[GHC] #12618: Add saturated constructor applications to Core

[GHC]

#12618: Add saturated constructor applications to Core
-------------------------------------+-------------------------------------
        Reporter:  simonpj           |                Owner:  nomeata
            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):
       Wiki Page:                    |
-------------------------------------+-------------------------------------

Comment (by nomeata):

 While I make progress with getting the tree to compiler properly again,
 here is one question that will need to be answered.

 Consider this rule:
 {{{
 "foldr/id" foldr (:) [] = \x  -> x
 }}}
 Because we desugar constructor in source to the ''wrapper'' (especially if
 they are unsaturated), but the wrapper is a function that will be marked
 as inlineable, the compiler now gives this error message:
 {{{
 libraries/base/GHC/Base.hs:855:1: warning: [-Winline-rule-shadowing]
     Rule "foldr/id" may never fire
       because ‘GHC.Types.$W:’ might inline first
 Probable fix: add an INLINE[n] or NOINLINE[n] pragma for ‘GHC.Types.$W:’
 }}}
 So at first I thought: Ok, no problem, I just force the inlining of data
 con wrappers after the desugaring of rule left-hand sides, and this might
 work for `[]`, but `(:)` is really used unsaturated here.

 What is the best way forward here?

 One way would be to disable this warning specifically for datacon workers,
 and then make the rule matcher smart enough to match both variants.

 Or alternatively, make the warning aware that an unsaturated use of a
 function with an unfolding will not inline, and it is thus ok to have
 something `INLINE` on the LHS of a rule, as long as it is unsaturated.

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/12618#comment:25>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler