#7015: Add support for 'static' -------------------------------------+------------------------------------- Reporter: edsko | Owner: Type: feature | Status: patch request | Milestone: 7.10.1 Priority: normal | Version: 7.4.2 Component: Compiler | Keywords: Resolution: | Architecture: Unknown/Multiple Operating System: | Difficulty: Unknown Unknown/Multiple | Blocked By: Type of failure: | Related Tickets: None/Unknown | Test Case: | Blocking: | Differential Revisions: Phab:D119 | -------------------------------------+------------------------------------- Comment (by simonpj): Edsko my remark about a "hack" was not intended as a criticism. Of course, if GHC is inadequate you must hack around its shortcomings. My intent was solely this: let's not enshrine a temporary hack into the design of the language that GHC implements. There are two things going on in this ticket: * What should be the specification of the feature that GHC implements? I think we are agreeing that: the argument of `static` should have no free type variables. Let's put that into the spec.[[BR]][[BR]] That leaves open the question of adding an impredicative-like feature, the ability to have terms of type `Static (forall a. [a] -> [a])`, which make perfect sense, have no free type variables, but which are disallowed at present. * Given the above restriction, and the absence of the impredicative support, how can we hack around the restriction to get something like polymorphism. This is the `Any` discussion on #9429. Even lacking a fix for poly-kinded `Any`, it's not a disaster because for any particular kind we can always do the job; it's just a bit inelegant. If we agree about these things, perhaps you can update the wiki page as above and we can go from there. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/7015#comment:26 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler