chard; GHC
| Subject: Re: Indexed constraint families
|
| Thanks for extending the wiki page.
|
| If I understand this correctly, then the *reason* to distinguish between
lifted and
| unlifted constraints is that equality constraints (in contrast to all the
others)
| have no *runtime* witness a
| To: Simon Peyton-Jones
| Cc: GHC; Dominic Orchard
| Subject: Re: Indexed constraint families
|
| This is fantastic stuff!! Thanks, Max!
|
| I wonder about the difference between lifted and unlifted constraints,
though. The
| wiki page doesn't seem to mention it.
|
| Manuel
|
|
| Simon Pe
Folks
Max has just pushed a patch that adds a major new feature: indexed constraint
families; see http://hackage.haskell.org/trac/ghc/wiki/KindFact
Internally, a big change is that there is no PredTy any more. So TypeRep has
gotten simpler (hooray).
Julien, Dimitrios: this will conflict
