[GHC] #11511: Type family producing infinite type accepted as injective

[GHC]

#11511: Type family producing infinite type accepted as injective
-------------------------------------+-------------------------------------
        Reporter:  jstolarek         |                Owner:
            Type:  bug               |               Status:  new
        Priority:  normal            |            Milestone:
       Component:  Compiler (Type    |              Version:  8.1
  checker)                           |             Keywords:  TypeFamilies,
      Resolution:                    |  Injective
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 jstolarek):

 Yes, certainly what I said is conservative - some injective functions
 would be rejected. What I'm wondering is whether our current approach is
 sound from a logical point of view. Perhaps assuming `F` injective when
 checking injectivity of `F` is the reason why Richard couldn't complete
 the proof we presented in the paper?

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11511#comment:7>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler