Forgot to copy Simon and the list: 2012/3/19 Simon Peyton-Jones <[email protected]>: > Simon, David > > > > I’m a little concerned that Haddock may not have quite the right idea. > > > > In Haddock.Convert I see > > > > -- the main function here! yay! > > tyThingToLHsDecl :: TyThing -> LHsDecl Name > > > > But for a family instance, the right place to start is from the [FamInst], > as presumably you do for the type class instances, which have a [ClsInst]. > Yet the tyThingToLHsDecl does stuff with CoAxioms... > > > > You should really think of a CoAxiom as a supporting thing that must be in > the type env, with the same relationship to its FamInst as a DFunId as to > its ClsInst. > > > > Happy to elaborate further... just flagging something that is, I think, not > quite right. > > > Simon >
OK, thanks! I will take a look at it. I'm not surprised if that code needs an overhaul. David _______________________________________________ Cvs-ghc mailing list [email protected] http://www.haskell.org/mailman/listinfo/cvs-ghc
