#16300: Make TH always reify data types with explicit return kinds -------------------------------------+------------------------------------- Reporter: RyanGlScott | Owner: (none) Type: feature request | Status: new Priority: normal | Milestone: Component: Template Haskell | Version: 8.6.3 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 RyanGlScott): Replying to [comment:4 goldfire]:
how about non-`Type`-kinded datatypes/newtypes always reify as GADTs? The choice of syntax in reification is immaterial, so let's just change it.
I can see the appeal behind this choice, but at the same time, I would be a little sad if we went in this direction. Generally speaking, I like keeping TH reification as honest as possible. Moreover, whether a data type is declared as a GADT or not has some consequences on its other properties. For instance, the derived `Show` instance for a data type with infix constructors changes depending on whether the type is a GADT, and unless consumers have a reliable way to reify whether or not a data type is a GADT, they wouldn't be able to replicate what `deriving Show` does with Template Haskell, which would be a bit sad. I really wish that we had top-level kind signatures, since then we could just reify `type Foo :: TYPE IntRep`, regardless of whether `Foo` was defined using Haskell98 or GADT syntax. I suppose we could just park this ticket until that becomes a reality. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/16300#comment:5 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler