[GHC] #9370: unfolding info as seen when building a module depends on flags in a previously-compiled module

[GHC]

#9370: unfolding info as seen when building a module depends on flags in a
previously-compiled module
-------------------------------------+-------------------------------------
        Reporter:  carter            |                Owner:
            Type:  bug               |               Status:  new
        Priority:  high              |            Milestone:  8.2.1
       Component:  Compiler          |              Version:  7.8.3
      Resolution:                    |             Keywords:  newcomer,
                                     |  Inlining
Operating System:  Unknown/Multiple  |         Architecture:
 Type of failure:  Compile-time      |  Unknown/Multiple
  performance bug                    |            Test Case:
      Blocked By:                    |             Blocking:
 Related Tickets:  #8635             |  Differential Rev(s):
       Wiki Page:                    |
-------------------------------------+-------------------------------------

Comment (by bgamari):

 Thanks Richard! I hope we see you again when you have time for some
 hacking.

 I had a look at this and the result is Phab:D2609. There is a bit more to
 it that comment:41 suggests. In addition to what I describe there I also
 needed to track down a few places outside of the simplifier where
 unfoldings were being force. The first place I encountered was
 `OccurAnalyse`, which wants to look a free variables of unfoldings.

 There are two ways I can envision fixing this,
  * Have occurrence analysis pretend there are no unfoldings if `-fignore-
 interface-pragmas` is passed. Note that this is a bit of a lie: we would
 be ignoring all unfoldings, not just those from interface files (since we
 can't tell the difference between the two under the current scheme).
  * Skip occurrence analysis altogether. If we are going to disable
 inlining entirely then (I think) it serves little purpose.

 Simon, perhaps you could weigh in here?

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/9370#comment:43>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler