[GHC] #9636: Function with type error accepted

[GHC]

#9636: Function with type error accepted
-------------------------------------+-------------------------------------
              Reporter:  augustss    |            Owner:
                  Type:  bug         |           Status:  new
              Priority:  normal      |        Milestone:
             Component:  Compiler    |          Version:  7.8.3
            Resolution:              |         Keywords:
      Operating System:              |     Architecture:  Unknown/Multiple
  Unknown/Multiple                   |       Difficulty:  Unknown
       Type of failure:              |       Blocked By:
  None/Unknown                       |  Related Tickets:
             Test Case:              |
              Blocking:              |
Differential Revisions:              |
-------------------------------------+-------------------------------------

Comment (by augustss):

 To me saying that `T Bool` is OK because it's inaccessible is akin to
 saying that type incorrect expressions are OK as long as they are
 inaccessible.  After all, if you don't use them, they can't cause any
 harm.  But for expressions we have decided that this isn't acceptable.

 I guess making `T a` behave would require something like kind classes.  We
 don't say that the expression `show x` is unconditionally type correct.
 It depends on the type of `x` belonging to the `Show` class.  In the same
 way, `T a` is not unconditionally type correct, it's only type correct if
 `a` is one of the types where `T` is well defined.  Until we have
 something like that I think you'll have to accept that the substitution
 lemma doesn't work.  You can pretend it works by saying `T Bool` is a
 type, if that makes you happier.  I just wonder which type it is. :)

 But I'm not asking for the moon.  :)  I'd just like the compiler to tell
 me when it finds something that is clearly not going to work, like `T
 Bool`.  Exactly under what conditions and how it tells me, I don't care.

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9636#comment:12>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler