[GHC] #947: ghc -O space leak: CSE between different CAFs

[GHC]

#947: ghc -O space leak: CSE between different CAFs
---------------------------+------------------------------------------------
  Reporter:  int-e@…       |          Owner:         
      Type:  bug           |         Status:  new    
  Priority:  normal        |      Milestone:  _|_    
 Component:  Compiler      |        Version:  6.5    
Resolution:                |       Keywords:         
        Os:  Linux         |   Architecture:  x86    
   Failure:  None/Unknown  |     Difficulty:  Unknown
  Testcase:                |      Blockedby:         
  Blocking:                |        Related:         
---------------------------+------------------------------------------------

Comment(by simonpj):

 John: presumably you've added yourself to the cc because this has bitten
 you?

 I have one other very simple suggestion: '''do no CSE for top-level
 CAFs'''.  Rationale:
  * It's not a good plan to rely on CSE for major sharing of really
 expensive expressions; if it really matters you'll want to do it yourself
 by let-binding the shared thing.  CSE is really intended for picking up
 nickles and dimes in inner loops.
  * Top-level things are evaluated at most once anyway, so the gain from
 sharing just dime's worth of work at top level is not worth having.

 I suspect that this simple change would deal with this ticket nicely.
 What do others think?

 Simon

-- 
Ticket URL: <http://hackage.haskell.org/trac/ghc/ticket/947#comment:9>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler