\n In the type family declaration for ‘MatchInt’\r\n}}}\r\n\r\nIf, however, I replace `x` in the pattern with `_`, the type checker is satisfied. If I give it a pattern signature,\r\n\r\n{{{#!hs\r\nMatchInt ('I# (x :: Int#)) = '()\r\n}}}\r\n\r\nI get a different (and equally unhelpful) error message,\r\n\r\n{{{\r\n • Expecting a lifted type, but ‘Int#’ is unlifted\r\n • In the kind ‘Int#’\r\n In the first argument of ‘ 'I#’, namely ‘(x :: Int#)’\r\n In the first argument of ‘MatchInt’, namely ‘( 'I# (x :: Int#))’\r\n}}}","type_of_failure":"OtherFailure","blocking":[]} >```hs
{# language TypeInType, TypeFamilies, MagicHash #}
import GHC.Exts
type family MatchInt (f :: Int) :: () where
MatchInt ('I# x) = '()
```
produces
```
<interactive>:2:59: error:
• Couldn't match a lifted type with an unlifted type
When matching kinds
k0 :: *
Int# :: TYPE 'IntRep
Expected kind ‘Int#’, but ‘x’ has kind ‘k0’
• In the first argument of ‘ 'I#’, namely ‘x’
In the first argument of ‘MatchInt’, namely ‘( 'I# x)’
In the type family declaration for ‘MatchInt’
```
If, however, I replace `x` in the pattern with `_`, the type checker is satisfied. If I give it a pattern signature,
```hs
MatchInt ('I# (x :: Int#)) = '()
```
I get a different (and equally unhelpful) error message,
```
• Expecting a lifted type, but ‘Int#’ is unlifted
• In the kind ‘Int#’
In the first argument of ‘ 'I#’, namely ‘(x :: Int#)’
In the first argument of ‘MatchInt’, namely ‘( 'I# (x :: Int#))’
```
<details><summary>Trac metadata</summary>
 Trac field  Value 
    
 Version  8.3 
 Type  Bug 
 TypeOfFailure  OtherFailure 
 Priority  normal 
 Resolution  Unresolved 
 Component  Compiler (Type checker) 
 Test case  
 Differential revisions  
 BlockedBy  
 Related  
 Blocking  
 CC  
 Operating system  
 Architecture  
</details>
<! {"blocked_by":[],"summary":"Strange/bad error message binding unboxed type variable","status":"New","operating_system":"","component":"Compiler (Type checker)","related":[],"milestone":"8.2.2","resolution":"Unresolved","owner":{"tag":"Unowned"},"version":"8.3","keywords":["TypeInType"],"differentials":[],"test_case":"","architecture":"","cc":[""],"type":"Bug","description":"{{{#!hs\r\n{# language TypeInType, TypeFamilies, MagicHash #}\r\n\r\nimport GHC.Exts\r\n\r\ntype family MatchInt (f :: Int) :: () where\r\n MatchInt ('I# x) = '()\r\n}}}\r\n\r\nproduces\r\n\r\n{{{\r\n<interactive>:2:59: error:\r\n • Couldn't match a lifted type with an unlifted type\r\n When matching kinds\r\n k0 :: *\r\n Int# :: TYPE 'IntRep\r\n Expected kind ‘Int#’, but ‘x’ has kind ‘k0’\r\n • In the first argument of ‘ 'I#’, namely ‘x’\r\n In the first argument of ‘MatchInt’, namely ‘( 'I# x)’\r\n In the type family declaration for ‘MatchInt’\r\n}}}\r\n\r\nIf, however, I replace `x` in the pattern with `_`, the type checker is satisfied. If I give it a pattern signature,\r\n\r\n{{{#!hs\r\nMatchInt ('I# (x :: Int#)) = '()\r\n}}}\r\n\r\nI get a different (and equally unhelpful) error message,\r\n\r\n{{{\r\n • Expecting a lifted type, but ‘Int#’ is unlifted\r\n • In the kind ‘Int#’\r\n In the first argument of ‘ 'I#’, namely ‘(x :: Int#)’\r\n In the first argument of ‘MatchInt’, namely ‘( 'I# (x :: Int#))’\r\n}}}","type_of_failure":"OtherFailure","blocking":[]} >https://gitlab.haskell.org/ghc/ghc//issues/14155GHC mentions unlifted types out of the blue (to me anyway)20200123T19:31:01ZIcelandjackGHC mentions unlifted types out of the blue (to me anyway)This one does me 'ead in, I accidentally type `Ran` instead of `Swap`
```hs
{# Language RankNTypes, DerivingStrategies, TypeApplications, ScopedTypeVariables, GADTs, GeneralizedNewtypeDeriving, InstanceSigs, PolyKinds #}
import Data.Coerce
newtype Ran g h a = Ran (forall b. (a > g b) > h b)
newtype Swap p f g a where
Swap :: p g f a > Swap p f g a
deriving newtype
Show
class IxPointed m where
ireturn :: a > m i i a
instance IxPointed f => IxPointed (Swap f) where
ireturn :: forall a i. a > Swap f i i a
ireturn a = Ran (ireturn a)
```
and get this error
```
$ ghci ignoredotghci /tmp/bug.hs
GHCi, version 8.3.20170605: http://www.haskell.org/ghc/ :? for help
[1 of 1] Compiling Main ( /tmp/bug.hs, interpreted )
/tmp/bug.hs:17:15: error:
• Couldn't match expected type ‘Swap f i i a’
with actual type ‘Ran g0 h0 a0’
• In the expression: Ran (ireturn a)
In an equation for ‘ireturn’: ireturn a = Ran (ireturn a)
In the instance declaration for ‘IxPointed (Swap f)’
• Relevant bindings include
a :: a (bound at /tmp/bug.hs:17:11)
ireturn :: a > Swap f i i a (bound at /tmp/bug.hs:17:3)

17  ireturn a = Ran (ireturn a)
 ^^^^^^^^^^^^^^^
/tmp/bug.hs:17:20: error:
• Couldn't match a lifted type with an unlifted type
Expected type: (a0 > g0 b) > h0 b
Actual type: (>) (a0 > g0 b) (a0 > g0 b) a
• In the first argument of ‘Ran’, namely ‘(ireturn a)’
In the expression: Ran (ireturn a)
In an equation for ‘ireturn’: ireturn a = Ran (ireturn a)

17  ireturn a = Ran (ireturn a)
 ^^^^^^^^^
Failed, modules loaded: none.
```
Is GHC right to bring up unlifted types? I would guess this is due to the newly added levity polymorphism of `(>) :: TYPE rep > TYPE rep' > Type`
<details><summary>Trac metadata</summary>
 Trac field  Value 
    
 Version  8.2.1 
 Type  Bug 
 TypeOfFailure  OtherFailure 
 Priority  normal 
 Resolution  Unresolved 
 Component  Compiler 
 Test case  
 Differential revisions  
 BlockedBy  
 Related  
 Blocking  
 CC  
 Operating system  
 Architecture  
</details>
<! {"blocked_by":[],"summary":"GHC mentions unlifted types out of the blue (to me anyway)","status":"New","operating_system":"","component":"Compiler","related":[],"milestone":"","resolution":"Unresolved","owner":{"tag":"Unowned"},"version":"8.2.1","keywords":[],"differentials":[],"test_case":"","architecture":"","cc":[""],"type":"Bug","description":"This one does me 'ead in, I accidentally type `Ran` instead of `Swap` \r\n\r\n{{{#!hs\r\n{# Language RankNTypes, DerivingStrategies, TypeApplications, ScopedTypeVariables, GADTs, GeneralizedNewtypeDeriving, InstanceSigs, PolyKinds #}\r\n\r\nimport Data.Coerce\r\n\r\nnewtype Ran g h a = Ran (forall b. (a > g b) > h b)\r\n\r\nnewtype Swap p f g a where\r\n Swap :: p g f a > Swap p f g a\r\n deriving newtype\r\n Show\r\n\r\nclass IxPointed m where\r\n ireturn :: a > m i i a\r\n\r\ninstance IxPointed f => IxPointed (Swap f) where\r\n ireturn :: forall a i. a > Swap f i i a\r\n ireturn a = Ran (ireturn a)\r\n}}}\r\n\r\nand get this error\r\n\r\n{{{\r\n$ ghci ignoredotghci /tmp/bug.hs\r\nGHCi, version 8.3.20170605: http://www.haskell.org/ghc/ :? for help\r\n[1 of 1] Compiling Main ( /tmp/bug.hs, interpreted )\r\n\r\n/tmp/bug.hs:17:15: error:\r\n • Couldn't match expected type ‘Swap f i i a’\r\n with actual type ‘Ran g0 h0 a0’\r\n • In the expression: Ran (ireturn a)\r\n In an equation for ‘ireturn’: ireturn a = Ran (ireturn a)\r\n In the instance declaration for ‘IxPointed (Swap f)’\r\n • Relevant bindings include\r\n a :: a (bound at /tmp/bug.hs:17:11)\r\n ireturn :: a > Swap f i i a (bound at /tmp/bug.hs:17:3)\r\n \r\n17  ireturn a = Ran (ireturn a)\r\n  ^^^^^^^^^^^^^^^\r\n\r\n/tmp/bug.hs:17:20: error:\r\n • Couldn't match a lifted type with an unlifted type\r\n Expected type: (a0 > g0 b) > h0 b\r\n Actual type: (>) (a0 > g0 b) (a0 > g0 b) a\r\n • In the first argument of ‘Ran’, namely ‘(ireturn a)’\r\n In the expression: Ran (ireturn a)\r\n In an equation for ‘ireturn’: ireturn a = Ran (ireturn a)\r\n \r\n17  ireturn a = Ran (ireturn a)\r\n  ^^^^^^^^^\r\nFailed, modules loaded: none.\r\n}}}\r\n\r\nIs GHC right to bring up unlifted types? I would guess this is due to the newly added levity polymorphism of `(>) :: TYPE rep > TYPE rep' > Type`","type_of_failure":"OtherFailure","blocking":[]} >This one does me 'ead in, I accidentally type `Ran` instead of `Swap`
```hs
{# Language RankNTypes, DerivingStrategies, TypeApplications, ScopedTypeVariables, GADTs, GeneralizedNewtypeDeriving, InstanceSigs, PolyKinds #}
import Data.Coerce
newtype Ran g h a = Ran (forall b. (a > g b) > h b)
newtype Swap p f g a where
Swap :: p g f a > Swap p f g a
deriving newtype
Show
class IxPointed m where
ireturn :: a > m i i a
instance IxPointed f => IxPointed (Swap f) where
ireturn :: forall a i. a > Swap f i i a
ireturn a = Ran (ireturn a)
```
and get this error
```
$ ghci ignoredotghci /tmp/bug.hs
GHCi, version 8.3.20170605: http://www.haskell.org/ghc/ :? for help
[1 of 1] Compiling Main ( /tmp/bug.hs, interpreted )
/tmp/bug.hs:17:15: error:
• Couldn't match expected type ‘Swap f i i a’
with actual type ‘Ran g0 h0 a0’
• In the expression: Ran (ireturn a)
In an equation for ‘ireturn’: ireturn a = Ran (ireturn a)
In the instance declaration for ‘IxPointed (Swap f)’
• Relevant bindings include
a :: a (bound at /tmp/bug.hs:17:11)
ireturn :: a > Swap f i i a (bound at /tmp/bug.hs:17:3)

17  ireturn a = Ran (ireturn a)
 ^^^^^^^^^^^^^^^
/tmp/bug.hs:17:20: error:
• Couldn't match a lifted type with an unlifted type
Expected type: (a0 > g0 b) > h0 b
Actual type: (>) (a0 > g0 b) (a0 > g0 b) a
• In the first argument of ‘Ran’, namely ‘(ireturn a)’
In the expression: Ran (ireturn a)
In an equation for ‘ireturn’: ireturn a = Ran (ireturn a)

17  ireturn a = Ran (ireturn a)
 ^^^^^^^^^
Failed, modules loaded: none.
```
Is GHC right to bring up unlifted types? I would guess this is due to the newly added levity polymorphism of `(>) :: TYPE rep > TYPE rep' > Type`
<details><summary>Trac metadata</summary>
 Trac field  Value 
    
 Version  8.2.1 
 Type  Bug 
 TypeOfFailure  OtherFailure 
 Priority  normal 
 Resolution  Unresolved 
 Component  Compiler 
 Test case  
 Differential revisions  
 BlockedBy  
 Related  
 Blocking  
 CC  
 Operating system  
 Architecture  
</details>
<! {"blocked_by":[],"summary":"GHC mentions unlifted types out of the blue (to me anyway)","status":"New","operating_system":"","component":"Compiler","related":[],"milestone":"","resolution":"Unresolved","owner":{"tag":"Unowned"},"version":"8.2.1","keywords":[],"differentials":[],"test_case":"","architecture":"","cc":[""],"type":"Bug","description":"This one does me 'ead in, I accidentally type `Ran` instead of `Swap` \r\n\r\n{{{#!hs\r\n{# Language RankNTypes, DerivingStrategies, TypeApplications, ScopedTypeVariables, GADTs, GeneralizedNewtypeDeriving, InstanceSigs, PolyKinds #}\r\n\r\nimport Data.Coerce\r\n\r\nnewtype Ran g h a = Ran (forall b. (a > g b) > h b)\r\n\r\nnewtype Swap p f g a where\r\n Swap :: p g f a > Swap p f g a\r\n deriving newtype\r\n Show\r\n\r\nclass IxPointed m where\r\n ireturn :: a > m i i a\r\n\r\ninstance IxPointed f => IxPointed (Swap f) where\r\n ireturn :: forall a i. a > Swap f i i a\r\n ireturn a = Ran (ireturn a)\r\n}}}\r\n\r\nand get this error\r\n\r\n{{{\r\n$ ghci ignoredotghci /tmp/bug.hs\r\nGHCi, version 8.3.20170605: http://www.haskell.org/ghc/ :? for help\r\n[1 of 1] Compiling Main ( /tmp/bug.hs, interpreted )\r\n\r\n/tmp/bug.hs:17:15: error:\r\n • Couldn't match expected type ‘Swap f i i a’\r\n with actual type ‘Ran g0 h0 a0’\r\n • In the expression: Ran (ireturn a)\r\n In an equation for ‘ireturn’: ireturn a = Ran (ireturn a)\r\n In the instance declaration for ‘IxPointed (Swap f)’\r\n • Relevant bindings include\r\n a :: a (bound at /tmp/bug.hs:17:11)\r\n ireturn :: a > Swap f i i a (bound at /tmp/bug.hs:17:3)\r\n \r\n17  ireturn a = Ran (ireturn a)\r\n  ^^^^^^^^^^^^^^^\r\n\r\n/tmp/bug.hs:17:20: error:\r\n • Couldn't match a lifted type with an unlifted type\r\n Expected type: (a0 > g0 b) > h0 b\r\n Actual type: (>) (a0 > g0 b) (a0 > g0 b) a\r\n • In the first argument of ‘Ran’, namely ‘(ireturn a)’\r\n In the expression: Ran (ireturn a)\r\n In an equation for ‘ireturn’: ireturn a = Ran (ireturn a)\r\n \r\n17  ireturn a = Ran (ireturn a)\r\n  ^^^^^^^^^\r\nFailed, modules loaded: none.\r\n}}}\r\n\r\nIs GHC right to bring up unlifted types? I would guess this is due to the newly added levity polymorphism of `(>) :: TYPE rep > TYPE rep' > Type`","type_of_failure":"OtherFailure","blocking":[]} >Richard Eisenbergrae@richarde.devRichard Eisenbergrae@richarde.devhttps://gitlab.haskell.org/ghc/ghc//issues/14040Typed holes regression in GHC 8.0.2: No skolem info: z_a1sY[sk:2]20200123T19:31:02ZRyan ScottTyped holes regression in GHC 8.0.2: No skolem info: z_a1sY[sk:2](Originally spun off from #13877.)
The following program gives a somewhat decent error message in GHC 8.0.1:
```hs
{# LANGUAGE GADTs #}
{# LANGUAGE RankNTypes #}
{# LANGUAGE ScopedTypeVariables #}
{# LANGUAGE TypeApplications #}
{# LANGUAGE TypeFamilies #}
{# LANGUAGE TypeInType #}
module Bug where
import Data.Kind
data family Sing (a :: k)
data WeirdList :: Type > Type where
WeirdNil :: WeirdList a
WeirdCons :: a > WeirdList (WeirdList a) > WeirdList a
data instance Sing (z :: WeirdList a) where
SWeirdNil :: Sing WeirdNil
SWeirdCons :: Sing w > Sing wws > Sing (WeirdCons w wws)
elimWeirdList :: forall (a :: Type) (wl :: WeirdList a)
(p :: forall (x :: Type). x > WeirdList x > Type).
Sing wl
> (forall (y :: Type). p _ WeirdNil)
> (forall (z :: Type) (x :: z) (xs :: WeirdList (WeirdList z)).
Sing x > Sing xs > p _ xs
> p _ (WeirdCons x xs))
> p _ wl
elimWeirdList SWeirdNil pWeirdNil _ = pWeirdNil
elimWeirdList (SWeirdCons (x :: Sing (x :: z))
(xs :: Sing (xs :: WeirdList (WeirdList z))))
pWeirdNil pWeirdCons
= pWeirdCons @z @x @xs x xs
(elimWeirdList @(WeirdList z) @xs @p xs pWeirdNil pWeirdCons)
```
```
$ /opt/ghc/8.0.1/bin/ghci Foo.hs
GHCi, version 8.0.1: http://www.haskell.org/ghc/ :? for help
Loaded GHCi configuration from /home/rgscott/.ghci
[1 of 1] Compiling Bug ( Foo.hs, interpreted )
Foo.hs:34:8: error:
• Cannot apply expression of type ‘Sing wl
> (forall y. p x0 t3 'WeirdNil)
> (forall z (x :: z) (xs :: WeirdList (WeirdList z)).
Sing x
> Sing xs
> p (WeirdList z) t2 xs
> p z t1 ('WeirdCons x xs))
> p a t0 wl’
to a visible type argument ‘WeirdList z’
• In the sixth argument of ‘pWeirdCons’, namely
‘(elimWeirdList @(WeirdList z) @xs @p xs pWeirdNil pWeirdCons)’
In the expression:
pWeirdCons
@z
@x
@xs
x
xs
(elimWeirdList @(WeirdList z) @xs @p xs pWeirdNil pWeirdCons)
In an equation for ‘elimWeirdList’:
elimWeirdList
(SWeirdCons (x :: Sing (x :: z))
(xs :: Sing (xs :: WeirdList (WeirdList z))))
pWeirdNil
pWeirdCons
= pWeirdCons
@z
@x
@xs
x
xs
(elimWeirdList @(WeirdList z) @xs @p xs pWeirdNil pWeirdCons)
```
But in GHC 8.0.2, 8.2.1, and HEAD, it panics to varying degrees:
```
$ /opt/ghc/8.0.2/bin/ghci Foo.hs
GHCi, version 8.0.2: http://www.haskell.org/ghc/ :? for help
Loaded GHCi configuration from /home/rgscott/.ghci
[1 of 1] Compiling Bug ( Foo.hs, interpreted )
Foo.hs:24:41: error:
• Found type wildcard ‘_’ standing for ‘t0’
Where: ‘t0’ is an ambiguous type variable
‘x0’ is an ambiguous type variable
To use the inferred type, enable PartialTypeSignatures
• In the type signature:
elimWeirdList :: forall (a :: Type)
(wl :: WeirdList a)
(p :: forall (x :: Type). x > WeirdList x > Type).
Sing wl
> (forall (y :: Type). p _ WeirdNil)
> (forall (z :: Type) (x :: z) (xs :: WeirdList (WeirdList z)).
Sing x > Sing xs > p _ xs > p _ (WeirdCons x xs))
> p _ wl
• Relevant bindings include
elimWeirdList :: Sing wl
> (forall y. p x0 t0 'WeirdNil)
> (forall z (x :: z) (xs :: WeirdList (WeirdList z)).
Sing x
> Sing xs > p (WeirdList z) t1 xs > p z t2 ('WeirdCons x xs))
> p a t3 wl
(bound at Foo.hs:29:1)
Foo.hs:26:44: error:ghc: panic! (the 'impossible' happened)
(GHC version 8.0.2 for x86_64unknownlinux):
No skolem info: z_a13X[sk]
```
```
$ /opt/ghc/8.2.1/bin/ghci Foo.hs
GHCi, version 8.2.1: http://www.haskell.org/ghc/ :? for help
Loaded GHCi configuration from /home/rgscott/.ghci
[1 of 1] Compiling Bug ( Foo.hs, interpreted )
Foo.hs:21:18: error:
• The kind of variable ‘wl1’, namely ‘WeirdList a1’,
depends on variable ‘a1’ from an inner scope
Perhaps bind ‘wl1’ sometime after binding ‘a1’
• In the type signature:
elimWeirdList :: forall (a :: Type)
(wl :: WeirdList a)
(p :: forall (x :: Type). x > WeirdList x > Type).
Sing wl
> (forall (y :: Type). p _ WeirdNil)
> (forall (z :: Type) (x :: z) (xs :: WeirdList (WeirdList z)).
Sing x > Sing xs > p _ xs > p _ (WeirdCons x xs))
> p _ wl

21  elimWeirdList :: forall (a :: Type) (wl :: WeirdList a)
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^...
Foo.hs:24:41: error:
• Found type wildcard ‘_’ standing for ‘w0’
Where: ‘w0’ is an ambiguous type variable
‘x0’ is an ambiguous type variable
To use the inferred type, enable PartialTypeSignatures
• In the type signature:
elimWeirdList :: forall (a :: Type)
(wl :: WeirdList a)
(p :: forall (x :: Type). x > WeirdList x > Type).
Sing wl
> (forall (y :: Type). p _ WeirdNil)
> (forall (z :: Type) (x :: z) (xs :: WeirdList (WeirdList z)).
Sing x > Sing xs > p _ xs > p _ (WeirdCons x xs))
> p _ wl

24  > (forall (y :: Type). p _ WeirdNil)
 ^
Foo.hs:26:44: error:ghc: panic! (the 'impossible' happened)
(GHC version 8.2.1 for x86_64unknownlinux):
No skolem info:
z_a1sY[sk:2]
Call stack:
CallStack (from HasCallStack):
prettyCurrentCallStack, called at compiler/utils/Outputable.hs:1133:58 in ghc:Outputable
callStackDoc, called at compiler/utils/Outputable.hs:1137:37 in ghc:Outputable
pprPanic, called at compiler/typecheck/TcErrors.hs:2653:5 in ghc:TcErrors
```
(The error messages from HEAD, at commit 791947db6db32ef7d4772a821a0823e558e3c05b, are the same as in GHC 8.2.1.)
<details><summary>Trac metadata</summary>
 Trac field  Value 
    
 Version  8.0.2 
 Type  Bug 
 TypeOfFailure  OtherFailure 
 Priority  normal 
 Resolution  Unresolved 
 Component  Compiler (Type checker) 
 Test case  
 Differential revisions  
 BlockedBy  
 Related  #13877 
 Blocking  
 CC  
 Operating system  
 Architecture  
</details>
<! {"blocked_by":[],"summary":"Typed holes regression in GHC 8.0.2: No skolem info: z_a1sY[sk:2]","status":"New","operating_system":"","component":"Compiler (Type checker)","related":[13877],"milestone":"","resolution":"Unresolved","owner":{"tag":"Unowned"},"version":"8.0.2","keywords":["PartialTypeSignatures","TypeFamilies,","TypeInType,"],"differentials":[],"test_case":"","architecture":"","cc":[""],"type":"Bug","description":"(Originally spun off from #13877.)\r\n\r\nThe following program gives a somewhat decent error message in GHC 8.0.1:\r\n\r\n{{{#!hs\r\n{# LANGUAGE GADTs #}\r\n{# LANGUAGE RankNTypes #}\r\n{# LANGUAGE ScopedTypeVariables #}\r\n{# LANGUAGE TypeApplications #}\r\n{# LANGUAGE TypeFamilies #}\r\n{# LANGUAGE TypeInType #}\r\nmodule Bug where\r\n\r\nimport Data.Kind\r\n\r\ndata family Sing (a :: k)\r\n\r\ndata WeirdList :: Type > Type where\r\n WeirdNil :: WeirdList a\r\n WeirdCons :: a > WeirdList (WeirdList a) > WeirdList a\r\n\r\ndata instance Sing (z :: WeirdList a) where\r\n SWeirdNil :: Sing WeirdNil\r\n SWeirdCons :: Sing w > Sing wws > Sing (WeirdCons w wws)\r\n\r\nelimWeirdList :: forall (a :: Type) (wl :: WeirdList a)\r\n (p :: forall (x :: Type). x > WeirdList x > Type).\r\n Sing wl\r\n > (forall (y :: Type). p _ WeirdNil)\r\n > (forall (z :: Type) (x :: z) (xs :: WeirdList (WeirdList z)).\r\n Sing x > Sing xs > p _ xs\r\n > p _ (WeirdCons x xs))\r\n > p _ wl\r\nelimWeirdList SWeirdNil pWeirdNil _ = pWeirdNil\r\nelimWeirdList (SWeirdCons (x :: Sing (x :: z))\r\n (xs :: Sing (xs :: WeirdList (WeirdList z))))\r\n pWeirdNil pWeirdCons\r\n = pWeirdCons @z @x @xs x xs\r\n (elimWeirdList @(WeirdList z) @xs @p xs pWeirdNil pWeirdCons)\r\n}}}\r\n\r\n{{{\r\n$ /opt/ghc/8.0.1/bin/ghci Foo.hs \r\nGHCi, version 8.0.1: http://www.haskell.org/ghc/ :? for help\r\nLoaded GHCi configuration from /home/rgscott/.ghci\r\n[1 of 1] Compiling Bug ( Foo.hs, interpreted )\r\n\r\nFoo.hs:34:8: error:\r\n • Cannot apply expression of type ‘Sing wl\r\n > (forall y. p x0 t3 'WeirdNil)\r\n > (forall z (x :: z) (xs :: WeirdList (WeirdList z)).\r\n Sing x\r\n > Sing xs\r\n > p (WeirdList z) t2 xs\r\n > p z t1 ('WeirdCons x xs))\r\n > p a t0 wl’\r\n to a visible type argument ‘WeirdList z’\r\n • In the sixth argument of ‘pWeirdCons’, namely\r\n ‘(elimWeirdList @(WeirdList z) @xs @p xs pWeirdNil pWeirdCons)’\r\n In the expression:\r\n pWeirdCons\r\n @z\r\n @x\r\n @xs\r\n x\r\n xs\r\n (elimWeirdList @(WeirdList z) @xs @p xs pWeirdNil pWeirdCons)\r\n In an equation for ‘elimWeirdList’:\r\n elimWeirdList\r\n (SWeirdCons (x :: Sing (x :: z))\r\n (xs :: Sing (xs :: WeirdList (WeirdList z))))\r\n pWeirdNil\r\n pWeirdCons\r\n = pWeirdCons\r\n @z\r\n @x\r\n @xs\r\n x\r\n xs\r\n (elimWeirdList @(WeirdList z) @xs @p xs pWeirdNil pWeirdCons)\r\n}}}\r\n\r\nBut in GHC 8.0.2, 8.2.1, and HEAD, it panics to varying degrees:\r\n\r\n{{{\r\n$ /opt/ghc/8.0.2/bin/ghci Foo.hs \r\nGHCi, version 8.0.2: http://www.haskell.org/ghc/ :? for help\r\nLoaded GHCi configuration from /home/rgscott/.ghci\r\n[1 of 1] Compiling Bug ( Foo.hs, interpreted )\r\n\r\nFoo.hs:24:41: error:\r\n • Found type wildcard ‘_’ standing for ‘t0’\r\n Where: ‘t0’ is an ambiguous type variable\r\n ‘x0’ is an ambiguous type variable\r\n To use the inferred type, enable PartialTypeSignatures\r\n • In the type signature:\r\n elimWeirdList :: forall (a :: Type)\r\n (wl :: WeirdList a)\r\n (p :: forall (x :: Type). x > WeirdList x > Type).\r\n Sing wl\r\n > (forall (y :: Type). p _ WeirdNil)\r\n > (forall (z :: Type) (x :: z) (xs :: WeirdList (WeirdList z)).\r\n Sing x > Sing xs > p _ xs > p _ (WeirdCons x xs))\r\n > p _ wl\r\n • Relevant bindings include\r\n elimWeirdList :: Sing wl\r\n > (forall y. p x0 t0 'WeirdNil)\r\n > (forall z (x :: z) (xs :: WeirdList (WeirdList z)).\r\n Sing x\r\n > Sing xs > p (WeirdList z) t1 xs > p z t2 ('WeirdCons x xs))\r\n > p a t3 wl\r\n (bound at Foo.hs:29:1)\r\n\r\nFoo.hs:26:44: error:ghc: panic! (the 'impossible' happened)\r\n (GHC version 8.0.2 for x86_64unknownlinux):\r\n\tNo skolem info: z_a13X[sk]\r\n}}}\r\n\r\n{{{\r\n$ /opt/ghc/8.2.1/bin/ghci Foo.hs \r\nGHCi, version 8.2.1: http://www.haskell.org/ghc/ :? for help\r\nLoaded GHCi configuration from /home/rgscott/.ghci\r\n[1 of 1] Compiling Bug ( Foo.hs, interpreted )\r\n\r\nFoo.hs:21:18: error:\r\n • The kind of variable ‘wl1’, namely ‘WeirdList a1’,\r\n depends on variable ‘a1’ from an inner scope\r\n Perhaps bind ‘wl1’ sometime after binding ‘a1’\r\n • In the type signature:\r\n elimWeirdList :: forall (a :: Type)\r\n (wl :: WeirdList a)\r\n (p :: forall (x :: Type). x > WeirdList x > Type).\r\n Sing wl\r\n > (forall (y :: Type). p _ WeirdNil)\r\n > (forall (z :: Type) (x :: z) (xs :: WeirdList (WeirdList z)).\r\n Sing x > Sing xs > p _ xs > p _ (WeirdCons x xs))\r\n > p _ wl\r\n \r\n21  elimWeirdList :: forall (a :: Type) (wl :: WeirdList a)\r\n  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^...\r\n\r\nFoo.hs:24:41: error:\r\n • Found type wildcard ‘_’ standing for ‘w0’\r\n Where: ‘w0’ is an ambiguous type variable\r\n ‘x0’ is an ambiguous type variable\r\n To use the inferred type, enable PartialTypeSignatures\r\n • In the type signature:\r\n elimWeirdList :: forall (a :: Type)\r\n (wl :: WeirdList a)\r\n (p :: forall (x :: Type). x > WeirdList x > Type).\r\n Sing wl\r\n > (forall (y :: Type). p _ WeirdNil)\r\n > (forall (z :: Type) (x :: z) (xs :: WeirdList (WeirdList z)).\r\n Sing x > Sing xs > p _ xs > p _ (WeirdCons x xs))\r\n > p _ wl\r\n \r\n24  > (forall (y :: Type). p _ WeirdNil)\r\n  ^\r\n\r\nFoo.hs:26:44: error:ghc: panic! (the 'impossible' happened)\r\n (GHC version 8.2.1 for x86_64unknownlinux):\r\n\tNo skolem info:\r\n z_a1sY[sk:2]\r\n Call stack:\r\n CallStack (from HasCallStack):\r\n prettyCurrentCallStack, called at compiler/utils/Outputable.hs:1133:58 in ghc:Outputable\r\n callStackDoc, called at compiler/utils/Outputable.hs:1137:37 in ghc:Outputable\r\n pprPanic, called at compiler/typecheck/TcErrors.hs:2653:5 in ghc:TcErrors\r\n}}}\r\n\r\n(The error messages from HEAD, at commit 791947db6db32ef7d4772a821a0823e558e3c05b, are the same as in GHC 8.2.1.)","type_of_failure":"OtherFailure","blocking":[]} >(Originally spun off from #13877.)
The following program gives a somewhat decent error message in GHC 8.0.1:
```hs
{# LANGUAGE GADTs #}
{# LANGUAGE RankNTypes #}
{# LANGUAGE ScopedTypeVariables #}
{# LANGUAGE TypeApplications #}
{# LANGUAGE TypeFamilies #}
{# LANGUAGE TypeInType #}
module Bug where
import Data.Kind
data family Sing (a :: k)
data WeirdList :: Type > Type where
WeirdNil :: WeirdList a
WeirdCons :: a > WeirdList (WeirdList a) > WeirdList a
data instance Sing (z :: WeirdList a) where
SWeirdNil :: Sing WeirdNil
SWeirdCons :: Sing w > Sing wws > Sing (WeirdCons w wws)
elimWeirdList :: forall (a :: Type) (wl :: WeirdList a)
(p :: forall (x :: Type). x > WeirdList x > Type).
Sing wl
> (forall (y :: Type). p _ WeirdNil)
> (forall (z :: Type) (x :: z) (xs :: WeirdList (WeirdList z)).
Sing x > Sing xs > p _ xs
> p _ (WeirdCons x xs))
> p _ wl
elimWeirdList SWeirdNil pWeirdNil _ = pWeirdNil
elimWeirdList (SWeirdCons (x :: Sing (x :: z))
(xs :: Sing (xs :: WeirdList (WeirdList z))))
pWeirdNil pWeirdCons
= pWeirdCons @z @x @xs x xs
(elimWeirdList @(WeirdList z) @xs @p xs pWeirdNil pWeirdCons)
```
```
$ /opt/ghc/8.0.1/bin/ghci Foo.hs
GHCi, version 8.0.1: http://www.haskell.org/ghc/ :? for help
Loaded GHCi configuration from /home/rgscott/.ghci
[1 of 1] Compiling Bug ( Foo.hs, interpreted )
Foo.hs:34:8: error:
• Cannot apply expression of type ‘Sing wl
> (forall y. p x0 t3 'WeirdNil)
> (forall z (x :: z) (xs :: WeirdList (WeirdList z)).
Sing x
> Sing xs
> p (WeirdList z) t2 xs
> p z t1 ('WeirdCons x xs))
> p a t0 wl’
to a visible type argument ‘WeirdList z’
• In the sixth argument of ‘pWeirdCons’, namely
‘(elimWeirdList @(WeirdList z) @xs @p xs pWeirdNil pWeirdCons)’
In the expression:
pWeirdCons
@z
@x
@xs
x
xs
(elimWeirdList @(WeirdList z) @xs @p xs pWeirdNil pWeirdCons)
In an equation for ‘elimWeirdList’:
elimWeirdList
(SWeirdCons (x :: Sing (x :: z))
(xs :: Sing (xs :: WeirdList (WeirdList z))))
pWeirdNil
pWeirdCons
= pWeirdCons
@z
@x
@xs
x
xs
(elimWeirdList @(WeirdList z) @xs @p xs pWeirdNil pWeirdCons)
```
But in GHC 8.0.2, 8.2.1, and HEAD, it panics to varying degrees:
```
$ /opt/ghc/8.0.2/bin/ghci Foo.hs
GHCi, version 8.0.2: http://www.haskell.org/ghc/ :? for help
Loaded GHCi configuration from /home/rgscott/.ghci
[1 of 1] Compiling Bug ( Foo.hs, interpreted )
Foo.hs:24:41: error:
• Found type wildcard ‘_’ standing for ‘t0’
Where: ‘t0’ is an ambiguous type variable
‘x0’ is an ambiguous type variable
To use the inferred type, enable PartialTypeSignatures
• In the type signature:
elimWeirdList :: forall (a :: Type)
(wl :: WeirdList a)
(p :: forall (x :: Type). x > WeirdList x > Type).
Sing wl
> (forall (y :: Type). p _ WeirdNil)
> (forall (z :: Type) (x :: z) (xs :: WeirdList (WeirdList z)).
Sing x > Sing xs > p _ xs > p _ (WeirdCons x xs))
> p _ wl
• Relevant bindings include
elimWeirdList :: Sing wl
> (forall y. p x0 t0 'WeirdNil)
> (forall z (x :: z) (xs :: WeirdList (WeirdList z)).
Sing x
> Sing xs > p (WeirdList z) t1 xs > p z t2 ('WeirdCons x xs))
> p a t3 wl
(bound at Foo.hs:29:1)
Foo.hs:26:44: error:ghc: panic! (the 'impossible' happened)
(GHC version 8.0.2 for x86_64unknownlinux):
No skolem info: z_a13X[sk]
```
```
$ /opt/ghc/8.2.1/bin/ghci Foo.hs
GHCi, version 8.2.1: http://www.haskell.org/ghc/ :? for help
Loaded GHCi configuration from /home/rgscott/.ghci
[1 of 1] Compiling Bug ( Foo.hs, interpreted )
Foo.hs:21:18: error:
• The kind of variable ‘wl1’, namely ‘WeirdList a1’,
depends on variable ‘a1’ from an inner scope
Perhaps bind ‘wl1’ sometime after binding ‘a1’
• In the type signature:
elimWeirdList :: forall (a :: Type)
(wl :: WeirdList a)
(p :: forall (x :: Type). x > WeirdList x > Type).
Sing wl
> (forall (y :: Type). p _ WeirdNil)
> (forall (z :: Type) (x :: z) (xs :: WeirdList (WeirdList z)).
Sing x > Sing xs > p _ xs > p _ (WeirdCons x xs))
> p _ wl

21  elimWeirdList :: forall (a :: Type) (wl :: WeirdList a)
 ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^...
Foo.hs:24:41: error:
• Found type wildcard ‘_’ standing for ‘w0’
Where: ‘w0’ is an ambiguous type variable
‘x0’ is an ambiguous type variable
To use the inferred type, enable PartialTypeSignatures
• In the type signature:
elimWeirdList :: forall (a :: Type)
(wl :: WeirdList a)
(p :: forall (x :: Type). x > WeirdList x > Type).
Sing wl
> (forall (y :: Type). p _ WeirdNil)
> (forall (z :: Type) (x :: z) (xs :: WeirdList (WeirdList z)).
Sing x > Sing xs > p _ xs > p _ (WeirdCons x xs))
> p _ wl

24  > (forall (y :: Type). p _ WeirdNil)
 ^
Foo.hs:26:44: error:ghc: panic! (the 'impossible' happened)
(GHC version 8.2.1 for x86_64unknownlinux):
No skolem info:
z_a1sY[sk:2]
Call stack:
CallStack (from HasCallStack):
prettyCurrentCallStack, called at compiler/utils/Outputable.hs:1133:58 in ghc:Outputable
callStackDoc, called at compiler/utils/Outputable.hs:1137:37 in ghc:Outputable
pprPanic, called at compiler/typecheck/TcErrors.hs:2653:5 in ghc:TcErrors
```
(The error messages from HEAD, at commit 791947db6db32ef7d4772a821a0823e558e3c05b, are the same as in GHC 8.2.1.)
<details><summary>Trac metadata</summary>
 Trac field  Value 
    
 Version  8.0.2 
 Type  Bug 
 TypeOfFailure  OtherFailure 
 Priority  normal 
 Resolution  Unresolved 
 Component  Compiler (Type checker) 
 Test case  
 Differential revisions  
 BlockedBy  
 Related  #13877 
 Blocking  
 CC  
 Operating system  
 Architecture  
</details>
<! {"blocked_by":[],"summary":"Typed holes regression in GHC 8.0.2: No skolem info: z_a1sY[sk:2]","status":"New","operating_system":"","component":"Compiler (Type checker)","related":[13877],"milestone":"","resolution":"Unresolved","owner":{"tag":"Unowned"},"version":"8.0.2","keywords":["PartialTypeSignatures","TypeFamilies,","TypeInType,"],"differentials":[],"test_case":"","architecture":"","cc":[""],"type":"Bug","description":"(Originally spun off from #13877.)\r\n\r\nThe following program gives a somewhat decent error message in GHC 8.0.1:\r\n\r\n{{{#!hs\r\n{# LANGUAGE GADTs #}\r\n{# LANGUAGE RankNTypes #}\r\n{# LANGUAGE ScopedTypeVariables #}\r\n{# LANGUAGE TypeApplications #}\r\n{# LANGUAGE TypeFamilies #}\r\n{# LANGUAGE TypeInType #}\r\nmodule Bug where\r\n\r\nimport Data.Kind\r\n\r\ndata family Sing (a :: k)\r\n\r\ndata WeirdList :: Type > Type where\r\n WeirdNil :: WeirdList a\r\n WeirdCons :: a > WeirdList (WeirdList a) > WeirdList a\r\n\r\ndata instance Sing (z :: WeirdList a) where\r\n SWeirdNil :: Sing WeirdNil\r\n SWeirdCons :: Sing w > Sing wws > Sing (WeirdCons w wws)\r\n\r\nelimWeirdList :: forall (a :: Type) (wl :: WeirdList a)\r\n (p :: forall (x :: Type). x > WeirdList x > Type).\r\n Sing wl\r\n > (forall (y :: Type). p _ WeirdNil)\r\n > (forall (z :: Type) (x :: z) (xs :: WeirdList (WeirdList z)).\r\n Sing x > Sing xs > p _ xs\r\n > p _ (WeirdCons x xs))\r\n > p _ wl\r\nelimWeirdList SWeirdNil pWeirdNil _ = pWeirdNil\r\nelimWeirdList (SWeirdCons (x :: Sing (x :: z))\r\n (xs :: Sing (xs :: WeirdList (WeirdList z))))\r\n pWeirdNil pWeirdCons\r\n = pWeirdCons @z @x @xs x xs\r\n (elimWeirdList @(WeirdList z) @xs @p xs pWeirdNil pWeirdCons)\r\n}}}\r\n\r\n{{{\r\n$ /opt/ghc/8.0.1/bin/ghci Foo.hs \r\nGHCi, version 8.0.1: http://www.haskell.org/ghc/ :? for help\r\nLoaded GHCi configuration from /home/rgscott/.ghci\r\n[1 of 1] Compiling Bug ( Foo.hs, interpreted )\r\n\r\nFoo.hs:34:8: error:\r\n • Cannot apply expression of type ‘Sing wl\r\n > (forall y. p x0 t3 'WeirdNil)\r\n > (forall z (x :: z) (xs :: WeirdList (WeirdList z)).\r\n Sing x\r\n > Sing xs\r\n > p (WeirdList z) t2 xs\r\n > p z t1 ('WeirdCons x xs))\r\n > p a t0 wl’\r\n to a visible type argument ‘WeirdList z’\r\n • In the sixth argument of ‘pWeirdCons’, namely\r\n ‘(elimWeirdList @(WeirdList z) @xs @p xs pWeirdNil pWeirdCons)’\r\n In the expression:\r\n pWeirdCons\r\n @z\r\n @x\r\n @xs\r\n x\r\n xs\r\n (elimWeirdList @(WeirdList z) @xs @p xs pWeirdNil pWeirdCons)\r\n In an equation for ‘elimWeirdList’:\r\n elimWeirdList\r\n (SWeirdCons (x :: Sing (x :: z))\r\n (xs :: Sing (xs :: WeirdList (WeirdList z))))\r\n pWeirdNil\r\n pWeirdCons\r\n = pWeirdCons\r\n @z\r\n @x\r\n @xs\r\n x\r\n xs\r\n (elimWeirdList @(WeirdList z) @xs @p xs pWeirdNil pWeirdCons)\r\n}}}\r\n\r\nBut in GHC 8.0.2, 8.2.1, and HEAD, it panics to varying degrees:\r\n\r\n{{{\r\n$ /opt/ghc/8.0.2/bin/ghci Foo.hs \r\nGHCi, version 8.0.2: http://www.haskell.org/ghc/ :? for help\r\nLoaded GHCi configuration from /home/rgscott/.ghci\r\n[1 of 1] Compiling Bug ( Foo.hs, interpreted )\r\n\r\nFoo.hs:24:41: error:\r\n • Found type wildcard ‘_’ standing for ‘t0’\r\n Where: ‘t0’ is an ambiguous type variable\r\n ‘x0’ is an ambiguous type variable\r\n To use the inferred type, enable PartialTypeSignatures\r\n • In the type signature:\r\n elimWeirdList :: forall (a :: Type)\r\n (wl :: WeirdList a)\r\n (p :: forall (x :: Type). x > WeirdList x > Type).\r\n Sing wl\r\n > (forall (y :: Type). p _ WeirdNil)\r\n > (forall (z :: Type) (x :: z) (xs :: WeirdList (WeirdList z)).\r\n Sing x > Sing xs > p _ xs > p _ (WeirdCons x xs))\r\n > p _ wl\r\n • Relevant bindings include\r\n elimWeirdList :: Sing wl\r\n > (forall y. p x0 t0 'WeirdNil)\r\n > (forall z (x :: z) (xs :: WeirdList (WeirdList z)).\r\n Sing x\r\n > Sing xs > p (WeirdList z) t1 xs > p z t2 ('WeirdCons x xs))\r\n > p a t3 wl\r\n (bound at Foo.hs:29:1)\r\n\r\nFoo.hs:26:44: error:ghc: panic! (the 'impossible' happened)\r\n (GHC version 8.0.2 for x86_64unknownlinux):\r\n\tNo skolem info: z_a13X[sk]\r\n}}}\r\n\r\n{{{\r\n$ /opt/ghc/8.2.1/bin/ghci Foo.hs \r\nGHCi, version 8.2.1: http://www.haskell.org/ghc/ :? for help\r\nLoaded GHCi configuration from /home/rgscott/.ghci\r\n[1 of 1] Compiling Bug ( Foo.hs, interpreted )\r\n\r\nFoo.hs:21:18: error:\r\n • The kind of variable ‘wl1’, namely ‘WeirdList a1’,\r\n depends on variable ‘a1’ from an inner scope\r\n Perhaps bind ‘wl1’ sometime after binding ‘a1’\r\n • In the type signature:\r\n elimWeirdList :: forall (a :: Type)\r\n (wl :: WeirdList a)\r\n (p :: forall (x :: Type). x > WeirdList x > Type).\r\n Sing wl\r\n > (forall (y :: Type). p _ WeirdNil)\r\n > (forall (z :: Type) (x :: z) (xs :: WeirdList (WeirdList z)).\r\n Sing x > Sing xs > p _ xs > p _ (WeirdCons x xs))\r\n > p _ wl\r\n \r\n21  elimWeirdList :: forall (a :: Type) (wl :: WeirdList a)\r\n  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^...\r\n\r\nFoo.hs:24:41: error:\r\n • Found type wildcard ‘_’ standing for ‘w0’\r\n Where: ‘w0’ is an ambiguous type variable\r\n ‘x0’ is an ambiguous type variable\r\n To use the inferred type, enable PartialTypeSignatures\r\n • In the type signature:\r\n elimWeirdList :: forall (a :: Type)\r\n (wl :: WeirdList a)\r\n (p :: forall (x :: Type). x > WeirdList x > Type).\r\n Sing wl\r\n > (forall (y :: Type). p _ WeirdNil)\r\n > (forall (z :: Type) (x :: z) (xs :: WeirdList (WeirdList z)).\r\n Sing x > Sing xs > p _ xs > p _ (WeirdCons x xs))\r\n > p _ wl\r\n \r\n24  > (forall (y :: Type). p _ WeirdNil)\r\n  ^\r\n\r\nFoo.hs:26:44: error:ghc: panic! (the 'impossible' happened)\r\n (GHC version 8.2.1 for x86_64unknownlinux):\r\n\tNo skolem info:\r\n z_a1sY[sk:2]\r\n Call stack:\r\n CallStack (from HasCallStack):\r\n prettyCurrentCallStack, called at compiler/utils/Outputable.hs:1133:58 in ghc:Outputable\r\n callStackDoc, called at compiler/utils/Outputable.hs:1137:37 in ghc:Outputable\r\n pprPanic, called at compiler/typecheck/TcErrors.hs:2653:5 in ghc:TcErrors\r\n}}}\r\n\r\n(The error messages from HEAD, at commit 791947db6db32ef7d4772a821a0823e558e3c05b, are the same as in GHC 8.2.1.)","type_of_failure":"OtherFailure","blocking":[]} >Richard Eisenbergrae@richarde.devRichard Eisenbergrae@richarde.devhttps://gitlab.haskell.org/ghc/ghc//issues/13933Support Typeable instances for types with coercions20190707T18:19:21ZRichard Eisenbergrae@richarde.devSupport Typeable instances for types with coercionsIf I say
```hs
{# LANGUAGE GADTs, TypeApplications, TypeInType #}
module Bug where
import Type.Reflection
data G a where
MkG :: a ~ Bool => G a
rep = typeRep @MkG
```
I get
```
Bug.hs:10:7: error:
• No instance for (Typeable <>) arising from a use of ‘typeRep’
• In the expression: typeRep @MkG
In an equation for ‘rep’: rep = typeRep @MkG

10  rep = typeRep @MkG

```
First off, the error message is confusing, mentioning the mysterious `<>`. But more importantly, it would be nice if the `Typeable` mechanism supported coercions.
<details><summary>Trac metadata</summary>
 Trac field  Value 
    
 Version  8.3 
 Type  Bug 
 TypeOfFailure  OtherFailure 
 Priority  normal 
 Resolution  Unresolved 
 Component  Compiler 
 Test case  
 Differential revisions  
 BlockedBy  
 Related  
 Blocking  
 CC  
 Operating system  
 Architecture  
</details>
<! {"blocked_by":[],"summary":"Support Typeable instances for types with coercions","status":"New","operating_system":"","component":"Compiler","related":[],"milestone":"","resolution":"Unresolved","owner":{"tag":"Unowned"},"version":"8.3","keywords":["TypeInType,","Typeable"],"differentials":[],"test_case":"","architecture":"","cc":[""],"type":"Bug","description":"If I say\r\n\r\n{{{#!hs\r\n{# LANGUAGE GADTs, TypeApplications, TypeInType #}\r\n\r\nmodule Bug where\r\n\r\nimport Type.Reflection\r\n\r\ndata G a where\r\n MkG :: a ~ Bool => G a\r\n\r\nrep = typeRep @MkG\r\n}}}\r\n\r\nI get\r\n\r\n{{{\r\nBug.hs:10:7: error:\r\n • No instance for (Typeable <>) arising from a use of ‘typeRep’\r\n • In the expression: typeRep @MkG\r\n In an equation for ‘rep’: rep = typeRep @MkG\r\n \r\n10  rep = typeRep @MkG\r\n  \r\n}}}\r\n\r\nFirst off, the error message is confusing, mentioning the mysterious `<>`. But more importantly, it would be nice if the `Typeable` mechanism supported coercions.","type_of_failure":"OtherFailure","blocking":[]} >If I say
```hs
{# LANGUAGE GADTs, TypeApplications, TypeInType #}
module Bug where
import Type.Reflection
data G a where
MkG :: a ~ Bool => G a
rep = typeRep @MkG
```
I get
```
Bug.hs:10:7: error:
• No instance for (Typeable <>) arising from a use of ‘typeRep’
• In the expression: typeRep @MkG
In an equation for ‘rep’: rep = typeRep @MkG

10  rep = typeRep @MkG

```
First off, the error message is confusing, mentioning the mysterious `<>`. But more importantly, it would be nice if the `Typeable` mechanism supported coercions.
<details><summary>Trac metadata</summary>
 Trac field  Value 
    
 Version  8.3 
 Type  Bug 
 TypeOfFailure  OtherFailure 
 Priority  normal 
 Resolution  Unresolved 
 Component  Compiler 
 Test case  
 Differential revisions  
 BlockedBy  
 Related  
 Blocking  
 CC  
 Operating system  
 Architecture  
</details>
<! {"blocked_by":[],"summary":"Support Typeable instances for types with coercions","status":"New","operating_system":"","component":"Compiler","related":[],"milestone":"","resolution":"Unresolved","owner":{"tag":"Unowned"},"version":"8.3","keywords":["TypeInType,","Typeable"],"differentials":[],"test_case":"","architecture":"","cc":[""],"type":"Bug","description":"If I say\r\n\r\n{{{#!hs\r\n{# LANGUAGE GADTs, TypeApplications, TypeInType #}\r\n\r\nmodule Bug where\r\n\r\nimport Type.Reflection\r\n\r\ndata G a where\r\n MkG :: a ~ Bool => G a\r\n\r\nrep = typeRep @MkG\r\n}}}\r\n\r\nI get\r\n\r\n{{{\r\nBug.hs:10:7: error:\r\n • No instance for (Typeable <>) arising from a use of ‘typeRep’\r\n • In the expression: typeRep @MkG\r\n In an equation for ‘rep’: rep = typeRep @MkG\r\n \r\n10  rep = typeRep @MkG\r\n  \r\n}}}\r\n\r\nFirst off, the error message is confusing, mentioning the mysterious `<>`. But more importantly, it would be nice if the `Typeable` mechanism supported coercions.","type_of_failure":"OtherFailure","blocking":[]} >https://gitlab.haskell.org/ghc/ghc//issues/12612Allow kinds of associated types to depend on earlier associated types20190707T18:25:50ZdavemenendezAllow kinds of associated types to depend on earlier associated typesGHC rejects the following code:
```hs
class C t where
type K t :: Type
type T t :: K t > Type
m :: t > T t a
```
with this error message
```
• Type constructor ‘K’ cannot be used here
(it is defined and used in the same recursive group)
• In the kind ‘K t > Type’
```
See [email discussion](https://mail.haskell.org/pipermail/glasgowhaskellusers/2016September/026402.html). This is connected to #12088, at least as far as defining instances of C.
<details><summary>Trac metadata</summary>
 Trac field  Value 
    
 Version  8.0.1 
 Type  Bug 
 TypeOfFailure  OtherFailure 
 Priority  normal 
 Resolution  Unresolved 
 Component  Compiler 
 Test case  
 Differential revisions  
 BlockedBy  
 Related  
 Blocking  
 CC  
 Operating system  
 Architecture  
</details>
<! {"blocked_by":[],"summary":"Allow kinds of associated types to depend on earlier associated types","status":"New","operating_system":"","component":"Compiler","related":[],"milestone":"","resolution":"Unresolved","owner":{"tag":"Unowned"},"version":"8.0.1","keywords":[],"differentials":[],"test_case":"","architecture":"","cc":[""],"type":"Bug","description":"GHC rejects the following code:\r\n\r\n{{{#!hs\r\nclass C t where\r\n type K t :: Type\r\n type T t :: K t > Type\r\n\r\n m :: t > T t a\r\n}}}\r\n\r\nwith this error message\r\n\r\n{{{\r\n • Type constructor ‘K’ cannot be used here\r\n (it is defined and used in the same recursive group)\r\n • In the kind ‘K t > Type’\r\n}}}\r\n\r\nSee [https://mail.haskell.org/pipermail/glasgowhaskellusers/2016September/026402.html email discussion]. This is connected to #12088, at least as far as defining instances of C.","type_of_failure":"OtherFailure","blocking":[]} >GHC rejects the following code:
```hs
class C t where
type K t :: Type
type T t :: K t > Type
m :: t > T t a
```
with this error message
```
• Type constructor ‘K’ cannot be used here
(it is defined and used in the same recursive group)
• In the kind ‘K t > Type’
```
See [email discussion](https://mail.haskell.org/pipermail/glasgowhaskellusers/2016September/026402.html). This is connected to #12088, at least as far as defining instances of C.
<details><summary>Trac metadata</summary>
 Trac field  Value 
    
 Version  8.0.1 
 Type  Bug 
 TypeOfFailure  OtherFailure 
 Priority  normal 
 Resolution  Unresolved 
 Component  Compiler 
 Test case  
 Differential revisions  
 BlockedBy  
 Related  
 Blocking  
 CC  
 Operating system  
 Architecture  
</details>
<! {"blocked_by":[],"summary":"Allow kinds of associated types to depend on earlier associated types","status":"New","operating_system":"","component":"Compiler","related":[],"milestone":"","resolution":"Unresolved","owner":{"tag":"Unowned"},"version":"8.0.1","keywords":[],"differentials":[],"test_case":"","architecture":"","cc":[""],"type":"Bug","description":"GHC rejects the following code:\r\n\r\n{{{#!hs\r\nclass C t where\r\n type K t :: Type\r\n type T t :: K t > Type\r\n\r\n m :: t > T t a\r\n}}}\r\n\r\nwith this error message\r\n\r\n{{{\r\n • Type constructor ‘K’ cannot be used here\r\n (it is defined and used in the same recursive group)\r\n • In the kind ‘K t > Type’\r\n}}}\r\n\r\nSee [https://mail.haskell.org/pipermail/glasgowhaskellusers/2016September/026402.html email discussion]. This is connected to #12088, at least as far as defining instances of C.","type_of_failure":"OtherFailure","blocking":[]} >https://gitlab.haskell.org/ghc/ghc//issues/17674Kill EQ120200518T17:02:46ZRichard Eisenbergrae@richarde.devKill EQ1Note [Respecting definitional equality] introduces invariant EQ1:
```
(EQ1) No decomposable CastTy to the left of an AppTy, where a decomposable
cast is one that relates either a FunTy to a FunTy or a
ForAllTy to a ForAllTy.
```
This invariant is a small pain to uphold. And perhaps we don't need to. If, instead, we modified `eqType` to look at the kind of arguments to `AppTy`s, then we wouldn't need EQ1, significantly simplifying `mkAppTy`. (Actually, it will just remove 2 lines. But it removes two calls to `decomposePiCos`. This will allow the one remaining call to this function to have more sway in the design of `decomposePiCos`, simplifying that bit of code, as well.)
Another nice benefit: by removing this invariant, we have an opportunity to explore different solutions to #17644, which hit a dead end in EQ1. See https://gitlab.haskell.org/ghc/ghc/issues/17644#note_245757.Note [Respecting definitional equality] introduces invariant EQ1:
```
(EQ1) No decomposable CastTy to the left of an AppTy, where a decomposable
cast is one that relates either a FunTy to a FunTy or a
ForAllTy to a ForAllTy.
```
This invariant is a small pain to uphold. And perhaps we don't need to. If, instead, we modified `eqType` to look at the kind of arguments to `AppTy`s, then we wouldn't need EQ1, significantly simplifying `mkAppTy`. (Actually, it will just remove 2 lines. But it removes two calls to `decomposePiCos`. This will allow the one remaining call to this function to have more sway in the design of `decomposePiCos`, simplifying that bit of code, as well.)
Another nice benefit: by removing this invariant, we have an opportunity to explore different solutions to #17644, which hit a dead end in EQ1. See https://gitlab.haskell.org/ghc/ghc/issues/17644#note_245757.Richard Eisenbergrae@richarde.devRichard Eisenbergrae@richarde.devhttps://gitlab.haskell.org/ghc/ghc//issues/17567Never `Any`ify during kind inference20200720T14:04:52ZRichard Eisenbergrae@richarde.devNever `Any`ify during kind inference#14198 concludes with a new plan: never `Any`ify during kind inference. This ticket tracks this particular problem, separate from #14198.
Here are some examples of `Any`ification during kind inference:
#17301:
```hs
data B a
data TySing ty where
SB :: TySing (B a)
data ATySing where
MkATySing :: TySing ty > ATySing
type family Forget ty :: ATySing where
Forget (B a) = MkATySing SB
```
The RHS of that type family equation is really `MkATySing @alpha (SB @alpha)`, and the `alpha` gets zonked to `Any`.
#14198:
```hs
type T = forall a. Proxy a
```
The RHS of the type synonym is really `forall (a :: kappa). Proxy @kappa a`, and the `kappa` gets zonked to `Any`.
#17562:
```hs
class (forall a. a b ~ a c) => C b c
```
The superclass constraint is really `forall (a :: Type > kappa). (~) @kappa (a b) (a c))`, and the `kappa` gets zonked to `Any`.
We want to stop zonking to `Any`, preferring to error instead. But how should we implement?
* Option A: Use a new variant of `ZonkFlexi`, a choice carried around in a `ZonkEnv` that says what to do with empty metavariables. The new choice would cause an error. This new form of `ZonkFlexi` would be used in the final zonks in e.g. TcTyClsDecls. Open question: how to get a decent error message? I think we'd have to pass around the original, toplevel type in order to report it. By the time we have just the unbound metavariable, we have no context to report.
* Option B: Similar to (A), but don't report an error in the zonker. Instead, the new variant of `ZonkFlexi` would insert some magical error type. Then, the validity checker could do an early pass, looking for the error type; it can then report a nice error message.
* Option C: Look for all cases where `Any`ification might happen, and detect each one separately. This can produce lovely error messages. The solution for #17562 in !2313 does this. Perhaps we can pair this choice with a new `ZonkFlexi` that panics. That way, we'll know if we've missed a case.
* Simon below proposes Option D: Zap to `Type` instead of `Any`. I (Richard) view D as an addon to any of the above plans. Because `Type` is not always wellkinded, we can only zap to `Type` sometimes, and we still need to decide what we do at other times. Personally, I prefer not to treat `Type` specially with XPolyKinds` enabled, and so I'd prefer that we don't do this.
* EDIT: Also, we can consider Option E: Report poor error messages, saying something about an unconstrained kind variable. Users would be helped only via the context (the "In the data declaration for `Wurble`" bits). This would mean building up useful contexts in the zonker.
* Option F: The new constructor for `ZonkFlexi` could carry the toplevel type we are trying to zonk. If we encounter an unconstrained metavariable, we just look at this bit of context to produce the error message. Perhaps this could be combined with the contextbuilding in Option E for good effect. This is the first solution I'm actually happy with.
Thoughts?#14198 concludes with a new plan: never `Any`ify during kind inference. This ticket tracks this particular problem, separate from #14198.
Here are some examples of `Any`ification during kind inference:
#17301:
```hs
data B a
data TySing ty where
SB :: TySing (B a)
data ATySing where
MkATySing :: TySing ty > ATySing
type family Forget ty :: ATySing where
Forget (B a) = MkATySing SB
```
The RHS of that type family equation is really `MkATySing @alpha (SB @alpha)`, and the `alpha` gets zonked to `Any`.
#14198:
```hs
type T = forall a. Proxy a
```
The RHS of the type synonym is really `forall (a :: kappa). Proxy @kappa a`, and the `kappa` gets zonked to `Any`.
#17562:
```hs
class (forall a. a b ~ a c) => C b c
```
The superclass constraint is really `forall (a :: Type > kappa). (~) @kappa (a b) (a c))`, and the `kappa` gets zonked to `Any`.
We want to stop zonking to `Any`, preferring to error instead. But how should we implement?
* Option A: Use a new variant of `ZonkFlexi`, a choice carried around in a `ZonkEnv` that says what to do with empty metavariables. The new choice would cause an error. This new form of `ZonkFlexi` would be used in the final zonks in e.g. TcTyClsDecls. Open question: how to get a decent error message? I think we'd have to pass around the original, toplevel type in order to report it. By the time we have just the unbound metavariable, we have no context to report.
* Option B: Similar to (A), but don't report an error in the zonker. Instead, the new variant of `ZonkFlexi` would insert some magical error type. Then, the validity checker could do an early pass, looking for the error type; it can then report a nice error message.
* Option C: Look for all cases where `Any`ification might happen, and detect each one separately. This can produce lovely error messages. The solution for #17562 in !2313 does this. Perhaps we can pair this choice with a new `ZonkFlexi` that panics. That way, we'll know if we've missed a case.
* Simon below proposes Option D: Zap to `Type` instead of `Any`. I (Richard) view D as an addon to any of the above plans. Because `Type` is not always wellkinded, we can only zap to `Type` sometimes, and we still need to decide what we do at other times. Personally, I prefer not to treat `Type` specially with XPolyKinds` enabled, and so I'd prefer that we don't do this.
* EDIT: Also, we can consider Option E: Report poor error messages, saying something about an unconstrained kind variable. Users would be helped only via the context (the "In the data declaration for `Wurble`" bits). This would mean building up useful contexts in the zonker.
* Option F: The new constructor for `ZonkFlexi` could carry the toplevel type we are trying to zonk. If we encounter an unconstrained metavariable, we just look at this bit of context to produce the error message. Perhaps this could be combined with the contextbuilding in Option E for good effect. This is the first solution I'm actually happy with.
Thoughts?https://gitlab.haskell.org/ghc/ghc//issues/17368Implement homogeneous equality20200310T14:21:59ZRichard Eisenbergrae@richarde.devImplement homogeneous equalityAs observed in [two](https://repository.brynmawr.edu/cgi/viewcontent.cgi?article=1076&context=compsci_pubs) [papers](https://richarde.dev/papers/2019/deproles/deproles.pdf), the primitive equality type in GHC can be made homogeneous. This is both a simplification over the status quo (heterogeneous equality) and an important step toward implementing dependent types.
This ticket is to track this change.
Step 1: Modify the typechecker to use predicates instead of types internally. This will essentially be a glorification of `PredTree` (renamed `Pred`), and a `CtEvidence` will now store a `Pred`, not a `PredType`.
See also https://gitlab.haskell.org/ghc/ghc/wikis/dependenthaskell/phase2, which has much discussion.As observed in [two](https://repository.brynmawr.edu/cgi/viewcontent.cgi?article=1076&context=compsci_pubs) [papers](https://richarde.dev/papers/2019/deproles/deproles.pdf), the primitive equality type in GHC can be made homogeneous. This is both a simplification over the status quo (heterogeneous equality) and an important step toward implementing dependent types.
This ticket is to track this change.
Step 1: Modify the typechecker to use predicates instead of types internally. This will essentially be a glorification of `PredTree` (renamed `Pred`), and a `CtEvidence` will now store a `Pred`, not a `PredType`.
See also https://gitlab.haskell.org/ghc/ghc/wikis/dependenthaskell/phase2, which has much discussion.Richard Eisenbergrae@richarde.devRichard Eisenbergrae@richarde.devhttps://gitlab.haskell.org/ghc/ghc//issues/14119Refactor type patterns20190707T18:18:23ZRichard Eisenbergrae@richarde.devRefactor type patternsType patterns are used in instance heads: class instances, type instances, and data/newtype instances. These type patterns differ from ordinary types in several ways:
 They must never mention a `forall`.
 They must never mention a context. (The context in a class instance head is a different part of the instance construct.)
 They must never mention a type family.
Types that appear as arguments on the LHS of a RULE should also be type patterns.
Type patterns are used in GHC differently than types as well: we should match only against patterns, never ordinary types.
I thus propose that a new datatype within GHC to store type patterns. I'll call it `TyPat` here, but perhaps a longer name is better. The matcher (in the `Unify` module) would then take a `TyPat` parameter, making clear which type is the template and which is the concrete instantiation.
We could have a new algorithm to typecheck/desugar source syntax into patterns. This new algorithm could accommodate issues like those mentioned in [ticket:14038\#comment:140786](https://gitlab.haskell.org//ghc/ghc/issues/14038#note_140786). It could also avoid ever putting a type family in a kind parameter, which would fix #12564. Separating out checking type patterns from proper types could also improve GADT pattern matching in types, which is currently more like "wobbly types" than OutsideIn.
<details><summary>Trac metadata</summary>
 Trac field  Value 
    
 Version  8.2.1 
 Type  Task 
 TypeOfFailure  OtherFailure 
 Priority  normal 
 Resolution  Unresolved 
 Component  Compiler 
 Test case  
 Differential revisions  
 BlockedBy  
 Related  
 Blocking  #12564, #14038 
 CC  
 Operating system  
 Architecture  
</details>
<! {"blocked_by":[],"summary":"Refactor type patterns","status":"New","operating_system":"","component":"Compiler","related":[],"milestone":"","resolution":"Unresolved","owner":{"tag":"Unowned"},"version":"8.2.1","keywords":["TypeInType"],"differentials":[],"test_case":"","architecture":"","cc":[""],"type":"Task","description":"Type patterns are used in instance heads: class instances, type instances, and data/newtype instances. These type patterns differ from ordinary types in several ways:\r\n\r\n * They must never mention a `forall`.\r\n * They must never mention a context. (The context in a class instance head is a different part of the instance construct.)\r\n * They must never mention a type family.\r\n\r\nTypes that appear as arguments on the LHS of a RULE should also be type patterns.\r\n\r\nType patterns are used in GHC differently than types as well: we should match only against patterns, never ordinary types.\r\n\r\nI thus propose that a new datatype within GHC to store type patterns. I'll call it `TyPat` here, but perhaps a longer name is better. The matcher (in the `Unify` module) would then take a `TyPat` parameter, making clear which type is the template and which is the concrete instantiation.\r\n\r\nWe could have a new algorithm to typecheck/desugar source syntax into patterns. This new algorithm could accommodate issues like those mentioned in comment:6:ticket:14038. It could also avoid ever putting a type family in a kind parameter, which would fix #12564. Separating out checking type patterns from proper types could also improve GADT pattern matching in types, which is currently more like \"wobbly types\" than OutsideIn.","type_of_failure":"OtherFailure","blocking":[12564,14038]} >Type patterns are used in instance heads: class instances, type instances, and data/newtype instances. These type patterns differ from ordinary types in several ways:
 They must never mention a `forall`.
 They must never mention a context. (The context in a class instance head is a different part of the instance construct.)
 They must never mention a type family.
Types that appear as arguments on the LHS of a RULE should also be type patterns.
Type patterns are used in GHC differently than types as well: we should match only against patterns, never ordinary types.
I thus propose that a new datatype within GHC to store type patterns. I'll call it `TyPat` here, but perhaps a longer name is better. The matcher (in the `Unify` module) would then take a `TyPat` parameter, making clear which type is the template and which is the concrete instantiation.
We could have a new algorithm to typecheck/desugar source syntax into patterns. This new algorithm could accommodate issues like those mentioned in [ticket:14038\#comment:140786](https://gitlab.haskell.org//ghc/ghc/issues/14038#note_140786). It could also avoid ever putting a type family in a kind parameter, which would fix #12564. Separating out checking type patterns from proper types could also improve GADT pattern matching in types, which is currently more like "wobbly types" than OutsideIn.
<details><summary>Trac metadata</summary>
 Trac field  Value 
    
 Version  8.2.1 
 Type  Task 
 TypeOfFailure  OtherFailure 
 Priority  normal 
 Resolution  Unresolved 
 Component  Compiler 
 Test case  
 Differential revisions  
 BlockedBy  
 Related  
 Blocking  #12564, #14038 
 CC  
 Operating system  
 Architecture  
</details>
<! {"blocked_by":[],"summary":"Refactor type patterns","status":"New","operating_system":"","component":"Compiler","related":[],"milestone":"","resolution":"Unresolved","owner":{"tag":"Unowned"},"version":"8.2.1","keywords":["TypeInType"],"differentials":[],"test_case":"","architecture":"","cc":[""],"type":"Task","description":"Type patterns are used in instance heads: class instances, type instances, and data/newtype instances. These type patterns differ from ordinary types in several ways:\r\n\r\n * They must never mention a `forall`.\r\n * They must never mention a context. (The context in a class instance head is a different part of the instance construct.)\r\n * They must never mention a type family.\r\n\r\nTypes that appear as arguments on the LHS of a RULE should also be type patterns.\r\n\r\nType patterns are used in GHC differently than types as well: we should match only against patterns, never ordinary types.\r\n\r\nI thus propose that a new datatype within GHC to store type patterns. I'll call it `TyPat` here, but perhaps a longer name is better. The matcher (in the `Unify` module) would then take a `TyPat` parameter, making clear which type is the template and which is the concrete instantiation.\r\n\r\nWe could have a new algorithm to typecheck/desugar source syntax into patterns. This new algorithm could accommodate issues like those mentioned in comment:6:ticket:14038. It could also avoid ever putting a type family in a kind parameter, which would fix #12564. Separating out checking type patterns from proper types could also improve GADT pattern matching in types, which is currently more like \"wobbly types\" than OutsideIn.","type_of_failure":"OtherFailure","blocking":[12564,14038]} >Richard Eisenbergrae@richarde.devRichard Eisenbergrae@richarde.devhttps://gitlab.haskell.org/ghc/ghc//issues/13408Consider inferring a higherrank kind for type synonyms20200925T17:22:29ZRichard Eisenbergrae@richarde.devConsider inferring a higherrank kind for type synonymsIn terms, a definition comprising one nonrecursive equation may have a higherrank type inferred. For example:
```hs
f :: (forall a. a > a > a) > Int
f z = z 0 1
g = f
```
`g` gets an inferred type equal to `f`'s. However, this fails at the type level:
```hs
type F (z :: forall a. a > a > a) = z 0 1
type G = F
```
If anything is strange here, it's that the termlevel definition is accepted. GHC should not be in the business of inferring higherrank types. But there is an exception for definitions comprising one nonrecursive equation.
This ticket proposes expanding this behavior to the type level, allowing `G` to be accepted.
This is spun off from #13399, but is not tightly coupled to that ticket.
<details><summary>Trac metadata</summary>
 Trac field  Value 
    
 Version  8.0.1 
 Type  FeatureRequest 
 TypeOfFailure  OtherFailure 
 Priority  normal 
 Resolution  Unresolved 
 Component  Compiler 
 Test case  
 Differential revisions  
 BlockedBy  
 Related  
 Blocking  
 CC  
 Operating system  
 Architecture  
</details>
<! {"blocked_by":[],"summary":"Consider inferring a higherrank kind for type synonyms","status":"New","operating_system":"","component":"Compiler","related":[],"milestone":"","resolution":"Unresolved","owner":{"tag":"Unowned"},"version":"8.0.1","keywords":["TypeInType"],"differentials":[],"test_case":"","architecture":"","cc":[""],"type":"FeatureRequest","description":"In terms, a definition comprising one nonrecursive equation may have a higherrank type inferred. For example:\r\n\r\n{{{#!hs\r\nf :: (forall a. a > a > a) > Int\r\nf z = z 0 1\r\n\r\ng = f\r\n}}}\r\n\r\n`g` gets an inferred type equal to `f`'s. However, this fails at the type level:\r\n\r\n{{{#!hs\r\ntype F (z :: forall a. a > a > a) = z 0 1\r\n\r\ntype G = F\r\n}}}\r\n\r\nIf anything is strange here, it's that the termlevel definition is accepted. GHC should not be in the business of inferring higherrank types. But there is an exception for definitions comprising one nonrecursive equation.\r\n\r\nThis ticket proposes expanding this behavior to the type level, allowing `G` to be accepted.\r\n\r\nThis is spun off from #13399, but is not tightly coupled to that ticket.","type_of_failure":"OtherFailure","blocking":[]} >In terms, a definition comprising one nonrecursive equation may have a higherrank type inferred. For example:
```hs
f :: (forall a. a > a > a) > Int
f z = z 0 1
g = f
```
`g` gets an inferred type equal to `f`'s. However, this fails at the type level:
```hs
type F (z :: forall a. a > a > a) = z 0 1
type G = F
```
If anything is strange here, it's that the termlevel definition is accepted. GHC should not be in the business of inferring higherrank types. But there is an exception for definitions comprising one nonrecursive equation.
This ticket proposes expanding this behavior to the type level, allowing `G` to be accepted.
This is spun off from #13399, but is not tightly coupled to that ticket.
<details><summary>Trac metadata</summary>
 Trac field  Value 
    
 Version  8.0.1 
 Type  FeatureRequest 
 TypeOfFailure  OtherFailure 
 Priority  normal 
 Resolution  Unresolved 
 Component  Compiler 
 Test case  
 Differential revisions  
 BlockedBy  
 Related  
 Blocking  
 CC  
 Operating system  
 Architecture  
</details>
<! {"blocked_by":[],"summary":"Consider inferring a higherrank kind for type synonyms","status":"New","operating_system":"","component":"Compiler","related":[],"milestone":"","resolution":"Unresolved","owner":{"tag":"Unowned"},"version":"8.0.1","keywords":["TypeInType"],"differentials":[],"test_case":"","architecture":"","cc":[""],"type":"FeatureRequest","description":"In terms, a definition comprising one nonrecursive equation may have a higherrank type inferred. For example:\r\n\r\n{{{#!hs\r\nf :: (forall a. a > a > a) > Int\r\nf z = z 0 1\r\n\r\ng = f\r\n}}}\r\n\r\n`g` gets an inferred type equal to `f`'s. However, this fails at the type level:\r\n\r\n{{{#!hs\r\ntype F (z :: forall a. a > a > a) = z 0 1\r\n\r\ntype G = F\r\n}}}\r\n\r\nIf anything is strange here, it's that the termlevel definition is accepted. GHC should not be in the business of inferring higherrank types. But there is an exception for definitions comprising one nonrecursive equation.\r\n\r\nThis ticket proposes expanding this behavior to the type level, allowing `G` to be accepted.\r\n\r\nThis is spun off from #13399, but is not tightly coupled to that ticket.","type_of_failure":"OtherFailure","blocking":[]} >https://gitlab.haskell.org/ghc/ghc//issues/11962Support induction recursion20200106T11:26:23ZRichard Eisenbergrae@richarde.devSupport induction recursionNow that we have `XTypeInType`, let's take it a step further and allow induction recursion. This feature exists in Agda and Idris. Here is a short example of what can be done in Agda:
```
mutual
 Codes for types.
data U : Set where
nat : U
pi : (a : U) → (El a → U) → U
 A function that interprets codes as actual types.
El : U → Set
El nat = ℕ
El (pi a b) = (x : El a) → El (b x)
```
Note that the `U` datatype and the `El` function depend on each other. But if you look more closely, the header for `U` does not depend on `El`; only the constructors of `U` depend on `El`. So if we typecheck `U : Set` first, then `El : U → Set`, then the constructors of `U`, then the equations of `El`, we're OK.
Translation into Haskell:
```
import Data.Kind
data family Sing (a :: k)  we still require singletons
data U :: Type where
Nat :: U
Pi :: Sing (a :: U) > (El a > U) > U
type family El (u :: U) :: Type where
El 'Nat = Int
El (Pi a b) = forall (x :: El a). Sing x > El (b x)
```
This currently errors with
```
• Type constructor ‘U’ cannot be used here
(it is defined and used in the same recursive group)
• In the kind ‘U’
```
It needn't error. (I'm cheating a bit here, because for unrelated reasons, we can't return a `forall` on the righthand side of a type family. But that's not the issue at hand.)
I have some very sketchy notes on how we might do this [here](https://gitlab.haskell.org/ghc/ghc/wikis/dependenthaskell/internal#separatingtypesignaturesfromdefinitions).
<details><summary>Trac metadata</summary>
 Trac field  Value 
    
 Version  8.1 
 Type  FeatureRequest 
 TypeOfFailure  OtherFailure 
 Priority  normal 
 Resolution  Unresolved 
 Component  Compiler 
 Test case  
 Differential revisions  
 BlockedBy  
 Related  
 Blocking  
 CC  
 Operating system  
 Architecture  
</details>
<! {"blocked_by":[],"summary":"Support induction recursion","status":"New","operating_system":"","component":"Compiler","related":[],"milestone":"","resolution":"Unresolved","owner":{"tag":"Unowned"},"version":"8.1","keywords":[],"differentials":[],"test_case":"","architecture":"","cc":[""],"type":"FeatureRequest","description":"Now that we have `XTypeInType`, let's take it a step further and allow induction recursion. This feature exists in Agda and Idris. Here is a short example of what can be done in Agda:\r\n\r\n{{{\r\n mutual\r\n  Codes for types.\r\n\r\n data U : Set where\r\n nat : U\r\n pi : (a : U) → (El a → U) → U\r\n\r\n  A function that interprets codes as actual types.\r\n\r\n El : U → Set\r\n El nat = ℕ\r\n El (pi a b) = (x : El a) → El (b x)\r\n}}}\r\n\r\nNote that the `U` datatype and the `El` function depend on each other. But if you look more closely, the header for `U` does not depend on `El`; only the constructors of `U` depend on `El`. So if we typecheck `U : Set` first, then `El : U → Set`, then the constructors of `U`, then the equations of `El`, we're OK.\r\n\r\nTranslation into Haskell:\r\n\r\n{{{\r\nimport Data.Kind\r\n\r\ndata family Sing (a :: k)  we still require singletons\r\n\r\ndata U :: Type where\r\n Nat :: U\r\n Pi :: Sing (a :: U) > (El a > U) > U\r\n\r\ntype family El (u :: U) :: Type where\r\n El 'Nat = Int\r\n El (Pi a b) = forall (x :: El a). Sing x > El (b x)\r\n}}}\r\n\r\nThis currently errors with\r\n\r\n{{{\r\n • Type constructor ‘U’ cannot be used here\r\n (it is defined and used in the same recursive group)\r\n • In the kind ‘U’\r\n}}}\r\n\r\nIt needn't error. (I'm cheating a bit here, because for unrelated reasons, we can't return a `forall` on the righthand side of a type family. But that's not the issue at hand.)\r\n\r\nI have some very sketchy notes on how we might do this [wiki:DependentHaskell/Internal#Separatingtypesignaturesfromdefinitions here].","type_of_failure":"OtherFailure","blocking":[]} >Now that we have `XTypeInType`, let's take it a step further and allow induction recursion. This feature exists in Agda and Idris. Here is a short example of what can be done in Agda:
```
mutual
 Codes for types.
data U : Set where
nat : U
pi : (a : U) → (El a → U) → U
 A function that interprets codes as actual types.
El : U → Set
El nat = ℕ
El (pi a b) = (x : El a) → El (b x)
```
Note that the `U` datatype and the `El` function depend on each other. But if you look more closely, the header for `U` does not depend on `El`; only the constructors of `U` depend on `El`. So if we typecheck `U : Set` first, then `El : U → Set`, then the constructors of `U`, then the equations of `El`, we're OK.
Translation into Haskell:
```
import Data.Kind
data family Sing (a :: k)  we still require singletons
data U :: Type where
Nat :: U
Pi :: Sing (a :: U) > (El a > U) > U
type family El (u :: U) :: Type where
El 'Nat = Int
El (Pi a b) = forall (x :: El a). Sing x > El (b x)
```
This currently errors with
```
• Type constructor ‘U’ cannot be used here
(it is defined and used in the same recursive group)
• In the kind ‘U’
```
It needn't error. (I'm cheating a bit here, because for unrelated reasons, we can't return a `forall` on the righthand side of a type family. But that's not the issue at hand.)
I have some very sketchy notes on how we might do this [here](https://gitlab.haskell.org/ghc/ghc/wikis/dependenthaskell/internal#separatingtypesignaturesfromdefinitions).
<details><summary>Trac metadata</summary>
 Trac field  Value 
    
 Version  8.1 
 Type  FeatureRequest 
 TypeOfFailure  OtherFailure 
 Priority  normal 
 Resolution  Unresolved 
 Component  Compiler 
 Test case  
 Differential revisions  
 BlockedBy  
 Related  
 Blocking  
 CC  
 Operating system  
 Architecture  
</details>
<! {"blocked_by":[],"summary":"Support induction recursion","status":"New","operating_system":"","component":"Compiler","related":[],"milestone":"","resolution":"Unresolved","owner":{"tag":"Unowned"},"version":"8.1","keywords":[],"differentials":[],"test_case":"","architecture":"","cc":[""],"type":"FeatureRequest","description":"Now that we have `XTypeInType`, let's take it a step further and allow induction recursion. This feature exists in Agda and Idris. Here is a short example of what can be done in Agda:\r\n\r\n{{{\r\n mutual\r\n  Codes for types.\r\n\r\n data U : Set where\r\n nat : U\r\n pi : (a : U) → (El a → U) → U\r\n\r\n  A function that interprets codes as actual types.\r\n\r\n El : U → Set\r\n El nat = ℕ\r\n El (pi a b) = (x : El a) → El (b x)\r\n}}}\r\n\r\nNote that the `U` datatype and the `El` function depend on each other. But if you look more closely, the header for `U` does not depend on `El`; only the constructors of `U` depend on `El`. So if we typecheck `U : Set` first, then `El : U → Set`, then the constructors of `U`, then the equations of `El`, we're OK.\r\n\r\nTranslation into Haskell:\r\n\r\n{{{\r\nimport Data.Kind\r\n\r\ndata family Sing (a :: k)  we still require singletons\r\n\r\ndata U :: Type where\r\n Nat :: U\r\n Pi :: Sing (a :: U) > (El a > U) > U\r\n\r\ntype family El (u :: U) :: Type where\r\n El 'Nat = Int\r\n El (Pi a b) = forall (x :: El a). Sing x > El (b x)\r\n}}}\r\n\r\nThis currently errors with\r\n\r\n{{{\r\n • Type constructor ‘U’ cannot be used here\r\n (it is defined and used in the same recursive group)\r\n • In the kind ‘U’\r\n}}}\r\n\r\nIt needn't error. (I'm cheating a bit here, because for unrelated reasons, we can't return a `forall` on the righthand side of a type family. But that's not the issue at hand.)\r\n\r\nI have some very sketchy notes on how we might do this [wiki:DependentHaskell/Internal#Separatingtypesignaturesfromdefinitions here].","type_of_failure":"OtherFailure","blocking":[]} >https://gitlab.haskell.org/ghc/ghc//issues/18753Tighten up the treatment of loose types in the solver20200927T02:25:14ZRichard Eisenbergrae@richarde.devTighten up the treatment of loose types in the solver`GHC.Tc.Solver.Monad` includes
```
Note [Use loose types in inert set]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Say we know (Eq (a > c1)) and we need (Eq (a > c2)). One is clearly
solvable from the other. So, we do lookup in the inert set using
loose types, which omits the kindcheck.
We must be careful when using the result of a lookup because it may
not match the requested info exactly!
```
There are several problems.
* We aren't always careful. For example, `lookupInInerts` tries `lookupSolvedDict` and then `lookupInertDict`, both of which use this "loose" lookup. Yet the result of `lookupInInerts` just uses the result, without any special handling to take the fact that the types might not match exactly.
* The Note doesn't actually make sense. The only difference between `Eq (a > c1)` and `Eq (a > c2)` is the coercion. But I sincerely hope we never care about the contents of a coercion. So even if we use evidence of type `Eq (a > c1)` where GHC is expecting `Eq (a > c2)` (assuming `c1` and `c2` have the same type), all will be well.
* What I think the Note is trying to say will never happen. That is, the actual implementation of the inertset lookup tries to match types while ignoring their kinds. (Normal type matching requires that the kind matches also.) So a better example would be something like comparing `Eq (a :: k1)` with `Eq (a :: k2)`. But that's impossible, even with casts: all type families and classes have *closed* kinds, meaning that any variables that appear in the kinds of arguments must themselves be earlier arguments. In other words, if I have welltyped `T blah1 (... :: kind1)` and `T blah2 (... :: kind2)`, then either `kind1` equals `kind2` or `blah1` and `blah2` must differ. We use this logic elsewhere  in particular, in the pure unifier.
Conclusions:
* Using "loose" matching (that is, ignoring kinds) in the solver is the right thing.
* We should update the Note with my argument above.
* No special care needs to be taken when using loose matching in this way. This means we can drop a few redundant equality checks (e.g. in `lookupFlatCache`).
I'm pretty confident about this all, but I'd like a doublecheck before I commit a fix.`GHC.Tc.Solver.Monad` includes
```
Note [Use loose types in inert set]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Say we know (Eq (a > c1)) and we need (Eq (a > c2)). One is clearly
solvable from the other. So, we do lookup in the inert set using
loose types, which omits the kindcheck.
We must be careful when using the result of a lookup because it may
not match the requested info exactly!
```
There are several problems.
* We aren't always careful. For example, `lookupInInerts` tries `lookupSolvedDict` and then `lookupInertDict`, both of which use this "loose" lookup. Yet the result of `lookupInInerts` just uses the result, without any special handling to take the fact that the types might not match exactly.
* The Note doesn't actually make sense. The only difference between `Eq (a > c1)` and `Eq (a > c2)` is the coercion. But I sincerely hope we never care about the contents of a coercion. So even if we use evidence of type `Eq (a > c1)` where GHC is expecting `Eq (a > c2)` (assuming `c1` and `c2` have the same type), all will be well.
* What I think the Note is trying to say will never happen. That is, the actual implementation of the inertset lookup tries to match types while ignoring their kinds. (Normal type matching requires that the kind matches also.) So a better example would be something like comparing `Eq (a :: k1)` with `Eq (a :: k2)`. But that's impossible, even with casts: all type families and classes have *closed* kinds, meaning that any variables that appear in the kinds of arguments must themselves be earlier arguments. In other words, if I have welltyped `T blah1 (... :: kind1)` and `T blah2 (... :: kind2)`, then either `kind1` equals `kind2` or `blah1` and `blah2` must differ. We use this logic elsewhere  in particular, in the pure unifier.
Conclusions:
* Using "loose" matching (that is, ignoring kinds) in the solver is the right thing.
* We should update the Note with my argument above.
* No special care needs to be taken when using loose matching in this way. This means we can drop a few redundant equality checks (e.g. in `lookupFlatCache`).
I'm pretty confident about this all, but I'd like a doublecheck before I commit a fix.https://gitlab.haskell.org/ghc/ghc//issues/18689Why check for fdefertypeerrors in metaTyVarUpdateOK?20200921T22:31:11ZRichard Eisenbergrae@richarde.devWhy check for fdefertypeerrors in metaTyVarUpdateOK?Function `metaTyVarUpdateOK` changes its behavior depending on the presence of `fdefertypeerrors` in an obscure case around heterogeneous equalities; see the code in `preCheck` that deals with `badCoercionHoleCo`. This is undocumented (in the code), and neither Simon nor I can figure out why it's done.
Task: figure this out, and either document or remove this behavior.Function `metaTyVarUpdateOK` changes its behavior depending on the presence of `fdefertypeerrors` in an obscure case around heterogeneous equalities; see the code in `preCheck` that deals with `badCoercionHoleCo`. This is undocumented (in the code), and neither Simon nor I can figure out why it's done.
Task: figure this out, and either document or remove this behavior.Richard Eisenbergrae@richarde.devRichard Eisenbergrae@richarde.dev