[GHC] #15927: Weird interaction between fundeps and overlappable instances

[GHC]

#15927: Weird interaction between fundeps and overlappable instances
-------------------------------------+-------------------------------------
        Reporter:  Darwin226         |                Owner:  (none)
            Type:  bug               |               Status:  new
        Priority:  normal            |            Milestone:  8.6.3
       Component:  Compiler          |              Version:  8.6.2
      Resolution:                    |             Keywords:
                                     |  FunctionalDependencies
Operating System:  Unknown/Multiple  |         Architecture:
 Type of failure:  GHC accepts       |  Unknown/Multiple
  invalid program                    |            Test Case:
      Blocked By:                    |             Blocking:
 Related Tickets:  10675, 15632      |  Differential Rev(s):
       Wiki Page:                    |
-------------------------------------+-------------------------------------

Comment (by AntC):

 Replying to [comment:6 Darwin226]:
 > Well, since the mtl default is to not only disallow multiple states of
 the same type, but multiple states of ANY type, I wouldn't really call
 this non-scaleable. It's strictly more flexible than mtl.

 I'm not following: the approach you're using 'allows' multiple states of
 same type in the sense you can stack a `String` on top of a `String`. But
 it's useless: `getMyState` as written will only access one of them. Then
 being "flexible" in this way, is only going to lead to confusion.

 Monad Transformers used to do this stacking. (I'm talking over 10 years
 ago.) The `l` in `mtl` stands for library: in general many modules might
 import `mtl` and declare instances of `MonadTrans`. If `mtl` were still
 relying on overlapping instances, it would be impossible/difficult to co-
 ordinate the instances to each use a unique type; and any code in modules
 using the stack would risk the well-known issues of 'orphan instances'.

 Perhaps your intended usage has a different pattern, and you're prepared
 to do the co-ordination manually. But at least for a sanity check,
 validate when you add a type to the stack that it's unique.

-- 
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/15927#comment:7>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler