Re: [GHC] #13615: Nondeterminism in ‘pure’ function w/ parallel evaluation & memo combinators

[GHC]

#13615: Nondeterminism in ‘pure’ function w/ parallel evaluation & memo combinators
-------------------------------------+-------------------------------------
        Reporter:  pacak             |                Owner:  (none)
            Type:  bug               |               Status:  new
        Priority:  highest           |            Milestone:  8.2.1
       Component:  Compiler          |              Version:  8.2.1-rc2
      Resolution:                    |             Keywords:
Operating System:  Unknown/Multiple  |         Architecture:
 Type of failure:  Incorrect result  |  Unknown/Multiple
  at runtime                         |            Test Case:
      Blocked By:                    |             Blocking:
 Related Tickets:                    |  Differential Rev(s):
       Wiki Page:                    |
-------------------------------------+-------------------------------------

Comment (by glguy):

 I believe I was experiencing this same bug in code that used
 Foreign.Marshal.Unsafe.unsafeLocalState and par together

 https://github.com/glguy/kami-solver/blob/master/src/Kami.hs#L172-L181

 unsafeLocalState is defined as unsafeDupablePerformIO. Switching to
 unsafePerformIO (which adds the noDuplicate) resolved the random
 segmentation faults.

 I'm just adding this comment here in case it becomes useful at some point.

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13615#comment:28>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler