[GHC] #4485: Per-class incoherence, and solve incoherent instances last (was: Unexplained change in type-inference + OverlappingInstances in)

[GHC]

#4485: Per-class incoherence, and solve incoherent instances last
---------------------------------+------------------------------------------
    Reporter:  JeremyShaw        |        Owner:                           
        Type:  bug               |       Status:  new                      
    Priority:  normal            |    Milestone:  7.2.1                    
   Component:  Compiler          |      Version:  7.0.1 RC1                
    Keywords:                    |     Testcase:                           
   Blockedby:                    |   Difficulty:                           
          Os:  Unknown/Multiple  |     Blocking:                           
Architecture:  Unknown/Multiple  |      Failure:  GHC rejects valid program
---------------------------------+------------------------------------------

Comment(by simonpj):

 This patch fixes the crash in overlap reporting. Please push to the 7.0
 branch.
 {{{
 Mon Nov 15 14:28:05 GMT 2010  simonpj at microsoft.com
   * Ensure that instance overlap errors are report properly

   This (annoyingly) requires us to re-flatten the class predicate.
   See Note [Flattening in error message generation]

     M ./compiler/typecheck/TcErrors.lhs -37 +91
 }}}
 I'll re-title the ticket to now ask for per-class incoherent instances
 (which is a feature that makes sense to me).  This links to people who
 want somewhat less draconian checking on instance overlap: #3877 and
 #4259.

 Also there's a case for making the solver solve (incoherent) type class
 instances ''last'', so that all the equality info we have is available.

 Simon

-- 
Ticket URL: <http://hackage.haskell.org/trac/ghc/ticket/4485#comment:8>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler