[GHC] #13148: Adding weak pointers to non-mutable unboxed values segfaults

[GHC]

#13148: Adding weak pointers to non-mutable unboxed values segfaults
-------------------------------------+-------------------------------------
        Reporter:  mboes             |                Owner:
            Type:  bug               |               Status:  new
        Priority:  normal            |            Milestone:
       Component:  Compiler          |              Version:  8.0.1
      Resolution:                    |             Keywords:
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 mboes):

 OK that makes sense. I was surprised that things worked with `MutVar#`
 because I assumed it was also unboxed, but I see now that's not the case.

 The types in `System.Mem.Weak` already preclude creating a weak pointer to
 an unboxed value. But that of `mkWeakNoFinalizer#` is not precise enough
 to rule that out. Is the rule that the first argument to this function is
 valid iff it is of unlifted (primitive) type? If so, I think it would be
 useful to document that there.

 This exploration arose from the need to retrofit a single finalizer to a
 codebase that makes extensive use of `Ptr` without having to switch to
 `ForeignPtr` everywhere. The latter being more inconvenient, since it
 can't appear in the type signature of foreign imports. I assume attaching
 a finalizer to a `Ptr` runs the risk that the finalizer might fire too
 early.

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13148#comment:2>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler