[GHC] #14137: Do more inlining into non-recursive join points

GHC ghc-devs at haskell.org
Fri Aug 18 13:53:48 UTC 2017


#14137: Do more inlining into non-recursive join points
-------------------------------------+-------------------------------------
        Reporter:  simonpj           |                Owner:  nomeata
            Type:  bug               |               Status:  new
        Priority:  normal            |            Milestone:
       Component:  Compiler          |              Version:  8.2.1
      Resolution:                    |             Keywords:  JoinPoints
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):

 I added the join-point related line here:
 {{{
 occAnalNonRecRhs env bndr bndrs body
   = occAnalLamOrRhs rhs_env bndrs body
   where
     -- See Note [Cascading inlines]
     env1 | certainly_inline = env
          | Just _ <- willBeJoinId_maybe bndr    = env
          | otherwise        = rhsCtxt env
 }}}
 to set `env` instead of `rhsCtxt`, and verified that it matches `lvl6` in
 the example above. But it did not cause any effect.

 (The desired effect would be that `ds5` gets inlined into `lvl6`, so that
 the remaining uses of `ds5`, all in `joinrec safe`, are tail-calls and
 `ds5` gets turned into a join point. Right?)

 I am attaching the Haskell’ification of the source code above that I was
 using for testing.

-- 
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/14137#comment:3>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler


More information about the ghc-tickets mailing list