{-# LANGUAGE BangPatterns #-} {-# LANGUAGE CPP #-} {-# LANGUAGE GADTs #-} {-# LANGUAGE MagicHash #-} #if __GLASGOW_HASKELL__ >= 800 {-# LANGUAGE TemplateHaskellQuotes #-} #else {-# LANGUAGE TemplateHaskell #-} #endif {-| Module: Data.Deriving.Internal Copyright: (C) 2015-2017 Ryan Scott License: BSD-style (see the file LICENSE) Maintainer: Ryan Scott Portability: Template Haskell Template Haskell-related utilities. Note: this is an internal module, and as such, the API presented here is not guaranteed to be stable, even between minor releases of this library. -} module Data.Deriving.Internal where import qualified Control.Applicative as App import Control.Monad (when, unless) import qualified Data.Foldable as F import Data.Functor.Classes ( Eq1(..), Ord1(..), Read1(..), Show1(..) #if MIN_VERSION_base(4,10,0) , liftReadListPrecDefault #endif ) #if !(MIN_VERSION_transformers(0,4,0)) || MIN_VERSION_transformers(0,5,0) import Data.Functor.Classes ( Eq2(..), Ord2(..), Read2(..), Show2(..) #if MIN_VERSION_base(4,10,0) , liftReadListPrec2Default #endif ) #endif import qualified Data.List as List import Data.List.NonEmpty (NonEmpty(..)) import qualified Data.Map as Map import Data.Map (Map) import Data.Maybe import Data.Monoid (Dual(..), Endo(..)) import qualified Data.Set as Set import Data.Set (Set) import qualified Data.Traversable as T import GHC.Arr (Ix(..)) import GHC.Base (getTag) import GHC.Exts import GHC.Read (choose, list, paren) import GHC.Show (showSpace) #if MIN_VERSION_base(4,19,0) import GHC.Int (Int8(..), Int16(..), Int32(..), Int64(..)) import GHC.Word (Word8(..), Word16(..), Word32(..), Word64(..)) #endif import Text.ParserCombinators.ReadPrec ( ReadPrec, (+++), pfail, prec, readPrec_to_S, readS_to_Prec , reset, step ) import Text.Read (Read(..), parens, readListPrecDefault) import qualified Text.Read.Lex as L import Text.Show (showListWith) #if MIN_VERSION_base(4,7,0) import GHC.Read (expectP) #else import GHC.Read (lexP) import Text.Read.Lex (Lexeme) #endif #if !(MIN_VERSION_base(4,8,0)) import Control.Applicative (Applicative(..)) import Data.Foldable (Foldable(..)) import Data.Functor (Functor(..)) import Data.Monoid (Monoid(..)) import Data.Traversable (Traversable(..)) #endif #if MIN_VERSION_base(4,10,0) import GHC.Show (showCommaSpace) #endif #if MIN_VERSION_base(4,11,0) import GHC.Read (readField, readSymField) #endif #if defined(MIN_VERSION_ghc_boot_th) import GHC.Lexeme (startsConSym, startsVarSym) #else import Data.Char (isSymbol, ord) #endif import Language.Haskell.TH.Datatype as Datatype import Language.Haskell.TH.Datatype.TyVarBndr import Language.Haskell.TH.Lib import Language.Haskell.TH.Ppr (pprint) import Language.Haskell.TH.Syntax -- Ensure, beyond a shadow of a doubt, that the instances are in-scope import Data.Functor () import Data.Functor.Classes () import Data.Foldable () import Data.Traversable () ------------------------------------------------------------------------------- -- Expanding type synonyms ------------------------------------------------------------------------------- applySubstitutionKind :: Map Name Kind -> Type -> Type #if MIN_VERSION_template_haskell(2,8,0) applySubstitutionKind = applySubstitution #else applySubstitutionKind _ t = t #endif substNameWithKind :: Name -> Kind -> Type -> Type substNameWithKind n k = applySubstitutionKind (Map.singleton n k) substNamesWithKindStar :: [Name] -> Type -> Type substNamesWithKindStar ns t = F.foldr' (flip substNameWithKind starK) t ns ------------------------------------------------------------------------------- -- Via ------------------------------------------------------------------------------- -- | A type-level modifier intended to be used in conjunction with 'deriveVia'. -- Refer to the documentation for 'deriveVia' for more details. data a `Via` b infix 0 `Via` ------------------------------------------------------------------------------- -- Type-specialized const functions ------------------------------------------------------------------------------- fmapConst :: f b -> (a -> b) -> f a -> f b fmapConst x _ _ = x {-# INLINE fmapConst #-} replaceConst :: f a -> a -> f b -> f a replaceConst x _ _ = x {-# INLINE replaceConst #-} foldrConst :: b -> (a -> b -> b) -> b -> t a -> b foldrConst x _ _ _ = x {-# INLINE foldrConst #-} foldMapConst :: m -> (a -> m) -> t a -> m foldMapConst x _ _ = x {-# INLINE foldMapConst #-} nullConst :: Bool -> t a -> Bool nullConst x _ = x {-# INLINE nullConst #-} traverseConst :: f (t b) -> (a -> f b) -> t a -> f (t b) traverseConst x _ _ = x {-# INLINE traverseConst #-} eqConst :: Bool -> a -> a -> Bool eqConst x _ _ = x {-# INLINE eqConst #-} eq1Const :: Bool -> f a -> f a-> Bool eq1Const x _ _ = x {-# INLINE eq1Const #-} liftEqConst :: Bool -> (a -> b -> Bool) -> f a -> f b -> Bool liftEqConst x _ _ _ = x {-# INLINE liftEqConst #-} liftEq2Const :: Bool -> (a -> b -> Bool) -> (c -> d -> Bool) -> f a c -> f b d -> Bool liftEq2Const x _ _ _ _ = x {-# INLINE liftEq2Const #-} compareConst :: Ordering -> a -> a -> Ordering compareConst x _ _ = x {-# INLINE compareConst #-} ltConst :: Bool -> a -> a -> Bool ltConst x _ _ = x {-# INLINE ltConst #-} compare1Const :: Ordering -> f a -> f a -> Ordering compare1Const x _ _ = x {-# INLINE compare1Const #-} liftCompareConst :: Ordering -> (a -> b -> Ordering) -> f a -> f b -> Ordering liftCompareConst x _ _ _ = x {-# INLINE liftCompareConst #-} liftCompare2Const :: Ordering -> (a -> b -> Ordering) -> (c -> d -> Ordering) -> f a c -> f b d -> Ordering liftCompare2Const x _ _ _ _ = x {-# INLINE liftCompare2Const #-} readsPrecConst :: ReadS a -> Int -> ReadS a readsPrecConst x _ = x {-# INLINE readsPrecConst #-} -- This isn't really necessary, but it makes for an easier implementation readPrecConst :: ReadPrec a -> ReadPrec a readPrecConst x = x {-# INLINE readPrecConst #-} readsPrec1Const :: ReadS (f a) -> Int -> ReadS (f a) readsPrec1Const x _ = x {-# INLINE readsPrec1Const #-} liftReadsPrecConst :: ReadS (f a) -> (Int -> ReadS a) -> ReadS [a] -> Int -> ReadS (f a) liftReadsPrecConst x _ _ _ = x {-# INLINE liftReadsPrecConst #-} liftReadPrecConst :: ReadPrec (f a) -> ReadPrec a -> ReadPrec [a] -> ReadPrec (f a) liftReadPrecConst x _ _ = x {-# INLINE liftReadPrecConst #-} liftReadsPrec2Const :: ReadS (f a b) -> (Int -> ReadS a) -> ReadS [a] -> (Int -> ReadS b) -> ReadS [b] -> Int -> ReadS (f a b) liftReadsPrec2Const x _ _ _ _ _ = x {-# INLINE liftReadsPrec2Const #-} liftReadPrec2Const :: ReadPrec (f a b) -> ReadPrec a -> ReadPrec [a] -> ReadPrec b -> ReadPrec [b] -> ReadPrec (f a b) liftReadPrec2Const x _ _ _ _ = x {-# INLINE liftReadPrec2Const #-} showsPrecConst :: ShowS -> Int -> a -> ShowS showsPrecConst x _ _ = x {-# INLINE showsPrecConst #-} showsPrec1Const :: ShowS -> Int -> f a -> ShowS showsPrec1Const x _ _ = x {-# INLINE showsPrec1Const #-} liftShowsPrecConst :: ShowS -> (Int -> a -> ShowS) -> ([a] -> ShowS) -> Int -> f a -> ShowS liftShowsPrecConst x _ _ _ _ = x {-# INLINE liftShowsPrecConst #-} liftShowsPrec2Const :: ShowS -> (Int -> a -> ShowS) -> ([a] -> ShowS) -> (Int -> b -> ShowS) -> ([b] -> ShowS) -> Int -> f a b -> ShowS liftShowsPrec2Const x _ _ _ _ _ _ = x {-# INLINE liftShowsPrec2Const #-} ------------------------------------------------------------------------------- -- StarKindStatus ------------------------------------------------------------------------------- -- | Whether a type is not of kind *, is of kind *, or is a kind variable. data StarKindStatus = NotKindStar | KindStar | IsKindVar Name deriving Eq -- | Does a Type have kind * or k (for some kind variable k)? canRealizeKindStar :: Type -> StarKindStatus canRealizeKindStar t | hasKindStar t = KindStar | otherwise = case t of #if MIN_VERSION_template_haskell(2,8,0) SigT _ (VarT k) -> IsKindVar k #endif _ -> NotKindStar -- | Returns 'Just' the kind variable 'Name' of a 'StarKindStatus' if it exists. -- Otherwise, returns 'Nothing'. starKindStatusToName :: StarKindStatus -> Maybe Name starKindStatusToName (IsKindVar n) = Just n starKindStatusToName _ = Nothing -- | Concat together all of the StarKindStatuses that are IsKindVar and extract -- the kind variables' Names out. catKindVarNames :: [StarKindStatus] -> [Name] catKindVarNames = mapMaybe starKindStatusToName ------------------------------------------------------------------------------- -- ClassRep ------------------------------------------------------------------------------- class ClassRep a where arity :: a -> Int allowExQuant :: a -> Bool fullClassName :: a -> Name classConstraint :: a -> Int -> Maybe Name ------------------------------------------------------------------------------- -- Template Haskell reifying and AST manipulation ------------------------------------------------------------------------------- -- For the given Types, generate an instance context and head. Coming up with -- the instance type isn't as simple as dropping the last types, as you need to -- be wary of kinds being instantiated with *. -- See Note [Type inference in derived instances] buildTypeInstance :: ClassRep a => a -- ^ The typeclass for which an instance should be derived -> Name -- ^ The type constructor or data family name -> Cxt -- ^ The datatype context -> [Type] -- ^ The types to instantiate the instance with -> DatatypeVariant -- ^ Are we dealing with a data family instance or not -> Q (Cxt, Type) buildTypeInstance cRep tyConName dataCxt varTysOrig variant = do -- Make sure to expand through type/kind synonyms! Otherwise, the -- eta-reduction check might get tripped up over type variables in a -- synonym that are actually dropped. -- (See GHC Trac #11416 for a scenario where this actually happened.) varTysExp <- T.mapM resolveTypeSynonyms varTysOrig let remainingLength :: Int remainingLength = length varTysOrig - arity cRep droppedTysExp :: [Type] droppedTysExp = drop remainingLength varTysExp droppedStarKindStati :: [StarKindStatus] droppedStarKindStati = map canRealizeKindStar droppedTysExp -- Check there are enough types to drop and that all of them are either of -- kind * or kind k (for some kind variable k). If not, throw an error. when (remainingLength < 0 || any (== NotKindStar) droppedStarKindStati) $ derivingKindError cRep tyConName let droppedKindVarNames :: [Name] droppedKindVarNames = catKindVarNames droppedStarKindStati -- Substitute kind * for any dropped kind variables varTysExpSubst :: [Type] varTysExpSubst = map (substNamesWithKindStar droppedKindVarNames) varTysExp remainingTysExpSubst, droppedTysExpSubst :: [Type] (remainingTysExpSubst, droppedTysExpSubst) = splitAt remainingLength varTysExpSubst -- All of the type variables mentioned in the dropped types -- (post-synonym expansion) droppedTyVarNames :: [Name] droppedTyVarNames = freeVariables droppedTysExpSubst -- If any of the dropped types were polykinded, ensure that they are of kind * -- after substituting * for the dropped kind variables. If not, throw an error. unless (all hasKindStar droppedTysExpSubst) $ derivingKindError cRep tyConName let preds :: [Maybe Pred] kvNames :: [[Name]] kvNames' :: [Name] -- Derive instance constraints (and any kind variables which are specialized -- to * in those constraints) (preds, kvNames) = unzip $ map (deriveConstraint cRep) remainingTysExpSubst kvNames' = concat kvNames -- Substitute the kind variables specialized in the constraints with * remainingTysExpSubst' :: [Type] remainingTysExpSubst' = map (substNamesWithKindStar kvNames') remainingTysExpSubst -- We now substitute all of the specialized-to-* kind variable names with -- *, but in the original types, not the synonym-expanded types. The reason -- we do this is a superficial one: we want the derived instance to resemble -- the datatype written in source code as closely as possible. For example, -- for the following data family instance: -- -- data family Fam a -- newtype instance Fam String = Fam String -- -- We'd want to generate the instance: -- -- instance C (Fam String) -- -- Not: -- -- instance C (Fam [Char]) remainingTysOrigSubst :: [Type] remainingTysOrigSubst = map (substNamesWithKindStar (List.union droppedKindVarNames kvNames')) $ take remainingLength varTysOrig isDataFamily <- case variant of Datatype -> return False Newtype -> return False DataInstance -> return True NewtypeInstance -> return True #if MIN_VERSION_th_abstraction(0,5,0) Datatype.TypeData -> typeDataError tyConName #endif let remainingTysOrigSubst' :: [Type] -- See Note [Kind signatures in derived instances] for an explanation -- of the isDataFamily check. remainingTysOrigSubst' = if isDataFamily then remainingTysOrigSubst else map unSigT remainingTysOrigSubst instanceCxt :: Cxt instanceCxt = catMaybes preds instanceType :: Type instanceType = AppT (ConT (fullClassName cRep)) $ applyTyCon tyConName remainingTysOrigSubst' -- If the datatype context mentions any of the dropped type variables, -- we can't derive an instance, so throw an error. when (any (`predMentionsName` droppedTyVarNames) dataCxt) $ datatypeContextError tyConName instanceType -- Also ensure the dropped types can be safely eta-reduced. Otherwise, -- throw an error. unless (canEtaReduce remainingTysExpSubst' droppedTysExpSubst) $ etaReductionError instanceType return (instanceCxt, instanceType) -- | Attempt to derive a constraint on a Type. If successful, return -- Just the constraint and any kind variable names constrained to *. -- Otherwise, return Nothing and the empty list. -- -- See Note [Type inference in derived instances] for the heuristics used to -- come up with constraints. deriveConstraint :: ClassRep a => a -> Type -> (Maybe Pred, [Name]) deriveConstraint cRep t | not (isTyVar t) = (Nothing, []) | hasKindStar t = ((`applyClass` tName) `fmap` classConstraint cRep 0, []) | otherwise = case hasKindVarChain 1 t of Just ns | cRepArity >= 1 -> ((`applyClass` tName) `fmap` classConstraint cRep 1, ns) _ -> case hasKindVarChain 2 t of Just ns | cRepArity == 2 -> ((`applyClass` tName) `fmap` classConstraint cRep 2, ns) _ -> (Nothing, []) where tName :: Name tName = varTToName t cRepArity :: Int cRepArity = arity cRep {- Note [Kind signatures in derived instances] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ It is possible to put explicit kind signatures into the derived instances, e.g., instance C a => C (Data (f :: * -> *)) where ... But it is preferable to avoid this if possible. If we come up with an incorrect kind signature (which is entirely possible, since our type inferencer is pretty unsophisticated - see Note [Type inference in derived instances]), then GHC will flat-out reject the instance, which is quite unfortunate. Plain old datatypes have the advantage that you can avoid using any kind signatures at all in their instances. This is because a datatype declaration uses all type variables, so the types that we use in a derived instance uniquely determine their kinds. As long as we plug in the right types, the kind inferencer can do the rest of the work. For this reason, we use unSigT to remove all kind signatures before splicing in the instance context and head. Data family instances are trickier, since a data family can have two instances that are distinguished by kind alone, e.g., data family Fam (a :: k) data instance Fam (a :: * -> *) data instance Fam (a :: *) If we dropped the kind signatures for C (Fam a), then GHC will have no way of knowing which instance we are talking about. To avoid this scenario, we always include explicit kind signatures in data family instances. There is a chance that the inferred kind signatures will be incorrect, but if so, we can always fall back on the make- functions. Note [Type inference in derived instances] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Type inference is can be tricky to get right, and we want to avoid recreating the entirety of GHC's type inferencer in Template Haskell. For this reason, we will probably never come up with derived instance contexts that are as accurate as GHC's. But that doesn't mean we can't do anything! There are a couple of simple things we can do to make instance contexts that work for 80% of use cases: 1. If one of the last type parameters is polykinded, then its kind will be specialized to * in the derived instance. We note what kind variable the type parameter had and substitute it with * in the other types as well. For example, imagine you had data Data (a :: k) (b :: k) Then you'd want to derived instance to be: instance C (Data (a :: *)) Not: instance C (Data (a :: k)) 2. We naïvely come up with instance constraints using the following criteria, using Show(1)(2) as the example typeclasses: (i) If there's a type parameter n of kind *, generate a Show n constraint. (ii) If there's a type parameter n of kind k1 -> k2 (where k1/k2 are * or kind variables), then generate a Show1 n constraint, and if k1/k2 are kind variables, then substitute k1/k2 with * elsewhere in the types. We must consider the case where they are kind variables because you might have a scenario like this: newtype Compose (f :: k2 -> *) (g :: k1 -> k2) (a :: k1) = Compose (f (g a)) Which would have a derived Show1 instance of: instance (Show1 f, Show1 g) => Show1 (Compose f g) where ... (iii) If there's a type parameter n of kind k1 -> k2 -> k3 (where k1/k2/k3 are * or kind variables), then generate a Show2 constraint and perform kind substitution as in the other cases. -} checkExistentialContext :: ClassRep a => a -> TyVarMap b -> Cxt -> Name -> Q c -> Q c checkExistentialContext cRep tvMap ctxt conName q = if (any (`predMentionsName` Map.keys tvMap) ctxt || Map.size tvMap < arity cRep) && not (allowExQuant cRep) then existentialContextError conName else q {- Note [Matching functions with GADT type variables] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ When deriving category-2 classes like Show2, there is a tricky corner case to consider: data Both a b where BothCon :: x -> x -> Both x x Which show functions should be applied to which arguments of BothCon? We have a choice, since both the function of type (Int -> a -> ShowS) and of type (Int -> b -> ShowS) can be applied to either argument. In such a scenario, the second show function takes precedence over the first show function, so the derived Show2 instance would be: instance Show2 Both where liftShowsPrec2 sp1 sp2 p (BothCon x1 x2) = showsParen (p > appPrec) $ showString "BothCon " . sp2 appPrec1 x1 . showSpace . sp2 appPrec1 x2 This is not an arbitrary choice, as this definition ensures that liftShowsPrec2 showsPrec = liftShowsPrec for a derived Show1 instance for Both. -} ------------------------------------------------------------------------------- -- Error messages ------------------------------------------------------------------------------- -- | The given datatype has no constructors, and we don't know what to do with it. noConstructorsError :: Q a noConstructorsError = fail "Must have at least one data constructor" -- | Either the given data type doesn't have enough type variables, or one of -- the type variables to be eta-reduced cannot realize kind *. derivingKindError :: ClassRep a => a -> Name -> Q b derivingKindError cRep tyConName = fail . showString "Cannot derive well-kinded instance of form ‘" . showString className . showChar ' ' . showParen True ( showString (nameBase tyConName) . showString " ..." ) . showString "‘\n\tClass " . showString className . showString " expects an argument of kind " . showString (pprint . createKindChain $ arity cRep) $ "" where className :: String className = nameBase $ fullClassName cRep -- | The last type variable appeared in a contravariant position -- when deriving Functor. contravarianceError :: Name -> Q a contravarianceError conName = fail . showString "Constructor ‘" . showString (nameBase conName) . showString "‘ must not use the last type variable in a function argument" $ "" -- | A constructor has a function argument in a derived Foldable or Traversable -- instance. noFunctionsError :: Name -> Q a noFunctionsError conName = fail . showString "Constructor ‘" . showString (nameBase conName) . showString "‘ must not contain function types" $ "" -- | One of the last type variables cannot be eta-reduced (see the canEtaReduce -- function for the criteria it would have to meet). etaReductionError :: Type -> Q a etaReductionError instanceType = fail $ "Cannot eta-reduce to an instance of form \n\tinstance (...) => " ++ pprint instanceType -- | The data type has a DatatypeContext which mentions one of the eta-reduced -- type variables. datatypeContextError :: Name -> Type -> Q a datatypeContextError dataName instanceType = fail . showString "Can't make a derived instance of ‘" . showString (pprint instanceType) . showString "‘:\n\tData type ‘" . showString (nameBase dataName) . showString "‘ must not have a class context involving the last type argument(s)" $ "" -- | The data type has an existential constraint which mentions one of the -- eta-reduced type variables. existentialContextError :: Name -> Q a existentialContextError conName = fail . showString "Constructor ‘" . showString (nameBase conName) . showString "‘ must be truly polymorphic in the last argument(s) of the data type" $ "" -- | The data type mentions one of the n eta-reduced type variables in a place other -- than the last nth positions of a data type in a constructor's field. outOfPlaceTyVarError :: ClassRep a => a -> Name -> Q b outOfPlaceTyVarError cRep conName = fail . showString "Constructor ‘" . showString (nameBase conName) . showString "‘ must only use its last " . shows n . showString " type variable(s) within the last " . shows n . showString " argument(s) of a data type" $ "" where n :: Int n = arity cRep enumerationError :: String -> Q a enumerationError = fail . enumerationErrorStr enumerationOrProductError :: String -> Q a enumerationOrProductError nb = fail $ unlines [ enumerationErrorStr nb , "\tor a product type (precisely one constructor)" ] enumerationErrorStr :: String -> String enumerationErrorStr nb = '\'':nb ++ "’ must be an enumeration type" ++ " (one or more nullary, non-GADT constructors)" typeDataError :: Name -> Q a typeDataError dataName = fail . showString "Cannot derive instance for ‘" . showString (nameBase dataName) . showString "‘, which is a ‘type data‘ declaration" $ "" ------------------------------------------------------------------------------- -- Assorted utilities ------------------------------------------------------------------------------- -- | A mapping of type variable Names to their auxiliary function Names. type TyVarMap a = Map Name (OneOrTwoNames a) type TyVarMap1 = TyVarMap One type TyVarMap2 = TyVarMap Two data OneOrTwoNames a where OneName :: Name -> OneOrTwoNames One TwoNames :: Name -> Name -> OneOrTwoNames Two data One data Two interleave :: [a] -> [a] -> [a] interleave (a1:a1s) (a2:a2s) = a1:a2:interleave a1s a2s interleave _ _ = [] #if !(MIN_VERSION_ghc_prim(0,3,1)) isTrue# :: Bool -> Bool isTrue# x = x {-# INLINE isTrue# #-} #endif -- filterByList, filterByLists, and partitionByList taken from GHC (BSD3-licensed) -- | 'filterByList' takes a list of Bools and a list of some elements and -- filters out these elements for which the corresponding value in the list of -- Bools is False. This function does not check whether the lists have equal -- length. filterByList :: [Bool] -> [a] -> [a] filterByList (True:bs) (x:xs) = x : filterByList bs xs filterByList (False:bs) (_:xs) = filterByList bs xs filterByList _ _ = [] -- | 'filterByLists' takes a list of Bools and two lists as input, and -- outputs a new list consisting of elements from the last two input lists. For -- each Bool in the list, if it is 'True', then it takes an element from the -- former list. If it is 'False', it takes an element from the latter list. -- The elements taken correspond to the index of the Bool in its list. -- For example: -- -- @ -- filterByLists [True, False, True, False] \"abcd\" \"wxyz\" = \"axcz\" -- @ -- -- This function does not check whether the lists have equal length. filterByLists :: [Bool] -> [a] -> [a] -> [a] filterByLists (True:bs) (x:xs) (_:ys) = x : filterByLists bs xs ys filterByLists (False:bs) (_:xs) (y:ys) = y : filterByLists bs xs ys filterByLists _ _ _ = [] -- | 'partitionByList' takes a list of Bools and a list of some elements and -- partitions the list according to the list of Bools. Elements corresponding -- to 'True' go to the left; elements corresponding to 'False' go to the right. -- For example, @partitionByList [True, False, True] [1,2,3] == ([1,3], [2])@ -- This function does not check whether the lists have equal -- length. partitionByList :: [Bool] -> [a] -> ([a], [a]) partitionByList = go [] [] where go trues falses (True : bs) (x : xs) = go (x:trues) falses bs xs go trues falses (False : bs) (x : xs) = go trues (x:falses) bs xs go trues falses _ _ = (reverse trues, reverse falses) integerE :: Int -> Q Exp integerE = litE . integerL . fromIntegral -- | Returns True if a Type has kind *. hasKindStar :: Type -> Bool hasKindStar VarT{} = True #if MIN_VERSION_template_haskell(2,8,0) hasKindStar (SigT _ StarT) = True #else hasKindStar (SigT _ StarK) = True #endif hasKindStar _ = False -- Returns True is a kind is equal to *, or if it is a kind variable. isStarOrVar :: Kind -> Bool #if MIN_VERSION_template_haskell(2,8,0) isStarOrVar StarT = True isStarOrVar VarT{} = True #else isStarOrVar StarK = True #endif isStarOrVar _ = False -- | @hasKindVarChain n kind@ Checks if @kind@ is of the form -- k_0 -> k_1 -> ... -> k_(n-1), where k0, k1, ..., and k_(n-1) can be * or -- kind variables. hasKindVarChain :: Int -> Type -> Maybe [Name] hasKindVarChain kindArrows t = let uk = uncurryKind (tyKind t) in if (length uk - 1 == kindArrows) && all isStarOrVar uk then Just (freeVariables uk) else Nothing -- | If a Type is a SigT, returns its kind signature. Otherwise, return *. tyKind :: Type -> Kind tyKind (SigT _ k) = k tyKind _ = starK zipWithAndUnzipM :: Monad m => (a -> b -> m (c, d)) -> [a] -> [b] -> m ([c], [d]) zipWithAndUnzipM f (x:xs) (y:ys) = do (c, d) <- f x y (cs, ds) <- zipWithAndUnzipM f xs ys return (c:cs, d:ds) zipWithAndUnzipM _ _ _ = return ([], []) {-# INLINE zipWithAndUnzipM #-} zipWith3AndUnzipM :: Monad m => (a -> b -> c -> m (d, e)) -> [a] -> [b] -> [c] -> m ([d], [e]) zipWith3AndUnzipM f (x:xs) (y:ys) (z:zs) = do (d, e) <- f x y z (ds, es) <- zipWith3AndUnzipM f xs ys zs return (d:ds, e:es) zipWith3AndUnzipM _ _ _ _ = return ([], []) {-# INLINE zipWith3AndUnzipM #-} thd3 :: (a, b, c) -> c thd3 (_, _, c) = c unsnoc :: [a] -> Maybe ([a], a) unsnoc [] = Nothing unsnoc (x:xs) = case unsnoc xs of Nothing -> Just ([], x) Just (a,b) -> Just (x:a, b) isNullaryCon :: ConstructorInfo -> Bool isNullaryCon (ConstructorInfo { constructorFields = tys }) = null tys -- | Returns the number of fields for the constructor. conArity :: ConstructorInfo -> Int conArity (ConstructorInfo { constructorFields = tys }) = length tys -- | Returns 'True' if it's a datatype with exactly one, non-existential constructor. isProductType :: NonEmpty ConstructorInfo -> Bool isProductType (con :| []) = null (constructorVars con) isProductType _ = False -- | Returns 'True' if it's a datatype with one or more nullary, non-GADT -- constructors. isEnumerationType :: NonEmpty ConstructorInfo -> Bool isEnumerationType cons = F.all (App.liftA2 (&&) isNullaryCon isVanillaCon) cons -- | Returns 'False' if we're dealing with existential quantification or GADTs. isVanillaCon :: ConstructorInfo -> Bool isVanillaCon (ConstructorInfo { constructorContext = ctxt, constructorVars = vars }) = null ctxt && null vars -- | Generate a list of fresh names with a common prefix, and numbered suffixes. newNameList :: String -> Int -> Q [Name] newNameList prefix n = T.mapM (newName . (prefix ++) . show) [1..n] -- | Extracts the kind from a TyVarBndr. tvbKind :: TyVarBndr_ flag -> Kind tvbKind = elimTV (\_ -> starK) (\_ k -> k) -- | Convert a TyVarBndr to a Type. tvbToType :: TyVarBndr_ flag -> Type tvbToType = elimTV VarT (\n k -> SigT (VarT n) k) -- | Applies a typeclass constraint to a type. applyClass :: Name -> Name -> Pred #if MIN_VERSION_template_haskell(2,10,0) applyClass con t = AppT (ConT con) (VarT t) #else applyClass con t = ClassP con [VarT t] #endif createKindChain :: Int -> Kind createKindChain = go starK where go :: Kind -> Int -> Kind go k !0 = k #if MIN_VERSION_template_haskell(2,8,0) go k !n = go (AppT (AppT ArrowT StarT) k) (n - 1) #else go k !n = go (ArrowK StarK k) (n - 1) #endif -- | Checks to see if the last types in a data family instance can be safely eta- -- reduced (i.e., dropped), given the other types. This checks for three conditions: -- -- (1) All of the dropped types are type variables -- (2) All of the dropped types are distinct -- (3) None of the remaining types mention any of the dropped types canEtaReduce :: [Type] -> [Type] -> Bool canEtaReduce remaining dropped = all isTyVar dropped && allDistinct droppedNames -- Make sure not to pass something of type [Type], since Type -- didn't have an Ord instance until template-haskell-2.10.0.0 && not (any (`mentionsName` droppedNames) remaining) where droppedNames :: [Name] droppedNames = map varTToName dropped -- | Extract the Name from a type constructor. If the argument Type is not a -- type variable, throw an error. conTToName :: Type -> Name conTToName (ConT n) = n conTToName (SigT t _) = conTToName t conTToName _ = error "Not a type constructor!" -- | Extract Just the Name from a type variable. If the argument Type is not a -- type variable, return Nothing. varTToName_maybe :: Type -> Maybe Name varTToName_maybe (VarT n) = Just n varTToName_maybe (SigT t _) = varTToName_maybe t varTToName_maybe _ = Nothing -- | Extract the Name from a type variable. If the argument Type is not a -- type variable, throw an error. varTToName :: Type -> Name varTToName = fromMaybe (error "Not a type variable!") . varTToName_maybe -- | Peel off a kind signature from a Type (if it has one). unSigT :: Type -> Type unSigT (SigT t _) = t unSigT t = t -- | Is the given type a variable? isTyVar :: Type -> Bool isTyVar (VarT _) = True isTyVar (SigT t _) = isTyVar t isTyVar _ = False -- | Detect if a Name in a list of provided Names occurs as an argument to some -- type family. This makes an effort to exclude /oversaturated/ arguments to -- type families. For instance, if one declared the following type family: -- -- @ -- type family F a :: Type -> Type -- @ -- -- Then in the type @F a b@, we would consider @a@ to be an argument to @F@, -- but not @b@. isInTypeFamilyApp :: [Name] -> Type -> [Type] -> Q Bool isInTypeFamilyApp names tyFun tyArgs = case tyFun of ConT tcName -> go tcName _ -> return False where go :: Name -> Q Bool go tcName = do info <- reify tcName case info of #if MIN_VERSION_template_haskell(2,11,0) FamilyI (OpenTypeFamilyD (TypeFamilyHead _ bndrs _ _)) _ -> withinFirstArgs bndrs #elif MIN_VERSION_template_haskell(2,7,0) FamilyI (FamilyD TypeFam _ bndrs _) _ -> withinFirstArgs bndrs #else TyConI (FamilyD TypeFam _ bndrs _) -> withinFirstArgs bndrs #endif #if MIN_VERSION_template_haskell(2,11,0) FamilyI (ClosedTypeFamilyD (TypeFamilyHead _ bndrs _ _) _) _ -> withinFirstArgs bndrs #elif MIN_VERSION_template_haskell(2,9,0) FamilyI (ClosedTypeFamilyD _ bndrs _ _) _ -> withinFirstArgs bndrs #endif _ -> return False where withinFirstArgs :: [a] -> Q Bool withinFirstArgs bndrs = let firstArgs = take (length bndrs) tyArgs argFVs = freeVariables firstArgs in return $ any (`elem` argFVs) names -- | Are all of the items in a list (which have an ordering) distinct? -- -- This uses Set (as opposed to nub) for better asymptotic time complexity. allDistinct :: Ord a => [a] -> Bool allDistinct = allDistinct' Set.empty where allDistinct' :: Ord a => Set a -> [a] -> Bool allDistinct' uniqs (x:xs) | x `Set.member` uniqs = False | otherwise = allDistinct' (Set.insert x uniqs) xs allDistinct' _ _ = True -- | Does the given type mention any of the Names in the list? mentionsName :: Type -> [Name] -> Bool mentionsName = go where go :: Type -> [Name] -> Bool go (AppT t1 t2) names = go t1 names || go t2 names go (SigT t _k) names = go t names #if MIN_VERSION_template_haskell(2,8,0) || go _k names #endif go (VarT n) names = n `elem` names go _ _ = False -- | Does an instance predicate mention any of the Names in the list? predMentionsName :: Pred -> [Name] -> Bool #if MIN_VERSION_template_haskell(2,10,0) predMentionsName = mentionsName #else predMentionsName (ClassP n tys) names = n `elem` names || any (`mentionsName` names) tys predMentionsName (EqualP t1 t2) names = mentionsName t1 names || mentionsName t2 names #endif -- | Construct a type via curried application. applyTy :: Type -> [Type] -> Type applyTy = List.foldl' AppT -- | Fully applies a type constructor to its type variables. applyTyCon :: Name -> [Type] -> Type applyTyCon = applyTy . ConT -- | Split an applied type into its individual components. For example, this: -- -- @ -- Either Int Char -- @ -- -- would split to this: -- -- @ -- [Either, Int, Char] -- @ unapplyTy :: Type -> (Type, [Type]) unapplyTy ty = go ty ty [] where go :: Type -> Type -> [Type] -> (Type, [Type]) go _ (AppT ty1 ty2) args = go ty1 ty1 (ty2:args) go origTy (SigT ty' _) args = go origTy ty' args #if MIN_VERSION_template_haskell(2,11,0) go origTy (InfixT ty1 n ty2) args = go origTy (ConT n `AppT` ty1 `AppT` ty2) args go origTy (ParensT ty') args = go origTy ty' args #endif go origTy _ args = (origTy, args) -- | Split a type signature by the arrows on its spine. For example, this: -- -- @ -- forall a b. (a ~ b) => (a -> b) -> Char -> () -- @ -- -- would split to this: -- -- @ -- (a ~ b, [a -> b, Char, ()]) -- @ uncurryTy :: Type -> (Cxt, [Type]) uncurryTy (AppT (AppT ArrowT t1) t2) = let (ctxt, tys) = uncurryTy t2 in (ctxt, t1:tys) uncurryTy (SigT t _) = uncurryTy t uncurryTy (ForallT _ ctxt t) = let (ctxt', tys) = uncurryTy t in (ctxt ++ ctxt', tys) uncurryTy t = ([], [t]) -- | Like uncurryType, except on a kind level. uncurryKind :: Kind -> [Kind] #if MIN_VERSION_template_haskell(2,8,0) uncurryKind = snd . uncurryTy #else uncurryKind (ArrowK k1 k2) = k1:uncurryKind k2 uncurryKind k = [k] #endif untagExpr :: [(Name, Name)] -> Q Exp -> Q Exp untagExpr [] e = e untagExpr ((untagThis, putTagHere) : more) e = caseE (varE getTagValName `appE` varE untagThis) [match (varP putTagHere) (normalB $ untagExpr more e) []] tag2ConExpr :: Type -> Q Exp tag2ConExpr ty = do iHash <- newName "i#" ty' <- freshenType ty lam1E (conP iHashDataName [varP iHash]) $ varE tagToEnumHashValName `appE` varE iHash `sigE` return (quantifyType ty') -- tagToEnum# is a hack, and won't typecheck unless it's in the -- immediate presence of a type ascription like so: -- -- tagToEnum# x :: Foo -- -- We have to be careful when dealing with datatypes with type -- variables, since Template Haskell might reject the type variables -- we use for being out-of-scope. To avoid this, we explicitly -- collect the type variable binders and shove them into a ForallT -- (using th-abstraction's quantifyType function). Also make sure -- to freshen the bound type variables to avoid shadowed variable -- warnings on old versions of GHC when -Wall is enabled. primOrdFunTbl :: Map Name (Name, Name, Name, Name, Name) primOrdFunTbl = Map.fromList [ (addrHashTypeName, ( ltAddrHashValName , leAddrHashValName , eqAddrHashValName , geAddrHashValName , gtAddrHashValName )) , (charHashTypeName, ( ltCharHashValName , leCharHashValName , eqCharHashValName , geCharHashValName , gtCharHashValName )) , (doubleHashTypeName, ( ltDoubleHashValName , leDoubleHashValName , eqDoubleHashValName , geDoubleHashValName , gtDoubleHashValName )) , (floatHashTypeName, ( ltFloatHashValName , leFloatHashValName , eqFloatHashValName , geFloatHashValName , gtFloatHashValName )) , (intHashTypeName, ( ltIntHashValName , leIntHashValName , eqIntHashValName , geIntHashValName , gtIntHashValName )) , (wordHashTypeName, ( ltWordHashValName , leWordHashValName , eqWordHashValName , geWordHashValName , gtWordHashValName )) #if MIN_VERSION_base(4,13,0) , (int8HashTypeName, ( ltInt8HashValName , leInt8HashValName , eqInt8HashValName , geInt8HashValName , gtInt8HashValName )) , (int16HashTypeName, ( ltInt16HashValName , leInt16HashValName , eqInt16HashValName , geInt16HashValName , gtInt16HashValName )) , (word8HashTypeName, ( ltWord8HashValName , leWord8HashValName , eqWord8HashValName , geWord8HashValName , gtWord8HashValName )) , (word16HashTypeName, ( ltWord16HashValName , leWord16HashValName , eqWord16HashValName , geWord16HashValName , gtWord16HashValName )) #endif #if MIN_VERSION_base(4,16,0) , (int32HashTypeName, ( ltInt32HashValName , leInt32HashValName , eqInt32HashValName , geInt32HashValName , gtInt32HashValName )) , (word32HashTypeName, ( ltWord32HashValName , leWord32HashValName , eqWord32HashValName , geWord32HashValName , gtWord32HashValName )) #endif ] removeClassApp :: Type -> Type removeClassApp (AppT _ t2) = t2 removeClassApp t = t -- This is an ugly, but unfortunately necessary hack on older versions of GHC which -- don't have a properly working newName. On those GHCs, even running newName on a -- variable isn't enought to avoid shadowed variable warnings, so we "fix" the issue by -- appending an uncommonly used string to the end of the name. This isn't foolproof, -- since a user could freshen a variable named x and still have another x_' variable in -- scope, but at least it's unlikely. freshen :: Name -> Q Name freshen n = newName (nameBase n ++ "_'") freshenType :: Type -> Q Type freshenType t = do let xs = [(n, VarT `fmap` freshen n) | n <- freeVariables t] subst <- T.sequence (Map.fromList xs) return (applySubstitution subst t) enumFromToExpr :: Q Exp -> Q Exp -> Q Exp enumFromToExpr f t = varE enumFromToValName `appE` f `appE` t primOpAppExpr :: Q Exp -> Name -> Q Exp -> Q Exp primOpAppExpr e1 op e2 = varE isTrueHashValName `appE` infixApp e1 (varE op) e2 -- | Checks if a 'Name' represents a tuple type constructor (other than '()') isNonUnitTuple :: Name -> Bool isNonUnitTuple = isNonUnitTupleString . nameBase -- | Checks if a 'String' represents a tuple (other than '()') isNonUnitTupleString :: String -> Bool isNonUnitTupleString ('(':',':_) = True isNonUnitTupleString _ = False -- | Checks if a 'String' names a valid Haskell infix data constructor (i.e., does -- it begin with a colon?). isInfixDataCon :: String -> Bool isInfixDataCon (':':_) = True isInfixDataCon _ = False isSym :: String -> Bool isSym "" = False isSym (c : _) = startsVarSym c || startsConSym c #if !defined(MIN_VERSION_ghc_boot_th) startsVarSym, startsConSym :: Char -> Bool startsVarSym c = startsVarSymASCII c || (ord c > 0x7f && isSymbol c) -- Infix Ids startsConSym c = c == ':' -- Infix data constructors startsVarSymASCII :: Char -> Bool startsVarSymASCII c = c `elem` "!#$%&*+./<=>?@\\^|~-" #endif ghc7'8OrLater :: Bool #if __GLASGOW_HASKELL__ >= 708 ghc7'8OrLater = True #else ghc7'8OrLater = False #endif ------------------------------------------------------------------------------- -- Quoted names ------------------------------------------------------------------------------- -- With GHC 8.0 or later, we can simply use TemplateHaskellQuotes to quote each -- name, which allows deriving-compat to be built with compilers that do not -- support Template Haskell (e.g., stage-1 cross compilers). Unfortunately, -- older versions of GHC must fall back on full-blown Template Haskell. isTrueHashValName :: Name isTrueHashValName = 'isTrue# fmapConstValName :: Name fmapConstValName = 'fmapConst replaceConstValName :: Name replaceConstValName = 'replaceConst foldrConstValName :: Name foldrConstValName = 'foldrConst foldMapConstValName :: Name foldMapConstValName = 'foldMapConst nullConstValName :: Name nullConstValName = 'nullConst traverseConstValName :: Name traverseConstValName = 'traverseConst eqConstValName :: Name eqConstValName = 'eqConst eq1ConstValName :: Name eq1ConstValName = 'eq1Const liftEqConstValName :: Name liftEqConstValName = 'liftEqConst liftEq2ConstValName :: Name liftEq2ConstValName = 'liftEq2Const compareConstValName :: Name compareConstValName = 'compareConst ltConstValName :: Name ltConstValName = 'ltConst compare1ConstValName :: Name compare1ConstValName = 'compare1Const liftCompareConstValName :: Name liftCompareConstValName = 'liftCompareConst liftCompare2ConstValName :: Name liftCompare2ConstValName = 'liftCompare2Const readsPrecConstValName :: Name readsPrecConstValName = 'readsPrecConst readPrecConstValName :: Name readPrecConstValName = 'readPrecConst readsPrec1ConstValName :: Name readsPrec1ConstValName = 'readsPrec1Const liftReadsPrecConstValName :: Name liftReadsPrecConstValName = 'liftReadsPrecConst liftReadPrecConstValName :: Name liftReadPrecConstValName = 'liftReadPrecConst liftReadsPrec2ConstValName :: Name liftReadsPrec2ConstValName = 'liftReadsPrec2Const liftReadPrec2ConstValName :: Name liftReadPrec2ConstValName = 'liftReadPrec2Const showsPrecConstValName :: Name showsPrecConstValName = 'showsPrecConst showsPrec1ConstValName :: Name showsPrec1ConstValName = 'showsPrec1Const liftShowsPrecConstValName :: Name liftShowsPrecConstValName = 'liftShowsPrecConst liftShowsPrec2ConstValName :: Name liftShowsPrec2ConstValName = 'liftShowsPrec2Const viaTypeName :: Name viaTypeName = ''Via cHashDataName :: Name cHashDataName = 'C# dHashDataName :: Name dHashDataName = 'D# fHashDataName :: Name fHashDataName = 'F# identDataName :: Name identDataName = 'L.Ident iHashDataName :: Name iHashDataName = 'I# puncDataName :: Name puncDataName = 'L.Punc symbolDataName :: Name symbolDataName = 'L.Symbol wrapMonadDataName :: Name wrapMonadDataName = 'App.WrapMonad addrHashTypeName :: Name addrHashTypeName = ''Addr# boundedTypeName :: Name boundedTypeName = ''Bounded charHashTypeName :: Name charHashTypeName = ''Char# doubleHashTypeName :: Name doubleHashTypeName = ''Double# enumTypeName :: Name enumTypeName = ''Enum floatHashTypeName :: Name floatHashTypeName = ''Float# foldableTypeName :: Name foldableTypeName = ''Foldable functorTypeName :: Name functorTypeName = ''Functor intTypeName :: Name intTypeName = ''Int intHashTypeName :: Name intHashTypeName = ''Int# ixTypeName :: Name ixTypeName = ''Ix readTypeName :: Name readTypeName = ''Read showTypeName :: Name showTypeName = ''Show traversableTypeName :: Name traversableTypeName = ''Traversable wordHashTypeName :: Name wordHashTypeName = ''Word# altValName :: Name altValName = '(+++) appendValName :: Name appendValName = '(++) chooseValName :: Name chooseValName = 'choose composeValName :: Name composeValName = '(.) constValName :: Name constValName = 'const enumFromValName :: Name enumFromValName = 'enumFrom enumFromThenValName :: Name enumFromThenValName = 'enumFromThen enumFromThenToValName :: Name enumFromThenToValName = 'enumFromThenTo enumFromToValName :: Name enumFromToValName = 'enumFromTo eqAddrHashValName :: Name eqAddrHashValName = 'eqAddr# eqCharHashValName :: Name eqCharHashValName = 'eqChar# eqDoubleHashValName :: Name eqDoubleHashValName = '(==##) eqFloatHashValName :: Name eqFloatHashValName = 'eqFloat# eqIntHashValName :: Name eqIntHashValName = '(==#) eqWordHashValName :: Name eqWordHashValName = 'eqWord# errorValName :: Name errorValName = 'error flipValName :: Name flipValName = 'flip fmapValName :: Name fmapValName = 'fmap foldrValName :: Name foldrValName = 'F.foldr foldMapValName :: Name foldMapValName = 'foldMap fromEnumValName :: Name fromEnumValName = 'fromEnum geAddrHashValName :: Name geAddrHashValName = 'geAddr# geCharHashValName :: Name geCharHashValName = 'geChar# geDoubleHashValName :: Name geDoubleHashValName = '(>=##) geFloatHashValName :: Name geFloatHashValName = 'geFloat# geIntHashValName :: Name geIntHashValName = '(>=#) getTagValName :: Name getTagValName = 'getTag geWordHashValName :: Name geWordHashValName = 'geWord# gtAddrHashValName :: Name gtAddrHashValName = 'gtAddr# gtCharHashValName :: Name gtCharHashValName = 'gtChar# gtDoubleHashValName :: Name gtDoubleHashValName = '(>##) gtFloatHashValName :: Name gtFloatHashValName = 'gtFloat# gtIntHashValName :: Name gtIntHashValName = '(>#) gtWordHashValName :: Name gtWordHashValName = 'gtWord# idValName :: Name idValName = 'id indexValName :: Name indexValName = 'index inRangeValName :: Name inRangeValName = 'inRange leAddrHashValName :: Name leAddrHashValName = 'leAddr# leCharHashValName :: Name leCharHashValName = 'leChar# leDoubleHashValName :: Name leDoubleHashValName = '(<=##) leFloatHashValName :: Name leFloatHashValName = 'leFloat# leIntHashValName :: Name leIntHashValName = '(<=#) leWordHashValName :: Name leWordHashValName = 'leWord# listValName :: Name listValName = 'list ltAddrHashValName :: Name ltAddrHashValName = 'ltAddr# ltCharHashValName :: Name ltCharHashValName = 'ltChar# ltDoubleHashValName :: Name ltDoubleHashValName = '(<##) ltFloatHashValName :: Name ltFloatHashValName = 'ltFloat# ltIntHashValName :: Name ltIntHashValName = '(<#) ltWordHashValName :: Name ltWordHashValName = 'ltWord# minBoundValName :: Name minBoundValName = 'minBound mapValName :: Name mapValName = 'map maxBoundValName :: Name maxBoundValName = 'maxBound minusIntHashValName :: Name minusIntHashValName = '(-#) neqIntHashValName :: Name neqIntHashValName = '(/=#) parenValName :: Name parenValName = 'paren parensValName :: Name parensValName = 'parens pfailValName :: Name pfailValName = 'pfail plusValName :: Name plusValName = '(+) precValName :: Name precValName = 'prec predValName :: Name predValName = 'pred rangeSizeValName :: Name rangeSizeValName = 'rangeSize rangeValName :: Name rangeValName = 'range readFieldHash :: String -> ReadPrec a -> ReadPrec a readFieldHash fieldName readVal = do expectP (L.Ident fieldName) expectP (L.Symbol "#") expectP (L.Punc "=") readVal {-# NOINLINE readFieldHash #-} readFieldHashValName :: Name readFieldHashValName = 'readFieldHash readListValName :: Name readListValName = 'readList readListPrecDefaultValName :: Name readListPrecDefaultValName = 'readListPrecDefault readListPrecValName :: Name readListPrecValName = 'readListPrec readPrec_to_SValName :: Name readPrec_to_SValName = 'readPrec_to_S readPrecValName :: Name readPrecValName = 'readPrec readS_to_PrecValName :: Name readS_to_PrecValName = 'readS_to_Prec readsPrecValName :: Name readsPrecValName = 'readsPrec replaceValName :: Name replaceValName = '(<$) resetValName :: Name resetValName = 'reset returnValName :: Name returnValName = 'return seqValName :: Name seqValName = 'seq showCharValName :: Name showCharValName = 'showChar showListValName :: Name showListValName = 'showList showListWithValName :: Name showListWithValName = 'showListWith showParenValName :: Name showParenValName = 'showParen showsPrecValName :: Name showsPrecValName = 'showsPrec showSpaceValName :: Name showSpaceValName = 'showSpace showStringValName :: Name showStringValName = 'showString stepValName :: Name stepValName = 'step succValName :: Name succValName = 'succ tagToEnumHashValName :: Name tagToEnumHashValName = 'tagToEnum# timesValName :: Name timesValName = '(*) toEnumValName :: Name toEnumValName = 'toEnum traverseValName :: Name traverseValName = 'traverse unsafeIndexValName :: Name unsafeIndexValName = 'unsafeIndex unsafeRangeSizeValName :: Name unsafeRangeSizeValName = 'unsafeRangeSize unwrapMonadValName :: Name unwrapMonadValName = 'App.unwrapMonad boolTypeName :: Name boolTypeName = ''Bool falseDataName :: Name falseDataName = 'False trueDataName :: Name trueDataName = 'True eqDataName :: Name eqDataName = 'EQ gtDataName :: Name gtDataName = 'GT ltDataName :: Name ltDataName = 'LT eqTypeName :: Name eqTypeName = ''Eq ordTypeName :: Name ordTypeName = ''Ord andValName :: Name andValName = '(&&) compareValName :: Name compareValName = 'compare eqValName :: Name eqValName = '(==) geValName :: Name geValName = '(>=) gtValName :: Name gtValName = '(>) leValName :: Name leValName = '(<=) ltValName :: Name ltValName = '(<) notValName :: Name notValName = 'not wHashDataName :: Name wHashDataName = 'W# #if !(MIN_VERSION_base(4,7,0)) expectP :: Lexeme -> ReadPrec () expectP lexeme = do thing <- lexP if thing == lexeme then return () else pfail #endif expectPValName :: Name expectPValName = 'expectP allValName :: Name allValName = 'all apValName :: Name apValName = '(<*>) pureValName :: Name pureValName = 'pure liftA2ValName :: Name liftA2ValName = 'App.liftA2 mappendValName :: Name mappendValName = 'mappend memptyValName :: Name memptyValName = 'mempty nullValName :: Name nullValName = 'null eq1TypeName :: Name eq1TypeName = ''Eq1 ord1TypeName :: Name ord1TypeName = ''Ord1 read1TypeName :: Name read1TypeName = ''Read1 show1TypeName :: Name show1TypeName = ''Show1 #if !(MIN_VERSION_transformers(0,4,0)) || MIN_VERSION_transformers(0,5,0) eq2TypeName :: Name eq2TypeName = ''Eq2 ord2TypeName :: Name ord2TypeName = ''Ord2 read2TypeName :: Name read2TypeName = ''Read2 show2TypeName :: Name show2TypeName = ''Show2 liftEqValName :: Name liftEqValName = 'liftEq liftEq2ValName :: Name liftEq2ValName = 'liftEq2 liftCompareValName :: Name liftCompareValName = 'liftCompare liftCompare2ValName :: Name liftCompare2ValName = 'liftCompare2 liftReadsPrecValName :: Name liftReadsPrecValName = 'liftReadsPrec liftReadListValName :: Name liftReadListValName = 'liftReadList liftReadsPrec2ValName :: Name liftReadsPrec2ValName = 'liftReadsPrec2 liftReadList2ValName :: Name liftReadList2ValName = 'liftReadList2 liftShowListValName :: Name liftShowListValName = 'liftShowList liftShowsPrecValName :: Name liftShowsPrecValName = 'liftShowsPrec liftShowList2ValName :: Name liftShowList2ValName = 'liftShowList2 liftShowsPrec2ValName :: Name liftShowsPrec2ValName = 'liftShowsPrec2 #else eq1ValName :: Name eq1ValName = 'eq1 compare1ValName :: Name compare1ValName = 'compare1 readsPrec1ValName :: Name readsPrec1ValName = 'readsPrec1 showsPrec1ValName :: Name showsPrec1ValName = 'showsPrec1 newtype Apply f a = Apply { unApply :: f a } instance (Eq1 f, Eq a) => Eq (Apply f a) where Apply x == Apply y = eq1 x y instance (Ord1 g, Ord a) => Ord (Apply g a) where compare (Apply x) (Apply y) = compare1 x y instance (Read1 f, Read a) => Read (Apply f a) where readsPrec d s = [(Apply a, t) | (a, t) <- readsPrec1 d s] instance (Show1 f, Show a) => Show (Apply f a) where showsPrec p (Apply x) = showsPrec1 p x makeFmapApplyNeg :: ClassRep a => a -> Name -> Type -> Name -> Q Exp makeFmapApplyNeg = makeFmapApply False makeFmapApplyPos :: ClassRep a => a -> Name -> Type -> Name -> Q Exp makeFmapApplyPos = makeFmapApply True makeFmapApply :: ClassRep a => Bool -> a -> Name -> Type -> Name -> Q Exp makeFmapApply pos cRep conName (SigT ty _) name = makeFmapApply pos cRep conName ty name makeFmapApply pos cRep conName t name = do let tyCon :: Type tyArgs :: [Type] (tyCon, tyArgs) = unapplyTy t numLastArgs :: Int numLastArgs = min (arity cRep) (length tyArgs) lhsArgs, rhsArgs :: [Type] (lhsArgs, rhsArgs) = splitAt (length tyArgs - numLastArgs) tyArgs inspectTy :: Type -> Q Exp inspectTy (SigT ty _) = inspectTy ty inspectTy (VarT a) | a == name = varE idValName inspectTy beta = varE fmapValName `appE` infixApp (if pos then makeFmapApply pos cRep conName beta name else conE applyDataName) (varE composeValName) (if pos then varE unApplyValName else makeFmapApply pos cRep conName beta name) itf <- isInTypeFamilyApp [name] tyCon tyArgs if any (`mentionsName` [name]) lhsArgs || itf then outOfPlaceTyVarError cRep conName else inspectTy (head rhsArgs) applyDataName :: Name applyDataName = 'Apply unApplyValName :: Name unApplyValName = 'unApply #endif #if MIN_VERSION_base(4,7,0) coerceValName :: Name coerceValName = 'coerce #endif #if MIN_VERSION_base(4,10,0) liftReadListPrecDefaultValName :: Name liftReadListPrecDefaultValName = 'liftReadListPrecDefault liftReadListPrec2DefaultValName :: Name liftReadListPrec2DefaultValName = 'liftReadListPrec2Default liftReadListPrecValName :: Name liftReadListPrecValName = 'liftReadListPrec liftReadListPrec2ValName :: Name liftReadListPrec2ValName = 'liftReadListPrec2 liftReadPrecValName :: Name liftReadPrecValName = 'liftReadPrec liftReadPrec2ValName :: Name liftReadPrec2ValName = 'liftReadPrec2 #else -- This is a gross hack to avoid needing to guard some uses of these two Names -- in Text.Read.Deriving.Internal with even grosser CPP. liftReadListPrecDefaultValName :: Name liftReadListPrecDefaultValName = error "using liftReadListPrecDefault before base-4.10.*" liftReadListPrec2DefaultValName :: Name liftReadListPrec2DefaultValName = error "using liftReadListPrec2Default before base-4.10.*" liftReadListPrecValName :: Name liftReadListPrecValName = error "using liftReadListPrec before base-4.10.*" liftReadListPrec2ValName :: Name liftReadListPrec2ValName = error "using liftReadListPrec2 before base-4.10.*" liftReadPrecValName :: Name liftReadPrecValName = error "using liftReadPrec before base-4.10.*" liftReadPrec2ValName :: Name liftReadPrec2ValName = error "using liftReadPrec2 before base-4.10.*" #endif #if !(MIN_VERSION_base(4,10,0)) showCommaSpace :: ShowS showCommaSpace = showString ", " #endif showCommaSpaceValName :: Name showCommaSpaceValName = 'showCommaSpace appEndoValName :: Name appEndoValName = 'appEndo dualDataName :: Name dualDataName = 'Dual endoDataName :: Name endoDataName = 'Endo getDualValName :: Name getDualValName = 'getDual #if !(MIN_VERSION_base(4,11,0)) readField :: String -> ReadPrec a -> ReadPrec a readField fieldName readVal = do expectP (L.Ident fieldName) expectP (L.Punc "=") readVal {-# NOINLINE readField #-} readSymField :: String -> ReadPrec a -> ReadPrec a readSymField fieldName readVal = do expectP (L.Punc "(") expectP (L.Symbol fieldName) expectP (L.Punc ")") expectP (L.Punc "=") readVal {-# NOINLINE readSymField #-} #endif readFieldValName :: Name readFieldValName = 'readField readSymFieldValName :: Name readSymFieldValName = 'readSymField #if MIN_VERSION_base(4,13,0) eqInt8HashValName :: Name eqInt8HashValName = 'eqInt8# eqInt16HashValName :: Name eqInt16HashValName = 'eqInt16# eqWord8HashValName :: Name eqWord8HashValName = 'eqWord8# eqWord16HashValName :: Name eqWord16HashValName = 'eqWord16# geInt8HashValName :: Name geInt8HashValName = 'geInt8# geInt16HashValName :: Name geInt16HashValName = 'geInt16# geWord8HashValName :: Name geWord8HashValName = 'geWord8# geWord16HashValName :: Name geWord16HashValName = 'geWord16# gtInt8HashValName :: Name gtInt8HashValName = 'gtInt8# gtInt16HashValName :: Name gtInt16HashValName = 'gtInt16# gtWord8HashValName :: Name gtWord8HashValName = 'gtWord8# gtWord16HashValName :: Name gtWord16HashValName = 'gtWord16# int8HashTypeName :: Name int8HashTypeName = ''Int8# int8ToIntHashValName :: Name int8ToIntHashValName = # if MIN_VERSION_base(4,16,0) 'int8ToInt# # else 'extendInt8# # endif int16HashTypeName :: Name int16HashTypeName = ''Int16# int16ToIntHashValName :: Name int16ToIntHashValName = # if MIN_VERSION_base(4,16,0) 'int16ToInt# # else 'extendInt16# # endif intToInt8HashValName :: Name intToInt8HashValName = # if MIN_VERSION_base(4,16,0) 'intToInt8# # else 'narrowInt8# # endif intToInt16HashValName :: Name intToInt16HashValName = # if MIN_VERSION_base(4,16,0) 'intToInt16# # else 'narrowInt16# # endif leInt8HashValName :: Name leInt8HashValName = 'leInt8# leInt16HashValName :: Name leInt16HashValName = 'leInt16# leWord8HashValName :: Name leWord8HashValName = 'leWord8# leWord16HashValName :: Name leWord16HashValName = 'leWord16# ltInt8HashValName :: Name ltInt8HashValName = 'ltInt8# ltInt16HashValName :: Name ltInt16HashValName = 'ltInt16# ltWord8HashValName :: Name ltWord8HashValName = 'ltWord8# ltWord16HashValName :: Name ltWord16HashValName = 'ltWord16# word8HashTypeName :: Name word8HashTypeName = ''Word8# word8ToWordHashValName :: Name word8ToWordHashValName = # if MIN_VERSION_base(4,16,0) 'word8ToWord# # else 'extendWord8# # endif word16HashTypeName :: Name word16HashTypeName = ''Word16# word16ToWordHashValName :: Name word16ToWordHashValName = # if MIN_VERSION_base(4,16,0) 'word16ToWord# # else 'extendWord16# # endif wordToWord8HashValName :: Name wordToWord8HashValName = # if MIN_VERSION_base(4,16,0) 'wordToWord8# # else 'narrowWord8# # endif wordToWord16HashValName :: Name wordToWord16HashValName = # if MIN_VERSION_base(4,16,0) 'wordToWord16# # else 'narrowWord16# # endif #endif #if MIN_VERSION_base(4,16,0) eqInt32HashValName :: Name eqInt32HashValName = 'eqInt32# eqWord32HashValName :: Name eqWord32HashValName = 'eqWord32# geInt32HashValName :: Name geInt32HashValName = 'geInt32# geWord32HashValName :: Name geWord32HashValName = 'geWord32# gtInt32HashValName :: Name gtInt32HashValName = 'gtInt32# gtWord32HashValName :: Name gtWord32HashValName = 'gtWord32# int32HashTypeName :: Name int32HashTypeName = ''Int32# int32ToIntHashValName :: Name int32ToIntHashValName = 'int32ToInt# intToInt32HashValName :: Name intToInt32HashValName = 'intToInt32# leInt32HashValName :: Name leInt32HashValName = 'leInt32# leWord32HashValName :: Name leWord32HashValName = 'leWord32# ltInt32HashValName :: Name ltInt32HashValName = 'ltInt32# ltWord32HashValName :: Name ltWord32HashValName = 'ltWord32# word32HashTypeName :: Name word32HashTypeName = ''Word32# word32ToWordHashValName :: Name word32ToWordHashValName = 'word32ToWord# wordToWord32HashValName :: Name wordToWord32HashValName = 'wordToWord32# #endif #if MIN_VERSION_base(4,19,0) i8HashDataName :: Name i8HashDataName = 'I8# i16HashDataName :: Name i16HashDataName = 'I16# i32HashDataName :: Name i32HashDataName = 'I32# i64HashDataName :: Name i64HashDataName = 'I64# int64HashTypeName :: Name int64HashTypeName = ''Int64# w8HashDataName :: Name w8HashDataName = 'W8# w16HashDataName :: Name w16HashDataName = 'W16# w32HashDataName :: Name w32HashDataName = 'W32# w64HashDataName :: Name w64HashDataName = 'W64# word64HashTypeName :: Name word64HashTypeName = ''Word64# #endif