#13305: static: check for identifiers should only consider term level variables -------------------------------------+------------------------------------- Reporter: edsko | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 8.0.2 Resolution: | Keywords: 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 simonpj): It is entirely unclear to me whether we should allow polymorphic static pointers. We certainly do not allow {{{ foo1 :: forall a. Num a => StaticPtr (a->a->a) foo1 = static (+) }}} so why should we allow `foo` in the `Description`? I think of the Static Pointer Table as containing `Dynamic`s: a pair of a value (just a code pointer) and a `TypeRep` describing its type. Now when we deserialise a static pointer we can (and jolly well should) do a dynamic type check. It might be possible to go further: see "Parametric polymorphism" in [wiki:StaticPointers]. But we have not even begun to implement it. So while the error message is not good, I'd like to rule out polymorphic static pointers altogether. I know that Facundo has use-cases where he wants polymorphism, but we should discuss those use-cases properly. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/13305#comment:2 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler