[GHC] #13143: NOINLINE and worker/wrapper

GHC ghc-devs at haskell.org
Mon Feb 6 17:19:46 UTC 2017


#13143: NOINLINE and worker/wrapper
-------------------------------------+-------------------------------------
        Reporter:  simonpj           |                Owner:
            Type:  bug               |               Status:  closed
        Priority:  normal            |            Milestone:  8.2.1
       Component:  Compiler          |              Version:  8.0.1
      Resolution:  fixed             |             Keywords:
Operating System:  Unknown/Multiple  |         Architecture:
                                     |  Unknown/Multiple
 Type of failure:  None/Unknown      |            Test Case:
      Blocked By:                    |             Blocking:
 Related Tickets:                    |  Differential Rev(s):  D3046
       Wiki Page:                    |
-------------------------------------+-------------------------------------

Comment (by nomeata):

 Hmm. I guess I am a bit late to the party, but I feel uneasy about this.
 On some level, to me, `NOINLINE` means “Dear compiler, in all uses of
 this, please treat this as a black box with precisely the interface I give
 here.” Allowing W/W moves a part of what the function does (e.g. taking a
 constructor apart) to the uses.

 I guess the test case did not come up with anything, but does this
 interfere with rules? A rule might mention the `NOINLINE` thing, but
 aftetr W/W, the worker remains, and a rule will no longer fire.

 `NOINLINE` is also very useful to prevent the compiler from optimizing
 test cases in our test suite too much (although I see that that should not
 be a driving factor).

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13143#comment:5>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler


More information about the ghc-tickets mailing list