{- (c) The University of Glasgow 2006 (c) The GRASP/AQUA Project, Glasgow University, 1992-1998 -} {-# LANGUAGE CPP #-} {-# LANGUAGE MultiWayIf #-} -- | Functions for inferring (and simplifying) the context for derived instances. module GHC.Tc.Deriv.Infer ( inferConstraints , simplifyInstanceContexts ) where #include "HsVersions.h" import GHC.Prelude import GHC.Data.Bag import GHC.Types.Basic import GHC.Core.Class import GHC.Core.DataCon import GHC.Utils.Error import GHC.Tc.Utils.Instantiate import GHC.Utils.Outputable import GHC.Utils.Panic import GHC.Data.Pair import GHC.Builtin.Names import GHC.Tc.Deriv.Utils import GHC.Tc.Utils.Env import GHC.Tc.Deriv.Generate import GHC.Tc.Deriv.Functor import GHC.Tc.Deriv.Generics import GHC.Tc.Utils.TcMType import GHC.Tc.Utils.Monad import GHC.Tc.Types.Origin import GHC.Tc.Types.Constraint import GHC.Core.Predicate import GHC.Tc.Utils.TcType import GHC.Core.TyCon import GHC.Core.TyCo.Ppr (pprTyVars) import GHC.Core.Type import GHC.Tc.Solver import GHC.Tc.Validity (validDerivPred) import GHC.Tc.Utils.Unify (buildImplicationFor, checkConstraints) import GHC.Builtin.Types (typeToTypeKind) import GHC.Core.Unify (tcUnifyTy) import GHC.Utils.Misc import GHC.Types.Var import GHC.Types.Var.Set import Control.Monad import Control.Monad.Trans.Class (lift) import Control.Monad.Trans.Reader (ask) import Data.List (sortBy) import Data.Maybe ---------------------- inferConstraints :: DerivSpecMechanism -> DerivM ([ThetaOrigin], [TyVar], [TcType]) -- inferConstraints figures out the constraints needed for the -- instance declaration generated by a 'deriving' clause on a -- data type declaration. It also returns the new in-scope type -- variables and instance types, in case they were changed due to -- the presence of functor-like constraints. -- See Note [Inferring the instance context] -- e.g. inferConstraints -- C Int (T [a]) -- Class and inst_tys -- :RTList a -- Rep tycon and its arg tys -- where T [a] ~R :RTList a -- -- Generate a sufficiently large set of constraints that typechecking the -- generated method definitions should succeed. This set will be simplified -- before being used in the instance declaration inferConstraints mechanism = do { DerivEnv { denv_tvs = tvs , denv_cls = main_cls , denv_inst_tys = inst_tys } <- ask ; wildcard <- isStandaloneWildcardDeriv ; let infer_constraints :: DerivM ([ThetaOrigin], [TyVar], [TcType]) infer_constraints = case mechanism of DerivSpecStock{dsm_stock_dit = dit} -> inferConstraintsStock dit DerivSpecAnyClass -> infer_constraints_simple inferConstraintsAnyclass DerivSpecNewtype { dsm_newtype_dit = DerivInstTys{dit_cls_tys = cls_tys} , dsm_newtype_rep_ty = rep_ty } -> infer_constraints_simple $ inferConstraintsCoerceBased cls_tys rep_ty DerivSpecVia { dsm_via_cls_tys = cls_tys , dsm_via_ty = via_ty } -> infer_constraints_simple $ inferConstraintsCoerceBased cls_tys via_ty -- Most deriving strategies do not need to do anything special to -- the type variables and arguments to the class in the derived -- instance, so they can pass through unchanged. The exception to -- this rule is stock deriving. See -- Note [Inferring the instance context]. infer_constraints_simple :: DerivM [ThetaOrigin] -> DerivM ([ThetaOrigin], [TyVar], [TcType]) infer_constraints_simple infer_thetas = do thetas <- infer_thetas pure (thetas, tvs, inst_tys) -- Constraints arising from superclasses -- See Note [Superclasses of derived instance] cls_tvs = classTyVars main_cls sc_constraints = ASSERT2( equalLength cls_tvs inst_tys , ppr main_cls <+> ppr inst_tys ) [ mkThetaOrigin (mkDerivOrigin wildcard) TypeLevel [] [] [] $ substTheta cls_subst (classSCTheta main_cls) ] cls_subst = ASSERT( equalLength cls_tvs inst_tys ) zipTvSubst cls_tvs inst_tys ; (inferred_constraints, tvs', inst_tys') <- infer_constraints ; lift $ traceTc "inferConstraints" $ vcat [ ppr main_cls <+> ppr inst_tys' , ppr inferred_constraints ] ; return ( sc_constraints ++ inferred_constraints , tvs', inst_tys' ) } -- | Like 'inferConstraints', but used only in the case of the @stock@ deriving -- strategy. The constraints are inferred by inspecting the fields of each data -- constructor. In this example: -- -- > data Foo = MkFoo Int Char deriving Show -- -- We would infer the following constraints ('ThetaOrigin's): -- -- > (Show Int, Show Char) -- -- Note that this function also returns the type variables ('TyVar's) and -- class arguments ('TcType's) for the resulting instance. This is because -- when deriving 'Functor'-like classes, we must sometimes perform kind -- substitutions to ensure the resulting instance is well kinded, which may -- affect the type variables and class arguments. In this example: -- -- > newtype Compose (f :: k -> Type) (g :: Type -> k) (a :: Type) = -- > Compose (f (g a)) deriving stock Functor -- -- We must unify @k@ with @Type@ in order for the resulting 'Functor' instance -- to be well kinded, so we return @[]@/@[Type, f, g]@ for the -- 'TyVar's/'TcType's, /not/ @[k]@/@[k, f, g]@. -- See Note [Inferring the instance context]. inferConstraintsStock :: DerivInstTys -> DerivM ([ThetaOrigin], [TyVar], [TcType]) inferConstraintsStock (DerivInstTys { dit_cls_tys = cls_tys , dit_tc = tc , dit_tc_args = tc_args , dit_rep_tc = rep_tc , dit_rep_tc_args = rep_tc_args }) = do DerivEnv { denv_tvs = tvs , denv_cls = main_cls , denv_inst_tys = inst_tys } <- ask wildcard <- isStandaloneWildcardDeriv let inst_ty = mkTyConApp tc tc_args tc_binders = tyConBinders rep_tc choose_level bndr | isNamedTyConBinder bndr = KindLevel | otherwise = TypeLevel t_or_ks = map choose_level tc_binders ++ repeat TypeLevel -- want to report *kind* errors when possible -- Constraints arising from the arguments of each constructor con_arg_constraints :: (CtOrigin -> TypeOrKind -> Type -> [([PredOrigin], Maybe TCvSubst)]) -> ([ThetaOrigin], [TyVar], [TcType]) con_arg_constraints get_arg_constraints = let (predss, mbSubsts) = unzip [ preds_and_mbSubst | data_con <- tyConDataCons rep_tc , (arg_n, arg_t_or_k, arg_ty) <- zip3 [1..] t_or_ks $ dataConInstOrigArgTys data_con all_rep_tc_args -- No constraints for unlifted types -- See Note [Deriving and unboxed types] , not (isUnliftedType (irrelevantMult arg_ty)) , let orig = DerivOriginDC data_con arg_n wildcard , preds_and_mbSubst <- get_arg_constraints orig arg_t_or_k (irrelevantMult arg_ty) ] preds = concat predss -- If the constraints require a subtype to be of kind -- (* -> *) (which is the case for functor-like -- constraints), then we explicitly unify the subtype's -- kinds with (* -> *). -- See Note [Inferring the instance context] subst = foldl' composeTCvSubst emptyTCvSubst (catMaybes mbSubsts) unmapped_tvs = filter (\v -> v `notElemTCvSubst` subst && not (v `isInScope` subst)) tvs (subst', _) = substTyVarBndrs subst unmapped_tvs preds' = map (substPredOrigin subst') preds inst_tys' = substTys subst' inst_tys tvs' = tyCoVarsOfTypesWellScoped inst_tys' in ([mkThetaOriginFromPreds preds'], tvs', inst_tys') is_generic = main_cls `hasKey` genClassKey is_generic1 = main_cls `hasKey` gen1ClassKey -- is_functor_like: see Note [Inferring the instance context] is_functor_like = tcTypeKind inst_ty `tcEqKind` typeToTypeKind || is_generic1 get_gen1_constraints :: Class -> CtOrigin -> TypeOrKind -> Type -> [([PredOrigin], Maybe TCvSubst)] get_gen1_constraints functor_cls orig t_or_k ty = mk_functor_like_constraints orig t_or_k functor_cls $ get_gen1_constrained_tys last_tv ty get_std_constrained_tys :: CtOrigin -> TypeOrKind -> Type -> [([PredOrigin], Maybe TCvSubst)] get_std_constrained_tys orig t_or_k ty | is_functor_like = mk_functor_like_constraints orig t_or_k main_cls $ deepSubtypesContaining last_tv ty | otherwise = [( [mk_cls_pred orig t_or_k main_cls ty] , Nothing )] mk_functor_like_constraints :: CtOrigin -> TypeOrKind -> Class -> [Type] -> [([PredOrigin], Maybe TCvSubst)] -- 'cls' is usually main_cls (Functor or Traversable etc), but if -- main_cls = Generic1, then 'cls' can be Functor; see -- get_gen1_constraints -- -- For each type, generate two constraints, -- [cls ty, kind(ty) ~ (*->*)], and a kind substitution that results -- from unifying kind(ty) with * -> *. If the unification is -- successful, it will ensure that the resulting instance is well -- kinded. If not, the second constraint will result in an error -- message which points out the kind mismatch. -- See Note [Inferring the instance context] mk_functor_like_constraints orig t_or_k cls = map $ \ty -> let ki = tcTypeKind ty in ( [ mk_cls_pred orig t_or_k cls ty , mkPredOrigin orig KindLevel (mkPrimEqPred ki typeToTypeKind) ] , tcUnifyTy ki typeToTypeKind ) rep_tc_tvs = tyConTyVars rep_tc last_tv = last rep_tc_tvs -- When we first gather up the constraints to solve, most of them -- contain rep_tc_tvs, i.e., the type variables from the derived -- datatype's type constructor. We don't want these type variables -- to appear in the final instance declaration, so we must -- substitute each type variable with its counterpart in the derived -- instance. rep_tc_args lists each of these counterpart types in -- the same order as the type variables. all_rep_tc_args = tyConInstArgTys rep_tc rep_tc_args -- Stupid constraints stupid_constraints = [ mkThetaOrigin deriv_origin TypeLevel [] [] [] $ substTheta tc_subst (tyConStupidTheta rep_tc) ] tc_subst = -- See the comment with all_rep_tc_args for an -- explanation of this assertion ASSERT( equalLength rep_tc_tvs all_rep_tc_args ) zipTvSubst rep_tc_tvs all_rep_tc_args -- Extra Data constraints -- The Data class (only) requires that for -- instance (...) => Data (T t1 t2) -- IF t1:*, t2:* -- THEN (Data t1, Data t2) are among the (...) constraints -- Reason: when the IF holds, we generate a method -- dataCast2 f = gcast2 f -- and we need the Data constraints to typecheck the method extra_constraints = [mkThetaOriginFromPreds constrs] where constrs | main_cls `hasKey` dataClassKey , all (isLiftedTypeKind . tcTypeKind) rep_tc_args = [ mk_cls_pred deriv_origin t_or_k main_cls ty | (t_or_k, ty) <- zip t_or_ks rep_tc_args] | otherwise = [] mk_cls_pred orig t_or_k cls ty -- Don't forget to apply to cls_tys' too = mkPredOrigin orig t_or_k (mkClassPred cls (cls_tys' ++ [ty])) cls_tys' | is_generic1 = [] -- In the awkward Generic1 case, cls_tys' should be -- empty, since we are applying the class Functor. | otherwise = cls_tys deriv_origin = mkDerivOrigin wildcard if -- Generic constraints are easy | is_generic -> return ([], tvs, inst_tys) -- Generic1 needs Functor -- See Note [Getting base classes] | is_generic1 -> ASSERT( rep_tc_tvs `lengthExceeds` 0 ) -- Generic1 has a single kind variable ASSERT( cls_tys `lengthIs` 1 ) do { functorClass <- lift $ tcLookupClass functorClassName ; pure $ con_arg_constraints $ get_gen1_constraints functorClass } -- The others are a bit more complicated | otherwise -> -- See the comment with all_rep_tc_args for an explanation of -- this assertion ASSERT2( equalLength rep_tc_tvs all_rep_tc_args , ppr main_cls <+> ppr rep_tc $$ ppr rep_tc_tvs $$ ppr all_rep_tc_args ) do { let (arg_constraints, tvs', inst_tys') = con_arg_constraints get_std_constrained_tys ; lift $ traceTc "inferConstraintsStock" $ vcat [ ppr main_cls <+> ppr inst_tys' , ppr arg_constraints ] ; return ( stupid_constraints ++ extra_constraints ++ arg_constraints , tvs', inst_tys') } -- | Like 'inferConstraints', but used only in the case of @DeriveAnyClass@, -- which gathers its constraints based on the type signatures of the class's -- methods instead of the types of the data constructor's field. -- -- See Note [Gathering and simplifying constraints for DeriveAnyClass] -- for an explanation of how these constraints are used to determine the -- derived instance context. inferConstraintsAnyclass :: DerivM [ThetaOrigin] inferConstraintsAnyclass = do { DerivEnv { denv_cls = cls , denv_inst_tys = inst_tys } <- ask ; wildcard <- isStandaloneWildcardDeriv ; let gen_dms = [ (sel_id, dm_ty) | (sel_id, Just (_, GenericDM dm_ty)) <- classOpItems cls ] cls_tvs = classTyVars cls do_one_meth :: (Id, Type) -> TcM ThetaOrigin -- (Id,Type) are the selector Id and the generic default method type -- NB: the latter is /not/ quantified over the class variables -- See Note [Gathering and simplifying constraints for DeriveAnyClass] do_one_meth (sel_id, gen_dm_ty) = do { let (sel_tvs, _cls_pred, meth_ty) = tcSplitMethodTy (varType sel_id) meth_ty' = substTyWith sel_tvs inst_tys meth_ty (meth_tvs, meth_theta, meth_tau) = tcSplitNestedSigmaTys meth_ty' gen_dm_ty' = substTyWith cls_tvs inst_tys gen_dm_ty (dm_tvs, dm_theta, dm_tau) = tcSplitNestedSigmaTys gen_dm_ty' tau_eq = mkPrimEqPred meth_tau dm_tau ; return (mkThetaOrigin (mkDerivOrigin wildcard) TypeLevel meth_tvs dm_tvs meth_theta (tau_eq:dm_theta)) } ; lift $ mapM do_one_meth gen_dms } -- Like 'inferConstraints', but used only for @GeneralizedNewtypeDeriving@ and -- @DerivingVia@. Since both strategies generate code involving 'coerce', the -- inferred constraints set up the scaffolding needed to typecheck those uses -- of 'coerce'. In this example: -- -- > newtype Age = MkAge Int deriving newtype Num -- -- We would infer the following constraints ('ThetaOrigin's): -- -- > (Num Int, Coercible Age Int) inferConstraintsCoerceBased :: [Type] -> Type -> DerivM [ThetaOrigin] inferConstraintsCoerceBased cls_tys rep_ty = do DerivEnv { denv_tvs = tvs , denv_cls = cls , denv_inst_tys = inst_tys } <- ask sa_wildcard <- isStandaloneWildcardDeriv let -- The following functions are polymorphic over the representation -- type, since we might either give it the underlying type of a -- newtype (for GeneralizedNewtypeDeriving) or a @via@ type -- (for DerivingVia). rep_tys ty = cls_tys ++ [ty] rep_pred ty = mkClassPred cls (rep_tys ty) rep_pred_o ty = mkPredOrigin deriv_origin TypeLevel (rep_pred ty) -- rep_pred is the representation dictionary, from where -- we are going to get all the methods for the final -- dictionary deriv_origin = mkDerivOrigin sa_wildcard -- Next we collect constraints for the class methods -- If there are no methods, we don't need any constraints -- Otherwise we need (C rep_ty), for the representation methods, -- and constraints to coerce each individual method meth_preds :: Type -> [PredOrigin] meth_preds ty | null meths = [] -- No methods => no constraints -- (#12814) | otherwise = rep_pred_o ty : coercible_constraints ty meths = classMethods cls coercible_constraints ty = [ mkPredOrigin (DerivOriginCoerce meth t1 t2 sa_wildcard) TypeLevel (mkReprPrimEqPred t1 t2) | meth <- meths , let (Pair t1 t2) = mkCoerceClassMethEqn cls tvs inst_tys ty meth ] all_thetas :: Type -> [ThetaOrigin] all_thetas ty = [mkThetaOriginFromPreds $ meth_preds ty] pure (all_thetas rep_ty) {- Note [Inferring the instance context] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ There are two sorts of 'deriving', as represented by the two constructors for DerivContext: * InferContext mb_wildcard: This can either be: - The deriving clause for a data type. (e.g, data T a = T1 a deriving( Eq )) In this case, mb_wildcard = Nothing. - A standalone declaration with an extra-constraints wildcard (e.g., deriving instance _ => Eq (Foo a)) In this case, mb_wildcard = Just loc, where loc is the location of the extra-constraints wildcard. Here we must infer an instance context, and generate instance declaration instance Eq a => Eq (T a) where ... * SupplyContext theta: standalone deriving deriving instance Eq a => Eq (T a) Here we only need to fill in the bindings; the instance context (theta) is user-supplied For the InferContext case, we must figure out the instance context (inferConstraintsStock). Suppose we are inferring the instance context for C t1 .. tn (T s1 .. sm) There are two cases * (T s1 .. sm) :: * (the normal case) Then we behave like Eq and guess (C t1 .. tn t) for each data constructor arg of type t. More details below. * (T s1 .. sm) :: * -> * (the functor-like case) Then we behave like Functor. In both cases we produce a bunch of un-simplified constraints and them simplify them in simplifyInstanceContexts; see Note [Simplifying the instance context]. In the functor-like case, we may need to unify some kind variables with * in order for the generated instance to be well-kinded. An example from #10524: newtype Compose (f :: k2 -> *) (g :: k1 -> k2) (a :: k1) = Compose (f (g a)) deriving Functor Earlier in the deriving pipeline, GHC unifies the kind of Compose f g (k1 -> *) with the kind of Functor's argument (* -> *), so k1 := *. But this alone isn't enough, since k2 wasn't unified with *: instance (Functor (f :: k2 -> *), Functor (g :: * -> k2)) => Functor (Compose f g) where ... The two Functor constraints are ill-kinded. To ensure this doesn't happen, we: 1. Collect all of a datatype's subtypes which require functor-like constraints. 2. For each subtype, create a substitution by unifying the subtype's kind with (* -> *). 3. Compose all the substitutions into one, then apply that substitution to all of the in-scope type variables and the instance types. Note [Getting base classes] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Functor and Typeable are defined in package 'base', and that is not available when compiling 'ghc-prim'. So we must be careful that 'deriving' for stuff in ghc-prim does not use Functor or Typeable implicitly via these lookups. Note [Deriving and unboxed types] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ We have some special hacks to support things like data T = MkT Int# deriving ( Show ) Specifically, we use GHC.Tc.Deriv.Generate.box to box the Int# into an Int (which we know how to show), and append a '#'. Parentheses are not required for unboxed values (`MkT -3#` is a valid expression). Note [Superclasses of derived instance] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ In general, a derived instance decl needs the superclasses of the derived class too. So if we have data T a = ...deriving( Ord ) then the initial context for Ord (T a) should include Eq (T a). Often this is redundant; we'll also generate an Ord constraint for each constructor argument, and that will probably generate enough constraints to make the Eq (T a) constraint be satisfied too. But not always; consider: data S a = S instance Eq (S a) instance Ord (S a) data T a = MkT (S a) deriving( Ord ) instance Num a => Eq (T a) The derived instance for (Ord (T a)) must have a (Num a) constraint! Similarly consider: data T a = MkT deriving( Data ) Here there *is* no argument field, but we must nevertheless generate a context for the Data instances: instance Typeable a => Data (T a) where ... ************************************************************************ * * Finding the fixed point of deriving equations * * ************************************************************************ Note [Simplifying the instance context] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Consider data T a b = C1 (Foo a) (Bar b) | C2 Int (T b a) | C3 (T a a) deriving (Eq) We want to come up with an instance declaration of the form instance (Ping a, Pong b, ...) => Eq (T a b) where x == y = ... It is pretty easy, albeit tedious, to fill in the code "...". The trick is to figure out what the context for the instance decl is, namely Ping, Pong and friends. Let's call the context reqd for the T instance of class C at types (a,b, ...) C (T a b). Thus: Eq (T a b) = (Ping a, Pong b, ...) Now we can get a (recursive) equation from the data decl. This part is done by inferConstraintsStock. Eq (T a b) = Eq (Foo a) u Eq (Bar b) -- From C1 u Eq (T b a) u Eq Int -- From C2 u Eq (T a a) -- From C3 Foo and Bar may have explicit instances for Eq, in which case we can just substitute for them. Alternatively, either or both may have their Eq instances given by deriving clauses, in which case they form part of the system of equations. Now all we need do is simplify and solve the equations, iterating to find the least fixpoint. This is done by simplifyInstanceConstraints. Notice that the order of the arguments can switch around, as here in the recursive calls to T. Let's suppose Eq (Foo a) = Eq a, and Eq (Bar b) = Ping b. We start with: Eq (T a b) = {} -- The empty set Next iteration: Eq (T a b) = Eq (Foo a) u Eq (Bar b) -- From C1 u Eq (T b a) u Eq Int -- From C2 u Eq (T a a) -- From C3 After simplification: = Eq a u Ping b u {} u {} u {} = Eq a u Ping b Next iteration: Eq (T a b) = Eq (Foo a) u Eq (Bar b) -- From C1 u Eq (T b a) u Eq Int -- From C2 u Eq (T a a) -- From C3 After simplification: = Eq a u Ping b u (Eq b u Ping a) u (Eq a u Ping a) = Eq a u Ping b u Eq b u Ping a The next iteration gives the same result, so this is the fixpoint. We need to make a canonical form of the RHS to ensure convergence. We do this by simplifying the RHS to a form in which - the classes constrain only tyvars - the list is sorted by tyvar (major key) and then class (minor key) - no duplicates, of course Note [Deterministic simplifyInstanceContexts] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Canonicalisation uses nonDetCmpType which is nondeterministic. Sorting with nonDetCmpType puts the returned lists in a nondeterministic order. If we were to return them, we'd get class constraints in nondeterministic order. Consider: data ADT a b = Z a b deriving Eq The generated code could be either: instance (Eq a, Eq b) => Eq (Z a b) where Or: instance (Eq b, Eq a) => Eq (Z a b) where To prevent the order from being nondeterministic we only canonicalize when comparing and return them in the same order as simplifyDeriv returned them. See also Note [nonDetCmpType nondeterminism] -} simplifyInstanceContexts :: [DerivSpec [ThetaOrigin]] -> TcM [DerivSpec ThetaType] -- Used only for deriving clauses or standalone deriving with an -- extra-constraints wildcard (InferContext) -- See Note [Simplifying the instance context] simplifyInstanceContexts [] = return [] simplifyInstanceContexts infer_specs = do { traceTc "simplifyInstanceContexts" $ vcat (map pprDerivSpec infer_specs) ; iterate_deriv 1 initial_solutions } where ------------------------------------------------------------------ -- The initial solutions for the equations claim that each -- instance has an empty context; this solution is certainly -- in canonical form. initial_solutions :: [ThetaType] initial_solutions = [ [] | _ <- infer_specs ] ------------------------------------------------------------------ -- iterate_deriv calculates the next batch of solutions, -- compares it with the current one; finishes if they are the -- same, otherwise recurses with the new solutions. -- It fails if any iteration fails iterate_deriv :: Int -> [ThetaType] -> TcM [DerivSpec ThetaType] iterate_deriv n current_solns | n > 20 -- Looks as if we are in an infinite loop -- This can happen if we have -XUndecidableInstances -- (See GHC.Tc.Solver.tcSimplifyDeriv.) = pprPanic "solveDerivEqns: probable loop" (vcat (map pprDerivSpec infer_specs) $$ ppr current_solns) | otherwise = do { -- Extend the inst info from the explicit instance decls -- with the current set of solutions, and simplify each RHS inst_specs <- zipWithM newDerivClsInst current_solns infer_specs ; new_solns <- checkNoErrs $ extendLocalInstEnv inst_specs $ mapM gen_soln infer_specs ; if (current_solns `eqSolution` new_solns) then return [ spec { ds_theta = soln } | (spec, soln) <- zip infer_specs current_solns ] else iterate_deriv (n+1) new_solns } eqSolution a b = eqListBy (eqListBy eqType) (canSolution a) (canSolution b) -- Canonicalise for comparison -- See Note [Deterministic simplifyInstanceContexts] canSolution = map (sortBy nonDetCmpType) ------------------------------------------------------------------ gen_soln :: DerivSpec [ThetaOrigin] -> TcM ThetaType gen_soln (DS { ds_loc = loc, ds_tvs = tyvars , ds_cls = clas, ds_tys = inst_tys, ds_theta = deriv_rhs }) = setSrcSpan loc $ addErrCtxt (derivInstCtxt the_pred) $ do { theta <- simplifyDeriv the_pred tyvars deriv_rhs -- checkValidInstance tyvars theta clas inst_tys -- Not necessary; see Note [Exotic derived instance contexts] ; traceTc "GHC.Tc.Deriv" (ppr deriv_rhs $$ ppr theta) -- Claim: the result instance declaration is guaranteed valid -- Hence no need to call: -- checkValidInstance tyvars theta clas inst_tys ; return theta } where the_pred = mkClassPred clas inst_tys derivInstCtxt :: PredType -> SDoc derivInstCtxt pred = text "When deriving the instance for" <+> parens (ppr pred) {- *********************************************************************************** * * * Simplify derived constraints * * *********************************************************************************** -} -- | Given @instance (wanted) => C inst_ty@, simplify 'wanted' as much -- as possible. Fail if not possible. simplifyDeriv :: PredType -- ^ @C inst_ty@, head of the instance we are -- deriving. Only used for SkolemInfo. -> [TyVar] -- ^ The tyvars bound by @inst_ty@. -> [ThetaOrigin] -- ^ Given and wanted constraints -> TcM ThetaType -- ^ Needed constraints (after simplification), -- i.e. @['PredType']@. simplifyDeriv pred tvs thetas = do { (skol_subst, tvs_skols) <- tcInstSkolTyVars tvs -- Skolemize -- The constraint solving machinery -- expects *TcTyVars* not TyVars. -- We use *non-overlappable* (vanilla) skolems -- See Note [Overlap and deriving] ; let skol_set = mkVarSet tvs_skols skol_info = DerivSkol pred doc = text "deriving" <+> parens (ppr pred) mk_given_ev :: PredType -> TcM EvVar mk_given_ev given = let given_pred = substTy skol_subst given in newEvVar given_pred emit_wanted_constraints :: [TyVar] -> [PredOrigin] -> TcM () emit_wanted_constraints metas_to_be preds = do { -- We instantiate metas_to_be with fresh meta type -- variables. Currently, these can only be type variables -- quantified in generic default type signatures. -- See Note [Gathering and simplifying constraints for -- DeriveAnyClass] (meta_subst, _meta_tvs) <- newMetaTyVars metas_to_be -- Now make a constraint for each of the instantiated predicates ; let wanted_subst = skol_subst `unionTCvSubst` meta_subst mk_wanted_ct (PredOrigin wanted orig t_or_k) = do { ev <- newWanted orig (Just t_or_k) $ substTyUnchecked wanted_subst wanted ; return (mkNonCanonical ev) } ; cts <- mapM mk_wanted_ct preds -- And emit them into the monad ; emitSimples (listToCts cts) } -- Create the implications we need to solve. For stock and newtype -- deriving, these implication constraints will be simple class -- constraints like (C a, Ord b). -- But with DeriveAnyClass, we make an implication constraint. -- See Note [Gathering and simplifying constraints for DeriveAnyClass] mk_wanteds :: ThetaOrigin -> TcM WantedConstraints mk_wanteds (ThetaOrigin { to_anyclass_skols = ac_skols , to_anyclass_metas = ac_metas , to_anyclass_givens = ac_givens , to_wanted_origins = preds }) = do { ac_given_evs <- mapM mk_given_ev ac_givens ; (_, wanteds) <- captureConstraints $ checkConstraints skol_info ac_skols ac_given_evs $ -- The checkConstraints bumps the TcLevel, and -- wraps the wanted constraints in an implication, -- when (but only when) necessary emit_wanted_constraints ac_metas preds ; pure wanteds } -- See [STEP DAC BUILD] -- Generate the implication constraints, one for each method, to solve -- with the skolemized variables. Start "one level down" because -- we are going to wrap the result in an implication with tvs_skols, -- in step [DAC RESIDUAL] ; (tc_lvl, wanteds) <- pushTcLevelM $ mapM mk_wanteds thetas ; traceTc "simplifyDeriv inputs" $ vcat [ pprTyVars tvs $$ ppr thetas $$ ppr wanteds, doc ] -- See [STEP DAC SOLVE] -- Simplify the constraints, starting at the same level at which -- they are generated (c.f. the call to runTcSWithEvBinds in -- simplifyInfer) ; solved_wanteds <- setTcLevel tc_lvl $ runTcSDeriveds $ solveWantedsAndDrop $ unionsWC wanteds -- It's not yet zonked! Obviously zonk it before peering at it ; solved_wanteds <- zonkWC solved_wanteds -- See [STEP DAC HOIST] -- From the simplified constraints extract a subset 'good' that will -- become the context 'min_theta' for the derived instance. ; let residual_simple = approximateWC True solved_wanteds good = mapMaybeBag get_good residual_simple -- Returns @Just p@ (where @p@ is the type of the Ct) if a Ct is -- suitable to be inferred in the context of a derived instance. -- Returns @Nothing@ if the Ct is too exotic. -- See Note [Exotic derived instance contexts] for what -- constitutes an exotic constraint. get_good :: Ct -> Maybe PredType get_good ct | validDerivPred skol_set p , isWantedCt ct = Just p -- TODO: This is wrong -- NB re 'isWantedCt': residual_wanted may contain -- unsolved CtDerived and we stick them into the -- bad set so that reportUnsolved may decide what -- to do with them | otherwise = Nothing where p = ctPred ct ; traceTc "simplifyDeriv outputs" $ vcat [ ppr tvs_skols, ppr residual_simple, ppr good ] -- Return the good unsolved constraints (unskolemizing on the way out.) ; let min_theta = mkMinimalBySCs id (bagToList good) -- An important property of mkMinimalBySCs (used above) is that in -- addition to removing constraints that are made redundant by -- superclass relationships, it also removes _duplicate_ -- constraints. -- See Note [Gathering and simplifying constraints for -- DeriveAnyClass] subst_skol = zipTvSubst tvs_skols $ mkTyVarTys tvs -- The reverse substitution (sigh) -- See [STEP DAC RESIDUAL] -- Ensure that min_theta is enough to solve /all/ the constraints in -- solved_wanteds, by solving the implication constraint -- -- forall tvs. min_theta => solved_wanteds ; min_theta_vars <- mapM newEvVar min_theta ; (leftover_implic, _) <- buildImplicationFor tc_lvl skol_info tvs_skols min_theta_vars solved_wanteds -- This call to simplifyTop is purely for error reporting -- See Note [Error reporting for deriving clauses] -- See also Note [Exotic derived instance contexts], which are caught -- in this line of code. ; simplifyTopImplic leftover_implic ; return (substTheta subst_skol min_theta) } {- Note [Overlap and deriving] ~~~~~~~~~~~~~~~~~~~~~~~~~~~ Consider some overlapping instances: instance Show a => Show [a] where .. instance Show [Char] where ... Now a data type with deriving: data T a = MkT [a] deriving( Show ) We want to get the derived instance instance Show [a] => Show (T a) where... and NOT instance Show a => Show (T a) where... so that the (Show (T Char)) instance does the Right Thing It's very like the situation when we're inferring the type of a function f x = show [x] and we want to infer f :: Show [a] => a -> String BOTTOM LINE: use vanilla, non-overlappable skolems when inferring the context for the derived instance. Hence tcInstSkolTyVars not tcInstSuperSkolTyVars Note [Gathering and simplifying constraints for DeriveAnyClass] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ DeriveAnyClass works quite differently from stock and newtype deriving in the way it gathers and simplifies constraints to be used in a derived instance's context. Stock and newtype deriving gather constraints by looking at the data constructors of the data type for which we are deriving an instance. But DeriveAnyClass doesn't need to know about a data type's definition at all! To see why, consider this example of DeriveAnyClass: class Foo a where bar :: forall b. Ix b => a -> b -> String default bar :: (Show a, Ix c) => a -> c -> String bar x y = show x ++ show (range (y,y)) baz :: Eq a => a -> a -> Bool default baz :: (Ord a, Show a) => a -> a -> Bool baz x y = compare x y == EQ Because 'bar' and 'baz' have default signatures, this generates a top-level definition for these generic default methods $gdm_bar :: forall a. Foo a => forall c. (Show a, Ix c) => a -> c -> String $gdm_bar x y = show x ++ show (range (y,y)) (and similarly for baz). Now consider a 'deriving' clause data Maybe s = ... deriving Foo This derives an instance of the form: instance (CX) => Foo (Maybe s) where bar = $gdm_bar baz = $gdm_baz Now it is GHC's job to fill in a suitable instance context (CX). If GHC were typechecking the binding bar = $gdm bar it would * skolemise the expected type of bar * instantiate the type of $gdm_bar with meta-type variables * build an implication constraint [STEP DAC BUILD] So that's what we do. We build the constraint (call it C1) forall[2] b. Ix b => (Show (Maybe s), Ix cc, Maybe s -> b -> String ~ Maybe s -> cc -> String) Here: * The level of this forall constraint is forall[2], because we are later going to wrap it in a forall[1] in [STEP DAC RESIDUAL] * The 'b' comes from the quantified type variable in the expected type of bar (i.e., 'to_anyclass_skols' in 'ThetaOrigin'). The 'cc' is a unification variable that comes from instantiating the quantified type variable 'c' in $gdm_bar's type (i.e., 'to_anyclass_metas' in 'ThetaOrigin). * The (Ix b) constraint comes from the context of bar's type (i.e., 'to_wanted_givens' in 'ThetaOrigin'). The (Show (Maybe s)) and (Ix cc) constraints come from the context of $gdm_bar's type (i.e., 'to_anyclass_givens' in 'ThetaOrigin'). * The equality constraint (Maybe s -> b -> String) ~ (Maybe s -> cc -> String) comes from marrying up the instantiated type of $gdm_bar with the specified type of bar. Notice that the type variables from the instance, 's' in this case, are global to this constraint. Note that it is vital that we instantiate the `c` in $gdm_bar's type with a new unification variable for each iteration of simplifyDeriv. If we re-use the same unification variable across multiple iterations, then bad things can happen, such as #14933. Similarly for 'baz', giving the constraint C2 forall[2]. Eq (Maybe s) => (Ord a, Show a, Maybe s -> Maybe s -> Bool ~ Maybe s -> Maybe s -> Bool) In this case baz has no local quantification, so the implication constraint has no local skolems and there are no unification variables. [STEP DAC SOLVE] We can combine these two implication constraints into a single constraint (C1, C2), and simplify, unifying cc:=b, to get: forall[2] b. Ix b => Show a /\ forall[2]. Eq (Maybe s) => (Ord a, Show a) [STEP DAC HOIST] Let's call that (C1', C2'). Now we need to hoist the unsolved constraints out of the implications to become our candidate for (CX). That is done by approximateWC, which will return: (Show a, Ord a, Show a) Now we can use mkMinimalBySCs to remove superclasses and duplicates, giving (Show a, Ord a) And that's what GHC uses for CX. [STEP DAC RESIDUAL] In this case we have solved all the leftover constraints, but what if we don't? Simple! We just form the final residual constraint forall[1] s. CX => (C1',C2') and simplify that. In simple cases it'll succeed easily, because CX literally contains the constraints in C1', C2', but if there is anything more complicated it will be reported in a civilised way. Note [Error reporting for deriving clauses] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ A surprisingly tricky aspect of deriving to get right is reporting sensible error messages. In particular, if simplifyDeriv reaches a constraint that it cannot solve, which might include: 1. Insoluble constraints 2. "Exotic" constraints (See Note [Exotic derived instance contexts]) Then we report an error immediately in simplifyDeriv. Another possible choice is to punt and let another part of the typechecker (e.g., simplifyInstanceContexts) catch the errors. But this tends to lead to worse error messages, so we do it directly in simplifyDeriv. simplifyDeriv checks for errors in a clever way. If the deriving machinery infers the context (Foo a)--that is, if this instance is to be generated: instance Foo a => ... Then we form an implication of the form: forall a. Foo a => <residual_wanted_constraints> And pass it to the simplifier. If the context (Foo a) is enough to discharge all the constraints in <residual_wanted_constraints>, then everything is hunky-dory. But if <residual_wanted_constraints> contains, say, an insoluble constraint, then (Foo a) won't be able to solve it, causing GHC to error. Note [Exotic derived instance contexts] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ In a 'derived' instance declaration, we *infer* the context. It's a bit unclear what rules we should apply for this; the Haskell report is silent. Obviously, constraints like (Eq a) are fine, but what about data T f a = MkT (f a) deriving( Eq ) where we'd get an Eq (f a) constraint. That's probably fine too. One could go further: consider data T a b c = MkT (Foo a b c) deriving( Eq ) instance (C Int a, Eq b, Eq c) => Eq (Foo a b c) Notice that this instance (just) satisfies the Paterson termination conditions. Then we *could* derive an instance decl like this: instance (C Int a, Eq b, Eq c) => Eq (T a b c) even though there is no instance for (C Int a), because there just *might* be an instance for, say, (C Int Bool) at a site where we need the equality instance for T's. However, this seems pretty exotic, and it's quite tricky to allow this, and yet give sensible error messages in the (much more common) case where we really want that instance decl for C. So for now we simply require that the derived instance context should have only type-variable constraints. Here is another example: data Fix f = In (f (Fix f)) deriving( Eq ) Here, if we are prepared to allow -XUndecidableInstances we could derive the instance instance Eq (f (Fix f)) => Eq (Fix f) but this is so delicate that I don't think it should happen inside 'deriving'. If you want this, write it yourself! NB: if you want to lift this condition, make sure you still meet the termination conditions! If not, the deriving mechanism generates larger and larger constraints. Example: data Succ a = S a data Seq a = Cons a (Seq (Succ a)) | Nil deriving Show Note the lack of a Show instance for Succ. First we'll generate instance (Show (Succ a), Show a) => Show (Seq a) and then instance (Show (Succ (Succ a)), Show (Succ a), Show a) => Show (Seq a) and so on. Instead we want to complain of no instance for (Show (Succ a)). The bottom line ~~~~~~~~~~~~~~~ Allow constraints which consist only of type variables, with no repeats. -}