[GHC] #10675: GHC does not check the functional dependency consistency condition correctly

[GHC]

#10675: GHC does not check the functional dependency consistency condition
correctly
-------------------------------------+-------------------------------------
        Reporter:  simonpj           |                Owner:  (none)
            Type:  bug               |               Status:  new
        Priority:  normal            |            Milestone:
       Component:  Compiler          |              Version:  7.10.1
      Resolution:                    |             Keywords:  FunDeps
Operating System:  Unknown/Multiple  |         Architecture:
                                     |  Unknown/Multiple
 Type of failure:  None/Unknown      |            Test Case:
      Blocked By:                    |             Blocking:
 Related Tickets:                    |  Differential Rev(s):
       Wiki Page:                    |
-------------------------------------+-------------------------------------

Comment (by AntC):

 hmm curioser and curioser. Going back to the O.P., I see this behaviour at
 both GHC 7.10 and 8.0.1:

 * Take out the `instance C Char ...`; and `f` is happy with any type of
 operand, not necessarily a `Maybe`.

 * Keep in the `instance C Char ...`; but give `f` a signature and again
 it's happy:
     `f :: (C Bool [a] [a]) => a -> [a]`

 * Keep in the `instance C Char ...`; give `f` this signature, and we're
 back needing `Maybe`:
     `f :: (C Bool [a] b) => a -> b`

 So there is something wrong: why is the `C Char ...` instance getting
 tangled up with the `C Bool ...` instance? The only connection would be
 checking conformance with FunDeps; but that should be only for validation,
 not 'type improvement' for function `f`(?)

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/10675#comment:14>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler