2 % (c) The University of Glasgow 2006
3 % (c) The GRASP/AQUA Project, Glasgow University, 1992-1998
10 tcSimplifyInfer, tcSimplifyInferCheck,
11 tcSimplifyCheck, tcSimplifyRestricted,
12 tcSimplifyRuleLhs, tcSimplifyIPs,
13 tcSimplifySuperClasses,
14 tcSimplifyTop, tcSimplifyInteractive,
15 tcSimplifyBracket, tcSimplifyCheckPat,
17 tcSimplifyDeriv, tcSimplifyDefault,
23 #include "HsVersions.h"
25 import {-# SOURCE #-} TcUnify( unifyType )
29 import TcHsSyn ( hsLPatType )
37 import DsUtils -- Big-tuple functions
66 %************************************************************************
70 %************************************************************************
72 --------------------------------------
73 Notes on functional dependencies (a bug)
74 --------------------------------------
81 instance D a b => C a b -- Undecidable
82 -- (Not sure if it's crucial to this eg)
83 f :: C a b => a -> Bool
86 g :: C a b => a -> Bool
89 Here f typechecks, but g does not!! Reason: before doing improvement,
90 we reduce the (C a b1) constraint from the call of f to (D a b1).
92 Here is a more complicated example:
95 > class Foo a b | a->b
97 > class Bar a b | a->b
101 > instance Bar Obj Obj
103 > instance (Bar a b) => Foo a b
105 > foo:: (Foo a b) => a -> String
108 > runFoo:: (forall a b. (Foo a b) => a -> w) -> w
114 Could not deduce (Bar a b) from the context (Foo a b)
115 arising from use of `foo' at <interactive>:1
117 Add (Bar a b) to the expected type of an expression
118 In the first argument of `runFoo', namely `foo'
119 In the definition of `it': it = runFoo foo
121 Why all of the sudden does GHC need the constraint Bar a b? The
122 function foo didn't ask for that...
125 The trouble is that to type (runFoo foo), GHC has to solve the problem:
127 Given constraint Foo a b
128 Solve constraint Foo a b'
130 Notice that b and b' aren't the same. To solve this, just do
131 improvement and then they are the same. But GHC currently does
136 That is usually fine, but it isn't here, because it sees that Foo a b is
137 not the same as Foo a b', and so instead applies the instance decl for
138 instance Bar a b => Foo a b. And that's where the Bar constraint comes
141 The Right Thing is to improve whenever the constraint set changes at
142 all. Not hard in principle, but it'll take a bit of fiddling to do.
144 Note [Choosing which variables to quantify]
145 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
146 Suppose we are about to do a generalisation step. We have in our hand
149 T the type of the RHS
150 C the constraints from that RHS
152 The game is to figure out
154 Q the set of type variables over which to quantify
155 Ct the constraints we will *not* quantify over
156 Cq the constraints we will quantify over
158 So we're going to infer the type
162 and float the constraints Ct further outwards.
164 Here are the things that *must* be true:
166 (A) Q intersect fv(G) = EMPTY limits how big Q can be
167 (B) Q superset fv(Cq union T) \ oclose(fv(G),C) limits how small Q can be
169 (A) says we can't quantify over a variable that's free in the environment.
170 (B) says we must quantify over all the truly free variables in T, else
171 we won't get a sufficiently general type.
173 We do not *need* to quantify over any variable that is fixed by the
174 free vars of the environment G.
176 BETWEEN THESE TWO BOUNDS, ANY Q WILL DO!
178 Example: class H x y | x->y where ...
180 fv(G) = {a} C = {H a b, H c d}
183 (A) Q intersect {a} is empty
184 (B) Q superset {a,b,c,d} \ oclose({a}, C) = {a,b,c,d} \ {a,b} = {c,d}
186 So Q can be {c,d}, {b,c,d}
188 In particular, it's perfectly OK to quantify over more type variables
189 than strictly necessary; there is no need to quantify over 'b', since
190 it is determined by 'a' which is free in the envt, but it's perfectly
191 OK to do so. However we must not quantify over 'a' itself.
193 Other things being equal, however, we'd like to quantify over as few
194 variables as possible: smaller types, fewer type applications, more
195 constraints can get into Ct instead of Cq. Here's a good way to
198 Q = grow( fv(T), C ) \ oclose( fv(G), C )
200 That is, quantify over all variable that that MIGHT be fixed by the
201 call site (which influences T), but which aren't DEFINITELY fixed by
202 G. This choice definitely quantifies over enough type variables,
203 albeit perhaps too many.
205 Why grow( fv(T), C ) rather than fv(T)? Consider
207 class H x y | x->y where ...
212 If we used fv(T) = {c} we'd get the type
214 forall c. H c d => c -> b
216 And then if the fn was called at several different c's, each of
217 which fixed d differently, we'd get a unification error, because
218 d isn't quantified. Solution: quantify d. So we must quantify
219 everything that might be influenced by c.
221 Why not oclose( fv(T), C )? Because we might not be able to see
222 all the functional dependencies yet:
224 class H x y | x->y where ...
225 instance H x y => Eq (T x y) where ...
230 Now oclose(fv(T),C) = {c}, because the functional dependency isn't
231 apparent yet, and that's wrong. We must really quantify over d too.
233 There really isn't any point in quantifying over any more than
234 grow( fv(T), C ), because the call sites can't possibly influence
235 any other type variables.
239 -------------------------------------
241 -------------------------------------
243 It's very hard to be certain when a type is ambiguous. Consider
247 instance H x y => K (x,y)
249 Is this type ambiguous?
250 forall a b. (K (a,b), Eq b) => a -> a
252 Looks like it! But if we simplify (K (a,b)) we get (H a b) and
253 now we see that a fixes b. So we can't tell about ambiguity for sure
254 without doing a full simplification. And even that isn't possible if
255 the context has some free vars that may get unified. Urgle!
257 Here's another example: is this ambiguous?
258 forall a b. Eq (T b) => a -> a
259 Not if there's an insance decl (with no context)
260 instance Eq (T b) where ...
262 You may say of this example that we should use the instance decl right
263 away, but you can't always do that:
265 class J a b where ...
266 instance J Int b where ...
268 f :: forall a b. J a b => a -> a
270 (Notice: no functional dependency in J's class decl.)
271 Here f's type is perfectly fine, provided f is only called at Int.
272 It's premature to complain when meeting f's signature, or even
273 when inferring a type for f.
277 However, we don't *need* to report ambiguity right away. It'll always
278 show up at the call site.... and eventually at main, which needs special
279 treatment. Nevertheless, reporting ambiguity promptly is an excellent thing.
281 So here's the plan. We WARN about probable ambiguity if
283 fv(Cq) is not a subset of oclose(fv(T) union fv(G), C)
285 (all tested before quantification).
286 That is, all the type variables in Cq must be fixed by the the variables
287 in the environment, or by the variables in the type.
289 Notice that we union before calling oclose. Here's an example:
291 class J a b c | a b -> c
295 forall b c. (J a b c) => b -> b
297 Only if we union {a} from G with {b} from T before using oclose,
298 do we see that c is fixed.
300 It's a bit vague exactly which C we should use for this oclose call. If we
301 don't fix enough variables we might complain when we shouldn't (see
302 the above nasty example). Nothing will be perfect. That's why we can
303 only issue a warning.
306 Can we ever be *certain* about ambiguity? Yes: if there's a constraint
308 c in C such that fv(c) intersect (fv(G) union fv(T)) = EMPTY
310 then c is a "bubble"; there's no way it can ever improve, and it's
311 certainly ambiguous. UNLESS it is a constant (sigh). And what about
316 instance H x y => K (x,y)
318 Is this type ambiguous?
319 forall a b. (K (a,b), Eq b) => a -> a
321 Urk. The (Eq b) looks "definitely ambiguous" but it isn't. What we are after
322 is a "bubble" that's a set of constraints
324 Cq = Ca union Cq' st fv(Ca) intersect (fv(Cq') union fv(T) union fv(G)) = EMPTY
326 Hence another idea. To decide Q start with fv(T) and grow it
327 by transitive closure in Cq (no functional dependencies involved).
328 Now partition Cq using Q, leaving the definitely-ambiguous and probably-ok.
329 The definitely-ambiguous can then float out, and get smashed at top level
330 (which squashes out the constants, like Eq (T a) above)
333 --------------------------------------
334 Notes on principal types
335 --------------------------------------
340 f x = let g y = op (y::Int) in True
342 Here the principal type of f is (forall a. a->a)
343 but we'll produce the non-principal type
344 f :: forall a. C Int => a -> a
347 --------------------------------------
348 The need for forall's in constraints
349 --------------------------------------
351 [Exchange on Haskell Cafe 5/6 Dec 2000]
353 class C t where op :: t -> Bool
354 instance C [t] where op x = True
356 p y = (let f :: c -> Bool; f x = op (y >> return x) in f, y ++ [])
357 q y = (y ++ [], let f :: c -> Bool; f x = op (y >> return x) in f)
359 The definitions of p and q differ only in the order of the components in
360 the pair on their right-hand sides. And yet:
362 ghc and "Typing Haskell in Haskell" reject p, but accept q;
363 Hugs rejects q, but accepts p;
364 hbc rejects both p and q;
365 nhc98 ... (Malcolm, can you fill in the blank for us!).
367 The type signature for f forces context reduction to take place, and
368 the results of this depend on whether or not the type of y is known,
369 which in turn depends on which component of the pair the type checker
372 Solution: if y::m a, float out the constraints
373 Monad m, forall c. C (m c)
374 When m is later unified with [], we can solve both constraints.
377 --------------------------------------
378 Notes on implicit parameters
379 --------------------------------------
381 Note [Inheriting implicit parameters]
382 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
387 where f is *not* a top-level binding.
388 From the RHS of f we'll get the constraint (?y::Int).
389 There are two types we might infer for f:
393 (so we get ?y from the context of f's definition), or
395 f :: (?y::Int) => Int -> Int
397 At first you might think the first was better, becuase then
398 ?y behaves like a free variable of the definition, rather than
399 having to be passed at each call site. But of course, the WHOLE
400 IDEA is that ?y should be passed at each call site (that's what
401 dynamic binding means) so we'd better infer the second.
403 BOTTOM LINE: when *inferring types* you *must* quantify
404 over implicit parameters. See the predicate isFreeWhenInferring.
407 Note [Implicit parameters and ambiguity]
408 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
409 Only a *class* predicate can give rise to ambiguity
410 An *implicit parameter* cannot. For example:
411 foo :: (?x :: [a]) => Int
413 is fine. The call site will suppply a particular 'x'
415 Furthermore, the type variables fixed by an implicit parameter
416 propagate to the others. E.g.
417 foo :: (Show a, ?x::[a]) => Int
419 The type of foo looks ambiguous. But it isn't, because at a call site
421 let ?x = 5::Int in foo
422 and all is well. In effect, implicit parameters are, well, parameters,
423 so we can take their type variables into account as part of the
424 "tau-tvs" stuff. This is done in the function 'FunDeps.grow'.
427 Question 2: type signatures
428 ~~~~~~~~~~~~~~~~~~~~~~~~~~~
429 BUT WATCH OUT: When you supply a type signature, we can't force you
430 to quantify over implicit parameters. For example:
434 This is perfectly reasonable. We do not want to insist on
436 (?x + 1) :: (?x::Int => Int)
438 That would be silly. Here, the definition site *is* the occurrence site,
439 so the above strictures don't apply. Hence the difference between
440 tcSimplifyCheck (which *does* allow implicit paramters to be inherited)
441 and tcSimplifyCheckBind (which does not).
443 What about when you supply a type signature for a binding?
444 Is it legal to give the following explicit, user type
445 signature to f, thus:
450 At first sight this seems reasonable, but it has the nasty property
451 that adding a type signature changes the dynamic semantics.
454 (let f x = (x::Int) + ?y
455 in (f 3, f 3 with ?y=5)) with ?y = 6
461 in (f 3, f 3 with ?y=5)) with ?y = 6
465 Indeed, simply inlining f (at the Haskell source level) would change the
468 Nevertheless, as Launchbury says (email Oct 01) we can't really give the
469 semantics for a Haskell program without knowing its typing, so if you
470 change the typing you may change the semantics.
472 To make things consistent in all cases where we are *checking* against
473 a supplied signature (as opposed to inferring a type), we adopt the
476 a signature does not need to quantify over implicit params.
478 [This represents a (rather marginal) change of policy since GHC 5.02,
479 which *required* an explicit signature to quantify over all implicit
480 params for the reasons mentioned above.]
482 But that raises a new question. Consider
484 Given (signature) ?x::Int
485 Wanted (inferred) ?x::Int, ?y::Bool
487 Clearly we want to discharge the ?x and float the ?y out. But
488 what is the criterion that distinguishes them? Clearly it isn't
489 what free type variables they have. The Right Thing seems to be
490 to float a constraint that
491 neither mentions any of the quantified type variables
492 nor any of the quantified implicit parameters
494 See the predicate isFreeWhenChecking.
497 Question 3: monomorphism
498 ~~~~~~~~~~~~~~~~~~~~~~~~
499 There's a nasty corner case when the monomorphism restriction bites:
503 The argument above suggests that we *must* generalise
504 over the ?y parameter, to get
505 z :: (?y::Int) => Int,
506 but the monomorphism restriction says that we *must not*, giving
508 Why does the momomorphism restriction say this? Because if you have
510 let z = x + ?y in z+z
512 you might not expect the addition to be done twice --- but it will if
513 we follow the argument of Question 2 and generalise over ?y.
516 Question 4: top level
517 ~~~~~~~~~~~~~~~~~~~~~
518 At the top level, monomorhism makes no sense at all.
521 main = let ?x = 5 in print foo
525 woggle :: (?x :: Int) => Int -> Int
528 We definitely don't want (foo :: Int) with a top-level implicit parameter
529 (?x::Int) becuase there is no way to bind it.
534 (A) Always generalise over implicit parameters
535 Bindings that fall under the monomorphism restriction can't
539 * Inlining remains valid
540 * No unexpected loss of sharing
541 * But simple bindings like
543 will be rejected, unless you add an explicit type signature
544 (to avoid the monomorphism restriction)
545 z :: (?y::Int) => Int
547 This seems unacceptable
549 (B) Monomorphism restriction "wins"
550 Bindings that fall under the monomorphism restriction can't
552 Always generalise over implicit parameters *except* for bindings
553 that fall under the monomorphism restriction
556 * Inlining isn't valid in general
557 * No unexpected loss of sharing
558 * Simple bindings like
560 accepted (get value of ?y from binding site)
562 (C) Always generalise over implicit parameters
563 Bindings that fall under the monomorphism restriction can't
564 be generalised, EXCEPT for implicit parameters
566 * Inlining remains valid
567 * Unexpected loss of sharing (from the extra generalisation)
568 * Simple bindings like
570 accepted (get value of ?y from occurrence sites)
575 None of these choices seems very satisfactory. But at least we should
576 decide which we want to do.
578 It's really not clear what is the Right Thing To Do. If you see
582 would you expect the value of ?y to be got from the *occurrence sites*
583 of 'z', or from the valuue of ?y at the *definition* of 'z'? In the
584 case of function definitions, the answer is clearly the former, but
585 less so in the case of non-fucntion definitions. On the other hand,
586 if we say that we get the value of ?y from the definition site of 'z',
587 then inlining 'z' might change the semantics of the program.
589 Choice (C) really says "the monomorphism restriction doesn't apply
590 to implicit parameters". Which is fine, but remember that every
591 innocent binding 'x = ...' that mentions an implicit parameter in
592 the RHS becomes a *function* of that parameter, called at each
593 use of 'x'. Now, the chances are that there are no intervening 'with'
594 clauses that bind ?y, so a decent compiler should common up all
595 those function calls. So I think I strongly favour (C). Indeed,
596 one could make a similar argument for abolishing the monomorphism
597 restriction altogether.
599 BOTTOM LINE: we choose (B) at present. See tcSimplifyRestricted
603 %************************************************************************
605 \subsection{tcSimplifyInfer}
607 %************************************************************************
609 tcSimplify is called when we *inferring* a type. Here's the overall game plan:
611 1. Compute Q = grow( fvs(T), C )
613 2. Partition C based on Q into Ct and Cq. Notice that ambiguous
614 predicates will end up in Ct; we deal with them at the top level
616 3. Try improvement, using functional dependencies
618 4. If Step 3 did any unification, repeat from step 1
619 (Unification can change the result of 'grow'.)
621 Note: we don't reduce dictionaries in step 2. For example, if we have
622 Eq (a,b), we don't simplify to (Eq a, Eq b). So Q won't be different
623 after step 2. However note that we may therefore quantify over more
624 type variables than we absolutely have to.
626 For the guts, we need a loop, that alternates context reduction and
627 improvement with unification. E.g. Suppose we have
629 class C x y | x->y where ...
631 and tcSimplify is called with:
633 Then improvement unifies a with b, giving
636 If we need to unify anything, we rattle round the whole thing all over
643 -> TcTyVarSet -- fv(T); type vars
645 -> TcM ([TcTyVar], -- Tyvars to quantify (zonked and quantified)
646 [Inst], -- Dict Ids that must be bound here (zonked)
647 TcDictBinds) -- Bindings
648 -- Any free (escaping) Insts are tossed into the environment
653 tcSimplifyInfer doc tau_tvs wanted
654 = do { tau_tvs1 <- zonkTcTyVarsAndFV (varSetElems tau_tvs)
655 ; wanted' <- mapM zonkInst wanted -- Zonk before deciding quantified tyvars
656 ; gbl_tvs <- tcGetGlobalTyVars
657 ; let preds1 = fdPredsOfInsts wanted'
658 gbl_tvs1 = oclose preds1 gbl_tvs
659 qtvs = grow preds1 tau_tvs1 `minusVarSet` gbl_tvs1
660 -- See Note [Choosing which variables to quantify]
662 -- To maximise sharing, remove from consideration any
663 -- constraints that don't mention qtvs at all
664 ; let (free, bound) = partition (isFreeWhenInferring qtvs) wanted'
667 -- To make types simple, reduce as much as possible
668 ; traceTc (text "infer" <+> (ppr preds1 $$ ppr (grow preds1 tau_tvs1) $$ ppr gbl_tvs $$
669 ppr gbl_tvs1 $$ ppr free $$ ppr bound))
670 ; (irreds1, binds1) <- tryHardCheckLoop doc bound
672 -- Note [Inference and implication constraints]
673 ; let want_dict d = tyVarsOfInst d `intersectsVarSet` qtvs
674 ; (irreds2, binds2) <- approximateImplications doc want_dict irreds1
676 -- Now work out all over again which type variables to quantify,
677 -- exactly in the same way as before, but starting from irreds2. Why?
678 -- a) By now improvment may have taken place, and we must *not*
679 -- quantify over any variable free in the environment
680 -- tc137 (function h inside g) is an example
682 -- b) Do not quantify over constraints that *now* do not
683 -- mention quantified type variables, because they are
684 -- simply ambiguous (or might be bound further out). Example:
685 -- f :: Eq b => a -> (a, b)
687 -- From the RHS of g we get the MethodInst f77 :: alpha -> (alpha, beta)
688 -- We decide to quantify over 'alpha' alone, but free1 does not include f77
689 -- because f77 mentions 'alpha'. Then reducing leaves only the (ambiguous)
690 -- constraint (Eq beta), which we dump back into the free set
691 -- See test tcfail181
693 -- c) irreds may contain type variables not previously mentioned,
694 -- e.g. instance D a x => Foo [a]
696 -- Then after simplifying we'll get (D a x), and x is fresh
697 -- We must quantify over x else it'll be totally unbound
698 ; tau_tvs2 <- zonkTcTyVarsAndFV (varSetElems tau_tvs1)
699 ; gbl_tvs2 <- zonkTcTyVarsAndFV (varSetElems gbl_tvs1)
700 -- Note that we start from gbl_tvs1
701 -- We use tcGetGlobalTyVars, then oclose wrt preds2, because
702 -- we've already put some of the original preds1 into frees
703 -- E.g. wanteds = C a b (where a->b)
706 -- Then b is fixed by gbl_tvs, so (C a b) will be in free, and
707 -- irreds2 will be empty. But we don't want to generalise over b!
708 ; let preds2 = fdPredsOfInsts irreds2 -- irreds2 is zonked
709 qtvs = grow preds2 tau_tvs2 `minusVarSet` oclose preds2 gbl_tvs2
710 ; let (free, irreds3) = partition (isFreeWhenInferring qtvs) irreds2
713 -- Turn the quantified meta-type variables into real type variables
714 ; qtvs2 <- zonkQuantifiedTyVars (varSetElems qtvs)
716 -- We can't abstract over any remaining unsolved
717 -- implications so instead just float them outwards. Ugh.
718 ; let (q_dicts0, implics) = partition isAbstractableInst irreds3
719 ; loc <- getInstLoc (ImplicOrigin doc)
720 ; implic_bind <- bindIrreds loc qtvs2 q_dicts0 implics
722 -- Prepare equality instances for quantification
723 ; let (q_eqs0,q_dicts) = partition isEqInst q_dicts0
724 ; q_eqs <- mapM finalizeEqInst q_eqs0
726 ; return (qtvs2, q_eqs ++ q_dicts, binds1 `unionBags` binds2 `unionBags` implic_bind) }
727 -- NB: when we are done, we might have some bindings, but
728 -- the final qtvs might be empty. See Note [NO TYVARS] below.
730 approximateImplications :: SDoc -> (Inst -> Bool) -> [Inst] -> TcM ([Inst], TcDictBinds)
731 -- Note [Inference and implication constraints]
732 -- Given a bunch of Dict and ImplicInsts, try to approximate the implications by
733 -- - fetching any dicts inside them that are free
734 -- - using those dicts as cruder constraints, to solve the implications
735 -- - returning the extra ones too
737 approximateImplications doc want_dict irreds
739 = return (irreds, emptyBag)
741 = do { extra_dicts' <- mapM cloneDict extra_dicts
742 ; tryHardCheckLoop doc (extra_dicts' ++ irreds) }
743 -- By adding extra_dicts', we make them
744 -- available to solve the implication constraints
746 extra_dicts = get_dicts (filter isImplicInst irreds)
748 get_dicts :: [Inst] -> [Inst] -- Returns only Dicts
749 -- Find the wanted constraints in implication constraints that satisfy
750 -- want_dict, and are not bound by forall's in the constraint itself
751 get_dicts ds = concatMap get_dict ds
753 get_dict d@(Dict {}) | want_dict d = [d]
755 get_dict (ImplicInst {tci_tyvars = tvs, tci_wanted = wanteds})
756 = [ d | let tv_set = mkVarSet tvs
757 , d <- get_dicts wanteds
758 , not (tyVarsOfInst d `intersectsVarSet` tv_set)]
759 get_dict i@(EqInst {}) | want_dict i = [i]
761 get_dict other = pprPanic "approximateImplications" (ppr other)
764 Note [Inference and implication constraints]
765 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
766 Suppose we have a wanted implication constraint (perhaps arising from
767 a nested pattern match) like
769 and we are now trying to quantify over 'a' when inferring the type for
770 a function. In principle it's possible that there might be an instance
771 instance (C a, E a) => D [a]
772 so the context (E a) would suffice. The Right Thing is to abstract over
773 the implication constraint, but we don't do that (a) because it'll be
774 surprising to programmers and (b) because we don't have the machinery to deal
775 with 'given' implications.
777 So our best approximation is to make (D [a]) part of the inferred
778 context, so we can use that to discharge the implication. Hence
779 the strange function get_dicts in approximateImplications.
781 The common cases are more clear-cut, when we have things like
783 Here, abstracting over (C b) is not an approximation at all -- but see
784 Note [Freeness and implications].
786 See Trac #1430 and test tc228.
790 -----------------------------------------------------------
791 -- tcSimplifyInferCheck is used when we know the constraints we are to simplify
792 -- against, but we don't know the type variables over which we are going to quantify.
793 -- This happens when we have a type signature for a mutually recursive group
796 -> TcTyVarSet -- fv(T)
799 -> TcM ([TyVar], -- Fully zonked, and quantified
800 TcDictBinds) -- Bindings
802 tcSimplifyInferCheck loc tau_tvs givens wanteds
803 = do { traceTc (text "tcSimplifyInferCheck <-" <+> ppr wanteds)
804 ; (irreds, binds) <- gentleCheckLoop loc givens wanteds
806 -- Figure out which type variables to quantify over
807 -- You might think it should just be the signature tyvars,
808 -- but in bizarre cases you can get extra ones
809 -- f :: forall a. Num a => a -> a
810 -- f x = fst (g (x, head [])) + 1
812 -- Here we infer g :: forall a b. a -> b -> (b,a)
813 -- We don't want g to be monomorphic in b just because
814 -- f isn't quantified over b.
815 ; let all_tvs = varSetElems (tau_tvs `unionVarSet` tyVarsOfInsts givens)
816 ; all_tvs <- zonkTcTyVarsAndFV all_tvs
817 ; gbl_tvs <- tcGetGlobalTyVars
818 ; let qtvs = varSetElems (all_tvs `minusVarSet` gbl_tvs)
819 -- We could close gbl_tvs, but its not necessary for
820 -- soundness, and it'll only affect which tyvars, not which
821 -- dictionaries, we quantify over
823 ; qtvs' <- zonkQuantifiedTyVars qtvs
825 -- Now we are back to normal (c.f. tcSimplCheck)
826 ; implic_bind <- bindIrreds loc qtvs' givens irreds
828 ; traceTc (text "tcSimplifyInferCheck ->" <+> ppr (implic_bind))
829 ; return (qtvs', binds `unionBags` implic_bind) }
832 Note [Squashing methods]
833 ~~~~~~~~~~~~~~~~~~~~~~~~~
834 Be careful if you want to float methods more:
835 truncate :: forall a. RealFrac a => forall b. Integral b => a -> b
836 From an application (truncate f i) we get
839 If we have also have a second occurrence of truncate, we get
842 When simplifying with i,f free, we might still notice that
843 t1=t3; but alas, the binding for t2 (which mentions t1)
844 may continue to float out!
849 class Y a b | a -> b where
852 instance Y [[a]] a where
855 k :: X a -> X a -> X a
857 g :: Num a => [X a] -> [X a]
860 h ys = ys ++ map (k (y [[0]])) xs
862 The excitement comes when simplifying the bindings for h. Initially
863 try to simplify {y @ [[t1]] t2, 0 @ t1}, with initial qtvs = {t2}.
864 From this we get t1:=:t2, but also various bindings. We can't forget
865 the bindings (because of [LOOP]), but in fact t1 is what g is
868 The net effect of [NO TYVARS]
871 isFreeWhenInferring :: TyVarSet -> Inst -> Bool
872 isFreeWhenInferring qtvs inst
873 = isFreeWrtTyVars qtvs inst -- Constrains no quantified vars
874 && isInheritableInst inst -- and no implicit parameter involved
875 -- see Note [Inheriting implicit parameters]
877 {- No longer used (with implication constraints)
878 isFreeWhenChecking :: TyVarSet -- Quantified tyvars
879 -> NameSet -- Quantified implicit parameters
881 isFreeWhenChecking qtvs ips inst
882 = isFreeWrtTyVars qtvs inst
883 && isFreeWrtIPs ips inst
886 isFreeWrtTyVars :: VarSet -> Inst -> Bool
887 isFreeWrtTyVars qtvs inst = tyVarsOfInst inst `disjointVarSet` qtvs
888 isFreeWrtIPs :: NameSet -> Inst -> Bool
889 isFreeWrtIPs ips inst = not (any (`elemNameSet` ips) (ipNamesOfInst inst))
893 %************************************************************************
895 \subsection{tcSimplifyCheck}
897 %************************************************************************
899 @tcSimplifyCheck@ is used when we know exactly the set of variables
900 we are going to quantify over. For example, a class or instance declaration.
903 -----------------------------------------------------------
904 -- tcSimplifyCheck is used when checking expression type signatures,
905 -- class decls, instance decls etc.
906 tcSimplifyCheck :: InstLoc
907 -> [TcTyVar] -- Quantify over these
910 -> TcM TcDictBinds -- Bindings
911 tcSimplifyCheck loc qtvs givens wanteds
912 = ASSERT( all isTcTyVar qtvs && all isSkolemTyVar qtvs )
913 do { traceTc (text "tcSimplifyCheck")
914 ; (irreds, binds) <- gentleCheckLoop loc givens wanteds
915 ; implic_bind <- bindIrreds loc qtvs givens irreds
916 ; return (binds `unionBags` implic_bind) }
918 -----------------------------------------------------------
919 -- tcSimplifyCheckPat is used for existential pattern match
920 tcSimplifyCheckPat :: InstLoc
921 -> [TcTyVar] -- Quantify over these
924 -> TcM TcDictBinds -- Bindings
925 tcSimplifyCheckPat loc qtvs givens wanteds
926 = ASSERT( all isTcTyVar qtvs && all isSkolemTyVar qtvs )
927 do { traceTc (text "tcSimplifyCheckPat")
928 ; (irreds, binds) <- gentleCheckLoop loc givens wanteds
929 ; implic_bind <- bindIrredsR loc qtvs givens irreds
930 ; return (binds `unionBags` implic_bind) }
932 -----------------------------------------------------------
933 bindIrreds :: InstLoc -> [TcTyVar]
936 bindIrreds loc qtvs givens irreds
937 = bindIrredsR loc qtvs givens irreds
939 bindIrredsR :: InstLoc -> [TcTyVar] -> [Inst] -> [Inst] -> TcM TcDictBinds
940 -- Make a binding that binds 'irreds', by generating an implication
941 -- constraint for them, *and* throwing the constraint into the LIE
942 bindIrredsR loc qtvs givens irreds
946 = do { let givens' = filter isAbstractableInst givens
947 -- The givens can (redundantly) include methods
948 -- We want to retain both EqInsts and Dicts
949 -- There should be no implicadtion constraints
950 -- See Note [Pruning the givens in an implication constraint]
952 -- If there are no 'givens', then it's safe to
953 -- partition the 'wanteds' by their qtvs, thereby trimming irreds
954 -- See Note [Freeness and implications]
955 ; irreds' <- if null givens'
957 { let qtv_set = mkVarSet qtvs
958 (frees, real_irreds) = partition (isFreeWrtTyVars qtv_set) irreds
960 ; return real_irreds }
963 ; (implics, bind) <- makeImplicationBind loc qtvs givens' irreds'
964 -- This call does the real work
965 -- If irreds' is empty, it does something sensible
970 makeImplicationBind :: InstLoc -> [TcTyVar]
972 -> TcM ([Inst], TcDictBinds)
973 -- Make a binding that binds 'irreds', by generating an implication
974 -- constraint for them, *and* throwing the constraint into the LIE
975 -- The binding looks like
976 -- (ir1, .., irn) = f qtvs givens
977 -- where f is (evidence for) the new implication constraint
978 -- f :: forall qtvs. {reft} givens => (ir1, .., irn)
979 -- qtvs includes coercion variables
981 -- This binding must line up the 'rhs' in reduceImplication
982 makeImplicationBind loc all_tvs
983 givens -- Guaranteed all Dicts
986 | null irreds -- If there are no irreds, we are done
987 = return ([], emptyBag)
988 | otherwise -- Otherwise we must generate a binding
989 = do { uniq <- newUnique
990 ; span <- getSrcSpanM
991 ; let (eq_givens, dict_givens) = partition isEqInst givens
992 eq_tyvar_cos = mkTyVarTys (varSetElems $ tyVarsOfTypes $ map eqInstType eq_givens)
993 -- Urgh! See line 2187 or thereabouts. I believe that all these
994 -- 'givens' must be a simple CoVar. This MUST be cleaned up.
996 ; let name = mkInternalName uniq (mkVarOcc "ic") span
997 implic_inst = ImplicInst { tci_name = name,
998 tci_tyvars = all_tvs,
999 tci_given = (eq_givens ++ dict_givens),
1000 tci_wanted = irreds, tci_loc = loc }
1001 ; let -- only create binder for dict_irreds
1002 (_, dict_irreds) = partition isEqInst irreds
1003 dict_irred_ids = map instToId dict_irreds
1004 lpat = mkBigLHsPatTup (map (L span . VarPat) dict_irred_ids)
1005 rhs = L span (mkHsWrap co (HsVar (instToId implic_inst)))
1006 co = mkWpApps (map instToId dict_givens)
1007 <.> mkWpTyApps eq_tyvar_cos
1008 <.> mkWpTyApps (mkTyVarTys all_tvs)
1009 bind | [dict_irred_id] <- dict_irred_ids = VarBind dict_irred_id rhs
1010 | otherwise = PatBind { pat_lhs = lpat,
1011 pat_rhs = unguardedGRHSs rhs,
1012 pat_rhs_ty = hsLPatType lpat,
1013 bind_fvs = placeHolderNames }
1014 ; traceTc $ text "makeImplicationBind" <+> ppr implic_inst
1015 ; return ([implic_inst], unitBag (L span bind))
1018 -----------------------------------------------------------
1019 tryHardCheckLoop :: SDoc
1021 -> TcM ([Inst], TcDictBinds)
1023 tryHardCheckLoop doc wanteds
1024 = do { (irreds,binds) <- checkLoop (mkRedEnv doc try_me []) wanteds
1025 ; return (irreds,binds)
1028 try_me _ = ReduceMe AddSCs
1029 -- Here's the try-hard bit
1031 -----------------------------------------------------------
1032 gentleCheckLoop :: InstLoc
1035 -> TcM ([Inst], TcDictBinds)
1037 gentleCheckLoop inst_loc givens wanteds
1038 = do { (irreds,binds) <- checkLoop env wanteds
1039 ; return (irreds,binds)
1042 env = mkRedEnv (pprInstLoc inst_loc) try_me givens
1044 try_me inst | isMethodOrLit inst = ReduceMe AddSCs
1046 -- When checking against a given signature
1047 -- we MUST be very gentle: Note [Check gently]
1049 gentleInferLoop :: SDoc -> [Inst]
1050 -> TcM ([Inst], TcDictBinds)
1051 gentleInferLoop doc wanteds
1052 = do { (irreds, binds) <- checkLoop env wanteds
1053 ; return (irreds, binds) }
1055 env = mkRedEnv doc try_me []
1056 try_me inst | isMethodOrLit inst = ReduceMe AddSCs
1061 ~~~~~~~~~~~~~~~~~~~~
1062 We have to very careful about not simplifying too vigorously
1067 f :: Show b => T b -> b
1068 f (MkT x) = show [x]
1070 Inside the pattern match, which binds (a:*, x:a), we know that
1072 Hence we have a dictionary for Show [a] available; and indeed we
1073 need it. We are going to build an implication contraint
1074 forall a. (b~[a]) => Show [a]
1075 Later, we will solve this constraint using the knowledge (Show b)
1077 But we MUST NOT reduce (Show [a]) to (Show a), else the whole
1078 thing becomes insoluble. So we simplify gently (get rid of literals
1079 and methods only, plus common up equal things), deferring the real
1080 work until top level, when we solve the implication constraint
1081 with tryHardCheckLooop.
1085 -----------------------------------------------------------
1088 -> TcM ([Inst], TcDictBinds)
1089 -- Precondition: givens are completely rigid
1090 -- Postcondition: returned Insts are zonked
1092 checkLoop env wanteds
1094 where go env wanteds
1095 = do { -- We do need to zonk the givens; cf Note [Zonking RedEnv]
1096 ; env' <- zonkRedEnv env
1097 ; wanteds' <- zonkInsts wanteds
1099 ; (improved, binds, irreds) <- reduceContext env' wanteds'
1101 ; if null irreds || not improved then
1102 return (irreds, binds)
1105 -- If improvement did some unification, we go round again.
1106 -- We start again with irreds, not wanteds
1107 -- Using an instance decl might have introduced a fresh type
1108 -- variable which might have been unified, so we'd get an
1109 -- infinite loop if we started again with wanteds!
1111 { (irreds1, binds1) <- go env' irreds
1112 ; return (irreds1, binds `unionBags` binds1) } }
1115 Note [Zonking RedEnv]
1116 ~~~~~~~~~~~~~~~~~~~~~
1117 It might appear as if the givens in RedEnv are always rigid, but that is not
1118 necessarily the case for programs involving higher-rank types that have class
1119 contexts constraining the higher-rank variables. An example from tc237 in the
1122 class Modular s a | s -> a
1124 wim :: forall a w. Integral a
1125 => a -> (forall s. Modular s a => M s w) -> w
1126 wim i k = error "urk"
1128 test5 :: (Modular s a, Integral a) => M s a
1131 test4 = wim 4 test4'
1133 Notice how the variable 'a' of (Modular s a) in the rank-2 type of wim is
1134 quantified further outside. When type checking test4, we have to check
1135 whether the signature of test5 is an instance of
1137 (forall s. Modular s a => M s w)
1139 Consequently, we will get (Modular s t_a), where t_a is a TauTv into the
1142 Given the FD of Modular in this example, class improvement will instantiate
1143 t_a to 'a', where 'a' is the skolem from test5's signatures (due to the
1144 Modular s a predicate in that signature). If we don't zonk (Modular s t_a) in
1145 the givens, we will get into a loop as improveOne uses the unification engine
1146 Unify.tcUnifyTys, which doesn't know about mutable type variables.
1151 class If b t e r | b t e -> r
1154 class Lte a b c | a b -> c where lte :: a -> b -> c
1156 instance (Lte a b l,If l b a c) => Max a b c
1158 Wanted: Max Z (S x) y
1160 Then we'll reduce using the Max instance to:
1161 (Lte Z (S x) l, If l (S x) Z y)
1162 and improve by binding l->T, after which we can do some reduction
1163 on both the Lte and If constraints. What we *can't* do is start again
1164 with (Max Z (S x) y)!
1168 %************************************************************************
1170 tcSimplifySuperClasses
1172 %************************************************************************
1174 Note [SUPERCLASS-LOOP 1]
1175 ~~~~~~~~~~~~~~~~~~~~~~~~
1176 We have to be very, very careful when generating superclasses, lest we
1177 accidentally build a loop. Here's an example:
1181 class S a => C a where { opc :: a -> a }
1182 class S b => D b where { opd :: b -> b }
1184 instance C Int where
1187 instance D Int where
1190 From (instance C Int) we get the constraint set {ds1:S Int, dd:D Int}
1191 Simplifying, we may well get:
1192 $dfCInt = :C ds1 (opd dd)
1195 Notice that we spot that we can extract ds1 from dd.
1197 Alas! Alack! We can do the same for (instance D Int):
1199 $dfDInt = :D ds2 (opc dc)
1203 And now we've defined the superclass in terms of itself.
1205 Solution: never generate a superclass selectors at all when
1206 satisfying the superclass context of an instance declaration.
1208 Two more nasty cases are in
1213 tcSimplifySuperClasses
1218 tcSimplifySuperClasses loc givens sc_wanteds
1219 = do { traceTc (text "tcSimplifySuperClasses")
1220 ; (irreds,binds1) <- checkLoop env sc_wanteds
1221 ; let (tidy_env, tidy_irreds) = tidyInsts irreds
1222 ; reportNoInstances tidy_env (Just (loc, givens)) tidy_irreds
1225 env = mkRedEnv (pprInstLoc loc) try_me givens
1226 try_me _ = ReduceMe NoSCs
1227 -- Like tryHardCheckLoop, but with NoSCs
1231 %************************************************************************
1233 \subsection{tcSimplifyRestricted}
1235 %************************************************************************
1237 tcSimplifyRestricted infers which type variables to quantify for a
1238 group of restricted bindings. This isn't trivial.
1241 We want to quantify over a to get id :: forall a. a->a
1244 We do not want to quantify over a, because there's an Eq a
1245 constraint, so we get eq :: a->a->Bool (notice no forall)
1248 RHS has type 'tau', whose free tyvars are tau_tvs
1249 RHS has constraints 'wanteds'
1252 Quantify over (tau_tvs \ ftvs(wanteds))
1253 This is bad. The constraints may contain (Monad (ST s))
1254 where we have instance Monad (ST s) where...
1255 so there's no need to be monomorphic in s!
1257 Also the constraint might be a method constraint,
1258 whose type mentions a perfectly innocent tyvar:
1259 op :: Num a => a -> b -> a
1260 Here, b is unconstrained. A good example would be
1262 We want to infer the polymorphic type
1263 foo :: forall b. b -> b
1266 Plan B (cunning, used for a long time up to and including GHC 6.2)
1267 Step 1: Simplify the constraints as much as possible (to deal
1268 with Plan A's problem). Then set
1269 qtvs = tau_tvs \ ftvs( simplify( wanteds ) )
1271 Step 2: Now simplify again, treating the constraint as 'free' if
1272 it does not mention qtvs, and trying to reduce it otherwise.
1273 The reasons for this is to maximise sharing.
1275 This fails for a very subtle reason. Suppose that in the Step 2
1276 a constraint (Foo (Succ Zero) (Succ Zero) b) gets thrown upstairs as 'free'.
1277 In the Step 1 this constraint might have been simplified, perhaps to
1278 (Foo Zero Zero b), AND THEN THAT MIGHT BE IMPROVED, to bind 'b' to 'T'.
1279 This won't happen in Step 2... but that in turn might prevent some other
1280 constraint (Baz [a] b) being simplified (e.g. via instance Baz [a] T where {..})
1281 and that in turn breaks the invariant that no constraints are quantified over.
1283 Test typecheck/should_compile/tc177 (which failed in GHC 6.2) demonstrates
1288 Step 1: Simplify the constraints as much as possible (to deal
1289 with Plan A's problem). Then set
1290 qtvs = tau_tvs \ ftvs( simplify( wanteds ) )
1291 Return the bindings from Step 1.
1294 A note about Plan C (arising from "bug" reported by George Russel March 2004)
1297 instance (HasBinary ty IO) => HasCodedValue ty
1299 foo :: HasCodedValue a => String -> IO a
1301 doDecodeIO :: HasCodedValue a => () -> () -> IO a
1302 doDecodeIO codedValue view
1303 = let { act = foo "foo" } in act
1305 You might think this should work becuase the call to foo gives rise to a constraint
1306 (HasCodedValue t), which can be satisfied by the type sig for doDecodeIO. But the
1307 restricted binding act = ... calls tcSimplifyRestricted, and PlanC simplifies the
1308 constraint using the (rather bogus) instance declaration, and now we are stuffed.
1310 I claim this is not really a bug -- but it bit Sergey as well as George. So here's
1314 Plan D (a variant of plan B)
1315 Step 1: Simplify the constraints as much as possible (to deal
1316 with Plan A's problem), BUT DO NO IMPROVEMENT. Then set
1317 qtvs = tau_tvs \ ftvs( simplify( wanteds ) )
1319 Step 2: Now simplify again, treating the constraint as 'free' if
1320 it does not mention qtvs, and trying to reduce it otherwise.
1322 The point here is that it's generally OK to have too few qtvs; that is,
1323 to make the thing more monomorphic than it could be. We don't want to
1324 do that in the common cases, but in wierd cases it's ok: the programmer
1325 can always add a signature.
1327 Too few qtvs => too many wanteds, which is what happens if you do less
1332 tcSimplifyRestricted -- Used for restricted binding groups
1333 -- i.e. ones subject to the monomorphism restriction
1336 -> [Name] -- Things bound in this group
1337 -> TcTyVarSet -- Free in the type of the RHSs
1338 -> [Inst] -- Free in the RHSs
1339 -> TcM ([TyVar], -- Tyvars to quantify (zonked and quantified)
1340 TcDictBinds) -- Bindings
1341 -- tcSimpifyRestricted returns no constraints to
1342 -- quantify over; by definition there are none.
1343 -- They are all thrown back in the LIE
1345 tcSimplifyRestricted doc top_lvl bndrs tau_tvs wanteds
1346 -- Zonk everything in sight
1347 = do { traceTc (text "tcSimplifyRestricted")
1348 ; wanteds' <- zonkInsts wanteds
1350 -- 'ReduceMe': Reduce as far as we can. Don't stop at
1351 -- dicts; the idea is to get rid of as many type
1352 -- variables as possible, and we don't want to stop
1353 -- at (say) Monad (ST s), because that reduces
1354 -- immediately, with no constraint on s.
1356 -- BUT do no improvement! See Plan D above
1357 -- HOWEVER, some unification may take place, if we instantiate
1358 -- a method Inst with an equality constraint
1359 ; let env = mkNoImproveRedEnv doc (\_ -> ReduceMe AddSCs)
1360 ; (_imp, _binds, constrained_dicts) <- reduceContext env wanteds'
1362 -- Next, figure out the tyvars we will quantify over
1363 ; tau_tvs' <- zonkTcTyVarsAndFV (varSetElems tau_tvs)
1364 ; gbl_tvs' <- tcGetGlobalTyVars
1365 ; constrained_dicts' <- zonkInsts constrained_dicts
1367 ; let qtvs1 = tau_tvs' `minusVarSet` oclose (fdPredsOfInsts constrained_dicts) gbl_tvs'
1368 -- As in tcSimplifyInfer
1370 -- Do not quantify over constrained type variables:
1371 -- this is the monomorphism restriction
1372 constrained_tvs' = tyVarsOfInsts constrained_dicts'
1373 qtvs = qtvs1 `minusVarSet` constrained_tvs'
1374 pp_bndrs = pprWithCommas (quotes . ppr) bndrs
1377 ; warn_mono <- doptM Opt_WarnMonomorphism
1378 ; warnTc (warn_mono && (constrained_tvs' `intersectsVarSet` qtvs1))
1379 (vcat[ ptext (sLit "the Monomorphism Restriction applies to the binding")
1380 <> plural bndrs <+> ptext (sLit "for") <+> pp_bndrs,
1381 ptext (sLit "Consider giving a type signature for") <+> pp_bndrs])
1383 ; traceTc (text "tcSimplifyRestricted" <+> vcat [
1384 pprInsts wanteds, pprInsts constrained_dicts',
1386 ppr constrained_tvs', ppr tau_tvs', ppr qtvs ])
1388 -- The first step may have squashed more methods than
1389 -- necessary, so try again, this time more gently, knowing the exact
1390 -- set of type variables to quantify over.
1392 -- We quantify only over constraints that are captured by qtvs;
1393 -- these will just be a subset of non-dicts. This in contrast
1394 -- to normal inference (using isFreeWhenInferring) in which we quantify over
1395 -- all *non-inheritable* constraints too. This implements choice
1396 -- (B) under "implicit parameter and monomorphism" above.
1398 -- Remember that we may need to do *some* simplification, to
1399 -- (for example) squash {Monad (ST s)} into {}. It's not enough
1400 -- just to float all constraints
1402 -- At top level, we *do* squash methods becuase we want to
1403 -- expose implicit parameters to the test that follows
1404 ; let is_nested_group = isNotTopLevel top_lvl
1405 try_me inst | isFreeWrtTyVars qtvs inst,
1406 (is_nested_group || isDict inst) = Stop
1407 | otherwise = ReduceMe AddSCs
1408 env = mkNoImproveRedEnv doc try_me
1409 ; (_imp, binds, irreds) <- reduceContext env wanteds'
1411 -- See "Notes on implicit parameters, Question 4: top level"
1412 ; ASSERT( all (isFreeWrtTyVars qtvs) irreds ) -- None should be captured
1413 if is_nested_group then
1415 else do { let (bad_ips, non_ips) = partition isIPDict irreds
1416 ; addTopIPErrs bndrs bad_ips
1417 ; extendLIEs non_ips }
1419 ; qtvs' <- zonkQuantifiedTyVars (varSetElems qtvs)
1420 ; return (qtvs', binds) }
1424 %************************************************************************
1428 %************************************************************************
1430 On the LHS of transformation rules we only simplify methods and constants,
1431 getting dictionaries. We want to keep all of them unsimplified, to serve
1432 as the available stuff for the RHS of the rule.
1434 Example. Consider the following left-hand side of a rule
1436 f (x == y) (y > z) = ...
1438 If we typecheck this expression we get constraints
1440 d1 :: Ord a, d2 :: Eq a
1442 We do NOT want to "simplify" to the LHS
1444 forall x::a, y::a, z::a, d1::Ord a.
1445 f ((==) (eqFromOrd d1) x y) ((>) d1 y z) = ...
1449 forall x::a, y::a, z::a, d1::Ord a, d2::Eq a.
1450 f ((==) d2 x y) ((>) d1 y z) = ...
1452 Here is another example:
1454 fromIntegral :: (Integral a, Num b) => a -> b
1455 {-# RULES "foo" fromIntegral = id :: Int -> Int #-}
1457 In the rule, a=b=Int, and Num Int is a superclass of Integral Int. But
1458 we *dont* want to get
1460 forall dIntegralInt.
1461 fromIntegral Int Int dIntegralInt (scsel dIntegralInt) = id Int
1463 because the scsel will mess up RULE matching. Instead we want
1465 forall dIntegralInt, dNumInt.
1466 fromIntegral Int Int dIntegralInt dNumInt = id Int
1470 g (x == y) (y == z) = ..
1472 where the two dictionaries are *identical*, we do NOT WANT
1474 forall x::a, y::a, z::a, d1::Eq a
1475 f ((==) d1 x y) ((>) d1 y z) = ...
1477 because that will only match if the dict args are (visibly) equal.
1478 Instead we want to quantify over the dictionaries separately.
1480 In short, tcSimplifyRuleLhs must *only* squash LitInst and MethInts, leaving
1481 all dicts unchanged, with absolutely no sharing. It's simpler to do this
1482 from scratch, rather than further parameterise simpleReduceLoop etc.
1483 Simpler, maybe, but alas not simple (see Trac #2494)
1485 * Type errors may give rise to an (unsatisfiable) equality constraint
1487 * Applications of a higher-rank function on the LHS may give
1488 rise to an implication constraint, esp if there are unsatisfiable
1489 equality constraints inside.
1492 tcSimplifyRuleLhs :: [Inst] -> TcM ([Inst], TcDictBinds)
1493 tcSimplifyRuleLhs wanteds
1494 = do { wanteds' <- zonkInsts wanteds
1495 ; (irreds, binds) <- go [] emptyBag wanteds'
1496 ; let (dicts, bad_irreds) = partition isDict irreds
1497 ; traceTc (text "tcSimplifyrulelhs" <+> pprInsts bad_irreds)
1498 ; addNoInstanceErrs (nub bad_irreds)
1499 -- The nub removes duplicates, which has
1500 -- not happened otherwise (see notes above)
1501 ; return (dicts, binds) }
1503 go :: [Inst] -> TcDictBinds -> [Inst] -> TcM ([Inst], TcDictBinds)
1505 = return (irreds, binds)
1506 go irreds binds (w:ws)
1508 = go (w:irreds) binds ws
1509 | isImplicInst w -- Have a go at reducing the implication
1510 = do { (binds1, irreds1) <- reduceImplication red_env w
1511 ; let (bad_irreds, ok_irreds) = partition isImplicInst irreds1
1512 ; go (bad_irreds ++ irreds)
1513 (binds `unionBags` binds1)
1516 = do { w' <- zonkInst w -- So that (3::Int) does not generate a call
1517 -- to fromInteger; this looks fragile to me
1518 ; lookup_result <- lookupSimpleInst w'
1519 ; case lookup_result of
1520 NoInstance -> go (w:irreds) binds ws
1521 GenInst ws' rhs -> go irreds binds' (ws' ++ ws)
1523 binds' = addInstToDictBind binds w rhs
1526 -- Sigh: we need to reduce inside implications
1527 red_env = mkRedEnv doc try_me []
1528 doc = ptext (sLit "Implication constraint in RULE lhs")
1529 try_me inst | isMethodOrLit inst = ReduceMe AddSCs
1530 | otherwise = Stop -- Be gentle
1533 tcSimplifyBracket is used when simplifying the constraints arising from
1534 a Template Haskell bracket [| ... |]. We want to check that there aren't
1535 any constraints that can't be satisfied (e.g. Show Foo, where Foo has no
1536 Show instance), but we aren't otherwise interested in the results.
1537 Nor do we care about ambiguous dictionaries etc. We will type check
1538 this bracket again at its usage site.
1541 tcSimplifyBracket :: [Inst] -> TcM ()
1542 tcSimplifyBracket wanteds
1543 = do { tryHardCheckLoop doc wanteds
1546 doc = text "tcSimplifyBracket"
1550 %************************************************************************
1552 \subsection{Filtering at a dynamic binding}
1554 %************************************************************************
1559 we must discharge all the ?x constraints from B. We also do an improvement
1560 step; if we have ?x::t1 and ?x::t2 we must unify t1, t2.
1562 Actually, the constraints from B might improve the types in ?x. For example
1564 f :: (?x::Int) => Char -> Char
1567 then the constraint (?x::Int) arising from the call to f will
1568 force the binding for ?x to be of type Int.
1571 tcSimplifyIPs :: [Inst] -- The implicit parameters bound here
1574 -- We need a loop so that we do improvement, and then
1575 -- (next time round) generate a binding to connect the two
1577 -- Here the two ?x's have different types, and improvement
1578 -- makes them the same.
1580 tcSimplifyIPs given_ips wanteds
1581 = do { wanteds' <- zonkInsts wanteds
1582 ; given_ips' <- zonkInsts given_ips
1583 -- Unusually for checking, we *must* zonk the given_ips
1585 ; let env = mkRedEnv doc try_me given_ips'
1586 ; (improved, binds, irreds) <- reduceContext env wanteds'
1588 ; if not improved then
1589 ASSERT( all is_free irreds )
1590 do { extendLIEs irreds
1593 tcSimplifyIPs given_ips wanteds }
1595 doc = text "tcSimplifyIPs" <+> ppr given_ips
1596 ip_set = mkNameSet (ipNamesOfInsts given_ips)
1597 is_free inst = isFreeWrtIPs ip_set inst
1599 -- Simplify any methods that mention the implicit parameter
1600 try_me inst | is_free inst = Stop
1601 | otherwise = ReduceMe NoSCs
1605 %************************************************************************
1607 \subsection[binds-for-local-funs]{@bindInstsOfLocalFuns@}
1609 %************************************************************************
1611 When doing a binding group, we may have @Insts@ of local functions.
1612 For example, we might have...
1614 let f x = x + 1 -- orig local function (overloaded)
1615 f.1 = f Int -- two instances of f
1620 The point is: we must drop the bindings for @f.1@ and @f.2@ here,
1621 where @f@ is in scope; those @Insts@ must certainly not be passed
1622 upwards towards the top-level. If the @Insts@ were binding-ified up
1623 there, they would have unresolvable references to @f@.
1625 We pass in an @init_lie@ of @Insts@ and a list of locally-bound @Ids@.
1626 For each method @Inst@ in the @init_lie@ that mentions one of the
1627 @Ids@, we create a binding. We return the remaining @Insts@ (in an
1628 @LIE@), as well as the @HsBinds@ generated.
1631 bindInstsOfLocalFuns :: [Inst] -> [TcId] -> TcM TcDictBinds
1632 -- Simlifies only MethodInsts, and generate only bindings of form
1634 -- We're careful not to even generate bindings of the form
1636 -- You'd think that'd be fine, but it interacts with what is
1637 -- arguably a bug in Match.tidyEqnInfo (see notes there)
1639 bindInstsOfLocalFuns wanteds local_ids
1640 | null overloaded_ids = do
1643 return emptyLHsBinds
1646 = do { (irreds, binds) <- gentleInferLoop doc for_me
1647 ; extendLIEs not_for_me
1651 doc = text "bindInsts" <+> ppr local_ids
1652 overloaded_ids = filter is_overloaded local_ids
1653 is_overloaded id = isOverloadedTy (idType id)
1654 (for_me, not_for_me) = partition (isMethodFor overloaded_set) wanteds
1656 overloaded_set = mkVarSet overloaded_ids -- There can occasionally be a lot of them
1657 -- so it's worth building a set, so that
1658 -- lookup (in isMethodFor) is faster
1662 %************************************************************************
1664 \subsection{Data types for the reduction mechanism}
1666 %************************************************************************
1668 The main control over context reduction is here
1672 = RedEnv { red_doc :: SDoc -- The context
1673 , red_try_me :: Inst -> WhatToDo
1674 , red_improve :: Bool -- True <=> do improvement
1675 , red_givens :: [Inst] -- All guaranteed rigid
1676 -- Always dicts & equalities
1677 -- but see Note [Rigidity]
1678 , red_stack :: (Int, [Inst]) -- Recursion stack (for err msg)
1679 -- See Note [RedStack]
1683 -- The red_givens are rigid so far as cmpInst is concerned.
1684 -- There is one case where they are not totally rigid, namely in tcSimplifyIPs
1685 -- let ?x = e in ...
1686 -- Here, the given is (?x::a), where 'a' is not necy a rigid type
1687 -- But that doesn't affect the comparison, which is based only on mame.
1690 -- The red_stack pair (n,insts) pair is just used for error reporting.
1691 -- 'n' is always the depth of the stack.
1692 -- The 'insts' is the stack of Insts being reduced: to produce X
1693 -- I had to produce Y, to produce Y I had to produce Z, and so on.
1696 mkRedEnv :: SDoc -> (Inst -> WhatToDo) -> [Inst] -> RedEnv
1697 mkRedEnv doc try_me givens
1698 = RedEnv { red_doc = doc, red_try_me = try_me,
1699 red_givens = givens,
1701 red_improve = True }
1703 mkNoImproveRedEnv :: SDoc -> (Inst -> WhatToDo) -> RedEnv
1704 -- Do not do improvement; no givens
1705 mkNoImproveRedEnv doc try_me
1706 = RedEnv { red_doc = doc, red_try_me = try_me,
1709 red_improve = True }
1712 = ReduceMe WantSCs -- Try to reduce this
1713 -- If there's no instance, add the inst to the
1714 -- irreductible ones, but don't produce an error
1715 -- message of any kind.
1716 -- It might be quite legitimate such as (Eq a)!
1718 | Stop -- Return as irreducible unless it can
1719 -- be reduced to a constant in one step
1720 -- Do not add superclasses; see
1722 data WantSCs = NoSCs | AddSCs -- Tells whether we should add the superclasses
1723 -- of a predicate when adding it to the avails
1724 -- The reason for this flag is entirely the super-class loop problem
1725 -- Note [SUPER-CLASS LOOP 1]
1727 zonkRedEnv :: RedEnv -> TcM RedEnv
1729 = do { givens' <- mapM zonkInst (red_givens env)
1730 ; return $ env {red_givens = givens'}
1735 %************************************************************************
1737 \subsection[reduce]{@reduce@}
1739 %************************************************************************
1741 Note [Ancestor Equalities]
1742 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1743 During context reduction, we add to the wanted equalities also those
1744 equalities that (transitively) occur in superclass contexts of wanted
1745 class constraints. Consider the following code
1747 class a ~ Int => C a
1750 If (C a) is wanted, we want to add (a ~ Int), which will be discharged by
1751 substituting Int for a. Hence, we ultimately want (C Int), which we
1752 discharge with the explicit instance.
1755 reduceContext :: RedEnv
1757 -> TcM (ImprovementDone,
1758 TcDictBinds, -- Dictionary bindings
1759 [Inst]) -- Irreducible
1761 reduceContext env wanteds0
1762 = do { traceTc (text "reduceContext" <+> (vcat [
1763 text "----------------------",
1765 text "given" <+> ppr (red_givens env),
1766 text "wanted" <+> ppr wanteds0,
1767 text "----------------------"
1770 -- We want to add as wanted equalities those that (transitively)
1771 -- occur in superclass contexts of wanted class constraints.
1772 -- See Note [Ancestor Equalities]
1773 ; ancestor_eqs <- ancestorEqualities wanteds0
1774 ; traceTc $ text "reduceContext: ancestor eqs" <+> ppr ancestor_eqs
1776 -- Normalise and solve all equality constraints as far as possible
1777 -- and normalise all dictionary constraints wrt to the reduced
1778 -- equalities. The returned wanted constraints include the
1779 -- irreducible wanted equalities.
1780 ; let wanteds = wanteds0 ++ ancestor_eqs
1781 givens = red_givens env
1785 eq_improved) <- tcReduceEqs givens wanteds
1786 ; traceTc $ text "reduceContext: tcReduceEqs result" <+> vcat
1787 [ppr givens', ppr wanteds', ppr normalise_binds]
1789 -- Build the Avail mapping from "given_dicts"
1790 ; (init_state, _) <- getLIE $ do
1791 { init_state <- foldlM addGiven emptyAvails givens'
1795 -- Solve the *wanted* *dictionary* constraints (not implications)
1796 -- This may expose some further equational constraints...
1797 ; let (wanted_implics, wanted_dicts) = partition isImplicInst wanteds'
1798 ; (avails, extra_eqs) <- getLIE (reduceList env wanted_dicts init_state)
1801 dict_irreds) <- extractResults avails wanted_dicts
1802 ; traceTc $ text "reduceContext: extractResults" <+> vcat
1803 [ppr avails, ppr wanted_dicts, ppr dict_binds]
1805 -- Solve the wanted *implications*. In doing so, we can provide
1806 -- as "given" all the dicts that were originally given,
1807 -- *or* for which we now have bindings,
1808 -- *or* which are now irreds
1809 -- NB: Equality irreds need to be converted, as the recursive
1810 -- invocation of the solver will still treat them as wanteds
1812 ; let implic_env = env { red_givens
1813 = givens ++ bound_dicts ++
1814 map wantedToLocalEqInst dict_irreds }
1815 ; (implic_binds_s, implic_irreds_s)
1816 <- mapAndUnzipM (reduceImplication implic_env) wanted_implics
1817 ; let implic_binds = unionManyBags implic_binds_s
1818 implic_irreds = concat implic_irreds_s
1820 -- Collect all irreducible instances, and determine whether we should
1821 -- go round again. We do so in either of two cases:
1822 -- (1) If dictionary reduction or equality solving led to
1823 -- improvement (i.e., instantiated type variables).
1824 -- (2) If we uncovered extra equalities. We will try to solve them
1825 -- in the next iteration.
1827 ; let all_irreds = dict_irreds ++ implic_irreds ++ extra_eqs
1828 avails_improved = availsImproved avails
1829 improvedFlexible = avails_improved || eq_improved
1830 extraEqs = (not . null) extra_eqs
1831 improved = improvedFlexible || extraEqs
1833 improvedHint = (if avails_improved then " [AVAILS]" else "") ++
1834 (if eq_improved then " [EQ]" else "") ++
1835 (if extraEqs then " [EXTRA EQS]" else "")
1837 ; traceTc (text "reduceContext end" <+> (vcat [
1838 text "----------------------",
1840 text "given" <+> ppr givens,
1841 text "wanted" <+> ppr wanteds0,
1843 text "avails" <+> pprAvails avails,
1844 text "improved =" <+> ppr improved <+> text improvedHint,
1845 text "(all) irreds = " <+> ppr all_irreds,
1846 text "dict-binds = " <+> ppr dict_binds,
1847 text "implic-binds = " <+> ppr implic_binds,
1848 text "----------------------"
1852 normalise_binds `unionBags` dict_binds
1853 `unionBags` implic_binds,
1857 tcImproveOne :: Avails -> Inst -> TcM ImprovementDone
1858 tcImproveOne avails inst
1859 | not (isDict inst) = return False
1861 = do { inst_envs <- tcGetInstEnvs
1862 ; let eqns = improveOne (classInstances inst_envs)
1863 (dictPred inst, pprInstArising inst)
1864 [ (dictPred p, pprInstArising p)
1865 | p <- availsInsts avails, isDict p ]
1866 -- Avails has all the superclasses etc (good)
1867 -- It also has all the intermediates of the deduction (good)
1868 -- It does not have duplicates (good)
1869 -- NB that (?x::t1) and (?x::t2) will be held separately in
1870 -- avails so that improve will see them separate
1871 ; traceTc (text "improveOne" <+> ppr inst)
1874 unifyEqns :: [(Equation, (PredType, SDoc), (PredType, SDoc))]
1875 -> TcM ImprovementDone
1876 unifyEqns [] = return False
1878 = do { traceTc (ptext (sLit "Improve:") <+> vcat (map pprEquationDoc eqns))
1879 ; improved <- mapM unify eqns
1880 ; return $ or improved
1883 unify ((qtvs, pairs), what1, what2)
1884 = addErrCtxtM (mkEqnMsg what1 what2) $
1885 do { let freeTyVars = unionVarSets (map tvs_pr pairs)
1887 ; (_, _, tenv) <- tcInstTyVars (varSetElems qtvs)
1888 ; mapM_ (unif_pr tenv) pairs
1889 ; anyM isFilledMetaTyVar $ varSetElems freeTyVars
1892 unif_pr tenv (ty1, ty2) = unifyType (substTy tenv ty1) (substTy tenv ty2)
1894 tvs_pr (ty1, ty2) = tyVarsOfType ty1 `unionVarSet` tyVarsOfType ty2
1896 pprEquationDoc :: (Equation, (PredType, SDoc), (PredType, SDoc)) -> SDoc
1897 pprEquationDoc (eqn, (p1, _), (p2, _))
1898 = vcat [pprEquation eqn, nest 2 (ppr p1), nest 2 (ppr p2)]
1900 mkEqnMsg :: (TcPredType, SDoc) -> (TcPredType, SDoc) -> TidyEnv
1901 -> TcM (TidyEnv, SDoc)
1902 mkEqnMsg (pred1,from1) (pred2,from2) tidy_env
1903 = do { pred1' <- zonkTcPredType pred1
1904 ; pred2' <- zonkTcPredType pred2
1905 ; let { pred1'' = tidyPred tidy_env pred1'
1906 ; pred2'' = tidyPred tidy_env pred2' }
1907 ; let msg = vcat [ptext (sLit "When using functional dependencies to combine"),
1908 nest 2 (sep [ppr pred1'' <> comma, nest 2 from1]),
1909 nest 2 (sep [ppr pred2'' <> comma, nest 2 from2])]
1910 ; return (tidy_env, msg) }
1913 The main context-reduction function is @reduce@. Here's its game plan.
1916 reduceList :: RedEnv -> [Inst] -> Avails -> TcM Avails
1917 reduceList env@(RedEnv {red_stack = (n,stk)}) wanteds state
1918 = do { traceTc (text "reduceList " <+> (ppr wanteds $$ ppr state))
1920 ; when (debugIsOn && (n > 8)) $ do
1921 debugDumpTcRn (hang (ptext (sLit "Interesting! Context reduction stack depth") <+> int n)
1922 2 (ifPprDebug (nest 2 (pprStack stk))))
1923 ; if n >= ctxtStkDepth dopts then
1924 failWithTc (reduceDepthErr n stk)
1928 go [] state = return state
1929 go (w:ws) state = do { state' <- reduce (env {red_stack = (n+1, w:stk)}) w state
1932 -- Base case: we're done!
1933 reduce :: RedEnv -> Inst -> Avails -> TcM Avails
1934 reduce env wanted avails
1936 -- We don't reduce equalities here (and they must not end up as irreds
1941 -- It's the same as an existing inst, or a superclass thereof
1942 | Just _ <- findAvail avails wanted
1943 = do { traceTc (text "reduce: found " <+> ppr wanted)
1948 = do { traceTc (text "reduce" <+> ppr wanted $$ ppr avails)
1949 ; case red_try_me env wanted of {
1950 Stop -> try_simple (addIrred NoSCs);
1951 -- See Note [No superclasses for Stop]
1953 ReduceMe want_scs -> do -- It should be reduced
1954 { (avails, lookup_result) <- reduceInst env avails wanted
1955 ; case lookup_result of
1956 NoInstance -> addIrred want_scs avails wanted
1957 -- Add it and its superclasses
1959 GenInst [] rhs -> addWanted want_scs avails wanted rhs []
1961 GenInst wanteds' rhs
1962 -> do { avails1 <- addIrred NoSCs avails wanted
1963 ; avails2 <- reduceList env wanteds' avails1
1964 ; addWanted want_scs avails2 wanted rhs wanteds' } }
1965 -- Temporarily do addIrred *before* the reduceList,
1966 -- which has the effect of adding the thing we are trying
1967 -- to prove to the database before trying to prove the things it
1968 -- needs. See note [RECURSIVE DICTIONARIES]
1969 -- NB: we must not do an addWanted before, because that adds the
1970 -- superclasses too, and that can lead to a spurious loop; see
1971 -- the examples in [SUPERCLASS-LOOP]
1972 -- So we do an addIrred before, and then overwrite it afterwards with addWanted
1975 -- First, see if the inst can be reduced to a constant in one step
1976 -- Works well for literals (1::Int) and constant dictionaries (d::Num Int)
1977 -- Don't bother for implication constraints, which take real work
1978 try_simple do_this_otherwise
1979 = do { res <- lookupSimpleInst wanted
1981 GenInst [] rhs -> addWanted AddSCs avails wanted rhs []
1982 _ -> do_this_otherwise avails wanted }
1986 Note [SUPERCLASS-LOOP 2]
1987 ~~~~~~~~~~~~~~~~~~~~~~~~
1988 But the above isn't enough. Suppose we are *given* d1:Ord a,
1989 and want to deduce (d2:C [a]) where
1991 class Ord a => C a where
1992 instance Ord [a] => C [a] where ...
1994 Then we'll use the instance decl to deduce C [a] from Ord [a], and then add the
1995 superclasses of C [a] to avails. But we must not overwrite the binding
1996 for Ord [a] (which is obtained from Ord a) with a superclass selection or we'll just
1999 Here's another variant, immortalised in tcrun020
2000 class Monad m => C1 m
2001 class C1 m => C2 m x
2002 instance C2 Maybe Bool
2003 For the instance decl we need to build (C1 Maybe), and it's no good if
2004 we run around and add (C2 Maybe Bool) and its superclasses to the avails
2005 before we search for C1 Maybe.
2007 Here's another example
2008 class Eq b => Foo a b
2009 instance Eq a => Foo [a] a
2013 we'll first deduce that it holds (via the instance decl). We must not
2014 then overwrite the Eq t constraint with a superclass selection!
2016 At first I had a gross hack, whereby I simply did not add superclass constraints
2017 in addWanted, though I did for addGiven and addIrred. This was sub-optimal,
2018 becuase it lost legitimate superclass sharing, and it still didn't do the job:
2019 I found a very obscure program (now tcrun021) in which improvement meant the
2020 simplifier got two bites a the cherry... so something seemed to be an Stop
2021 first time, but reducible next time.
2023 Now we implement the Right Solution, which is to check for loops directly
2024 when adding superclasses. It's a bit like the occurs check in unification.
2027 Note [RECURSIVE DICTIONARIES]
2028 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2030 data D r = ZeroD | SuccD (r (D r));
2032 instance (Eq (r (D r))) => Eq (D r) where
2033 ZeroD == ZeroD = True
2034 (SuccD a) == (SuccD b) = a == b
2037 equalDC :: D [] -> D [] -> Bool;
2040 We need to prove (Eq (D [])). Here's how we go:
2044 by instance decl, holds if
2048 by instance decl of Eq, holds if
2050 where d2 = dfEqList d3
2053 But now we can "tie the knot" to give
2059 and it'll even run! The trick is to put the thing we are trying to prove
2060 (in this case Eq (D []) into the database before trying to prove its
2061 contributing clauses.
2064 %************************************************************************
2066 Reducing a single constraint
2068 %************************************************************************
2071 ---------------------------------------------
2072 reduceInst :: RedEnv -> Avails -> Inst -> TcM (Avails, LookupInstResult)
2073 reduceInst _ avails other_inst
2074 = do { result <- lookupSimpleInst other_inst
2075 ; return (avails, result) }
2078 Note [Equational Constraints in Implication Constraints]
2079 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2081 An implication constraint is of the form
2083 where Given and Wanted may contain both equational and dictionary
2084 constraints. The delay and reduction of these two kinds of constraints
2087 -) In the generated code, wanted Dictionary constraints are wrapped up in an
2088 implication constraint that is created at the code site where the wanted
2089 dictionaries can be reduced via a let-binding. This let-bound implication
2090 constraint is deconstructed at the use-site of the wanted dictionaries.
2092 -) While the reduction of equational constraints is also delayed, the delay
2093 is not manifest in the generated code. The required evidence is generated
2094 in the code directly at the use-site. There is no let-binding and deconstruction
2095 necessary. The main disadvantage is that we cannot exploit sharing as the
2096 same evidence may be generated at multiple use-sites. However, this disadvantage
2097 is limited because it only concerns coercions which are erased.
2099 The different treatment is motivated by the different in representation. Dictionary
2100 constraints require manifest runtime dictionaries, while equations require coercions
2104 ---------------------------------------------
2105 reduceImplication :: RedEnv
2107 -> TcM (TcDictBinds, [Inst])
2110 Suppose we are simplifying the constraint
2111 forall bs. extras => wanted
2112 in the context of an overall simplification problem with givens 'givens'.
2115 * The 'givens' need not mention any of the quantified type variables
2116 e.g. forall {}. Eq a => Eq [a]
2117 forall {}. C Int => D (Tree Int)
2119 This happens when you have something like
2121 T1 :: Eq a => a -> T a
2124 f x = ...(case x of { T1 v -> v==v })...
2127 -- ToDo: should we instantiate tvs? I think it's not necessary
2129 -- Note on coercion variables:
2131 -- The extra given coercion variables are bound at two different sites:
2132 -- -) in the creation context of the implication constraint
2133 -- the solved equational constraints use these binders
2135 -- -) at the solving site of the implication constraint
2136 -- the solved dictionaries use these binders
2137 -- these binders are generated by reduceImplication
2139 reduceImplication env
2140 orig_implic@(ImplicInst { tci_name = name, tci_loc = inst_loc,
2142 tci_given = extra_givens, tci_wanted = wanteds })
2143 = do { -- Solve the sub-problem
2144 ; let try_me _ = ReduceMe AddSCs -- Note [Freeness and implications]
2145 env' = env { red_givens = extra_givens ++ red_givens env
2146 , red_doc = sep [ptext (sLit "reduceImplication for")
2148 nest 2 (parens $ ptext (sLit "within")
2150 , red_try_me = try_me }
2152 ; traceTc (text "reduceImplication" <+> vcat
2153 [ ppr (red_givens env), ppr extra_givens,
2155 ; (irreds, binds) <- checkLoop env' wanteds
2156 ; let (extra_eq_givens, extra_dict_givens) = partition isEqInst extra_givens
2157 -- SLPJ Sept 07: I think this is bogus; currently
2158 -- there are no Eqinsts in extra_givens
2159 dict_ids = map instToId extra_dict_givens
2161 -- Note [Reducing implication constraints]
2162 -- Tom -- update note, put somewhere!
2164 ; traceTc (text "reduceImplication result" <+> vcat
2165 [ppr irreds, ppr binds])
2167 ; -- extract superclass binds
2168 -- (sc_binds,_) <- extractResults avails []
2169 -- ; traceTc (text "reduceImplication sc_binds" <+> vcat
2170 -- [ppr sc_binds, ppr avails])
2173 -- SLPJ Sept 07: what if improvement happened inside the checkLoop?
2174 -- Then we must iterate the outer loop too!
2176 ; let backOff = isEmptyLHsBinds binds && -- no new bindings
2177 (not $ null irreds) && -- but still some irreds
2178 all (not . isEqInst) wanteds
2179 -- we may have instantiated a cotv
2180 -- => must make a new implication constraint!
2182 ; traceTc $ text "reduceImplication condition" <+> ppr backOff
2184 -- Progress is no longer measered by the number of bindings
2185 ; if backOff then -- No progress
2186 -- If there are any irreds, we back off and do nothing
2187 return (emptyBag, [orig_implic])
2189 { (simpler_implic_insts, bind)
2190 <- makeImplicationBind inst_loc tvs extra_givens irreds
2191 -- This binding is useless if the recursive simplification
2192 -- made no progress; but currently we don't try to optimise that
2193 -- case. After all, we only try hard to reduce at top level, or
2194 -- when inferring types.
2196 ; let dict_wanteds = filter (not . isEqInst) wanteds
2197 -- TOMDO: given equational constraints bug!
2198 -- we need a different evidence for given
2199 -- equations depending on whether we solve
2200 -- dictionary constraints or equational constraints
2202 eq_tyvars = varSetElems $ tyVarsOfTypes $ map eqInstType extra_eq_givens
2203 -- SLPJ Sept07: this looks Utterly Wrong to me, but I think
2204 -- that current extra_givens has no EqInsts, so
2205 -- it makes no difference
2206 co = wrap_inline -- Note [Always inline implication constraints]
2208 <.> mkWpLams eq_tyvars
2209 <.> mkWpLams dict_ids
2210 <.> WpLet (binds `unionBags` bind)
2211 wrap_inline | null dict_ids = idHsWrapper
2212 | otherwise = WpInline
2213 rhs = mkLHsWrap co payload
2214 loc = instLocSpan inst_loc
2215 payload = mkBigLHsTup (map (L loc . HsVar . instToId) dict_wanteds)
2218 ; traceTc (vcat [text "reduceImplication" <+> ppr name,
2219 ppr simpler_implic_insts,
2220 text "->" <+> ppr rhs])
2221 ; return (unitBag (L loc (VarBind (instToId orig_implic) rhs)),
2222 simpler_implic_insts)
2225 reduceImplication _ i = pprPanic "reduceImplication" (ppr i)
2228 Note [Always inline implication constraints]
2229 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2230 Suppose an implication constraint floats out of an INLINE function.
2231 Then although the implication has a single call site, it won't be
2232 inlined. And that is bad because it means that even if there is really
2233 *no* overloading (type signatures specify the exact types) there will
2234 still be dictionary passing in the resulting code. To avert this,
2235 we mark the implication constraints themselves as INLINE, at least when
2236 there is no loss of sharing as a result.
2238 Note [Freeness and implications]
2239 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2240 It's hard to say when an implication constraint can be floated out. Consider
2241 forall {} Eq a => Foo [a]
2242 The (Foo [a]) doesn't mention any of the quantified variables, but it
2243 still might be partially satisfied by the (Eq a).
2245 There is a useful special case when it *is* easy to partition the
2246 constraints, namely when there are no 'givens'. Consider
2247 forall {a}. () => Bar b
2248 There are no 'givens', and so there is no reason to capture (Bar b).
2249 We can let it float out. But if there is even one constraint we
2250 must be much more careful:
2251 forall {a}. C a b => Bar (m b)
2252 because (C a b) might have a superclass (D b), from which we might
2253 deduce (Bar [b]) when m later gets instantiated to []. Ha!
2255 Here is an even more exotic example
2257 Now consider the constraint
2258 forall b. D Int b => C Int
2259 We can satisfy the (C Int) from the superclass of D, so we don't want
2260 to float the (C Int) out, even though it mentions no type variable in
2263 One more example: the constraint
2265 instance (C a, E c) => E (a,c)
2267 constraint: forall b. D Int b => E (Int,c)
2269 You might think that the (D Int b) can't possibly contribute
2270 to solving (E (Int,c)), since the latter mentions 'c'. But
2271 in fact it can, because solving the (E (Int,c)) constraint needs
2274 and the (C Int) can be satisfied from the superclass of (D Int b).
2275 So we must still not float (E (Int,c)) out.
2277 To think about: special cases for unary type classes?
2279 Note [Pruning the givens in an implication constraint]
2280 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2281 Suppose we are about to form the implication constraint
2282 forall tvs. Eq a => Ord b
2283 The (Eq a) cannot contribute to the (Ord b), because it has no access to
2284 the type variable 'b'. So we could filter out the (Eq a) from the givens.
2285 But BE CAREFUL of the examples above in [Freeness and implications].
2287 Doing so would be a bit tidier, but all the implication constraints get
2288 simplified away by the optimiser, so it's no great win. So I don't take
2289 advantage of that at the moment.
2291 If you do, BE CAREFUL of wobbly type variables.
2294 %************************************************************************
2296 Avails and AvailHow: the pool of evidence
2298 %************************************************************************
2302 data Avails = Avails !ImprovementDone !AvailEnv
2304 type ImprovementDone = Bool -- True <=> some unification has happened
2305 -- so some Irreds might now be reducible
2306 -- keys that are now
2308 type AvailEnv = FiniteMap Inst AvailHow
2310 = IsIrred -- Used for irreducible dictionaries,
2311 -- which are going to be lambda bound
2313 | Given Inst -- Used for dictionaries for which we have a binding
2314 -- e.g. those "given" in a signature
2316 | Rhs -- Used when there is a RHS
2317 (LHsExpr TcId) -- The RHS
2318 [Inst] -- Insts free in the RHS; we need these too
2320 instance Outputable Avails where
2323 pprAvails :: Avails -> SDoc
2324 pprAvails (Avails imp avails)
2325 = vcat [ ptext (sLit "Avails") <> (if imp then ptext (sLit "[improved]") else empty)
2327 vcat [ sep [ppr inst, nest 2 (equals <+> ppr avail)]
2328 | (inst,avail) <- fmToList avails ]]
2330 instance Outputable AvailHow where
2333 -------------------------
2334 pprAvail :: AvailHow -> SDoc
2335 pprAvail IsIrred = text "Irred"
2336 pprAvail (Given x) = text "Given" <+> ppr x
2337 pprAvail (Rhs rhs bs) = sep [text "Rhs" <+> ppr bs,
2340 -------------------------
2341 extendAvailEnv :: AvailEnv -> Inst -> AvailHow -> AvailEnv
2342 extendAvailEnv env inst avail = addToFM env inst avail
2344 findAvailEnv :: AvailEnv -> Inst -> Maybe AvailHow
2345 findAvailEnv env wanted = lookupFM env wanted
2346 -- NB 1: the Ord instance of Inst compares by the class/type info
2347 -- *not* by unique. So
2348 -- d1::C Int == d2::C Int
2350 emptyAvails :: Avails
2351 emptyAvails = Avails False emptyFM
2353 findAvail :: Avails -> Inst -> Maybe AvailHow
2354 findAvail (Avails _ avails) wanted = findAvailEnv avails wanted
2356 elemAvails :: Inst -> Avails -> Bool
2357 elemAvails wanted (Avails _ avails) = wanted `elemFM` avails
2359 extendAvails :: Avails -> Inst -> AvailHow -> TcM Avails
2361 extendAvails avails@(Avails imp env) inst avail
2362 = do { imp1 <- tcImproveOne avails inst -- Do any improvement
2363 ; return (Avails (imp || imp1) (extendAvailEnv env inst avail)) }
2365 availsInsts :: Avails -> [Inst]
2366 availsInsts (Avails _ avails) = keysFM avails
2368 availsImproved :: Avails -> ImprovementDone
2369 availsImproved (Avails imp _) = imp
2372 Extracting the bindings from a bunch of Avails.
2373 The bindings do *not* come back sorted in dependency order.
2374 We assume that they'll be wrapped in a big Rec, so that the
2375 dependency analyser can sort them out later
2378 type DoneEnv = FiniteMap Inst [Id]
2379 -- Tracks which things we have evidence for
2381 extractResults :: Avails
2383 -> TcM (TcDictBinds, -- Bindings
2384 [Inst], -- The insts bound by the bindings
2385 [Inst]) -- Irreducible ones
2386 -- Note [Reducing implication constraints]
2388 extractResults (Avails _ avails) wanteds
2389 = go emptyBag [] [] emptyFM wanteds
2391 go :: TcDictBinds -- Bindings for dicts
2392 -> [Inst] -- Bound by the bindings
2394 -> DoneEnv -- Has an entry for each inst in the above three sets
2396 -> TcM (TcDictBinds, [Inst], [Inst])
2397 go binds bound_dicts irreds _ []
2398 = return (binds, bound_dicts, irreds)
2400 go binds bound_dicts irreds done (w:ws)
2402 = go binds bound_dicts (w:irreds) done' ws
2404 | Just done_ids@(done_id : rest_done_ids) <- lookupFM done w
2405 = if w_id `elem` done_ids then
2406 go binds bound_dicts irreds done ws
2408 go (add_bind (nlHsVar done_id)) bound_dicts irreds
2409 (addToFM done w (done_id : w_id : rest_done_ids)) ws
2411 | otherwise -- Not yet done
2412 = case findAvailEnv avails w of
2413 Nothing -> pprTrace "Urk: extractResults" (ppr w) $
2414 go binds bound_dicts irreds done ws
2416 Just IsIrred -> go binds bound_dicts (w:irreds) done' ws
2418 Just (Rhs rhs ws') -> go (add_bind rhs) (w:bound_dicts) irreds done' (ws' ++ ws)
2420 Just (Given g) -> go binds' bound_dicts irreds (addToFM done w [g_id]) ws
2423 binds' | w_id == g_id = binds
2424 | otherwise = add_bind (nlHsVar g_id)
2427 done' = addToFM done w [w_id]
2428 add_bind rhs = addInstToDictBind binds w rhs
2432 Note [No superclasses for Stop]
2433 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2434 When we decide not to reduce an Inst -- the 'WhatToDo' --- we still
2435 add it to avails, so that any other equal Insts will be commoned up
2436 right here. However, we do *not* add superclasses. If we have
2439 but a is not bound here, then we *don't* want to derive dn from df
2440 here lest we lose sharing.
2443 addWanted :: WantSCs -> Avails -> Inst -> LHsExpr TcId -> [Inst] -> TcM Avails
2444 addWanted want_scs avails wanted rhs_expr wanteds
2445 = addAvailAndSCs want_scs avails wanted avail
2447 avail = Rhs rhs_expr wanteds
2449 addGiven :: Avails -> Inst -> TcM Avails
2450 addGiven avails given = addAvailAndSCs AddSCs avails given (Given given)
2451 -- Always add superclasses for 'givens'
2453 -- No ASSERT( not (given `elemAvails` avails) ) because in an instance
2454 -- decl for Ord t we can add both Ord t and Eq t as 'givens',
2455 -- so the assert isn't true
2459 addIrred :: WantSCs -> Avails -> Inst -> TcM Avails
2460 addIrred want_scs avails irred = ASSERT2( not (irred `elemAvails` avails), ppr irred $$ ppr avails )
2461 addAvailAndSCs want_scs avails irred IsIrred
2463 addAvailAndSCs :: WantSCs -> Avails -> Inst -> AvailHow -> TcM Avails
2464 addAvailAndSCs want_scs avails inst avail
2465 | not (isClassDict inst) = extendAvails avails inst avail
2466 | NoSCs <- want_scs = extendAvails avails inst avail
2467 | otherwise = do { traceTc (text "addAvailAndSCs" <+> vcat [ppr inst, ppr deps])
2468 ; avails' <- extendAvails avails inst avail
2469 ; addSCs is_loop avails' inst }
2471 is_loop pred = any (`tcEqType` mkPredTy pred) dep_tys
2472 -- Note: this compares by *type*, not by Unique
2473 deps = findAllDeps (unitVarSet (instToVar inst)) avail
2474 dep_tys = map idType (varSetElems deps)
2476 findAllDeps :: IdSet -> AvailHow -> IdSet
2477 -- Find all the Insts that this one depends on
2478 -- See Note [SUPERCLASS-LOOP 2]
2479 -- Watch out, though. Since the avails may contain loops
2480 -- (see Note [RECURSIVE DICTIONARIES]), so we need to track the ones we've seen so far
2481 findAllDeps so_far (Rhs _ kids) = foldl find_all so_far kids
2482 findAllDeps so_far _ = so_far
2484 find_all :: IdSet -> Inst -> IdSet
2486 | isEqInst kid = so_far
2487 | kid_id `elemVarSet` so_far = so_far
2488 | Just avail <- findAvail avails kid = findAllDeps so_far' avail
2489 | otherwise = so_far'
2491 so_far' = extendVarSet so_far kid_id -- Add the new kid to so_far
2492 kid_id = instToId kid
2494 addSCs :: (TcPredType -> Bool) -> Avails -> Inst -> TcM Avails
2495 -- Add all the superclasses of the Inst to Avails
2496 -- The first param says "don't do this because the original thing
2497 -- depends on this one, so you'd build a loop"
2498 -- Invariant: the Inst is already in Avails.
2500 addSCs is_loop avails dict
2501 = ASSERT( isDict dict )
2502 do { sc_dicts <- newDictBndrs (instLoc dict) sc_theta'
2503 ; foldlM add_sc avails (zipEqual "add_scs" sc_dicts sc_sels) }
2505 (clas, tys) = getDictClassTys dict
2506 (tyvars, sc_theta, sc_sels, _) = classBigSig clas
2507 sc_theta' = filter (not . isEqPred) $
2508 substTheta (zipTopTvSubst tyvars tys) sc_theta
2510 add_sc avails (sc_dict, sc_sel)
2511 | is_loop (dictPred sc_dict) = return avails -- See Note [SUPERCLASS-LOOP 2]
2512 | is_given sc_dict = return avails
2513 | otherwise = do { avails' <- extendAvails avails sc_dict (Rhs sc_sel_rhs [dict])
2514 ; addSCs is_loop avails' sc_dict }
2516 sc_sel_rhs = L (instSpan dict) (HsWrap co_fn (HsVar sc_sel))
2517 co_fn = WpApp (instToVar dict) <.> mkWpTyApps tys
2519 is_given :: Inst -> Bool
2520 is_given sc_dict = case findAvail avails sc_dict of
2521 Just (Given _) -> True -- Given is cheaper than superclass selection
2524 -- From the a set of insts obtain all equalities that (transitively) occur in
2525 -- superclass contexts of class constraints (aka the ancestor equalities).
2527 ancestorEqualities :: [Inst] -> TcM [Inst]
2529 = mapM mkWantedEqInst -- turn only equality predicates..
2530 . filter isEqPred -- ..into wanted equality insts
2532 . addAEsToBag emptyBag -- collect the superclass constraints..
2533 . map dictPred -- ..of all predicates in a bag
2534 . filter isClassDict
2536 addAEsToBag :: Bag PredType -> [PredType] -> Bag PredType
2537 addAEsToBag bag [] = bag
2538 addAEsToBag bag (pred:preds)
2539 | pred `elemBag` bag = addAEsToBag bag preds
2540 | isEqPred pred = addAEsToBag bagWithPred preds
2541 | isClassPred pred = addAEsToBag bagWithPred predsWithSCs
2542 | otherwise = addAEsToBag bag preds
2544 bagWithPred = bag `snocBag` pred
2545 predsWithSCs = preds ++ substTheta (zipTopTvSubst tyvars tys) sc_theta
2547 (tyvars, sc_theta, _, _) = classBigSig clas
2548 (clas, tys) = getClassPredTys pred
2552 %************************************************************************
2554 \section{tcSimplifyTop: defaulting}
2556 %************************************************************************
2559 @tcSimplifyTop@ is called once per module to simplify all the constant
2560 and ambiguous Insts.
2562 We need to be careful of one case. Suppose we have
2564 instance Num a => Num (Foo a b) where ...
2566 and @tcSimplifyTop@ is given a constraint (Num (Foo x y)). Then it'll simplify
2567 to (Num x), and default x to Int. But what about y??
2569 It's OK: the final zonking stage should zap y to (), which is fine.
2573 tcSimplifyTop, tcSimplifyInteractive :: [Inst] -> TcM TcDictBinds
2574 tcSimplifyTop wanteds
2575 = tc_simplify_top doc False wanteds
2577 doc = text "tcSimplifyTop"
2579 tcSimplifyInteractive wanteds
2580 = tc_simplify_top doc True wanteds
2582 doc = text "tcSimplifyInteractive"
2584 -- The TcLclEnv should be valid here, solely to improve
2585 -- error message generation for the monomorphism restriction
2586 tc_simplify_top :: SDoc -> Bool -> [Inst] -> TcM (Bag (LHsBind TcId))
2587 tc_simplify_top doc interactive wanteds
2588 = do { dflags <- getDOpts
2589 ; wanteds <- zonkInsts wanteds
2590 ; mapM_ zonkTopTyVar (varSetElems (tyVarsOfInsts wanteds))
2592 ; traceTc (text "tc_simplify_top 0: " <+> ppr wanteds)
2593 ; (irreds1, binds1) <- tryHardCheckLoop doc1 wanteds
2594 -- ; (irreds1, binds1) <- gentleInferLoop doc1 wanteds
2595 ; traceTc (text "tc_simplify_top 1: " <+> ppr irreds1)
2596 ; (irreds2, binds2) <- approximateImplications doc2 (\_ -> True) irreds1
2597 ; traceTc (text "tc_simplify_top 2: " <+> ppr irreds2)
2599 -- Use the defaulting rules to do extra unification
2600 -- NB: irreds2 are already zonked
2601 ; (irreds3, binds3) <- disambiguate doc3 interactive dflags irreds2
2603 -- Deal with implicit parameters
2604 ; let (bad_ips, non_ips) = partition isIPDict irreds3
2605 (ambigs, others) = partition isTyVarDict non_ips
2607 ; topIPErrs bad_ips -- Can arise from f :: Int -> Int
2609 ; addNoInstanceErrs others
2610 ; addTopAmbigErrs ambigs
2612 ; return (binds1 `unionBags` binds2 `unionBags` binds3) }
2614 doc1 = doc <+> ptext (sLit "(first round)")
2615 doc2 = doc <+> ptext (sLit "(approximate)")
2616 doc3 = doc <+> ptext (sLit "(disambiguate)")
2619 If a dictionary constrains a type variable which is
2620 * not mentioned in the environment
2621 * and not mentioned in the type of the expression
2622 then it is ambiguous. No further information will arise to instantiate
2623 the type variable; nor will it be generalised and turned into an extra
2624 parameter to a function.
2626 It is an error for this to occur, except that Haskell provided for
2627 certain rules to be applied in the special case of numeric types.
2629 * at least one of its classes is a numeric class, and
2630 * all of its classes are numeric or standard
2631 then the type variable can be defaulted to the first type in the
2632 default-type list which is an instance of all the offending classes.
2634 So here is the function which does the work. It takes the ambiguous
2635 dictionaries and either resolves them (producing bindings) or
2636 complains. It works by splitting the dictionary list by type
2637 variable, and using @disambigOne@ to do the real business.
2639 @disambigOne@ assumes that its arguments dictionaries constrain all
2640 the same type variable.
2642 ADR Comment 20/6/94: I've changed the @CReturnable@ case to default to
2643 @()@ instead of @Int@. I reckon this is the Right Thing to do since
2644 the most common use of defaulting is code like:
2646 _ccall_ foo `seqPrimIO` bar
2648 Since we're not using the result of @foo@, the result if (presumably)
2652 disambiguate :: SDoc -> Bool -> DynFlags -> [Inst] -> TcM ([Inst], TcDictBinds)
2653 -- Just does unification to fix the default types
2654 -- The Insts are assumed to be pre-zonked
2655 disambiguate doc interactive dflags insts
2657 = return (insts, emptyBag)
2659 | null defaultable_groups
2660 = do { traceTc (text "disambigutate, no defaultable groups" <+> vcat [ppr unaries, ppr insts, ppr bad_tvs, ppr defaultable_groups])
2661 ; return (insts, emptyBag) }
2664 = do { -- Figure out what default types to use
2665 default_tys <- getDefaultTys extended_defaulting ovl_strings
2667 ; traceTc (text "disambiguate1" <+> vcat [ppr insts, ppr unaries, ppr bad_tvs, ppr defaultable_groups])
2668 ; mapM_ (disambigGroup default_tys) defaultable_groups
2670 -- disambigGroup does unification, hence try again
2671 ; tryHardCheckLoop doc insts }
2674 extended_defaulting = interactive || dopt Opt_ExtendedDefaultRules dflags
2675 ovl_strings = dopt Opt_OverloadedStrings dflags
2677 unaries :: [(Inst, Class, TcTyVar)] -- (C tv) constraints
2678 bad_tvs :: TcTyVarSet -- Tyvars mentioned by *other* constraints
2679 (unaries, bad_tvs_s) = partitionWith find_unary insts
2680 bad_tvs = unionVarSets bad_tvs_s
2682 -- Finds unary type-class constraints
2683 find_unary d@(Dict {tci_pred = ClassP cls [ty]})
2684 | Just tv <- tcGetTyVar_maybe ty = Left (d,cls,tv)
2685 find_unary inst = Right (tyVarsOfInst inst)
2687 -- Group by type variable
2688 defaultable_groups :: [[(Inst,Class,TcTyVar)]]
2689 defaultable_groups = filter defaultable_group (equivClasses cmp_tv unaries)
2690 cmp_tv (_,_,tv1) (_,_,tv2) = tv1 `compare` tv2
2692 defaultable_group :: [(Inst,Class,TcTyVar)] -> Bool
2693 defaultable_group ds@((_,_,tv):_)
2694 = isTyConableTyVar tv -- Note [Avoiding spurious errors]
2695 && not (tv `elemVarSet` bad_tvs)
2696 && defaultable_classes [c | (_,c,_) <- ds]
2697 defaultable_group [] = panic "defaultable_group"
2699 defaultable_classes clss
2700 | extended_defaulting = any isInteractiveClass clss
2701 | otherwise = all is_std_class clss && (any is_num_class clss)
2703 -- In interactive mode, or with -XExtendedDefaultRules,
2704 -- we default Show a to Show () to avoid graututious errors on "show []"
2705 isInteractiveClass cls
2706 = is_num_class cls || (classKey cls `elem` [showClassKey, eqClassKey, ordClassKey])
2708 is_num_class cls = isNumericClass cls || (ovl_strings && (cls `hasKey` isStringClassKey))
2709 -- is_num_class adds IsString to the standard numeric classes,
2710 -- when -foverloaded-strings is enabled
2712 is_std_class cls = isStandardClass cls || (ovl_strings && (cls `hasKey` isStringClassKey))
2713 -- Similarly is_std_class
2715 -----------------------
2716 disambigGroup :: [Type] -- The default types
2717 -> [(Inst,Class,TcTyVar)] -- All standard classes of form (C a)
2718 -> TcM () -- Just does unification, to fix the default types
2720 disambigGroup default_tys dicts
2721 = try_default default_tys
2723 (_,_,tyvar) = ASSERT(not (null dicts)) head dicts -- Should be non-empty
2724 classes = [c | (_,c,_) <- dicts]
2726 try_default [] = return ()
2727 try_default (default_ty : default_tys)
2728 = tryTcLIE_ (try_default default_tys) $
2729 do { tcSimplifyDefault [mkClassPred clas [default_ty] | clas <- classes]
2730 -- This may fail; then the tryTcLIE_ kicks in
2731 -- Failure here is caused by there being no type in the
2732 -- default list which can satisfy all the ambiguous classes.
2733 -- For example, if Real a is reqd, but the only type in the
2734 -- default list is Int.
2736 -- After this we can't fail
2737 ; warnDefault dicts default_ty
2738 ; unifyType default_ty (mkTyVarTy tyvar)
2739 ; return () -- TOMDO: do something with the coercion
2743 -----------------------
2744 getDefaultTys :: Bool -> Bool -> TcM [Type]
2745 getDefaultTys extended_deflts ovl_strings
2746 = do { mb_defaults <- getDeclaredDefaultTys
2747 ; case mb_defaults of {
2748 Just tys -> return tys ; -- User-supplied defaults
2751 -- No use-supplied default
2752 -- Use [Integer, Double], plus modifications
2753 { integer_ty <- tcMetaTy integerTyConName
2754 ; checkWiredInTyCon doubleTyCon
2755 ; string_ty <- tcMetaTy stringTyConName
2756 ; return (opt_deflt extended_deflts unitTy
2757 -- Note [Default unitTy]
2759 [integer_ty,doubleTy]
2761 opt_deflt ovl_strings string_ty) } } }
2763 opt_deflt True ty = [ty]
2764 opt_deflt False _ = []
2767 Note [Default unitTy]
2768 ~~~~~~~~~~~~~~~~~~~~~
2769 In interative mode (or with -XExtendedDefaultRules) we add () as the first type we
2770 try when defaulting. This has very little real impact, except in the following case.
2772 Text.Printf.printf "hello"
2773 This has type (forall a. IO a); it prints "hello", and returns 'undefined'. We don't
2774 want the GHCi repl loop to try to print that 'undefined'. The neatest thing is to
2775 default the 'a' to (), rather than to Integer (which is what would otherwise happen;
2776 and then GHCi doesn't attempt to print the (). So in interactive mode, we add
2777 () to the list of defaulting types. See Trac #1200.
2779 Note [Avoiding spurious errors]
2780 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2781 When doing the unification for defaulting, we check for skolem
2782 type variables, and simply don't default them. For example:
2783 f = (*) -- Monomorphic
2784 g :: Num a => a -> a
2786 Here, we get a complaint when checking the type signature for g,
2787 that g isn't polymorphic enough; but then we get another one when
2788 dealing with the (Num a) context arising from f's definition;
2789 we try to unify a with Int (to default it), but find that it's
2790 already been unified with the rigid variable from g's type sig
2793 %************************************************************************
2795 \subsection[simple]{@Simple@ versions}
2797 %************************************************************************
2799 Much simpler versions when there are no bindings to make!
2801 @tcSimplifyThetas@ simplifies class-type constraints formed by
2802 @deriving@ declarations and when specialising instances. We are
2803 only interested in the simplified bunch of class/type constraints.
2805 It simplifies to constraints of the form (C a b c) where
2806 a,b,c are type variables. This is required for the context of
2807 instance declarations.
2810 tcSimplifyDeriv :: InstOrigin
2812 -> ThetaType -- Wanted
2813 -> TcM ThetaType -- Needed
2814 -- Given instance (wanted) => C inst_ty
2815 -- Simplify 'wanted' as much as possible
2817 tcSimplifyDeriv orig tyvars theta
2818 = do { (tvs, _, tenv) <- tcInstTyVars tyvars
2819 -- The main loop may do unification, and that may crash if
2820 -- it doesn't see a TcTyVar, so we have to instantiate. Sigh
2821 -- ToDo: what if two of them do get unified?
2822 ; wanteds <- newDictBndrsO orig (substTheta tenv theta)
2823 ; (irreds, _) <- tryHardCheckLoop doc wanteds
2825 ; let (tv_dicts, others) = partition ok irreds
2826 ; addNoInstanceErrs others
2827 -- See Note [Exotic derived instance contexts] in TcMType
2829 ; let rev_env = zipTopTvSubst tvs (mkTyVarTys tyvars)
2830 simpl_theta = substTheta rev_env (map dictPred tv_dicts)
2831 -- This reverse-mapping is a pain, but the result
2832 -- should mention the original TyVars not TcTyVars
2834 ; return simpl_theta }
2836 doc = ptext (sLit "deriving classes for a data type")
2838 ok dict | isDict dict = validDerivPred (dictPred dict)
2843 @tcSimplifyDefault@ just checks class-type constraints, essentially;
2844 used with \tr{default} declarations. We are only interested in
2845 whether it worked or not.
2848 tcSimplifyDefault :: ThetaType -- Wanted; has no type variables in it
2851 tcSimplifyDefault theta = do
2852 wanteds <- newDictBndrsO DefaultOrigin theta
2853 (irreds, _) <- tryHardCheckLoop doc wanteds
2854 addNoInstanceErrs irreds
2858 traceTc (ptext (sLit "tcSimplifyDefault failing")) >> failM
2860 doc = ptext (sLit "default declaration")
2864 %************************************************************************
2866 \section{Errors and contexts}
2868 %************************************************************************
2870 ToDo: for these error messages, should we note the location as coming
2871 from the insts, or just whatever seems to be around in the monad just
2875 groupErrs :: ([Inst] -> TcM ()) -- Deal with one group
2876 -> [Inst] -- The offending Insts
2878 -- Group together insts with the same origin
2879 -- We want to report them together in error messages
2883 groupErrs report_err (inst:insts)
2884 = do { do_one (inst:friends)
2885 ; groupErrs report_err others }
2887 -- (It may seem a bit crude to compare the error messages,
2888 -- but it makes sure that we combine just what the user sees,
2889 -- and it avoids need equality on InstLocs.)
2890 (friends, others) = partition is_friend insts
2891 loc_msg = showSDoc (pprInstLoc (instLoc inst))
2892 is_friend friend = showSDoc (pprInstLoc (instLoc friend)) == loc_msg
2893 do_one insts = addInstCtxt (instLoc (head insts)) (report_err insts)
2894 -- Add location and context information derived from the Insts
2896 -- Add the "arising from..." part to a message about bunch of dicts
2897 addInstLoc :: [Inst] -> Message -> Message
2898 addInstLoc insts msg = msg $$ nest 2 (pprInstArising (head insts))
2900 addTopIPErrs :: [Name] -> [Inst] -> TcM ()
2903 addTopIPErrs bndrs ips
2904 = do { dflags <- getDOpts
2905 ; addErrTcM (tidy_env, mk_msg dflags tidy_ips) }
2907 (tidy_env, tidy_ips) = tidyInsts ips
2909 = vcat [sep [ptext (sLit "Implicit parameters escape from"),
2910 nest 2 (ptext (sLit "the monomorphic top-level binding")
2911 <> plural bndrs <+> ptext (sLit "of")
2912 <+> pprBinders bndrs <> colon)],
2913 nest 2 (vcat (map ppr_ip ips)),
2914 monomorphism_fix dflags]
2915 ppr_ip ip = pprPred (dictPred ip) <+> pprInstArising ip
2917 topIPErrs :: [Inst] -> TcM ()
2919 = groupErrs report tidy_dicts
2921 (tidy_env, tidy_dicts) = tidyInsts dicts
2922 report dicts = addErrTcM (tidy_env, mk_msg dicts)
2923 mk_msg dicts = addInstLoc dicts (ptext (sLit "Unbound implicit parameter") <>
2924 plural tidy_dicts <+> pprDictsTheta tidy_dicts)
2926 addNoInstanceErrs :: [Inst] -- Wanted (can include implications)
2928 addNoInstanceErrs insts
2929 = do { let (tidy_env, tidy_insts) = tidyInsts insts
2930 ; reportNoInstances tidy_env Nothing tidy_insts }
2934 -> Maybe (InstLoc, [Inst]) -- Context
2935 -- Nothing => top level
2936 -- Just (d,g) => d describes the construct
2938 -> [Inst] -- What is wanted (can include implications)
2941 reportNoInstances tidy_env mb_what insts
2942 = groupErrs (report_no_instances tidy_env mb_what) insts
2944 report_no_instances :: TidyEnv -> Maybe (InstLoc, [Inst]) -> [Inst] -> TcM ()
2945 report_no_instances tidy_env mb_what insts
2946 = do { inst_envs <- tcGetInstEnvs
2947 ; let (implics, insts1) = partition isImplicInst insts
2948 (insts2, overlaps) = partitionWith (check_overlap inst_envs) insts1
2949 (eqInsts, insts3) = partition isEqInst insts2
2950 ; traceTc (text "reportNoInstances" <+> vcat
2951 [ppr insts, ppr implics, ppr insts1, ppr insts2])
2952 ; mapM_ complain_implic implics
2953 ; mapM_ (\doc -> addErrTcM (tidy_env, doc)) overlaps
2954 ; groupErrs complain_no_inst insts3
2955 ; mapM_ (addErrTcM . mk_eq_err) eqInsts
2958 complain_no_inst insts = addErrTcM (tidy_env, mk_no_inst_err insts)
2960 complain_implic inst -- Recurse!
2961 = reportNoInstances tidy_env
2962 (Just (tci_loc inst, tci_given inst))
2965 check_overlap :: (InstEnv,InstEnv) -> Inst -> Either Inst SDoc
2966 -- Right msg => overlap message
2967 -- Left inst => no instance
2968 check_overlap inst_envs wanted
2969 | not (isClassDict wanted) = Left wanted
2971 = case lookupInstEnv inst_envs clas tys of
2972 ([], _) -> Left wanted -- No match
2973 -- The case of exactly one match and no unifiers means a
2974 -- successful lookup. That can't happen here, because dicts
2975 -- only end up here if they didn't match in Inst.lookupInst
2977 | debugIsOn -> pprPanic "reportNoInstance" (ppr wanted)
2978 res -> Right (mk_overlap_msg wanted res)
2980 (clas,tys) = getDictClassTys wanted
2982 mk_overlap_msg dict (matches, unifiers)
2983 = ASSERT( not (null matches) )
2984 vcat [ addInstLoc [dict] ((ptext (sLit "Overlapping instances for")
2985 <+> pprPred (dictPred dict))),
2986 sep [ptext (sLit "Matching instances") <> colon,
2987 nest 2 (vcat [pprInstances ispecs, pprInstances unifiers])],
2988 if not (isSingleton matches)
2989 then -- Two or more matches
2991 else -- One match, plus some unifiers
2992 ASSERT( not (null unifiers) )
2993 parens (vcat [ptext (sLit "The choice depends on the instantiation of") <+>
2994 quotes (pprWithCommas ppr (varSetElems (tyVarsOfInst dict))),
2995 ptext (sLit "To pick the first instance above, use -XIncoherentInstances"),
2996 ptext (sLit "when compiling the other instance declarations")])]
2998 ispecs = [ispec | (ispec, _) <- matches]
3000 mk_eq_err :: Inst -> (TidyEnv, SDoc)
3001 mk_eq_err inst = misMatchMsg tidy_env (eqInstTys inst)
3003 mk_no_inst_err insts
3004 | null insts = empty
3006 | Just (loc, givens) <- mb_what, -- Nested (type signatures, instance decls)
3007 not (isEmptyVarSet (tyVarsOfInsts insts))
3008 = vcat [ addInstLoc insts $
3009 sep [ ptext (sLit "Could not deduce") <+> pprDictsTheta insts
3010 , nest 2 $ ptext (sLit "from the context") <+> pprDictsTheta givens]
3011 , show_fixes (fix1 loc : fixes2) ]
3013 | otherwise -- Top level
3014 = vcat [ addInstLoc insts $
3015 ptext (sLit "No instance") <> plural insts
3016 <+> ptext (sLit "for") <+> pprDictsTheta insts
3017 , show_fixes fixes2 ]
3020 fix1 loc = sep [ ptext (sLit "add") <+> pprDictsTheta insts
3021 <+> ptext (sLit "to the context of"),
3022 nest 2 (ppr (instLocOrigin loc)) ]
3023 -- I'm not sure it helps to add the location
3024 -- nest 2 (ptext (sLit "at") <+> ppr (instLocSpan loc)) ]
3026 fixes2 | null instance_dicts = []
3027 | otherwise = [sep [ptext (sLit "add an instance declaration for"),
3028 pprDictsTheta instance_dicts]]
3029 instance_dicts = [d | d <- insts, isClassDict d, not (isTyVarDict d)]
3030 -- Insts for which it is worth suggesting an adding an instance declaration
3031 -- Exclude implicit parameters, and tyvar dicts
3033 show_fixes :: [SDoc] -> SDoc
3034 show_fixes [] = empty
3035 show_fixes (f:fs) = sep [ptext (sLit "Possible fix:"),
3036 nest 2 (vcat (f : map (ptext (sLit "or") <+>) fs))]
3038 addTopAmbigErrs :: [Inst] -> TcRn ()
3039 addTopAmbigErrs dicts
3040 -- Divide into groups that share a common set of ambiguous tyvars
3041 = ifErrsM (return ()) $ -- Only report ambiguity if no other errors happened
3042 -- See Note [Avoiding spurious errors]
3043 mapM_ report (equivClasses cmp [(d, tvs_of d) | d <- tidy_dicts])
3045 (tidy_env, tidy_dicts) = tidyInsts dicts
3047 tvs_of :: Inst -> [TcTyVar]
3048 tvs_of d = varSetElems (tyVarsOfInst d)
3049 cmp (_,tvs1) (_,tvs2) = tvs1 `compare` tvs2
3051 report :: [(Inst,[TcTyVar])] -> TcM ()
3052 report pairs@((inst,tvs) : _) = do -- The pairs share a common set of ambiguous tyvars
3053 (tidy_env, mono_msg) <- mkMonomorphismMsg tidy_env tvs
3054 setSrcSpan (instSpan inst) $
3055 -- the location of the first one will do for the err message
3056 addErrTcM (tidy_env, msg $$ mono_msg)
3058 dicts = map fst pairs
3059 msg = sep [text "Ambiguous type variable" <> plural tvs <+>
3060 pprQuotedList tvs <+> in_msg,
3061 nest 2 (pprDictsInFull dicts)]
3062 in_msg = text "in the constraint" <> plural dicts <> colon
3063 report [] = panic "addTopAmbigErrs"
3066 mkMonomorphismMsg :: TidyEnv -> [TcTyVar] -> TcM (TidyEnv, Message)
3067 -- There's an error with these Insts; if they have free type variables
3068 -- it's probably caused by the monomorphism restriction.
3069 -- Try to identify the offending variable
3070 -- ASSUMPTION: the Insts are fully zonked
3071 mkMonomorphismMsg tidy_env inst_tvs
3072 = do { dflags <- getDOpts
3073 ; (tidy_env, docs) <- findGlobals (mkVarSet inst_tvs) tidy_env
3074 ; return (tidy_env, mk_msg dflags docs) }
3076 mk_msg _ _ | any isRuntimeUnk inst_tvs
3077 = vcat [ptext (sLit "Cannot resolve unknown runtime types:") <+>
3078 (pprWithCommas ppr inst_tvs),
3079 ptext (sLit "Use :print or :force to determine these types")]
3080 mk_msg _ [] = ptext (sLit "Probable fix: add a type signature that fixes these type variable(s)")
3081 -- This happens in things like
3082 -- f x = show (read "foo")
3083 -- where monomorphism doesn't play any role
3085 = vcat [ptext (sLit "Possible cause: the monomorphism restriction applied to the following:"),
3087 monomorphism_fix dflags]
3089 monomorphism_fix :: DynFlags -> SDoc
3090 monomorphism_fix dflags
3091 = ptext (sLit "Probable fix:") <+> vcat
3092 [ptext (sLit "give these definition(s) an explicit type signature"),
3093 if dopt Opt_MonomorphismRestriction dflags
3094 then ptext (sLit "or use -XNoMonomorphismRestriction")
3095 else empty] -- Only suggest adding "-XNoMonomorphismRestriction"
3096 -- if it is not already set!
3098 warnDefault :: [(Inst, Class, Var)] -> Type -> TcM ()
3099 warnDefault ups default_ty = do
3100 warn_flag <- doptM Opt_WarnTypeDefaults
3101 addInstCtxt (instLoc (head (dicts))) (warnTc warn_flag warn_msg)
3103 dicts = [d | (d,_,_) <- ups]
3106 (_, tidy_dicts) = tidyInsts dicts
3107 warn_msg = vcat [ptext (sLit "Defaulting the following constraint(s) to type") <+>
3108 quotes (ppr default_ty),
3109 pprDictsInFull tidy_dicts]
3111 reduceDepthErr :: Int -> [Inst] -> SDoc
3112 reduceDepthErr n stack
3113 = vcat [ptext (sLit "Context reduction stack overflow; size =") <+> int n,
3114 ptext (sLit "Use -fcontext-stack=N to increase stack size to N"),
3115 nest 4 (pprStack stack)]
3117 pprStack :: [Inst] -> SDoc
3118 pprStack stack = vcat (map pprInstInFull stack)