desugaring type function application to constraint makes bug disappear

When considering type functions, I find it helpful to desugar their applications into additional constraints (tf a becomes tf a~tfa=>tfa). But consider this example from a recent haskell-cafe thread:

{-# OPTIONS_GHC -fglasgow-exts #-}

class Blah f a where blah :: a -> T f f a
class A f where type T f :: (* -> *) -> * -> *

-- wrapper :: forall a f tf . (Blah f a,T f~tf) => a -> tf f a
wrapper :: forall a f . (Blah f a) => a -> T f f a
wrapper x = blah x

for which GHCi, version 6.9.20080514 yields:

C:\Documents and Settings\cr3\Desktop\TF.hs:8:12:
    Could not deduce (Blah f a) from the context (Blah f1 a)
      arising from a use of `blah'
                   at C:\Documents and Settings\cr3\Desktop\TF.hs:8:12-17
    Possible fix:
      add (Blah f a) to the context of the type signature for `wrapper'
    In the expression: blah x
    In the definition of `wrapper': wrapper x = blah x

C:\Documents and Settings\cr3\Desktop\TF.hs:8:12:
    Couldn't match expected type `T f1 f1 a'
           against inferred type `T f f a'
    In the expression: blah x
    In the definition of `wrapper': wrapper x = blah x
Failed, modules loaded: none.

Switching to the desugared version of the wrapper signature makes the error go away, so the "desugared" and original version are not equivalent in the current implementation of type families! See also http://www.haskell.org/pipermail/haskell-cafe/2008-July/044914.html.

For added fun, GHCi reports the sugared type when using the desugared one:

*Main> :t wrapper
wrapper :: (Blah f a) => a -> T f f a

Trac metadata

Trac field	Value
Version	6.9
Type	Bug
TypeOfFailure	OtherFailure
Priority	normal
Resolution	Unresolved
Component	Compiler (Type checker)
Test case
Differential revisions
BlockedBy
Related
Blocking
CC
Operating system	Unknown
Architecture	Unknown