[GHC] #11715: Constraint vs *

[GHC]

#11715: Constraint vs *
-------------------------------------+-------------------------------------
        Reporter:  bgamari           |                Owner:
            Type:  bug               |               Status:  new
        Priority:  high              |            Milestone:  8.2.1
       Component:  Compiler (Type    |              Version:  8.0.1-rc1
  checker)                           |
      Resolution:                    |             Keywords:  Typeable
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 goldfire):

 Perhaps the user guide is deficient with regard to implicit parameter
 shadowing. Maybe it's not even reliable -- I'm not sure.

 As for your first paragraph, I'm not convinced yet: `Int` is not like any
 other constraint, in that it's not a class.

 In GHC 8.0, we have a few forms of constraint:
 1. Fully-applied classes
 2. Implicit parameters
 3. Tuples (which is actually special syntax for (1))
 4. Equality constraints
 5. Other built-in constraints (`Coercible`, `Typeable`, are there more?)
 6. Type families and synonyms which produce constraints

 You are advocating for something new:
 7. Other Haskell types

 We thus can choose how to handle case (7) independently of how we handle
 other constraints, just like (2) has different "shadowing" behavior than
 (1). So `Int` will always be different from `Given Int`, and I think
 that's OK.

 As long as implicit parameters are around, then we've already lost the
 thinness of constraints.

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/11715#comment:43>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler