2 % (c) The University of Glasgow 2006
3 % (c) The GRASP/AQUA Project, Glasgow University, 1992-1998
10 -- The above warning supression flag is a temporary kludge.
11 -- While working on this module you are encouraged to remove it and fix
12 -- any warnings in the module. See
13 -- http://hackage.haskell.org/trac/ghc/wiki/Commentary/CodingStyle#Warnings
17 tcSimplifyInfer, tcSimplifyInferCheck,
18 tcSimplifyCheck, tcSimplifyRestricted,
19 tcSimplifyRuleLhs, tcSimplifyIPs,
20 tcSimplifySuperClasses,
21 tcSimplifyTop, tcSimplifyInteractive,
22 tcSimplifyBracket, tcSimplifyCheckPat,
24 tcSimplifyDeriv, tcSimplifyDefault,
30 #include "HsVersions.h"
32 import {-# SOURCE #-} TcUnify( unifyType )
73 %************************************************************************
77 %************************************************************************
79 --------------------------------------
80 Notes on functional dependencies (a bug)
81 --------------------------------------
88 instance D a b => C a b -- Undecidable
89 -- (Not sure if it's crucial to this eg)
90 f :: C a b => a -> Bool
93 g :: C a b => a -> Bool
96 Here f typechecks, but g does not!! Reason: before doing improvement,
97 we reduce the (C a b1) constraint from the call of f to (D a b1).
99 Here is a more complicated example:
101 | > class Foo a b | a->b
103 | > class Bar a b | a->b
107 | > instance Bar Obj Obj
109 | > instance (Bar a b) => Foo a b
111 | > foo:: (Foo a b) => a -> String
114 | > runFoo:: (forall a b. (Foo a b) => a -> w) -> w
120 | Could not deduce (Bar a b) from the context (Foo a b)
121 | arising from use of `foo' at <interactive>:1
123 | Add (Bar a b) to the expected type of an expression
124 | In the first argument of `runFoo', namely `foo'
125 | In the definition of `it': it = runFoo foo
127 | Why all of the sudden does GHC need the constraint Bar a b? The
128 | function foo didn't ask for that...
130 The trouble is that to type (runFoo foo), GHC has to solve the problem:
132 Given constraint Foo a b
133 Solve constraint Foo a b'
135 Notice that b and b' aren't the same. To solve this, just do
136 improvement and then they are the same. But GHC currently does
141 That is usually fine, but it isn't here, because it sees that Foo a b is
142 not the same as Foo a b', and so instead applies the instance decl for
143 instance Bar a b => Foo a b. And that's where the Bar constraint comes
146 The Right Thing is to improve whenever the constraint set changes at
147 all. Not hard in principle, but it'll take a bit of fiddling to do.
149 Note [Choosing which variables to quantify]
150 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
151 Suppose we are about to do a generalisation step. We have in our hand
154 T the type of the RHS
155 C the constraints from that RHS
157 The game is to figure out
159 Q the set of type variables over which to quantify
160 Ct the constraints we will *not* quantify over
161 Cq the constraints we will quantify over
163 So we're going to infer the type
167 and float the constraints Ct further outwards.
169 Here are the things that *must* be true:
171 (A) Q intersect fv(G) = EMPTY limits how big Q can be
172 (B) Q superset fv(Cq union T) \ oclose(fv(G),C) limits how small Q can be
174 (A) says we can't quantify over a variable that's free in the environment.
175 (B) says we must quantify over all the truly free variables in T, else
176 we won't get a sufficiently general type.
178 We do not *need* to quantify over any variable that is fixed by the
179 free vars of the environment G.
181 BETWEEN THESE TWO BOUNDS, ANY Q WILL DO!
183 Example: class H x y | x->y where ...
185 fv(G) = {a} C = {H a b, H c d}
188 (A) Q intersect {a} is empty
189 (B) Q superset {a,b,c,d} \ oclose({a}, C) = {a,b,c,d} \ {a,b} = {c,d}
191 So Q can be {c,d}, {b,c,d}
193 In particular, it's perfectly OK to quantify over more type variables
194 than strictly necessary; there is no need to quantify over 'b', since
195 it is determined by 'a' which is free in the envt, but it's perfectly
196 OK to do so. However we must not quantify over 'a' itself.
198 Other things being equal, however, we'd like to quantify over as few
199 variables as possible: smaller types, fewer type applications, more
200 constraints can get into Ct instead of Cq. Here's a good way to
203 Q = grow( fv(T), C ) \ oclose( fv(G), C )
205 That is, quantify over all variable that that MIGHT be fixed by the
206 call site (which influences T), but which aren't DEFINITELY fixed by
207 G. This choice definitely quantifies over enough type variables,
208 albeit perhaps too many.
210 Why grow( fv(T), C ) rather than fv(T)? Consider
212 class H x y | x->y where ...
217 If we used fv(T) = {c} we'd get the type
219 forall c. H c d => c -> b
221 And then if the fn was called at several different c's, each of
222 which fixed d differently, we'd get a unification error, because
223 d isn't quantified. Solution: quantify d. So we must quantify
224 everything that might be influenced by c.
226 Why not oclose( fv(T), C )? Because we might not be able to see
227 all the functional dependencies yet:
229 class H x y | x->y where ...
230 instance H x y => Eq (T x y) where ...
235 Now oclose(fv(T),C) = {c}, because the functional dependency isn't
236 apparent yet, and that's wrong. We must really quantify over d too.
238 There really isn't any point in quantifying over any more than
239 grow( fv(T), C ), because the call sites can't possibly influence
240 any other type variables.
244 -------------------------------------
246 -------------------------------------
248 It's very hard to be certain when a type is ambiguous. Consider
252 instance H x y => K (x,y)
254 Is this type ambiguous?
255 forall a b. (K (a,b), Eq b) => a -> a
257 Looks like it! But if we simplify (K (a,b)) we get (H a b) and
258 now we see that a fixes b. So we can't tell about ambiguity for sure
259 without doing a full simplification. And even that isn't possible if
260 the context has some free vars that may get unified. Urgle!
262 Here's another example: is this ambiguous?
263 forall a b. Eq (T b) => a -> a
264 Not if there's an insance decl (with no context)
265 instance Eq (T b) where ...
267 You may say of this example that we should use the instance decl right
268 away, but you can't always do that:
270 class J a b where ...
271 instance J Int b where ...
273 f :: forall a b. J a b => a -> a
275 (Notice: no functional dependency in J's class decl.)
276 Here f's type is perfectly fine, provided f is only called at Int.
277 It's premature to complain when meeting f's signature, or even
278 when inferring a type for f.
282 However, we don't *need* to report ambiguity right away. It'll always
283 show up at the call site.... and eventually at main, which needs special
284 treatment. Nevertheless, reporting ambiguity promptly is an excellent thing.
286 So here's the plan. We WARN about probable ambiguity if
288 fv(Cq) is not a subset of oclose(fv(T) union fv(G), C)
290 (all tested before quantification).
291 That is, all the type variables in Cq must be fixed by the the variables
292 in the environment, or by the variables in the type.
294 Notice that we union before calling oclose. Here's an example:
296 class J a b c | a b -> c
300 forall b c. (J a b c) => b -> b
302 Only if we union {a} from G with {b} from T before using oclose,
303 do we see that c is fixed.
305 It's a bit vague exactly which C we should use for this oclose call. If we
306 don't fix enough variables we might complain when we shouldn't (see
307 the above nasty example). Nothing will be perfect. That's why we can
308 only issue a warning.
311 Can we ever be *certain* about ambiguity? Yes: if there's a constraint
313 c in C such that fv(c) intersect (fv(G) union fv(T)) = EMPTY
315 then c is a "bubble"; there's no way it can ever improve, and it's
316 certainly ambiguous. UNLESS it is a constant (sigh). And what about
321 instance H x y => K (x,y)
323 Is this type ambiguous?
324 forall a b. (K (a,b), Eq b) => a -> a
326 Urk. The (Eq b) looks "definitely ambiguous" but it isn't. What we are after
327 is a "bubble" that's a set of constraints
329 Cq = Ca union Cq' st fv(Ca) intersect (fv(Cq') union fv(T) union fv(G)) = EMPTY
331 Hence another idea. To decide Q start with fv(T) and grow it
332 by transitive closure in Cq (no functional dependencies involved).
333 Now partition Cq using Q, leaving the definitely-ambiguous and probably-ok.
334 The definitely-ambiguous can then float out, and get smashed at top level
335 (which squashes out the constants, like Eq (T a) above)
338 --------------------------------------
339 Notes on principal types
340 --------------------------------------
345 f x = let g y = op (y::Int) in True
347 Here the principal type of f is (forall a. a->a)
348 but we'll produce the non-principal type
349 f :: forall a. C Int => a -> a
352 --------------------------------------
353 The need for forall's in constraints
354 --------------------------------------
356 [Exchange on Haskell Cafe 5/6 Dec 2000]
358 class C t where op :: t -> Bool
359 instance C [t] where op x = True
361 p y = (let f :: c -> Bool; f x = op (y >> return x) in f, y ++ [])
362 q y = (y ++ [], let f :: c -> Bool; f x = op (y >> return x) in f)
364 The definitions of p and q differ only in the order of the components in
365 the pair on their right-hand sides. And yet:
367 ghc and "Typing Haskell in Haskell" reject p, but accept q;
368 Hugs rejects q, but accepts p;
369 hbc rejects both p and q;
370 nhc98 ... (Malcolm, can you fill in the blank for us!).
372 The type signature for f forces context reduction to take place, and
373 the results of this depend on whether or not the type of y is known,
374 which in turn depends on which component of the pair the type checker
377 Solution: if y::m a, float out the constraints
378 Monad m, forall c. C (m c)
379 When m is later unified with [], we can solve both constraints.
382 --------------------------------------
383 Notes on implicit parameters
384 --------------------------------------
386 Note [Inheriting implicit parameters]
387 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
392 where f is *not* a top-level binding.
393 From the RHS of f we'll get the constraint (?y::Int).
394 There are two types we might infer for f:
398 (so we get ?y from the context of f's definition), or
400 f :: (?y::Int) => Int -> Int
402 At first you might think the first was better, becuase then
403 ?y behaves like a free variable of the definition, rather than
404 having to be passed at each call site. But of course, the WHOLE
405 IDEA is that ?y should be passed at each call site (that's what
406 dynamic binding means) so we'd better infer the second.
408 BOTTOM LINE: when *inferring types* you *must* quantify
409 over implicit parameters. See the predicate isFreeWhenInferring.
412 Note [Implicit parameters and ambiguity]
413 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
414 Only a *class* predicate can give rise to ambiguity
415 An *implicit parameter* cannot. For example:
416 foo :: (?x :: [a]) => Int
418 is fine. The call site will suppply a particular 'x'
420 Furthermore, the type variables fixed by an implicit parameter
421 propagate to the others. E.g.
422 foo :: (Show a, ?x::[a]) => Int
424 The type of foo looks ambiguous. But it isn't, because at a call site
426 let ?x = 5::Int in foo
427 and all is well. In effect, implicit parameters are, well, parameters,
428 so we can take their type variables into account as part of the
429 "tau-tvs" stuff. This is done in the function 'FunDeps.grow'.
432 Question 2: type signatures
433 ~~~~~~~~~~~~~~~~~~~~~~~~~~~
434 BUT WATCH OUT: When you supply a type signature, we can't force you
435 to quantify over implicit parameters. For example:
439 This is perfectly reasonable. We do not want to insist on
441 (?x + 1) :: (?x::Int => Int)
443 That would be silly. Here, the definition site *is* the occurrence site,
444 so the above strictures don't apply. Hence the difference between
445 tcSimplifyCheck (which *does* allow implicit paramters to be inherited)
446 and tcSimplifyCheckBind (which does not).
448 What about when you supply a type signature for a binding?
449 Is it legal to give the following explicit, user type
450 signature to f, thus:
455 At first sight this seems reasonable, but it has the nasty property
456 that adding a type signature changes the dynamic semantics.
459 (let f x = (x::Int) + ?y
460 in (f 3, f 3 with ?y=5)) with ?y = 6
466 in (f 3, f 3 with ?y=5)) with ?y = 6
470 Indeed, simply inlining f (at the Haskell source level) would change the
473 Nevertheless, as Launchbury says (email Oct 01) we can't really give the
474 semantics for a Haskell program without knowing its typing, so if you
475 change the typing you may change the semantics.
477 To make things consistent in all cases where we are *checking* against
478 a supplied signature (as opposed to inferring a type), we adopt the
481 a signature does not need to quantify over implicit params.
483 [This represents a (rather marginal) change of policy since GHC 5.02,
484 which *required* an explicit signature to quantify over all implicit
485 params for the reasons mentioned above.]
487 But that raises a new question. Consider
489 Given (signature) ?x::Int
490 Wanted (inferred) ?x::Int, ?y::Bool
492 Clearly we want to discharge the ?x and float the ?y out. But
493 what is the criterion that distinguishes them? Clearly it isn't
494 what free type variables they have. The Right Thing seems to be
495 to float a constraint that
496 neither mentions any of the quantified type variables
497 nor any of the quantified implicit parameters
499 See the predicate isFreeWhenChecking.
502 Question 3: monomorphism
503 ~~~~~~~~~~~~~~~~~~~~~~~~
504 There's a nasty corner case when the monomorphism restriction bites:
508 The argument above suggests that we *must* generalise
509 over the ?y parameter, to get
510 z :: (?y::Int) => Int,
511 but the monomorphism restriction says that we *must not*, giving
513 Why does the momomorphism restriction say this? Because if you have
515 let z = x + ?y in z+z
517 you might not expect the addition to be done twice --- but it will if
518 we follow the argument of Question 2 and generalise over ?y.
521 Question 4: top level
522 ~~~~~~~~~~~~~~~~~~~~~
523 At the top level, monomorhism makes no sense at all.
526 main = let ?x = 5 in print foo
530 woggle :: (?x :: Int) => Int -> Int
533 We definitely don't want (foo :: Int) with a top-level implicit parameter
534 (?x::Int) becuase there is no way to bind it.
539 (A) Always generalise over implicit parameters
540 Bindings that fall under the monomorphism restriction can't
544 * Inlining remains valid
545 * No unexpected loss of sharing
546 * But simple bindings like
548 will be rejected, unless you add an explicit type signature
549 (to avoid the monomorphism restriction)
550 z :: (?y::Int) => Int
552 This seems unacceptable
554 (B) Monomorphism restriction "wins"
555 Bindings that fall under the monomorphism restriction can't
557 Always generalise over implicit parameters *except* for bindings
558 that fall under the monomorphism restriction
561 * Inlining isn't valid in general
562 * No unexpected loss of sharing
563 * Simple bindings like
565 accepted (get value of ?y from binding site)
567 (C) Always generalise over implicit parameters
568 Bindings that fall under the monomorphism restriction can't
569 be generalised, EXCEPT for implicit parameters
571 * Inlining remains valid
572 * Unexpected loss of sharing (from the extra generalisation)
573 * Simple bindings like
575 accepted (get value of ?y from occurrence sites)
580 None of these choices seems very satisfactory. But at least we should
581 decide which we want to do.
583 It's really not clear what is the Right Thing To Do. If you see
587 would you expect the value of ?y to be got from the *occurrence sites*
588 of 'z', or from the valuue of ?y at the *definition* of 'z'? In the
589 case of function definitions, the answer is clearly the former, but
590 less so in the case of non-fucntion definitions. On the other hand,
591 if we say that we get the value of ?y from the definition site of 'z',
592 then inlining 'z' might change the semantics of the program.
594 Choice (C) really says "the monomorphism restriction doesn't apply
595 to implicit parameters". Which is fine, but remember that every
596 innocent binding 'x = ...' that mentions an implicit parameter in
597 the RHS becomes a *function* of that parameter, called at each
598 use of 'x'. Now, the chances are that there are no intervening 'with'
599 clauses that bind ?y, so a decent compiler should common up all
600 those function calls. So I think I strongly favour (C). Indeed,
601 one could make a similar argument for abolishing the monomorphism
602 restriction altogether.
604 BOTTOM LINE: we choose (B) at present. See tcSimplifyRestricted
608 %************************************************************************
610 \subsection{tcSimplifyInfer}
612 %************************************************************************
614 tcSimplify is called when we *inferring* a type. Here's the overall game plan:
616 1. Compute Q = grow( fvs(T), C )
618 2. Partition C based on Q into Ct and Cq. Notice that ambiguous
619 predicates will end up in Ct; we deal with them at the top level
621 3. Try improvement, using functional dependencies
623 4. If Step 3 did any unification, repeat from step 1
624 (Unification can change the result of 'grow'.)
626 Note: we don't reduce dictionaries in step 2. For example, if we have
627 Eq (a,b), we don't simplify to (Eq a, Eq b). So Q won't be different
628 after step 2. However note that we may therefore quantify over more
629 type variables than we absolutely have to.
631 For the guts, we need a loop, that alternates context reduction and
632 improvement with unification. E.g. Suppose we have
634 class C x y | x->y where ...
636 and tcSimplify is called with:
638 Then improvement unifies a with b, giving
641 If we need to unify anything, we rattle round the whole thing all over
648 -> TcTyVarSet -- fv(T); type vars
650 -> TcM ([TcTyVar], -- Tyvars to quantify (zonked and quantified)
651 [Inst], -- Dict Ids that must be bound here (zonked)
652 TcDictBinds) -- Bindings
653 -- Any free (escaping) Insts are tossed into the environment
658 tcSimplifyInfer doc tau_tvs wanted
659 = do { tau_tvs1 <- zonkTcTyVarsAndFV (varSetElems tau_tvs)
660 ; wanted' <- mappM zonkInst wanted -- Zonk before deciding quantified tyvars
661 ; gbl_tvs <- tcGetGlobalTyVars
662 ; let preds1 = fdPredsOfInsts wanted'
663 gbl_tvs1 = oclose preds1 gbl_tvs
664 qtvs = grow preds1 tau_tvs1 `minusVarSet` gbl_tvs1
665 -- See Note [Choosing which variables to quantify]
667 -- To maximise sharing, remove from consideration any
668 -- constraints that don't mention qtvs at all
669 ; let (free, bound) = partition (isFreeWhenInferring qtvs) wanted'
672 -- To make types simple, reduce as much as possible
673 ; traceTc (text "infer" <+> (ppr preds1 $$ ppr (grow preds1 tau_tvs1) $$ ppr gbl_tvs $$
674 ppr gbl_tvs1 $$ ppr free $$ ppr bound))
675 ; (irreds1, binds1) <- tryHardCheckLoop doc bound
677 -- Note [Inference and implication constraints]
678 ; let want_dict d = tyVarsOfInst d `intersectsVarSet` qtvs
679 ; (irreds2, binds2) <- approximateImplications doc want_dict irreds1
681 -- Now work out all over again which type variables to quantify,
682 -- exactly in the same way as before, but starting from irreds2. Why?
683 -- a) By now improvment may have taken place, and we must *not*
684 -- quantify over any variable free in the environment
685 -- tc137 (function h inside g) is an example
687 -- b) Do not quantify over constraints that *now* do not
688 -- mention quantified type variables, because they are
689 -- simply ambiguous (or might be bound further out). Example:
690 -- f :: Eq b => a -> (a, b)
692 -- From the RHS of g we get the MethodInst f77 :: alpha -> (alpha, beta)
693 -- We decide to quantify over 'alpha' alone, but free1 does not include f77
694 -- because f77 mentions 'alpha'. Then reducing leaves only the (ambiguous)
695 -- constraint (Eq beta), which we dump back into the free set
696 -- See test tcfail181
698 -- c) irreds may contain type variables not previously mentioned,
699 -- e.g. instance D a x => Foo [a]
701 -- Then after simplifying we'll get (D a x), and x is fresh
702 -- We must quantify over x else it'll be totally unbound
703 ; tau_tvs2 <- zonkTcTyVarsAndFV (varSetElems tau_tvs1)
704 ; gbl_tvs2 <- zonkTcTyVarsAndFV (varSetElems gbl_tvs1)
705 -- Note that we start from gbl_tvs1
706 -- We use tcGetGlobalTyVars, then oclose wrt preds2, because
707 -- we've already put some of the original preds1 into frees
708 -- E.g. wanteds = C a b (where a->b)
711 -- Then b is fixed by gbl_tvs, so (C a b) will be in free, and
712 -- irreds2 will be empty. But we don't want to generalise over b!
713 ; let preds2 = fdPredsOfInsts irreds2 -- irreds2 is zonked
714 qtvs = grow preds2 tau_tvs2 `minusVarSet` oclose preds2 gbl_tvs2
715 ; let (free, irreds3) = partition (isFreeWhenInferring qtvs) irreds2
718 -- Turn the quantified meta-type variables into real type variables
719 ; qtvs2 <- zonkQuantifiedTyVars (varSetElems qtvs)
721 -- We can't abstract over any remaining unsolved
722 -- implications so instead just float them outwards. Ugh.
723 ; let (q_dicts0, implics) = partition isAbstractableInst irreds3
724 ; loc <- getInstLoc (ImplicOrigin doc)
725 ; implic_bind <- bindIrreds loc qtvs2 q_dicts0 implics
727 -- Prepare equality instances for quantification
728 ; let (q_eqs0,q_dicts) = partition isEqInst q_dicts0
729 ; q_eqs <- mappM finalizeEqInst q_eqs0
731 ; return (qtvs2, q_eqs ++ q_dicts, binds1 `unionBags` binds2 `unionBags` implic_bind) }
732 -- NB: when we are done, we might have some bindings, but
733 -- the final qtvs might be empty. See Note [NO TYVARS] below.
735 approximateImplications :: SDoc -> (Inst -> Bool) -> [Inst] -> TcM ([Inst], TcDictBinds)
736 -- Note [Inference and implication constraints]
737 -- Given a bunch of Dict and ImplicInsts, try to approximate the implications by
738 -- - fetching any dicts inside them that are free
739 -- - using those dicts as cruder constraints, to solve the implications
740 -- - returning the extra ones too
742 approximateImplications doc want_dict irreds
744 = return (irreds, emptyBag)
746 = do { extra_dicts' <- mapM cloneDict extra_dicts
747 ; tryHardCheckLoop doc (extra_dicts' ++ irreds) }
748 -- By adding extra_dicts', we make them
749 -- available to solve the implication constraints
751 extra_dicts = get_dicts (filter isImplicInst irreds)
753 get_dicts :: [Inst] -> [Inst] -- Returns only Dicts
754 -- Find the wanted constraints in implication constraints that satisfy
755 -- want_dict, and are not bound by forall's in the constraint itself
756 get_dicts ds = concatMap get_dict ds
758 get_dict d@(Dict {}) | want_dict d = [d]
760 get_dict (ImplicInst {tci_tyvars = tvs, tci_wanted = wanteds})
761 = [ d | let tv_set = mkVarSet tvs
762 , d <- get_dicts wanteds
763 , not (tyVarsOfInst d `intersectsVarSet` tv_set)]
764 get_dict i@(EqInst {}) | want_dict i = [i]
766 get_dict other = pprPanic "approximateImplications" (ppr other)
769 Note [Inference and implication constraints]
770 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
771 Suppose we have a wanted implication constraint (perhaps arising from
772 a nested pattern match) like
774 and we are now trying to quantify over 'a' when inferring the type for
775 a function. In principle it's possible that there might be an instance
776 instance (C a, E a) => D [a]
777 so the context (E a) would suffice. The Right Thing is to abstract over
778 the implication constraint, but we don't do that (a) because it'll be
779 surprising to programmers and (b) because we don't have the machinery to deal
780 with 'given' implications.
782 So our best approximation is to make (D [a]) part of the inferred
783 context, so we can use that to discharge the implication. Hence
784 the strange function get_dicts in approximateImplications.
786 The common cases are more clear-cut, when we have things like
788 Here, abstracting over (C b) is not an approximation at all -- but see
789 Note [Freeness and implications].
791 See Trac #1430 and test tc228.
795 -----------------------------------------------------------
796 -- tcSimplifyInferCheck is used when we know the constraints we are to simplify
797 -- against, but we don't know the type variables over which we are going to quantify.
798 -- This happens when we have a type signature for a mutually recursive group
801 -> TcTyVarSet -- fv(T)
804 -> TcM ([TyVar], -- Fully zonked, and quantified
805 TcDictBinds) -- Bindings
807 tcSimplifyInferCheck loc tau_tvs givens wanteds
808 = do { traceTc (text "tcSimplifyInferCheck <-" <+> ppr wanteds)
809 ; (irreds, binds) <- gentleCheckLoop loc givens wanteds
811 -- Figure out which type variables to quantify over
812 -- You might think it should just be the signature tyvars,
813 -- but in bizarre cases you can get extra ones
814 -- f :: forall a. Num a => a -> a
815 -- f x = fst (g (x, head [])) + 1
817 -- Here we infer g :: forall a b. a -> b -> (b,a)
818 -- We don't want g to be monomorphic in b just because
819 -- f isn't quantified over b.
820 ; let all_tvs = varSetElems (tau_tvs `unionVarSet` tyVarsOfInsts givens)
821 ; all_tvs <- zonkTcTyVarsAndFV all_tvs
822 ; gbl_tvs <- tcGetGlobalTyVars
823 ; let qtvs = varSetElems (all_tvs `minusVarSet` gbl_tvs)
824 -- We could close gbl_tvs, but its not necessary for
825 -- soundness, and it'll only affect which tyvars, not which
826 -- dictionaries, we quantify over
828 ; qtvs' <- zonkQuantifiedTyVars qtvs
830 -- Now we are back to normal (c.f. tcSimplCheck)
831 ; implic_bind <- bindIrreds loc qtvs' givens irreds
833 ; traceTc (text "tcSimplifyInferCheck ->" <+> ppr (implic_bind))
834 ; return (qtvs', binds `unionBags` implic_bind) }
837 Note [Squashing methods]
838 ~~~~~~~~~~~~~~~~~~~~~~~~~
839 Be careful if you want to float methods more:
840 truncate :: forall a. RealFrac a => forall b. Integral b => a -> b
841 From an application (truncate f i) we get
844 If we have also have a second occurrence of truncate, we get
847 When simplifying with i,f free, we might still notice that
848 t1=t3; but alas, the binding for t2 (which mentions t1)
849 may continue to float out!
854 class Y a b | a -> b where
857 instance Y [[a]] a where
860 k :: X a -> X a -> X a
862 g :: Num a => [X a] -> [X a]
865 h ys = ys ++ map (k (y [[0]])) xs
867 The excitement comes when simplifying the bindings for h. Initially
868 try to simplify {y @ [[t1]] t2, 0 @ t1}, with initial qtvs = {t2}.
869 From this we get t1:=:t2, but also various bindings. We can't forget
870 the bindings (because of [LOOP]), but in fact t1 is what g is
873 The net effect of [NO TYVARS]
876 isFreeWhenInferring :: TyVarSet -> Inst -> Bool
877 isFreeWhenInferring qtvs inst
878 = isFreeWrtTyVars qtvs inst -- Constrains no quantified vars
879 && isInheritableInst inst -- and no implicit parameter involved
880 -- see Note [Inheriting implicit parameters]
882 {- No longer used (with implication constraints)
883 isFreeWhenChecking :: TyVarSet -- Quantified tyvars
884 -> NameSet -- Quantified implicit parameters
886 isFreeWhenChecking qtvs ips inst
887 = isFreeWrtTyVars qtvs inst
888 && isFreeWrtIPs ips inst
891 isFreeWrtTyVars qtvs inst = tyVarsOfInst inst `disjointVarSet` qtvs
892 isFreeWrtIPs ips inst = not (any (`elemNameSet` ips) (ipNamesOfInst inst))
896 %************************************************************************
898 \subsection{tcSimplifyCheck}
900 %************************************************************************
902 @tcSimplifyCheck@ is used when we know exactly the set of variables
903 we are going to quantify over. For example, a class or instance declaration.
906 -----------------------------------------------------------
907 -- tcSimplifyCheck is used when checking expression type signatures,
908 -- class decls, instance decls etc.
909 tcSimplifyCheck :: InstLoc
910 -> [TcTyVar] -- Quantify over these
913 -> TcM TcDictBinds -- Bindings
914 tcSimplifyCheck loc qtvs givens wanteds
915 = ASSERT( all isTcTyVar qtvs && all isSkolemTyVar qtvs )
916 do { traceTc (text "tcSimplifyCheck")
917 ; (irreds, binds) <- gentleCheckLoop loc givens wanteds
918 ; implic_bind <- bindIrreds loc qtvs givens irreds
919 ; return (binds `unionBags` implic_bind) }
921 -----------------------------------------------------------
922 -- tcSimplifyCheckPat is used for existential pattern match
923 tcSimplifyCheckPat :: InstLoc
924 -> [CoVar] -> Refinement
925 -> [TcTyVar] -- Quantify over these
928 -> TcM TcDictBinds -- Bindings
929 tcSimplifyCheckPat loc co_vars reft qtvs givens wanteds
930 = ASSERT( all isTcTyVar qtvs && all isSkolemTyVar qtvs )
931 do { traceTc (text "tcSimplifyCheckPat")
932 ; (irreds, binds) <- gentleCheckLoop loc givens wanteds
933 ; implic_bind <- bindIrredsR loc qtvs co_vars reft
935 ; return (binds `unionBags` implic_bind) }
937 -----------------------------------------------------------
938 bindIrreds :: InstLoc -> [TcTyVar]
941 bindIrreds loc qtvs givens irreds
942 = bindIrredsR loc qtvs [] emptyRefinement givens irreds
944 bindIrredsR :: InstLoc -> [TcTyVar] -> [CoVar]
945 -> Refinement -> [Inst] -> [Inst]
947 -- Make a binding that binds 'irreds', by generating an implication
948 -- constraint for them, *and* throwing the constraint into the LIE
949 bindIrredsR loc qtvs co_vars reft givens irreds
953 = do { let givens' = filter isAbstractableInst givens
954 -- The givens can (redundantly) include methods
955 -- We want to retain both EqInsts and Dicts
956 -- There should be no implicadtion constraints
957 -- See Note [Pruning the givens in an implication constraint]
959 -- If there are no 'givens' *and* the refinement is empty
960 -- (the refinement is like more givens), then it's safe to
961 -- partition the 'wanteds' by their qtvs, thereby trimming irreds
962 -- See Note [Freeness and implications]
963 ; irreds' <- if null givens' && isEmptyRefinement reft
965 { let qtv_set = mkVarSet qtvs
966 (frees, real_irreds) = partition (isFreeWrtTyVars qtv_set) irreds
968 ; return real_irreds }
971 ; let all_tvs = qtvs ++ co_vars -- Abstract over all these
972 ; (implics, bind) <- makeImplicationBind loc all_tvs reft givens' irreds'
973 -- This call does the real work
974 -- If irreds' is empty, it does something sensible
979 makeImplicationBind :: InstLoc -> [TcTyVar] -> Refinement
981 -> TcM ([Inst], TcDictBinds)
982 -- Make a binding that binds 'irreds', by generating an implication
983 -- constraint for them, *and* throwing the constraint into the LIE
984 -- The binding looks like
985 -- (ir1, .., irn) = f qtvs givens
986 -- where f is (evidence for) the new implication constraint
987 -- f :: forall qtvs. {reft} givens => (ir1, .., irn)
988 -- qtvs includes coercion variables
990 -- This binding must line up the 'rhs' in reduceImplication
991 makeImplicationBind loc all_tvs reft
992 givens -- Guaranteed all Dicts
995 | null irreds -- If there are no irreds, we are done
996 = return ([], emptyBag)
997 | otherwise -- Otherwise we must generate a binding
998 = do { uniq <- newUnique
999 ; span <- getSrcSpanM
1000 ; let (eq_givens, dict_givens) = partition isEqInst givens
1001 eq_tyvar_cos = mkTyVarTys (varSetElems $ tyVarsOfTypes $ map eqInstType eq_givens)
1002 -- Urgh! See line 2187 or thereabouts. I believe that all these
1003 -- 'givens' must be a simple CoVar. This MUST be cleaned up.
1005 ; let name = mkInternalName uniq (mkVarOcc "ic") span
1006 implic_inst = ImplicInst { tci_name = name, tci_reft = reft,
1007 tci_tyvars = all_tvs,
1008 tci_given = (eq_givens ++ dict_givens),
1009 tci_wanted = irreds, tci_loc = loc }
1010 ; let -- only create binder for dict_irreds
1011 (eq_irreds, dict_irreds) = partition isEqInst irreds
1012 n_dict_irreds = length dict_irreds
1013 dict_irred_ids = map instToId dict_irreds
1014 tup_ty = mkTupleTy Boxed n_dict_irreds (map idType dict_irred_ids)
1015 pat = TuplePat (map nlVarPat dict_irred_ids) Boxed tup_ty
1016 rhs = L span (mkHsWrap co (HsVar (instToId implic_inst)))
1017 co = mkWpApps (map instToId dict_givens) <.> mkWpTyApps eq_tyvar_cos <.> mkWpTyApps (mkTyVarTys all_tvs)
1018 bind | [dict_irred_id] <- dict_irred_ids = VarBind dict_irred_id rhs
1019 | otherwise = PatBind { pat_lhs = L span pat,
1020 pat_rhs = unguardedGRHSs rhs,
1021 pat_rhs_ty = tup_ty,
1022 bind_fvs = placeHolderNames }
1023 ; -- pprTrace "Make implic inst" (ppr (implic_inst,irreds,dict_irreds,tup_ty)) $
1024 return ([implic_inst], unitBag (L span bind)) }
1026 -----------------------------------------------------------
1027 tryHardCheckLoop :: SDoc
1029 -> TcM ([Inst], TcDictBinds)
1031 tryHardCheckLoop doc wanteds
1032 = do { (irreds,binds,_) <- checkLoop (mkRedEnv doc try_me []) wanteds
1033 ; return (irreds,binds)
1036 try_me inst = ReduceMe AddSCs
1037 -- Here's the try-hard bit
1039 -----------------------------------------------------------
1040 gentleCheckLoop :: InstLoc
1043 -> TcM ([Inst], TcDictBinds)
1045 gentleCheckLoop inst_loc givens wanteds
1046 = do { (irreds,binds,_) <- checkLoop env wanteds
1047 ; return (irreds,binds)
1050 env = mkRedEnv (pprInstLoc inst_loc) try_me givens
1052 try_me inst | isMethodOrLit inst = ReduceMe AddSCs
1054 -- When checking against a given signature
1055 -- we MUST be very gentle: Note [Check gently]
1057 gentleInferLoop :: SDoc -> [Inst]
1058 -> TcM ([Inst], TcDictBinds)
1059 gentleInferLoop doc wanteds
1060 = do { (irreds, binds, _) <- checkLoop env wanteds
1061 ; return (irreds, binds) }
1063 env = mkRedEnv doc try_me []
1064 try_me inst | isMethodOrLit inst = ReduceMe AddSCs
1069 ~~~~~~~~~~~~~~~~~~~~
1070 We have to very careful about not simplifying too vigorously
1075 f :: Show b => T b -> b
1076 f (MkT x) = show [x]
1078 Inside the pattern match, which binds (a:*, x:a), we know that
1080 Hence we have a dictionary for Show [a] available; and indeed we
1081 need it. We are going to build an implication contraint
1082 forall a. (b~[a]) => Show [a]
1083 Later, we will solve this constraint using the knowledge (Show b)
1085 But we MUST NOT reduce (Show [a]) to (Show a), else the whole
1086 thing becomes insoluble. So we simplify gently (get rid of literals
1087 and methods only, plus common up equal things), deferring the real
1088 work until top level, when we solve the implication constraint
1089 with tryHardCheckLooop.
1093 -----------------------------------------------------------
1096 -> TcM ([Inst], TcDictBinds,
1097 [Inst]) -- needed givens
1098 -- Precondition: givens are completely rigid
1099 -- Postcondition: returned Insts are zonked
1101 checkLoop env wanteds
1103 where go env wanteds needed_givens
1104 = do { -- We do need to zonk the givens; cf Note [Zonking RedEnv]
1105 ; env' <- zonkRedEnv env
1106 ; wanteds' <- zonkInsts wanteds
1108 ; (improved, binds, irreds, more_needed_givens) <- reduceContext env' wanteds'
1110 ; let all_needed_givens = needed_givens ++ more_needed_givens
1112 ; if not improved then
1113 return (irreds, binds, all_needed_givens)
1116 -- If improvement did some unification, we go round again.
1117 -- We start again with irreds, not wanteds
1118 -- Using an instance decl might have introduced a fresh type variable
1119 -- which might have been unified, so we'd get an infinite loop
1120 -- if we started again with wanteds! See Note [LOOP]
1121 { (irreds1, binds1, all_needed_givens1) <- go env' irreds all_needed_givens
1122 ; return (irreds1, binds `unionBags` binds1, all_needed_givens1) } }
1125 Note [Zonking RedEnv]
1126 ~~~~~~~~~~~~~~~~~~~~~
1127 It might appear as if the givens in RedEnv are always rigid, but that is not
1128 necessarily the case for programs involving higher-rank types that have class
1129 contexts constraining the higher-rank variables. An example from tc237 in the
1132 class Modular s a | s -> a
1134 wim :: forall a w. Integral a
1135 => a -> (forall s. Modular s a => M s w) -> w
1136 wim i k = error "urk"
1138 test5 :: (Modular s a, Integral a) => M s a
1141 test4 = wim 4 test4'
1143 Notice how the variable 'a' of (Modular s a) in the rank-2 type of wim is
1144 quantified further outside. When type checking test4, we have to check
1145 whether the signature of test5 is an instance of
1147 (forall s. Modular s a => M s w)
1149 Consequently, we will get (Modular s t_a), where t_a is a TauTv into the
1152 Given the FD of Modular in this example, class improvement will instantiate
1153 t_a to 'a', where 'a' is the skolem from test5's signatures (due to the
1154 Modular s a predicate in that signature). If we don't zonk (Modular s t_a) in
1155 the givens, we will get into a loop as improveOne uses the unification engine
1156 TcGadt.tcUnifyTys, which doesn't know about mutable type variables.
1161 class If b t e r | b t e -> r
1164 class Lte a b c | a b -> c where lte :: a -> b -> c
1166 instance (Lte a b l,If l b a c) => Max a b c
1168 Wanted: Max Z (S x) y
1170 Then we'll reduce using the Max instance to:
1171 (Lte Z (S x) l, If l (S x) Z y)
1172 and improve by binding l->T, after which we can do some reduction
1173 on both the Lte and If constraints. What we *can't* do is start again
1174 with (Max Z (S x) y)!
1178 %************************************************************************
1180 tcSimplifySuperClasses
1182 %************************************************************************
1184 Note [SUPERCLASS-LOOP 1]
1185 ~~~~~~~~~~~~~~~~~~~~~~~~
1186 We have to be very, very careful when generating superclasses, lest we
1187 accidentally build a loop. Here's an example:
1191 class S a => C a where { opc :: a -> a }
1192 class S b => D b where { opd :: b -> b }
1194 instance C Int where
1197 instance D Int where
1200 From (instance C Int) we get the constraint set {ds1:S Int, dd:D Int}
1201 Simplifying, we may well get:
1202 $dfCInt = :C ds1 (opd dd)
1205 Notice that we spot that we can extract ds1 from dd.
1207 Alas! Alack! We can do the same for (instance D Int):
1209 $dfDInt = :D ds2 (opc dc)
1213 And now we've defined the superclass in terms of itself.
1215 Solution: never generate a superclass selectors at all when
1216 satisfying the superclass context of an instance declaration.
1218 Two more nasty cases are in
1223 tcSimplifySuperClasses
1228 tcSimplifySuperClasses loc givens sc_wanteds
1229 = do { traceTc (text "tcSimplifySuperClasses")
1230 ; (irreds,binds1,_) <- checkLoop env sc_wanteds
1231 ; let (tidy_env, tidy_irreds) = tidyInsts irreds
1232 ; reportNoInstances tidy_env (Just (loc, givens)) tidy_irreds
1235 env = mkRedEnv (pprInstLoc loc) try_me givens
1236 try_me inst = ReduceMe NoSCs
1237 -- Like tryHardCheckLoop, but with NoSCs
1241 %************************************************************************
1243 \subsection{tcSimplifyRestricted}
1245 %************************************************************************
1247 tcSimplifyRestricted infers which type variables to quantify for a
1248 group of restricted bindings. This isn't trivial.
1251 We want to quantify over a to get id :: forall a. a->a
1254 We do not want to quantify over a, because there's an Eq a
1255 constraint, so we get eq :: a->a->Bool (notice no forall)
1258 RHS has type 'tau', whose free tyvars are tau_tvs
1259 RHS has constraints 'wanteds'
1262 Quantify over (tau_tvs \ ftvs(wanteds))
1263 This is bad. The constraints may contain (Monad (ST s))
1264 where we have instance Monad (ST s) where...
1265 so there's no need to be monomorphic in s!
1267 Also the constraint might be a method constraint,
1268 whose type mentions a perfectly innocent tyvar:
1269 op :: Num a => a -> b -> a
1270 Here, b is unconstrained. A good example would be
1272 We want to infer the polymorphic type
1273 foo :: forall b. b -> b
1276 Plan B (cunning, used for a long time up to and including GHC 6.2)
1277 Step 1: Simplify the constraints as much as possible (to deal
1278 with Plan A's problem). Then set
1279 qtvs = tau_tvs \ ftvs( simplify( wanteds ) )
1281 Step 2: Now simplify again, treating the constraint as 'free' if
1282 it does not mention qtvs, and trying to reduce it otherwise.
1283 The reasons for this is to maximise sharing.
1285 This fails for a very subtle reason. Suppose that in the Step 2
1286 a constraint (Foo (Succ Zero) (Succ Zero) b) gets thrown upstairs as 'free'.
1287 In the Step 1 this constraint might have been simplified, perhaps to
1288 (Foo Zero Zero b), AND THEN THAT MIGHT BE IMPROVED, to bind 'b' to 'T'.
1289 This won't happen in Step 2... but that in turn might prevent some other
1290 constraint (Baz [a] b) being simplified (e.g. via instance Baz [a] T where {..})
1291 and that in turn breaks the invariant that no constraints are quantified over.
1293 Test typecheck/should_compile/tc177 (which failed in GHC 6.2) demonstrates
1298 Step 1: Simplify the constraints as much as possible (to deal
1299 with Plan A's problem). Then set
1300 qtvs = tau_tvs \ ftvs( simplify( wanteds ) )
1301 Return the bindings from Step 1.
1304 A note about Plan C (arising from "bug" reported by George Russel March 2004)
1307 instance (HasBinary ty IO) => HasCodedValue ty
1309 foo :: HasCodedValue a => String -> IO a
1311 doDecodeIO :: HasCodedValue a => () -> () -> IO a
1312 doDecodeIO codedValue view
1313 = let { act = foo "foo" } in act
1315 You might think this should work becuase the call to foo gives rise to a constraint
1316 (HasCodedValue t), which can be satisfied by the type sig for doDecodeIO. But the
1317 restricted binding act = ... calls tcSimplifyRestricted, and PlanC simplifies the
1318 constraint using the (rather bogus) instance declaration, and now we are stuffed.
1320 I claim this is not really a bug -- but it bit Sergey as well as George. So here's
1324 Plan D (a variant of plan B)
1325 Step 1: Simplify the constraints as much as possible (to deal
1326 with Plan A's problem), BUT DO NO IMPROVEMENT. Then set
1327 qtvs = tau_tvs \ ftvs( simplify( wanteds ) )
1329 Step 2: Now simplify again, treating the constraint as 'free' if
1330 it does not mention qtvs, and trying to reduce it otherwise.
1332 The point here is that it's generally OK to have too few qtvs; that is,
1333 to make the thing more monomorphic than it could be. We don't want to
1334 do that in the common cases, but in wierd cases it's ok: the programmer
1335 can always add a signature.
1337 Too few qtvs => too many wanteds, which is what happens if you do less
1342 tcSimplifyRestricted -- Used for restricted binding groups
1343 -- i.e. ones subject to the monomorphism restriction
1346 -> [Name] -- Things bound in this group
1347 -> TcTyVarSet -- Free in the type of the RHSs
1348 -> [Inst] -- Free in the RHSs
1349 -> TcM ([TyVar], -- Tyvars to quantify (zonked and quantified)
1350 TcDictBinds) -- Bindings
1351 -- tcSimpifyRestricted returns no constraints to
1352 -- quantify over; by definition there are none.
1353 -- They are all thrown back in the LIE
1355 tcSimplifyRestricted doc top_lvl bndrs tau_tvs wanteds
1356 -- Zonk everything in sight
1357 = do { traceTc (text "tcSimplifyRestricted")
1358 ; wanteds' <- zonkInsts wanteds
1360 -- 'ReduceMe': Reduce as far as we can. Don't stop at
1361 -- dicts; the idea is to get rid of as many type
1362 -- variables as possible, and we don't want to stop
1363 -- at (say) Monad (ST s), because that reduces
1364 -- immediately, with no constraint on s.
1366 -- BUT do no improvement! See Plan D above
1367 -- HOWEVER, some unification may take place, if we instantiate
1368 -- a method Inst with an equality constraint
1369 ; let env = mkNoImproveRedEnv doc (\i -> ReduceMe AddSCs)
1370 ; (_imp, _binds, constrained_dicts, _) <- reduceContext env wanteds'
1372 -- Next, figure out the tyvars we will quantify over
1373 ; tau_tvs' <- zonkTcTyVarsAndFV (varSetElems tau_tvs)
1374 ; gbl_tvs' <- tcGetGlobalTyVars
1375 ; constrained_dicts' <- zonkInsts constrained_dicts
1377 ; let qtvs1 = tau_tvs' `minusVarSet` oclose (fdPredsOfInsts constrained_dicts) gbl_tvs'
1378 -- As in tcSimplifyInfer
1380 -- Do not quantify over constrained type variables:
1381 -- this is the monomorphism restriction
1382 constrained_tvs' = tyVarsOfInsts constrained_dicts'
1383 qtvs = qtvs1 `minusVarSet` constrained_tvs'
1384 pp_bndrs = pprWithCommas (quotes . ppr) bndrs
1387 ; warn_mono <- doptM Opt_WarnMonomorphism
1388 ; warnTc (warn_mono && (constrained_tvs' `intersectsVarSet` qtvs1))
1389 (vcat[ ptext SLIT("the Monomorphism Restriction applies to the binding")
1390 <> plural bndrs <+> ptext SLIT("for") <+> pp_bndrs,
1391 ptext SLIT("Consider giving a type signature for") <+> pp_bndrs])
1393 ; traceTc (text "tcSimplifyRestricted" <+> vcat [
1394 pprInsts wanteds, pprInsts constrained_dicts',
1396 ppr constrained_tvs', ppr tau_tvs', ppr qtvs ])
1398 -- The first step may have squashed more methods than
1399 -- necessary, so try again, this time more gently, knowing the exact
1400 -- set of type variables to quantify over.
1402 -- We quantify only over constraints that are captured by qtvs;
1403 -- these will just be a subset of non-dicts. This in contrast
1404 -- to normal inference (using isFreeWhenInferring) in which we quantify over
1405 -- all *non-inheritable* constraints too. This implements choice
1406 -- (B) under "implicit parameter and monomorphism" above.
1408 -- Remember that we may need to do *some* simplification, to
1409 -- (for example) squash {Monad (ST s)} into {}. It's not enough
1410 -- just to float all constraints
1412 -- At top level, we *do* squash methods becuase we want to
1413 -- expose implicit parameters to the test that follows
1414 ; let is_nested_group = isNotTopLevel top_lvl
1415 try_me inst | isFreeWrtTyVars qtvs inst,
1416 (is_nested_group || isDict inst) = Stop
1417 | otherwise = ReduceMe AddSCs
1418 env = mkNoImproveRedEnv doc try_me
1419 ; (_imp, binds, irreds, _) <- reduceContext env wanteds'
1421 -- See "Notes on implicit parameters, Question 4: top level"
1422 ; ASSERT( all (isFreeWrtTyVars qtvs) irreds ) -- None should be captured
1423 if is_nested_group then
1425 else do { let (bad_ips, non_ips) = partition isIPDict irreds
1426 ; addTopIPErrs bndrs bad_ips
1427 ; extendLIEs non_ips }
1429 ; qtvs' <- zonkQuantifiedTyVars (varSetElems qtvs)
1430 ; return (qtvs', binds) }
1434 %************************************************************************
1438 %************************************************************************
1440 On the LHS of transformation rules we only simplify methods and constants,
1441 getting dictionaries. We want to keep all of them unsimplified, to serve
1442 as the available stuff for the RHS of the rule.
1444 Example. Consider the following left-hand side of a rule
1446 f (x == y) (y > z) = ...
1448 If we typecheck this expression we get constraints
1450 d1 :: Ord a, d2 :: Eq a
1452 We do NOT want to "simplify" to the LHS
1454 forall x::a, y::a, z::a, d1::Ord a.
1455 f ((==) (eqFromOrd d1) x y) ((>) d1 y z) = ...
1459 forall x::a, y::a, z::a, d1::Ord a, d2::Eq a.
1460 f ((==) d2 x y) ((>) d1 y z) = ...
1462 Here is another example:
1464 fromIntegral :: (Integral a, Num b) => a -> b
1465 {-# RULES "foo" fromIntegral = id :: Int -> Int #-}
1467 In the rule, a=b=Int, and Num Int is a superclass of Integral Int. But
1468 we *dont* want to get
1470 forall dIntegralInt.
1471 fromIntegral Int Int dIntegralInt (scsel dIntegralInt) = id Int
1473 because the scsel will mess up RULE matching. Instead we want
1475 forall dIntegralInt, dNumInt.
1476 fromIntegral Int Int dIntegralInt dNumInt = id Int
1480 g (x == y) (y == z) = ..
1482 where the two dictionaries are *identical*, we do NOT WANT
1484 forall x::a, y::a, z::a, d1::Eq a
1485 f ((==) d1 x y) ((>) d1 y z) = ...
1487 because that will only match if the dict args are (visibly) equal.
1488 Instead we want to quantify over the dictionaries separately.
1490 In short, tcSimplifyRuleLhs must *only* squash LitInst and MethInts, leaving
1491 all dicts unchanged, with absolutely no sharing. It's simpler to do this
1492 from scratch, rather than further parameterise simpleReduceLoop etc
1495 tcSimplifyRuleLhs :: [Inst] -> TcM ([Inst], TcDictBinds)
1496 tcSimplifyRuleLhs wanteds
1497 = go [] emptyBag wanteds
1500 = return (dicts, binds)
1501 go dicts binds (w:ws)
1503 = go (w:dicts) binds ws
1505 = do { w' <- zonkInst w -- So that (3::Int) does not generate a call
1506 -- to fromInteger; this looks fragile to me
1507 ; lookup_result <- lookupSimpleInst w'
1508 ; case lookup_result of
1510 go dicts (addInstToDictBind binds w rhs) (ws' ++ ws)
1511 NoInstance -> pprPanic "tcSimplifyRuleLhs" (ppr w)
1515 tcSimplifyBracket is used when simplifying the constraints arising from
1516 a Template Haskell bracket [| ... |]. We want to check that there aren't
1517 any constraints that can't be satisfied (e.g. Show Foo, where Foo has no
1518 Show instance), but we aren't otherwise interested in the results.
1519 Nor do we care about ambiguous dictionaries etc. We will type check
1520 this bracket again at its usage site.
1523 tcSimplifyBracket :: [Inst] -> TcM ()
1524 tcSimplifyBracket wanteds
1525 = do { tryHardCheckLoop doc wanteds
1528 doc = text "tcSimplifyBracket"
1532 %************************************************************************
1534 \subsection{Filtering at a dynamic binding}
1536 %************************************************************************
1541 we must discharge all the ?x constraints from B. We also do an improvement
1542 step; if we have ?x::t1 and ?x::t2 we must unify t1, t2.
1544 Actually, the constraints from B might improve the types in ?x. For example
1546 f :: (?x::Int) => Char -> Char
1549 then the constraint (?x::Int) arising from the call to f will
1550 force the binding for ?x to be of type Int.
1553 tcSimplifyIPs :: [Inst] -- The implicit parameters bound here
1556 -- We need a loop so that we do improvement, and then
1557 -- (next time round) generate a binding to connect the two
1559 -- Here the two ?x's have different types, and improvement
1560 -- makes them the same.
1562 tcSimplifyIPs given_ips wanteds
1563 = do { wanteds' <- zonkInsts wanteds
1564 ; given_ips' <- zonkInsts given_ips
1565 -- Unusually for checking, we *must* zonk the given_ips
1567 ; let env = mkRedEnv doc try_me given_ips'
1568 ; (improved, binds, irreds, _) <- reduceContext env wanteds'
1570 ; if not improved then
1571 ASSERT( all is_free irreds )
1572 do { extendLIEs irreds
1575 tcSimplifyIPs given_ips wanteds }
1577 doc = text "tcSimplifyIPs" <+> ppr given_ips
1578 ip_set = mkNameSet (ipNamesOfInsts given_ips)
1579 is_free inst = isFreeWrtIPs ip_set inst
1581 -- Simplify any methods that mention the implicit parameter
1582 try_me inst | is_free inst = Stop
1583 | otherwise = ReduceMe NoSCs
1587 %************************************************************************
1589 \subsection[binds-for-local-funs]{@bindInstsOfLocalFuns@}
1591 %************************************************************************
1593 When doing a binding group, we may have @Insts@ of local functions.
1594 For example, we might have...
1596 let f x = x + 1 -- orig local function (overloaded)
1597 f.1 = f Int -- two instances of f
1602 The point is: we must drop the bindings for @f.1@ and @f.2@ here,
1603 where @f@ is in scope; those @Insts@ must certainly not be passed
1604 upwards towards the top-level. If the @Insts@ were binding-ified up
1605 there, they would have unresolvable references to @f@.
1607 We pass in an @init_lie@ of @Insts@ and a list of locally-bound @Ids@.
1608 For each method @Inst@ in the @init_lie@ that mentions one of the
1609 @Ids@, we create a binding. We return the remaining @Insts@ (in an
1610 @LIE@), as well as the @HsBinds@ generated.
1613 bindInstsOfLocalFuns :: [Inst] -> [TcId] -> TcM TcDictBinds
1614 -- Simlifies only MethodInsts, and generate only bindings of form
1616 -- We're careful not to even generate bindings of the form
1618 -- You'd think that'd be fine, but it interacts with what is
1619 -- arguably a bug in Match.tidyEqnInfo (see notes there)
1621 bindInstsOfLocalFuns wanteds local_ids
1622 | null overloaded_ids
1624 = extendLIEs wanteds `thenM_`
1625 returnM emptyLHsBinds
1628 = do { (irreds, binds) <- gentleInferLoop doc for_me
1629 ; extendLIEs not_for_me
1633 doc = text "bindInsts" <+> ppr local_ids
1634 overloaded_ids = filter is_overloaded local_ids
1635 is_overloaded id = isOverloadedTy (idType id)
1636 (for_me, not_for_me) = partition (isMethodFor overloaded_set) wanteds
1638 overloaded_set = mkVarSet overloaded_ids -- There can occasionally be a lot of them
1639 -- so it's worth building a set, so that
1640 -- lookup (in isMethodFor) is faster
1644 %************************************************************************
1646 \subsection{Data types for the reduction mechanism}
1648 %************************************************************************
1650 The main control over context reduction is here
1654 = RedEnv { red_doc :: SDoc -- The context
1655 , red_try_me :: Inst -> WhatToDo
1656 , red_improve :: Bool -- True <=> do improvement
1657 , red_givens :: [Inst] -- All guaranteed rigid
1659 -- but see Note [Rigidity]
1660 , red_reft :: Refinement -- The refinement to apply to the 'givens'
1661 -- You should think of it as 'given equalities'
1662 , red_stack :: (Int, [Inst]) -- Recursion stack (for err msg)
1663 -- See Note [RedStack]
1667 -- The red_givens are rigid so far as cmpInst is concerned.
1668 -- There is one case where they are not totally rigid, namely in tcSimplifyIPs
1669 -- let ?x = e in ...
1670 -- Here, the given is (?x::a), where 'a' is not necy a rigid type
1671 -- But that doesn't affect the comparison, which is based only on mame.
1674 -- The red_stack pair (n,insts) pair is just used for error reporting.
1675 -- 'n' is always the depth of the stack.
1676 -- The 'insts' is the stack of Insts being reduced: to produce X
1677 -- I had to produce Y, to produce Y I had to produce Z, and so on.
1680 mkRedEnv :: SDoc -> (Inst -> WhatToDo) -> [Inst] -> RedEnv
1681 mkRedEnv doc try_me givens
1682 = RedEnv { red_doc = doc, red_try_me = try_me,
1683 red_givens = givens,
1684 red_reft = emptyRefinement,
1686 red_improve = True }
1688 mkNoImproveRedEnv :: SDoc -> (Inst -> WhatToDo) -> RedEnv
1689 -- Do not do improvement; no givens
1690 mkNoImproveRedEnv doc try_me
1691 = RedEnv { red_doc = doc, red_try_me = try_me,
1692 red_givens = [], red_reft = emptyRefinement,
1694 red_improve = True }
1697 = ReduceMe WantSCs -- Try to reduce this
1698 -- If there's no instance, add the inst to the
1699 -- irreductible ones, but don't produce an error
1700 -- message of any kind.
1701 -- It might be quite legitimate such as (Eq a)!
1703 | Stop -- Return as irreducible unless it can
1704 -- be reduced to a constant in one step
1705 -- Do not add superclasses; see
1707 data WantSCs = NoSCs | AddSCs -- Tells whether we should add the superclasses
1708 -- of a predicate when adding it to the avails
1709 -- The reason for this flag is entirely the super-class loop problem
1710 -- Note [SUPER-CLASS LOOP 1]
1712 zonkRedEnv :: RedEnv -> TcM RedEnv
1714 = do { givens' <- mappM zonkInst (red_givens env)
1715 ; return $ env {red_givens = givens'}
1720 %************************************************************************
1722 \subsection[reduce]{@reduce@}
1724 %************************************************************************
1726 Note [Ancestor Equalities]
1727 ~~~~~~~~~~~~~~~~~~~~~~~~~~
1728 During context reduction, we add to the wanted equalities also those
1729 equalities that (transitively) occur in superclass contexts of wanted
1730 class constraints. Consider the following code
1732 class a ~ Int => C a
1735 If (C a) is wanted, we want to add (a ~ Int), which will be discharged by
1736 substituting Int for a. Hence, we ultimately want (C Int), which we
1737 discharge with the explicit instance.
1740 reduceContext :: RedEnv
1742 -> TcM (ImprovementDone,
1743 TcDictBinds, -- Dictionary bindings
1744 [Inst], -- Irreducible
1745 [Inst]) -- Needed givens
1747 reduceContext env wanteds
1748 = do { traceTc (text "reduceContext" <+> (vcat [
1749 text "----------------------",
1751 text "given" <+> ppr (red_givens env),
1752 text "wanted" <+> ppr wanteds,
1753 text "----------------------"
1757 ; let givens = red_givens env
1758 (given_eqs0, given_dicts0) = partition isEqInst givens
1759 (wanted_eqs0, wanted_dicts0) = partition isEqInst wanteds
1761 -- We want to add as wanted equalities those that (transitively)
1762 -- occur in superclass contexts of wanted class constraints.
1763 -- See Note [Ancestor Equalities]
1764 ; ancestor_eqs <- ancestorEqualities wanted_dicts0
1765 ; let wanted_eqs = wanted_eqs0 ++ ancestor_eqs
1766 ; traceTc $ text "reduceContext: ancestor eqs" <+> ppr ancestor_eqs
1768 -- 1. Normalise the *given* *equality* constraints
1769 ; (given_eqs, eliminate_skolems) <- normaliseGivenEqs given_eqs0
1771 -- 2. Normalise the *given* *dictionary* constraints
1772 -- wrt. the toplevel and given equations
1773 ; (given_dicts, given_binds) <- normaliseGivenDicts given_eqs
1776 -- 5. Build the Avail mapping from "given_dicts"
1777 -- Add dicts refined by the current type refinement
1778 ; (init_state, extra_givens) <- getLIE $ do
1779 { init_state <- foldlM addGiven emptyAvails given_dicts
1780 ; let reft = red_reft env
1781 ; if isEmptyRefinement reft then return init_state
1782 else foldlM (addRefinedGiven reft)
1783 init_state given_dicts }
1785 -- *** ToDo: what to do with the "extra_givens"? For the
1786 -- moment I'm simply discarding them, which is probably wrong
1788 -- 7. Normalise the *wanted* *dictionary* constraints
1789 -- wrt. the toplevel and given equations
1790 -- NB: normalisation includes zonking as part of what it does
1791 -- so it's important to do it after any unifications
1792 -- that happened as a result of the addGivens
1793 ; (wanted_dicts,normalise_binds1) <- normaliseWantedDicts given_eqs wanted_dicts0
1795 -- 6. Solve the *wanted* *dictionary* constraints
1796 -- This may expose some further equational constraints...
1797 ; (avails, extra_eqs) <- getLIE (reduceList env wanted_dicts init_state)
1798 ; let (binds, irreds1, needed_givens) = extractResults avails wanted_dicts
1799 ; traceTc $ text "reduceContext extractresults" <+> vcat
1800 [ppr avails,ppr wanted_dicts,ppr binds,ppr needed_givens]
1802 -- *** ToDo: what to do with the "extra_eqs"? For the
1803 -- moment I'm simply discarding them, which is probably wrong
1805 -- 3. Solve the *wanted* *equation* constraints
1806 ; eq_irreds0 <- solveWantedEqs given_eqs wanted_eqs
1808 -- 4. Normalise the *wanted* equality constraints with respect to
1810 ; eq_irreds <- normaliseWantedEqs eq_irreds0
1812 -- 8. Substitute the wanted *equations* in the wanted *dictionaries*
1813 ; (irreds,normalise_binds2) <- substEqInDictInsts eq_irreds irreds1
1815 -- 9. eliminate the artificial skolem constants introduced in 1.
1818 -- Figure out whether we should go round again
1819 -- My current plan is to see if any of the mutable tyvars in
1820 -- givens or irreds has been filled in by improvement.
1821 -- If so, there is merit in going around again, because
1822 -- we may make further progress
1824 -- ToDo: is it only mutable stuff? We may have exposed new
1825 -- equality constraints and should probably go round again
1826 -- then as well. But currently we are dropping them on the
1829 ; let all_irreds = irreds ++ eq_irreds
1830 ; improved <- anyM isFilledMetaTyVar $ varSetElems $
1831 tyVarsOfInsts (givens ++ all_irreds)
1833 -- The old plan (fragile)
1834 -- improveed = availsImproved avails
1835 -- || (not $ isEmptyBag normalise_binds1)
1836 -- || (not $ isEmptyBag normalise_binds2)
1837 -- || (any isEqInst irreds)
1839 ; traceTc (text "reduceContext end" <+> (vcat [
1840 text "----------------------",
1842 text "given" <+> ppr givens,
1843 text "given_eqs" <+> ppr given_eqs,
1844 text "wanted" <+> ppr wanteds,
1845 text "wanted_dicts" <+> ppr wanted_dicts,
1847 text "avails" <+> pprAvails avails,
1848 text "improved =" <+> ppr improved,
1849 text "irreds = " <+> ppr irreds,
1850 text "binds = " <+> ppr binds,
1851 text "needed givens = " <+> ppr needed_givens,
1852 text "----------------------"
1856 given_binds `unionBags` normalise_binds1
1857 `unionBags` normalise_binds2
1863 tcImproveOne :: Avails -> Inst -> TcM ImprovementDone
1864 tcImproveOne avails inst
1865 | not (isDict inst) = return False
1867 = do { inst_envs <- tcGetInstEnvs
1868 ; let eqns = improveOne (classInstances inst_envs)
1869 (dictPred inst, pprInstArising inst)
1870 [ (dictPred p, pprInstArising p)
1871 | p <- availsInsts avails, isDict p ]
1872 -- Avails has all the superclasses etc (good)
1873 -- It also has all the intermediates of the deduction (good)
1874 -- It does not have duplicates (good)
1875 -- NB that (?x::t1) and (?x::t2) will be held separately in avails
1876 -- so that improve will see them separate
1877 ; traceTc (text "improveOne" <+> ppr inst)
1880 unifyEqns :: [(Equation,(PredType,SDoc),(PredType,SDoc))]
1881 -> TcM ImprovementDone
1882 unifyEqns [] = return False
1884 = do { traceTc (ptext SLIT("Improve:") <+> vcat (map pprEquationDoc eqns))
1888 unify ((qtvs, pairs), what1, what2)
1889 = addErrCtxtM (mkEqnMsg what1 what2) $
1890 tcInstTyVars (varSetElems qtvs) `thenM` \ (_, _, tenv) ->
1891 mapM_ (unif_pr tenv) pairs
1892 unif_pr tenv (ty1,ty2) = unifyType (substTy tenv ty1) (substTy tenv ty2)
1894 pprEquationDoc (eqn, (p1,w1), (p2,w2)) = vcat [pprEquation eqn, nest 2 (ppr p1), nest 2 (ppr p2)]
1896 mkEqnMsg (pred1,from1) (pred2,from2) tidy_env
1897 = do { pred1' <- zonkTcPredType pred1; pred2' <- zonkTcPredType pred2
1898 ; let { pred1'' = tidyPred tidy_env pred1'; pred2'' = tidyPred tidy_env pred2' }
1899 ; let msg = vcat [ptext SLIT("When using functional dependencies to combine"),
1900 nest 2 (sep [ppr pred1'' <> comma, nest 2 from1]),
1901 nest 2 (sep [ppr pred2'' <> comma, nest 2 from2])]
1902 ; return (tidy_env, msg) }
1905 The main context-reduction function is @reduce@. Here's its game plan.
1908 reduceList :: RedEnv -> [Inst] -> Avails -> TcM Avails
1909 reduceList env@(RedEnv {red_stack = (n,stk)}) wanteds state
1910 = do { traceTc (text "reduceList " <+> (ppr wanteds $$ ppr state))
1914 dumpTcRn (hang (ptext SLIT("Interesting! Context reduction stack depth") <+> int n)
1915 2 (ifPprDebug (nest 2 (pprStack stk))))
1918 ; if n >= ctxtStkDepth dopts then
1919 failWithTc (reduceDepthErr n stk)
1923 go [] state = return state
1924 go (w:ws) state = do { state' <- reduce (env {red_stack = (n+1, w:stk)}) w state
1927 -- Base case: we're done!
1928 reduce env wanted avails
1929 -- It's the same as an existing inst, or a superclass thereof
1930 | Just avail <- findAvail avails wanted
1931 = do { traceTc (text "reduce: found " <+> ppr wanted)
1936 = do { traceTc (text "reduce" <+> ppr wanted $$ ppr avails)
1937 ; case red_try_me env wanted of {
1938 Stop -> try_simple (addIrred NoSCs);
1939 -- See Note [No superclasses for Stop]
1941 ReduceMe want_scs -> do -- It should be reduced
1942 { (avails, lookup_result) <- reduceInst env avails wanted
1943 ; case lookup_result of
1944 NoInstance -> addIrred want_scs avails wanted
1945 -- Add it and its superclasses
1947 GenInst [] rhs -> addWanted want_scs avails wanted rhs []
1949 GenInst wanteds' rhs
1950 -> do { avails1 <- addIrred NoSCs avails wanted
1951 ; avails2 <- reduceList env wanteds' avails1
1952 ; addWanted want_scs avails2 wanted rhs wanteds' } }
1953 -- Temporarily do addIrred *before* the reduceList,
1954 -- which has the effect of adding the thing we are trying
1955 -- to prove to the database before trying to prove the things it
1956 -- needs. See note [RECURSIVE DICTIONARIES]
1957 -- NB: we must not do an addWanted before, because that adds the
1958 -- superclasses too, and that can lead to a spurious loop; see
1959 -- the examples in [SUPERCLASS-LOOP]
1960 -- So we do an addIrred before, and then overwrite it afterwards with addWanted
1963 -- First, see if the inst can be reduced to a constant in one step
1964 -- Works well for literals (1::Int) and constant dictionaries (d::Num Int)
1965 -- Don't bother for implication constraints, which take real work
1966 try_simple do_this_otherwise
1967 = do { res <- lookupSimpleInst wanted
1969 GenInst [] rhs -> addWanted AddSCs avails wanted rhs []
1970 other -> do_this_otherwise avails wanted }
1974 Note [SUPERCLASS-LOOP 2]
1975 ~~~~~~~~~~~~~~~~~~~~~~~~
1976 But the above isn't enough. Suppose we are *given* d1:Ord a,
1977 and want to deduce (d2:C [a]) where
1979 class Ord a => C a where
1980 instance Ord [a] => C [a] where ...
1982 Then we'll use the instance decl to deduce C [a] from Ord [a], and then add the
1983 superclasses of C [a] to avails. But we must not overwrite the binding
1984 for Ord [a] (which is obtained from Ord a) with a superclass selection or we'll just
1987 Here's another variant, immortalised in tcrun020
1988 class Monad m => C1 m
1989 class C1 m => C2 m x
1990 instance C2 Maybe Bool
1991 For the instance decl we need to build (C1 Maybe), and it's no good if
1992 we run around and add (C2 Maybe Bool) and its superclasses to the avails
1993 before we search for C1 Maybe.
1995 Here's another example
1996 class Eq b => Foo a b
1997 instance Eq a => Foo [a] a
2001 we'll first deduce that it holds (via the instance decl). We must not
2002 then overwrite the Eq t constraint with a superclass selection!
2004 At first I had a gross hack, whereby I simply did not add superclass constraints
2005 in addWanted, though I did for addGiven and addIrred. This was sub-optimal,
2006 becuase it lost legitimate superclass sharing, and it still didn't do the job:
2007 I found a very obscure program (now tcrun021) in which improvement meant the
2008 simplifier got two bites a the cherry... so something seemed to be an Stop
2009 first time, but reducible next time.
2011 Now we implement the Right Solution, which is to check for loops directly
2012 when adding superclasses. It's a bit like the occurs check in unification.
2015 Note [RECURSIVE DICTIONARIES]
2016 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2018 data D r = ZeroD | SuccD (r (D r));
2020 instance (Eq (r (D r))) => Eq (D r) where
2021 ZeroD == ZeroD = True
2022 (SuccD a) == (SuccD b) = a == b
2025 equalDC :: D [] -> D [] -> Bool;
2028 We need to prove (Eq (D [])). Here's how we go:
2032 by instance decl, holds if
2036 by instance decl of Eq, holds if
2038 where d2 = dfEqList d3
2041 But now we can "tie the knot" to give
2047 and it'll even run! The trick is to put the thing we are trying to prove
2048 (in this case Eq (D []) into the database before trying to prove its
2049 contributing clauses.
2052 %************************************************************************
2054 Reducing a single constraint
2056 %************************************************************************
2059 ---------------------------------------------
2060 reduceInst :: RedEnv -> Avails -> Inst -> TcM (Avails, LookupInstResult)
2061 reduceInst env avails (ImplicInst { tci_name = name,
2062 tci_tyvars = tvs, tci_reft = reft, tci_loc = loc,
2063 tci_given = extra_givens, tci_wanted = wanteds })
2064 = reduceImplication env avails name reft tvs extra_givens wanteds loc
2066 reduceInst env avails other_inst
2067 = do { result <- lookupSimpleInst other_inst
2068 ; return (avails, result) }
2071 Note [Equational Constraints in Implication Constraints]
2072 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2074 An implication constraint is of the form
2076 where Given and Wanted may contain both equational and dictionary
2077 constraints. The delay and reduction of these two kinds of constraints
2080 -) In the generated code, wanted Dictionary constraints are wrapped up in an
2081 implication constraint that is created at the code site where the wanted
2082 dictionaries can be reduced via a let-binding. This let-bound implication
2083 constraint is deconstructed at the use-site of the wanted dictionaries.
2085 -) While the reduction of equational constraints is also delayed, the delay
2086 is not manifest in the generated code. The required evidence is generated
2087 in the code directly at the use-site. There is no let-binding and deconstruction
2088 necessary. The main disadvantage is that we cannot exploit sharing as the
2089 same evidence may be generated at multiple use-sites. However, this disadvantage
2090 is limited because it only concerns coercions which are erased.
2092 The different treatment is motivated by the different in representation. Dictionary
2093 constraints require manifest runtime dictionaries, while equations require coercions
2097 ---------------------------------------------
2098 reduceImplication :: RedEnv
2101 -> Refinement -- May refine the givens; often empty
2102 -> [TcTyVar] -- Quantified type variables; all skolems
2103 -> [Inst] -- Extra givens; all rigid
2106 -> TcM (Avails, LookupInstResult)
2109 Suppose we are simplifying the constraint
2110 forall bs. extras => wanted
2111 in the context of an overall simplification problem with givens 'givens',
2112 and refinment 'reft'.
2115 * The refinement is often empty
2117 * The 'extra givens' need not mention any of the quantified type variables
2118 e.g. forall {}. Eq a => Eq [a]
2119 forall {}. C Int => D (Tree Int)
2121 This happens when you have something like
2123 T1 :: Eq a => a -> T a
2126 f x = ...(case x of { T1 v -> v==v })...
2129 -- ToDo: should we instantiate tvs? I think it's not necessary
2131 -- Note on coercion variables:
2133 -- The extra given coercion variables are bound at two different sites:
2134 -- -) in the creation context of the implication constraint
2135 -- the solved equational constraints use these binders
2137 -- -) at the solving site of the implication constraint
2138 -- the solved dictionaries use these binders
2139 -- these binders are generated by reduceImplication
2141 reduceImplication env orig_avails name reft tvs extra_givens wanteds inst_loc
2142 = do { -- Add refined givens, and the extra givens
2144 -- (refined_red_givens,refined_avails)
2145 -- <- if isEmptyRefinement reft then return (red_givens env,orig_avails)
2146 -- else foldlM (addRefinedGiven reft) ([],orig_avails) (red_givens env)
2147 -- Commented out SLPJ Sept 07; see comment with extractLocalResults below
2148 let refined_red_givens = []
2150 -- Solve the sub-problem
2151 ; let try_me inst = ReduceMe AddSCs -- Note [Freeness and implications]
2152 env' = env { red_givens = extra_givens ++ availsInsts orig_avails
2154 , red_doc = sep [ptext SLIT("reduceImplication for") <+> ppr name,
2155 nest 2 (parens $ ptext SLIT("within") <+> red_doc env)]
2156 , red_try_me = try_me }
2158 ; traceTc (text "reduceImplication" <+> vcat
2160 ppr (red_givens env), ppr extra_givens,
2161 ppr reft, ppr wanteds])
2162 ; (irreds,binds,needed_givens0) <- checkLoop env' wanteds
2163 ; let (extra_eq_givens, extra_dict_givens) = partition isEqInst extra_givens
2164 -- SLPJ Sept 07: I think this is bogus; currently
2165 -- there are no Eqinsts in extra_givens
2166 dict_ids = map instToId extra_dict_givens
2168 -- needed_givens0 is the free vars of the bindings
2169 -- Remove the ones we are going to lambda-bind
2170 -- Use the actual dictionary identity *not* equality on Insts
2171 -- (Mind you, it should make no difference here.)
2172 ; let needed_givens = [ng | ng <- needed_givens0
2173 , instToVar ng `notElem` dict_ids]
2175 -- Note [Reducing implication constraints]
2176 -- Tom -- update note, put somewhere!
2178 ; traceTc (text "reduceImplication result" <+> vcat
2179 [ppr irreds, ppr binds, ppr needed_givens])
2181 ; -- extract superclass binds
2182 -- (sc_binds,_) <- extractResults avails []
2183 -- ; traceTc (text "reduceImplication sc_binds" <+> vcat
2184 -- [ppr sc_binds, ppr avails])
2187 -- We always discard the extra avails we've generated;
2188 -- but we remember if we have done any (global) improvement
2189 -- ; let ret_avails = avails
2190 ; let ret_avails = orig_avails
2191 -- ; let ret_avails = updateImprovement orig_avails avails
2193 -- SLPJ Sept 07: what if improvement happened inside the checkLoop?
2194 -- Then we must iterate the outer loop too!
2196 ; traceTc (text "reduceImplication condition" <+> ppr ((isEmptyLHsBinds binds) || (null irreds)))
2198 -- Progress is no longer measered by the number of bindings
2199 ; if (isEmptyLHsBinds binds) && (not $ null irreds) then -- No progress
2200 -- If there are any irreds, we back off and return NoInstance
2201 return (ret_avails, NoInstance)
2203 { (implic_insts, bind) <- makeImplicationBind inst_loc tvs reft extra_givens irreds
2204 -- This binding is useless if the recursive simplification
2205 -- made no progress; but currently we don't try to optimise that
2206 -- case. After all, we only try hard to reduce at top level, or
2207 -- when inferring types.
2209 ; let dict_wanteds = filter (not . isEqInst) wanteds
2210 -- TOMDO: given equational constraints bug!
2211 -- we need a different evidence for given
2212 -- equations depending on whether we solve
2213 -- dictionary constraints or equational constraints
2215 eq_tyvars = varSetElems $ tyVarsOfTypes $ map eqInstType extra_eq_givens
2216 -- SLPJ Sept07: this looks Utterly Wrong to me, but I think
2217 -- that current extra_givens has no EqInsts, so
2218 -- it makes no difference
2219 -- dict_ids = map instToId extra_givens
2220 co = mkWpTyLams tvs <.> mkWpTyLams eq_tyvars <.> mkWpLams dict_ids <.> WpLet (binds `unionBags` bind)
2221 rhs = mkHsWrap co payload
2222 loc = instLocSpan inst_loc
2223 payload | [dict_wanted] <- dict_wanteds = HsVar (instToId dict_wanted)
2224 | otherwise = ExplicitTuple (map (L loc . HsVar . instToId) dict_wanteds) Boxed
2227 ; traceTc (vcat [text "reduceImplication" <+> ppr name,
2229 text "->" <+> sep [ppr needed_givens, ppr rhs]])
2230 ; return (ret_avails, GenInst (implic_insts ++ needed_givens) (L loc rhs))
2235 Note [Reducing implication constraints]
2236 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2237 Suppose we are trying to simplify
2239 ic: (forall b. C a b => (W [a] b, D c b)) )
2241 instance (C a b, Ord a) => W [a] b
2242 When solving the implication constraint, we'll start with
2244 in the Avails. Then we add (C a b -> Given) and solve. Extracting
2245 the results gives us a binding for the (W [a] b), with an Irred of
2246 (Ord a, D c b). Now, the (Ord a) comes from "outside" the implication,
2247 but the (D d b) is from "inside". So we want to generate a GenInst
2252 ic' :: forall b. C a b => D c b]
2253 (/\b \(dc:C a b). (df a b dc do, ic' b dc))
2255 The first arg of GenInst gives the free dictionary variables of the
2256 second argument -- the "needed givens". And that list in turn is
2257 vital because it's used to determine what other dicts must be solved.
2258 This very list ends up in the second field of the Rhs, and drives
2261 The need for this field is why we have to return "needed givens"
2262 from extractResults, reduceContext, checkLoop, and so on.
2264 NB: the "needed givens" in a GenInst or Rhs, may contain two dicts
2265 with the same type but different Ids, e.g. [d12 :: Eq a, d81 :: Eq a]
2266 That says we must generate a binding for both d12 and d81.
2268 The "inside" and "outside" distinction is what's going on with 'inner' and
2269 'outer' in reduceImplication
2272 Note [Freeness and implications]
2273 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2274 It's hard to say when an implication constraint can be floated out. Consider
2275 forall {} Eq a => Foo [a]
2276 The (Foo [a]) doesn't mention any of the quantified variables, but it
2277 still might be partially satisfied by the (Eq a).
2279 There is a useful special case when it *is* easy to partition the
2280 constraints, namely when there are no 'givens'. Consider
2281 forall {a}. () => Bar b
2282 There are no 'givens', and so there is no reason to capture (Bar b).
2283 We can let it float out. But if there is even one constraint we
2284 must be much more careful:
2285 forall {a}. C a b => Bar (m b)
2286 because (C a b) might have a superclass (D b), from which we might
2287 deduce (Bar [b]) when m later gets instantiated to []. Ha!
2289 Here is an even more exotic example
2291 Now consider the constraint
2292 forall b. D Int b => C Int
2293 We can satisfy the (C Int) from the superclass of D, so we don't want
2294 to float the (C Int) out, even though it mentions no type variable in
2297 Note [Pruning the givens in an implication constraint]
2298 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2299 Suppose we are about to form the implication constraint
2300 forall tvs. Eq a => Ord b
2301 The (Eq a) cannot contribute to the (Ord b), because it has no access to
2302 the type variable 'b'. So we could filter out the (Eq a) from the givens.
2304 Doing so would be a bit tidier, but all the implication constraints get
2305 simplified away by the optimiser, so it's no great win. So I don't take
2306 advantage of that at the moment.
2308 If you do, BE CAREFUL of wobbly type variables.
2311 %************************************************************************
2313 Avails and AvailHow: the pool of evidence
2315 %************************************************************************
2319 data Avails = Avails !ImprovementDone !AvailEnv
2321 type ImprovementDone = Bool -- True <=> some unification has happened
2322 -- so some Irreds might now be reducible
2323 -- keys that are now
2325 type AvailEnv = FiniteMap Inst AvailHow
2327 = IsIrred -- Used for irreducible dictionaries,
2328 -- which are going to be lambda bound
2330 | Given Inst -- Used for dictionaries for which we have a binding
2331 -- e.g. those "given" in a signature
2333 | Rhs -- Used when there is a RHS
2334 (LHsExpr TcId) -- The RHS
2335 [Inst] -- Insts free in the RHS; we need these too
2337 instance Outputable Avails where
2340 pprAvails (Avails imp avails)
2341 = vcat [ ptext SLIT("Avails") <> (if imp then ptext SLIT("[improved]") else empty)
2343 vcat [ sep [ppr inst, nest 2 (equals <+> ppr avail)]
2344 | (inst,avail) <- fmToList avails ]]
2346 instance Outputable AvailHow where
2349 -------------------------
2350 pprAvail :: AvailHow -> SDoc
2351 pprAvail IsIrred = text "Irred"
2352 pprAvail (Given x) = text "Given" <+> ppr x
2353 pprAvail (Rhs rhs bs) = sep [text "Rhs" <+> ppr bs,
2356 -------------------------
2357 extendAvailEnv :: AvailEnv -> Inst -> AvailHow -> AvailEnv
2358 extendAvailEnv env inst avail = addToFM env inst avail
2360 findAvailEnv :: AvailEnv -> Inst -> Maybe AvailHow
2361 findAvailEnv env wanted = lookupFM env wanted
2362 -- NB 1: the Ord instance of Inst compares by the class/type info
2363 -- *not* by unique. So
2364 -- d1::C Int == d2::C Int
2366 emptyAvails :: Avails
2367 emptyAvails = Avails False emptyFM
2369 findAvail :: Avails -> Inst -> Maybe AvailHow
2370 findAvail (Avails _ avails) wanted = findAvailEnv avails wanted
2372 elemAvails :: Inst -> Avails -> Bool
2373 elemAvails wanted (Avails _ avails) = wanted `elemFM` avails
2375 extendAvails :: Avails -> Inst -> AvailHow -> TcM Avails
2377 extendAvails avails@(Avails imp env) inst avail
2378 = do { imp1 <- tcImproveOne avails inst -- Do any improvement
2379 ; return (Avails (imp || imp1) (extendAvailEnv env inst avail)) }
2381 availsInsts :: Avails -> [Inst]
2382 availsInsts (Avails _ avails) = keysFM avails
2384 availsImproved (Avails imp _) = imp
2386 updateImprovement :: Avails -> Avails -> Avails
2387 -- (updateImprovement a1 a2) sets a1's improvement flag from a2
2388 updateImprovement (Avails _ avails1) (Avails imp2 _) = Avails imp2 avails1
2391 Extracting the bindings from a bunch of Avails.
2392 The bindings do *not* come back sorted in dependency order.
2393 We assume that they'll be wrapped in a big Rec, so that the
2394 dependency analyser can sort them out later
2397 type DoneEnv = FiniteMap Inst [Id]
2398 -- Tracks which things we have evidence for
2400 extractResults :: Avails
2402 -> (TcDictBinds, -- Bindings
2403 [Inst], -- Irreducible ones
2404 [Inst]) -- Needed givens, i.e. ones used in the bindings
2405 -- Postcondition: needed-givens = free vars( binds ) \ irreds
2406 -- needed-gives is subset of Givens in incoming Avails
2407 -- Note [Reducing implication constraints]
2409 extractResults (Avails _ avails) wanteds
2410 = go emptyBag [] [] emptyFM wanteds
2412 go :: TcDictBinds -- Bindings for dicts
2414 -> [Inst] -- Needed givens
2415 -> DoneEnv -- Has an entry for each inst in the above three sets
2417 -> (TcDictBinds, [Inst], [Inst])
2418 go binds irreds givens done []
2419 = (binds, irreds, givens)
2421 go binds irreds givens done (w:ws)
2422 | Just done_ids@(done_id : rest_done_ids) <- lookupFM done w
2423 = if w_id `elem` done_ids then
2424 go binds irreds givens done ws
2426 go (add_bind (nlHsVar done_id)) irreds givens
2427 (addToFM done w (done_id : w_id : rest_done_ids)) ws
2429 | otherwise -- Not yet done
2430 = case findAvailEnv avails w of
2431 Nothing -> pprTrace "Urk: extractResults" (ppr w) $
2432 go binds irreds givens done ws
2434 Just IsIrred -> go binds (w:irreds) givens done' ws
2436 Just (Rhs rhs ws') -> go (add_bind rhs) irreds givens done' (ws' ++ ws)
2438 Just (Given g) -> go binds' irreds (g:givens) (addToFM done w [g_id]) ws
2441 binds' | w_id == g_id = binds
2442 | otherwise = add_bind (nlHsVar g_id)
2445 done' = addToFM done w [w_id]
2446 add_bind rhs = addInstToDictBind binds w rhs
2450 Note [No superclasses for Stop]
2451 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2452 When we decide not to reduce an Inst -- the 'WhatToDo' --- we still
2453 add it to avails, so that any other equal Insts will be commoned up
2454 right here. However, we do *not* add superclasses. If we have
2457 but a is not bound here, then we *don't* want to derive dn from df
2458 here lest we lose sharing.
2461 addWanted :: WantSCs -> Avails -> Inst -> LHsExpr TcId -> [Inst] -> TcM Avails
2462 addWanted want_scs avails wanted rhs_expr wanteds
2463 = addAvailAndSCs want_scs avails wanted avail
2465 avail = Rhs rhs_expr wanteds
2467 addGiven :: Avails -> Inst -> TcM Avails
2468 addGiven avails given = addAvailAndSCs AddSCs avails given (Given given)
2469 -- Always add superclasses for 'givens'
2471 -- No ASSERT( not (given `elemAvails` avails) ) because in an instance
2472 -- decl for Ord t we can add both Ord t and Eq t as 'givens',
2473 -- so the assert isn't true
2475 addRefinedGiven :: Refinement -> Avails -> Inst -> TcM Avails
2476 addRefinedGiven reft avails given
2477 | isDict given -- We sometimes have 'given' methods, but they
2478 -- are always optional, so we can drop them
2479 , let pred = dictPred given
2480 , isRefineablePred pred -- See Note [ImplicInst rigidity]
2481 , Just (co, pred) <- refinePred reft pred
2482 = do { new_given <- newDictBndr (instLoc given) pred
2483 ; let rhs = L (instSpan given) $
2484 HsWrap (WpCo co) (HsVar (instToId given))
2485 ; addAvailAndSCs AddSCs avails new_given (Rhs rhs [given]) }
2486 -- ToDo: the superclasses of the original given all exist in Avails
2487 -- so we could really just cast them, but it's more awkward to do,
2488 -- and hopefully the optimiser will spot the duplicated work
2493 Note [ImplicInst rigidity]
2494 ~~~~~~~~~~~~~~~~~~~~~~~~~~
2496 C :: forall ab. (Eq a, Ord b) => b -> T a
2498 ...(case x of C v -> <body>)...
2500 From the case (where x::T ty) we'll get an implication constraint
2501 forall b. (Eq ty, Ord b) => <body-constraints>
2502 Now suppose <body-constraints> itself has an implication constraint
2504 forall c. <reft> => <payload>
2505 Then, we can certainly apply the refinement <reft> to the Ord b, becuase it is
2506 existential, but we probably should not apply it to the (Eq ty) because it may
2507 be wobbly. Hence the isRigidInst
2509 @Insts@ are ordered by their class/type info, rather than by their
2510 unique. This allows the context-reduction mechanism to use standard finite
2511 maps to do their stuff. It's horrible that this code is here, rather
2512 than with the Avails handling stuff in TcSimplify
2515 addIrred :: WantSCs -> Avails -> Inst -> TcM Avails
2516 addIrred want_scs avails irred = ASSERT2( not (irred `elemAvails` avails), ppr irred $$ ppr avails )
2517 addAvailAndSCs want_scs avails irred IsIrred
2519 addAvailAndSCs :: WantSCs -> Avails -> Inst -> AvailHow -> TcM Avails
2520 addAvailAndSCs want_scs avails inst avail
2521 | not (isClassDict inst) = extendAvails avails inst avail
2522 | NoSCs <- want_scs = extendAvails avails inst avail
2523 | otherwise = do { traceTc (text "addAvailAndSCs" <+> vcat [ppr inst, ppr deps])
2524 ; avails' <- extendAvails avails inst avail
2525 ; addSCs is_loop avails' inst }
2527 is_loop pred = any (`tcEqType` mkPredTy pred) dep_tys
2528 -- Note: this compares by *type*, not by Unique
2529 deps = findAllDeps (unitVarSet (instToVar inst)) avail
2530 dep_tys = map idType (varSetElems deps)
2532 findAllDeps :: IdSet -> AvailHow -> IdSet
2533 -- Find all the Insts that this one depends on
2534 -- See Note [SUPERCLASS-LOOP 2]
2535 -- Watch out, though. Since the avails may contain loops
2536 -- (see Note [RECURSIVE DICTIONARIES]), so we need to track the ones we've seen so far
2537 findAllDeps so_far (Rhs _ kids) = foldl find_all so_far kids
2538 findAllDeps so_far other = so_far
2540 find_all :: IdSet -> Inst -> IdSet
2542 | isEqInst kid = so_far
2543 | kid_id `elemVarSet` so_far = so_far
2544 | Just avail <- findAvail avails kid = findAllDeps so_far' avail
2545 | otherwise = so_far'
2547 so_far' = extendVarSet so_far kid_id -- Add the new kid to so_far
2548 kid_id = instToId kid
2550 addSCs :: (TcPredType -> Bool) -> Avails -> Inst -> TcM Avails
2551 -- Add all the superclasses of the Inst to Avails
2552 -- The first param says "don't do this because the original thing
2553 -- depends on this one, so you'd build a loop"
2554 -- Invariant: the Inst is already in Avails.
2556 addSCs is_loop avails dict
2557 = ASSERT( isDict dict )
2558 do { sc_dicts <- newDictBndrs (instLoc dict) sc_theta'
2559 ; foldlM add_sc avails (zipEqual "add_scs" sc_dicts sc_sels) }
2561 (clas, tys) = getDictClassTys dict
2562 (tyvars, sc_theta, sc_sels, _) = classBigSig clas
2563 sc_theta' = filter (not . isEqPred) $
2564 substTheta (zipTopTvSubst tyvars tys) sc_theta
2566 add_sc avails (sc_dict, sc_sel)
2567 | is_loop (dictPred sc_dict) = return avails -- See Note [SUPERCLASS-LOOP 2]
2568 | is_given sc_dict = return avails
2569 | otherwise = do { avails' <- extendAvails avails sc_dict (Rhs sc_sel_rhs [dict])
2570 ; addSCs is_loop avails' sc_dict }
2572 sc_sel_rhs = L (instSpan dict) (HsWrap co_fn (HsVar sc_sel))
2573 co_fn = WpApp (instToVar dict) <.> mkWpTyApps tys
2575 is_given :: Inst -> Bool
2576 is_given sc_dict = case findAvail avails sc_dict of
2577 Just (Given _) -> True -- Given is cheaper than superclass selection
2580 -- From the a set of insts obtain all equalities that (transitively) occur in
2581 -- superclass contexts of class constraints (aka the ancestor equalities).
2583 ancestorEqualities :: [Inst] -> TcM [Inst]
2585 = mapM mkWantedEqInst -- turn only equality predicates..
2586 . filter isEqPred -- ..into wanted equality insts
2588 . addAEsToBag emptyBag -- collect the superclass constraints..
2589 . map dictPred -- ..of all predicates in a bag
2590 . filter isClassDict
2592 addAEsToBag :: Bag PredType -> [PredType] -> Bag PredType
2593 addAEsToBag bag [] = bag
2594 addAEsToBag bag (pred:preds)
2595 | pred `elemBag` bag = addAEsToBag bag preds
2596 | isEqPred pred = addAEsToBag bagWithPred preds
2597 | isClassPred pred = addAEsToBag bagWithPred predsWithSCs
2598 | otherwise = addAEsToBag bag preds
2600 bagWithPred = bag `snocBag` pred
2601 predsWithSCs = preds ++ substTheta (zipTopTvSubst tyvars tys) sc_theta
2603 (tyvars, sc_theta, _, _) = classBigSig clas
2604 (clas, tys) = getClassPredTys pred
2608 %************************************************************************
2610 \section{tcSimplifyTop: defaulting}
2612 %************************************************************************
2615 @tcSimplifyTop@ is called once per module to simplify all the constant
2616 and ambiguous Insts.
2618 We need to be careful of one case. Suppose we have
2620 instance Num a => Num (Foo a b) where ...
2622 and @tcSimplifyTop@ is given a constraint (Num (Foo x y)). Then it'll simplify
2623 to (Num x), and default x to Int. But what about y??
2625 It's OK: the final zonking stage should zap y to (), which is fine.
2629 tcSimplifyTop, tcSimplifyInteractive :: [Inst] -> TcM TcDictBinds
2630 tcSimplifyTop wanteds
2631 = tc_simplify_top doc False wanteds
2633 doc = text "tcSimplifyTop"
2635 tcSimplifyInteractive wanteds
2636 = tc_simplify_top doc True wanteds
2638 doc = text "tcSimplifyInteractive"
2640 -- The TcLclEnv should be valid here, solely to improve
2641 -- error message generation for the monomorphism restriction
2642 tc_simplify_top doc interactive wanteds
2643 = do { dflags <- getDOpts
2644 ; wanteds <- zonkInsts wanteds
2645 ; mapM_ zonkTopTyVar (varSetElems (tyVarsOfInsts wanteds))
2647 ; traceTc (text "tc_simplify_top 0: " <+> ppr wanteds)
2648 ; (irreds1, binds1) <- tryHardCheckLoop doc1 wanteds
2649 -- ; (irreds1, binds1) <- gentleInferLoop doc1 wanteds
2650 ; traceTc (text "tc_simplify_top 1: " <+> ppr irreds1)
2651 ; (irreds2, binds2) <- approximateImplications doc2 (\d -> True) irreds1
2652 ; traceTc (text "tc_simplify_top 2: " <+> ppr irreds2)
2654 -- Use the defaulting rules to do extra unification
2655 -- NB: irreds2 are already zonked
2656 ; (irreds3, binds3) <- disambiguate doc3 interactive dflags irreds2
2658 -- Deal with implicit parameters
2659 ; let (bad_ips, non_ips) = partition isIPDict irreds3
2660 (ambigs, others) = partition isTyVarDict non_ips
2662 ; topIPErrs bad_ips -- Can arise from f :: Int -> Int
2664 ; addNoInstanceErrs others
2665 ; addTopAmbigErrs ambigs
2667 ; return (binds1 `unionBags` binds2 `unionBags` binds3) }
2669 doc1 = doc <+> ptext SLIT("(first round)")
2670 doc2 = doc <+> ptext SLIT("(approximate)")
2671 doc3 = doc <+> ptext SLIT("(disambiguate)")
2674 If a dictionary constrains a type variable which is
2675 * not mentioned in the environment
2676 * and not mentioned in the type of the expression
2677 then it is ambiguous. No further information will arise to instantiate
2678 the type variable; nor will it be generalised and turned into an extra
2679 parameter to a function.
2681 It is an error for this to occur, except that Haskell provided for
2682 certain rules to be applied in the special case of numeric types.
2684 * at least one of its classes is a numeric class, and
2685 * all of its classes are numeric or standard
2686 then the type variable can be defaulted to the first type in the
2687 default-type list which is an instance of all the offending classes.
2689 So here is the function which does the work. It takes the ambiguous
2690 dictionaries and either resolves them (producing bindings) or
2691 complains. It works by splitting the dictionary list by type
2692 variable, and using @disambigOne@ to do the real business.
2694 @disambigOne@ assumes that its arguments dictionaries constrain all
2695 the same type variable.
2697 ADR Comment 20/6/94: I've changed the @CReturnable@ case to default to
2698 @()@ instead of @Int@. I reckon this is the Right Thing to do since
2699 the most common use of defaulting is code like:
2701 _ccall_ foo `seqPrimIO` bar
2703 Since we're not using the result of @foo@, the result if (presumably)
2707 disambiguate :: SDoc -> Bool -> DynFlags -> [Inst] -> TcM ([Inst], TcDictBinds)
2708 -- Just does unification to fix the default types
2709 -- The Insts are assumed to be pre-zonked
2710 disambiguate doc interactive dflags insts
2712 = return (insts, emptyBag)
2714 | null defaultable_groups
2715 = do { traceTc (text "disambigutate, no defaultable groups" <+> vcat [ppr unaries, ppr insts, ppr bad_tvs, ppr defaultable_groups])
2716 ; return (insts, emptyBag) }
2719 = do { -- Figure out what default types to use
2720 default_tys <- getDefaultTys extended_defaulting ovl_strings
2722 ; traceTc (text "disambiguate1" <+> vcat [ppr insts, ppr unaries, ppr bad_tvs, ppr defaultable_groups])
2723 ; mapM_ (disambigGroup default_tys) defaultable_groups
2725 -- disambigGroup does unification, hence try again
2726 ; tryHardCheckLoop doc insts }
2729 extended_defaulting = interactive || dopt Opt_ExtendedDefaultRules dflags
2730 ovl_strings = dopt Opt_OverloadedStrings dflags
2732 unaries :: [(Inst, Class, TcTyVar)] -- (C tv) constraints
2733 bad_tvs :: TcTyVarSet -- Tyvars mentioned by *other* constraints
2734 (unaries, bad_tvs_s) = partitionWith find_unary insts
2735 bad_tvs = unionVarSets bad_tvs_s
2737 -- Finds unary type-class constraints
2738 find_unary d@(Dict {tci_pred = ClassP cls [ty]})
2739 | Just tv <- tcGetTyVar_maybe ty = Left (d,cls,tv)
2740 find_unary inst = Right (tyVarsOfInst inst)
2742 -- Group by type variable
2743 defaultable_groups :: [[(Inst,Class,TcTyVar)]]
2744 defaultable_groups = filter defaultable_group (equivClasses cmp_tv unaries)
2745 cmp_tv (_,_,tv1) (_,_,tv2) = tv1 `compare` tv2
2747 defaultable_group :: [(Inst,Class,TcTyVar)] -> Bool
2748 defaultable_group ds@((_,_,tv):_)
2749 = isTyConableTyVar tv -- Note [Avoiding spurious errors]
2750 && not (tv `elemVarSet` bad_tvs)
2751 && defaultable_classes [c | (_,c,_) <- ds]
2752 defaultable_group [] = panic "defaultable_group"
2754 defaultable_classes clss
2755 | extended_defaulting = any isInteractiveClass clss
2756 | otherwise = all is_std_class clss && (any is_num_class clss)
2758 -- In interactive mode, or with -fextended-default-rules,
2759 -- we default Show a to Show () to avoid graututious errors on "show []"
2760 isInteractiveClass cls
2761 = is_num_class cls || (classKey cls `elem` [showClassKey, eqClassKey, ordClassKey])
2763 is_num_class cls = isNumericClass cls || (ovl_strings && (cls `hasKey` isStringClassKey))
2764 -- is_num_class adds IsString to the standard numeric classes,
2765 -- when -foverloaded-strings is enabled
2767 is_std_class cls = isStandardClass cls || (ovl_strings && (cls `hasKey` isStringClassKey))
2768 -- Similarly is_std_class
2770 -----------------------
2771 disambigGroup :: [Type] -- The default types
2772 -> [(Inst,Class,TcTyVar)] -- All standard classes of form (C a)
2773 -> TcM () -- Just does unification, to fix the default types
2775 disambigGroup default_tys dicts
2776 = try_default default_tys
2778 (_,_,tyvar) = ASSERT(not (null dicts)) head dicts -- Should be non-empty
2779 classes = [c | (_,c,_) <- dicts]
2781 try_default [] = return ()
2782 try_default (default_ty : default_tys)
2783 = tryTcLIE_ (try_default default_tys) $
2784 do { tcSimplifyDefault [mkClassPred clas [default_ty] | clas <- classes]
2785 -- This may fail; then the tryTcLIE_ kicks in
2786 -- Failure here is caused by there being no type in the
2787 -- default list which can satisfy all the ambiguous classes.
2788 -- For example, if Real a is reqd, but the only type in the
2789 -- default list is Int.
2791 -- After this we can't fail
2792 ; warnDefault dicts default_ty
2793 ; unifyType default_ty (mkTyVarTy tyvar)
2794 ; return () -- TOMDO: do something with the coercion
2798 -----------------------
2799 getDefaultTys :: Bool -> Bool -> TcM [Type]
2800 getDefaultTys extended_deflts ovl_strings
2801 = do { mb_defaults <- getDeclaredDefaultTys
2802 ; case mb_defaults of {
2803 Just tys -> return tys ; -- User-supplied defaults
2806 -- No use-supplied default
2807 -- Use [Integer, Double], plus modifications
2808 { integer_ty <- tcMetaTy integerTyConName
2809 ; checkWiredInTyCon doubleTyCon
2810 ; string_ty <- tcMetaTy stringTyConName
2811 ; return (opt_deflt extended_deflts unitTy
2812 -- Note [Default unitTy]
2814 [integer_ty,doubleTy]
2816 opt_deflt ovl_strings string_ty) } } }
2818 opt_deflt True ty = [ty]
2819 opt_deflt False ty = []
2822 Note [Default unitTy]
2823 ~~~~~~~~~~~~~~~~~~~~~
2824 In interative mode (or with -fextended-default-rules) we add () as the first type we
2825 try when defaulting. This has very little real impact, except in the following case.
2827 Text.Printf.printf "hello"
2828 This has type (forall a. IO a); it prints "hello", and returns 'undefined'. We don't
2829 want the GHCi repl loop to try to print that 'undefined'. The neatest thing is to
2830 default the 'a' to (), rather than to Integer (which is what would otherwise happen;
2831 and then GHCi doesn't attempt to print the (). So in interactive mode, we add
2832 () to the list of defaulting types. See Trac #1200.
2834 Note [Avoiding spurious errors]
2835 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
2836 When doing the unification for defaulting, we check for skolem
2837 type variables, and simply don't default them. For example:
2838 f = (*) -- Monomorphic
2839 g :: Num a => a -> a
2841 Here, we get a complaint when checking the type signature for g,
2842 that g isn't polymorphic enough; but then we get another one when
2843 dealing with the (Num a) context arising from f's definition;
2844 we try to unify a with Int (to default it), but find that it's
2845 already been unified with the rigid variable from g's type sig
2848 %************************************************************************
2850 \subsection[simple]{@Simple@ versions}
2852 %************************************************************************
2854 Much simpler versions when there are no bindings to make!
2856 @tcSimplifyThetas@ simplifies class-type constraints formed by
2857 @deriving@ declarations and when specialising instances. We are
2858 only interested in the simplified bunch of class/type constraints.
2860 It simplifies to constraints of the form (C a b c) where
2861 a,b,c are type variables. This is required for the context of
2862 instance declarations.
2865 tcSimplifyDeriv :: InstOrigin
2867 -> ThetaType -- Wanted
2868 -> TcM ThetaType -- Needed
2869 -- Given instance (wanted) => C inst_ty
2870 -- Simplify 'wanted' as much as possible
2872 tcSimplifyDeriv orig tyvars theta
2873 = do { (tvs, _, tenv) <- tcInstTyVars tyvars
2874 -- The main loop may do unification, and that may crash if
2875 -- it doesn't see a TcTyVar, so we have to instantiate. Sigh
2876 -- ToDo: what if two of them do get unified?
2877 ; wanteds <- newDictBndrsO orig (substTheta tenv theta)
2878 ; (irreds, _) <- tryHardCheckLoop doc wanteds
2880 ; let (tv_dicts, others) = partition ok irreds
2881 ; addNoInstanceErrs others
2882 -- See Note [Exotic derived instance contexts] in TcMType
2884 ; let rev_env = zipTopTvSubst tvs (mkTyVarTys tyvars)
2885 simpl_theta = substTheta rev_env (map dictPred tv_dicts)
2886 -- This reverse-mapping is a pain, but the result
2887 -- should mention the original TyVars not TcTyVars
2889 ; return simpl_theta }
2891 doc = ptext SLIT("deriving classes for a data type")
2893 ok dict | isDict dict = validDerivPred (dictPred dict)
2898 @tcSimplifyDefault@ just checks class-type constraints, essentially;
2899 used with \tr{default} declarations. We are only interested in
2900 whether it worked or not.
2903 tcSimplifyDefault :: ThetaType -- Wanted; has no type variables in it
2906 tcSimplifyDefault theta
2907 = newDictBndrsO DefaultOrigin theta `thenM` \ wanteds ->
2908 tryHardCheckLoop doc wanteds `thenM` \ (irreds, _) ->
2909 addNoInstanceErrs irreds `thenM_`
2913 traceTc (ptext SLIT("tcSimplifyDefault failing")) >> failM
2915 doc = ptext SLIT("default declaration")
2919 %************************************************************************
2921 \section{Errors and contexts}
2923 %************************************************************************
2925 ToDo: for these error messages, should we note the location as coming
2926 from the insts, or just whatever seems to be around in the monad just
2930 groupErrs :: ([Inst] -> TcM ()) -- Deal with one group
2931 -> [Inst] -- The offending Insts
2933 -- Group together insts with the same origin
2934 -- We want to report them together in error messages
2936 groupErrs report_err []
2938 groupErrs report_err (inst:insts)
2939 = do { do_one (inst:friends)
2940 ; groupErrs report_err others }
2942 -- (It may seem a bit crude to compare the error messages,
2943 -- but it makes sure that we combine just what the user sees,
2944 -- and it avoids need equality on InstLocs.)
2945 (friends, others) = partition is_friend insts
2946 loc_msg = showSDoc (pprInstLoc (instLoc inst))
2947 is_friend friend = showSDoc (pprInstLoc (instLoc friend)) == loc_msg
2948 do_one insts = addInstCtxt (instLoc (head insts)) (report_err insts)
2949 -- Add location and context information derived from the Insts
2951 -- Add the "arising from..." part to a message about bunch of dicts
2952 addInstLoc :: [Inst] -> Message -> Message
2953 addInstLoc insts msg = msg $$ nest 2 (pprInstArising (head insts))
2955 addTopIPErrs :: [Name] -> [Inst] -> TcM ()
2956 addTopIPErrs bndrs []
2958 addTopIPErrs bndrs ips
2959 = do { dflags <- getDOpts
2960 ; addErrTcM (tidy_env, mk_msg dflags tidy_ips) }
2962 (tidy_env, tidy_ips) = tidyInsts ips
2964 = vcat [sep [ptext SLIT("Implicit parameters escape from"),
2965 nest 2 (ptext SLIT("the monomorphic top-level binding")
2966 <> plural bndrs <+> ptext SLIT("of")
2967 <+> pprBinders bndrs <> colon)],
2968 nest 2 (vcat (map ppr_ip ips)),
2969 monomorphism_fix dflags]
2970 ppr_ip ip = pprPred (dictPred ip) <+> pprInstArising ip
2972 topIPErrs :: [Inst] -> TcM ()
2974 = groupErrs report tidy_dicts
2976 (tidy_env, tidy_dicts) = tidyInsts dicts
2977 report dicts = addErrTcM (tidy_env, mk_msg dicts)
2978 mk_msg dicts = addInstLoc dicts (ptext SLIT("Unbound implicit parameter") <>
2979 plural tidy_dicts <+> pprDictsTheta tidy_dicts)
2981 addNoInstanceErrs :: [Inst] -- Wanted (can include implications)
2983 addNoInstanceErrs insts
2984 = do { let (tidy_env, tidy_insts) = tidyInsts insts
2985 ; reportNoInstances tidy_env Nothing tidy_insts }
2989 -> Maybe (InstLoc, [Inst]) -- Context
2990 -- Nothing => top level
2991 -- Just (d,g) => d describes the construct
2993 -> [Inst] -- What is wanted (can include implications)
2996 reportNoInstances tidy_env mb_what insts
2997 = groupErrs (report_no_instances tidy_env mb_what) insts
2999 report_no_instances tidy_env mb_what insts
3000 = do { inst_envs <- tcGetInstEnvs
3001 ; let (implics, insts1) = partition isImplicInst insts
3002 (insts2, overlaps) = partitionWith (check_overlap inst_envs) insts1
3003 (eqInsts, insts3) = partition isEqInst insts2
3004 ; traceTc (text "reportNoInstances" <+> vcat
3005 [ppr insts, ppr implics, ppr insts1, ppr insts2])
3006 ; mapM_ complain_implic implics
3007 ; mapM_ (\doc -> addErrTcM (tidy_env, doc)) overlaps
3008 ; groupErrs complain_no_inst insts3
3009 ; mapM_ (addErrTcM . mk_eq_err) eqInsts
3012 complain_no_inst insts = addErrTcM (tidy_env, mk_no_inst_err insts)
3014 complain_implic inst -- Recurse!
3015 = reportNoInstances tidy_env
3016 (Just (tci_loc inst, tci_given inst))
3019 check_overlap :: (InstEnv,InstEnv) -> Inst -> Either Inst SDoc
3020 -- Right msg => overlap message
3021 -- Left inst => no instance
3022 check_overlap inst_envs wanted
3023 | not (isClassDict wanted) = Left wanted
3025 = case lookupInstEnv inst_envs clas tys of
3026 -- The case of exactly one match and no unifiers means a
3027 -- successful lookup. That can't happen here, because dicts
3028 -- only end up here if they didn't match in Inst.lookupInst
3030 ([m],[]) -> pprPanic "reportNoInstance" (ppr wanted)
3032 ([], _) -> Left wanted -- No match
3033 res -> Right (mk_overlap_msg wanted res)
3035 (clas,tys) = getDictClassTys wanted
3037 mk_overlap_msg dict (matches, unifiers)
3038 = ASSERT( not (null matches) )
3039 vcat [ addInstLoc [dict] ((ptext SLIT("Overlapping instances for")
3040 <+> pprPred (dictPred dict))),
3041 sep [ptext SLIT("Matching instances") <> colon,
3042 nest 2 (vcat [pprInstances ispecs, pprInstances unifiers])],
3043 if not (isSingleton matches)
3044 then -- Two or more matches
3046 else -- One match, plus some unifiers
3047 ASSERT( not (null unifiers) )
3048 parens (vcat [ptext SLIT("The choice depends on the instantiation of") <+>
3049 quotes (pprWithCommas ppr (varSetElems (tyVarsOfInst dict))),
3050 ptext SLIT("To pick the first instance above, use -fallow-incoherent-instances"),
3051 ptext SLIT("when compiling the other instance declarations")])]
3053 ispecs = [ispec | (ispec, _) <- matches]
3055 mk_eq_err :: Inst -> (TidyEnv, SDoc)
3056 mk_eq_err inst = misMatchMsg tidy_env (eqInstTys inst)
3058 mk_no_inst_err insts
3059 | null insts = empty
3061 | Just (loc, givens) <- mb_what, -- Nested (type signatures, instance decls)
3062 not (isEmptyVarSet (tyVarsOfInsts insts))
3063 = vcat [ addInstLoc insts $
3064 sep [ ptext SLIT("Could not deduce") <+> pprDictsTheta insts
3065 , nest 2 $ ptext SLIT("from the context") <+> pprDictsTheta givens]
3066 , show_fixes (fix1 loc : fixes2) ]
3068 | otherwise -- Top level
3069 = vcat [ addInstLoc insts $
3070 ptext SLIT("No instance") <> plural insts
3071 <+> ptext SLIT("for") <+> pprDictsTheta insts
3072 , show_fixes fixes2 ]
3075 fix1 loc = sep [ ptext SLIT("add") <+> pprDictsTheta insts
3076 <+> ptext SLIT("to the context of"),
3077 nest 2 (ppr (instLocOrigin loc)) ]
3078 -- I'm not sure it helps to add the location
3079 -- nest 2 (ptext SLIT("at") <+> ppr (instLocSpan loc)) ]
3081 fixes2 | null instance_dicts = []
3082 | otherwise = [sep [ptext SLIT("add an instance declaration for"),
3083 pprDictsTheta instance_dicts]]
3084 instance_dicts = [d | d <- insts, isClassDict d, not (isTyVarDict d)]
3085 -- Insts for which it is worth suggesting an adding an instance declaration
3086 -- Exclude implicit parameters, and tyvar dicts
3088 show_fixes :: [SDoc] -> SDoc
3089 show_fixes [] = empty
3090 show_fixes (f:fs) = sep [ptext SLIT("Possible fix:"),
3091 nest 2 (vcat (f : map (ptext SLIT("or") <+>) fs))]
3093 addTopAmbigErrs dicts
3094 -- Divide into groups that share a common set of ambiguous tyvars
3095 = ifErrsM (return ()) $ -- Only report ambiguity if no other errors happened
3096 -- See Note [Avoiding spurious errors]
3097 mapM_ report (equivClasses cmp [(d, tvs_of d) | d <- tidy_dicts])
3099 (tidy_env, tidy_dicts) = tidyInsts dicts
3101 tvs_of :: Inst -> [TcTyVar]
3102 tvs_of d = varSetElems (tyVarsOfInst d)
3103 cmp (_,tvs1) (_,tvs2) = tvs1 `compare` tvs2
3105 report :: [(Inst,[TcTyVar])] -> TcM ()
3106 report pairs@((inst,tvs) : _) -- The pairs share a common set of ambiguous tyvars
3107 = mkMonomorphismMsg tidy_env tvs `thenM` \ (tidy_env, mono_msg) ->
3108 setSrcSpan (instSpan inst) $
3109 -- the location of the first one will do for the err message
3110 addErrTcM (tidy_env, msg $$ mono_msg)
3112 dicts = map fst pairs
3113 msg = sep [text "Ambiguous type variable" <> plural tvs <+>
3114 pprQuotedList tvs <+> in_msg,
3115 nest 2 (pprDictsInFull dicts)]
3116 in_msg = text "in the constraint" <> plural dicts <> colon
3117 report [] = panic "addTopAmbigErrs"
3120 mkMonomorphismMsg :: TidyEnv -> [TcTyVar] -> TcM (TidyEnv, Message)
3121 -- There's an error with these Insts; if they have free type variables
3122 -- it's probably caused by the monomorphism restriction.
3123 -- Try to identify the offending variable
3124 -- ASSUMPTION: the Insts are fully zonked
3125 mkMonomorphismMsg tidy_env inst_tvs
3126 = do { dflags <- getDOpts
3127 ; (tidy_env, docs) <- findGlobals (mkVarSet inst_tvs) tidy_env
3128 ; return (tidy_env, mk_msg dflags docs) }
3130 mk_msg _ _ | any isRuntimeUnk inst_tvs
3131 = vcat [ptext SLIT("Cannot resolve unknown runtime types:") <+>
3132 (pprWithCommas ppr inst_tvs),
3133 ptext SLIT("Use :print or :force to determine these types")]
3134 mk_msg _ [] = ptext SLIT("Probable fix: add a type signature that fixes these type variable(s)")
3135 -- This happens in things like
3136 -- f x = show (read "foo")
3137 -- where monomorphism doesn't play any role
3139 = vcat [ptext SLIT("Possible cause: the monomorphism restriction applied to the following:"),
3141 monomorphism_fix dflags]
3143 monomorphism_fix :: DynFlags -> SDoc
3144 monomorphism_fix dflags
3145 = ptext SLIT("Probable fix:") <+> vcat
3146 [ptext SLIT("give these definition(s) an explicit type signature"),
3147 if dopt Opt_MonomorphismRestriction dflags
3148 then ptext SLIT("or use -fno-monomorphism-restriction")
3149 else empty] -- Only suggest adding "-fno-monomorphism-restriction"
3150 -- if it is not already set!
3152 warnDefault ups default_ty
3153 = doptM Opt_WarnTypeDefaults `thenM` \ warn_flag ->
3154 addInstCtxt (instLoc (head (dicts))) (warnTc warn_flag warn_msg)
3156 dicts = [d | (d,_,_) <- ups]
3159 (_, tidy_dicts) = tidyInsts dicts
3160 warn_msg = vcat [ptext SLIT("Defaulting the following constraint(s) to type") <+>
3161 quotes (ppr default_ty),
3162 pprDictsInFull tidy_dicts]
3164 reduceDepthErr n stack
3165 = vcat [ptext SLIT("Context reduction stack overflow; size =") <+> int n,
3166 ptext SLIT("Use -fcontext-stack=N to increase stack size to N"),
3167 nest 4 (pprStack stack)]
3169 pprStack stack = vcat (map pprInstInFull stack)