[GHC] #12045: Visible kind application

[GHC]

#12045: Visible kind application
-------------------------------------+-------------------------------------
        Reporter:  Iceland_jack      |                Owner:
            Type:  feature request   |               Status:  new
        Priority:  normal            |            Milestone:
       Component:  Compiler          |              Version:  8.1
      Resolution:                    |             Keywords:
                                     |  TypeApplications TypeInType
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):

 This is one of the great many things I would love to work on if I had
 time.

 Popping up a level: you've written a great many bug reports / feature
 requests recently. These are very helpful! Thanks! A good deal of them
 would naturally fall to me to implement/fix, but I'm dreadfully short of
 time these days. (And will be until September, at least.) So: Would you
 care to try to fix some of these infelicities? You have a great grasp of
 GHC's type system and how to abuse it (I mean that, quite surely, as a
 compliment), and I imagine you would have fun getting into the gearbox and
 getting your hands dirty.

 I conjecture that this ticket is actually a good place to start. My rule
 of thumb is that writing a new feature is easier than debugging someone
 else's mistake. For a user-facing feature like this, just start by
 extending the abstract syntax (in hsSyn/HsType.hs, to be specific), add
 some rules to the parser, and then follow the errors / warnings that
 ensue. Unlike adding `TypeApplications` to expressions, the type-level
 type checker (er, kind checker) already does lazy instantiation, so this
 shouldn't be a terribly invasive change.

 I'm quite happy to advise, even being short of time. One mode of
 collaboration might be to start a patch and submit it to Phabricator, and
 then I can offer feedback.

 (PS: I sometimes let tickets / Phab requests slip by me these days. If you
 want to be sure to get a response to something, email me. My contact info
 is [http://www.cis.upenn.edu/~eir here].)

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/12045#comment:2>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler