[GHC] #13242: Panic "StgCmmEnv: variable not found" with ApplicativeDo and ExistentialQuantification

[GHC]

#13242: Panic "StgCmmEnv: variable not found" with ApplicativeDo and
ExistentialQuantification
-------------------------------------+-------------------------------------
        Reporter:  ljli              |                Owner:  simonmar
            Type:  bug               |               Status:  new
        Priority:  highest           |            Milestone:  8.0.3
       Component:  Compiler          |              Version:  8.1
      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 simonmar):

 > But the rule is terribly unsatisfactory.

 Yes it is.  I think I'd also be fine with disabling ApplicativeDo for all
 statements following an existential pattern match.  Or perhaps @rwbarton's
 suggestion:

 > whenever there is a pattern match that binds invisible stuff, just
 assume that that stuff is used in all following statements

 But by itself this doesn't solve the problem that the example in this
 ticket raises, because the A type doesn't bind any dictionaries or
 constraints.

 By all means refactor the implementation of `tcApplicativeStmts`!
 Provided the tests pass it should be fine.  I remember trying quite hard
 to simplify the implementation to use fewer fresh type variables, but I
 couldn't figure out how to do it.

--
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13242#comment:11>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler