[GHC] #5561: assertion overriden by other exceptions

[GHC]

#5561: assertion overriden by other exceptions
------------------------------------------+---------------------------------
  Reporter:  MikolajKonarski              |          Owner:                
      Type:  bug                          |         Status:  new           
  Priority:  normal                       |      Milestone:                
 Component:  Compiler                     |        Version:  7.3           
Resolution:                               |       Keywords:                
  Testcase:                               |      Blockedby:                
Difficulty:                               |             Os:  Linux         
  Blocking:                               |   Architecture:  x86_64 (amd64)
   Failure:  Incorrect result at runtime  |  
------------------------------------------+---------------------------------

Comment(by igloo):

 We've been debating this on IRC. I think we're all agreed that the current
 behaviour is correct, in the sense that that's how imprecise exceptions
 work. But we could "fix" ''this case'', as above, to behave as expected.

 The "fixed" version would still be correct in the sense of imprecise
 exceptions. It would behave differently to user-defined functions that
 aren't similarly fixed, although you can already get differences depending
 on whether GHC decides to inline assertError or not.

 Note that performance of optimised code won't generally suffer, as
 normally optimisations mean assert is disabled anyway.

 Mikolaj and I think it is worth fixing this case. Duncan thinks it will
 just add to the confusion. Any other opinions?

-- 
Ticket URL: <http://hackage.haskell.org/trac/ghc/ticket/5561#comment:4>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler