[GHC] #12162: Concatenation of type level symbols missing

[GHC]

#12162: Concatenation of type level symbols missing
-------------------------------------+-------------------------------------
        Reporter:  augustss          |                Owner:  phadej
            Type:  feature request   |               Status:  new
        Priority:  normal            |            Milestone:
       Component:  Compiler          |              Version:  8.0.1
      Resolution:                    |             Keywords:
Operating System:  Unknown/Multiple  |         Architecture:
                                     |  Unknown/Multiple
 Type of failure:  None/Unknown      |            Test Case:
      Blocked By:                    |             Blocking:
 Related Tickets:                    |  Differential Rev(s):  Phab:D2632
       Wiki Page:                    |
-------------------------------------+-------------------------------------

Comment (by simonpj):

 Generally the patch looks good to me, thank you.  But can we first agree
 about goals and nomenclature.

 * Minor: you use the term "concat" in the implementation, but the list
 `concat` function has type `[[a]] -> [a]` not `[a] -> [a] -> [a]`.  We
 generally use "append" for the latter; e.g. `appendFS`.

 * In order to do this you have figured out how to add a new type-level
 operator to GHC.  Is that written up on the wiki anywhere?  If not it'd be
 great if you could do so?   There are a number of separate things you have
 to do. [wiki:TypeNats] would be a good place to start, despite its Nat-
 sounding title.

 * More significant. You are introducing a type-level operator `(<>) ::
 Symbol -> Symbol -> Symbol`.  Wouldn't `(++)` be more appropriate?  It's
 type is very like `(++)`.

   The other use of `(<>)` that I know of is in the pretty printing
 library, where it has type `Doc -> Doc -> Doc` so there is precedent
 there.

   I think it partly depends on whether we are regarding `Symbol` as an
 opaque type; in which case we might want to keep `(++)` for appending
 type-level lists (or does it exist already?).  But #11342 is (I think)
 arguing for a less opaque representation.

 * Incidentally I realise that I don't know how to find the Haddock docs
 for the functions available over `Symbol`....

 I'm copying in the core libraries committee, since this is really their
 territory.

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/12162#comment:8>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler