#14728: Is (GeneralizedNewtypeDeriving + associated type classes) completely bogus? -------------------------------------+------------------------------------- Reporter: RyanGlScott | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: Component: Compiler (Type | Version: 8.2.2 checker) | Keywords: deriving, Resolution: | TypeFamilies Operating System: Unknown/Multiple | Architecture: | Unknown/Multiple Type of failure: None/Unknown | Test Case: Blocked By: | Blocking: Related Tickets: | Differential Rev(s): Wiki Page: | -------------------------------------+------------------------------------- Comment (by RyanGlScott): After talking to goldfire about this, I can answer my earlier question in comment:7: Replying to [comment:7 RyanGlScott]:
Is this blocked (at least partially) on getting your branch merged and/or fixing #14729?
The short answer is: no. The long answer is: what goldfire was suggesting in comment:3 was that an intrepid GHC hacker could look at the innards of his GHC fork at https://github.com/goldfirere/ghc and, conceivably, adapt the logic he uses in `TcFlatten.flatten_args` to come up with an algorithm that would fix the problem in this ticket. (I had mistakenly though he meant that `liftCoSubst` //was// this algorithm, but in fact, it's only one component of it.) That being said, the performance of `TcFlatten.flatten_args` is apparently pretty bad, so adapting it in its current state might not be the wisest course of action. In light of this, I think I'll hold off on this idea for now and proceed with option (2) in comment:2 as a stopgap solution. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/14728#comment:9 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler