#13549: GHC 8.2.1's typechecker rejects code generated by singletons that 8.0 accepts -------------------------------------+------------------------------------- Reporter: RyanGlScott | Owner: (none) Type: bug | Status: closed Priority: normal | Milestone: 8.4.1 Component: Compiler (Type | Version: 8.1 checker) | Resolution: duplicate | Keywords: TypeInType 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):
This is just fine for type inference, but it's bad for kind inference.
I don't understand this. With `MonoLocalBinds` we don't generalise and all is well, no?
So, here is the characterization of the infelicity in singletons: if a local function's type would be different depending on the presence of MonoLocalBinds, then it needs a type signature in order for singletons to work.
I don't understand this either. Could you give a small example? Looking at the `Bug3.hs` it looks as if there ''is'' a type signature, generated by Template Haskell. Perhaps you are saying that this type signature is insufficiently kind-generalised? In which case shouldn't the TH code simply add the necessary kind generalisation?
I'm going to close the ticket as a dup of #13555.
But #13555's conclusion was "not a bug". So is this not a bug goo? If so what about the singletons library? Very confused! -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/13549#comment:14 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler