[commit: ghc] master's head updated: Merge branch 'master' of http://darcs.haskell.org/ghc into ghc-new-flavor (1a4ae30)

[dimitris at microsoft.com]

Repository : ssh://darcs.haskell.org//srv/darcs/ghc

Branch 'master' now includes:

     cc2d2e1... Midstream check-in on    (i) Replaced a lot of clunky and fragile EvVar handling code with        a more uniform ``flavor transformer'' API in the canonicalizer        and the interaction solver. Now EvVars are just fields inside        the CtFlavors.    (ii) Significantly simplified our caching story This patch does not validate yet and more refactoring is on the way.
     b4b7647... Merge branch 'master' of http://darcs.haskell.org/ghc into ghc-new-flavor
     c6411c3... Slightly more informal message when EvCast fails.
     3be60fa... Bugfixes: (i) wrong evidence generetation when using cached solved family equation           (ii) evidence loops because of caching fixed
     4e0f1a0... Small bugfix (for indexed_types/should_compile/T2291.hs). Because our inert sets do not tolerate more than a single constraint per family head we have to allow family interactions /with the inerts/ (not with top-level) when the context says simplEqsOnly.
     1cc92e9... defaultKind the kind of one of the types when you create a higher-dimensional equality proof, pretty much as we do for mkTcEqPred. Ugly and delicate but we have to rething the subkinding story. For now I am doing the same thing as mkTcEqPred.
     81faaee... Merge branch 'master' of http://darcs.haskell.org/ghc into ghc-new-flavor
     bf334d2... Add handling of LitTy. Still a TODO to deal properly with the forall case in there.
     e74fbf0... Started investigating how performance is affected and how intensively our caches are used.
     f30f90b... More informative commentary as to why we allow family interactions with inerts when simplEqsOnly is on.
     3f42011... Changing the orientation of a generated equality
     f84ea58... Deleting A LOT of commented out code!
     d86e6c0... A little tracing.
     6785820... Mostly commentary to follow up after discussions with SPJ on several open tickets.
     b3f6a8b... Merge branch 'master' of http://darcs.haskell.org/ghc into ghc-new-flavor
     8580b97... Deleting more old commented code.
     5802ebd... Dropping the idempotence restriction on the inert substitution, for efficiency. More documentation to follow.
     f9da524... Commentary, following the relaxation of idempotence of the inert substitution.
     d29e277... Bugfix in emitKindConstraint (must not cache orig constraint as Solved!); added short note to document an example.
     8737d47... When rewriting a flavor with rewriteCtFlavor, in the case of reflexivity, avoid setting the flavor PredType to be the new PredType if the types exactly match, so that in case of failure we report errors that do not have expanded type synonyms.
     2183c87... More informative tracing for ddump-cs-trace
     e3e740d... Wibble: using original type for the dictionary if it has not changed (to preserve synonyms)
     d3020a7... Adding a forgotten pre-canonicalication cache-lookup stage.
     0dd6cc5... Merge branch 'master' of http://darcs.haskell.org/ghc into ghc-new-flavor
     022c1d2... Merge branch 'master' of http://darcs.haskell.org/ghc into ghc-new-flavor
     1a4ae30... Merge branch 'master' of http://darcs.haskell.org/ghc into ghc-new-flavor