2 % (c) The University of Glasgow 2006
3 % (c) The AQUA Project, Glasgow University, 1998
6 This module contains definitions for the IdInfo for things that
7 have a standard form, namely:
11 - method and superclass selectors
12 - primitive operations
15 {-# OPTIONS -fno-warn-missing-signatures #-}
16 -- The above warning supression flag is a temporary kludge.
17 -- While working on this module you are encouraged to remove it and fix
18 -- any warnings in the module. See
19 -- <http://hackage.haskell.org/trac/ghc/wiki/Commentary/CodingStyle#Warnings>
23 mkDictFunId, mkDefaultMethodId,
27 mkPrimOpId, mkFCallId, mkTickBoxOpId, mkBreakPointOpId,
29 mkReboxingAlt, wrapNewTypeBody, unwrapNewTypeBody,
30 wrapFamInstBody, unwrapFamInstScrut,
31 mkUnpackCase, mkProductBox,
33 -- And some particular Ids; see below for why they are wired in
34 wiredInIds, ghcPrimIds,
35 unsafeCoerceId, realWorldPrimId, voidArgId, nullAddrId, seqId,
38 mkRuntimeErrorApp, mkImpossibleExpr,
39 rEC_CON_ERROR_ID, iRREFUT_PAT_ERROR_ID, rUNTIME_ERROR_ID,
40 nON_EXHAUSTIVE_GUARDS_ERROR_ID, nO_METHOD_BINDING_ERROR_ID,
41 pAT_ERROR_ID, eRROR_ID, rEC_SEL_ERROR_ID,
46 #include "HsVersions.h"
55 import CoreUtils ( exprType, mkCoerce )
66 import Var ( Var, TyVar, mkCoVar, mkExportedLocalVar )
72 import BasicTypes hiding ( SuccessFlag(..) )
80 %************************************************************************
82 \subsection{Wired in Ids}
84 %************************************************************************
88 There are several reasons why an Id might appear in the wiredInIds:
90 (1) The ghcPrimIds are wired in because they can't be defined in
91 Haskell at all, although the can be defined in Core. They have
92 compulsory unfoldings, so they are always inlined and they have
93 no definition site. Their home module is GHC.Prim, so they
94 also have a description in primops.txt.pp, where they are called
97 (2) The 'error' function, eRROR_ID, is wired in because we don't yet have
98 a way to express in an interface file that the result type variable
99 is 'open'; that is can be unified with an unboxed type
101 [The interface file format now carry such information, but there's
102 no way yet of expressing at the definition site for these
103 error-reporting functions that they have an 'open'
104 result type. -- sof 1/99]
106 (3) Other error functions (rUNTIME_ERROR_ID) are wired in (a) because
107 the desugarer generates code that mentiones them directly, and
108 (b) for the same reason as eRROR_ID
110 (4) lazyId is wired in because the wired-in version overrides the
111 strictness of the version defined in GHC.Base
113 In cases (2-4), the function has a definition in a library module, and
114 can be called; but the wired-in version means that the details are
115 never read from that module's interface file; instead, the full definition
123 eRROR_ID, -- This one isn't used anywhere else in the compiler
124 -- But we still need it in wiredInIds so that when GHC
125 -- compiles a program that mentions 'error' we don't
126 -- import its type from the interface file; we just get
127 -- the Id defined here. Which has an 'open-tyvar' type.
130 iRREFUT_PAT_ERROR_ID,
131 nON_EXHAUSTIVE_GUARDS_ERROR_ID,
132 nO_METHOD_BINDING_ERROR_ID,
140 -- These Ids are exported from GHC.Prim
143 = [ -- These can't be defined in Haskell, but they have
144 -- perfectly reasonable unfoldings in Core
152 %************************************************************************
154 \subsection{Data constructors}
156 %************************************************************************
158 The wrapper for a constructor is an ordinary top-level binding that evaluates
159 any strict args, unboxes any args that are going to be flattened, and calls
162 We're going to build a constructor that looks like:
164 data (Data a, C b) => T a b = T1 !a !Int b
167 \d1::Data a, d2::C b ->
168 \p q r -> case p of { p ->
170 Con T1 [a,b] [p,q,r]}}
174 * d2 is thrown away --- a context in a data decl is used to make sure
175 one *could* construct dictionaries at the site the constructor
176 is used, but the dictionary isn't actually used.
178 * We have to check that we can construct Data dictionaries for
179 the types a and Int. Once we've done that we can throw d1 away too.
181 * We use (case p of q -> ...) to evaluate p, rather than "seq" because
182 all that matters is that the arguments are evaluated. "seq" is
183 very careful to preserve evaluation order, which we don't need
186 You might think that we could simply give constructors some strictness
187 info, like PrimOps, and let CoreToStg do the let-to-case transformation.
188 But we don't do that because in the case of primops and functions strictness
189 is a *property* not a *requirement*. In the case of constructors we need to
190 do something active to evaluate the argument.
192 Making an explicit case expression allows the simplifier to eliminate
193 it in the (common) case where the constructor arg is already evaluated.
195 Note [Wrappers for data instance tycons]
196 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
197 In the case of data instances, the wrapper also applies the coercion turning
198 the representation type into the family instance type to cast the result of
199 the wrapper. For example, consider the declarations
201 data family Map k :: * -> *
202 data instance Map (a, b) v = MapPair (Map a (Pair b v))
204 The tycon to which the datacon MapPair belongs gets a unique internal
205 name of the form :R123Map, and we call it the representation tycon.
206 In contrast, Map is the family tycon (accessible via
207 tyConFamInst_maybe). A coercion allows you to move between
208 representation and family type. It is accessible from :R123Map via
209 tyConFamilyCoercion_maybe and has kind
211 Co123Map a b v :: {Map (a, b) v ~ :R123Map a b v}
213 The wrapper and worker of MapPair get the types
216 $WMapPair :: forall a b v. Map a (Map a b v) -> Map (a, b) v
217 $WMapPair a b v = MapPair a b v `cast` sym (Co123Map a b v)
220 MapPair :: forall a b v. Map a (Map a b v) -> :R123Map a b v
222 This coercion is conditionally applied by wrapFamInstBody.
224 It's a bit more complicated if the data instance is a GADT as well!
226 data instance T [a] where
227 T1 :: forall b. b -> T [Maybe b]
229 Co7T a :: T [a] ~ :R7T a
234 $WT1 :: forall b. b -> T [Maybe b]
235 $WT1 b v = T1 (Maybe b) b (Maybe b) v
236 `cast` sym (Co7T (Maybe b))
239 T1 :: forall c b. (c ~ Maybe b) => b -> :R7T c
242 mkDataConIds :: Name -> Name -> DataCon -> DataConIds
243 mkDataConIds wrap_name wkr_name data_con
244 | isNewTyCon tycon -- Newtype, only has a worker
245 = DCIds Nothing nt_work_id
247 | any isBanged all_strict_marks -- Algebraic, needs wrapper
248 || not (null eq_spec) -- NB: LoadIface.ifaceDeclSubBndrs
249 || isFamInstTyCon tycon -- depends on this test
250 = DCIds (Just alg_wrap_id) wrk_id
252 | otherwise -- Algebraic, no wrapper
253 = DCIds Nothing wrk_id
255 (univ_tvs, ex_tvs, eq_spec,
256 eq_theta, dict_theta, orig_arg_tys, res_ty) = dataConFullSig data_con
257 tycon = dataConTyCon data_con -- The representation TyCon (not family)
259 ----------- Worker (algebraic data types only) --------------
260 -- The *worker* for the data constructor is the function that
261 -- takes the representation arguments and builds the constructor.
262 wrk_id = mkGlobalId (DataConWorkId data_con) wkr_name
263 (dataConRepType data_con) wkr_info
265 wkr_arity = dataConRepArity data_con
266 wkr_info = noCafIdInfo
267 `setArityInfo` wkr_arity
268 `setStrictnessInfo` Just wkr_sig
269 `setUnfoldingInfo` evaldUnfolding -- Record that it's evaluated,
272 wkr_sig = mkStrictSig (mkTopDmdType (replicate wkr_arity topDmd) cpr_info)
273 -- Note [Data-con worker strictness]
274 -- Notice that we do *not* say the worker is strict
275 -- even if the data constructor is declared strict
276 -- e.g. data T = MkT !(Int,Int)
277 -- Why? Because the *wrapper* is strict (and its unfolding has case
278 -- expresssions that do the evals) but the *worker* itself is not.
279 -- If we pretend it is strict then when we see
280 -- case x of y -> $wMkT y
281 -- the simplifier thinks that y is "sure to be evaluated" (because
282 -- $wMkT is strict) and drops the case. No, $wMkT is not strict.
284 -- When the simplifer sees a pattern
285 -- case e of MkT x -> ...
286 -- it uses the dataConRepStrictness of MkT to mark x as evaluated;
287 -- but that's fine... dataConRepStrictness comes from the data con
288 -- not from the worker Id.
290 cpr_info | isProductTyCon tycon &&
293 wkr_arity <= mAX_CPR_SIZE = retCPR
295 -- RetCPR is only true for products that are real data types;
296 -- that is, not unboxed tuples or [non-recursive] newtypes
298 ----------- Workers for newtypes --------------
299 nt_work_id = mkGlobalId (DataConWrapId data_con) wkr_name wrap_ty nt_work_info
300 nt_work_info = noCafIdInfo -- The NoCaf-ness is set by noCafIdInfo
301 `setArityInfo` 1 -- Arity 1
302 `setUnfoldingInfo` newtype_unf
303 id_arg1 = mkTemplateLocal 1 (head orig_arg_tys)
304 newtype_unf = ASSERT2( isVanillaDataCon data_con &&
305 isSingleton orig_arg_tys, ppr data_con )
306 -- Note [Newtype datacons]
307 mkCompulsoryUnfolding $
308 mkLams wrap_tvs $ Lam id_arg1 $
309 wrapNewTypeBody tycon res_ty_args (Var id_arg1)
312 ----------- Wrapper --------------
313 -- We used to include the stupid theta in the wrapper's args
314 -- but now we don't. Instead the type checker just injects these
315 -- extra constraints where necessary.
316 wrap_tvs = (univ_tvs `minusList` map fst eq_spec) ++ ex_tvs
317 res_ty_args = substTyVars (mkTopTvSubst eq_spec) univ_tvs
318 eq_tys = mkPredTys eq_theta
319 dict_tys = mkPredTys dict_theta
320 wrap_ty = mkForAllTys wrap_tvs $ mkFunTys eq_tys $ mkFunTys dict_tys $
321 mkFunTys orig_arg_tys $ res_ty
322 -- NB: watch out here if you allow user-written equality
323 -- constraints in data constructor signatures
325 ----------- Wrappers for algebraic data types --------------
326 alg_wrap_id = mkGlobalId (DataConWrapId data_con) wrap_name wrap_ty alg_wrap_info
327 alg_wrap_info = noCafIdInfo -- The NoCaf-ness is set by noCafIdInfo
328 `setArityInfo` wrap_arity
329 -- It's important to specify the arity, so that partial
330 -- applications are treated as values
331 `setUnfoldingInfo` wrap_unf
332 `setStrictnessInfo` Just wrap_sig
334 all_strict_marks = dataConExStricts data_con ++ dataConStrictMarks data_con
335 wrap_sig = mkStrictSig (mkTopDmdType arg_dmds cpr_info)
336 arg_dmds = map mk_dmd all_strict_marks
337 mk_dmd str | isBanged str = evalDmd
338 | otherwise = lazyDmd
339 -- The Cpr info can be important inside INLINE rhss, where the
340 -- wrapper constructor isn't inlined.
341 -- And the argument strictness can be important too; we
342 -- may not inline a contructor when it is partially applied.
344 -- data W = C !Int !Int !Int
345 -- ...(let w = C x in ...(w p q)...)...
346 -- we want to see that w is strict in its two arguments
348 wrap_unf = mkInlineRule wrap_rhs (Just (length dict_args + length id_args))
349 wrap_rhs = mkLams wrap_tvs $
351 mkLams dict_args $ mkLams id_args $
352 foldr mk_case con_app
353 (zip (dict_args ++ id_args) all_strict_marks)
356 con_app _ rep_ids = wrapFamInstBody tycon res_ty_args $
357 Var wrk_id `mkTyApps` res_ty_args
359 -- Equality evidence:
360 `mkTyApps` map snd eq_spec
362 `mkVarApps` reverse rep_ids
364 (dict_args,i2) = mkLocals 1 dict_tys
365 (id_args,i3) = mkLocals i2 orig_arg_tys
367 (eq_args,_) = mkCoVarLocals i3 eq_tys
369 mkCoVarLocals i [] = ([],i)
370 mkCoVarLocals i (x:xs) = let (ys,j) = mkCoVarLocals (i+1) xs
371 y = mkCoVar (mkSysTvName (mkBuiltinUnique i) (fsLit "dc_co")) x
375 :: (Id, HsBang) -- Arg, strictness
376 -> (Int -> [Id] -> CoreExpr) -- Body
377 -> Int -- Next rep arg id
378 -> [Id] -- Rep args so far, reversed
380 mk_case (arg,strict) body i rep_args
382 HsNoBang -> body i (arg:rep_args)
383 HsUnpack -> unboxProduct i (Var arg) (idType arg) the_body
385 the_body i con_args = body i (reverse con_args ++ rep_args)
386 _other -- HsUnpackFailed and HsStrict
387 | isUnLiftedType (idType arg) -> body i (arg:rep_args)
388 | otherwise -> Case (Var arg) arg res_ty
389 [(DEFAULT,[], body i (arg:rep_args))]
391 mAX_CPR_SIZE :: Arity
393 -- We do not treat very big tuples as CPR-ish:
394 -- a) for a start we get into trouble because there aren't
395 -- "enough" unboxed tuple types (a tiresome restriction,
397 -- b) more importantly, big unboxed tuples get returned mainly
398 -- on the stack, and are often then allocated in the heap
399 -- by the caller. So doing CPR for them may in fact make
402 mkLocals i tys = (zipWith mkTemplateLocal [i..i+n-1] tys, i+n)
407 Note [Newtype datacons]
408 ~~~~~~~~~~~~~~~~~~~~~~~
409 The "data constructor" for a newtype should always be vanilla. At one
410 point this wasn't true, because the newtype arising from
413 newtype T:D a = D:D (C a)
414 so the data constructor for T:C had a single argument, namely the
415 predicate (C a). But now we treat that as an ordinary argument, not
416 part of the theta-type, so all is well.
419 %************************************************************************
421 \subsection{Dictionary selectors}
423 %************************************************************************
425 Selecting a field for a dictionary. If there is just one field, then
426 there's nothing to do.
428 Dictionary selectors may get nested forall-types. Thus:
431 op :: forall b. Ord b => a -> b -> b
433 Then the top-level type for op is
435 op :: forall a. Foo a =>
439 This is unlike ordinary record selectors, which have all the for-alls
440 at the outside. When dealing with classes it's very convenient to
441 recover the original type signature from the class op selector.
444 mkDictSelId :: Bool -- True <=> don't include the unfolding
445 -- Little point on imports without -O, because the
446 -- dictionary itself won't be visible
447 -> Name -> Class -> Id
448 mkDictSelId no_unf name clas
449 = mkGlobalId (ClassOpId clas) name sel_ty info
451 sel_ty = mkForAllTys tyvars (mkFunTy (idType dict_id) (idType the_arg_id))
452 -- We can't just say (exprType rhs), because that would give a type
454 -- for a single-op class (after all, the selector is the identity)
455 -- But it's type must expose the representation of the dictionary
456 -- to get (say) C a -> (a -> a)
458 base_info = noCafIdInfo
460 `setStrictnessInfo` Just strict_sig
461 `setUnfoldingInfo` (if no_unf then noUnfolding
462 else mkImplicitUnfolding rhs)
463 -- In module where class op is defined, we must add
464 -- the unfolding, even though it'll never be inlined
465 -- becuase we use that to generate a top-level binding
468 info = base_info `setSpecInfo` mkSpecInfo [rule]
469 `setInlinePragInfo` neverInlinePragma
470 -- Add a magic BuiltinRule, and never inline it
471 -- so that the rule is always available to fire.
472 -- See Note [ClassOp/DFun selection] in TcInstDcls
474 n_ty_args = length tyvars
476 -- This is the built-in rule that goes
477 -- op (dfT d1 d2) ---> opT d1 d2
478 rule = BuiltinRule { ru_name = fsLit "Class op " `appendFS`
479 occNameFS (getOccName name)
481 , ru_nargs = n_ty_args + 1
482 , ru_try = dictSelRule index n_ty_args }
484 -- The strictness signature is of the form U(AAAVAAAA) -> T
485 -- where the V depends on which item we are selecting
486 -- It's worth giving one, so that absence info etc is generated
487 -- even if the selector isn't inlined
488 strict_sig = mkStrictSig (mkTopDmdType [arg_dmd] TopRes)
489 arg_dmd | new_tycon = evalDmd
490 | otherwise = Eval (Prod [ if the_arg_id == id then evalDmd else Abs
493 tycon = classTyCon clas
494 new_tycon = isNewTyCon tycon
495 [data_con] = tyConDataCons tycon
496 tyvars = dataConUnivTyVars data_con
497 arg_tys = {- ASSERT( isVanillaDataCon data_con ) -} dataConRepArgTys data_con
498 eq_theta = dataConEqTheta data_con
499 index = assoc "MkId.mkDictSelId" (map idName (classSelIds clas) `zip` [0..]) name
500 the_arg_id = arg_ids !! index
502 pred = mkClassPred clas (mkTyVarTys tyvars)
503 dict_id = mkTemplateLocal 1 $ mkPredTy pred
504 (eq_ids,n) = mkCoVarLocals 2 $ mkPredTys eq_theta
505 arg_ids = mkTemplateLocalsNum n arg_tys
507 mkCoVarLocals i [] = ([],i)
508 mkCoVarLocals i (x:xs) = let (ys,j) = mkCoVarLocals (i+1) xs
509 y = mkCoVar (mkSysTvName (mkBuiltinUnique i) (fsLit "dc_co")) x
512 rhs = mkLams tyvars (Lam dict_id rhs_body)
513 rhs_body | new_tycon = unwrapNewTypeBody tycon (map mkTyVarTy tyvars) (Var dict_id)
514 | otherwise = Case (Var dict_id) dict_id (idType the_arg_id)
515 [(DataAlt data_con, eq_ids ++ arg_ids, Var the_arg_id)]
517 dictSelRule :: Int -> Arity -> IdUnfoldingFun -> [CoreExpr] -> Maybe CoreExpr
519 -- op_i t1..tk (df s1..sn d1..dm) = op_i_helper s1..sn d1..dm
520 -- op_i t1..tk (D t1..tk op1 ... opm) = opi
522 -- NB: the data constructor has the same number of type args as the class op
524 dictSelRule index n_ty_args id_unf args
525 | (dict_arg : _) <- drop n_ty_args args
526 , Just (_, _, val_args) <- exprIsConApp_maybe id_unf dict_arg
527 = Just (val_args !! index)
533 %************************************************************************
537 %************************************************************************
540 -- unbox a product type...
541 -- we will recurse into newtypes, casting along the way, and unbox at the
542 -- first product data constructor we find. e.g.
544 -- data PairInt = PairInt Int Int
545 -- newtype S = MkS PairInt
548 -- If we have e = MkT (MkS (PairInt 0 1)) and some body expecting a list of
549 -- ids, we get (modulo int passing)
551 -- case (e `cast` CoT) `cast` CoS of
552 -- PairInt a b -> body [a,b]
554 -- The Ints passed around are just for creating fresh locals
555 unboxProduct :: Int -> CoreExpr -> Type -> (Int -> [Id] -> CoreExpr) -> CoreExpr
556 unboxProduct i arg arg_ty body
559 result = mkUnpackCase the_id arg con_args boxing_con rhs
560 (_tycon, _tycon_args, boxing_con, tys) = deepSplitProductType "unboxProduct" arg_ty
561 ([the_id], i') = mkLocals i [arg_ty]
562 (con_args, i'') = mkLocals i' tys
563 rhs = body i'' con_args
565 mkUnpackCase :: Id -> CoreExpr -> [Id] -> DataCon -> CoreExpr -> CoreExpr
566 -- (mkUnpackCase x e args Con body)
568 -- case (e `cast` ...) of bndr { Con args -> body }
570 -- the type of the bndr passed in is irrelevent
571 mkUnpackCase bndr arg unpk_args boxing_con body
572 = Case cast_arg (setIdType bndr bndr_ty) (exprType body) [(DataAlt boxing_con, unpk_args, body)]
574 (cast_arg, bndr_ty) = go (idType bndr) arg
576 | (tycon, tycon_args, _, _) <- splitProductType "mkUnpackCase" ty
577 , isNewTyCon tycon && not (isRecursiveTyCon tycon)
578 = go (newTyConInstRhs tycon tycon_args)
579 (unwrapNewTypeBody tycon tycon_args arg)
580 | otherwise = (arg, ty)
583 reboxProduct :: [Unique] -- uniques to create new local binders
584 -> Type -- type of product to box
585 -> ([Unique], -- remaining uniques
586 CoreExpr, -- boxed product
587 [Id]) -- Ids being boxed into product
590 (_tycon, _tycon_args, _pack_con, con_arg_tys) = deepSplitProductType "reboxProduct" ty
592 us' = dropList con_arg_tys us
594 arg_ids = zipWith (mkSysLocal (fsLit "rb")) us con_arg_tys
596 bind_rhs = mkProductBox arg_ids ty
599 (us', bind_rhs, arg_ids)
601 mkProductBox :: [Id] -> Type -> CoreExpr
602 mkProductBox arg_ids ty
605 (tycon, tycon_args, pack_con, _con_arg_tys) = splitProductType "mkProductBox" ty
608 | isNewTyCon tycon && not (isRecursiveTyCon tycon)
609 = wrap (mkProductBox arg_ids (newTyConInstRhs tycon tycon_args))
610 | otherwise = mkConApp pack_con (map Type tycon_args ++ map Var arg_ids)
612 wrap expr = wrapNewTypeBody tycon tycon_args expr
615 -- (mkReboxingAlt us con xs rhs) basically constructs the case
616 -- alternative (con, xs, rhs)
617 -- but it does the reboxing necessary to construct the *source*
618 -- arguments, xs, from the representation arguments ys.
620 -- data T = MkT !(Int,Int) Bool
622 -- mkReboxingAlt MkT [x,b] r
623 -- = (DataAlt MkT, [y::Int,z::Int,b], let x = (y,z) in r)
625 -- mkDataAlt should really be in DataCon, but it can't because
626 -- it manipulates CoreSyn.
629 :: [Unique] -- Uniques for the new Ids
631 -> [Var] -- Source-level args, including existential dicts
635 mkReboxingAlt us con args rhs
636 | not (any isMarkedUnboxed stricts)
637 = (DataAlt con, args, rhs)
641 (binds, args') = go args stricts us
643 (DataAlt con, args', mkLets binds rhs)
646 stricts = dataConExStricts con ++ dataConStrictMarks con
648 go [] _stricts _us = ([], [])
650 -- Type variable case
651 go (arg:args) stricts us
653 = let (binds, args') = go args stricts us
654 in (binds, arg:args')
656 -- Term variable case
657 go (arg:args) (str:stricts) us
658 | isMarkedUnboxed str
660 let (binds, unpacked_args') = go args stricts us'
661 (us', bind_rhs, unpacked_args) = reboxProduct us (idType arg)
663 (NonRec arg bind_rhs : binds, unpacked_args ++ unpacked_args')
665 = let (binds, args') = go args stricts us
666 in (binds, arg:args')
667 go (_ : _) [] _ = panic "mkReboxingAlt"
671 %************************************************************************
673 Wrapping and unwrapping newtypes and type families
675 %************************************************************************
678 wrapNewTypeBody :: TyCon -> [Type] -> CoreExpr -> CoreExpr
679 -- The wrapper for the data constructor for a newtype looks like this:
680 -- newtype T a = MkT (a,Int)
681 -- MkT :: forall a. (a,Int) -> T a
682 -- MkT = /\a. \(x:(a,Int)). x `cast` sym (CoT a)
683 -- where CoT is the coercion TyCon assoicated with the newtype
685 -- The call (wrapNewTypeBody T [a] e) returns the
686 -- body of the wrapper, namely
687 -- e `cast` (CoT [a])
689 -- If a coercion constructor is provided in the newtype, then we use
690 -- it, otherwise the wrap/unwrap are both no-ops
692 -- If the we are dealing with a newtype *instance*, we have a second coercion
693 -- identifying the family instance with the constructor of the newtype
694 -- instance. This coercion is applied in any case (ie, composed with the
695 -- coercion constructor of the newtype or applied by itself).
697 wrapNewTypeBody tycon args result_expr
698 = wrapFamInstBody tycon args inner
701 | Just co_con <- newTyConCo_maybe tycon
702 = mkCoerce (mkSymCoercion (mkTyConApp co_con args)) result_expr
706 -- When unwrapping, we do *not* apply any family coercion, because this will
707 -- be done via a CoPat by the type checker. We have to do it this way as
708 -- computing the right type arguments for the coercion requires more than just
709 -- a spliting operation (cf, TcPat.tcConPat).
711 unwrapNewTypeBody :: TyCon -> [Type] -> CoreExpr -> CoreExpr
712 unwrapNewTypeBody tycon args result_expr
713 | Just co_con <- newTyConCo_maybe tycon
714 = mkCoerce (mkTyConApp co_con args) result_expr
718 -- If the type constructor is a representation type of a data instance, wrap
719 -- the expression into a cast adjusting the expression type, which is an
720 -- instance of the representation type, to the corresponding instance of the
721 -- family instance type.
722 -- See Note [Wrappers for data instance tycons]
723 wrapFamInstBody :: TyCon -> [Type] -> CoreExpr -> CoreExpr
724 wrapFamInstBody tycon args body
725 | Just co_con <- tyConFamilyCoercion_maybe tycon
726 = mkCoerce (mkSymCoercion (mkTyConApp co_con args)) body
730 unwrapFamInstScrut :: TyCon -> [Type] -> CoreExpr -> CoreExpr
731 unwrapFamInstScrut tycon args scrut
732 | Just co_con <- tyConFamilyCoercion_maybe tycon
733 = mkCoerce (mkTyConApp co_con args) scrut
739 %************************************************************************
741 \subsection{Primitive operations}
743 %************************************************************************
746 mkPrimOpId :: PrimOp -> Id
750 (tyvars,arg_tys,res_ty, arity, strict_sig) = primOpSig prim_op
751 ty = mkForAllTys tyvars (mkFunTys arg_tys res_ty)
752 name = mkWiredInName gHC_PRIM (primOpOcc prim_op)
753 (mkPrimOpIdUnique (primOpTag prim_op))
755 id = mkGlobalId (PrimOpId prim_op) name ty info
758 `setSpecInfo` mkSpecInfo (primOpRules prim_op name)
760 `setStrictnessInfo` Just strict_sig
762 -- For each ccall we manufacture a separate CCallOpId, giving it
763 -- a fresh unique, a type that is correct for this particular ccall,
764 -- and a CCall structure that gives the correct details about calling
767 -- The *name* of this Id is a local name whose OccName gives the full
768 -- details of the ccall, type and all. This means that the interface
769 -- file reader can reconstruct a suitable Id
771 mkFCallId :: Unique -> ForeignCall -> Type -> Id
772 mkFCallId uniq fcall ty
773 = ASSERT( isEmptyVarSet (tyVarsOfType ty) )
774 -- A CCallOpId should have no free type variables;
775 -- when doing substitutions won't substitute over it
776 mkGlobalId (FCallId fcall) name ty info
778 occ_str = showSDoc (braces (ppr fcall <+> ppr ty))
779 -- The "occurrence name" of a ccall is the full info about the
780 -- ccall; it is encoded, but may have embedded spaces etc!
782 name = mkFCallName uniq occ_str
786 `setStrictnessInfo` Just strict_sig
788 (_, tau) = tcSplitForAllTys ty
789 (arg_tys, _) = tcSplitFunTys tau
790 arity = length arg_tys
791 strict_sig = mkStrictSig (mkTopDmdType (replicate arity evalDmd) TopRes)
793 -- Tick boxes and breakpoints are both represented as TickBoxOpIds,
794 -- except for the type:
796 -- a plain HPC tick box has type (State# RealWorld)
797 -- a breakpoint Id has type forall a.a
799 -- The breakpoint Id will be applied to a list of arbitrary free variables,
800 -- which is why it needs a polymorphic type.
802 mkTickBoxOpId :: Unique -> Module -> TickBoxId -> Id
803 mkTickBoxOpId uniq mod ix = mkTickBox' uniq mod ix realWorldStatePrimTy
805 mkBreakPointOpId :: Unique -> Module -> TickBoxId -> Id
806 mkBreakPointOpId uniq mod ix = mkTickBox' uniq mod ix ty
807 where ty = mkSigmaTy [openAlphaTyVar] [] openAlphaTy
809 mkTickBox' uniq mod ix ty = mkGlobalId (TickBoxOpId tickbox) name ty info
811 tickbox = TickBox mod ix
812 occ_str = showSDoc (braces (ppr tickbox))
813 name = mkTickBoxOpName uniq occ_str
818 %************************************************************************
820 \subsection{DictFuns and default methods}
822 %************************************************************************
824 Important notes about dict funs and default methods
825 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
826 Dict funs and default methods are *not* ImplicitIds. Their definition
827 involves user-written code, so we can't figure out their strictness etc
828 based on fixed info, as we can for constructors and record selectors (say).
830 We build them as LocalIds, but with External Names. This ensures that
831 they are taken to account by free-variable finding and dependency
832 analysis (e.g. CoreFVs.exprFreeVars).
834 Why shouldn't they be bound as GlobalIds? Because, in particular, if
835 they are globals, the specialiser floats dict uses above their defns,
836 which prevents good simplifications happening. Also the strictness
837 analyser treats a occurrence of a GlobalId as imported and assumes it
838 contains strictness in its IdInfo, which isn't true if the thing is
839 bound in the same module as the occurrence.
841 It's OK for dfuns to be LocalIds, because we form the instance-env to
842 pass on to the next module (md_insts) in CoreTidy, afer tidying
843 and globalising the top-level Ids.
845 BUT make sure they are *exported* LocalIds (mkExportedLocalId) so
846 that they aren't discarded by the occurrence analyser.
849 mkDefaultMethodId dm_name ty = mkExportedLocalId dm_name ty
851 mkDictFunId :: Name -- Name to use for the dict fun;
858 mkDictFunId dfun_name inst_tyvars dfun_theta clas inst_tys
859 = mkExportedLocalVar (DFunId is_nt) dfun_name dfun_ty vanillaIdInfo
861 is_nt = isNewTyCon (classTyCon clas)
862 dfun_ty = mkSigmaTy inst_tyvars dfun_theta (mkDictTy clas inst_tys)
866 %************************************************************************
868 \subsection{Un-definable}
870 %************************************************************************
872 These Ids can't be defined in Haskell. They could be defined in
873 unfoldings in the wired-in GHC.Prim interface file, but we'd have to
874 ensure that they were definitely, definitely inlined, because there is
875 no curried identifier for them. That's what mkCompulsoryUnfolding
876 does. If we had a way to get a compulsory unfolding from an interface
877 file, we could do that, but we don't right now.
879 unsafeCoerce# isn't so much a PrimOp as a phantom identifier, that
880 just gets expanded into a type coercion wherever it occurs. Hence we
881 add it as a built-in Id with an unfolding here.
883 The type variables we use here are "open" type variables: this means
884 they can unify with both unlifted and lifted types. Hence we provide
885 another gun with which to shoot yourself in the foot.
888 mkWiredInIdName mod fs uniq id
889 = mkWiredInName mod (mkOccNameFS varName fs) uniq (AnId id) UserSyntax
891 unsafeCoerceName = mkWiredInIdName gHC_PRIM (fsLit "unsafeCoerce#") unsafeCoerceIdKey unsafeCoerceId
892 nullAddrName = mkWiredInIdName gHC_PRIM (fsLit "nullAddr#") nullAddrIdKey nullAddrId
893 seqName = mkWiredInIdName gHC_PRIM (fsLit "seq") seqIdKey seqId
894 realWorldName = mkWiredInIdName gHC_PRIM (fsLit "realWorld#") realWorldPrimIdKey realWorldPrimId
895 lazyIdName = mkWiredInIdName gHC_BASE (fsLit "lazy") lazyIdKey lazyId
897 errorName = mkWiredInIdName gHC_ERR (fsLit "error") errorIdKey eRROR_ID
898 recSelErrorName = mkWiredInIdName cONTROL_EXCEPTION_BASE (fsLit "recSelError") recSelErrorIdKey rEC_SEL_ERROR_ID
899 runtimeErrorName = mkWiredInIdName cONTROL_EXCEPTION_BASE (fsLit "runtimeError") runtimeErrorIdKey rUNTIME_ERROR_ID
900 irrefutPatErrorName = mkWiredInIdName cONTROL_EXCEPTION_BASE (fsLit "irrefutPatError") irrefutPatErrorIdKey iRREFUT_PAT_ERROR_ID
901 recConErrorName = mkWiredInIdName cONTROL_EXCEPTION_BASE (fsLit "recConError") recConErrorIdKey rEC_CON_ERROR_ID
902 patErrorName = mkWiredInIdName cONTROL_EXCEPTION_BASE (fsLit "patError") patErrorIdKey pAT_ERROR_ID
903 noMethodBindingErrorName = mkWiredInIdName cONTROL_EXCEPTION_BASE (fsLit "noMethodBindingError")
904 noMethodBindingErrorIdKey nO_METHOD_BINDING_ERROR_ID
905 nonExhaustiveGuardsErrorName
906 = mkWiredInIdName cONTROL_EXCEPTION_BASE (fsLit "nonExhaustiveGuardsError")
907 nonExhaustiveGuardsErrorIdKey nON_EXHAUSTIVE_GUARDS_ERROR_ID
911 ------------------------------------------------
912 -- unsafeCoerce# :: forall a b. a -> b
914 = pcMiscPrelId unsafeCoerceName ty info
916 info = noCafIdInfo `setUnfoldingInfo` mkCompulsoryUnfolding rhs
919 ty = mkForAllTys [openAlphaTyVar,openBetaTyVar]
920 (mkFunTy openAlphaTy openBetaTy)
921 [x] = mkTemplateLocals [openAlphaTy]
922 rhs = mkLams [openAlphaTyVar,openBetaTyVar,x] $
923 Cast (Var x) (mkUnsafeCoercion openAlphaTy openBetaTy)
925 ------------------------------------------------
927 -- nullAddr# :: Addr#
928 -- The reason is is here is because we don't provide
929 -- a way to write this literal in Haskell.
930 nullAddrId = pcMiscPrelId nullAddrName addrPrimTy info
932 info = noCafIdInfo `setUnfoldingInfo`
933 mkCompulsoryUnfolding (Lit nullAddrLit)
935 ------------------------------------------------
936 seqId :: Id -- See Note [seqId magic]
937 seqId = pcMiscPrelId seqName ty info
939 info = noCafIdInfo `setUnfoldingInfo` mkCompulsoryUnfolding rhs
940 `setSpecInfo` mkSpecInfo [seq_cast_rule]
943 ty = mkForAllTys [alphaTyVar,openBetaTyVar]
944 (mkFunTy alphaTy (mkFunTy openBetaTy openBetaTy))
945 [x,y] = mkTemplateLocals [alphaTy, openBetaTy]
946 rhs = mkLams [alphaTyVar,openBetaTyVar,x,y] (Case (Var x) x openBetaTy [(DEFAULT, [], Var y)])
948 -- See Note [Built-in RULES for seq]
949 seq_cast_rule = BuiltinRule { ru_name = fsLit "seq of cast"
952 , ru_try = match_seq_of_cast
955 match_seq_of_cast :: IdUnfoldingFun -> [CoreExpr] -> Maybe CoreExpr
956 -- See Note [Built-in RULES for seq]
957 match_seq_of_cast _ [Type _, Type res_ty, Cast scrut co, expr]
958 = Just (Var seqId `mkApps` [Type (fst (coercionKind co)), Type res_ty,
960 match_seq_of_cast _ _ = Nothing
962 ------------------------------------------------
963 lazyId :: Id -- See Note [lazyId magic]
964 lazyId = pcMiscPrelId lazyIdName ty info
967 ty = mkForAllTys [alphaTyVar] (mkFunTy alphaTy alphaTy)
972 'GHC.Prim.seq' is special in several ways.
974 a) Its second arg can have an unboxed type
977 b) Its fixity is set in LoadIface.ghcPrimIface
979 c) It has quite a bit of desugaring magic.
980 See DsUtils.lhs Note [Desugaring seq (1)] and (2) and (3)
982 d) There is some special rule handing: Note [User-defined RULES for seq]
984 Note [User-defined RULES for seq]
985 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
986 Roman found situations where he had
988 where he knew that f (which was strict in n) would terminate if n did.
989 Notice that the result of (f n) is discarded. So it makes sense to
993 Rather than attempt some general analysis to support this, I've added
994 enough support that you can do this using a rewrite rule:
996 RULE "f/seq" forall n. seq (f n) e = seq n e
998 You write that rule. When GHC sees a case expression that discards
999 its result, it mentally transforms it to a call to 'seq' and looks for
1000 a RULE. (This is done in Simplify.rebuildCase.) As usual, the
1001 correctness of the rule is up to you.
1003 To make this work, we need to be careful that the magical desugaring
1004 done in Note [seqId magic] item (c) is *not* done on the LHS of a rule.
1005 Or rather, we arrange to un-do it, in DsBinds.decomposeRuleLhs.
1007 Note [Built-in RULES for seq]
1008 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1009 We also have the following built-in rule for seq
1011 seq (x `cast` co) y = seq x y
1013 This eliminates unnecessary casts and also allows other seq rules to
1014 match more often. Notably,
1016 seq (f x `cast` co) y --> seq (f x) y
1018 and now a user-defined rule for seq (see Note [User-defined RULES for seq])
1024 lazy :: forall a?. a? -> a? (i.e. works for unboxed types too)
1026 Used to lazify pseq: pseq a b = a `seq` lazy b
1028 Also, no strictness: by being a built-in Id, all the info about lazyId comes from here,
1029 not from GHC.Base.hi. This is important, because the strictness
1030 analyser will spot it as strict!
1032 Also no unfolding in lazyId: it gets "inlined" by a HACK in CorePrep.
1033 It's very important to do this inlining *after* unfoldings are exposed
1034 in the interface file. Otherwise, the unfolding for (say) pseq in the
1035 interface file will not mention 'lazy', so if we inline 'pseq' we'll totally
1036 miss the very thing that 'lazy' was there for in the first place.
1037 See Trac #3259 for a real world example.
1039 lazyId is defined in GHC.Base, so we don't *have* to inline it. If it
1040 appears un-applied, we'll end up just calling it.
1042 -------------------------------------------------------------
1043 @realWorld#@ used to be a magic literal, \tr{void#}. If things get
1044 nasty as-is, change it back to a literal (@Literal@).
1046 voidArgId is a Local Id used simply as an argument in functions
1047 where we just want an arg to avoid having a thunk of unlifted type.
1049 x = \ void :: State# RealWorld -> (# p, q #)
1051 This comes up in strictness analysis
1054 realWorldPrimId -- :: State# RealWorld
1055 = pcMiscPrelId realWorldName realWorldStatePrimTy
1056 (noCafIdInfo `setUnfoldingInfo` evaldUnfolding)
1057 -- The evaldUnfolding makes it look that realWorld# is evaluated
1058 -- which in turn makes Simplify.interestingArg return True,
1059 -- which in turn makes INLINE things applied to realWorld# likely
1063 voidArgId -- :: State# RealWorld
1064 = mkSysLocal (fsLit "void") voidArgIdKey realWorldStatePrimTy
1068 %************************************************************************
1070 \subsection[PrelVals-error-related]{@error@ and friends; @trace@}
1072 %************************************************************************
1074 GHC randomly injects these into the code.
1076 @patError@ is just a version of @error@ for pattern-matching
1077 failures. It knows various ``codes'' which expand to longer
1078 strings---this saves space!
1080 @absentErr@ is a thing we put in for ``absent'' arguments. They jolly
1081 well shouldn't be yanked on, but if one is, then you will get a
1082 friendly message from @absentErr@ (rather than a totally random
1085 @parError@ is a special version of @error@ which the compiler does
1086 not know to be a bottoming Id. It is used in the @_par_@ and @_seq_@
1087 templates, but we don't ever expect to generate code for it.
1091 :: Id -- Should be of type (forall a. Addr# -> a)
1092 -- where Addr# points to a UTF8 encoded string
1093 -> Type -- The type to instantiate 'a'
1094 -> String -- The string to print
1097 mkRuntimeErrorApp err_id res_ty err_msg
1098 = mkApps (Var err_id) [Type res_ty, err_string]
1100 err_string = Lit (mkMachString err_msg)
1102 mkImpossibleExpr :: Type -> CoreExpr
1103 mkImpossibleExpr res_ty
1104 = mkRuntimeErrorApp rUNTIME_ERROR_ID res_ty "Impossible case alternative"
1106 rEC_SEL_ERROR_ID = mkRuntimeErrorId recSelErrorName
1107 rUNTIME_ERROR_ID = mkRuntimeErrorId runtimeErrorName
1108 iRREFUT_PAT_ERROR_ID = mkRuntimeErrorId irrefutPatErrorName
1109 rEC_CON_ERROR_ID = mkRuntimeErrorId recConErrorName
1110 pAT_ERROR_ID = mkRuntimeErrorId patErrorName
1111 nO_METHOD_BINDING_ERROR_ID = mkRuntimeErrorId noMethodBindingErrorName
1112 nON_EXHAUSTIVE_GUARDS_ERROR_ID = mkRuntimeErrorId nonExhaustiveGuardsErrorName
1114 -- The runtime error Ids take a UTF8-encoded string as argument
1116 mkRuntimeErrorId :: Name -> Id
1117 mkRuntimeErrorId name = pc_bottoming_Id name runtimeErrorTy
1119 runtimeErrorTy :: Type
1120 runtimeErrorTy = mkSigmaTy [openAlphaTyVar] [] (mkFunTy addrPrimTy openAlphaTy)
1124 eRROR_ID = pc_bottoming_Id errorName errorTy
1127 errorTy = mkSigmaTy [openAlphaTyVar] [] (mkFunTys [mkListTy charTy] openAlphaTy)
1128 -- Notice the openAlphaTyVar. It says that "error" can be applied
1129 -- to unboxed as well as boxed types. This is OK because it never
1130 -- returns, so the return type is irrelevant.
1134 %************************************************************************
1136 \subsection{Utilities}
1138 %************************************************************************
1141 pcMiscPrelId :: Name -> Type -> IdInfo -> Id
1142 pcMiscPrelId name ty info
1143 = mkVanillaGlobalWithInfo name ty info
1144 -- We lie and say the thing is imported; otherwise, we get into
1145 -- a mess with dependency analysis; e.g., core2stg may heave in
1146 -- random calls to GHCbase.unpackPS__. If GHCbase is the module
1147 -- being compiled, then it's just a matter of luck if the definition
1148 -- will be in "the right place" to be in scope.
1150 pc_bottoming_Id :: Name -> Type -> Id
1151 -- Function of arity 1, which diverges after being given one argument
1152 pc_bottoming_Id name ty
1153 = pcMiscPrelId name ty bottoming_info
1155 bottoming_info = vanillaIdInfo `setStrictnessInfo` Just strict_sig
1157 -- Make arity and strictness agree
1159 -- Do *not* mark them as NoCafRefs, because they can indeed have
1160 -- CAF refs. For example, pAT_ERROR_ID calls GHC.Err.untangle,
1161 -- which has some CAFs
1162 -- In due course we may arrange that these error-y things are
1163 -- regarded by the GC as permanently live, in which case we
1164 -- can give them NoCaf info. As it is, any function that calls
1165 -- any pc_bottoming_Id will itself have CafRefs, which bloats
1168 strict_sig = mkStrictSig (mkTopDmdType [evalDmd] BotRes)
1169 -- These "bottom" out, no matter what their arguments