2 % (c) The AQUA Project, Glasgow University, 1993-1998
4 \section[Simplify]{The main module of the simplifier}
7 module Simplify ( simplTopBinds, simplExpr ) where
9 #include "HsVersions.h"
13 import Type hiding ( substTy, extendTvSubst )
16 import FamInstEnv ( FamInstEnv )
18 import MkId ( mkImpossibleExpr, seqId )
21 import Name ( mkSystemVarName )
23 import FamInstEnv ( topNormaliseType )
24 import DataCon ( DataCon, dataConWorkId, dataConRepStrictness )
26 import Demand ( isStrictDmd, splitStrictSig )
27 import PprCore ( pprParendExpr, pprCoreExpr )
28 import CoreUnfold ( mkUnfolding, mkCoreUnfolding, mkInlineRule,
29 exprIsConApp_maybe, callSiteInline, CallCtxt(..) )
31 import qualified CoreSubst
32 import CoreArity ( exprArity )
33 import Rules ( lookupRule, getRules )
34 import BasicTypes ( isMarkedStrict, Arity )
35 import CostCentre ( currentCCS, pushCCisNop )
36 import TysPrim ( realWorldStatePrimTy )
37 import PrelInfo ( realWorldPrimId )
38 import BasicTypes ( TopLevelFlag(..), isTopLevel, RecFlag(..) )
39 import MonadUtils ( foldlM, mapAccumLM )
40 import Maybes ( orElse )
41 import Data.List ( mapAccumL )
47 The guts of the simplifier is in this module, but the driver loop for
48 the simplifier is in SimplCore.lhs.
51 -----------------------------------------
52 *** IMPORTANT NOTE ***
53 -----------------------------------------
54 The simplifier used to guarantee that the output had no shadowing, but
55 it does not do so any more. (Actually, it never did!) The reason is
56 documented with simplifyArgs.
59 -----------------------------------------
60 *** IMPORTANT NOTE ***
61 -----------------------------------------
62 Many parts of the simplifier return a bunch of "floats" as well as an
63 expression. This is wrapped as a datatype SimplUtils.FloatsWith.
65 All "floats" are let-binds, not case-binds, but some non-rec lets may
66 be unlifted (with RHS ok-for-speculation).
70 -----------------------------------------
71 ORGANISATION OF FUNCTIONS
72 -----------------------------------------
74 - simplify all top-level binders
75 - for NonRec, call simplRecOrTopPair
76 - for Rec, call simplRecBind
79 ------------------------------
80 simplExpr (applied lambda) ==> simplNonRecBind
81 simplExpr (Let (NonRec ...) ..) ==> simplNonRecBind
82 simplExpr (Let (Rec ...) ..) ==> simplify binders; simplRecBind
84 ------------------------------
85 simplRecBind [binders already simplfied]
86 - use simplRecOrTopPair on each pair in turn
88 simplRecOrTopPair [binder already simplified]
89 Used for: recursive bindings (top level and nested)
90 top-level non-recursive bindings
92 - check for PreInlineUnconditionally
96 Used for: non-top-level non-recursive bindings
97 beta reductions (which amount to the same thing)
98 Because it can deal with strict arts, it takes a
99 "thing-inside" and returns an expression
101 - check for PreInlineUnconditionally
102 - simplify binder, including its IdInfo
111 simplNonRecX: [given a *simplified* RHS, but an *unsimplified* binder]
112 Used for: binding case-binder and constr args in a known-constructor case
113 - check for PreInLineUnconditionally
117 ------------------------------
118 simplLazyBind: [binder already simplified, RHS not]
119 Used for: recursive bindings (top level and nested)
120 top-level non-recursive bindings
121 non-top-level, but *lazy* non-recursive bindings
122 [must not be strict or unboxed]
123 Returns floats + an augmented environment, not an expression
124 - substituteIdInfo and add result to in-scope
125 [so that rules are available in rec rhs]
128 - float if exposes constructor or PAP
132 completeNonRecX: [binder and rhs both simplified]
133 - if the the thing needs case binding (unlifted and not ok-for-spec)
139 completeBind: [given a simplified RHS]
140 [used for both rec and non-rec bindings, top level and not]
141 - try PostInlineUnconditionally
142 - add unfolding [this is the only place we add an unfolding]
147 Right hand sides and arguments
148 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
149 In many ways we want to treat
150 (a) the right hand side of a let(rec), and
151 (b) a function argument
152 in the same way. But not always! In particular, we would
153 like to leave these arguments exactly as they are, so they
154 will match a RULE more easily.
159 It's harder to make the rule match if we ANF-ise the constructor,
160 or eta-expand the PAP:
162 f (let { a = g x; b = h x } in (a,b))
165 On the other hand if we see the let-defns
170 then we *do* want to ANF-ise and eta-expand, so that p and q
171 can be safely inlined.
173 Even floating lets out is a bit dubious. For let RHS's we float lets
174 out if that exposes a value, so that the value can be inlined more vigorously.
177 r = let x = e in (x,x)
179 Here, if we float the let out we'll expose a nice constructor. We did experiments
180 that showed this to be a generally good thing. But it was a bad thing to float
181 lets out unconditionally, because that meant they got allocated more often.
183 For function arguments, there's less reason to expose a constructor (it won't
184 get inlined). Just possibly it might make a rule match, but I'm pretty skeptical.
185 So for the moment we don't float lets out of function arguments either.
190 For eta expansion, we want to catch things like
192 case e of (a,b) -> \x -> case a of (p,q) -> \y -> r
194 If the \x was on the RHS of a let, we'd eta expand to bring the two
195 lambdas together. And in general that's a good thing to do. Perhaps
196 we should eta expand wherever we find a (value) lambda? Then the eta
197 expansion at a let RHS can concentrate solely on the PAP case.
200 %************************************************************************
202 \subsection{Bindings}
204 %************************************************************************
207 simplTopBinds :: SimplEnv -> [InBind] -> SimplM SimplEnv
209 simplTopBinds env0 binds0
210 = do { -- Put all the top-level binders into scope at the start
211 -- so that if a transformation rule has unexpectedly brought
212 -- anything into scope, then we don't get a complaint about that.
213 -- It's rather as if the top-level binders were imported.
214 ; env1 <- simplRecBndrs env0 (bindersOfBinds binds0)
215 ; dflags <- getDOptsSmpl
216 ; let dump_flag = dopt Opt_D_dump_inlinings dflags ||
217 dopt Opt_D_dump_rule_firings dflags
218 ; env2 <- simpl_binds dump_flag env1 binds0
219 ; freeTick SimplifierDone
222 -- We need to track the zapped top-level binders, because
223 -- they should have their fragile IdInfo zapped (notably occurrence info)
224 -- That's why we run down binds and bndrs' simultaneously.
226 -- The dump-flag emits a trace for each top-level binding, which
227 -- helps to locate the tracing for inlining and rule firing
228 simpl_binds :: Bool -> SimplEnv -> [InBind] -> SimplM SimplEnv
229 simpl_binds _ env [] = return env
230 simpl_binds dump env (bind:binds) = do { env' <- trace_bind dump bind $
232 ; simpl_binds dump env' binds }
234 trace_bind True bind = pprTrace "SimplBind" (ppr (bindersOf bind))
235 trace_bind False _ = \x -> x
237 simpl_bind env (Rec pairs) = simplRecBind env TopLevel pairs
238 simpl_bind env (NonRec b r) = simplRecOrTopPair env' TopLevel b b' r
240 (env', b') = addBndrRules env b (lookupRecBndr env b)
244 %************************************************************************
246 \subsection{Lazy bindings}
248 %************************************************************************
250 simplRecBind is used for
251 * recursive bindings only
254 simplRecBind :: SimplEnv -> TopLevelFlag
257 simplRecBind env0 top_lvl pairs0
258 = do { let (env_with_info, triples) = mapAccumL add_rules env0 pairs0
259 ; env1 <- go (zapFloats env_with_info) triples
260 ; return (env0 `addRecFloats` env1) }
261 -- addFloats adds the floats from env1,
262 -- _and_ updates env0 with the in-scope set from env1
264 add_rules :: SimplEnv -> (InBndr,InExpr) -> (SimplEnv, (InBndr, OutBndr, InExpr))
265 -- Add the (substituted) rules to the binder
266 add_rules env (bndr, rhs) = (env', (bndr, bndr', rhs))
268 (env', bndr') = addBndrRules env bndr (lookupRecBndr env bndr)
270 go env [] = return env
272 go env ((old_bndr, new_bndr, rhs) : pairs)
273 = do { env' <- simplRecOrTopPair env top_lvl old_bndr new_bndr rhs
277 simplOrTopPair is used for
278 * recursive bindings (whether top level or not)
279 * top-level non-recursive bindings
281 It assumes the binder has already been simplified, but not its IdInfo.
284 simplRecOrTopPair :: SimplEnv
286 -> InId -> OutBndr -> InExpr -- Binder and rhs
287 -> SimplM SimplEnv -- Returns an env that includes the binding
289 simplRecOrTopPair env top_lvl old_bndr new_bndr rhs
290 | preInlineUnconditionally env top_lvl old_bndr rhs -- Check for unconditional inline
291 = do { tick (PreInlineUnconditionally old_bndr)
292 ; return (extendIdSubst env old_bndr (mkContEx env rhs)) }
295 = simplLazyBind env top_lvl Recursive old_bndr new_bndr rhs env
296 -- May not actually be recursive, but it doesn't matter
300 simplLazyBind is used for
301 * [simplRecOrTopPair] recursive bindings (whether top level or not)
302 * [simplRecOrTopPair] top-level non-recursive bindings
303 * [simplNonRecE] non-top-level *lazy* non-recursive bindings
306 1. It assumes that the binder is *already* simplified,
307 and is in scope, and its IdInfo too, except unfolding
309 2. It assumes that the binder type is lifted.
311 3. It does not check for pre-inline-unconditionallly;
312 that should have been done already.
315 simplLazyBind :: SimplEnv
316 -> TopLevelFlag -> RecFlag
317 -> InId -> OutId -- Binder, both pre-and post simpl
318 -- The OutId has IdInfo, except arity, unfolding
319 -> InExpr -> SimplEnv -- The RHS and its environment
322 simplLazyBind env top_lvl is_rec bndr bndr1 rhs rhs_se
323 = do { let rhs_env = rhs_se `setInScope` env
324 (tvs, body) = case collectTyBinders rhs of
325 (tvs, body) | not_lam body -> (tvs,body)
326 | otherwise -> ([], rhs)
327 not_lam (Lam _ _) = False
329 -- Do not do the "abstract tyyvar" thing if there's
330 -- a lambda inside, becuase it defeats eta-reduction
331 -- f = /\a. \x. g a x
334 ; (body_env, tvs') <- simplBinders rhs_env tvs
335 -- See Note [Floating and type abstraction] in SimplUtils
338 ; (body_env1, body1) <- simplExprF body_env body mkRhsStop
339 -- ANF-ise a constructor or PAP rhs
340 ; (body_env2, body2) <- prepareRhs body_env1 bndr1 body1
343 <- if not (doFloatFromRhs top_lvl is_rec False body2 body_env2)
344 then -- No floating, just wrap up!
345 do { rhs' <- mkLam env tvs' (wrapFloats body_env2 body2)
346 ; return (env, rhs') }
348 else if null tvs then -- Simple floating
349 do { tick LetFloatFromLet
350 ; return (addFloats env body_env2, body2) }
352 else -- Do type-abstraction first
353 do { tick LetFloatFromLet
354 ; (poly_binds, body3) <- abstractFloats tvs' body_env2 body2
355 ; rhs' <- mkLam env tvs' body3
356 ; env' <- foldlM (addPolyBind top_lvl) env poly_binds
357 ; return (env', rhs') }
359 ; completeBind env' top_lvl bndr bndr1 rhs' }
362 A specialised variant of simplNonRec used when the RHS is already simplified,
363 notably in knownCon. It uses case-binding where necessary.
366 simplNonRecX :: SimplEnv
367 -> InId -- Old binder
368 -> OutExpr -- Simplified RHS
371 simplNonRecX env bndr new_rhs
372 | isDeadBinder bndr -- Not uncommon; e.g. case (a,b) of b { (p,q) -> p }
373 = return env -- Here b is dead, and we avoid creating
374 | otherwise -- the binding b = (a,b)
375 = do { (env', bndr') <- simplBinder env bndr
376 ; completeNonRecX env' (isStrictId bndr) bndr bndr' new_rhs }
378 completeNonRecX :: SimplEnv
380 -> InId -- Old binder
381 -> OutId -- New binder
382 -> OutExpr -- Simplified RHS
385 completeNonRecX env is_strict old_bndr new_bndr new_rhs
386 = do { (env1, rhs1) <- prepareRhs (zapFloats env) new_bndr new_rhs
388 if doFloatFromRhs NotTopLevel NonRecursive is_strict rhs1 env1
389 then do { tick LetFloatFromLet
390 ; return (addFloats env env1, rhs1) } -- Add the floats to the main env
391 else return (env, wrapFloats env1 rhs1) -- Wrap the floats around the RHS
392 ; completeBind env2 NotTopLevel old_bndr new_bndr rhs2 }
395 {- No, no, no! Do not try preInlineUnconditionally in completeNonRecX
396 Doing so risks exponential behaviour, because new_rhs has been simplified once already
397 In the cases described by the folowing commment, postInlineUnconditionally will
398 catch many of the relevant cases.
399 -- This happens; for example, the case_bndr during case of
400 -- known constructor: case (a,b) of x { (p,q) -> ... }
401 -- Here x isn't mentioned in the RHS, so we don't want to
402 -- create the (dead) let-binding let x = (a,b) in ...
404 -- Similarly, single occurrences can be inlined vigourously
405 -- e.g. case (f x, g y) of (a,b) -> ....
406 -- If a,b occur once we can avoid constructing the let binding for them.
408 Furthermore in the case-binding case preInlineUnconditionally risks extra thunks
409 -- Consider case I# (quotInt# x y) of
410 -- I# v -> let w = J# v in ...
411 -- If we gaily inline (quotInt# x y) for v, we end up building an
413 -- let w = J# (quotInt# x y) in ...
414 -- because quotInt# can fail.
416 | preInlineUnconditionally env NotTopLevel bndr new_rhs
417 = thing_inside (extendIdSubst env bndr (DoneEx new_rhs))
420 ----------------------------------
421 prepareRhs takes a putative RHS, checks whether it's a PAP or
422 constructor application and, if so, converts it to ANF, so that the
423 resulting thing can be inlined more easily. Thus
430 We also want to deal well cases like this
431 v = (f e1 `cast` co) e2
432 Here we want to make e1,e2 trivial and get
433 x1 = e1; x2 = e2; v = (f x1 `cast` co) v2
434 That's what the 'go' loop in prepareRhs does
437 prepareRhs :: SimplEnv -> OutId -> OutExpr -> SimplM (SimplEnv, OutExpr)
438 -- Adds new floats to the env iff that allows us to return a good RHS
439 prepareRhs env id (Cast rhs co) -- Note [Float coercions]
440 | (ty1, _ty2) <- coercionKind co -- Do *not* do this if rhs has an unlifted type
441 , not (isUnLiftedType ty1) -- see Note [Float coercions (unlifted)]
442 = do { (env', rhs') <- makeTrivialWithInfo env sanitised_info rhs
443 ; return (env', Cast rhs' co) }
445 sanitised_info = vanillaIdInfo `setStrictnessInfo` strictnessInfo info
446 `setDemandInfo` demandInfo info
449 prepareRhs env0 _ rhs0
450 = do { (_is_exp, env1, rhs1) <- go 0 env0 rhs0
451 ; return (env1, rhs1) }
453 go n_val_args env (Cast rhs co)
454 = do { (is_exp, env', rhs') <- go n_val_args env rhs
455 ; return (is_exp, env', Cast rhs' co) }
456 go n_val_args env (App fun (Type ty))
457 = do { (is_exp, env', rhs') <- go n_val_args env fun
458 ; return (is_exp, env', App rhs' (Type ty)) }
459 go n_val_args env (App fun arg)
460 = do { (is_exp, env', fun') <- go (n_val_args+1) env fun
462 True -> do { (env'', arg') <- makeTrivial env' arg
463 ; return (True, env'', App fun' arg') }
464 False -> return (False, env, App fun arg) }
465 go n_val_args env (Var fun)
466 = return (is_exp, env, Var fun)
468 is_exp = isExpandableApp fun n_val_args -- The fun a constructor or PAP
469 -- See Note [CONLIKE pragma] in BasicTypes
470 -- The definition of is_exp should match that in
471 -- OccurAnal.occAnalApp
474 = return (False, env, other)
478 Note [Float coercions]
479 ~~~~~~~~~~~~~~~~~~~~~~
480 When we find the binding
482 we'd like to transform it to
484 x = x `cast` co -- A trivial binding
485 There's a chance that e will be a constructor application or function, or something
486 like that, so moving the coerion to the usage site may well cancel the coersions
487 and lead to further optimisation. Example:
490 data instance T Int = T Int
492 foo :: Int -> Int -> Int
497 go n = case x of { T m -> go (n-m) }
498 -- This case should optimise
500 Note [Preserve strictness when floating coercions]
501 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
502 In the Note [Float coercions] transformation, keep the strictness info.
504 f = e `cast` co -- f has strictness SSL
506 f' = e -- f' also has strictness SSL
507 f = f' `cast` co -- f still has strictness SSL
509 Its not wrong to drop it on the floor, but better to keep it.
511 Note [Float coercions (unlifted)]
512 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
513 BUT don't do [Float coercions] if 'e' has an unlifted type.
516 foo :: Int = (error (# Int,Int #) "urk")
517 `cast` CoUnsafe (# Int,Int #) Int
519 If do the makeTrivial thing to the error call, we'll get
520 foo = case error (# Int,Int #) "urk" of v -> v `cast` ...
521 But 'v' isn't in scope!
523 These strange casts can happen as a result of case-of-case
524 bar = case (case x of { T -> (# 2,3 #); F -> error "urk" }) of
529 makeTrivial :: SimplEnv -> OutExpr -> SimplM (SimplEnv, OutExpr)
530 -- Binds the expression to a variable, if it's not trivial, returning the variable
531 makeTrivial env expr = makeTrivialWithInfo env vanillaIdInfo expr
533 makeTrivialWithInfo :: SimplEnv -> IdInfo -> OutExpr -> SimplM (SimplEnv, OutExpr)
534 -- Propagate strictness and demand info to the new binder
535 -- Note [Preserve strictness when floating coercions]
536 makeTrivialWithInfo env info expr
539 | otherwise -- See Note [Take care] below
540 = do { uniq <- getUniqueM
541 ; let name = mkSystemVarName uniq (fsLit "a")
542 var = mkLocalIdWithInfo name (exprType expr) info
543 ; env' <- completeNonRecX env False var var expr
544 ; return (env', substExpr env' (Var var)) }
545 -- The substitution is needed becase we're constructing a new binding
547 -- And if rhs is of form (rhs1 |> co), then we might get
550 -- and now a's RHS is trivial and can be substituted out, and that
551 -- is what completeNonRecX will do
555 %************************************************************************
557 \subsection{Completing a lazy binding}
559 %************************************************************************
562 * deals only with Ids, not TyVars
563 * takes an already-simplified binder and RHS
564 * is used for both recursive and non-recursive bindings
565 * is used for both top-level and non-top-level bindings
567 It does the following:
568 - tries discarding a dead binding
569 - tries PostInlineUnconditionally
570 - add unfolding [this is the only place we add an unfolding]
573 It does *not* attempt to do let-to-case. Why? Because it is used for
574 - top-level bindings (when let-to-case is impossible)
575 - many situations where the "rhs" is known to be a WHNF
576 (so let-to-case is inappropriate).
578 Nor does it do the atomic-argument thing
581 completeBind :: SimplEnv
582 -> TopLevelFlag -- Flag stuck into unfolding
583 -> InId -- Old binder
584 -> OutId -> OutExpr -- New binder and RHS
586 -- completeBind may choose to do its work
587 -- * by extending the substitution (e.g. let x = y in ...)
588 -- * or by adding to the floats in the envt
590 completeBind env top_lvl old_bndr new_bndr new_rhs
591 = do { let old_info = idInfo old_bndr
592 old_unf = unfoldingInfo old_info
593 occ_info = occInfo old_info
595 ; new_unfolding <- simplUnfolding env top_lvl old_bndr occ_info new_rhs old_unf
597 ; if postInlineUnconditionally env top_lvl new_bndr occ_info new_rhs new_unfolding
598 -- Inline and discard the binding
599 then do { tick (PostInlineUnconditionally old_bndr)
600 ; -- pprTrace "postInlineUnconditionally" (ppr old_bndr <+> equals <+> ppr new_rhs) $
601 return (extendIdSubst env old_bndr (DoneEx new_rhs)) }
602 -- Use the substitution to make quite, quite sure that the
603 -- substitution will happen, since we are going to discard the binding
605 else return (addNonRecWithUnf env new_bndr new_rhs new_unfolding) }
607 ------------------------------
608 addPolyBind :: TopLevelFlag -> SimplEnv -> OutBind -> SimplM SimplEnv
609 -- Add a new binding to the environment, complete with its unfolding
610 -- but *do not* do postInlineUnconditionally, because we have already
611 -- processed some of the scope of the binding
612 -- We still want the unfolding though. Consider
614 -- x = /\a. let y = ... in Just y
616 -- Then we float the y-binding out (via abstractFloats and addPolyBind)
617 -- but 'x' may well then be inlined in 'body' in which case we'd like the
618 -- opportunity to inline 'y' too.
620 addPolyBind top_lvl env (NonRec poly_id rhs)
621 = do { unfolding <- simplUnfolding env top_lvl poly_id NoOccInfo rhs noUnfolding
622 -- Assumes that poly_id did not have an INLINE prag
623 -- which is perhaps wrong. ToDo: think about this
624 ; return (addNonRecWithUnf env poly_id rhs unfolding) }
626 addPolyBind _ env bind@(Rec _) = return (extendFloats env bind)
627 -- Hack: letrecs are more awkward, so we extend "by steam"
628 -- without adding unfoldings etc. At worst this leads to
629 -- more simplifier iterations
631 ------------------------------
632 addNonRecWithUnf :: SimplEnv
633 -> OutId -> OutExpr -- New binder and RHS
634 -> Unfolding -- New unfolding
636 addNonRecWithUnf env new_bndr new_rhs new_unfolding
637 = let new_arity = exprArity new_rhs
638 old_arity = idArity new_bndr
639 info1 = idInfo new_bndr `setArityInfo` new_arity
641 -- Unfolding info: Note [Setting the new unfolding]
642 info2 = info1 `setUnfoldingInfo` new_unfolding
644 -- Demand info: Note [Setting the demand info]
645 info3 | isEvaldUnfolding new_unfolding = zapDemandInfo info2 `orElse` info2
648 final_id = new_bndr `setIdInfo` info3
649 dmd_arity = length $ fst $ splitStrictSig $ idStrictness new_bndr
651 ASSERT( isId new_bndr )
652 WARN( new_arity < old_arity || new_arity < dmd_arity,
653 (ptext (sLit "Arity decrease:") <+> ppr final_id <+> ppr old_arity
654 <+> ppr new_arity <+> ppr dmd_arity) )
655 -- Note [Arity decrease]
657 final_id `seq` -- This seq forces the Id, and hence its IdInfo,
658 -- and hence any inner substitutions
659 -- pprTrace "Binding" (ppr final_id <+> ppr unfolding) $
660 addNonRec env final_id new_rhs
661 -- The addNonRec adds it to the in-scope set too
663 ------------------------------
664 simplUnfolding :: SimplEnv-> TopLevelFlag
666 -> OccInfo -> OutExpr
667 -> Unfolding -> SimplM Unfolding
668 -- Note [Setting the new unfolding]
669 simplUnfolding env _ _ _ _ (DFunUnfolding con ops)
670 = return (DFunUnfolding con ops')
672 ops' = map (CoreSubst.substExpr (mkCoreSubst env)) ops
674 simplUnfolding env top_lvl _ _ _
675 (CoreUnfolding { uf_tmpl = expr, uf_arity = arity
676 , uf_src = src, uf_guidance = guide })
677 | isInlineRuleSource src
678 = do { expr' <- simplExpr (updMode updModeForInlineRules env) expr
679 -- See Note [Simplifying gently inside InlineRules] in SimplUtils
680 ; let src' = CoreSubst.substUnfoldingSource (mkCoreSubst env) src
681 ; return (mkCoreUnfolding (isTopLevel top_lvl) src' expr' arity guide) }
682 -- See Note [Top-level flag on inline rules] in CoreUnfold
684 simplUnfolding _ top_lvl id _occ_info new_rhs _
685 = return (mkUnfolding (isTopLevel top_lvl) (isBottomingId id) new_rhs)
686 -- We make an unfolding *even for loop-breakers*.
687 -- Reason: (a) It might be useful to know that they are WHNF
688 -- (b) In TidyPgm we currently assume that, if we want to
689 -- expose the unfolding then indeed we *have* an unfolding
690 -- to expose. (We could instead use the RHS, but currently
691 -- we don't.) The simple thing is always to have one.
694 Note [Arity decrease]
695 ~~~~~~~~~~~~~~~~~~~~~
696 Generally speaking the arity of a binding should not decrease. But it *can*
697 legitimately happen becuase of RULES. Eg
699 where g has arity 2, will have arity 2. But if there's a rewrite rule
701 where h has arity 1, then f's arity will decrease. Here's a real-life example,
702 which is in the output of Specialise:
705 $dm {Arity 2} = \d.\x. op d
706 {-# RULES forall d. $dm Int d = $s$dm #-}
708 dInt = MkD .... opInt ...
709 opInt {Arity 1} = $dm dInt
711 $s$dm {Arity 0} = \x. op dInt }
713 Here opInt has arity 1; but when we apply the rule its arity drops to 0.
714 That's why Specialise goes to a little trouble to pin the right arity
715 on specialised functions too.
717 Note [Setting the new unfolding]
718 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
719 * If there's an INLINE pragma, we simplify the RHS gently. Maybe we
720 should do nothing at all, but simplifying gently might get rid of
723 * If not, we make an unfolding from the new RHS. But *only* for
724 non-loop-breakers. Making loop breakers not have an unfolding at all
725 means that we can avoid tests in exprIsConApp, for example. This is
726 important: if exprIsConApp says 'yes' for a recursive thing, then we
727 can get into an infinite loop
729 If there's an InlineRule on a loop breaker, we hang on to the inlining.
730 It's pretty dodgy, but the user did say 'INLINE'. May need to revisit
733 Note [Setting the demand info]
734 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
735 If the unfolding is a value, the demand info may
736 go pear-shaped, so we nuke it. Example:
738 case x of (p,q) -> h p q x
739 Here x is certainly demanded. But after we've nuked
740 the case, we'll get just
741 let x = (a,b) in h a b x
742 and now x is not demanded (I'm assuming h is lazy)
743 This really happens. Similarly
744 let f = \x -> e in ...f..f...
745 After inlining f at some of its call sites the original binding may
746 (for example) be no longer strictly demanded.
747 The solution here is a bit ad hoc...
750 %************************************************************************
752 \subsection[Simplify-simplExpr]{The main function: simplExpr}
754 %************************************************************************
756 The reason for this OutExprStuff stuff is that we want to float *after*
757 simplifying a RHS, not before. If we do so naively we get quadratic
758 behaviour as things float out.
760 To see why it's important to do it after, consider this (real) example:
774 a -- Can't inline a this round, cos it appears twice
778 Each of the ==> steps is a round of simplification. We'd save a
779 whole round if we float first. This can cascade. Consider
784 let f = let d1 = ..d.. in \y -> e
788 in \x -> ...(\y ->e)...
790 Only in this second round can the \y be applied, and it
791 might do the same again.
795 simplExpr :: SimplEnv -> CoreExpr -> SimplM CoreExpr
796 simplExpr env expr = simplExprC env expr mkBoringStop
798 simplExprC :: SimplEnv -> CoreExpr -> SimplCont -> SimplM CoreExpr
799 -- Simplify an expression, given a continuation
800 simplExprC env expr cont
801 = -- pprTrace "simplExprC" (ppr expr $$ ppr cont {- $$ ppr (seIdSubst env) -} $$ ppr (seFloats env) ) $
802 do { (env', expr') <- simplExprF (zapFloats env) expr cont
803 ; -- pprTrace "simplExprC ret" (ppr expr $$ ppr expr') $
804 -- pprTrace "simplExprC ret3" (ppr (seInScope env')) $
805 -- pprTrace "simplExprC ret4" (ppr (seFloats env')) $
806 return (wrapFloats env' expr') }
808 --------------------------------------------------
809 simplExprF :: SimplEnv -> InExpr -> SimplCont
810 -> SimplM (SimplEnv, OutExpr)
812 simplExprF env e cont
813 = -- pprTrace "simplExprF" (ppr e $$ ppr cont $$ ppr (seTvSubst env) $$ ppr (seIdSubst env) {- $$ ppr (seFloats env) -} ) $
814 simplExprF' env e cont
816 simplExprF' :: SimplEnv -> InExpr -> SimplCont
817 -> SimplM (SimplEnv, OutExpr)
818 simplExprF' env (Var v) cont = simplVar env v cont
819 simplExprF' env (Lit lit) cont = rebuild env (Lit lit) cont
820 simplExprF' env (Note n expr) cont = simplNote env n expr cont
821 simplExprF' env (Cast body co) cont = simplCast env body co cont
822 simplExprF' env (App fun arg) cont = simplExprF env fun $
823 ApplyTo NoDup arg env cont
825 simplExprF' env expr@(Lam _ _) cont
826 = simplLam env (map zap bndrs) body cont
827 -- The main issue here is under-saturated lambdas
828 -- (\x1. \x2. e) arg1
829 -- Here x1 might have "occurs-once" occ-info, because occ-info
830 -- is computed assuming that a group of lambdas is applied
831 -- all at once. If there are too few args, we must zap the
834 n_args = countArgs cont
835 n_params = length bndrs
836 (bndrs, body) = collectBinders expr
837 zap | n_args >= n_params = \b -> b
838 | otherwise = \b -> if isTyVar b then b
840 -- NB: we count all the args incl type args
841 -- so we must count all the binders (incl type lambdas)
843 simplExprF' env (Type ty) cont
844 = ASSERT( contIsRhsOrArg cont )
845 do { ty' <- simplCoercion env ty
846 ; rebuild env (Type ty') cont }
848 simplExprF' env (Case scrut bndr _ alts) cont
849 | not (switchIsOn (getSwitchChecker env) NoCaseOfCase)
850 = -- Simplify the scrutinee with a Select continuation
851 simplExprF env scrut (Select NoDup bndr alts env cont)
854 = -- If case-of-case is off, simply simplify the case expression
855 -- in a vanilla Stop context, and rebuild the result around it
856 do { case_expr' <- simplExprC env scrut case_cont
857 ; rebuild env case_expr' cont }
859 case_cont = Select NoDup bndr alts env mkBoringStop
861 simplExprF' env (Let (Rec pairs) body) cont
862 = do { env' <- simplRecBndrs env (map fst pairs)
863 -- NB: bndrs' don't have unfoldings or rules
864 -- We add them as we go down
866 ; env'' <- simplRecBind env' NotTopLevel pairs
867 ; simplExprF env'' body cont }
869 simplExprF' env (Let (NonRec bndr rhs) body) cont
870 = simplNonRecE env bndr (rhs, env) ([], body) cont
872 ---------------------------------
873 simplType :: SimplEnv -> InType -> SimplM OutType
874 -- Kept monadic just so we can do the seqType
876 = -- pprTrace "simplType" (ppr ty $$ ppr (seTvSubst env)) $
877 seqType new_ty `seq` return new_ty
879 new_ty = substTy env ty
881 ---------------------------------
882 simplCoercion :: SimplEnv -> InType -> SimplM OutType
883 -- The InType isn't *necessarily* a coercion, but it might be
884 -- (in a type application, say) and optCoercion is a no-op on types
886 = seqType new_co `seq` return new_co
888 new_co = optCoercion (getTvSubst env) co
892 %************************************************************************
894 \subsection{The main rebuilder}
896 %************************************************************************
899 rebuild :: SimplEnv -> OutExpr -> SimplCont -> SimplM (SimplEnv, OutExpr)
900 -- At this point the substitution in the SimplEnv should be irrelevant
901 -- only the in-scope set and floats should matter
902 rebuild env expr cont0
903 = -- pprTrace "rebuild" (ppr expr $$ ppr cont0 $$ ppr (seFloats env)) $
905 Stop {} -> return (env, expr)
906 CoerceIt co cont -> rebuild env (mkCoerce co expr) cont
907 Select _ bndr alts se cont -> rebuildCase (se `setFloats` env) expr bndr alts cont
908 StrictArg info _ cont -> rebuildCall env (info `addArgTo` expr) cont
909 StrictBind b bs body se cont -> do { env' <- simplNonRecX (se `setFloats` env) b expr
910 ; simplLam env' bs body cont }
911 ApplyTo _ arg se cont -> do { arg' <- simplExpr (se `setInScope` env) arg
912 ; rebuild env (App expr arg') cont }
916 %************************************************************************
920 %************************************************************************
923 simplCast :: SimplEnv -> InExpr -> Coercion -> SimplCont
924 -> SimplM (SimplEnv, OutExpr)
925 simplCast env body co0 cont0
926 = do { co1 <- simplCoercion env co0
927 ; simplExprF env body (addCoerce co1 cont0) }
929 addCoerce co cont = add_coerce co (coercionKind co) cont
931 add_coerce _co (s1, k1) cont -- co :: ty~ty
932 | s1 `coreEqType` k1 = cont -- is a no-op
934 add_coerce co1 (s1, _k2) (CoerceIt co2 cont)
935 | (_l1, t1) <- coercionKind co2
936 -- e |> (g1 :: S1~L) |> (g2 :: L~T1)
939 -- e |> (g1 . g2 :: S1~T1) otherwise
941 -- For example, in the initial form of a worker
942 -- we may find (coerce T (coerce S (\x.e))) y
943 -- and we'd like it to simplify to e[y/x] in one round
945 , s1 `coreEqType` t1 = cont -- The coerces cancel out
946 | otherwise = CoerceIt (mkTransCoercion co1 co2) cont
948 add_coerce co (s1s2, _t1t2) (ApplyTo dup (Type arg_ty) arg_se cont)
949 -- (f |> g) ty ---> (f ty) |> (g @ ty)
950 -- This implements the PushT and PushC rules from the paper
951 | Just (tyvar,_) <- splitForAllTy_maybe s1s2
953 (new_arg_ty, new_cast)
954 | isCoVar tyvar = (new_arg_co, mkCselRCoercion co) -- PushC rule
955 | otherwise = (ty', mkInstCoercion co ty') -- PushT rule
957 ApplyTo dup (Type new_arg_ty) (zapSubstEnv arg_se) (addCoerce new_cast cont)
959 ty' = substTy (arg_se `setInScope` env) arg_ty
960 new_arg_co = mkCsel1Coercion co `mkTransCoercion`
961 ty' `mkTransCoercion`
962 mkSymCoercion (mkCsel2Coercion co)
964 add_coerce co (s1s2, _t1t2) (ApplyTo dup arg arg_se cont)
965 | not (isTypeArg arg) -- This implements the Push rule from the paper
966 , isFunTy s1s2 -- t1t2 must be a function type, becuase it's applied
967 -- (e |> (g :: s1s2 ~ t1->t2)) f
969 -- (e (f |> (arg g :: t1~s1))
970 -- |> (res g :: s2->t2)
972 -- t1t2 must be a function type, t1->t2, because it's applied
973 -- to something but s1s2 might conceivably not be
975 -- When we build the ApplyTo we can't mix the out-types
976 -- with the InExpr in the argument, so we simply substitute
977 -- to make it all consistent. It's a bit messy.
978 -- But it isn't a common case.
980 -- Example of use: Trac #995
981 = ApplyTo dup new_arg (zapSubstEnv arg_se) (addCoerce co2 cont)
983 -- we split coercion t1->t2 ~ s1->s2 into t1 ~ s1 and
984 -- t2 ~ s2 with left and right on the curried form:
985 -- (->) t1 t2 ~ (->) s1 s2
986 [co1, co2] = decomposeCo 2 co
987 new_arg = mkCoerce (mkSymCoercion co1) arg'
988 arg' = substExpr (arg_se `setInScope` env) arg
990 add_coerce co _ cont = CoerceIt co cont
994 %************************************************************************
998 %************************************************************************
1001 simplLam :: SimplEnv -> [InId] -> InExpr -> SimplCont
1002 -> SimplM (SimplEnv, OutExpr)
1004 simplLam env [] body cont = simplExprF env body cont
1007 simplLam env (bndr:bndrs) body (ApplyTo _ arg arg_se cont)
1008 = do { tick (BetaReduction bndr)
1009 ; simplNonRecE env bndr (arg, arg_se) (bndrs, body) cont }
1011 -- Not enough args, so there are real lambdas left to put in the result
1012 simplLam env bndrs body cont
1013 = do { (env', bndrs') <- simplLamBndrs env bndrs
1014 ; body' <- simplExpr env' body
1015 ; new_lam <- mkLam env' bndrs' body'
1016 ; rebuild env' new_lam cont }
1019 simplNonRecE :: SimplEnv
1020 -> InBndr -- The binder
1021 -> (InExpr, SimplEnv) -- Rhs of binding (or arg of lambda)
1022 -> ([InBndr], InExpr) -- Body of the let/lambda
1025 -> SimplM (SimplEnv, OutExpr)
1027 -- simplNonRecE is used for
1028 -- * non-top-level non-recursive lets in expressions
1031 -- It deals with strict bindings, via the StrictBind continuation,
1032 -- which may abort the whole process
1034 -- The "body" of the binding comes as a pair of ([InId],InExpr)
1035 -- representing a lambda; so we recurse back to simplLam
1036 -- Why? Because of the binder-occ-info-zapping done before
1037 -- the call to simplLam in simplExprF (Lam ...)
1039 -- First deal with type applications and type lets
1040 -- (/\a. e) (Type ty) and (let a = Type ty in e)
1041 simplNonRecE env bndr (Type ty_arg, rhs_se) (bndrs, body) cont
1042 = ASSERT( isTyVar bndr )
1043 do { ty_arg' <- simplType (rhs_se `setInScope` env) ty_arg
1044 ; simplLam (extendTvSubst env bndr ty_arg') bndrs body cont }
1046 simplNonRecE env bndr (rhs, rhs_se) (bndrs, body) cont
1047 | preInlineUnconditionally env NotTopLevel bndr rhs
1048 = do { tick (PreInlineUnconditionally bndr)
1049 ; simplLam (extendIdSubst env bndr (mkContEx rhs_se rhs)) bndrs body cont }
1052 = do { simplExprF (rhs_se `setFloats` env) rhs
1053 (StrictBind bndr bndrs body env cont) }
1056 = ASSERT( not (isTyVar bndr) )
1057 do { (env1, bndr1) <- simplNonRecBndr env bndr
1058 ; let (env2, bndr2) = addBndrRules env1 bndr bndr1
1059 ; env3 <- simplLazyBind env2 NotTopLevel NonRecursive bndr bndr2 rhs rhs_se
1060 ; simplLam env3 bndrs body cont }
1064 %************************************************************************
1068 %************************************************************************
1071 -- Hack alert: we only distinguish subsumed cost centre stacks for the
1072 -- purposes of inlining. All other CCCSs are mapped to currentCCS.
1073 simplNote :: SimplEnv -> Note -> CoreExpr -> SimplCont
1074 -> SimplM (SimplEnv, OutExpr)
1075 simplNote env (SCC cc) e cont
1076 | pushCCisNop cc (getEnclosingCC env) -- scc "f" (...(scc "f" e)...)
1077 = simplExprF env e cont -- ==> scc "f" (...e...)
1079 = do { e' <- simplExpr (setEnclosingCC env currentCCS) e
1080 ; rebuild env (mkSCC cc e') cont }
1082 simplNote env (CoreNote s) e cont
1083 = do { e' <- simplExpr env e
1084 ; rebuild env (Note (CoreNote s) e') cont }
1088 %************************************************************************
1090 \subsection{Dealing with calls}
1092 %************************************************************************
1095 simplVar :: SimplEnv -> Id -> SimplCont -> SimplM (SimplEnv, OutExpr)
1096 simplVar env var cont
1097 = case substId env var of
1098 DoneEx e -> simplExprF (zapSubstEnv env) e cont
1099 ContEx tvs ids e -> simplExprF (setSubstEnv env tvs ids) e cont
1100 DoneId var1 -> completeCall env var1 cont
1101 -- Note [zapSubstEnv]
1102 -- The template is already simplified, so don't re-substitute.
1103 -- This is VITAL. Consider
1105 -- let y = \z -> ...x... in
1107 -- We'll clone the inner \x, adding x->x' in the id_subst
1108 -- Then when we inline y, we must *not* replace x by x' in
1109 -- the inlined copy!!
1111 ---------------------------------------------------------
1112 -- Dealing with a call site
1114 completeCall :: SimplEnv -> Id -> SimplCont -> SimplM (SimplEnv, OutExpr)
1115 completeCall env var cont
1116 = do { ------------- Try inlining ----------------
1117 dflags <- getDOptsSmpl
1118 ; let (args,call_cont) = contArgs cont
1119 -- The args are OutExprs, obtained by *lazily* substituting
1120 -- in the args found in cont. These args are only examined
1121 -- to limited depth (unless a rule fires). But we must do
1122 -- the substitution; rule matching on un-simplified args would
1125 arg_infos = [interestingArg arg | arg <- args, isValArg arg]
1126 n_val_args = length arg_infos
1127 interesting_cont = interestingCallContext call_cont
1128 unfolding = activeUnfolding env var
1129 maybe_inline = callSiteInline dflags var unfolding
1130 (null args) arg_infos interesting_cont
1131 ; case maybe_inline of {
1132 Just unfolding -- There is an inlining!
1133 -> do { tick (UnfoldingDone var)
1134 ; (if dopt Opt_D_dump_inlinings dflags then
1135 pprTrace ("Inlining done: " ++ showSDoc (ppr var)) (vcat [
1136 text "Before:" <+> ppr var <+> sep (map pprParendExpr args),
1137 text "Inlined fn: " <+> nest 2 (ppr unfolding),
1138 text "Cont: " <+> ppr call_cont])
1141 simplExprF (zapSubstEnv env) unfolding cont }
1143 ; Nothing -> do -- No inlining!
1145 { rule_base <- getSimplRules
1146 ; let info = mkArgInfo var (getRules rule_base var) n_val_args call_cont
1147 ; rebuildCall env info cont
1150 rebuildCall :: SimplEnv
1153 -> SimplM (SimplEnv, OutExpr)
1154 rebuildCall env (ArgInfo { ai_fun = fun, ai_args = rev_args, ai_strs = [] }) cont
1155 -- When we run out of strictness args, it means
1156 -- that the call is definitely bottom; see SimplUtils.mkArgInfo
1157 -- Then we want to discard the entire strict continuation. E.g.
1158 -- * case (error "hello") of { ... }
1159 -- * (error "Hello") arg
1160 -- * f (error "Hello") where f is strict
1162 -- Then, especially in the first of these cases, we'd like to discard
1163 -- the continuation, leaving just the bottoming expression. But the
1164 -- type might not be right, so we may have to add a coerce.
1165 | not (contIsTrivial cont) -- Only do this if there is a non-trivial
1166 = return (env, mk_coerce res) -- contination to discard, else we do it
1167 where -- again and again!
1168 res = mkApps (Var fun) (reverse rev_args)
1169 res_ty = exprType res
1170 cont_ty = contResultType env res_ty cont
1171 co = mkUnsafeCoercion res_ty cont_ty
1172 mk_coerce expr | cont_ty `coreEqType` res_ty = expr
1173 | otherwise = mkCoerce co expr
1175 rebuildCall env info (ApplyTo _ (Type arg_ty) se cont)
1176 = do { ty' <- simplCoercion (se `setInScope` env) arg_ty
1177 ; rebuildCall env (info `addArgTo` Type ty') cont }
1179 rebuildCall env info@(ArgInfo { ai_encl = encl_rules
1180 , ai_strs = str:strs, ai_discs = disc:discs })
1181 (ApplyTo _ arg arg_se cont)
1182 | str -- Strict argument
1183 = -- pprTrace "Strict Arg" (ppr arg $$ ppr (seIdSubst env) $$ ppr (seInScope env)) $
1184 simplExprF (arg_se `setFloats` env) arg
1185 (StrictArg info' cci cont)
1188 | otherwise -- Lazy argument
1189 -- DO NOT float anything outside, hence simplExprC
1190 -- There is no benefit (unlike in a let-binding), and we'd
1191 -- have to be very careful about bogus strictness through
1192 -- floating a demanded let.
1193 = do { arg' <- simplExprC (arg_se `setInScope` env) arg
1195 ; rebuildCall env (addArgTo info' arg') cont }
1197 info' = info { ai_strs = strs, ai_discs = discs }
1198 cci | encl_rules || disc > 0 = ArgCtxt encl_rules -- Be keener here
1199 | otherwise = BoringCtxt -- Nothing interesting
1201 rebuildCall env (ArgInfo { ai_fun = fun, ai_args = rev_args, ai_rules = rules }) cont
1202 = do { -- We've accumulated a simplified call in <fun,rev_args>
1203 -- so try rewrite rules; see Note [RULEs apply to simplified arguments]
1204 -- See also Note [Rules for recursive functions]
1205 ; let args = reverse rev_args
1206 env' = zapSubstEnv env
1207 ; mb_rule <- tryRules env rules fun args cont
1209 Just (n_args, rule_rhs) -> simplExprF env' rule_rhs $
1210 pushArgs env' (drop n_args args) cont ;
1211 -- n_args says how many args the rule consumed
1212 ; Nothing -> rebuild env (mkApps (Var fun) args) cont -- No rules
1216 Note [RULES apply to simplified arguments]
1217 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1218 It's very desirable to try RULES once the arguments have been simplified, because
1219 doing so ensures that rule cascades work in one pass. Consider
1220 {-# RULES g (h x) = k x
1223 Then we want to rewrite (g (h x)) to (k x) and only then try f's rules. If
1224 we match f's rules against the un-simplified RHS, it won't match. This
1225 makes a particularly big difference when superclass selectors are involved:
1226 op ($p1 ($p2 (df d)))
1227 We want all this to unravel in one sweeep.
1231 This part of the simplifier may break the no-shadowing invariant
1233 f (...(\a -> e)...) (case y of (a,b) -> e')
1234 where f is strict in its second arg
1235 If we simplify the innermost one first we get (...(\a -> e)...)
1236 Simplifying the second arg makes us float the case out, so we end up with
1237 case y of (a,b) -> f (...(\a -> e)...) e'
1238 So the output does not have the no-shadowing invariant. However, there is
1239 no danger of getting name-capture, because when the first arg was simplified
1240 we used an in-scope set that at least mentioned all the variables free in its
1241 static environment, and that is enough.
1243 We can't just do innermost first, or we'd end up with a dual problem:
1244 case x of (a,b) -> f e (...(\a -> e')...)
1246 I spent hours trying to recover the no-shadowing invariant, but I just could
1247 not think of an elegant way to do it. The simplifier is already knee-deep in
1248 continuations. We have to keep the right in-scope set around; AND we have
1249 to get the effect that finding (error "foo") in a strict arg position will
1250 discard the entire application and replace it with (error "foo"). Getting
1251 all this at once is TOO HARD!
1254 %************************************************************************
1258 %************************************************************************
1261 tryRules :: SimplEnv -> [CoreRule]
1262 -> Id -> [OutExpr] -> SimplCont
1263 -> SimplM (Maybe (Arity, CoreExpr)) -- The arity is the number of
1264 -- args consumed by the rule
1265 tryRules env rules fn args call_cont
1269 = do { dflags <- getDOptsSmpl
1270 ; case activeRule dflags env of {
1271 Nothing -> return Nothing ; -- No rules apply
1273 case lookupRule act_fn (activeUnfInRule env) (getInScope env) fn args rules of {
1274 Nothing -> return Nothing ; -- No rule matches
1275 Just (rule, rule_rhs) ->
1277 do { tick (RuleFired (ru_name rule))
1278 ; (if dopt Opt_D_dump_rule_firings dflags then
1279 pprTrace "Rule fired" (vcat [
1280 text "Rule:" <+> ftext (ru_name rule),
1281 text "Before:" <+> ppr fn <+> sep (map pprParendExpr args),
1282 text "After: " <+> pprCoreExpr rule_rhs,
1283 text "Cont: " <+> ppr call_cont])
1286 return (Just (ruleArity rule, rule_rhs)) }}}}
1289 Note [Rules for recursive functions]
1290 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1291 You might think that we shouldn't apply rules for a loop breaker:
1292 doing so might give rise to an infinite loop, because a RULE is
1293 rather like an extra equation for the function:
1294 RULE: f (g x) y = x+y
1297 But it's too drastic to disable rules for loop breakers.
1298 Even the foldr/build rule would be disabled, because foldr
1299 is recursive, and hence a loop breaker:
1300 foldr k z (build g) = g k z
1301 So it's up to the programmer: rules can cause divergence
1304 %************************************************************************
1306 Rebuilding a cse expression
1308 %************************************************************************
1310 Note [Case elimination]
1311 ~~~~~~~~~~~~~~~~~~~~~~~
1312 The case-elimination transformation discards redundant case expressions.
1313 Start with a simple situation:
1315 case x# of ===> e[x#/y#]
1318 (when x#, y# are of primitive type, of course). We can't (in general)
1319 do this for algebraic cases, because we might turn bottom into
1322 The code in SimplUtils.prepareAlts has the effect of generalise this
1323 idea to look for a case where we're scrutinising a variable, and we
1324 know that only the default case can match. For example:
1328 DEFAULT -> ...(case x of
1332 Here the inner case is first trimmed to have only one alternative, the
1333 DEFAULT, after which it's an instance of the previous case. This
1334 really only shows up in eliminating error-checking code.
1336 We also make sure that we deal with this very common case:
1341 Here we are using the case as a strict let; if x is used only once
1342 then we want to inline it. We have to be careful that this doesn't
1343 make the program terminate when it would have diverged before, so we
1345 - e is already evaluated (it may so if e is a variable)
1346 - x is used strictly, or
1348 Lastly, the code in SimplUtils.mkCase combines identical RHSs. So
1350 case e of ===> case e of DEFAULT -> r
1354 Now again the case may be elminated by the CaseElim transformation.
1357 Further notes about case elimination
1358 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1359 Consider: test :: Integer -> IO ()
1362 Turns out that this compiles to:
1365 eta1 :: State# RealWorld ->
1366 case PrelNum.< eta PrelNum.zeroInteger of wild { __DEFAULT ->
1368 (PrelNum.jtos eta ($w[] @ Char))
1370 of wild1 { (# new_s, a4 #) -> PrelIO.lvl23 new_s }}
1372 Notice the strange '<' which has no effect at all. This is a funny one.
1373 It started like this:
1375 f x y = if x < 0 then jtos x
1376 else if y==0 then "" else jtos x
1378 At a particular call site we have (f v 1). So we inline to get
1380 if v < 0 then jtos x
1381 else if 1==0 then "" else jtos x
1383 Now simplify the 1==0 conditional:
1385 if v<0 then jtos v else jtos v
1387 Now common-up the two branches of the case:
1389 case (v<0) of DEFAULT -> jtos v
1391 Why don't we drop the case? Because it's strict in v. It's technically
1392 wrong to drop even unnecessary evaluations, and in practice they
1393 may be a result of 'seq' so we *definitely* don't want to drop those.
1394 I don't really know how to improve this situation.
1397 ---------------------------------------------------------
1398 -- Eliminate the case if possible
1400 rebuildCase, reallyRebuildCase
1402 -> OutExpr -- Scrutinee
1403 -> InId -- Case binder
1404 -> [InAlt] -- Alternatives (inceasing order)
1406 -> SimplM (SimplEnv, OutExpr)
1408 --------------------------------------------------
1409 -- 1. Eliminate the case if there's a known constructor
1410 --------------------------------------------------
1412 rebuildCase env scrut case_bndr alts cont
1413 | Lit lit <- scrut -- No need for same treatment as constructors
1414 -- because literals are inlined more vigorously
1415 = do { tick (KnownBranch case_bndr)
1416 ; case findAlt (LitAlt lit) alts of
1417 Nothing -> missingAlt env case_bndr alts cont
1418 Just (_, bs, rhs) -> simple_rhs bs rhs }
1420 | Just (con, ty_args, other_args) <- exprIsConApp_maybe (activeUnfInRule env) scrut
1421 -- Works when the scrutinee is a variable with a known unfolding
1422 -- as well as when it's an explicit constructor application
1423 = do { tick (KnownBranch case_bndr)
1424 ; case findAlt (DataAlt con) alts of
1425 Nothing -> missingAlt env case_bndr alts cont
1426 Just (DEFAULT, bs, rhs) -> simple_rhs bs rhs
1427 Just (_, bs, rhs) -> knownCon env scrut con ty_args other_args
1428 case_bndr bs rhs cont
1431 simple_rhs bs rhs = ASSERT( null bs )
1432 do { env' <- simplNonRecX env case_bndr scrut
1433 ; simplExprF env' rhs cont }
1436 --------------------------------------------------
1437 -- 2. Eliminate the case if scrutinee is evaluated
1438 --------------------------------------------------
1440 rebuildCase env scrut case_bndr [(_, bndrs, rhs)] cont
1441 -- See if we can get rid of the case altogether
1442 -- See Note [Case eliminiation]
1443 -- mkCase made sure that if all the alternatives are equal,
1444 -- then there is now only one (DEFAULT) rhs
1445 | all isDeadBinder bndrs -- bndrs are [InId]
1447 -- Check that the scrutinee can be let-bound instead of case-bound
1448 , exprOkForSpeculation scrut
1449 -- OK not to evaluate it
1450 -- This includes things like (==# a# b#)::Bool
1451 -- so that we simplify
1452 -- case ==# a# b# of { True -> x; False -> x }
1455 -- This particular example shows up in default methods for
1456 -- comparision operations (e.g. in (>=) for Int.Int32)
1457 || exprIsHNF scrut -- It's already evaluated
1458 || var_demanded_later scrut -- It'll be demanded later
1460 -- || not opt_SimplPedanticBottoms) -- Or we don't care!
1461 -- We used to allow improving termination by discarding cases, unless -fpedantic-bottoms was on,
1462 -- but that breaks badly for the dataToTag# primop, which relies on a case to evaluate
1463 -- its argument: case x of { y -> dataToTag# y }
1464 -- Here we must *not* discard the case, because dataToTag# just fetches the tag from
1465 -- the info pointer. So we'll be pedantic all the time, and see if that gives any
1467 -- Also we don't want to discard 'seq's
1468 = do { tick (CaseElim case_bndr)
1469 ; env' <- simplNonRecX env case_bndr scrut
1470 ; simplExprF env' rhs cont }
1472 -- The case binder is going to be evaluated later,
1473 -- and the scrutinee is a simple variable
1474 var_demanded_later (Var v) = isStrictDmd (idDemandInfo case_bndr)
1475 && not (isTickBoxOp v)
1476 -- ugly hack; covering this case is what
1477 -- exprOkForSpeculation was intended for.
1478 var_demanded_later _ = False
1480 --------------------------------------------------
1481 -- 3. Try seq rules; see Note [User-defined RULES for seq] in MkId
1482 --------------------------------------------------
1484 rebuildCase env scrut case_bndr alts@[(_, bndrs, rhs)] cont
1485 | all isDeadBinder (case_bndr : bndrs) -- So this is just 'seq'
1486 = do { let rhs' = substExpr env rhs
1487 out_args = [Type (substTy env (idType case_bndr)),
1488 Type (exprType rhs'), scrut, rhs']
1489 -- Lazily evaluated, so we don't do most of this
1491 ; rule_base <- getSimplRules
1492 ; mb_rule <- tryRules env (getRules rule_base seqId) seqId out_args cont
1494 Just (n_args, res) -> simplExprF (zapSubstEnv env)
1495 (mkApps res (drop n_args out_args))
1497 Nothing -> reallyRebuildCase env scrut case_bndr alts cont }
1499 rebuildCase env scrut case_bndr alts cont
1500 = reallyRebuildCase env scrut case_bndr alts cont
1502 --------------------------------------------------
1503 -- 3. Catch-all case
1504 --------------------------------------------------
1506 reallyRebuildCase env scrut case_bndr alts cont
1507 = do { -- Prepare the continuation;
1508 -- The new subst_env is in place
1509 (env', dup_cont, nodup_cont) <- prepareCaseCont env alts cont
1511 -- Simplify the alternatives
1512 ; (scrut', case_bndr', alts') <- simplAlts env' scrut case_bndr alts dup_cont
1514 -- Check for empty alternatives
1515 ; if null alts' then missingAlt env case_bndr alts cont
1517 { dflags <- getDOptsSmpl
1518 ; case_expr <- mkCase dflags scrut' case_bndr' alts'
1520 -- Notice that rebuild gets the in-scope set from env', not alt_env
1521 -- (which in any case is only build in simplAlts)
1522 -- The case binder *not* scope over the whole returned case-expression
1523 ; rebuild env' case_expr nodup_cont } }
1526 simplCaseBinder checks whether the scrutinee is a variable, v. If so,
1527 try to eliminate uses of v in the RHSs in favour of case_bndr; that
1528 way, there's a chance that v will now only be used once, and hence
1531 Historical note: we use to do the "case binder swap" in the Simplifier
1532 so there were additional complications if the scrutinee was a variable.
1533 Now the binder-swap stuff is done in the occurrence analyer; see
1534 OccurAnal Note [Binder swap].
1538 If the case binder is not dead, then neither are the pattern bound
1540 case <any> of x { (a,b) ->
1541 case x of { (p,q) -> p } }
1542 Here (a,b) both look dead, but come alive after the inner case is eliminated.
1543 The point is that we bring into the envt a binding
1545 after the outer case, and that makes (a,b) alive. At least we do unless
1546 the case binder is guaranteed dead.
1548 In practice, the scrutinee is almost always a variable, so we pretty
1549 much always zap the OccInfo of the binders. It doesn't matter much though.
1554 Consider case (v `cast` co) of x { I# y ->
1555 ... (case (v `cast` co) of {...}) ...
1556 We'd like to eliminate the inner case. We can get this neatly by
1557 arranging that inside the outer case we add the unfolding
1558 v |-> x `cast` (sym co)
1559 to v. Then we should inline v at the inner case, cancel the casts, and away we go
1561 Note [Improving seq]
1564 type family F :: * -> *
1565 type instance F Int = Int
1567 ... case e of x { DEFAULT -> rhs } ...
1569 where x::F Int. Then we'd like to rewrite (F Int) to Int, getting
1571 case e `cast` co of x'::Int
1572 I# x# -> let x = x' `cast` sym co
1575 so that 'rhs' can take advantage of the form of x'.
1577 Notice that Note [Case of cast] may then apply to the result.
1579 Nota Bene: We only do the [Improving seq] transformation if the
1580 case binder 'x' is actually used in the rhs; that is, if the case
1581 is *not* a *pure* seq.
1582 a) There is no point in adding the cast to a pure seq.
1583 b) There is a good reason not to: doing so would interfere
1584 with seq rules (Note [Built-in RULES for seq] in MkId).
1585 In particular, this [Improving seq] thing *adds* a cast
1586 while [Built-in RULES for seq] *removes* one, so they
1589 You might worry about
1590 case v of x { __DEFAULT ->
1591 ... case (v `cast` co) of y { I# -> ... }}
1592 This is a pure seq (since x is unused), so [Improving seq] won't happen.
1593 But it's ok: the simplifier will replace 'v' by 'x' in the rhs to get
1594 case v of x { __DEFAULT ->
1595 ... case (x `cast` co) of y { I# -> ... }}
1596 Now the outer case is not a pure seq, so [Improving seq] will happen,
1597 and then the inner case will disappear.
1599 The need for [Improving seq] showed up in Roman's experiments. Example:
1600 foo :: F Int -> Int -> Int
1601 foo t n = t `seq` bar n
1604 bar n = bar (n - case t of TI i -> i)
1605 Here we'd like to avoid repeated evaluating t inside the loop, by
1606 taking advantage of the `seq`.
1608 At one point I did transformation in LiberateCase, but it's more
1609 robust here. (Otherwise, there's a danger that we'll simply drop the
1610 'seq' altogether, before LiberateCase gets to see it.)
1613 simplAlts :: SimplEnv
1615 -> InId -- Case binder
1616 -> [InAlt] -- Non-empty
1618 -> SimplM (OutExpr, OutId, [OutAlt]) -- Includes the continuation
1619 -- Like simplExpr, this just returns the simplified alternatives;
1620 -- it does not return an environment
1622 simplAlts env scrut case_bndr alts cont'
1623 = -- pprTrace "simplAlts" (ppr alts $$ ppr (seIdSubst env)) $
1624 do { let env0 = zapFloats env
1626 ; (env1, case_bndr1) <- simplBinder env0 case_bndr
1628 ; fam_envs <- getFamEnvs
1629 ; (alt_env', scrut', case_bndr') <- improveSeq fam_envs env1 scrut
1630 case_bndr case_bndr1 alts
1632 ; (imposs_deflt_cons, in_alts) <- prepareAlts scrut' case_bndr' alts
1634 ; alts' <- mapM (simplAlt alt_env' imposs_deflt_cons case_bndr' cont') in_alts
1635 ; return (scrut', case_bndr', alts') }
1638 ------------------------------------
1639 improveSeq :: (FamInstEnv, FamInstEnv) -> SimplEnv
1640 -> OutExpr -> InId -> OutId -> [InAlt]
1641 -> SimplM (SimplEnv, OutExpr, OutId)
1642 -- Note [Improving seq]
1643 improveSeq fam_envs env scrut case_bndr case_bndr1 [(DEFAULT,_,_)]
1644 | not (isDeadBinder case_bndr) -- Not a pure seq! See the Note!
1645 , Just (co, ty2) <- topNormaliseType fam_envs (idType case_bndr1)
1646 = do { case_bndr2 <- newId (fsLit "nt") ty2
1647 ; let rhs = DoneEx (Var case_bndr2 `Cast` mkSymCoercion co)
1648 env2 = extendIdSubst env case_bndr rhs
1649 ; return (env2, scrut `Cast` co, case_bndr2) }
1651 improveSeq _ env scrut _ case_bndr1 _
1652 = return (env, scrut, case_bndr1)
1655 ------------------------------------
1656 simplAlt :: SimplEnv
1657 -> [AltCon] -- These constructors can't be present when
1658 -- matching the DEFAULT alternative
1659 -> OutId -- The case binder
1664 simplAlt env imposs_deflt_cons case_bndr' cont' (DEFAULT, bndrs, rhs)
1665 = ASSERT( null bndrs )
1666 do { let env' = addBinderOtherCon env case_bndr' imposs_deflt_cons
1667 -- Record the constructors that the case-binder *can't* be.
1668 ; rhs' <- simplExprC env' rhs cont'
1669 ; return (DEFAULT, [], rhs') }
1671 simplAlt env _ case_bndr' cont' (LitAlt lit, bndrs, rhs)
1672 = ASSERT( null bndrs )
1673 do { let env' = addBinderUnfolding env case_bndr' (Lit lit)
1674 ; rhs' <- simplExprC env' rhs cont'
1675 ; return (LitAlt lit, [], rhs') }
1677 simplAlt env _ case_bndr' cont' (DataAlt con, vs, rhs)
1678 = do { -- Deal with the pattern-bound variables
1679 -- Mark the ones that are in ! positions in the
1680 -- data constructor as certainly-evaluated.
1681 -- NB: simplLamBinders preserves this eval info
1682 let vs_with_evals = add_evals (dataConRepStrictness con)
1683 ; (env', vs') <- simplLamBndrs env vs_with_evals
1685 -- Bind the case-binder to (con args)
1686 ; let inst_tys' = tyConAppArgs (idType case_bndr')
1687 con_args = map Type inst_tys' ++ varsToCoreExprs vs'
1688 env'' = addBinderUnfolding env' case_bndr'
1689 (mkConApp con con_args)
1691 ; rhs' <- simplExprC env'' rhs cont'
1692 ; return (DataAlt con, vs', rhs') }
1694 -- add_evals records the evaluated-ness of the bound variables of
1695 -- a case pattern. This is *important*. Consider
1696 -- data T = T !Int !Int
1698 -- case x of { T a b -> T (a+1) b }
1700 -- We really must record that b is already evaluated so that we don't
1701 -- go and re-evaluate it when constructing the result.
1702 -- See Note [Data-con worker strictness] in MkId.lhs
1707 go (v:vs') strs | isTyVar v = v : go vs' strs
1708 go (v:vs') (str:strs)
1709 | isMarkedStrict str = evald_v : go vs' strs
1710 | otherwise = zapped_v : go vs' strs
1712 zapped_v = zap_occ_info v
1713 evald_v = zapped_v `setIdUnfolding` evaldUnfolding
1714 go _ _ = pprPanic "cat_evals" (ppr con $$ ppr vs $$ ppr the_strs)
1716 -- See Note [zapOccInfo]
1717 -- zap_occ_info: if the case binder is alive, then we add the unfolding
1719 -- to the envt; so vs are now very much alive
1720 -- Note [Aug06] I can't see why this actually matters, but it's neater
1721 -- case e of t { (a,b) -> ...(case t of (p,q) -> p)... }
1722 -- ==> case e of t { (a,b) -> ...(a)... }
1723 -- Look, Ma, a is alive now.
1724 zap_occ_info = zapCasePatIdOcc case_bndr'
1726 addBinderUnfolding :: SimplEnv -> Id -> CoreExpr -> SimplEnv
1727 addBinderUnfolding env bndr rhs
1728 = modifyInScope env (bndr `setIdUnfolding` mkUnfolding False False rhs)
1730 addBinderOtherCon :: SimplEnv -> Id -> [AltCon] -> SimplEnv
1731 addBinderOtherCon env bndr cons
1732 = modifyInScope env (bndr `setIdUnfolding` mkOtherCon cons)
1734 zapCasePatIdOcc :: Id -> Id -> Id
1735 -- Consider case e of b { (a,b) -> ... }
1736 -- Then if we bind b to (a,b) in "...", and b is not dead,
1737 -- then we must zap the deadness info on a,b
1738 zapCasePatIdOcc case_bndr
1739 | isDeadBinder case_bndr = \ pat_id -> pat_id
1740 | otherwise = \ pat_id -> zapIdOccInfo pat_id
1744 %************************************************************************
1746 \subsection{Known constructor}
1748 %************************************************************************
1750 We are a bit careful with occurrence info. Here's an example
1752 (\x* -> case x of (a*, b) -> f a) (h v, e)
1754 where the * means "occurs once". This effectively becomes
1755 case (h v, e) of (a*, b) -> f a)
1757 let a* = h v; b = e in f a
1761 All this should happen in one sweep.
1764 knownCon :: SimplEnv
1765 -> OutExpr -- The scrutinee
1766 -> DataCon -> [OutType] -> [OutExpr] -- The scrutinee (in pieces)
1767 -> InId -> [InBndr] -> InExpr -- The alternative
1769 -> SimplM (SimplEnv, OutExpr)
1771 knownCon env scrut dc dc_ty_args dc_args bndr bs rhs cont
1772 = do { env' <- bind_args env bs dc_args
1774 -- It's useful to bind bndr to scrut, rather than to a fresh
1775 -- binding x = Con arg1 .. argn
1776 -- because very often the scrut is a variable, so we avoid
1777 -- creating, and then subsequently eliminating, a let-binding
1778 -- BUT, if scrut is a not a variable, we must be careful
1779 -- about duplicating the arg redexes; in that case, make
1780 -- a new con-app from the args
1781 bndr_rhs | exprIsTrivial scrut = scrut
1782 | otherwise = con_app
1783 con_app = Var (dataConWorkId dc)
1784 `mkTyApps` dc_ty_args
1785 `mkApps` [substExpr env' (varToCoreExpr b) | b <- bs]
1786 -- dc_ty_args are aready OutTypes, but bs are InBndrs
1788 ; env'' <- simplNonRecX env' bndr bndr_rhs
1789 ; simplExprF env'' rhs cont }
1791 zap_occ = zapCasePatIdOcc bndr -- bndr is an InId
1794 bind_args env' [] _ = return env'
1796 bind_args env' (b:bs') (Type ty : args)
1797 = ASSERT( isTyVar b )
1798 bind_args (extendTvSubst env' b ty) bs' args
1800 bind_args env' (b:bs') (arg : args)
1802 do { let b' = zap_occ b
1803 -- Note that the binder might be "dead", because it doesn't
1804 -- occur in the RHS; and simplNonRecX may therefore discard
1805 -- it via postInlineUnconditionally.
1806 -- Nevertheless we must keep it if the case-binder is alive,
1807 -- because it may be used in the con_app. See Note [zapOccInfo]
1808 ; env'' <- simplNonRecX env' b' arg
1809 ; bind_args env'' bs' args }
1812 pprPanic "bind_args" $ ppr dc $$ ppr bs $$ ppr dc_args $$
1813 text "scrut:" <+> ppr scrut
1816 missingAlt :: SimplEnv -> Id -> [InAlt] -> SimplCont -> SimplM (SimplEnv, OutExpr)
1817 -- This isn't strictly an error, although it is unusual.
1818 -- It's possible that the simplifer might "see" that
1819 -- an inner case has no accessible alternatives before
1820 -- it "sees" that the entire branch of an outer case is
1821 -- inaccessible. So we simply put an error case here instead.
1822 missingAlt env case_bndr alts cont
1823 = WARN( True, ptext (sLit "missingAlt") <+> ppr case_bndr )
1824 return (env, mkImpossibleExpr res_ty)
1826 res_ty = contResultType env (substTy env (coreAltsType alts)) cont
1830 %************************************************************************
1832 \subsection{Duplicating continuations}
1834 %************************************************************************
1837 prepareCaseCont :: SimplEnv
1838 -> [InAlt] -> SimplCont
1839 -> SimplM (SimplEnv, SimplCont,SimplCont)
1840 -- Return a duplicatable continuation, a non-duplicable part
1841 -- plus some extra bindings (that scope over the entire
1844 -- No need to make it duplicatable if there's only one alternative
1845 prepareCaseCont env [_] cont = return (env, cont, mkBoringStop)
1846 prepareCaseCont env _ cont = mkDupableCont env cont
1850 mkDupableCont :: SimplEnv -> SimplCont
1851 -> SimplM (SimplEnv, SimplCont, SimplCont)
1853 mkDupableCont env cont
1854 | contIsDupable cont
1855 = return (env, cont, mkBoringStop)
1857 mkDupableCont _ (Stop {}) = panic "mkDupableCont" -- Handled by previous eqn
1859 mkDupableCont env (CoerceIt ty cont)
1860 = do { (env', dup, nodup) <- mkDupableCont env cont
1861 ; return (env', CoerceIt ty dup, nodup) }
1863 mkDupableCont env cont@(StrictBind {})
1864 = return (env, mkBoringStop, cont)
1865 -- See Note [Duplicating StrictBind]
1867 mkDupableCont env (StrictArg info cci cont)
1868 -- See Note [Duplicating StrictArg]
1869 = do { (env', dup, nodup) <- mkDupableCont env cont
1870 ; (env'', args') <- mapAccumLM makeTrivial env' (ai_args info)
1871 ; return (env'', StrictArg (info { ai_args = args' }) cci dup, nodup) }
1873 mkDupableCont env (ApplyTo _ arg se cont)
1874 = -- e.g. [...hole...] (...arg...)
1876 -- let a = ...arg...
1877 -- in [...hole...] a
1878 do { (env', dup_cont, nodup_cont) <- mkDupableCont env cont
1879 ; arg' <- simplExpr (se `setInScope` env') arg
1880 ; (env'', arg'') <- makeTrivial env' arg'
1881 ; let app_cont = ApplyTo OkToDup arg'' (zapSubstEnv env'') dup_cont
1882 ; return (env'', app_cont, nodup_cont) }
1884 mkDupableCont env cont@(Select _ case_bndr [(_, bs, _rhs)] _ _)
1885 -- See Note [Single-alternative case]
1886 -- | not (exprIsDupable rhs && contIsDupable case_cont)
1887 -- | not (isDeadBinder case_bndr)
1888 | all isDeadBinder bs -- InIds
1889 && not (isUnLiftedType (idType case_bndr))
1890 -- Note [Single-alternative-unlifted]
1891 = return (env, mkBoringStop, cont)
1893 mkDupableCont env (Select _ case_bndr alts se cont)
1894 = -- e.g. (case [...hole...] of { pi -> ei })
1896 -- let ji = \xij -> ei
1897 -- in case [...hole...] of { pi -> ji xij }
1898 do { tick (CaseOfCase case_bndr)
1899 ; (env', dup_cont, nodup_cont) <- mkDupableCont env cont
1900 -- NB: call mkDupableCont here, *not* prepareCaseCont
1901 -- We must make a duplicable continuation, whereas prepareCaseCont
1902 -- doesn't when there is a single case branch
1904 ; let alt_env = se `setInScope` env'
1905 ; (alt_env', case_bndr') <- simplBinder alt_env case_bndr
1906 ; alts' <- mapM (simplAlt alt_env' [] case_bndr' dup_cont) alts
1907 -- Safe to say that there are no handled-cons for the DEFAULT case
1908 -- NB: simplBinder does not zap deadness occ-info, so
1909 -- a dead case_bndr' will still advertise its deadness
1910 -- This is really important because in
1911 -- case e of b { (# p,q #) -> ... }
1912 -- b is always dead, and indeed we are not allowed to bind b to (# p,q #),
1913 -- which might happen if e was an explicit unboxed pair and b wasn't marked dead.
1914 -- In the new alts we build, we have the new case binder, so it must retain
1916 -- NB: we don't use alt_env further; it has the substEnv for
1917 -- the alternatives, and we don't want that
1919 ; (env'', alts'') <- mkDupableAlts env' case_bndr' alts'
1920 ; return (env'', -- Note [Duplicated env]
1921 Select OkToDup case_bndr' alts'' (zapSubstEnv env'') mkBoringStop,
1925 mkDupableAlts :: SimplEnv -> OutId -> [InAlt]
1926 -> SimplM (SimplEnv, [InAlt])
1927 -- Absorbs the continuation into the new alternatives
1929 mkDupableAlts env case_bndr' the_alts
1932 go env0 [] = return (env0, [])
1934 = do { (env1, alt') <- mkDupableAlt env0 case_bndr' alt
1935 ; (env2, alts') <- go env1 alts
1936 ; return (env2, alt' : alts' ) }
1938 mkDupableAlt :: SimplEnv -> OutId -> (AltCon, [CoreBndr], CoreExpr)
1939 -> SimplM (SimplEnv, (AltCon, [CoreBndr], CoreExpr))
1940 mkDupableAlt env case_bndr (con, bndrs', rhs')
1941 | exprIsDupable rhs' -- Note [Small alternative rhs]
1942 = return (env, (con, bndrs', rhs'))
1944 = do { let rhs_ty' = exprType rhs'
1945 scrut_ty = idType case_bndr
1948 DEFAULT -> case_bndr
1949 DataAlt dc -> setIdUnfolding case_bndr unf
1951 -- See Note [Case binders and join points]
1952 unf = mkInlineRule needSaturated rhs 0
1953 rhs = mkConApp dc (map Type (tyConAppArgs scrut_ty)
1954 ++ varsToCoreExprs bndrs')
1956 LitAlt {} -> WARN( True, ptext (sLit "mkDupableAlt")
1957 <+> ppr case_bndr <+> ppr con )
1959 -- The case binder is alive but trivial, so why has
1960 -- it not been substituted away?
1962 used_bndrs' | isDeadBinder case_bndr = filter abstract_over bndrs'
1963 | otherwise = bndrs' ++ [case_bndr_w_unf]
1966 | isTyVar bndr = True -- Abstract over all type variables just in case
1967 | otherwise = not (isDeadBinder bndr)
1968 -- The deadness info on the new Ids is preserved by simplBinders
1970 ; (final_bndrs', final_args) -- Note [Join point abstraction]
1971 <- if (any isId used_bndrs')
1972 then return (used_bndrs', varsToCoreExprs used_bndrs')
1973 else do { rw_id <- newId (fsLit "w") realWorldStatePrimTy
1974 ; return ([rw_id], [Var realWorldPrimId]) }
1976 ; join_bndr <- newId (fsLit "$j") (mkPiTypes final_bndrs' rhs_ty')
1977 -- Note [Funky mkPiTypes]
1979 ; let -- We make the lambdas into one-shot-lambdas. The
1980 -- join point is sure to be applied at most once, and doing so
1981 -- prevents the body of the join point being floated out by
1982 -- the full laziness pass
1983 really_final_bndrs = map one_shot final_bndrs'
1984 one_shot v | isId v = setOneShotLambda v
1986 join_rhs = mkLams really_final_bndrs rhs'
1987 join_call = mkApps (Var join_bndr) final_args
1989 ; env' <- addPolyBind NotTopLevel env (NonRec join_bndr join_rhs)
1990 ; return (env', (con, bndrs', join_call)) }
1991 -- See Note [Duplicated env]
1994 Note [Case binders and join points]
1995 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
1997 case (case .. ) of c {
2000 If we make a join point with c but not c# we get
2001 $j = \c -> ....c....
2003 But if later inlining scrutines the c, thus
2005 $j = \c -> ... case c of { I# y -> ... } ...
2007 we won't see that 'c' has already been scrutinised. This actually
2008 happens in the 'tabulate' function in wave4main, and makes a significant
2009 difference to allocation.
2011 An alternative plan is this:
2013 $j = \c# -> let c = I# c# in ...c....
2015 but that is bad if 'c' is *not* later scrutinised.
2017 So instead we do both: we pass 'c' and 'c#' , and record in c's inlining
2018 that it's really I# c#, thus
2020 $j = \c# -> \c[=I# c#] -> ...c....
2022 Absence analysis may later discard 'c'.
2025 Note [Duplicated env]
2026 ~~~~~~~~~~~~~~~~~~~~~
2027 Some of the alternatives are simplified, but have not been turned into a join point
2028 So they *must* have an zapped subst-env. So we can't use completeNonRecX to
2029 bind the join point, because it might to do PostInlineUnconditionally, and
2030 we'd lose that when zapping the subst-env. We could have a per-alt subst-env,
2031 but zapping it (as we do in mkDupableCont, the Select case) is safe, and
2032 at worst delays the join-point inlining.
2034 Note [Small alternative rhs]
2035 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2036 It is worth checking for a small RHS because otherwise we
2037 get extra let bindings that may cause an extra iteration of the simplifier to
2038 inline back in place. Quite often the rhs is just a variable or constructor.
2039 The Ord instance of Maybe in PrelMaybe.lhs, for example, took several extra
2040 iterations because the version with the let bindings looked big, and so wasn't
2041 inlined, but after the join points had been inlined it looked smaller, and so
2044 NB: we have to check the size of rhs', not rhs.
2045 Duplicating a small InAlt might invalidate occurrence information
2046 However, if it *is* dupable, we return the *un* simplified alternative,
2047 because otherwise we'd need to pair it up with an empty subst-env....
2048 but we only have one env shared between all the alts.
2049 (Remember we must zap the subst-env before re-simplifying something).
2050 Rather than do this we simply agree to re-simplify the original (small) thing later.
2052 Note [Funky mkPiTypes]
2053 ~~~~~~~~~~~~~~~~~~~~~~
2054 Notice the funky mkPiTypes. If the contructor has existentials
2055 it's possible that the join point will be abstracted over
2056 type varaibles as well as term variables.
2057 Example: Suppose we have
2058 data T = forall t. C [t]
2060 case (case e of ...) of
2062 We get the join point
2063 let j :: forall t. [t] -> ...
2064 j = /\t \xs::[t] -> rhs
2066 case (case e of ...) of
2067 C t xs::[t] -> j t xs
2069 Note [Join point abstaction]
2070 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2071 If we try to lift a primitive-typed something out
2072 for let-binding-purposes, we will *caseify* it (!),
2073 with potentially-disastrous strictness results. So
2074 instead we turn it into a function: \v -> e
2075 where v::State# RealWorld#. The value passed to this function
2076 is realworld#, which generates (almost) no code.
2078 There's a slight infelicity here: we pass the overall
2079 case_bndr to all the join points if it's used in *any* RHS,
2080 because we don't know its usage in each RHS separately
2082 We used to say "&& isUnLiftedType rhs_ty'" here, but now
2083 we make the join point into a function whenever used_bndrs'
2084 is empty. This makes the join-point more CPR friendly.
2085 Consider: let j = if .. then I# 3 else I# 4
2086 in case .. of { A -> j; B -> j; C -> ... }
2088 Now CPR doesn't w/w j because it's a thunk, so
2089 that means that the enclosing function can't w/w either,
2090 which is a lose. Here's the example that happened in practice:
2091 kgmod :: Int -> Int -> Int
2092 kgmod x y = if x > 0 && y < 0 || x < 0 && y > 0
2096 I have seen a case alternative like this:
2098 It's a bit silly to add the realWorld dummy arg in this case, making
2101 (the \v alone is enough to make CPR happy) but I think it's rare
2103 Note [Duplicating StrictArg]
2104 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2105 The original plan had (where E is a big argument)
2107 ==> let $j = \a -> f E a
2110 But this is terrible! Here's an example:
2111 && E (case x of { T -> F; F -> T })
2112 Now, && is strict so we end up simplifying the case with
2113 an ArgOf continuation. If we let-bind it, we get
2114 let $j = \v -> && E v
2115 in simplExpr (case x of { T -> F; F -> T })
2117 And after simplifying more we get
2118 let $j = \v -> && E v
2119 in case x of { T -> $j F; F -> $j T }
2120 Which is a Very Bad Thing
2122 What we do now is this
2126 Now if the thing in the hole is a case expression (which is when
2127 we'll call mkDupableCont), we'll push the function call into the
2128 branches, which is what we want. Now RULES for f may fire, and
2129 call-pattern specialisation. Here's an example from Trac #3116
2132 _ -> Chunk p fpc (o+1) (l-1) bs')
2133 If we can push the call for 'go' inside the case, we get
2134 call-pattern specialisation for 'go', which is *crucial* for
2137 Here is the (&&) example:
2138 && E (case x of { T -> F; F -> T })
2140 case x of { T -> && a F; F -> && a T }
2144 * Arguments to f *after* the strict one are handled by
2145 the ApplyTo case of mkDupableCont. Eg
2148 * We can only do the let-binding of E because the function
2149 part of a StrictArg continuation is an explicit syntax
2150 tree. In earlier versions we represented it as a function
2151 (CoreExpr -> CoreEpxr) which we couldn't take apart.
2153 Do *not* duplicate StrictBind and StritArg continuations. We gain
2154 nothing by propagating them into the expressions, and we do lose a
2157 The desire not to duplicate is the entire reason that
2158 mkDupableCont returns a pair of continuations.
2160 Note [Duplicating StrictBind]
2161 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2162 Unlike StrictArg, there doesn't seem anything to gain from
2163 duplicating a StrictBind continuation, so we don't.
2165 The desire not to duplicate is the entire reason that
2166 mkDupableCont returns a pair of continuations.
2169 Note [Single-alternative cases]
2170 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2171 This case is just like the ArgOf case. Here's an example:
2175 case (case x of I# x' ->
2177 True -> I# (negate# x')
2178 False -> I# x') of y {
2180 Because the (case x) has only one alternative, we'll transform to
2182 case (case x' <# 0# of
2183 True -> I# (negate# x')
2184 False -> I# x') of y {
2186 But now we do *NOT* want to make a join point etc, giving
2188 let $j = \y -> MkT y
2190 True -> $j (I# (negate# x'))
2192 In this case the $j will inline again, but suppose there was a big
2193 strict computation enclosing the orginal call to MkT. Then, it won't
2194 "see" the MkT any more, because it's big and won't get duplicated.
2195 And, what is worse, nothing was gained by the case-of-case transform.
2197 When should use this case of mkDupableCont?
2198 However, matching on *any* single-alternative case is a *disaster*;
2199 e.g. case (case ....) of (a,b) -> (# a,b #)
2200 We must push the outer case into the inner one!
2203 * Match [(DEFAULT,_,_)], but in the common case of Int,
2204 the alternative-filling-in code turned the outer case into
2205 case (...) of y { I# _ -> MkT y }
2207 * Match on single alternative plus (not (isDeadBinder case_bndr))
2208 Rationale: pushing the case inwards won't eliminate the construction.
2209 But there's a risk of
2210 case (...) of y { (a,b) -> let z=(a,b) in ... }
2211 Now y looks dead, but it'll come alive again. Still, this
2212 seems like the best option at the moment.
2214 * Match on single alternative plus (all (isDeadBinder bndrs))
2215 Rationale: this is essentially seq.
2217 * Match when the rhs is *not* duplicable, and hence would lead to a
2218 join point. This catches the disaster-case above. We can test
2219 the *un-simplified* rhs, which is fine. It might get bigger or
2220 smaller after simplification; if it gets smaller, this case might
2221 fire next time round. NB also that we must test contIsDupable
2222 case_cont *btoo, because case_cont might be big!
2224 HOWEVER: I found that this version doesn't work well, because
2225 we can get let x = case (...) of { small } in ...case x...
2226 When x is inlined into its full context, we find that it was a bad
2227 idea to have pushed the outer case inside the (...) case.
2229 Note [Single-alternative-unlifted]
2230 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2231 Here's another single-alternative where we really want to do case-of-case:
2239 case y_s6X of tpl_s7m {
2240 M1.Mk1 ipv_s70 -> ipv_s70;
2241 M1.Mk2 ipv_s72 -> ipv_s72;
2247 case x_s74 of tpl_s7n {
2248 M1.Mk1 ipv_s77 -> ipv_s77;
2249 M1.Mk2 ipv_s79 -> ipv_s79;
2253 { __DEFAULT -> ==# [wild1_s7b wild_s7c];
2257 So the outer case is doing *nothing at all*, other than serving as a
2258 join-point. In this case we really want to do case-of-case and decide
2259 whether to use a real join point or just duplicate the continuation.
2261 Hence: check whether the case binder's type is unlifted, because then
2262 the outer case is *not* a seq.