[GHC] #13104: runRW# ruins join points

[GHC]

#13104: runRW# ruins join points
-------------------------------------+-------------------------------------
        Reporter:  lukemaurer        |                Owner:  chessai
            Type:  bug               |               Status:  infoneeded
        Priority:  normal            |            Milestone:  8.8.1
       Component:  Compiler          |              Version:  8.1
      Resolution:                    |             Keywords:  JoinPoints
Operating System:  Unknown/Multiple  |         Architecture:
 Type of failure:  Runtime           |  Unknown/Multiple
  performance bug                    |            Test Case:
      Blocked By:                    |             Blocking:
 Related Tickets:                    |  Differential Rev(s):
       Wiki Page:                    |
-------------------------------------+-------------------------------------
Changes (by sgraf):

 * status:  new => infoneeded


Comment:

 Actually, I don't think this reproduces any longer. What's actually the
 regression test here? The argument to `runST` in #12781 is properly eta-
 expanded in GHC 8.4.3. Any attempt to manually reproduce the issue with
 similar code as in the OP, e.g.

 {{{
 f = runST (loop (0 :: Int))
   where
     loop 1000 = return 42
     loop n = loop (n+1)
 }}}

 leads to this Core for `f`:

 {{{
 Rec {
 -- RHS size: {terms: 14, types: 11, coercions: 0, joins: 0/0}
 $wloop
 $wloop
   = \ ww_s2DT w_s2DQ ->
       case ww_s2DT of wild_Xk {
         __DEFAULT -> $wloop (+# wild_Xk 1#) w_s2DQ;
         1000# -> (# w_s2DQ, lvl_r2Eo #)
       }
 end Rec }

 -- RHS size: {terms: 4, types: 2, coercions: 0, joins: 0/0}
 f1
 f1 = \ w_s2DQ -> $wloop 0# w_s2DQ

 -- RHS size: {terms: 5, types: 30, coercions: 0, joins: 0/0}
 f
 f = case runRW# f1 of { (# ipv_a2Bw, ipv1_a2Bx #) -> ipv1_a2Bx }
 }}}

-- 
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/13104#comment:21>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler