[GHC] #12463: SPECIALIZABLE pragma?

[GHC]

#12463: SPECIALIZABLE pragma?
-------------------------------------+-------------------------------------
        Reporter:  bgamari           |                Owner:
            Type:  feature request   |               Status:  new
        Priority:  low               |            Milestone:
       Component:  Compiler          |              Version:  8.0.1
      Resolution:                    |             Keywords:  Inlining
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 mpickering):

 Here is my understanding of what Ben means.

 ---

 We introduce a new top-level pragma which is introduced the the syntax
 `{-# RECURSIVE_SPECIALISABLE varid #-}`.

 Only top-level functions can be marked with this pragma.

 For a function `f` which is marked with `RECURSIVE_SPECIALISABLE`:

 1. When `f` is exported, `f`'s unfolding is included in the interface
 file. (As if `f` was marked `INLINABLE`).
 2. When `f` is used in the definition of another function `g`, `g`'s
 unfolding is included in the interface file when `g` is exported. (As if
 `g` was marked `INLINABLE`).

 ---

 Is that what you mean Ben? My questions are

 1. Why would you mark your function as `INLINABLE` rather than
 `RECURSIVE_SPECIALISABLE`?
 2. What advantages does this pragma have over including the unfoldings of
 all polymorphic functions ?

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/12463#comment:14>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler