Type checker rejects ambiguous top level declaration.

It doesn't seem like a serious problem, but I report it anyway.

With stack's project template, I just removed main and someFunc's type signature and changed someFunc's definition to return (). stack build worked well, but stack repl emitted this error message.

D:\ghc-corePrepPgm\app\Main.hs:6:1: warning: [-Wdeferred-type-errors]
    • Couldn't match type ‘ghc-prim-0.5.0.0:GHC.Prim.Any’ with ‘IO’
      Expected type: IO ()
        Actual type: ghc-prim-0.5.0.0:GHC.Prim.Any ()
    • In the expression: main
      When checking the type of the IO action ‘main’
ghc.EXE: panic! (the 'impossible' happened)
  (GHC version 8.0.2 for x86_64-unknown-mingw32):
        corePrepPgm
  [False] cobox_r1bo = typeError
                         @ 'VoidRep
                         @ (Any :: (* -> *)) ~# (IO :: (* -> *))
                         "D:\\ghc-corePrepPgm\\app\\Main.hs:6:1: error:\n\
                         \    \\226\\128\\162 Couldn't match type \\226\\128\\152ghc-prim-0.5.0.0:GHC.Prim.Any\\226\\128\\153 with \\226\\128\\152IO\\226\\128\\153\n\
                         \      Expected type: IO ()\n\
                         \        Actual type: ghc-prim-0.5.0.0:GHC.Prim.Any ()\n\
                         \    \\226\\128\\162 In the expression: main\n\
                         \      When checking the type of the IO action \\226\\128\\152main\\226\\128\\153\n\
                         \(deferred type error)"#

Please report this as a GHC bug:  http://www.haskell.org/ghc/reportabug

Trac metadata

Trac field	Value
Version	8.0.2
Type	Bug
TypeOfFailure	OtherFailure
Priority	normal
Resolution	Unresolved
Component	Compiler
Test case
Differential revisions
BlockedBy
Related
Blocking
CC
Operating system	Windows
Architecture

changed weight to 5

added Tbug Trac import labels

Attached file ghc-corePrepPgm.7z (download).

changed title from panic! (the 'impossible' happened): corePrepPgm to Type checker rejects ambiguous top level declaration.

Hi,

Thanks for the report. I am not sure that this is a bug or at least not a new one. It seems more like a limitation of the type inferencing.

The confusing error message comes from the handling of packages in the cabal file.

The error simplified is:

$ ~/ghc/inplace/bin/ghc-stage2.exe Main.hs
[1 of 2] Compiling Lib              ( Lib.hs, Lib.o )

Lib.hs:6:12: error:
    * Ambiguous type variable `m0' arising from a use of `return'
      prevents the constraint `(Monad m0)' from being solved.
      Relevant bindings include someFunc :: m0 () (bound at Lib.hs:6:1)
      Probable fix: use a type annotation to specify what `m0' should be.
      These potential instances exist:
        instance Monad IO -- Defined in `GHC.Base'
        instance Monad Maybe -- Defined in `GHC.Base'
        instance Monad ((->) r) -- Defined in `GHC.Base'
        ...plus two others
        (use -fprint-potential-instances to see them all)
    * In the expression: return ()
      In an equation for `someFunc': someFunc = return ()
  |
6 | someFunc = return ()
  |            ^^^^^^^^^

Essentially (I think) because there's so little context for the type inference it can't disambiguate between the possible types someFunc can have.

So it can't chose which type is the most general type for someFunc simply because they're all valid and can be the most general type.

So it needs to be told, either by adding a type signature, or using someFunc in a way that allows it to determine what the most general type should be.

e.g.

module Lib
    ( someFunc
    ) where

-- someFunc :: Monad m => m ()
someFunc = moreFunc

moreFunc :: IO ()
moreFunc = someFunc

But I'm not that familiar with GHC's typechecker so I'll reclassify and leave it for someone more knowledgeable in this area. I also don't know why stack repl would have worked.

Trac metadata

Trac field	Value
Component	Compiler → Compiler (Type checker)
Operating system	Windows → Unknown/Multiple

added typechecker label

The issue isn't the warning/error, it's the fact that GHC panicked afterward.

It would be helpful to have steps to reproduce that don't involve stack. E.g. stack repl is presumably invoking ghci somehow, what command exactly does it run?

This is the command which causes the panic.

ghc --interactive -i -isrc -hide-all-packages -package-id=base-4.9.1.0
-- in ghci
:add Lib app/Main.hs

Thanks. -fdefer-type-errors is also needed to reproduce (presumably you have this set in a .ghci file).

Oh sorry, I completely missed the panic!

Trac metadata

Trac field	Value
CC	bgamari → -

mentioned in commit 3c62b1d6

closed

Thanks. There was a real bug here.

Simon

Trac metadata

Trac field	Value
Resolution	Unresolved → ResolvedFixed
Test case	- → typecheck/should_fail/T13292

mentioned in issue #13510 (closed)

mentioned in issue #13838 (closed)

added Pnormal label