[GHC] #14827: Recognize when inlining would create a join point

[GHC]

#14827: Recognize when inlining would create a join point
-------------------------------------+-------------------------------------
        Reporter:  ersetzen          |                Owner:  (none)
            Type:  feature request   |               Status:  new
        Priority:  normal            |            Milestone:
       Component:  Compiler          |              Version:  8.2.2
      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:                    |
-------------------------------------+-------------------------------------

Comment (by simonpj):

 I've got lost in this ticket.

 1. comment:13 shows (correctly) that if you specify phases wrong you can
 get bad results.  Yes, it's possible we could warn about this.

 2. comment:11, I think, shows a case where an INLINE pragmas makes things
 worse.  I have not dug into it.

 3. The original thing in the Description got tangled with stuff about
 loopification, so I'm not sure of its status.

 Are (1) and (2) connected?  If not, can you make a separate ticket for
 (1)?

 Is (30 dealt with now?  That is, are we left with (2) only?

-- 
Ticket URL: <http://ghc.haskell.org/trac/ghc/ticket/14827#comment:14>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler