#13018: TH-spliced pattern synonym declaration fails to typecheck -------------------------------------+------------------------------------- Reporter: RyanGlScott | Owner: Type: bug | Status: new Priority: normal | Milestone: Component: Compiler (Type | Version: 8.0.1 checker) | Keywords: Resolution: | PatternSynonyms Operating System: Unknown/Multiple | Architecture: Type of failure: GHC rejects | Unknown/Multiple valid program | Test Case: Blocked By: | Blocking: Related Tickets: | Differential Rev(s): Wiki Page: | -------------------------------------+------------------------------------- Comment (by simonpj):
Agreed that the splitting rule must be implemented in the TH quoting mechanism.
I do not agree. The business of TH is to reflect Haskell ''source'' code. In this case the Haskell source code signature is {{{ pattern P :: b -> T a }}} That generates a `LHsSigType` for the type. '''We should be able to round- trip that `LHsSigType` through TH and get back the exact same thing'''. If we do that then, because it's the exact same thing, if it typechecked before it'll typecheck after the round trip. What is happening, I think, is that we are getting back a `LHsSigType` with more explicit quantification that the original one had. That's bad! We should get back the exact same thing. I haven't dug deeper, but that must be possible, right? I think it's the ''wrong'' thing for the TH (syntax-level) conversions to attempt to figure out what is quantified where. Just reproduce what the user wrote! Simon -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/13018#comment:12 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler