Re: [GHC] #13415: Instance declarations don't recognize named wildcards

20 Mar 2017

      #13415: Instance declarations don't recognize named wildcards
-------------------------------------+-------------------------------------
        Reporter:  RyanGlScott       |                Owner:  (none)
            Type:  bug               |               Status:  patch
        Priority:  normal            |            Milestone:  8.4.1
       Component:  Compiler (Type    |              Version:  8.0.1
  checker)                           |
      Resolution:                    |             Keywords:
Operating System:  Unknown/Multiple  |         Architecture:
 Type of failure:  Poor/confusing    |  Unknown/Multiple
  error message                      |            Test Case:
      Blocked By:                    |             Blocking:
 Related Tickets:  #13324            |  Differential Rev(s):  Phab:D3332
       Wiki Page:                    |
-------------------------------------+-------------------------------------

Comment (by RyanGlScott):

 I feel like we might not be on the same page here, so let me try to clear
 things up.

 I might be operating under a misconception here, but isn't it impossible
 to properly detect wildcards in a type (and integrate it with
 `-XPartialTypeSignatures`) unless it's a `LHsSigWcType`? That is why the
 error message you get when you use named wildcards in instance contexts
 currently is so bizarre--it doesn't even properly //detect// that it's a
 wildcard, but instead misinterprets it as a type variable (leading to the
 not-quite-on-the-mark `-XConstraintKinds` error).

 So given this, if we are to be able to support `-XPartialTypeSignatures`
 eventually, we need to first transition from the use of `LHsSigType` in
 instance declarations to `LHsSigWcType`. I chose to do this first in its
 own Diff since:

 * It's far easier than adding `-XPartialTypeSignatures` support all in ome
 go. This is an easily identifiable and necessary component that still
 requires changing quite a few files, so splitting this task out will make
 the Diff(s) that //do// deal with `-XPartialTypeSignatures` support less
 noisy.
 * It requires a Haddock change, which is somewhat ugly, and I'd rather get
 it out of the way upfront.
 * It has the immediate benefit of getting GHC to recognize named wildcards
 in instance declarations, as noted above.

 Does that make sense?

--
Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/13415#comment:9
GHC http://www.haskell.org/ghc/
The Glasgow Haskell Compiler