TcCanonical.hs 66.7 KB
Newer Older
1 2
{-# LANGUAGE CPP #-}

3
module TcCanonical(
4 5 6 7 8
     canonicalize,
     unifyDerived,

     StopOrContinue(..), stopWith, continueWith
  ) where
9 10 11 12 13

#include "HsVersions.h"

import TcRnTypes
import TcType
14
import Type
dreixel's avatar
dreixel committed
15
import Kind
16 17
import TcFlatten
import TcSMonad
18
import TcEvidence
19 20 21
import Class
import TyCon
import TypeRep
22 23 24
import Coercion
import FamInstEnv ( FamInstEnvs )
import FamInst ( tcTopNormaliseNewTypeTF_maybe )
25
import Var
26
import Name( isSystemName )
27
import OccName( OccName )
28
import Outputable
29
import DynFlags( DynFlags )
30
import VarSet
31
import RdrName
32
import DataCon ( dataConName )
33

34
import Pair
35
import Util
36
import Bag
37 38
import MonadUtils ( zipWith3M, zipWith3M_ )
import Data.List  ( zip4 )
39
import BasicTypes
40
import FastString
41

Austin Seipp's avatar
Austin Seipp committed
42 43 44 45 46 47
{-
************************************************************************
*                                                                      *
*                      The Canonicaliser                               *
*                                                                      *
************************************************************************
48

49 50
Note [Canonicalization]
~~~~~~~~~~~~~~~~~~~~~~~
51

52
Canonicalization converts a simple constraint to a canonical form. It is
53 54 55
unary (i.e. treats individual constraints one at a time), does not do
any zonking, but lives in TcS monad because it needs to create fresh
variables (for flattening) and consult the inerts (for efficiency).
56

57
The execution plan for canonicalization is the following:
Simon Peyton Jones's avatar
Simon Peyton Jones committed
58 59

  1) Decomposition of equalities happens as necessary until we reach a
60
     variable or type family in one side. There is no decomposition step
Simon Peyton Jones's avatar
Simon Peyton Jones committed
61
     for other forms of constraints.
62

Simon Peyton Jones's avatar
Simon Peyton Jones committed
63 64 65 66
  2) If, when we decompose, we discover a variable on the head then we
     look at inert_eqs from the current inert for a substitution for this
     variable and contine decomposing. Hence we lazily apply the inert
     substitution if it is needed.
67

68 69
  3) If no more decomposition is possible, we deeply apply the substitution
     from the inert_eqs and continue with flattening.
70

Simon Peyton Jones's avatar
Simon Peyton Jones committed
71 72 73 74 75
  4) During flattening, we examine whether we have already flattened some
     function application by looking at all the CTyFunEqs with the same
     function in the inert set. The reason for deeply applying the inert
     substitution at step (3) is to maximise our chances of matching an
     already flattened family application in the inert.
76

Simon Peyton Jones's avatar
Simon Peyton Jones committed
77 78
The net result is that a constraint coming out of the canonicalization
phase cannot be rewritten any further from the inerts (but maybe /it/ can
79 80
rewrite an inert or still interact with an inert in a further phase in the
simplifier.
dimitris's avatar
dimitris committed
81

82
Note [Caching for canonicals]
Simon Peyton Jones's avatar
Simon Peyton Jones committed
83
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
84 85 86 87
Our plan with pre-canonicalization is to be able to solve a constraint
really fast from existing bindings in TcEvBinds. So one may think that
the condition (isCNonCanonical) is not necessary.  However consider
the following setup:
88

Simon Peyton Jones's avatar
Simon Peyton Jones committed
89 90
InertSet = { [W] d1 : Num t }
WorkList = { [W] d2 : Num t, [W] c : t ~ Int}
91

92 93 94 95 96
Now, we prioritize equalities, but in our concrete example
(should_run/mc17.hs) the first (d2) constraint is dealt with first,
because (t ~ Int) is an equality that only later appears in the
worklist since it is pulled out from a nested implication
constraint. So, let's examine what happens:
Simon Peyton Jones's avatar
Simon Peyton Jones committed
97

98 99
   - We encounter work item (d2 : Num t)

Simon Peyton Jones's avatar
Simon Peyton Jones committed
100
   - Nothing is yet in EvBinds, so we reach the interaction with inerts
101
     and set:
Simon Peyton Jones's avatar
Simon Peyton Jones committed
102
              d2 := d1
103 104
    and we discard d2 from the worklist. The inert set remains unaffected.

105 106 107
   - Now the equation ([W] c : t ~ Int) is encountered and kicks-out
     (d1 : Num t) from the inerts.  Then that equation gets
     spontaneously solved, perhaps. We end up with:
108
        InertSet : { [G] c : t ~ Int }
Simon Peyton Jones's avatar
Simon Peyton Jones committed
109
        WorkList : { [W] d1 : Num t}
110

111 112
   - Now we examine (d1), we observe that there is a binding for (Num
     t) in the evidence binds and we set:
Simon Peyton Jones's avatar
Simon Peyton Jones committed
113
             d1 := d2
114 115
     and end up in a loop!

116 117 118 119 120 121 122 123
Now, the constraints that get kicked out from the inert set are always
Canonical, so by restricting the use of the pre-canonicalizer to
NonCanonical constraints we eliminate this danger. Moreover, for
canonical constraints we already have good caching mechanisms
(effectively the interaction solver) and we are interested in reducing
things like superclasses of the same non-canonical constraint being
generated hence I don't expect us to lose a lot by introducing the
(isCNonCanonical) restriction.
124

125 126 127 128 129 130 131
A similar situation can arise in TcSimplify, at the end of the
solve_wanteds function, where constraints from the inert set are
returned as new work -- our substCt ensures however that if they are
not rewritten by subst, they remain canonical and hence we will not
attempt to solve them from the EvBinds. If on the other hand they did
get rewritten and are now non-canonical they will still not match the
EvBinds, so we are again good.
Austin Seipp's avatar
Austin Seipp committed
132
-}
133

134 135 136
-- Top-level canonicalization
-- ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

137
canonicalize :: Ct -> TcS (StopOrContinue Ct)
138
canonicalize ct@(CNonCanonical { cc_ev = ev })
139
  = do { traceTcS "canonicalize (non-canonical)" (ppr ct)
140
       ; {-# SCC "canEvVar" #-}
141
         canEvNC ev }
142

143
canonicalize (CDictCan { cc_ev = ev
144 145
                       , cc_class  = cls
                       , cc_tyargs = xis })
146
  = {-# SCC "canClass" #-}
147 148
    canClass ev cls xis -- Do not add any superclasses
canonicalize (CTyEqCan { cc_ev = ev
149
                       , cc_tyvar  = tv
150 151
                       , cc_rhs    = xi
                       , cc_eq_rel = eq_rel })
152
  = {-# SCC "canEqLeafTyVarEq" #-}
153 154 155
    canEqNC ev eq_rel (mkTyVarTy tv) xi
      -- NB: Don't use canEqTyVar because that expects flattened types,
      -- and tv and xi may not be flat w.r.t. an updated inert set
156

157
canonicalize (CFunEqCan { cc_ev = ev
158 159
                        , cc_fun    = fn
                        , cc_tyargs = xis1
160
                        , cc_fsk    = fsk })
Simon Peyton Jones's avatar
Simon Peyton Jones committed
161
  = {-# SCC "canEqLeafFunEq" #-}
162
    canCFunEqCan ev fn xis1 fsk
163

164 165
canonicalize (CIrredEvCan { cc_ev = ev })
  = canIrred ev
thomasw's avatar
thomasw committed
166 167
canonicalize (CHoleCan { cc_ev = ev, cc_occ = occ, cc_hole = hole })
  = canHole ev occ hole
168

169
canEvNC :: CtEvidence -> TcS (StopOrContinue Ct)
Simon Peyton Jones's avatar
Simon Peyton Jones committed
170
-- Called only for non-canonical EvVars
171
canEvNC ev
172
  = case classifyPredType (ctEvPred ev) of
173 174 175 176 177 178
      ClassPred cls tys     -> do traceTcS "canEvNC:cls" (ppr cls <+> ppr tys)
                                  canClassNC ev cls tys
      EqPred eq_rel ty1 ty2 -> do traceTcS "canEvNC:eq" (ppr ty1 $$ ppr ty2)
                                  canEqNC    ev eq_rel ty1 ty2
      IrredPred {}          -> do traceTcS "canEvNC:irred" (ppr (ctEvPred ev))
                                  canIrred   ev
Austin Seipp's avatar
Austin Seipp committed
179 180 181 182 183 184 185
{-
************************************************************************
*                                                                      *
*                      Class Canonicalization
*                                                                      *
************************************************************************
-}
186

Simon Peyton Jones's avatar
Simon Peyton Jones committed
187
canClass, canClassNC
188
   :: CtEvidence
189
   -> Class -> [Type] -> TcS (StopOrContinue Ct)
Simon Peyton Jones's avatar
Simon Peyton Jones committed
190
-- Precondition: EvVar is class evidence
191 192 193 194 195 196

-- The canClassNC version is used on non-canonical constraints
-- and adds superclasses.  The plain canClass version is used
-- for already-canonical class constraints (but which might have
-- been subsituted or somthing), and hence do not need superclasses

197 198
canClassNC ev cls tys
  = canClass ev cls tys
199 200
    `andWhenContinue` emitSuperclasses

201
canClass ev cls tys
202 203
  =   -- all classes do *nominal* matching
    ASSERT2( ctEvRole ev == Nominal, ppr ev $$ ppr cls $$ ppr tys )
204
    do { (xis, cos) <- flattenManyNom ev tys
Joachim Breitner's avatar
Joachim Breitner committed
205
       ; let co = mkTcTyConAppCo Nominal (classTyCon cls) cos
206
             xi = mkClassPred cls xis
207 208
             mk_ct new_ev = CDictCan { cc_ev = new_ev
                                     , cc_tyargs = xis, cc_class = cls }
209
       ; mb <- rewriteEvidence ev xi co
Simon Peyton Jones's avatar
Simon Peyton Jones committed
210
       ; traceTcS "canClass" (vcat [ ppr ev <+> ppr cls <+> ppr tys
Simon Peyton Jones's avatar
Simon Peyton Jones committed
211
                                   , ppr xi, ppr mb ])
212
       ; return (fmap mk_ct mb) }
dimitris's avatar
dimitris committed
213

214
emitSuperclasses :: Ct -> TcS (StopOrContinue Ct)
215
emitSuperclasses ct@(CDictCan { cc_ev = ev , cc_tyargs = xis_new, cc_class = cls })
Simon Peyton Jones's avatar
Simon Peyton Jones committed
216 217
            -- Add superclasses of this one here, See Note [Adding superclasses].
            -- But only if we are not simplifying the LHS of a rule.
218
 = do { newSCWorkFromFlavored ev cls xis_new
Simon Peyton Jones's avatar
Simon Peyton Jones committed
219
      -- Arguably we should "seq" the coercions if they are derived,
220
      -- as we do below for emit_kind_constraint, to allow errors in
Simon Peyton Jones's avatar
Simon Peyton Jones committed
221
      -- superclasses to be executed if deferred to runtime!
222 223
      ; continueWith ct }
emitSuperclasses _ = panic "emit_superclasses of non-class!"
224

225 226
{- Note [Adding superclasses]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
227 228 229 230 231
Since dictionaries are canonicalized only once in their lifetime, the
place to add their superclasses is canonicalisation.  See Note [Add
superclasses only during canonicalisation].  Here is what we do:

  Givens:   Add all their superclasses as Givens.
232
            They may be needed to prove Wanteds.
233

234 235
  Wanteds/Derived:
            Add all their superclasses as Derived.
236 237
            The sole reason is to expose functional dependencies
            in superclasses or equality superclasses.
238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257

Examples of how adding superclasses as Derived is useful

    --- Example 1
        class C a b | a -> b
    Suppose we want to solve
         [G] C a b
         [W] C a beta
    Then adding [D] beta~b will let us solve it.

    -- Example 2 (similar but using a type-equality superclass)
        class (F a ~ b) => C a b
    And try to sllve:
         [G] C a b
         [W] C a beta
    Follow the superclass rules to add
         [G] F a ~ b
         [D] F a ~ beta
    Now we we get [D] beta ~ b, and can solve that.

258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275
    -- Example (tcfail138)
      class L a b | a -> b
      class (G a, L a b) => C a b

      instance C a b' => G (Maybe a)
      instance C a b  => C (Maybe a) a
      instance L (Maybe a) a

    When solving the superclasses of the (C (Maybe a) a) instance, we get
      [G] C a b, and hance by superclasses, [G] G a, [G] L a b
      [W] G (Maybe a)
    Use the instance decl to get
      [W] C a beta
    Generate its derived superclass
      [D] L a beta.  Now using fundeps, combine with [G] L a b to get
      [D] beta ~ b
    which is what we want.

276
---------- Historical note -----------
277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295
Example of why adding superclass of a Wanted as a Given would
be terrible, see Note [Do not add superclasses of solved dictionaries]
in TcSMonad, which has this example:
        class Ord a => C a where
        instance Ord [a] => C [a] where ...
Suppose we are trying to solve
  [G] d1 : Ord a
  [W] d2 : C [a]
If we (bogusly) added the superclass of d2 as Gievn we'd have
  [G] d1 : Ord a
  [W] d2 : C [a]
  [G] d3 : Ord [a]   -- Superclass of d2, bogus

Then we'll use the instance decl to give
  [G] d1 : Ord a     Solved: d2 : C [a] = $dfCList d4
  [G] d3 : Ord [a]   -- Superclass of d2, bogus
  [W] d4: Ord [a]

ANd now we could bogusly solve d4 from d3.
296
---------- End of historical note -----------
297 298 299 300 301 302 303 304

Note [Add superclasses only during canonicalisation]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
We add superclasses only during canonicalisation, on the passage
from CNonCanonical to CDictCan.  A class constraint can be repeatedly
rewritten, and there's no point in repeatedly adding its superclasses.

Here's a serious, but now out-dated example, from Trac #4497:
Simon Peyton Jones's avatar
Simon Peyton Jones committed
305

306 307 308
   class Num (RealOf t) => Normed t
   type family RealOf x

Simon Peyton Jones's avatar
Simon Peyton Jones committed
309
Assume the generated wanted constraint is:
310 311 312
   [W] RealOf e ~ e
   [W] Normed e

Simon Peyton Jones's avatar
Simon Peyton Jones committed
313
If we were to be adding the superclasses during simplification we'd get:
314 315 316 317
   [W] RealOf e ~ e
   [W] Normed e
   [D] RealOf e ~ fuv
   [D] Num fuv
Simon Peyton Jones's avatar
Simon Peyton Jones committed
318
==>
319
   e := fuv, Num fuv, Normed fuv, RealOf fuv ~ fuv
Simon Peyton Jones's avatar
Simon Peyton Jones committed
320

321 322 323
While looks exactly like our original constraint. If we add the
superclass of (Normed fuv) again we'd loop.  By adding superclasses
definitely only once, during canonicalisation, this situation can't
324
happen.
325 326 327 328

Mind you, now that Wanteds cannot rewrite Derived, I think this particular
situation can't happen.
  -}
329

330
newSCWorkFromFlavored :: CtEvidence -> Class -> [Xi] -> TcS ()
331
-- Returns superclasses, see Note [Adding superclasses]
332
newSCWorkFromFlavored flavor cls xis
333
  | CtGiven { ctev_evar = evar, ctev_loc = loc } <- flavor
Simon Peyton Jones's avatar
Simon Peyton Jones committed
334 335 336 337 338 339 340 341
  = do { let size = sizeTypes xis
             loc' | isCTupleClass cls
                  = loc   -- For tuple predicates, just take them apart, without
                          -- adding their (large) size into the chain.  When we
                          -- get down to a base predicate, we'll include its size.
                          -- Trac #10335
                  | otherwise
                  = case ctLocOrigin loc of
342 343 344 345 346 347 348 349 350
                       GivenOrigin InstSkol
                         -> loc { ctl_origin = GivenOrigin (InstSC size) }
                       GivenOrigin (InstSC n)
                         -> loc { ctl_origin = GivenOrigin (InstSC (n `max` size)) }
                       _ -> loc
                    -- See Note [Solving superclass constraints] in TcInstDcls
                    -- for explantation of loc'

       ; given_evs <- newGivenEvVars loc' (mkEvScSelectors (EvId evar) cls xis)
351
       ; emitWorkNC given_evs }
dimitris's avatar
dimitris committed
352 353

  | isEmptyVarSet (tyVarsOfTypes xis)
354
  = return () -- Wanteds with no variables yield no deriveds.
355
              -- See Note [Improvement from Ground Wanteds]
356

357
  | otherwise -- Derived case, just add those SC that can lead to improvement.
Simon Peyton Jones's avatar
Simon Peyton Jones committed
358
  = do { let sc_rec_theta = transSuperClasses cls xis
359
             impr_theta   = filter isImprovementPred sc_rec_theta
360
             loc          = ctEvLoc flavor
361
       ; traceTcS "newSCWork/Derived" $ text "impr_theta =" <+> ppr impr_theta
362
       ; emitNewDeriveds loc impr_theta }
363 364


Austin Seipp's avatar
Austin Seipp committed
365 366 367 368 369 370 371
{-
************************************************************************
*                                                                      *
*                      Irreducibles canonicalization
*                                                                      *
************************************************************************
-}
372

373
canIrred :: CtEvidence -> TcS (StopOrContinue Ct)
374
-- Precondition: ty not a tuple and no other evidence form
375
canIrred old_ev
Simon Peyton Jones's avatar
Simon Peyton Jones committed
376 377
  = do { let old_ty = ctEvPred old_ev
       ; traceTcS "can_pred" (text "IrredPred = " <+> ppr old_ty)
378
       ; (xi,co) <- flatten FM_FlattenAll old_ev old_ty -- co :: xi ~ old_ty
379
       ; rewriteEvidence old_ev xi co `andWhenContinue` \ new_ev ->
380 381
    do { -- Re-classify, in case flattening has improved its shape
       ; case classifyPredType (ctEvPred new_ev) of
382 383 384 385
           ClassPred cls tys     -> canClassNC new_ev cls tys
           EqPred eq_rel ty1 ty2 -> canEqNC new_ev eq_rel ty1 ty2
           _                     -> continueWith $
                                    CIrredEvCan { cc_ev = new_ev } } }
386

thomasw's avatar
thomasw committed
387 388
canHole :: CtEvidence -> OccName -> HoleSort -> TcS (StopOrContinue Ct)
canHole ev occ hole_sort
389 390
  = do { let ty = ctEvPred ev
       ; (xi,co) <- flatten FM_SubstOnly ev ty -- co :: xi ~ ty
391 392 393 394 395
       ; rewriteEvidence ev xi co `andWhenContinue` \ new_ev ->
    do { emitInsoluble (CHoleCan { cc_ev = new_ev
                                 , cc_occ = occ
                                 , cc_hole = hole_sort })
       ; stopWith new_ev "Emit insoluble hole" } }
396

Austin Seipp's avatar
Austin Seipp committed
397 398 399 400 401 402
{-
************************************************************************
*                                                                      *
*        Equalities
*                                                                      *
************************************************************************
403 404 405 406 407 408 409 410 411 412 413 414 415

Note [Canonicalising equalities]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
In order to canonicalise an equality, we look at the structure of the
two types at hand, looking for similarities. A difficulty is that the
types may look dissimilar before flattening but similar after flattening.
However, we don't just want to jump in and flatten right away, because
this might be wasted effort. So, after looking for similarities and failing,
we flatten and then try again. Of course, we don't want to loop, so we
track whether or not we've already flattened.

It is conceivable to do a better job at tracking whether or not a type
is flattened, but this is left as future work. (Mar '15)
Austin Seipp's avatar
Austin Seipp committed
416
-}
417

418 419
canEqNC :: CtEvidence -> EqRel -> Type -> Type -> TcS (StopOrContinue Ct)
canEqNC ev eq_rel ty1 ty2
420
  = can_eq_nc False ev eq_rel ty1 ty1 ty2 ty2
421

422
can_eq_nc
423 424
   :: Bool            -- True => both types are flat
   -> CtEvidence
425
   -> EqRel
Austin Seipp's avatar
Austin Seipp committed
426 427
   -> Type -> Type    -- LHS, after and before type-synonym expansion, resp
   -> Type -> Type    -- RHS, after and before type-synonym expansion, resp
428
   -> TcS (StopOrContinue Ct)
429
can_eq_nc flat ev eq_rel ty1 ps_ty1 ty2 ps_ty2
Austin Seipp's avatar
Austin Seipp committed
430
  = do { traceTcS "can_eq_nc" $
431 432 433
         vcat [ ppr ev, ppr eq_rel, ppr ty1, ppr ps_ty1, ppr ty2, ppr ps_ty2 ]
       ; rdr_env <- getGlobalRdrEnvTcS
       ; fam_insts <- getFamInstEnvs
434
       ; can_eq_nc' flat rdr_env fam_insts ev eq_rel ty1 ps_ty1 ty2 ps_ty2 }
435 436

can_eq_nc'
437 438
   :: Bool           -- True => both input types are flattened
   -> GlobalRdrEnv   -- needed to see which newtypes are in scope
439 440 441 442 443 444
   -> FamInstEnvs    -- needed to unwrap data instances
   -> CtEvidence
   -> EqRel
   -> Type -> Type    -- LHS, after and before type-synonym expansion, resp
   -> Type -> Type    -- RHS, after and before type-synonym expansion, resp
   -> TcS (StopOrContinue Ct)
445 446

-- Expand synonyms first; see Note [Type synonyms and canonicalization]
447 448 449 450 451
can_eq_nc' flat _rdr_env _envs ev eq_rel ty1 ps_ty1 ty2 ps_ty2
  | Just ty1' <- tcView ty1 = can_eq_nc flat ev eq_rel ty1' ps_ty1 ty2  ps_ty2
  | Just ty2' <- tcView ty2 = can_eq_nc flat ev eq_rel ty1  ps_ty1 ty2' ps_ty2

-- need to check for reflexivity in the ReprEq case.
eir@cis.upenn.edu's avatar
eir@cis.upenn.edu committed
452
-- See Note [Eager reflexivity check]
453 454 455 456 457 458
can_eq_nc' _flat _rdr_env _envs ev ReprEq ty1 _ ty2 _
  | ty1 `eqType` ty2
  = canEqReflexive ev ReprEq ty1

-- When working with ReprEq, unwrap newtypes.
can_eq_nc' _flat rdr_env envs ev ReprEq ty1 _ ty2 ps_ty2
459 460
  | Just (co, ty1') <- tcTopNormaliseNewTypeTF_maybe envs rdr_env ty1
  = can_eq_newtype_nc rdr_env ev NotSwapped co ty1 ty1' ty2 ps_ty2
461
can_eq_nc' _flat rdr_env envs ev ReprEq ty1 ps_ty1 ty2 _
462 463
  | Just (co, ty2') <- tcTopNormaliseNewTypeTF_maybe envs rdr_env ty2
  = can_eq_newtype_nc rdr_env ev IsSwapped  co ty2 ty2' ty1 ps_ty1
464 465 466 467 468 469

----------------------
-- Otherwise try to decompose
----------------------

-- Literals
470
can_eq_nc' _flat _rdr_env _envs ev eq_rel ty1@(LitTy l1) _ (LitTy l2) _
471
 | l1 == l2
472 473
  = do { setEvBindIfWanted ev (EvCoercion $
                               mkTcReflCo (eqRelRole eq_rel) ty1)
474
       ; stopWith ev "Equal LitTy" }
475

Simon Peyton Jones's avatar
Simon Peyton Jones committed
476 477
-- Try to decompose type constructor applications
-- Including FunTy (s -> t)
478 479 480 481 482
can_eq_nc' _flat _rdr_env _envs ev eq_rel ty1 _ ty2 _
  | Just (tc1, tys1) <- tcSplitTyConApp_maybe ty1
  , Just (tc2, tys2) <- tcSplitTyConApp_maybe ty2
  , not (isTypeFamilyTyCon tc1)
  , not (isTypeFamilyTyCon tc2)
483
  = canTyConApp ev eq_rel tc1 tys1 tc2 tys2
484

485 486
can_eq_nc' _flat _rdr_env _envs ev eq_rel
           s1@(ForAllTy {}) _ s2@(ForAllTy {}) _
487
 | CtWanted { ctev_loc = loc, ctev_evar = orig_ev } <- ev
488 489
 = do { let (tvs1,body1) = tcSplitForAllTys s1
            (tvs2,body2) = tcSplitForAllTys s2
Simon Peyton Jones's avatar
Simon Peyton Jones committed
490
      ; if not (equalLength tvs1 tvs2) then
491
          canEqHardFailure ev eq_rel s1 s2
492
        else
493
          do { traceTcS "Creating implication for polytype equality" $ ppr ev
494 495
             ; ev_term <- deferTcSForAllEq (eqRelRole eq_rel)
                                           loc (tvs1,body1) (tvs2,body2)
496
             ; setWantedEvBind orig_ev ev_term
497
             ; stopWith ev "Deferred polytype equality" } }
498
 | otherwise
Simon Peyton Jones's avatar
Simon Peyton Jones committed
499
 = do { traceTcS "Ommitting decomposition of given polytype equality" $
500
        pprEq s1 s2    -- See Note [Do not decompose given polytype equalities]
501
      ; stopWith ev "Discard given polytype equality" }
502

503
-- See Note [Canonicalising type applications] about why we require flat types
eir@cis.upenn.edu's avatar
eir@cis.upenn.edu committed
504
can_eq_nc' True _rdr_env _envs ev eq_rel (AppTy t1 s1) _ ty2 _
505
  | Just (t2, s2) <- tcSplitAppTy_maybe ty2
eir@cis.upenn.edu's avatar
eir@cis.upenn.edu committed
506 507
  = can_eq_app ev eq_rel t1 s1 t2 s2
can_eq_nc' True _rdr_env _envs ev eq_rel ty1 _ (AppTy t2 s2) _
508
  | Just (t1, s1) <- tcSplitAppTy_maybe ty1
eir@cis.upenn.edu's avatar
eir@cis.upenn.edu committed
509
  = can_eq_app ev eq_rel t1 s1 t2 s2
510 511 512 513 514 515

-- No similarity in type structure detected. Flatten and try again!
can_eq_nc' False rdr_env envs ev eq_rel _ ps_ty1 _ ps_ty2
  = do { (xi1, co1) <- flatten FM_FlattenAll ev ps_ty1
       ; (xi2, co2) <- flatten FM_FlattenAll ev ps_ty2
       ; rewriteEqEvidence ev eq_rel NotSwapped xi1 xi2 co1 co2
516
         `andWhenContinue` \ new_ev ->
517 518 519 520 521 522 523 524 525 526 527 528 529
         can_eq_nc' True rdr_env envs new_ev eq_rel xi1 xi1 xi2 xi2 }

-- Type variable on LHS or RHS are last. We want only flat types sent
-- to canEqTyVar.
-- See also Note [No top-level newtypes on RHS of representational equalities]
can_eq_nc' True _rdr_env _envs ev eq_rel (TyVarTy tv1) _ _ ps_ty2
  = canEqTyVar ev eq_rel NotSwapped tv1 ps_ty2
can_eq_nc' True _rdr_env _envs ev eq_rel _ ps_ty1 (TyVarTy tv2) _
  = canEqTyVar ev eq_rel IsSwapped  tv2 ps_ty1

-- We've flattened and the types don't match. Give up.
can_eq_nc' True _rdr_env _envs ev eq_rel _ ps_ty1 _ ps_ty2
  = canEqHardFailure ev eq_rel ps_ty1 ps_ty2
530

531
{-
532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549
Note [Newtypes can blow the stack]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Suppose we have

  newtype X = MkX (Int -> X)
  newtype Y = MkY (Int -> Y)

and now wish to prove

  [W] X ~R Y

This Wanted will loop, expanding out the newtypes ever deeper looking
for a solid match or a solid discrepancy. Indeed, there is something
appropriate to this looping, because X and Y *do* have the same representation,
in the limit -- they're both (Fix ((->) Int)). However, no finitely-sized
coercion will ever witness it. This loop won't actually cause GHC to hang,
though, because we check our depth when unwrapping newtypes.

550 551 552 553 554 555 556 557 558
Note [Eager reflexivity check]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Suppose we have

  newtype X = MkX (Int -> X)

and

  [W] X ~R X
559

560 561 562 563 564
Naively, we would start unwrapping X and end up in a loop. Instead,
we do this eager reflexivity check. This is necessary only for representational
equality because the flattener technology deals with the similar case
(recursive type families) for nominal equality.

565 566
Note that this check does not catch all cases, but it will catch the cases
we're most worried about, types like X above that are actually inhabited.
567

eir@cis.upenn.edu's avatar
eir@cis.upenn.edu committed
568
Here's another place where this reflexivity check is key:
569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595
Consider trying to prove (f a) ~R (f a). The AppTys in there can't
be decomposed, because representational equality isn't congruent with respect
to AppTy. So, when canonicalising the equality above, we get stuck and
would normally produce a CIrredEvCan. However, we really do want to
be able to solve (f a) ~R (f a). So, in the representational case only,
we do a reflexivity check.

(This would be sound in the nominal case, but unnecessary, and I [Richard
E.] am worried that it would slow down the common case.)
-}

------------------------
-- | We're able to unwrap a newtype. Update the bits accordingly.
can_eq_newtype_nc :: GlobalRdrEnv
                  -> CtEvidence           -- ^ :: ty1 ~ ty2
                  -> SwapFlag
                  -> TcCoercion           -- ^ :: ty1 ~ ty1'
                  -> TcType               -- ^ ty1
                  -> TcType               -- ^ ty1'
                  -> TcType               -- ^ ty2
                  -> TcType               -- ^ ty2, with type synonyms
                  -> TcS (StopOrContinue Ct)
can_eq_newtype_nc rdr_env ev swapped co ty1 ty1' ty2 ps_ty2
  = do { traceTcS "can_eq_newtype_nc" $
         vcat [ ppr ev, ppr swapped, ppr co, ppr ty1', ppr ty2 ]

         -- check for blowing our stack:
596 597 598
         -- See Note [Newtypes can blow the stack]
       ; checkReductionDepth (ctEvLoc ev) ty1
       ; markDataConsAsUsed rdr_env (tyConAppTyCon ty1)
599 600 601 602 603 604
           -- we have actually used the newtype constructor here, so
           -- make sure we don't warn about importing it!

       ; rewriteEqEvidence ev ReprEq swapped ty1' ps_ty2
                           (mkTcSymCo co) (mkTcReflCo Representational ps_ty2)
         `andWhenContinue` \ new_ev ->
605
         can_eq_nc False new_ev ReprEq ty1' ty1' ty2 ps_ty2 }
606 607 608 609 610

-- | Mark all the datacons of the given 'TyCon' as used in this module,
-- avoiding "redundant import" warnings.
markDataConsAsUsed :: GlobalRdrEnv -> TyCon -> TcS ()
markDataConsAsUsed rdr_env tc = addUsedRdrNamesTcS
611
  [ greUsedRdrName gre
612
  | dc <- tyConDataCons tc
613 614
  , gre : _  <- return $ lookupGRE_Name rdr_env (dataConName dc)
  , not (isLocalGRE gre) ]
615

616
---------
eir@cis.upenn.edu's avatar
eir@cis.upenn.edu committed
617
-- ^ Decompose a type application.
618
-- All input types must be flat. See Note [Canonicalising type applications]
eir@cis.upenn.edu's avatar
eir@cis.upenn.edu committed
619 620
can_eq_app :: CtEvidence       -- :: s1 t1 ~r s2 t2
           -> EqRel            -- r
621 622 623
           -> Xi -> Xi         -- s1 t1
           -> Xi -> Xi         -- s2 t2
           -> TcS (StopOrContinue Ct)
eir@cis.upenn.edu's avatar
eir@cis.upenn.edu committed
624 625

-- AppTys only decompose for nominal equality, so this case just leads
eir@cis.upenn.edu's avatar
eir@cis.upenn.edu committed
626
-- to an irreducible constraint; see typecheck/should_compile/T10494
Simon Peyton Jones's avatar
Simon Peyton Jones committed
627
-- See Note [Decomposing equality], note {4}
eir@cis.upenn.edu's avatar
eir@cis.upenn.edu committed
628 629 630
can_eq_app ev ReprEq _ _ _ _
  = do { traceTcS "failing to decompose representational AppTy equality" (ppr ev)
       ; continueWith (CIrredEvCan { cc_ev = ev }) }
eir@cis.upenn.edu's avatar
eir@cis.upenn.edu committed
631 632
          -- no need to call canEqFailure, because that flattens, and the
          -- types involved here are already flat
eir@cis.upenn.edu's avatar
eir@cis.upenn.edu committed
633 634

can_eq_app ev NomEq s1 t1 s2 t2
635
  | CtDerived { ctev_loc = loc } <- ev
636
  = do { emitNewDerivedEq loc (mkTcEqPred t1 t2)
637 638 639 640 641 642 643
       ; canEqNC ev NomEq s1 s2 }
  | CtWanted { ctev_evar = evar, ctev_loc = loc } <- ev
  = do { ev_s <- newWantedEvVarNC loc (mkTcEqPred s1 s2)
       ; co_t <- unifyWanted loc Nominal t1 t2
       ; let co = mkTcAppCo (ctEvCoercion ev_s) co_t
       ; setWantedEvBind evar (EvCoercion co)
       ; canEqNC ev_s NomEq s1 s2 }
644 645
  | CtGiven { ctev_evar = evar, ctev_loc = loc } <- ev
  = do { let co   = mkTcCoVarCo evar
646 647 648 649 650 651 652 653
             co_s = mkTcLRCo CLeft  co
             co_t = mkTcLRCo CRight co
       ; evar_s <- newGivenEvVar loc (mkTcEqPred s1 s2, EvCoercion co_s)
       ; evar_t <- newGivenEvVar loc (mkTcEqPred t1 t2, EvCoercion co_t)
       ; emitWorkNC [evar_t]
       ; canEqNC evar_s NomEq s1 s2 }
  | otherwise  -- Can't happen
  = error "can_eq_app"
654

655
------------------------
656 657 658 659
canTyConApp :: CtEvidence -> EqRel
            -> TyCon -> [TcType]
            -> TyCon -> [TcType]
            -> TcS (StopOrContinue Ct)
660
-- See Note [Decomposing TyConApps]
661
canTyConApp ev eq_rel tc1 tys1 tc2 tys2
662
  | tc1 == tc2
663
  , length tys1 == length tys2
664
  = do { inerts <- getTcSInerts
665 666
       ; if can_decompose inerts
         then do { traceTcS "canTyConApp"
667 668 669 670
                       (ppr ev $$ ppr eq_rel $$ ppr tc1 $$ ppr tys1 $$ ppr tys2)
                 ; canDecomposableTyConAppOK ev eq_rel tc1 tys1 tys2
                 ; stopWith ev "Decomposed TyConApp" }
         else canEqFailure ev eq_rel ty1 ty2 }
671

672 673
  -- Fail straight away for better error messages
  -- See Note [Use canEqFailure in canDecomposableTyConApp]
674 675
  | eq_rel == ReprEq && not (isGenerativeTyCon tc1 Representational &&
                             isGenerativeTyCon tc2 Representational)
676 677 678 679 680 681 682
  = canEqFailure ev eq_rel ty1 ty2
  | otherwise
  = canEqHardFailure ev eq_rel ty1 ty2
  where
    ty1 = mkTyConApp tc1 tys1
    ty2 = mkTyConApp tc2 tys2

683 684 685
    loc  = ctEvLoc ev
    pred = ctEvPred ev

686 687 688 689 690
     -- See Note [Decomposing equality]
    can_decompose inerts
      =  isInjectiveTyCon tc1 (eqRelRole eq_rel)
      || (ctEvFlavour ev /= Given && isEmptyBag (matchableGivens loc pred inerts))

691 692 693 694 695
{-
Note [Use canEqFailure in canDecomposableTyConApp]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
We must use canEqFailure, not canEqHardFailure here, because there is
the possibility of success if working with a representational equality.
eir@cis.upenn.edu's avatar
eir@cis.upenn.edu committed
696
Here is one case:
697 698 699 700 701

  type family TF a where TF Char = Bool
  data family DF a
  newtype instance DF Bool = MkDF Int

eir@cis.upenn.edu's avatar
eir@cis.upenn.edu committed
702
Suppose we are canonicalising (Int ~R DF (TF a)), where we don't yet
703 704
know `a`. This is *not* a hard failure, because we might soon learn
that `a` is, in fact, Char, and then the equality succeeds.
705

eir@cis.upenn.edu's avatar
eir@cis.upenn.edu committed
706 707
Here is another case:

eir@cis.upenn.edu's avatar
eir@cis.upenn.edu committed
708
  [G] Age ~R Int
eir@cis.upenn.edu's avatar
eir@cis.upenn.edu committed
709 710 711 712

where Age's constructor is not in scope. We don't want to report
an "inaccessible code" error in the context of this Given!

eir@cis.upenn.edu's avatar
eir@cis.upenn.edu committed
713 714 715 716 717 718 719 720 721 722
For example, see typecheck/should_compile/T10493, repeated here:

  import Data.Ord (Down)  -- no constructor

  foo :: Coercible (Down Int) Int => Down Int -> Int
  foo = coerce

That should compile, but only because we use canEqFailure and not
canEqHardFailure.

723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761
Note [Decomposing equality]
~~~~~~~~~~~~~~~~~~~~~~~~~~~
If we have a constraint (of any flavour and role) that looks like
T tys1 ~ T tys2, what can we conclude about tys1 and tys2? The answer,
of course, is "it depends". This Note spells it all out.

In this Note, "decomposition" refers to taking the constraint
  [fl] (T tys1 ~X T tys2)
(for some flavour fl and some role X) and replacing it with
  [fls'] (tys1 ~Xs' tys2)
where that notation indicates a list of new constraints, where the
new constraints may have different flavours and different roles.

The key property to consider is injectivity. When decomposing a Given the
decomposition is sound if and only if T is injective in all of its type
arguments. When decomposing a Wanted, the decomposition is sound (assuming the
correct roles in the produced equality constraints), but it may be a guess --
that is, an unforced decision by the constraint solver. Decomposing Wanteds
over injective TyCons does not entail guessing. But sometimes we want to
decompose a Wanted even when the TyCon involved is not injective! (See below.)

So, in broad strokes, we want this rule:

(*) Decompose a constraint (T tys1 ~X T tys2) if and only if T is injective
at role X.

Pursuing the details requires exploring three axes:
* Flavour: Given vs. Derived vs. Wanted
* Role: Nominal vs. Representational
* TyCon species: datatype vs. newtype vs. data family vs. type family vs. type variable

(So a type variable isn't a TyCon, but it's convenient to put the AppTy case
in the same table.)

Right away, we can say that Derived behaves just as Wanted for the purposes
of decomposition. The difference between Derived and Wanted is the handling of
evidence. Since decomposition in these cases isn't a matter of soundness but of
guessing, we want the same behavior regardless of evidence.

Simon Peyton Jones's avatar
Simon Peyton Jones committed
762 763 764 765 766 767
Here is a table (discussion following) detailing where decomposition of
   (T s1 ... sn) ~r (T t1 .. tn)
is allowed.  The first four lines (Data types ... type family) refer
to TyConApps with various TyCons T; the last line is for AppTy, where
there is presumably a type variable at the head, so it's actually
   (s s1 ... sn) ~r (t t1 .. tn)
768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794

NOMINAL               GIVEN                       WANTED

Datatype               YES                         YES
Newtype                YES                         YES
Data family            YES                         YES
Type family            YES, in injective args{1}   YES, in injective args{1}
Type variable          YES                         YES

REPRESENTATIONAL      GIVEN                       WANTED

Datatype               YES                         YES
Newtype                NO{2}                      MAYBE{2}
Data family            NO{3}                      MAYBE{3}
Type family             NO                          NO
Type variable          NO{4}                       NO{4}

{1}: Type families can be injective in some, but not all, of their arguments,
so we want to do partial decomposition. This is quite different than the way
other decomposition is done, where the decomposed equalities replace the original
one. We thus proceed much like we do with superclasses: emitting new Givens
when "decomposing" a partially-injective type family Given and new Deriveds
when "decomposing" a partially-injective type family Wanted. (As of the time of
writing, 13 June 2015, the implementation of injective type families has not
been merged, but it should be soon. Please delete this parenthetical if the
implementation is indeed merged.)

Simon Peyton Jones's avatar
Simon Peyton Jones committed
795
{2}: See Note [Decomposing newtypes at representational role]
796

Simon Peyton Jones's avatar
Simon Peyton Jones committed
797 798 799
{3}: Because of the possibility of newtype instances, we must treat
data families like newtypes. See also Note [Decomposing newtypes at
representational role]. See #10534 and test case
800
typecheck/should_fail/T10534.
801 802 803 804 805 806 807 808 809 810 811

{4}: Because type variables can stand in for newtypes, we conservatively do not
decompose AppTys over representational equality.

In the implementation of can_eq_nc and friends, we don't directly pattern
match using lines like in the tables above, as those tables don't cover
all cases (what about PrimTyCon? tuples?). Instead we just ask about injectivity,
boiling the tables above down to rule (*). The exceptions to rule (*) are for
injective type families, which are handled separately from other decompositions,
and the MAYBE entries above.

Simon Peyton Jones's avatar
Simon Peyton Jones committed
812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871
Note [Decomposing newtypes at representational role]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
This note discusses the 'newtype' line in the REPRESENTATIONAL table
in Note [Decomposing equality]. (At nominal role, newtypes are fully
decomposable.)

Here is a representative example of why representational equality over
newtypes is tricky:

  newtype Nt a = Mk Bool         -- NB: a is not used in the RHS,
  type role Nt representational  -- but the user gives it an R role anyway

If we have [W] Nt alpha ~R Nt beta, we *don't* want to decompose to
[W] alpha ~R beta, because it's possible that alpha and beta aren't
representationally equal. Here's another example.

  newtype Nt a = MkNt (Id a)
  type family Id a where Id a = a

  [W] Nt Int ~R Nt Age

Because of its use of a type family, Nt's parameter will get inferred to have
a nominal role. Thus, decomposing the wanted will yield [W] Int ~N Age, which
is unsatisfiable. Unwrapping, though, leads to a solution.

Conclusion:
 * Unwrap newtypes before attempting to decompose them.
   This is done in can_eq_nc'.

It all comes from the fact that newtypes aren't necessarily injective
w.r.t. representational equality.

Furthermore, as explained in Note [NthCo and newtypes] in Coercion, we can't use
NthCo on representational coercions over newtypes. NthCo comes into play
only when decomposing givens.

Conclusion:
 * Do not decompose [G] N s ~R N t

Is it sensible to decompose *Wanted* constraints over newtypes?  Yes!
It's the only way we could ever prove (IO Int ~R IO Age), recalling
that IO is a newtype.

However we must be careful.  Consider

  type role Nt representational

  [G] Nt a ~R Nt b       (1)
  [W] NT alpha ~R Nt b   (2)
  [W] alpha ~ a          (3)

If we focus on (3) first, we'll substitute in (2), and now it's
identical to the given (1), so we succeed.  But if we focus on (2)
first, and decompose it, we'll get (alpha ~R b), which is not soluble.
This is exactly like the question of overlapping Givens for class
constraints: see Note [Instance and Given overlap] in TcInteract.

Conclusion:
  * Decompose [W] N s ~R N t  iff there no given constraint that could
    later solve it.
872 873 874
-}

canDecomposableTyConAppOK :: CtEvidence -> EqRel
875
                          -> TyCon -> [TcType] -> [TcType]
876 877
                          -> TcS ()
-- Precondition: tys1 and tys2 are the same length, hence "OK"
878
canDecomposableTyConAppOK ev eq_rel tc tys1 tys2
879 880
  = case ev of
     CtDerived { ctev_loc = loc }
881
        -> unifyDeriveds loc tc_roles tys1 tys2
882 883

     CtWanted { ctev_evar = evar, ctev_loc = loc }
884
        -> do { cos <- zipWith3M (unifyWanted loc) tc_roles tys1 tys2
885
              ; setWantedEvBind evar (EvCoercion (mkTcTyConAppCo role tc cos)) }
886

887 888
     CtGiven { ctev_evar = evar, ctev_loc = loc }
        -> do { let ev_co = mkTcCoVarCo evar
889 890 891 892 893
              ; given_evs <- newGivenEvVars loc $
                             [ ( mkTcEqPredRole r ty1 ty2
                               , EvCoercion (mkTcNthCo i ev_co) )
                             | (r, ty1, ty2, i) <- zip4 tc_roles tys1 tys2 [0..]
                             , r /= Phantom ]
894 895
              ; emitWorkNC given_evs }
  where
896 897 898 899 900
    role     = eqRelRole eq_rel
    tc_roles = tyConRolesX role tc

-- | Call when canonicalizing an equality fails, but if the equality is
-- representational, there is some hope for the future.
901
-- Examples in Note [Use canEqFailure in canDecomposableTyConApp]
902 903
canEqFailure :: CtEvidence -> EqRel
             -> TcType -> TcType -> TcS (StopOrContinue Ct)
904 905
canEqFailure ev NomEq ty1 ty2
  = canEqHardFailure ev NomEq ty1 ty2
906
canEqFailure ev ReprEq ty1 ty2
907
  = do { (xi1, co1) <- flatten FM_FlattenAll ev ty1
908
       ; (xi2, co2) <- flatten FM_FlattenAll ev ty2
909 910 911
            -- We must flatten the types before putting them in the
            -- inert set, so that we are sure to kick them out when
            -- new equalities become available
912 913
       ; traceTcS "canEqFailure with ReprEq" $
         vcat [ ppr ev, ppr ty1, ppr ty2, ppr xi1, ppr xi2 ]
eir@cis.upenn.edu's avatar
eir@cis.upenn.edu committed
914 915 916
       ; rewriteEqEvidence ev ReprEq NotSwapped xi1 xi2 co1 co2
         `andWhenContinue` \ new_ev ->
         continueWith (CIrredEvCan { cc_ev = new_ev }) }
917 918 919 920

-- | Call when canonicalizing an equality fails with utterly no hope.
canEqHardFailure :: CtEvidence -> EqRel
                 -> TcType -> TcType -> TcS (StopOrContinue Ct)
921
-- See Note [Make sure that insolubles are fully rewritten]
922
canEqHardFailure ev eq_rel ty1 ty2
923 924
  = do { (s1, co1) <- flatten FM_SubstOnly ev ty1
       ; (s2, co2) <- flatten FM_SubstOnly ev ty2
925 926 927 928
       ; rewriteEqEvidence ev eq_rel NotSwapped s1 s2 co1 co2
         `andWhenContinue` \ new_ev ->
    do { emitInsoluble (mkNonCanonical new_ev)
       ; stopWith new_ev "Definitely not equal" }}
929

Austin Seipp's avatar
Austin Seipp committed
930
{-
931 932 933 934 935 936 937 938 939 940 941 942 943
Note [Decomposing TyConApps]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~
If we see (T s1 t1 ~ T s2 t2), then we can just decompose to
  (s1 ~ s2, t1 ~ t2)
and push those back into the work list.  But if
  s1 = K k1    s2 = K k2
then we will jus decomopose s1~s2, and it might be better to
do so on the spot.  An important special case is where s1=s2,
and we get just Refl.

So canDecomposableTyCon is a fast-path decomposition that uses
unifyWanted etc to short-cut that work.

944 945 946
Note [Canonicalising type applications]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Given (s1 t1) ~ ty2, how should we proceed?
Austin Seipp's avatar
Austin Seipp committed
947
The simple things is to see if ty2 is of form (s2 t2), and
948
decompose.  By this time s1 and s2 can't be saturated type
Austin Seipp's avatar
Austin Seipp committed
949 950
function applications, because those have been dealt with
by an earlier equation in can_eq_nc, so it is always sound to
951 952
decompose.

Austin Seipp's avatar
Austin Seipp committed
953
However, over-eager decomposition gives bad error messages
954 955 956 957 958 959 960
for things like
   a b ~ Maybe c
   e f ~ p -> q
Suppose (in the first example) we already know a~Array.  Then if we
decompose the application eagerly, yielding
   a ~ Maybe
   b ~ c
Austin Seipp's avatar
Austin Seipp committed
961
we get an error        "Can't match Array ~ Maybe",
962 963
but we'd prefer to get "Can't match Array b ~ Maybe c".

964 965 966
So instead can_eq_wanted_app flattens the LHS and RHS, in the hope of
replacing (a b) by (Array b), before using try_decompose_app to
decompose it.
967

968 969
Note [Make sure that insolubles are fully rewritten]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Simon Peyton Jones's avatar
Simon Peyton Jones committed
970 971
When an equality fails, we still want to rewrite the equality
all the way down, so that it accurately reflects
972 973
 (a) the mutable reference substitution in force at start of solving
 (b) any ty-binds in force at this point in solving
974
See Note [Kick out insolubles] in TcSMonad.
Simon Peyton Jones's avatar
Simon Peyton Jones committed
975
And if we don't do this there is a bad danger that
976 977 978
TcSimplify.applyTyVarDefaulting will find a variable
that has in fact been substituted.

979
Note [Do not decompose Given polytype equalities]
980 981
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Consider [G] (forall a. t1 ~ forall a. t2).  Can we decompose this?
982
No -- what would the evidence look like?  So instead we simply discard
Simon Peyton Jones's avatar
Simon Peyton Jones committed
983
this given evidence.
984 985


986 987 988 989 990 991 992 993 994 995 996 997 998
Note [Combining insoluble constraints]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
As this point we have an insoluble constraint, like Int~Bool.

 * If it is Wanted, delete it from the cache, so that subsequent
   Int~Bool constraints give rise to separate error messages

 * But if it is Derived, DO NOT delete from cache.  A class constraint
   may get kicked out of the inert set, and then have its functional
   dependency Derived constraints generated a second time. In that
   case we don't want to get two (or more) error messages by
   generating two (or more) insoluble fundep constraints from the same
   class constraint.
999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009

Note [No top-level newtypes on RHS of representational equalities]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Suppose we're in this situation:

 work item:  [W] c1 : a ~R b
     inert:  [G] c2 : b ~R Id a

where
  newtype Id a = Id a

1010 1011 1012 1013
We want to make sure canEqTyVar sees [W] a ~R a, after b is flattened
and the Id newtype is unwrapped. This is assured by requiring only flat
types in canEqTyVar *and* having the newtype-unwrapping check above
the tyvar check in can_eq_nc.
1014

Austin Seipp's avatar
Austin Seipp committed
1015
-}
Simon Peyton Jones's avatar
Simon Peyton Jones committed
1016

Austin Seipp's avatar
Austin Seipp committed
1017
canCFunEqCan :: CtEvidence
1018
             -> TyCon -> [TcType]   -- LHS
1019 1020
             -> TcTyVar             -- RHS
             -> TcS (StopOrContinue Ct)
Austin Seipp's avatar
Austin Seipp committed
1021 1022
-- ^ Canonicalise a CFunEqCan.  We know that
--     the arg types are already flat,
1023 1024 1025
-- and the RHS is a fsk, which we must *not* substitute.
-- So just substitute in the LHS
canCFunEqCan ev fn tys fsk
1026
  = do { (tys', cos) <- flattenManyNom ev tys
1027 1028 1029 1030 1031
                        -- cos :: tys' ~ tys
       ; let lhs_co  = mkTcTyConAppCo Nominal fn cos
                        -- :: F tys' ~ F tys
             new_lhs = mkTyConApp fn tys'
             fsk_ty  = mkTyVarTy fsk
1032 1033 1034 1035
       ; rewriteEqEvidence ev NomEq NotSwapped new_lhs fsk_ty
                           lhs_co (mkTcNomReflCo fsk_ty)
         `andWhenContinue` \ ev' ->
    do { extendFlatCache fn tys' (ctEvCoercion ev', fsk_ty, ctEvFlavour ev')
1036
       ; continueWith (CFunEqCan { cc_ev = ev', cc_fun = fn
1037
                                 , cc_tyargs = tys', cc_fsk = fsk }) } }
1038 1039

---------------------
1040
canEqTyVar :: CtEvidence -> EqRel -> SwapFlag
1041 1042
           -> TcTyVar             -- already flat
           -> TcType              -- already flat
1043
           -> TcS (StopOrContinue Ct)
1044
-- A TyVar on LHS, but so far un-zonked
1045 1046 1047
canEqTyVar ev eq_rel swapped tv1 ps_ty2              -- ev :: tv ~ s2
  = do { traceTcS "canEqTyVar" (ppr tv1 $$ ppr ps_ty2 $$ ppr swapped)
         -- FM_Avoid commented out: see Note [Lazy flattening] in TcFlatten
1048 1049 1050 1051
         -- let fmode = FE { fe_ev = ev, fe_mode = FM_Avoid tv1' True }
         -- Flatten the RHS less vigorously, to avoid gratuitous flattening
         -- True <=> xi2 should not itself be a type-function application
       ; dflags <- getDynFlags
1052
       ; canEqTyVar2 dflags ev eq_rel swapped tv1 ps_ty2 }
1053 1054

canEqTyVar2 :: DynFlags
1055
            -> CtEvidence   -- lhs ~ rhs (or, if swapped, orhs ~ olhs)
1056
            -> EqRel
1057
            -> SwapFlag
1058 1059
            -> TcTyVar      -- lhs, flat
            -> TcType       -- rhs, flat
1060
            -> TcS (StopOrContinue Ct)
Austin Seipp's avatar
Austin Seipp committed
1061
-- LHS is an inert type variable,
1062
-- and RHS is fully rewritten, but with type synonyms
1063
-- preserved as much as possible
1064

1065
canEqTyVar2 dflags ev eq_rel swapped tv1 xi2
1066
  | Just tv2 <- getTyVar_maybe xi2
1067
  = canEqTyVarTyVar ev eq_rel swapped tv1 tv2
1068

1069
  | OC_OK xi2' <- occurCheckExpand dflags tv1 xi2  -- No occurs check
1070 1071 1072 1073
     -- We use xi2' on the RHS of the new CTyEqCan, a ~ xi2'
     -- to establish the invariant that a does not appear in the
     -- rhs of the CTyEqCan. This is guaranteed by occurCheckExpand;
     -- see Note [Occurs check expansion] in TcType
1074 1075
  = do { let k1 = tyVarKind tv1
             k2 = typeKind xi2'
1076
       ; rewriteEqEvidence ev eq_rel swapped xi1 xi2' co1 (mkTcReflCo role xi2')
1077 1078 1079
         `andWhenContinue` \ new_ev ->
         if k2 `isSubKind` k1
         then   -- Establish CTyEqCan kind invariant
1080 1081
                -- Reorientation has done its best, but the kinds might
                -- simply be incompatible
1082 1083 1084 1085
               continueWith (CTyEqCan { cc_ev = new_ev
                                      , cc_tyvar  = tv1, cc_rhs = xi2'
                                      , cc_eq_rel = eq_rel })
         else incompatibleKind new_ev xi1 k1 xi2' k2 }
1086 1087

  | otherwise  -- Occurs check error
1088 1089 1090
  = rewriteEqEvidence ev eq_rel swapped xi1 xi2 co1 co2
    `andWhenContinue` \ new_ev ->
    case eq_rel of
1091
      NomEq  -> do { emitInsoluble (mkNonCanonical new_ev)
1092 1093 1094
              -- If we have a ~ [a], it is not canonical, and in particular
              -- we don't want to rewrite existing inerts with it, otherwise
              -- we'd risk divergence in the constraint solver
1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105
                   ; stopWith new_ev "Occurs check" }

        -- A representational equality with an occurs-check problem isn't
        -- insoluble! For example:
        --   a ~R b a
        -- We might learn that b is the newtype Id.
        -- But, the occurs-check certainly prevents the equality from being
        -- canonical, and we might loop if we were to use it in rewriting.
      ReprEq -> do { traceTcS "Occurs-check in representational equality"
                              (ppr xi1 $$ ppr xi2)
                   ; continueWith (CIrredEvCan { cc_ev = new_ev }) }
1106
  where
1107 1108 1109 1110
    role = eqRelRole eq_rel
    xi1  = mkTyVarTy tv1
    co1  = mkTcReflCo role xi1
    co2  = mkTcReflCo role xi2
1111

1112
canEqTyVarTyVar :: CtEvidence           -- tv1 ~ rhs (or rhs ~ tv1, if swapped)
1113
                -> EqRel
1114
                -> SwapFlag
1115
                -> TcTyVar -> TcTyVar   -- tv1, tv2
1116
                -> TcS (StopOrContinue Ct)
1117
-- Both LHS and RHS rewrote to a type variable
1118
-- See Note [Canonical orientation for tyvar/tyvar equality constraints]
1119
canEqTyVarTyVar ev eq_rel swapped tv1 tv2
1120
  | tv1 == tv2
1121
  = do { setEvBindIfWanted ev (EvCoercion $ mkTcReflCo role xi1)
1122 1123
       ; stopWith ev "Equal tyvars" }

1124 1125 1126 1127 1128 1129 1130
  | incompat_kind   = incompatibleKind ev xi1 k1 xi2 k2

-- We don't do this any more
-- See Note [Orientation of equalities with fmvs] in TcSMonad
--  | isFmvTyVar tv1  = do_fmv swapped            tv1 xi1 xi2 co1 co2
--  | isFmvTyVar tv2  = do_fmv (flipSwap swapped) tv2 xi2 xi1 co2 co1

1131 1132 1133
  | same_kind       = if swap_over then do_swap else no_swap
  | k1_sub_k2       = do_swap   -- Note [Kind orientation for CTyEqCan]
  | otherwise       = no_swap   -- k2_sub_k1
1134
  where
1135
    role = eqRelRole eq_rel
1136
    xi1 = mkTyVarTy tv1
1137
    co1 = mkTcReflCo role xi1
1138
    xi2 = mkTyVarTy tv2
1139
    co2 = mkTcReflCo role xi2
1140 1141
    k1  = tyVarKind tv1
    k2  = tyVarKind tv2
1142 1143 1144 1145 1146 1147 1148 1149 1150
    k1_sub_k2     = k1 `isSubKind` k2
    k2_sub_k1     = k2 `isSubKind` k1
    same_kind     = k1_sub_k2 && k2_sub_k1
    incompat_kind = not (k1_sub_k2 || k2_sub_k1)

    no_swap = canon_eq swapped            tv1 xi1 xi2 co1 co2
    do_swap = canon_eq (flipSwap swapped) tv2 xi2 xi1 co2 co1

    canon_eq swapped tv1 xi1 xi2 co1 co2
1151 1152 1153 1154 1155
        -- ev  : tv1 ~ rhs  (not swapped) or   rhs ~ tv1   (swapped)
      = rewriteEqEvidence ev eq_rel swapped xi1 xi2 co1 co2
        `andWhenContinue` \ new_ev ->
        continueWith (CTyEqCan { cc_ev = new_ev, cc_tyvar = tv1
                               , cc_rhs = xi2, cc_eq_rel = eq_rel })
1156

1157 1158 1159
{- We don't do this any more
   See Note [Orientation of equalities with fmvs] in TcSMonad
    -- tv1 is the flatten meta-var
1160 1161 1162
    do_fmv swapped tv1 xi1 xi2 co1 co2
      | same_kind
      = canon_eq swapped tv1 xi1 xi2 co1 co2