#8707: Kind inference fails in data instance definition -------------------------------------+------------------------------------ Reporter: goldfire | Owner: jstolarek Type: bug | Status: new Priority: normal | Milestone: Component: Compiler | Version: 7.7 Resolution: | Keywords: Operating System: Unknown/Multiple | Architecture: Unknown/Multiple Type of failure: None/Unknown | Difficulty: Unknown Test Case: | Blocked By: Blocking: | Related Tickets: -------------------------------------+------------------------------------ Comment (by jstolarek): Thanks for all the detail Simon. I've been looking at this for the last two days trying to make sense of what is going on. I have a few newbie questions: 1. Kind checking type patterns (`TcTyClsDecls.tcFamTyPat` function called from `TcInstDcls.tcDataFamInstDecl`). Does the term "type patterns" refer to type parameters of a data family? 2. From looking at `traceTc` output I see that kind checking type patterns includes type checking (and also inferring?) type of `SFalse` constructor and its components (`False` and `Ctor`). Then GHC typechecks constructor declarations (`TcTyClsDecls.tcConDecls` called from `TcInstDcls.tcDataFamInstDecl`) where it again typechecks `SFalse` and its components. Why is this done twice? I believe I can implement solution to this ticket if someone more experienced helps me to figure out what is the right thing to do here.
Then my question remains: what's wrong with the status quo, for question (2)?
I believe that Richard's point was that GHC 7.6.3 displayed kind parameters, which made it easier for the user to deduce what went wrong (ie. seeing incorrect `Bool (k -> *)` kind params instead of correct `Bool (Bool -> *)`. -- Ticket URL: http://ghc.haskell.org/trac/ghc/ticket/8707#comment:8 GHC http://www.haskell.org/ghc/ The Glasgow Haskell Compiler