2 % (c) The GRASP/AQUA Project, Glasgow University, 1992-1998
4 \section{Tidying up Core}
7 module TidyPgm( mkBootModDetailsDs, mkBootModDetailsTc,
8 tidyProgram, globaliseAndTidyId ) where
10 #include "HsVersions.h"
22 import CoreArity ( exprArity )
23 import Class ( classSelIds )
45 import FastBool hiding ( fastOr )
48 import Data.List ( sortBy )
49 import Data.IORef ( IORef, readIORef, writeIORef )
53 Constructing the TypeEnv, Instances, Rules from which the ModIface is
54 constructed, and which goes on to subsequent modules in --make mode.
56 Most of the interface file is obtained simply by serialising the
57 TypeEnv. One important consequence is that if the *interface file*
58 has pragma info if and only if the final TypeEnv does. This is not so
59 important for *this* module, but it's essential for ghc --make:
60 subsequent compilations must not see (e.g.) the arity if the interface
61 file does not contain arity If they do, they'll exploit the arity;
62 then the arity might change, but the iface file doesn't change =>
63 recompilation does not happen => disaster.
65 For data types, the final TypeEnv will have a TyThing for the TyCon,
66 plus one for each DataCon; the interface file will contain just one
67 data type declaration, but it is de-serialised back into a collection
70 %************************************************************************
74 %************************************************************************
77 Plan A: mkBootModDetails: omit pragmas, make interfaces small
78 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
81 * Drop all WiredIn things from the TypeEnv
82 (we never want them in interface files)
84 * Retain all TyCons and Classes in the TypeEnv, to avoid
85 having to find which ones are mentioned in the
88 * Trim off the constructors of non-exported TyCons, both
89 from the TyCon and from the TypeEnv
91 * Drop non-exported Ids from the TypeEnv
93 * Tidy the types of the DFunIds of Instances,
94 make them into GlobalIds, (they already have External Names)
95 and add them to the TypeEnv
97 * Tidy the types of the (exported) Ids in the TypeEnv,
98 make them into GlobalIds (they already have External Names)
100 * Drop rules altogether
102 * Tidy the bindings, to ensure that the Caf and Arity
103 information is correct for each top-level binder; the
104 code generator needs it. And to ensure that local names have
105 distinct OccNames in case of object-file splitting
108 -- This is Plan A: make a small type env when typechecking only,
109 -- or when compiling a hs-boot file, or simply when not using -O
111 -- We don't look at the bindings at all -- there aren't any
114 mkBootModDetailsTc :: HscEnv -> TcGblEnv -> IO ModDetails
115 mkBootModDetailsTc hsc_env
116 TcGblEnv{ tcg_exports = exports,
117 tcg_type_env = type_env,
119 tcg_fam_insts = fam_insts
121 = mkBootModDetails hsc_env exports type_env insts fam_insts
123 mkBootModDetailsDs :: HscEnv -> ModGuts -> IO ModDetails
124 mkBootModDetailsDs hsc_env
125 ModGuts{ mg_exports = exports,
128 mg_fam_insts = fam_insts
130 = mkBootModDetails hsc_env exports type_env insts fam_insts
132 mkBootModDetails :: HscEnv -> [AvailInfo] -> NameEnv TyThing
133 -> [Instance] -> [FamInstEnv.FamInst] -> IO ModDetails
134 mkBootModDetails hsc_env exports type_env insts fam_insts
135 = do { let dflags = hsc_dflags hsc_env
136 ; showPass dflags "Tidy [hoot] type env"
138 ; let { insts' = tidyInstances globaliseAndTidyId insts
139 ; dfun_ids = map instanceDFunId insts'
140 ; type_env1 = tidyBootTypeEnv (availsToNameSet exports) type_env
141 ; type_env' = extendTypeEnvWithIds type_env1 dfun_ids
143 ; return (ModDetails { md_types = type_env'
145 , md_fam_insts = fam_insts
148 , md_exports = exports
149 , md_vect_info = noVectInfo
154 tidyBootTypeEnv :: NameSet -> TypeEnv -> TypeEnv
155 tidyBootTypeEnv exports type_env
156 = tidyTypeEnv True False exports type_env final_ids
158 -- Find the LocalIds in the type env that are exported
159 -- Make them into GlobalIds, and tidy their types
161 -- It's very important to remove the non-exported ones
162 -- because we don't tidy the OccNames, and if we don't remove
163 -- the non-exported ones we'll get many things with the
164 -- same name in the interface file, giving chaos.
165 final_ids = [ globaliseAndTidyId id
166 | id <- typeEnvIds type_env
170 -- default methods have their export flag set, but everything
171 -- else doesn't (yet), because this is pre-desugaring, so we
173 keep_it id = isExportedId id || idName id `elemNameSet` exports
177 globaliseAndTidyId :: Id -> Id
178 -- Takes an LocalId with an External Name,
179 -- makes it into a GlobalId
180 -- * unchanged Name (might be Internal or External)
181 -- * unchanged details
182 -- * VanillaIdInfo (makes a conservative assumption about Caf-hood)
183 globaliseAndTidyId id
184 = Id.setIdType (globaliseId id) tidy_type
186 tidy_type = tidyTopType (idType id)
190 %************************************************************************
192 Plan B: tidy bindings, make TypeEnv full of IdInfo
194 %************************************************************************
196 Plan B: include pragmas, make interfaces
197 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
198 * Figure out which Ids are externally visible
200 * Tidy the bindings, externalising appropriate Ids
202 * Drop all Ids from the TypeEnv, and add all the External Ids from
203 the bindings. (This adds their IdInfo to the TypeEnv; and adds
204 floated-out Ids that weren't even in the TypeEnv before.)
206 Step 1: Figure out external Ids
207 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
208 Note [choosing external names]
210 See also the section "Interface stability" in the
211 RecompilationAvoidance commentary:
212 http://hackage.haskell.org/trac/ghc/wiki/Commentary/Compiler/RecompilationAvoidance
214 First we figure out which Ids are "external" Ids. An
215 "external" Id is one that is visible from outside the compilation
217 a) the user exported ones
218 b) ones mentioned in the unfoldings, workers,
219 or rules of externally-visible ones
221 While figuring out which Ids are external, we pick a "tidy" OccName
222 for each one. That is, we make its OccName distinct from the other
223 external OccNames in this module, so that in interface files and
224 object code we can refer to it unambiguously by its OccName. The
225 OccName for each binder is prefixed by the name of the exported Id
226 that references it; e.g. if "f" references "x" in its unfolding, then
227 "x" is renamed to "f_x". This helps distinguish the different "x"s
228 from each other, and means that if "f" is later removed, things that
229 depend on the other "x"s will not need to be recompiled. Of course,
230 if there are multiple "f_x"s, then we have to disambiguate somehow; we
231 use "f_x0", "f_x1" etc.
233 As far as possible we should assign names in a deterministic fashion.
234 Each time this module is compiled with the same options, we should end
235 up with the same set of external names with the same types. That is,
236 the ABI hash in the interface should not change. This turns out to be
237 quite tricky, since the order of the bindings going into the tidy
238 phase is already non-deterministic, as it is based on the ordering of
239 Uniques, which are assigned unpredictably.
241 To name things in a stable way, we do a depth-first-search of the
242 bindings, starting from the exports sorted by name. This way, as long
243 as the bindings themselves are deterministic (they sometimes aren't!),
244 the order in which they are presented to the tidying phase does not
245 affect the names we assign.
247 Step 2: Tidy the program
248 ~~~~~~~~~~~~~~~~~~~~~~~~
249 Next we traverse the bindings top to bottom. For each *top-level*
252 1. Make it into a GlobalId; its IdDetails becomes VanillaGlobal,
253 reflecting the fact that from now on we regard it as a global,
256 2. Give it a system-wide Unique.
257 [Even non-exported things need system-wide Uniques because the
258 byte-code generator builds a single Name->BCO symbol table.]
260 We use the NameCache kept in the HscEnv as the
261 source of such system-wide uniques.
263 For external Ids, use the original-name cache in the NameCache
264 to ensure that the unique assigned is the same as the Id had
265 in any previous compilation run.
267 3. Rename top-level Ids according to the names we chose in step 1.
268 If it's an external Id, make it have a External Name, otherwise
269 make it have an Internal Name. This is used by the code generator
270 to decide whether to make the label externally visible
272 4. Give it its UTTERLY FINAL IdInfo; in ptic,
273 * its unfolding, if it should have one
275 * its arity, computed from the number of visible lambdas
277 * its CAF info, computed from what is free in its RHS
280 Finally, substitute these new top-level binders consistently
281 throughout, including in unfoldings. We also tidy binders in
282 RHSs, so that they print nicely in interfaces.
285 tidyProgram :: HscEnv -> ModGuts -> IO (CgGuts, ModDetails)
286 tidyProgram hsc_env (ModGuts { mg_module = mod, mg_exports = exports,
288 mg_insts = insts, mg_fam_insts = fam_insts,
290 mg_rules = imp_rules,
291 mg_vect_info = vect_info,
292 mg_dir_imps = dir_imps,
295 mg_foreign = foreign_stubs,
296 mg_hpc_info = hpc_info,
297 mg_modBreaks = modBreaks })
299 = do { let dflags = hsc_dflags hsc_env
300 ; showPass dflags "Tidy Core"
302 ; let { omit_prags = dopt Opt_OmitInterfacePragmas dflags
303 ; th = dopt Opt_TemplateHaskell dflags
306 ; (unfold_env, tidy_occ_env)
307 <- chooseExternalIds hsc_env mod omit_prags binds
311 | otherwise = findExternalRules binds imp_rules unfold_env
312 -- findExternalRules filters imp_rules to avoid binders that
313 -- aren't externally visible; but the externally-visible binders
314 -- are computed (by findExternalIds) assuming that all orphan
315 -- rules are exported (they get their Exported flag set in the desugarer)
316 -- So in fact we may export more than we need.
317 -- (It's a sort of mutual recursion.)
320 ; let { (tidy_env, tidy_binds)
321 = tidyTopBinds hsc_env unfold_env tidy_occ_env binds }
323 ; let { export_set = availsToNameSet exports
324 ; final_ids = [ id | id <- bindersOfBinds tidy_binds,
325 isExternalName (idName id)]
326 ; tidy_type_env = tidyTypeEnv omit_prags th export_set
328 ; tidy_insts = tidyInstances (lookup_dfun tidy_type_env) insts
329 -- A DFunId will have a binding in tidy_binds, and so
330 -- will now be in final_env, replete with IdInfo
331 -- Its name will be unchanged since it was born, but
332 -- we want Global, IdInfo-rich (or not) DFunId in the
335 ; tidy_rules = tidyRules tidy_env ext_rules
336 -- You might worry that the tidy_env contains IdInfo-rich stuff
337 -- and indeed it does, but if omit_prags is on, ext_rules is
340 -- See Note [Injecting implicit bindings]
341 ; implicit_binds = getImplicitBinds type_env
342 ; all_tidy_binds = implicit_binds ++ tidy_binds
344 ; alg_tycons = filter isAlgTyCon (typeEnvTyCons type_env)
347 ; endPass dflags "Tidy Core" Opt_D_dump_simpl all_tidy_binds
348 ; dumpIfSet_core dflags Opt_D_dump_simpl
350 (pprRules tidy_rules)
352 ; let dir_imp_mods = moduleEnvKeys dir_imps
354 ; return (CgGuts { cg_module = mod,
355 cg_tycons = alg_tycons,
356 cg_binds = all_tidy_binds,
357 cg_dir_imps = dir_imp_mods,
358 cg_foreign = foreign_stubs,
359 cg_dep_pkgs = dep_pkgs deps,
360 cg_hpc_info = hpc_info,
361 cg_modBreaks = modBreaks },
363 ModDetails { md_types = tidy_type_env,
364 md_rules = tidy_rules,
365 md_insts = tidy_insts,
366 md_fam_insts = fam_insts,
367 md_exports = exports,
368 md_anns = anns, -- are already tidy
369 md_vect_info = vect_info --
373 lookup_dfun :: TypeEnv -> Var -> Id
374 lookup_dfun type_env dfun_id
375 = case lookupTypeEnv type_env (idName dfun_id) of
376 Just (AnId dfun_id') -> dfun_id'
377 _other -> pprPanic "lookup_dfun" (ppr dfun_id)
379 --------------------------
380 tidyTypeEnv :: Bool -- Compiling without -O, so omit prags
381 -> Bool -- Template Haskell is on
382 -> NameSet -> TypeEnv -> [Id] -> TypeEnv
384 -- The competed type environment is gotten from
385 -- Dropping any wired-in things, and then
386 -- a) keeping the types and classes
387 -- b) removing all Ids,
388 -- c) adding Ids with correct IdInfo, including unfoldings,
389 -- gotten from the bindings
390 -- From (c) we keep only those Ids with External names;
391 -- the CoreTidy pass makes sure these are all and only
392 -- the externally-accessible ones
393 -- This truncates the type environment to include only the
394 -- exported Ids and things needed from them, which saves space
396 tidyTypeEnv omit_prags th exports type_env final_ids
397 = let type_env1 = filterNameEnv keep_it type_env
398 type_env2 = extendTypeEnvWithIds type_env1 final_ids
399 type_env3 | omit_prags = mapNameEnv (trimThing th exports) type_env2
400 | otherwise = type_env2
404 -- We keep GlobalIds, because they won't appear
405 -- in the bindings from which final_ids are derived!
406 -- (The bindings bind LocalIds.)
407 keep_it thing | isWiredInThing thing = False
408 keep_it (AnId id) = isGlobalId id -- Keep GlobalIds (e.g. class ops)
409 keep_it _other = True -- Keep all TyCons, DataCons, and Classes
411 --------------------------
412 isWiredInThing :: TyThing -> Bool
413 isWiredInThing thing = isWiredInName (getName thing)
415 --------------------------
416 trimThing :: Bool -> NameSet -> TyThing -> TyThing
417 -- Trim off inessentials, for boot files and no -O
418 trimThing th exports (ATyCon tc)
419 | not th && not (mustExposeTyCon exports tc)
420 = ATyCon (makeTyConAbstract tc) -- Note [Trimming and Template Haskell]
422 trimThing _th _exports (AnId id)
423 | not (isImplicitId id)
424 = AnId (id `setIdInfo` vanillaIdInfo)
426 trimThing _th _exports other_thing
430 {- Note [Trimming and Template Haskell]
431 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
432 Consider (Trac #2386) this
433 module M(T, makeOne) where
435 makeOne = [| Yay "Yep" |]
436 Notice that T is exported abstractly, but makeOne effectively exports it too!
437 A module that splices in $(makeOne) will then look for a declartion of Yay,
438 so it'd better be there. Hence, brutally but simply, we switch off type
439 constructor trimming if TH is enabled in this module. -}
442 mustExposeTyCon :: NameSet -- Exports
443 -> TyCon -- The tycon
444 -> Bool -- Can its rep be hidden?
445 -- We are compiling without -O, and thus trying to write as little as
446 -- possible into the interface file. But we must expose the details of
447 -- any data types whose constructors or fields are exported
448 mustExposeTyCon exports tc
449 | not (isAlgTyCon tc) -- Synonyms
451 | isEnumerationTyCon tc -- For an enumeration, exposing the constructors
452 = True -- won't lead to the need for further exposure
453 -- (This includes data types with no constructors.)
454 | isOpenTyCon tc -- Open type family
457 | otherwise -- Newtype, datatype
458 = any exported_con (tyConDataCons tc)
459 -- Expose rep if any datacon or field is exported
461 || (isNewTyCon tc && isFFITy (snd (newTyConRhs tc)))
462 -- Expose the rep for newtypes if the rep is an FFI type.
463 -- For a very annoying reason. 'Foreign import' is meant to
464 -- be able to look through newtypes transparently, but it
465 -- can only do that if it can "see" the newtype representation
467 exported_con con = any (`elemNameSet` exports)
468 (dataConName con : dataConFieldLabels con)
470 tidyInstances :: (DFunId -> DFunId) -> [Instance] -> [Instance]
471 tidyInstances tidy_dfun ispecs
474 tidy ispec = setInstanceDFunId ispec $
475 tidy_dfun (instanceDFunId ispec)
479 %************************************************************************
483 %************************************************************************
485 Note [Injecting implicit bindings]
486 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
487 We inject the implict bindings right at the end, in CoreTidy.
488 Some of these bindings, notably record selectors, are not
489 constructed in an optimised form. E.g. record selector for
490 data T = MkT { x :: {-# UNPACK #-} !Int }
491 Then the unfolding looks like
492 x = \t. case t of MkT x1 -> let x = I# x1 in x
493 This generates bad code unless it's first simplified a bit. That is
494 why CoreUnfold.mkImplicitUnfolding uses simleExprOpt to do a bit of
495 optimisation first. (Only matters when the selector is used curried;
496 eg map x ys.) See Trac #2070.
498 At one time I tried injecting the implicit bindings *early*, at the
499 beginning of SimplCore. But that gave rise to real difficulty,
500 becuase GlobalIds are supposed to have *fixed* IdInfo, but the
501 simplifier and other core-to-core passes mess with IdInfo all the
502 time. The straw that broke the camels back was when a class selector
503 got the wrong arity -- ie the simplifier gave it arity 2, whereas
504 importing modules were expecting it to have arity 1 (Trac #2844).
505 It's much safer just to inject them right at the end, after tidying.
507 Oh: two other reasons for injecting them late:
508 - If implicit Ids are already in the bindings when we start TidyPgm,
509 we'd have to be careful not to treat them as external Ids (in
510 the sense of findExternalIds); else the Ids mentioned in *their*
511 RHSs will be treated as external and you get an interface file
513 but nothing refererring to a18 (because the implicit Id is the
516 - More seriously, the tidied type-envt will include the implicit
517 Id replete with a18 in its unfolding; but we won't take account
518 of a18 when computing a fingerprint for the class; result chaos.
522 getImplicitBinds :: TypeEnv -> [CoreBind]
523 getImplicitBinds type_env
524 = map get_defn (concatMap implicit_ids (typeEnvElts type_env))
526 implicit_ids (ATyCon tc) = mapCatMaybes dataConWrapId_maybe (tyConDataCons tc)
527 implicit_ids (AClass cls) = classSelIds cls
530 get_defn :: Id -> CoreBind
531 get_defn id = NonRec id (unfoldingTemplate (idUnfolding id))
535 %************************************************************************
537 \subsection{Step 1: finding externals}
539 %************************************************************************
541 Sete Note [choosing external names].
544 type UnfoldEnv = IdEnv (Name{-new name-}, Bool {-show unfolding-})
545 -- maps each top-level Id to its new Name (the Id is tidied in step 2)
546 -- The Unique is unchanged. If the new Id is external, it will be
547 -- visible in the interface file.
549 -- Bool => expose unfolding or not.
551 chooseExternalIds :: HscEnv
555 -> IO (UnfoldEnv, TidyOccEnv)
556 -- Step 1 from the notes above
558 chooseExternalIds hsc_env mod omit_prags binds
560 (unfold_env1,occ_env1)
561 <- search (zip sorted_exports sorted_exports) emptyVarEnv init_occ_env
562 let internal_ids = filter (not . (`elemVarEnv` unfold_env1)) binders
563 tidy_internal internal_ids unfold_env1 occ_env1
565 nc_var = hsc_NC hsc_env
567 -- the exports, sorted by OccName. This is a deterministic list of
568 -- Ids (i.e. it's the same list every time this module is compiled),
569 -- in contrast to the bindings, which are ordered
570 -- non-deterministically.
572 -- This list will serve as a starting point for finding a
573 -- deterministic, tidy, renaming for all external Ids in this
575 sorted_exports = sortBy (compare `on` getOccName) $
576 filter isExportedId binders
578 binders = bindersOfBinds binds
580 bind_env :: IdEnv CoreExpr
581 bind_env = mkVarEnv (flattenBinds binds)
583 avoids = [getOccName name | bndr <- binders,
584 let name = idName bndr,
585 isExternalName name ]
586 -- In computing our "avoids" list, we must include
588 -- all things with global names (assigned once and for
589 -- all by the renamer)
590 -- since their names are "taken".
591 -- The type environment is a convenient source of such things.
593 -- We also make sure to avoid any exported binders. Consider
594 -- f{-u1-} = 1 -- Local decl
596 -- f{-u2-} = 2 -- Exported decl
598 -- The second exported decl must 'get' the name 'f', so we
599 -- have to put 'f' in the avoids list before we get to the first
600 -- decl. tidyTopId then does a no-op on exported binders.
601 init_occ_env = initTidyOccEnv avoids
604 search :: [(Id,Id)] -- (external id, referrring id)
605 -> UnfoldEnv -- id -> (new Name, show_unfold)
606 -> TidyOccEnv -- occ env for choosing new Names
607 -> IO (UnfoldEnv, TidyOccEnv)
609 search [] unfold_env occ_env = return (unfold_env, occ_env)
611 search ((id,referrer) : rest) unfold_env occ_env
612 | id `elemVarEnv` unfold_env = search rest unfold_env occ_env
614 (occ_env', name') <- tidyTopName mod nc_var (Just referrer) occ_env id
616 rhs = expectJust (showSDoc (text "chooseExternalIds: " <> ppr id)) $ lookupVarEnv bind_env id
617 (new_ids, show_unfold)
618 | omit_prags = ([], False)
619 | otherwise = addExternal id rhs
620 unfold_env' = extendVarEnv unfold_env id (name',show_unfold)
621 referrer' | isExportedId id = id
622 | otherwise = referrer
624 search (zip new_ids (repeat referrer') ++ rest) unfold_env' occ_env'
626 tidy_internal :: [Id] -> UnfoldEnv -> TidyOccEnv
627 -> IO (UnfoldEnv, TidyOccEnv)
628 tidy_internal [] unfold_env occ_env = return (unfold_env,occ_env)
629 tidy_internal (id:ids) unfold_env occ_env = do
630 (occ_env', name') <- tidyTopName mod nc_var Nothing occ_env id
631 let unfold_env' = extendVarEnv unfold_env id (name',False)
632 tidy_internal ids unfold_env' occ_env'
634 addExternal :: Id -> CoreExpr -> ([Id],Bool)
635 addExternal id rhs = (new_needed_ids, show_unfold)
637 new_needed_ids = unfold_ids ++
638 filter (\id -> isLocalId id &&
639 not (id `elemVarSet` unfold_set))
640 (varSetElems worker_ids ++
641 varSetElems spec_ids) -- XXX non-det ordering
644 dont_inline = isNeverActive (inlinePragmaActivation (inlinePragInfo idinfo))
645 loop_breaker = isNonRuleLoopBreaker (occInfo idinfo)
646 bottoming_fn = isBottomingSig (newStrictnessInfo idinfo `orElse` topSig)
647 spec_ids = specInfoFreeVars (specInfo idinfo)
648 worker_info = workerInfo idinfo
650 -- Stuff to do with the Id's unfolding
651 -- The simplifier has put an up-to-date unfolding
652 -- in the IdInfo, but the RHS will do just as well
653 unfolding = unfoldingInfo idinfo
654 rhs_is_small = not (neverUnfold unfolding)
656 -- We leave the unfolding there even if there is a worker
657 -- In GHCI the unfolding is used by importers
658 -- When writing an interface file, we omit the unfolding
659 -- if there is a worker
660 show_unfold = not bottoming_fn && -- Not necessary
663 rhs_is_small -- Small enough
665 (unfold_set, unfold_ids)
666 | show_unfold = freeVarsInDepthFirstOrder rhs
667 | otherwise = (emptyVarSet, [])
669 worker_ids = case worker_info of
670 HasWorker work_id _ -> unitVarSet work_id
671 _otherwise -> emptyVarSet
674 -- We want a deterministic free-variable list. exprFreeVars gives us
675 -- a VarSet, which is in a non-deterministic order when converted to a
676 -- list. Hence, here we define a free-variable finder that returns
677 -- the free variables in the order that they are encountered.
679 -- Note [choosing external names]
681 freeVarsInDepthFirstOrder :: CoreExpr -> (VarSet, [Id])
682 freeVarsInDepthFirstOrder e =
684 DFFV m -> case m emptyVarSet [] of
685 (set,ids,_) -> (set,ids)
687 newtype DFFV a = DFFV (VarSet -> [Var] -> (VarSet,[Var],a))
689 instance Monad DFFV where
690 return a = DFFV $ \set ids -> (set, ids, a)
691 (DFFV m) >>= k = DFFV $ \set ids ->
693 (set',ids',a) -> case k a of
694 DFFV f -> f set' ids'
696 insert :: Var -> DFFV ()
697 insert v = DFFV $ \ set ids -> case () of
698 _ | v `elemVarSet` set -> (set,ids,())
699 | otherwise -> (extendVarSet set v, v:ids, ())
701 dffvExpr :: CoreExpr -> DFFV ()
702 dffvExpr e = go emptyVarSet e
704 go scope e = case e of
705 Var v | isLocalId v && not (v `elemVarSet` scope) -> insert v
706 App e1 e2 -> do go scope e1; go scope e2
707 Lam v e -> go (extendVarSet scope v) e
708 Note _ e -> go scope e
709 Cast e _ -> go scope e
710 Let (NonRec x r) e -> do go scope r; go (extendVarSet scope x) e
711 Let (Rec prs) e -> do let scope' = extendVarSetList scope (map fst prs)
712 mapM_ (go scope') (map snd prs)
714 Case e b _ as -> do go scope e
715 mapM_ (go_alt (extendVarSet scope b)) as
718 go_alt scope (_,xs,r) = go (extendVarSetList scope xs) r
722 --------------------------------------------------------------------
724 -- This is where we set names to local/global based on whether they really are
725 -- externally visible (see comment at the top of this module). If the name
726 -- was previously local, we have to give it a unique occurrence name if
727 -- we intend to externalise it.
730 tidyTopName :: Module -> IORef NameCache -> Maybe Id -> TidyOccEnv
731 -> Id -> IO (TidyOccEnv, Name)
732 tidyTopName mod nc_var maybe_ref occ_env id
733 | global && internal = return (occ_env, localiseName name)
735 | global && external = return (occ_env, name)
736 -- Global names are assumed to have been allocated by the renamer,
737 -- so they already have the "right" unique
738 -- And it's a system-wide unique too
740 -- Now we get to the real reason that all this is in the IO Monad:
741 -- we have to update the name cache in a nice atomic fashion
743 | local && internal = do { nc <- readIORef nc_var
744 ; let (nc', new_local_name) = mk_new_local nc
745 ; writeIORef nc_var nc'
746 ; return (occ_env', new_local_name) }
747 -- Even local, internal names must get a unique occurrence, because
748 -- if we do -split-objs we externalise the name later, in the code generator
750 -- Similarly, we must make sure it has a system-wide Unique, because
751 -- the byte-code generator builds a system-wide Name->BCO symbol table
753 | local && external = do { nc <- readIORef nc_var
754 ; let (nc', new_external_name) = mk_new_external nc
755 ; writeIORef nc_var nc'
756 ; return (occ_env', new_external_name) }
758 | otherwise = panic "tidyTopName"
761 external = isJust maybe_ref
762 global = isExternalName name
764 internal = not external
765 loc = nameSrcSpan name
767 old_occ = nameOccName name
769 | Just ref <- maybe_ref, ref /= id =
770 mkOccName (occNameSpace old_occ) $
772 ref_str = occNameString (getOccName ref)
773 occ_str = occNameString old_occ
777 -- workers: the worker for a function already
778 -- includes the occname for its parent, so there's
779 -- no need to prepend the referrer.
780 _other | isSystemName name -> ref_str
781 | otherwise -> ref_str ++ '_' : occ_str
782 -- If this name was system-generated, then don't bother
783 -- to retain its OccName, just use the referrer. These
784 -- system-generated names will become "f1", "f2", etc. for
786 | otherwise = old_occ
788 (occ_env', occ') = tidyOccName occ_env new_occ
790 mk_new_local nc = (nc { nsUniqs = us2 }, mkInternalName uniq occ' loc)
792 (us1, us2) = splitUniqSupply (nsUniqs nc)
793 uniq = uniqFromSupply us1
795 mk_new_external nc = allocateGlobalBinder nc mod occ' loc
796 -- If we want to externalise a currently-local name, check
797 -- whether we have already assigned a unique for it.
798 -- If so, use it; if not, extend the table.
799 -- All this is done by allcoateGlobalBinder.
800 -- This is needed when *re*-compiling a module in GHCi; we must
801 -- use the same name for externally-visible things as we did before.
805 findExternalRules :: [CoreBind]
806 -> [CoreRule] -- Non-local rules (i.e. ones for imported fns)
807 -> UnfoldEnv -- Ids that are exported, so we need their rules
809 -- The complete rules are gotten by combining
810 -- a) the non-local rules
811 -- b) rules embedded in the top-level Ids
812 findExternalRules binds non_local_rules unfold_env
813 = filter (not . internal_rule) (non_local_rules ++ local_rules)
816 | id <- bindersOfBinds binds,
818 rule <- idCoreRules id
822 = any (not . external_id) (varSetElems (ruleLhsFreeIds rule))
823 -- Don't export a rule whose LHS mentions a locally-defined
824 -- Id that is completely internal (i.e. not visible to an
828 | Just (name,_) <- lookupVarEnv unfold_env id = isExternalName name
834 %************************************************************************
836 \subsection{Step 2: top-level tidying}
838 %************************************************************************
842 -- TopTidyEnv: when tidying we need to know
843 -- * nc_var: The NameCache, containing a unique supply and any pre-ordained Names.
844 -- These may have arisen because the
845 -- renamer read in an interface file mentioning M.$wf, say,
846 -- and assigned it unique r77. If, on this compilation, we've
847 -- invented an Id whose name is $wf (but with a different unique)
848 -- we want to rename it to have unique r77, so that we can do easy
849 -- comparisons with stuff from the interface file
851 -- * occ_env: The TidyOccEnv, which tells us which local occurrences
854 -- * subst_env: A Var->Var mapping that substitutes the new Var for the old
856 tidyTopBinds :: HscEnv
860 -> (TidyEnv, [CoreBind])
862 tidyTopBinds hsc_env unfold_env init_occ_env binds
863 = tidy init_env binds
865 init_env = (init_occ_env, emptyVarEnv)
867 this_pkg = thisPackage (hsc_dflags hsc_env)
869 tidy env [] = (env, [])
870 tidy env (b:bs) = let (env1, b') = tidyTopBind this_pkg unfold_env env b
871 (env2, bs') = tidy env1 bs
875 ------------------------
876 tidyTopBind :: PackageId
880 -> (TidyEnv, CoreBind)
882 tidyTopBind this_pkg unfold_env (occ_env,subst1) (NonRec bndr rhs)
883 = (tidy_env2, NonRec bndr' rhs')
885 Just (name',show_unfold) = lookupVarEnv unfold_env bndr
886 caf_info = hasCafRefs this_pkg subst1 (idArity bndr) rhs
887 (bndr', rhs') = tidyTopPair show_unfold tidy_env2 caf_info name' (bndr, rhs)
888 subst2 = extendVarEnv subst1 bndr bndr'
889 tidy_env2 = (occ_env, subst2)
891 tidyTopBind this_pkg unfold_env (occ_env,subst1) (Rec prs)
892 = (tidy_env2, Rec prs')
894 prs' = [ tidyTopPair show_unfold tidy_env2 caf_info name' (id,rhs)
896 let (name',show_unfold) =
897 expectJust "tidyTopBind" $ lookupVarEnv unfold_env id
900 subst2 = extendVarEnvList subst1 (bndrs `zip` map fst prs')
901 tidy_env2 = (occ_env, subst2)
905 -- the CafInfo for a recursive group says whether *any* rhs in
906 -- the group may refer indirectly to a CAF (because then, they all do).
908 | or [ mayHaveCafRefs (hasCafRefs this_pkg subst1 (idArity bndr) rhs)
909 | (bndr,rhs) <- prs ] = MayHaveCafRefs
910 | otherwise = NoCafRefs
912 -----------------------------------------------------------
913 tidyTopPair :: Bool -- show unfolding
914 -> TidyEnv -- The TidyEnv is used to tidy the IdInfo
915 -- It is knot-tied: don't look at it!
918 -> (Id, CoreExpr) -- Binder and RHS before tidying
920 -- This function is the heart of Step 2
921 -- The rec_tidy_env is the one to use for the IdInfo
922 -- It's necessary because when we are dealing with a recursive
923 -- group, a variable late in the group might be mentioned
924 -- in the IdInfo of one early in the group
926 tidyTopPair show_unfold rhs_tidy_env caf_info name' (bndr, rhs)
929 bndr' = mkGlobalId details name' ty' idinfo'
930 details = idDetails bndr -- Preserve the IdDetails
931 ty' = tidyTopType (idType bndr)
932 rhs' = tidyExpr rhs_tidy_env rhs
934 idinfo' = tidyTopIdInfo (isExternalName name')
935 idinfo unfold_info worker_info
938 unfold_info | show_unfold = mkTopUnfolding rhs'
939 | otherwise = noUnfolding
940 worker_info = tidyWorker rhs_tidy_env show_unfold (workerInfo idinfo)
942 -- Usually the Id will have an accurate arity on it, because
943 -- the simplifier has just run, but not always.
944 -- One case I found was when the last thing the simplifier
945 -- did was to let-bind a non-atomic argument and then float
946 -- it to the top level. So it seems more robust just to
948 arity = exprArity rhs
951 -- tidyTopIdInfo creates the final IdInfo for top-level
952 -- binders. There are two delicate pieces:
954 -- * Arity. After CoreTidy, this arity must not change any more.
955 -- Indeed, CorePrep must eta expand where necessary to make
956 -- the manifest arity equal to the claimed arity.
958 -- * CAF info. This must also remain valid through to code generation.
959 -- We add the info here so that it propagates to all
960 -- occurrences of the binders in RHSs, and hence to occurrences in
961 -- unfoldings, which are inside Ids imported by GHCi. Ditto RULES.
962 -- CoreToStg makes use of this when constructing SRTs.
963 tidyTopIdInfo :: Bool -> IdInfo -> Unfolding
964 -> WorkerInfo -> ArityInfo -> CafInfo
966 tidyTopIdInfo is_external idinfo unfold_info worker_info arity caf_info
967 | not is_external -- For internal Ids (not externally visible)
968 = vanillaIdInfo -- we only need enough info for code generation
969 -- Arity and strictness info are enough;
970 -- c.f. CoreTidy.tidyLetBndr
971 `setCafInfo` caf_info
973 `setAllStrictnessInfo` newStrictnessInfo idinfo
975 | otherwise -- Externally-visible Ids get the whole lot
977 `setCafInfo` caf_info
979 `setAllStrictnessInfo` newStrictnessInfo idinfo
980 `setInlinePragInfo` inlinePragInfo idinfo
981 `setUnfoldingInfo` unfold_info
982 `setWorkerInfo` worker_info
983 -- NB: we throw away the Rules
984 -- They have already been extracted by findExternalRules
988 ------------ Worker --------------
989 tidyWorker :: TidyEnv -> Bool -> WorkerInfo -> WorkerInfo
990 tidyWorker _tidy_env _show_unfold NoWorker
992 tidyWorker tidy_env show_unfold (HasWorker work_id wrap_arity)
993 | show_unfold = HasWorker (tidyVarOcc tidy_env work_id) wrap_arity
994 | otherwise = NoWorker
995 -- NB: do *not* expose the worker if show_unfold is off,
996 -- because that means this thing is a loop breaker or
997 -- marked NOINLINE or something like that
998 -- This is important: if you expose the worker for a loop-breaker
999 -- then you can make the simplifier go into an infinite loop, because
1000 -- in effect the unfolding is exposed. See Trac #1709
1002 -- You might think that if show_unfold is False, then the thing should
1003 -- not be w/w'd in the first place. But a legitimate reason is this:
1004 -- the function returns bottom
1005 -- In this case, show_unfold will be false (we don't expose unfoldings
1006 -- for bottoming functions), but we might still have a worker/wrapper
1007 -- split (see Note [Worker-wrapper for bottoming functions] in WorkWrap.lhs
1010 %************************************************************************
1012 \subsection{Figuring out CafInfo for an expression}
1014 %************************************************************************
1016 hasCafRefs decides whether a top-level closure can point into the dynamic heap.
1017 We mark such things as `MayHaveCafRefs' because this information is
1018 used to decide whether a particular closure needs to be referenced
1021 There are two reasons for setting MayHaveCafRefs:
1022 a) The RHS is a CAF: a top-level updatable thunk.
1023 b) The RHS refers to something that MayHaveCafRefs
1025 Possible improvement: In an effort to keep the number of CAFs (and
1026 hence the size of the SRTs) down, we could also look at the expression and
1027 decide whether it requires a small bounded amount of heap, so we can ignore
1028 it as a CAF. In these cases however, we would need to use an additional
1029 CAF list to keep track of non-collectable CAFs.
1032 hasCafRefs :: PackageId -> VarEnv Var -> Arity -> CoreExpr -> CafInfo
1033 hasCafRefs this_pkg p arity expr
1034 | is_caf || mentions_cafs
1036 | otherwise = NoCafRefs
1038 mentions_cafs = isFastTrue (cafRefs p expr)
1039 is_caf = not (arity > 0 || rhsIsStatic this_pkg expr)
1041 -- NB. we pass in the arity of the expression, which is expected
1042 -- to be calculated by exprArity. This is because exprArity
1043 -- knows how much eta expansion is going to be done by
1044 -- CorePrep later on, and we don't want to duplicate that
1045 -- knowledge in rhsIsStatic below.
1047 cafRefs :: VarEnv Id -> Expr a -> FastBool
1049 -- imported Ids first:
1050 | not (isLocalId id) = fastBool (mayHaveCafRefs (idCafInfo id))
1051 -- now Ids local to this module:
1053 case lookupVarEnv p id of
1054 Just id' -> fastBool (mayHaveCafRefs (idCafInfo id'))
1055 Nothing -> fastBool False
1057 cafRefs _ (Lit _) = fastBool False
1058 cafRefs p (App f a) = fastOr (cafRefs p f) (cafRefs p) a
1059 cafRefs p (Lam _ e) = cafRefs p e
1060 cafRefs p (Let b e) = fastOr (cafRefss p (rhssOfBind b)) (cafRefs p) e
1061 cafRefs p (Case e _bndr _ alts) = fastOr (cafRefs p e) (cafRefss p) (rhssOfAlts alts)
1062 cafRefs p (Note _n e) = cafRefs p e
1063 cafRefs p (Cast e _co) = cafRefs p e
1064 cafRefs _ (Type _) = fastBool False
1066 cafRefss :: VarEnv Id -> [Expr a] -> FastBool
1067 cafRefss _ [] = fastBool False
1068 cafRefss p (e:es) = fastOr (cafRefs p e) (cafRefss p) es
1070 fastOr :: FastBool -> (a -> FastBool) -> a -> FastBool
1071 -- hack for lazy-or over FastBool.
1072 fastOr a f x = fastBool (isFastTrue a || isFastTrue (f x))