{- (c) The University of Glasgow 2006 (c) The AQUA Project, Glasgow University, 1998 This module contains definitions for the IdInfo for things that have a standard form, namely: - data constructors - record selectors - method and superclass selectors - primitive operations -} {-# LANGUAGE CPP #-} module MkId ( mkDictFunId, mkDictFunTy, mkDictSelId, mkDictSelRhs, mkPrimOpId, mkFCallId, unwrapNewTypeBody, wrapFamInstBody, DataConBoxer(..), mkDataConRep, mkDataConWorkId, -- And some particular Ids; see below for why they are wired in wiredInIds, ghcPrimIds, unsafeCoerceName, unsafeCoerceId, realWorldPrimId, voidPrimId, voidArgId, nullAddrId, seqId, lazyId, lazyIdKey, coercionTokenId, magicDictId, coerceId, proxyHashId, noinlineId, noinlineIdName, -- Re-export error Ids module PrelRules ) where #include "HsVersions.h" import GhcPrelude import Rules import TysPrim import TysWiredIn import PrelRules import Type import FamInstEnv import Coercion import TcType import MkCore import CoreUtils ( exprType, mkCast ) import CoreUnfold import Literal import TyCon import Class import NameSet import Name import PrimOp import ForeignCall import DataCon import Id import IdInfo import Demand import CoreSyn import Unique import UniqSupply import PrelNames import BasicTypes hiding ( SuccessFlag(..) ) import Util import Pair import DynFlags import Outputable import FastString import ListSetOps import qualified GHC.LanguageExtensions as LangExt import Data.Maybe ( maybeToList ) {- ************************************************************************ * * \subsection{Wired in Ids} * * ************************************************************************ Note [Wired-in Ids] ~~~~~~~~~~~~~~~~~~~ A "wired-in" Id can be referred to directly in GHC (e.g. 'voidPrimId') rather than by looking it up its name in some environment or fetching it from an interface file. There are several reasons why an Id might appear in the wiredInIds: * ghcPrimIds: see Note [ghcPrimIds (aka pseudoops)] * magicIds: see Note [magicIds] * errorIds, defined in coreSyn/MkCore.hs. These error functions (e.g. rUNTIME_ERROR_ID) are wired in because the desugarer generates code that mentions them directly In all cases except ghcPrimIds, there is a definition site in a library module, which may be called (e.g. in higher order situations); but the wired-in version means that the details are never read from that module's interface file; instead, the full definition is right here. Note [ghcPrimIds (aka pseudoops)] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ The ghcPrimIds * Are exported from GHC.Prim * Can't be defined in Haskell, and hence no Haskell binding site, but have perfectly reasonable unfoldings in Core * Either have a CompulsoryUnfolding (hence always inlined), or of an EvaldUnfolding and void representation (e.g. void#) * Are (or should be) defined in primops.txt.pp as 'pseudoop' Reason: that's how we generate documentation for them Note [magicIds] ~~~~~~~~~~~~~~~ The magicIds * Are exported from GHC.Magic * Can be defined in Haskell (and are, in ghc-prim:GHC/Magic.hs). This definition at least generates Haddock documentation for them. * May or may not have a CompulsoryUnfolding. * But have some special behaviour that can't be done via an unfolding from an interface file -} wiredInIds :: [Id] wiredInIds = magicIds ++ ghcPrimIds ++ errorIds -- Defined in MkCore magicIds :: [Id] -- See Note [magicIds] magicIds = [lazyId, oneShotId, noinlineId] ghcPrimIds :: [Id] -- See Note [ghcPrimIds (aka pseudoops)] ghcPrimIds = [ realWorldPrimId , voidPrimId , unsafeCoerceId , nullAddrId , seqId , magicDictId , coerceId , proxyHashId ] {- ************************************************************************ * * \subsection{Data constructors} * * ************************************************************************ The wrapper for a constructor is an ordinary top-level binding that evaluates any strict args, unboxes any args that are going to be flattened, and calls the worker. We're going to build a constructor that looks like: data (Data a, C b) => T a b = T1 !a !Int b T1 = /\ a b -> \d1::Data a, d2::C b -> \p q r -> case p of { p -> case q of { q -> Con T1 [a,b] [p,q,r]}} Notice that * d2 is thrown away --- a context in a data decl is used to make sure one *could* construct dictionaries at the site the constructor is used, but the dictionary isn't actually used. * We have to check that we can construct Data dictionaries for the types a and Int. Once we've done that we can throw d1 away too. * We use (case p of q -> ...) to evaluate p, rather than "seq" because all that matters is that the arguments are evaluated. "seq" is very careful to preserve evaluation order, which we don't need to be here. You might think that we could simply give constructors some strictness info, like PrimOps, and let CoreToStg do the let-to-case transformation. But we don't do that because in the case of primops and functions strictness is a *property* not a *requirement*. In the case of constructors we need to do something active to evaluate the argument. Making an explicit case expression allows the simplifier to eliminate it in the (common) case where the constructor arg is already evaluated. Note [Wrappers for data instance tycons] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ In the case of data instances, the wrapper also applies the coercion turning the representation type into the family instance type to cast the result of the wrapper. For example, consider the declarations data family Map k :: * -> * data instance Map (a, b) v = MapPair (Map a (Pair b v)) The tycon to which the datacon MapPair belongs gets a unique internal name of the form :R123Map, and we call it the representation tycon. In contrast, Map is the family tycon (accessible via tyConFamInst_maybe). A coercion allows you to move between representation and family type. It is accessible from :R123Map via tyConFamilyCoercion_maybe and has kind Co123Map a b v :: {Map (a, b) v ~ :R123Map a b v} The wrapper and worker of MapPair get the types -- Wrapper $WMapPair :: forall a b v. Map a (Map a b v) -> Map (a, b) v $WMapPair a b v = MapPair a b v `cast` sym (Co123Map a b v) -- Worker MapPair :: forall a b v. Map a (Map a b v) -> :R123Map a b v This coercion is conditionally applied by wrapFamInstBody. It's a bit more complicated if the data instance is a GADT as well! data instance T [a] where T1 :: forall b. b -> T [Maybe b] Hence we translate to -- Wrapper $WT1 :: forall b. b -> T [Maybe b] $WT1 b v = T1 (Maybe b) b (Maybe b) v `cast` sym (Co7T (Maybe b)) -- Worker T1 :: forall c b. (c ~ Maybe b) => b -> :R7T c -- Coercion from family type to representation type Co7T a :: T [a] ~ :R7T a Newtype instances through an additional wrinkle into the mix. Consider the following example (adapted from #15318, comment:2): data family T a newtype instance T [a] = MkT [a] Within the newtype instance, there are three distinct types at play: 1. The newtype's underlying type, [a]. 2. The instance's representation type, TList a (where TList is the representation tycon). 3. The family type, T [a]. We need two coercions in order to cast from (1) to (3): (a) A newtype coercion axiom: axiom coTList a :: TList a ~ [a] (Where TList is the representation tycon of the newtype instance.) (b) A data family instance coercion axiom: axiom coT a :: T [a] ~ TList a When we translate the newtype instance to Core, we obtain: -- Wrapper $WMkT :: forall a. [a] -> T [a] $WMkT a x = MkT a x |> Sym (coT a) -- Worker MkT :: forall a. [a] -> TList [a] MkT a x = x |> Sym (coTList a) Unlike for data instances, the worker for a newtype instance is actually an executable function which expands to a cast, but otherwise, the general strategy is essentially the same as for data instances. Also note that we have a wrapper, which is unusual for a newtype, but we make GHC produce one anyway for symmetry with the way data instances are handled. Note [Newtype datacons] ~~~~~~~~~~~~~~~~~~~~~~~ The "data constructor" for a newtype should always be vanilla. At one point this wasn't true, because the newtype arising from class C a => D a looked like newtype T:D a = D:D (C a) so the data constructor for T:C had a single argument, namely the predicate (C a). But now we treat that as an ordinary argument, not part of the theta-type, so all is well. ************************************************************************ * * \subsection{Dictionary selectors} * * ************************************************************************ Selecting a field for a dictionary. If there is just one field, then there's nothing to do. Dictionary selectors may get nested forall-types. Thus: class Foo a where op :: forall b. Ord b => a -> b -> b Then the top-level type for op is op :: forall a. Foo a => forall b. Ord b => a -> b -> b -} mkDictSelId :: Name -- Name of one of the *value* selectors -- (dictionary superclass or method) -> Class -> Id mkDictSelId name clas = mkGlobalId (ClassOpId clas) name sel_ty info where tycon = classTyCon clas sel_names = map idName (classAllSelIds clas) new_tycon = isNewTyCon tycon [data_con] = tyConDataCons tycon tyvars = dataConUserTyVarBinders data_con n_ty_args = length tyvars arg_tys = dataConRepArgTys data_con -- Includes the dictionary superclasses val_index = assoc "MkId.mkDictSelId" (sel_names `zip` [0..]) name sel_ty = mkForAllTys tyvars $ mkFunTy (mkClassPred clas (mkTyVarTys (binderVars tyvars))) $ getNth arg_tys val_index base_info = noCafIdInfo `setArityInfo` 1 `setStrictnessInfo` strict_sig `setLevityInfoWithType` sel_ty info | new_tycon = base_info `setInlinePragInfo` alwaysInlinePragma `setUnfoldingInfo` mkInlineUnfoldingWithArity 1 (mkDictSelRhs clas val_index) -- See Note [Single-method classes] in TcInstDcls -- for why alwaysInlinePragma | otherwise = base_info `setRuleInfo` mkRuleInfo [rule] -- Add a magic BuiltinRule, but no unfolding -- so that the rule is always available to fire. -- See Note [ClassOp/DFun selection] in TcInstDcls -- This is the built-in rule that goes -- op (dfT d1 d2) ---> opT d1 d2 rule = BuiltinRule { ru_name = fsLit "Class op " `appendFS` occNameFS (getOccName name) , ru_fn = name , ru_nargs = n_ty_args + 1 , ru_try = dictSelRule val_index n_ty_args } -- The strictness signature is of the form U(AAAVAAAA) -> T -- where the V depends on which item we are selecting -- It's worth giving one, so that absence info etc is generated -- even if the selector isn't inlined strict_sig = mkClosedStrictSig [arg_dmd] topRes arg_dmd | new_tycon = evalDmd | otherwise = mkManyUsedDmd $ mkProdDmd [ if name == sel_name then evalDmd else absDmd | sel_name <- sel_names ] mkDictSelRhs :: Class -> Int -- 0-indexed selector among (superclasses ++ methods) -> CoreExpr mkDictSelRhs clas val_index = mkLams tyvars (Lam dict_id rhs_body) where tycon = classTyCon clas new_tycon = isNewTyCon tycon [data_con] = tyConDataCons tycon tyvars = dataConUnivTyVars data_con arg_tys = dataConRepArgTys data_con -- Includes the dictionary superclasses the_arg_id = getNth arg_ids val_index pred = mkClassPred clas (mkTyVarTys tyvars) dict_id = mkTemplateLocal 1 pred arg_ids = mkTemplateLocalsNum 2 arg_tys rhs_body | new_tycon = unwrapNewTypeBody tycon (mkTyVarTys tyvars) (Var dict_id) | otherwise = Case (Var dict_id) dict_id (idType the_arg_id) [(DataAlt data_con, arg_ids, varToCoreExpr the_arg_id)] -- varToCoreExpr needed for equality superclass selectors -- sel a b d = case x of { MkC _ (g:a~b) _ -> CO g } dictSelRule :: Int -> Arity -> RuleFun -- Tries to persuade the argument to look like a constructor -- application, using exprIsConApp_maybe, and then selects -- from it -- sel_i t1..tk (D t1..tk op1 ... opm) = opi -- dictSelRule val_index n_ty_args _ id_unf _ args | (dict_arg : _) <- drop n_ty_args args , Just (_, _, con_args) <- exprIsConApp_maybe id_unf dict_arg = Just (getNth con_args val_index) | otherwise = Nothing {- ************************************************************************ * * Data constructors * * ************************************************************************ -} mkDataConWorkId :: Name -> DataCon -> Id mkDataConWorkId wkr_name data_con | isNewTyCon tycon = mkGlobalId (DataConWrapId data_con) wkr_name wkr_ty nt_work_info | otherwise = mkGlobalId (DataConWorkId data_con) wkr_name wkr_ty alg_wkr_info where tycon = dataConTyCon data_con -- The representation TyCon wkr_ty = dataConRepType data_con ----------- Workers for data types -------------- alg_wkr_info = noCafIdInfo `setArityInfo` wkr_arity `setStrictnessInfo` wkr_sig `setUnfoldingInfo` evaldUnfolding -- Record that it's evaluated, -- even if arity = 0 `setLevityInfoWithType` wkr_ty -- NB: unboxed tuples have workers, so we can't use -- setNeverLevPoly wkr_arity = dataConRepArity data_con wkr_sig = mkClosedStrictSig (replicate wkr_arity topDmd) (dataConCPR data_con) -- Note [Data-con worker strictness] -- Notice that we do *not* say the worker Id is strict -- even if the data constructor is declared strict -- e.g. data T = MkT !(Int,Int) -- Why? Because the *wrapper* $WMkT is strict (and its unfolding has -- case expressions that do the evals) but the *worker* MkT itself is -- not. If we pretend it is strict then when we see -- case x of y -> MkT y -- the simplifier thinks that y is "sure to be evaluated" (because -- the worker MkT is strict) and drops the case. No, the workerId -- MkT is not strict. -- -- However, the worker does have StrictnessMarks. When the simplifier -- sees a pattern -- case e of MkT x -> ... -- it uses the dataConRepStrictness of MkT to mark x as evaluated; -- but that's fine... dataConRepStrictness comes from the data con -- not from the worker Id. ----------- Workers for newtypes -------------- univ_tvs = dataConUnivTyVars data_con arg_tys = dataConRepArgTys data_con -- Should be same as dataConOrigArgTys nt_work_info = noCafIdInfo -- The NoCaf-ness is set by noCafIdInfo `setArityInfo` 1 -- Arity 1 `setInlinePragInfo` alwaysInlinePragma `setUnfoldingInfo` newtype_unf `setLevityInfoWithType` wkr_ty id_arg1 = mkTemplateLocal 1 (head arg_tys) res_ty_args = mkTyVarTys univ_tvs newtype_unf = ASSERT2( isVanillaDataCon data_con && isSingleton arg_tys , ppr data_con ) -- Note [Newtype datacons] mkCompulsoryUnfolding $ mkLams univ_tvs $ Lam id_arg1 $ wrapNewTypeBody tycon res_ty_args (Var id_arg1) dataConCPR :: DataCon -> DmdResult dataConCPR con | isDataTyCon tycon -- Real data types only; that is, -- not unboxed tuples or newtypes , null (dataConExTyVars con) -- No existentials , wkr_arity > 0 , wkr_arity <= mAX_CPR_SIZE = if is_prod then vanillaCprProdRes (dataConRepArity con) else cprSumRes (dataConTag con) | otherwise = topRes where is_prod = isProductTyCon tycon tycon = dataConTyCon con wkr_arity = dataConRepArity con mAX_CPR_SIZE :: Arity mAX_CPR_SIZE = 10 -- We do not treat very big tuples as CPR-ish: -- a) for a start we get into trouble because there aren't -- "enough" unboxed tuple types (a tiresome restriction, -- but hard to fix), -- b) more importantly, big unboxed tuples get returned mainly -- on the stack, and are often then allocated in the heap -- by the caller. So doing CPR for them may in fact make -- things worse. {- ------------------------------------------------- -- Data constructor representation -- -- This is where we decide how to wrap/unwrap the -- constructor fields -- -------------------------------------------------- -} type Unboxer = Var -> UniqSM ([Var], CoreExpr -> CoreExpr) -- Unbox: bind rep vars by decomposing src var data Boxer = UnitBox | Boxer (TCvSubst -> UniqSM ([Var], CoreExpr)) -- Box: build src arg using these rep vars -- | Data Constructor Boxer newtype DataConBoxer = DCB ([Type] -> [Var] -> UniqSM ([Var], [CoreBind])) -- Bind these src-level vars, returning the -- rep-level vars to bind in the pattern {- Note [Inline partially-applied constructor wrappers] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ We allow the wrapper to inline when partially applied to avoid boxing values unnecessarily. For example, consider data Foo a = Foo !Int a instance Traversable Foo where traverse f (Foo i a) = Foo i <$> f a This desugars to traverse f foo = case foo of Foo i# a -> let i = I# i# in map ($WFoo i) (f a) If the wrapper `$WFoo` is not inlined, we get a fruitless reboxing of `i`. But if we inline the wrapper, we get map (\a. case i of I# i# a -> Foo i# a) (f a) and now case-of-known-constructor eliminates the redundant allocation. -} mkDataConRep :: DynFlags -> FamInstEnvs -> Name -> Maybe [HsImplBang] -- See Note [Bangs on imported data constructors] -> DataCon -> UniqSM DataConRep mkDataConRep dflags fam_envs wrap_name mb_bangs data_con | not wrapper_reqd = return NoDataConRep | otherwise = do { wrap_args <- mapM newLocal wrap_arg_tys ; wrap_body <- mk_rep_app (wrap_args `zip` dropList eq_spec unboxers) initial_wrap_app ; let wrap_id = mkGlobalId (DataConWrapId data_con) wrap_name wrap_ty wrap_info wrap_info = noCafIdInfo `setArityInfo` wrap_arity -- It's important to specify the arity, so that partial -- applications are treated as values `setInlinePragInfo` wrap_prag `setUnfoldingInfo` wrap_unf `setStrictnessInfo` wrap_sig -- We need to get the CAF info right here because TidyPgm -- does not tidy the IdInfo of implicit bindings (like the wrapper) -- so it not make sure that the CAF info is sane `setNeverLevPoly` wrap_ty wrap_sig = mkClosedStrictSig wrap_arg_dmds (dataConCPR data_con) wrap_arg_dmds = replicate (length theta) topDmd ++ map mk_dmd arg_ibangs -- Don't forget the dictionary arguments when building -- the strictness signature (#14290). mk_dmd str | isBanged str = evalDmd | otherwise = topDmd wrap_prag = alwaysInlinePragma `setInlinePragmaActivation` ActiveAfter NoSourceText 2 -- See Note [Activation for data constructor wrappers] -- The wrapper will usually be inlined (see wrap_unf), so its -- strictness and CPR info is usually irrelevant. But this is -- not always the case; GHC may choose not to inline it. In -- particular, the wrapper constructor is not inlined inside -- an INLINE rhs or when it is not applied to any arguments. -- See Note [Inline partially-applied constructor wrappers] -- Passing Nothing here allows the wrapper to inline when -- unsaturated. wrap_unf = mkInlineUnfolding wrap_rhs wrap_rhs = mkLams wrap_tvs $ mkLams wrap_args $ wrapFamInstBody tycon res_ty_args $ wrap_body ; return (DCR { dcr_wrap_id = wrap_id , dcr_boxer = mk_boxer boxers , dcr_arg_tys = rep_tys , dcr_stricts = rep_strs , dcr_bangs = arg_ibangs }) } where (univ_tvs, ex_tvs, eq_spec, theta, orig_arg_tys, _orig_res_ty) = dataConFullSig data_con wrap_tvs = dataConUserTyVars data_con res_ty_args = substTyVars (mkTvSubstPrs (map eqSpecPair eq_spec)) univ_tvs tycon = dataConTyCon data_con -- The representation TyCon (not family) wrap_ty = dataConUserType data_con ev_tys = eqSpecPreds eq_spec ++ theta all_arg_tys = ev_tys ++ orig_arg_tys ev_ibangs = map (const HsLazy) ev_tys orig_bangs = dataConSrcBangs data_con wrap_arg_tys = theta ++ orig_arg_tys wrap_arity = length wrap_arg_tys -- The wrap_args are the arguments *other than* the eq_spec -- Because we are going to apply the eq_spec args manually in the -- wrapper arg_ibangs = case mb_bangs of Nothing -> zipWith (dataConSrcToImplBang dflags fam_envs) orig_arg_tys orig_bangs Just bangs -> bangs (rep_tys_w_strs, wrappers) = unzip (zipWith dataConArgRep all_arg_tys (ev_ibangs ++ arg_ibangs)) (unboxers, boxers) = unzip wrappers (rep_tys, rep_strs) = unzip (concat rep_tys_w_strs) wrapper_reqd = (not (isNewTyCon tycon) -- (Most) newtypes have only a worker, with the exception -- of some newtypes written with GADT syntax. See below. && (any isBanged (ev_ibangs ++ arg_ibangs) -- Some forcing/unboxing (includes eq_spec) || (not $ null eq_spec))) -- GADT || isFamInstTyCon tycon -- Cast result || dataConUserTyVarsArePermuted data_con -- If the data type was written with GADT syntax and -- orders the type variables differently from what the -- worker expects, it needs a data con wrapper to reorder -- the type variables. -- See Note [Data con wrappers and GADT syntax]. initial_wrap_app = Var (dataConWorkId data_con) `mkTyApps` res_ty_args `mkVarApps` ex_tvs `mkCoApps` map (mkReflCo Nominal . eqSpecType) eq_spec mk_boxer :: [Boxer] -> DataConBoxer mk_boxer boxers = DCB (\ ty_args src_vars -> do { let (ex_vars, term_vars) = splitAtList ex_tvs src_vars subst1 = zipTvSubst univ_tvs ty_args subst2 = extendTvSubstList subst1 ex_tvs (mkTyVarTys ex_vars) ; (rep_ids, binds) <- go subst2 boxers term_vars ; return (ex_vars ++ rep_ids, binds) } ) go _ [] src_vars = ASSERT2( null src_vars, ppr data_con ) return ([], []) go subst (UnitBox : boxers) (src_var : src_vars) = do { (rep_ids2, binds) <- go subst boxers src_vars ; return (src_var : rep_ids2, binds) } go subst (Boxer boxer : boxers) (src_var : src_vars) = do { (rep_ids1, arg) <- boxer subst ; (rep_ids2, binds) <- go subst boxers src_vars ; return (rep_ids1 ++ rep_ids2, NonRec src_var arg : binds) } go _ (_:_) [] = pprPanic "mk_boxer" (ppr data_con) mk_rep_app :: [(Id,Unboxer)] -> CoreExpr -> UniqSM CoreExpr mk_rep_app [] con_app = return con_app mk_rep_app ((wrap_arg, unboxer) : prs) con_app = do { (rep_ids, unbox_fn) <- unboxer wrap_arg ; expr <- mk_rep_app prs (mkVarApps con_app rep_ids) ; return (unbox_fn expr) } {- Note [Activation for data constructor wrappers] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ The Activation on a data constructor wrapper allows it to inline in Phase 2 and later (1, 0). But not in the InitialPhase. That gives rewrite rules a chance to fire (in the InitialPhase) if they mention a data constructor on the left RULE "foo" f (K a b) = ... Since the LHS of rules are simplified with InitialPhase, we won't inline the wrapper on the LHS either. People have asked for this before, but now that even the InitialPhase does some inlining, it has become important. Note [Bangs on imported data constructors] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ We pass Maybe [HsImplBang] to mkDataConRep to make use of HsImplBangs from imported modules. - Nothing <=> use HsSrcBangs - Just bangs <=> use HsImplBangs For imported types we can't work it all out from the HsSrcBangs, because we want to be very sure to follow what the original module (where the data type was declared) decided, and that depends on what flags were enabled when it was compiled. So we record the decisions in the interface file. The HsImplBangs passed are in 1-1 correspondence with the dataConOrigArgTys of the DataCon. Note [Data con wrappers and unlifted types] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Consider data T = MkT !Int# We certainly do not want to make a wrapper $WMkT x = case x of y { DEFAULT -> MkT y } For a start, it's still to generate a no-op. But worse, since wrappers are currently injected at TidyCore, we don't even optimise it away! So the stupid case expression stays there. This actually happened for the Integer data type (see Trac #1600 comment:66)! Note [Data con wrappers and GADT syntax] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Consider these two very similar data types: data T1 a b = MkT1 b data T2 a b where MkT2 :: forall b a. b -> T2 a b Despite their similar appearance, T2 will have a data con wrapper but T1 will not. What sets them apart? The types of their constructors, which are: MkT1 :: forall a b. b -> T1 a b MkT2 :: forall b a. b -> T2 a b MkT2's use of GADT syntax allows it to permute the order in which `a` and `b` would normally appear. See Note [DataCon user type variable binders] in DataCon for further discussion on this topic. The worker data cons for T1 and T2, however, both have types such that `a` is expected to come before `b` as arguments. Because MkT2 permutes this order, it needs a data con wrapper to swizzle around the type variables to be in the order the worker expects. A somewhat surprising consequence of this is that *newtypes* can have data con wrappers! After all, a newtype can also be written with GADT syntax: newtype T3 a b where MkT3 :: forall b a. b -> T3 a b Again, this needs a wrapper data con to reorder the type variables. It does mean that this newtype constructor requires another level of indirection when being called, but the inliner should make swift work of that. -} ------------------------- newLocal :: Type -> UniqSM Var newLocal ty = do { uniq <- getUniqueM ; return (mkSysLocalOrCoVar (fsLit "dt") uniq ty) } -- | Unpack/Strictness decisions from source module dataConSrcToImplBang :: DynFlags -> FamInstEnvs -> Type -> HsSrcBang -> HsImplBang dataConSrcToImplBang dflags fam_envs arg_ty (HsSrcBang ann unpk NoSrcStrict) | xopt LangExt.StrictData dflags -- StrictData => strict field = dataConSrcToImplBang dflags fam_envs arg_ty (HsSrcBang ann unpk SrcStrict) | otherwise -- no StrictData => lazy field = HsLazy dataConSrcToImplBang _ _ _ (HsSrcBang _ _ SrcLazy) = HsLazy dataConSrcToImplBang dflags fam_envs arg_ty (HsSrcBang _ unpk_prag SrcStrict) | isUnliftedType arg_ty = HsLazy -- For !Int#, say, use HsLazy -- See Note [Data con wrappers and unlifted types] | not (gopt Opt_OmitInterfacePragmas dflags) -- Don't unpack if -fomit-iface-pragmas -- Don't unpack if we aren't optimising; rather arbitrarily, -- we use -fomit-iface-pragmas as the indication , let mb_co = topNormaliseType_maybe fam_envs arg_ty -- Unwrap type families and newtypes arg_ty' = case mb_co of { Just (_,ty) -> ty; Nothing -> arg_ty } , isUnpackableType dflags fam_envs arg_ty' , (rep_tys, _) <- dataConArgUnpack arg_ty' , case unpk_prag of NoSrcUnpack -> gopt Opt_UnboxStrictFields dflags || (gopt Opt_UnboxSmallStrictFields dflags && rep_tys `lengthAtMost` 1) -- See Note [Unpack one-wide fields] srcUnpack -> isSrcUnpacked srcUnpack = case mb_co of Nothing -> HsUnpack Nothing Just (co,_) -> HsUnpack (Just co) | otherwise -- Record the strict-but-no-unpack decision = HsStrict -- | Wrappers/Workers and representation following Unpack/Strictness -- decisions dataConArgRep :: Type -> HsImplBang -> ([(Type,StrictnessMark)] -- Rep types ,(Unboxer,Boxer)) dataConArgRep arg_ty HsLazy = ([(arg_ty, NotMarkedStrict)], (unitUnboxer, unitBoxer)) dataConArgRep arg_ty HsStrict = ([(arg_ty, MarkedStrict)], (seqUnboxer, unitBoxer)) dataConArgRep arg_ty (HsUnpack Nothing) | (rep_tys, wrappers) <- dataConArgUnpack arg_ty = (rep_tys, wrappers) dataConArgRep _ (HsUnpack (Just co)) | let co_rep_ty = pSnd (coercionKind co) , (rep_tys, wrappers) <- dataConArgUnpack co_rep_ty = (rep_tys, wrapCo co co_rep_ty wrappers) ------------------------- wrapCo :: Coercion -> Type -> (Unboxer, Boxer) -> (Unboxer, Boxer) wrapCo co rep_ty (unbox_rep, box_rep) -- co :: arg_ty ~ rep_ty = (unboxer, boxer) where unboxer arg_id = do { rep_id <- newLocal rep_ty ; (rep_ids, rep_fn) <- unbox_rep rep_id ; let co_bind = NonRec rep_id (Var arg_id `Cast` co) ; return (rep_ids, Let co_bind . rep_fn) } boxer = Boxer $ \ subst -> do { (rep_ids, rep_expr) <- case box_rep of UnitBox -> do { rep_id <- newLocal (TcType.substTy subst rep_ty) ; return ([rep_id], Var rep_id) } Boxer boxer -> boxer subst ; let sco = substCoUnchecked subst co ; return (rep_ids, rep_expr `Cast` mkSymCo sco) } ------------------------ seqUnboxer :: Unboxer seqUnboxer v = return ([v], \e -> Case (Var v) v (exprType e) [(DEFAULT, [], e)]) unitUnboxer :: Unboxer unitUnboxer v = return ([v], \e -> e) unitBoxer :: Boxer unitBoxer = UnitBox ------------------------- dataConArgUnpack :: Type -> ( [(Type, StrictnessMark)] -- Rep types , (Unboxer, Boxer) ) dataConArgUnpack arg_ty | Just (tc, tc_args) <- splitTyConApp_maybe arg_ty , Just con <- tyConSingleAlgDataCon_maybe tc -- NB: check for an *algebraic* data type -- A recursive newtype might mean that -- 'arg_ty' is a newtype , let rep_tys = dataConInstArgTys con tc_args = ASSERT( null (dataConExTyVars con) ) -- Note [Unpacking GADTs and existentials] ( rep_tys `zip` dataConRepStrictness con ,( \ arg_id -> do { rep_ids <- mapM newLocal rep_tys ; let unbox_fn body = Case (Var arg_id) arg_id (exprType body) [(DataAlt con, rep_ids, body)] ; return (rep_ids, unbox_fn) } , Boxer $ \ subst -> do { rep_ids <- mapM (newLocal . TcType.substTyUnchecked subst) rep_tys ; return (rep_ids, Var (dataConWorkId con) `mkTyApps` (substTysUnchecked subst tc_args) `mkVarApps` rep_ids ) } ) ) | otherwise = pprPanic "dataConArgUnpack" (ppr arg_ty) -- An interface file specified Unpacked, but we couldn't unpack it isUnpackableType :: DynFlags -> FamInstEnvs -> Type -> Bool -- True if we can unpack the UNPACK the argument type -- See Note [Recursive unboxing] -- We look "deeply" inside rather than relying on the DataCons -- we encounter on the way, because otherwise we might well -- end up relying on ourselves! isUnpackableType dflags fam_envs ty | Just data_con <- unpackable_type ty = ok_con_args emptyNameSet data_con | otherwise = False where ok_con_args dcs con | dc_name `elemNameSet` dcs = False | otherwise = all (ok_arg dcs') (dataConOrigArgTys con `zip` dataConSrcBangs con) -- NB: dataConSrcBangs gives the *user* request; -- We'd get a black hole if we used dataConImplBangs where dc_name = getName con dcs' = dcs `extendNameSet` dc_name ok_arg dcs (ty, bang) = not (attempt_unpack bang) || ok_ty dcs norm_ty where norm_ty = topNormaliseType fam_envs ty ok_ty dcs ty | Just data_con <- unpackable_type ty = ok_con_args dcs data_con | otherwise = True -- NB True here, in contrast to False at top level attempt_unpack (HsSrcBang _ SrcUnpack NoSrcStrict) = xopt LangExt.StrictData dflags attempt_unpack (HsSrcBang _ SrcUnpack SrcStrict) = True attempt_unpack (HsSrcBang _ NoSrcUnpack SrcStrict) = True -- Be conservative attempt_unpack (HsSrcBang _ NoSrcUnpack NoSrcStrict) = xopt LangExt.StrictData dflags -- Be conservative attempt_unpack _ = False unpackable_type :: Type -> Maybe DataCon -- Works just on a single level unpackable_type ty | Just (tc, _) <- splitTyConApp_maybe ty , Just data_con <- tyConSingleAlgDataCon_maybe tc , null (dataConExTyVars data_con) -- See Note [Unpacking GADTs and existentials] = Just data_con | otherwise = Nothing {- Note [Unpacking GADTs and existentials] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ There is nothing stopping us unpacking a data type with equality components, like data Equal a b where Equal :: Equal a a And it'd be fine to unpack a product type with existential components too, but that would require a bit more plumbing, so currently we don't. So for now we require: null (dataConExTyVars data_con) See Trac #14978 Note [Unpack one-wide fields] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ The flag UnboxSmallStrictFields ensures that any field that can (safely) be unboxed to a word-sized unboxed field, should be so unboxed. For example: data A = A Int# newtype B = B A data C = C !B data D = D !C data E = E !() data F = F !D data G = G !F !F All of these should have an Int# as their representation, except G which should have two Int#s. However data T = T !(S Int) data S = S !a Here we can represent T with an Int#. Note [Recursive unboxing] ~~~~~~~~~~~~~~~~~~~~~~~~~ Consider data R = MkR {-# UNPACK #-} !S Int data S = MkS {-# UNPACK #-} !Int The representation arguments of MkR are the *representation* arguments of S (plus Int); the rep args of MkS are Int#. This is all fine. But be careful not to try to unbox this! data T = MkT {-# UNPACK #-} !T Int Because then we'd get an infinite number of arguments. Here is a more complicated case: data S = MkS {-# UNPACK #-} !T Int data T = MkT {-# UNPACK #-} !S Int Each of S and T must decide independently whether to unpack and they had better not both say yes. So they must both say no. Also behave conservatively when there is no UNPACK pragma data T = MkS !T Int with -funbox-strict-fields or -funbox-small-strict-fields we need to behave as if there was an UNPACK pragma there. But it's the *argument* type that matters. This is fine: data S = MkS S !Int because Int is non-recursive. ************************************************************************ * * Wrapping and unwrapping newtypes and type families * * ************************************************************************ -} wrapNewTypeBody :: TyCon -> [Type] -> CoreExpr -> CoreExpr -- The wrapper for the data constructor for a newtype looks like this: -- newtype T a = MkT (a,Int) -- MkT :: forall a. (a,Int) -> T a -- MkT = /\a. \(x:(a,Int)). x `cast` sym (CoT a) -- where CoT is the coercion TyCon associated with the newtype -- -- The call (wrapNewTypeBody T [a] e) returns the -- body of the wrapper, namely -- e `cast` (CoT [a]) -- -- If a coercion constructor is provided in the newtype, then we use -- it, otherwise the wrap/unwrap are both no-ops wrapNewTypeBody tycon args result_expr = ASSERT( isNewTyCon tycon ) mkCast result_expr (mkSymCo co) where co = mkUnbranchedAxInstCo Representational (newTyConCo tycon) args [] -- When unwrapping, we do *not* apply any family coercion, because this will -- be done via a CoPat by the type checker. We have to do it this way as -- computing the right type arguments for the coercion requires more than just -- a spliting operation (cf, TcPat.tcConPat). unwrapNewTypeBody :: TyCon -> [Type] -> CoreExpr -> CoreExpr unwrapNewTypeBody tycon args result_expr = ASSERT( isNewTyCon tycon ) mkCast result_expr (mkUnbranchedAxInstCo Representational (newTyConCo tycon) args []) -- If the type constructor is a representation type of a data instance, wrap -- the expression into a cast adjusting the expression type, which is an -- instance of the representation type, to the corresponding instance of the -- family instance type. -- See Note [Wrappers for data instance tycons] wrapFamInstBody :: TyCon -> [Type] -> CoreExpr -> CoreExpr wrapFamInstBody tycon args body | Just co_con <- tyConFamilyCoercion_maybe tycon = mkCast body (mkSymCo (mkUnbranchedAxInstCo Representational co_con args [])) | otherwise = body {- ************************************************************************ * * \subsection{Primitive operations} * * ************************************************************************ -} mkPrimOpId :: PrimOp -> Id mkPrimOpId prim_op = id where (tyvars,arg_tys,res_ty, arity, strict_sig) = primOpSig prim_op ty = mkSpecForAllTys tyvars (mkFunTys arg_tys res_ty) name = mkWiredInName gHC_PRIM (primOpOcc prim_op) (mkPrimOpIdUnique (primOpTag prim_op)) (AnId id) UserSyntax id = mkGlobalId (PrimOpId prim_op) name ty info info = noCafIdInfo `setRuleInfo` mkRuleInfo (maybeToList $ primOpRules name prim_op) `setArityInfo` arity `setStrictnessInfo` strict_sig `setInlinePragInfo` neverInlinePragma `setLevityInfoWithType` res_ty -- We give PrimOps a NOINLINE pragma so that we don't -- get silly warnings from Desugar.dsRule (the inline_shadows_rule -- test) about a RULE conflicting with a possible inlining -- cf Trac #7287 -- For each ccall we manufacture a separate CCallOpId, giving it -- a fresh unique, a type that is correct for this particular ccall, -- and a CCall structure that gives the correct details about calling -- convention etc. -- -- The *name* of this Id is a local name whose OccName gives the full -- details of the ccall, type and all. This means that the interface -- file reader can reconstruct a suitable Id mkFCallId :: DynFlags -> Unique -> ForeignCall -> Type -> Id mkFCallId dflags uniq fcall ty = ASSERT( noFreeVarsOfType ty ) -- A CCallOpId should have no free type variables; -- when doing substitutions won't substitute over it mkGlobalId (FCallId fcall) name ty info where occ_str = showSDoc dflags (braces (ppr fcall <+> ppr ty)) -- The "occurrence name" of a ccall is the full info about the -- ccall; it is encoded, but may have embedded spaces etc! name = mkFCallName uniq occ_str info = noCafIdInfo `setArityInfo` arity `setStrictnessInfo` strict_sig `setLevityInfoWithType` ty (bndrs, _) = tcSplitPiTys ty arity = count isAnonTyBinder bndrs strict_sig = mkClosedStrictSig (replicate arity topDmd) topRes -- the call does not claim to be strict in its arguments, since they -- may be lifted (foreign import prim) and the called code doesn't -- necessarily force them. See Trac #11076. {- ************************************************************************ * * \subsection{DictFuns and default methods} * * ************************************************************************ Note [Dict funs and default methods] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Dict funs and default methods are *not* ImplicitIds. Their definition involves user-written code, so we can't figure out their strictness etc based on fixed info, as we can for constructors and record selectors (say). NB: See also Note [Exported LocalIds] in Id -} mkDictFunId :: Name -- Name to use for the dict fun; -> [TyVar] -> ThetaType -> Class -> [Type] -> Id -- Implements the DFun Superclass Invariant (see TcInstDcls) -- See Note [Dict funs and default methods] mkDictFunId dfun_name tvs theta clas tys = mkExportedLocalId (DFunId is_nt) dfun_name dfun_ty where is_nt = isNewTyCon (classTyCon clas) dfun_ty = mkDictFunTy tvs theta clas tys mkDictFunTy :: [TyVar] -> ThetaType -> Class -> [Type] -> Type mkDictFunTy tvs theta clas tys = mkSpecSigmaTy tvs theta (mkClassPred clas tys) {- ************************************************************************ * * \subsection{Un-definable} * * ************************************************************************ These Ids can't be defined in Haskell. They could be defined in unfoldings in the wired-in GHC.Prim interface file, but we'd have to ensure that they were definitely, definitely inlined, because there is no curried identifier for them. That's what mkCompulsoryUnfolding does. If we had a way to get a compulsory unfolding from an interface file, we could do that, but we don't right now. unsafeCoerce# isn't so much a PrimOp as a phantom identifier, that just gets expanded into a type coercion wherever it occurs. Hence we add it as a built-in Id with an unfolding here. The type variables we use here are "open" type variables: this means they can unify with both unlifted and lifted types. Hence we provide another gun with which to shoot yourself in the foot. -} unsafeCoerceName, nullAddrName, seqName, realWorldName, voidPrimIdName, coercionTokenName, magicDictName, coerceName, proxyName :: Name unsafeCoerceName = mkWiredInIdName gHC_PRIM (fsLit "unsafeCoerce#") unsafeCoerceIdKey unsafeCoerceId nullAddrName = mkWiredInIdName gHC_PRIM (fsLit "nullAddr#") nullAddrIdKey nullAddrId seqName = mkWiredInIdName gHC_PRIM (fsLit "seq") seqIdKey seqId realWorldName = mkWiredInIdName gHC_PRIM (fsLit "realWorld#") realWorldPrimIdKey realWorldPrimId voidPrimIdName = mkWiredInIdName gHC_PRIM (fsLit "void#") voidPrimIdKey voidPrimId coercionTokenName = mkWiredInIdName gHC_PRIM (fsLit "coercionToken#") coercionTokenIdKey coercionTokenId magicDictName = mkWiredInIdName gHC_PRIM (fsLit "magicDict") magicDictKey magicDictId coerceName = mkWiredInIdName gHC_PRIM (fsLit "coerce") coerceKey coerceId proxyName = mkWiredInIdName gHC_PRIM (fsLit "proxy#") proxyHashKey proxyHashId lazyIdName, oneShotName, noinlineIdName :: Name lazyIdName = mkWiredInIdName gHC_MAGIC (fsLit "lazy") lazyIdKey lazyId oneShotName = mkWiredInIdName gHC_MAGIC (fsLit "oneShot") oneShotKey oneShotId noinlineIdName = mkWiredInIdName gHC_MAGIC (fsLit "noinline") noinlineIdKey noinlineId ------------------------------------------------ proxyHashId :: Id proxyHashId = pcMiscPrelId proxyName ty (noCafIdInfo `setUnfoldingInfo` evaldUnfolding -- Note [evaldUnfoldings] `setNeverLevPoly` ty ) where -- proxy# :: forall k (a:k). Proxy# k a bndrs = mkTemplateKiTyVars [liftedTypeKind] (\ks -> ks) [k,t] = mkTyVarTys bndrs ty = mkSpecForAllTys bndrs (mkProxyPrimTy k t) ------------------------------------------------ unsafeCoerceId :: Id unsafeCoerceId = pcMiscPrelId unsafeCoerceName ty info where info = noCafIdInfo `setInlinePragInfo` alwaysInlinePragma `setUnfoldingInfo` mkCompulsoryUnfolding rhs -- unsafeCoerce# :: forall (r1 :: RuntimeRep) (r2 :: RuntimeRep) -- (a :: TYPE r1) (b :: TYPE r2). -- a -> b bndrs = mkTemplateKiTyVars [runtimeRepTy, runtimeRepTy] (\ks -> map tYPE ks) [_, _, a, b] = mkTyVarTys bndrs ty = mkSpecForAllTys bndrs (mkFunTy a b) [x] = mkTemplateLocals [a] rhs = mkLams (bndrs ++ [x]) $ Cast (Var x) (mkUnsafeCo Representational a b) ------------------------------------------------ nullAddrId :: Id -- nullAddr# :: Addr# -- The reason it is here is because we don't provide -- a way to write this literal in Haskell. nullAddrId = pcMiscPrelId nullAddrName addrPrimTy info where info = noCafIdInfo `setInlinePragInfo` alwaysInlinePragma `setUnfoldingInfo` mkCompulsoryUnfolding (Lit nullAddrLit) `setNeverLevPoly` addrPrimTy ------------------------------------------------ seqId :: Id -- See Note [seqId magic] seqId = pcMiscPrelId seqName ty info where info = noCafIdInfo `setInlinePragInfo` inline_prag `setUnfoldingInfo` mkCompulsoryUnfolding rhs `setNeverLevPoly` ty inline_prag = alwaysInlinePragma `setInlinePragmaActivation` ActiveAfter NoSourceText 0 -- Make 'seq' not inline-always, so that simpleOptExpr -- (see CoreSubst.simple_app) won't inline 'seq' on the -- LHS of rules. That way we can have rules for 'seq'; -- see Note [seqId magic] ty = mkSpecForAllTys [alphaTyVar,betaTyVar] (mkFunTy alphaTy (mkFunTy betaTy betaTy)) [x,y] = mkTemplateLocals [alphaTy, betaTy] rhs = mkLams [alphaTyVar,betaTyVar,x,y] (Case (Var x) x betaTy [(DEFAULT, [], Var y)]) ------------------------------------------------ lazyId :: Id -- See Note [lazyId magic] lazyId = pcMiscPrelId lazyIdName ty info where info = noCafIdInfo `setNeverLevPoly` ty ty = mkSpecForAllTys [alphaTyVar] (mkFunTy alphaTy alphaTy) noinlineId :: Id -- See Note [noinlineId magic] noinlineId = pcMiscPrelId noinlineIdName ty info where info = noCafIdInfo `setNeverLevPoly` ty ty = mkSpecForAllTys [alphaTyVar] (mkFunTy alphaTy alphaTy) oneShotId :: Id -- See Note [The oneShot function] oneShotId = pcMiscPrelId oneShotName ty info where info = noCafIdInfo `setInlinePragInfo` alwaysInlinePragma `setUnfoldingInfo` mkCompulsoryUnfolding rhs ty = mkSpecForAllTys [ runtimeRep1TyVar, runtimeRep2TyVar , openAlphaTyVar, openBetaTyVar ] (mkFunTy fun_ty fun_ty) fun_ty = mkFunTy openAlphaTy openBetaTy [body, x] = mkTemplateLocals [fun_ty, openAlphaTy] x' = setOneShotLambda x -- Here is the magic bit! rhs = mkLams [ runtimeRep1TyVar, runtimeRep2TyVar , openAlphaTyVar, openBetaTyVar , body, x'] $ Var body `App` Var x -------------------------------------------------------------------------------- magicDictId :: Id -- See Note [magicDictId magic] magicDictId = pcMiscPrelId magicDictName ty info where info = noCafIdInfo `setInlinePragInfo` neverInlinePragma `setNeverLevPoly` ty ty = mkSpecForAllTys [alphaTyVar] alphaTy -------------------------------------------------------------------------------- coerceId :: Id coerceId = pcMiscPrelId coerceName ty info where info = noCafIdInfo `setInlinePragInfo` alwaysInlinePragma `setUnfoldingInfo` mkCompulsoryUnfolding rhs `setNeverLevPoly` ty eqRTy = mkTyConApp coercibleTyCon [ liftedTypeKind , alphaTy, betaTy ] eqRPrimTy = mkTyConApp eqReprPrimTyCon [ liftedTypeKind , liftedTypeKind , alphaTy, betaTy ] ty = mkSpecForAllTys [alphaTyVar, betaTyVar] $ mkFunTys [eqRTy, alphaTy] betaTy [eqR,x,eq] = mkTemplateLocals [eqRTy, alphaTy, eqRPrimTy] rhs = mkLams [alphaTyVar, betaTyVar, eqR, x] $ mkWildCase (Var eqR) eqRTy betaTy $ [(DataAlt coercibleDataCon, [eq], Cast (Var x) (mkCoVarCo eq))] {- Note [Unsafe coerce magic] ~~~~~~~~~~~~~~~~~~~~~~~~~~ We define a *primitive* GHC.Prim.unsafeCoerce# and then in the base library we define the ordinary function Unsafe.Coerce.unsafeCoerce :: forall (a:*) (b:*). a -> b unsafeCoerce x = unsafeCoerce# x Notice that unsafeCoerce has a civilized (albeit still dangerous) polymorphic type, whose type args have kind *. So you can't use it on unboxed values (unsafeCoerce 3#). In contrast unsafeCoerce# is even more dangerous because you *can* use it on unboxed things, (unsafeCoerce# 3#) :: Int. Its type is forall (r1 :: RuntimeRep) (r2 :: RuntimeRep) (a: TYPE r1) (b: TYPE r2). a -> b Note [seqId magic] ~~~~~~~~~~~~~~~~~~ 'GHC.Prim.seq' is special in several ways. a) In source Haskell its second arg can have an unboxed type x `seq` (v +# w) But see Note [Typing rule for seq] in TcExpr, which explains why we give seq itself an ordinary type seq :: forall a b. a -> b -> b and treat it as a language construct from a typing point of view. b) Its fixity is set in LoadIface.ghcPrimIface c) It has quite a bit of desugaring magic. See DsUtils.hs Note [Desugaring seq (1)] and (2) and (3) d) There is some special rule handing: Note [User-defined RULES for seq] Note [User-defined RULES for seq] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Roman found situations where he had case (f n) of _ -> e where he knew that f (which was strict in n) would terminate if n did. Notice that the result of (f n) is discarded. So it makes sense to transform to case n of _ -> e Rather than attempt some general analysis to support this, I've added enough support that you can do this using a rewrite rule: RULE "f/seq" forall n. seq (f n) = seq n You write that rule. When GHC sees a case expression that discards its result, it mentally transforms it to a call to 'seq' and looks for a RULE. (This is done in Simplify.trySeqRules.) As usual, the correctness of the rule is up to you. VERY IMPORTANT: to make this work, we give the RULE an arity of 1, not 2. If we wrote RULE "f/seq" forall n e. seq (f n) e = seq n e with rule arity 2, then two bad things would happen: - The magical desugaring done in Note [seqId magic] item (c) for saturated application of 'seq' would turn the LHS into a case expression! - The code in Simplify.rebuildCase would need to actually supply the value argument, which turns out to be awkward. Note [lazyId magic] ~~~~~~~~~~~~~~~~~~~ lazy :: forall a?. a? -> a? (i.e. works for unboxed types too) 'lazy' is used to make sure that a sub-expression, and its free variables, are truly used call-by-need, with no code motion. Key examples: * pseq: pseq a b = a `seq` lazy b We want to make sure that the free vars of 'b' are not evaluated before 'a', even though the expression is plainly strict in 'b'. * catch: catch a b = catch# (lazy a) b Again, it's clear that 'a' will be evaluated strictly (and indeed applied to a state token) but we want to make sure that any exceptions arising from the evaluation of 'a' are caught by the catch (see Trac #11555). Implementing 'lazy' is a bit tricky: * It must not have a strictness signature: by being a built-in Id, all the info about lazyId comes from here, not from GHC.Base.hi. This is important, because the strictness analyser will spot it as strict! * It must not have an unfolding: it gets "inlined" by a HACK in CorePrep. It's very important to do this inlining *after* unfoldings are exposed in the interface file. Otherwise, the unfolding for (say) pseq in the interface file will not mention 'lazy', so if we inline 'pseq' we'll totally miss the very thing that 'lazy' was there for in the first place. See Trac #3259 for a real world example. * Suppose CorePrep sees (catch# (lazy e) b). At all costs we must avoid using call by value here: case e of r -> catch# r b Avoiding that is the whole point of 'lazy'. So in CorePrep (which generate the 'case' expression for a call-by-value call) we must spot the 'lazy' on the arg (in CorePrep.cpeApp), and build a 'let' instead. * lazyId is defined in GHC.Base, so we don't *have* to inline it. If it appears un-applied, we'll end up just calling it. Note [noinlineId magic] ~~~~~~~~~~~~~~~~~~~~~~~ noinline :: forall a. a -> a 'noinline' is used to make sure that a function f is never inlined, e.g., as in 'noinline f x'. Ordinarily, the identity function with NOINLINE could be used to achieve this effect; however, this has the unfortunate result of leaving a (useless) call to noinline at runtime. So we have a little bit of magic to optimize away 'noinline' after we are done running the simplifier. 'noinline' needs to be wired-in because it gets inserted automatically when we serialize an expression to the interface format, and we DON'T want use its fingerprints. Note [The oneShot function] ~~~~~~~~~~~~~~~~~~~~~~~~~~~ In the context of making left-folds fuse somewhat okish (see ticket #7994 and Note [Left folds via right fold]) it was determined that it would be useful if library authors could explicitly tell the compiler that a certain lambda is called at most once. The oneShot function allows that. 'oneShot' is levity-polymorphic, i.e. the type variables can refer to unlifted types as well (Trac #10744); e.g. oneShot (\x:Int# -> x +# 1#) Like most magic functions it has a compulsory unfolding, so there is no need for a real definition somewhere. We have one in GHC.Magic for the convenience of putting the documentation there. It uses `setOneShotLambda` on the lambda's binder. That is the whole magic: A typical call looks like oneShot (\y. e) after unfolding the definition `oneShot = \f \x[oneshot]. f x` we get (\f \x[oneshot]. f x) (\y. e) --> \x[oneshot]. ((\y.e) x) --> \x[oneshot] e[x/y] which is what we want. It is only effective if the one-shot info survives as long as possible; in particular it must make it into the interface in unfoldings. See Note [Preserve OneShotInfo] in CoreTidy. Also see https://ghc.haskell.org/trac/ghc/wiki/OneShot. Note [magicDictId magic] ~~~~~~~~~~~~~~~~~~~~~~~~~ The identifier `magicDict` is just a place-holder, which is used to implement a primitive that we cannot define in Haskell but we can write in Core. It is declared with a place-holder type: magicDict :: forall a. a The intention is that the identifier will be used in a very specific way, to create dictionaries for classes with a single method. Consider a class like this: class C a where f :: T a We are going to use `magicDict`, in conjunction with a built-in Prelude rule, to cast values of type `T a` into dictionaries for `C a`. To do this, we define a function like this in the library: data WrapC a b = WrapC (C a => Proxy a -> b) withT :: (C a => Proxy a -> b) -> T a -> Proxy a -> b withT f x y = magicDict (WrapC f) x y The purpose of `WrapC` is to avoid having `f` instantiated. Also, it avoids impredicativity, because `magicDict`'s type cannot be instantiated with a forall. The field of `WrapC` contains a `Proxy` parameter which is used to link the type of the constraint, `C a`, with the type of the `Wrap` value being made. Next, we add a built-in Prelude rule (see prelude/PrelRules.hs), which will replace the RHS of this definition with the appropriate definition in Core. The rewrite rule works as follows: magicDict @t (wrap @a @b f) x y ----> f (x `cast` co a) y The `co` coercion is the newtype-coercion extracted from the type-class. The type class is obtain by looking at the type of wrap. ------------------------------------------------------------- @realWorld#@ used to be a magic literal, \tr{void#}. If things get nasty as-is, change it back to a literal (@Literal@). voidArgId is a Local Id used simply as an argument in functions where we just want an arg to avoid having a thunk of unlifted type. E.g. x = \ void :: Void# -> (# p, q #) This comes up in strictness analysis Note [evaldUnfoldings] ~~~~~~~~~~~~~~~~~~~~~~ The evaldUnfolding makes it look that some primitive value is evaluated, which in turn makes Simplify.interestingArg return True, which in turn makes INLINE things applied to said value likely to be inlined. -} realWorldPrimId :: Id -- :: State# RealWorld realWorldPrimId = pcMiscPrelId realWorldName realWorldStatePrimTy (noCafIdInfo `setUnfoldingInfo` evaldUnfolding -- Note [evaldUnfoldings] `setOneShotInfo` stateHackOneShot `setNeverLevPoly` realWorldStatePrimTy) voidPrimId :: Id -- Global constant :: Void# voidPrimId = pcMiscPrelId voidPrimIdName voidPrimTy (noCafIdInfo `setUnfoldingInfo` evaldUnfolding -- Note [evaldUnfoldings] `setNeverLevPoly` voidPrimTy) voidArgId :: Id -- Local lambda-bound :: Void# voidArgId = mkSysLocal (fsLit "void") voidArgIdKey voidPrimTy coercionTokenId :: Id -- :: () ~ () coercionTokenId -- Used to replace Coercion terms when we go to STG = pcMiscPrelId coercionTokenName (mkTyConApp eqPrimTyCon [liftedTypeKind, liftedTypeKind, unitTy, unitTy]) noCafIdInfo pcMiscPrelId :: Name -> Type -> IdInfo -> Id pcMiscPrelId name ty info = mkVanillaGlobalWithInfo name ty info -- We lie and say the thing is imported; otherwise, we get into -- a mess with dependency analysis; e.g., core2stg may heave in -- random calls to GHCbase.unpackPS__. If GHCbase is the module -- being compiled, then it's just a matter of luck if the definition -- will be in "the right place" to be in scope.