#15359: Quantified constraints do not work with equality constraints -------------------------------------+------------------------------------- Reporter: goldfire | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: 8.6.1 Component: Compiler | Version: 8.5 Resolution: | Keywords: | QuantifiedConstraints 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 AntC): Replying to [comment:1 RyanGlScott]:
It's very strange to me that we allow quantified `Coercible` constraints but not quantified `(~)` constraints. What is the rationale behind this decision choice?
Yes, I'm gravely disappointed to learn that: I've been posting lots of suggestions using `~` inside quantified (implication) constraints. {{{ • Class ‘~’ does not support user-specified instances }}} Before there was `~`, there were plenty of ways for a user to specify a unifiability constraint -- see for example `TypeCast` in the HList paper. (Using bi-directional FunDeps.) And other examples of Quantified Constraints are using FunDeps for type improvement. So just do that instead of `~`? -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/15359#comment:3 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler