[GHC] #15155: How untagged pointers sneak into banged fields

[GHC]

#15155: How untagged pointers sneak into banged fields
-------------------------------------+-------------------------------------
        Reporter:  heisenbug         |                Owner:  (none)
            Type:  bug               |               Status:  new
        Priority:  normal            |            Milestone:  8.6.1
       Component:  Compiler          |              Version:  8.4.2
      Resolution:                    |             Keywords:
Operating System:  Unknown/Multiple  |         Architecture:
                                     |  Unknown/Multiple
 Type of failure:  None/Unknown      |            Test Case:
      Blocked By:                    |             Blocking:  14677
 Related Tickets:                    |  Differential Rev(s):
       Wiki Page:                    |
-------------------------------------+-------------------------------------

Comment (by simonmar):

 I didn't really follow the description here.  Naively I'd expect `c` to
 compile into a CAF, like `c = case b of b' -> C b'` and then we'd be fine.
 If the compiler is assuming that `b` is already a value and thus avoiding
 evaluating it, that would be a false assumption. Where is the false
 assumption being made?

 I think the `IND_STATIC` stuff is a red herring. There's a top-level
 binding `b = a` which is compiled into an `IND_STATIC` as an optimisation,
 but it could also be compiled into a CAF, this is just a back-end code-
 generation choice.

-- 
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/15155#comment:3>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler