[GHC] #10892: ApplicativeDo should use *> and <*

[GHC]

#10892: ApplicativeDo should use *> and <*
-------------------------------------+-------------------------------------
        Reporter:  simonmar          |                Owner:  bollu
            Type:  task              |               Status:  new
        Priority:  high              |            Milestone:  8.4.1
       Component:  Compiler          |              Version:  7.11
      Resolution:                    |             Keywords:  ApplicativeDo
Operating System:  Unknown/Multiple  |         Architecture:
                                     |  Unknown/Multiple
 Type of failure:  None/Unknown      |            Test Case:
      Blocked By:                    |             Blocking:  12143
 Related Tickets:  #13309            |  Differential Rev(s):
       Wiki Page:                    |
-------------------------------------+-------------------------------------

Comment (by AaronFriel):

 @ekmett, @simonpj, @bgamari

 I am working on this issue and have a preliminary patch that explores the
 design space for this. For the exploring implementation, I added a new
 constructor of `ApplicativeArg`, `ApplicativeArgNil`, which lacks a
 pattern. To explore the design space and verify my type checking, I
 modified the renamer to use `<$` and `<*` (in place of `fmap` and `<*>`)
 if and only if every statement in the segment was an `ApplicativeArgNil`:


 {{{#!hs
 mkApplicativeStmt ctxt args need_join body_stmts
   | all isAppArgNil args
   = do { (replace_op, fvs1) <- lookupStmtName ctxt replaceFName
        ; (but_op, fvs2) <- lookupStmtName ctxt butAName
   ...
 }}}

 For the sake of faithfully implementing your proposed desugaring, I need
 to ask about this:

 {{{#!hs
 -- Example:
 f   = do foo;bar;baz; x <- quux; y <- quaffle; return (xyzzy x y)

 -- Desugaring:
 f'  = foo *> bar *> baz *> (xyzzy <$> quux <*> quaffle)
 }}}

 Is there a reason that desugaring is strictly better than:

 {{{#!hs
 -- Desugaring:
 f'' = xyzzy <$> (foo *> bar *> baz *> quux) <*> quaffle
 }}}

 I don't think it'd be too difficult to move the `*>` "then" operators to
 the beginning, but it would involve changing more of the existing
 applicative code to do so. I think that this style is more suited to
 addressing #13309.

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/10892#comment:19>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler