#15828: Type family equation foralls allow strange re-quantification of class-bound type variables -------------------------------------+------------------------------------- Reporter: RyanGlScott | Owner: (none) Type: bug | Status: new Priority: normal | Milestone: 8.8.1 Component: Compiler | Version: 8.7 Resolution: | Keywords: TypeFamilies Operating System: Unknown/Multiple | Architecture: | Unknown/Multiple Type of failure: None/Unknown | Test Case: | rename/should_fail/T15828 Blocked By: | Blocking: Related Tickets: | Differential Rev(s): Phab:D5283 Wiki Page: | -------------------------------------+------------------------------------- Changes (by RyanGlScott): * cc: simonpj (added) * status: closed => new * resolution: fixed => Comment: Hm... I might have been a little too hasty in closing this ticket. This patch was undoubtedly an improvement over the status quo, as we now throw a proper error instead of silently accepting the original program. However, the error message leaves something to be desired: {{{ T15828.hs:9:3: error: • Type indexes must match class instance head Expected: T (Maybe a) _ Actual: T (Maybe a) b -- Defined at T15828.hs:9:20 • In the type instance declaration for ‘T’ In the instance declaration for ‘C (Maybe a)’ }}} Why did this fail? It's extremely hard to tell, but it's because the two `a` variables aren't the same! I suspect that we're missing some necessary tidying here. Simon, this behavior appears to be new as of your commit 2257a86daa72db382eb927df12a718669d5491f8 (`Taming the Kind Inference Monster`). In particular, this change looks suspicious: {{{#!diff + pp_expected_ty = pprIfaceType (toIfaceType (mkTyConApp fam_tc expected_args)) + -- Do /not/ tidy, because that will rename all those "_" + -- variables we have put in. And (I think) the intance type + -- is already tidy }}} -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/15828#comment:9 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler