TcSimplify.lhs 52.8 KB
Newer Older
 simonm committed Jan 08, 1998 1 \begin{code}  Ian Lynagh committed Nov 04, 2011 2 3 4 5 6 7 8 {-# OPTIONS -fno-warn-tabs #-} -- The above warning supression flag is a temporary kludge. -- While working on this module you are encouraged to remove it and -- detab the module (please do the detabbing in a separate patch). See -- http://hackage.haskell.org/trac/ghc/wiki/Commentary/CodingStyle#TabsvsSpaces -- for details  simonpj@microsoft.com committed Sep 13, 2010 9 module TcSimplify(  Simon Peyton Jones committed Aug 16, 2011 10  simplifyInfer, simplifyAmbiguityCheck,  11  simplifyDefault, simplifyDeriv,  Simon Peyton Jones committed Sep 03, 2012 12 13  simplifyRule, simplifyTop, simplifyInteractive, solveWantedsTcM  simonpj@microsoft.com committed Sep 13, 2010 14  ) where  partain committed Jan 08, 1996 15   simonm committed Jan 08, 1998 16 #include "HsVersions.h"  partain committed Mar 19, 1996 17   Simon Peyton Jones committed Aug 28, 2012 18 import TcRnTypes  simonpj committed Sep 13, 2002 19 import TcRnMonad  simonpj@microsoft.com committed Sep 13, 2010 20 import TcErrors  simonpj@microsoft.com committed Jan 02, 2007 21 import TcMType  simonpj@microsoft.com committed Sep 13, 2010 22 23 import TcType import TcSMonad  dimitris committed Nov 16, 2011 24 import TcInteract  simonpj@microsoft.com committed Sep 13, 2010 25 import Inst  Simon Peyton Jones committed Aug 28, 2012 26 27 import Type ( classifyPredType, PredTree(..), getClassPredTys_maybe ) import Class ( Class )  Simon Marlow committed Oct 11, 2006 28 import Var  dimitris committed Jun 08, 2012 29 import Unique  simonm committed Dec 02, 1998 30 import VarSet  simonpj@microsoft.com committed Nov 12, 2010 31 import VarEnv  Simon Peyton Jones committed Dec 05, 2011 32 import TcEvidence  simonpj@microsoft.com committed Nov 12, 2010 33 import TypeRep  simonpj@microsoft.com committed Sep 13, 2010 34 import Name  simonmar committed Dec 10, 2003 35 import Bag  Simon Marlow committed Oct 11, 2006 36 37 import ListSetOps import Util  simonpj@microsoft.com committed Sep 13, 2010 38 39 40 import PrelInfo import PrelNames import Class ( classKey )  Simon Peyton Jones committed Aug 16, 2011 41 import BasicTypes ( RuleName )  simonpj@microsoft.com committed Sep 13, 2010 42 import Outputable  Ian Lynagh committed Mar 29, 2008 43 import FastString  dimitris committed Mar 28, 2012 44 import TrieMap () -- DV: for now  partain committed Jan 08, 1996 45 46 47 \end{code}  simonpj@microsoft.com committed Sep 13, 2010 48 49 50 51 52 ********************************************************************************* * * * External interface * * * *********************************************************************************  simonpj committed Jan 25, 2001 53   simonpj@microsoft.com committed Sep 13, 2010 54 55 56 \begin{code} simplifyTop :: WantedConstraints -> TcM (Bag EvBind) -- Simplify top-level constraints  simonpj@microsoft.com committed Dec 13, 2010 57 58 59 -- Usually these will be implications, -- but when there is nothing to quantify we don't wrap -- in a degenerate implication, so we do that here instead  dimitris committed Sep 06, 2012 60 61 62 63 64 simplifyTop wanteds = do { traceTc "simplifyTop {" $text "wanted = " <+> ppr wanteds ; ev_binds_var <- newTcEvBinds ; zonked_final_wc <- solveWantedsTcMWithEvBinds ev_binds_var wanteds simpl_top ; binds1 <- TcRnMonad.getTcEvBinds ev_binds_var  Simon Peyton Jones committed Sep 01, 2012 65 66 67  ; traceTc "End simplifyTop }" empty ; traceTc "reportUnsolved {" empty  Simon Peyton Jones committed Sep 17, 2012 68  ; binds2 <- reportUnsolved zonked_final_wc  Simon Peyton Jones committed Sep 01, 2012 69  ; traceTc "reportUnsolved }" empty  dimitris committed Sep 06, 2012 70   Simon Peyton Jones committed Sep 01, 2012 71  ; return (binds1 unionBags binds2) }  Simon Peyton Jones committed Aug 28, 2012 72   Simon Peyton Jones committed Sep 01, 2012 73 74  where -- See Note [Top-level Defaulting Plan]  Simon Peyton Jones committed Sep 18, 2012 75  simpl_top :: WantedConstraints -> TcS WantedConstraints  Simon Peyton Jones committed Sep 01, 2012 76  simpl_top wanteds  dimitris committed Sep 06, 2012 77  = do { wc_first_go <- nestTcS (solve_wanteds_and_drop wanteds)  Simon Peyton Jones committed Sep 01, 2012 78 79  ; applyTyVarDefaulting wc_first_go ; simpl_top_loop wc_first_go }  dimitris committed Jun 08, 2012 80   Simon Peyton Jones committed Sep 01, 2012 81  simpl_top_loop wc  Simon Peyton Jones committed Oct 04, 2012 82 83 84  | isEmptyWC wc || insolubleWC wc -- Don't do type-class defaulting if there are insolubles -- Doing so is not going to solve the insolubles  Simon Peyton Jones committed Sep 01, 2012 85 86  = return wc | otherwise  dimitris committed Sep 06, 2012 87  = do { wc_residual <- nestTcS (solve_wanteds_and_drop wc)  Simon Peyton Jones committed Sep 01, 2012 88 89 90 91 92 93 94  ; let wc_flat_approximate = approximateWC wc_residual ; something_happened <- applyDefaultingRules wc_flat_approximate -- See Note [Top-level Defaulting Plan] ; if something_happened then simpl_top_loop wc_residual else return wc_residual }  dimitris committed Jun 08, 2012 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 \end{code} Note [Top-level Defaulting Plan] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ We have considered two design choices for where/when to apply defaulting. (i) Do it in SimplCheck mode only /whenever/ you try to solve some flat constraints, maybe deep inside the context of implications. This used to be the case in GHC 7.4.1. (ii) Do it in a tight loop at simplifyTop, once all other constraint has finished. This is the current story. Option (i) had many disadvantages: a) First it was deep inside the actual solver, b) Second it was dependent on the context (Infer a type signature, or Check a type signature, or Interactive) since we did not want to always start defaulting when inferring (though there is an exception to this see Note [Default while Inferring]) c) It plainly did not work. Consider typecheck/should_compile/DfltProb2.hs: f :: Int -> Bool f x = const True (\y -> let w :: a -> a w a = const a (y+1) in w y) We will get an implication constraint (for beta the type of y): [untch=beta] forall a. 0 => Num beta which we really cannot default /while solving/ the implication, since beta is untouchable. Instead our new defaulting story is to pull defaulting out of the solver loop and go with option (i), implemented at SimplifyTop. Namely: - First have a go at solving the residual constraint of the whole program - Try to approximate it with a flat constraint - Figure out derived defaulting equations for that flat constraint - Go round the loop again if you did manage to get some equations Now, that has to do with class defaulting. However there exists type variable /kind/ defaulting. Again this is done at the top-level and the plan is: - At the top-level, once you had a go at solving the constraint, do figure out /all/ the touchable unification variables of the wanted contraints. - Apply defaulting to their kinds More details in Note [DefaultTyVar]. \begin{code}  simonpj@microsoft.com committed Sep 13, 2010 139   Simon Peyton Jones committed Aug 16, 2011 140 141 142 ------------------ simplifyAmbiguityCheck :: Name -> WantedConstraints -> TcM (Bag EvBind) simplifyAmbiguityCheck name wanteds  dimitris committed Jun 08, 2012 143  = traceTc "simplifyAmbiguityCheck" (text "name =" <+> ppr name) >>  Simon Peyton Jones committed Sep 03, 2012 144  simplifyTop wanteds -- NB: must be simplifyTop so that we  145 146  -- do ambiguity resolution. -- See Note [Impedence matching] in TcBinds.  dimitris committed Nov 16, 2011 147   simonpj@microsoft.com committed Sep 13, 2010 148 149 150 ------------------ simplifyInteractive :: WantedConstraints -> TcM (Bag EvBind) simplifyInteractive wanteds  dimitris committed Jun 08, 2012 151 152  = traceTc "simplifyInteractive" empty >> simplifyTop wanteds  simonpj@microsoft.com committed Sep 13, 2010 153 154 155 156 157  ------------------ simplifyDefault :: ThetaType -- Wanted; has no type variables in it -> TcM () -- Succeeds iff the constraint is soluble simplifyDefault theta  dimitris committed Jun 08, 2012 158 159  = do { traceTc "simplifyInteractive" empty ; wanted <- newFlatWanteds DefaultOrigin theta  Simon Peyton Jones committed Sep 03, 2012 160 161 162 163  ; (unsolved, _binds) <- solveWantedsTcM (mkFlatWC wanted) ; traceTc "reportUnsolved {" empty -- See Note [Deferring coercion errors to runtime]  Simon Peyton Jones committed Sep 17, 2012 164  ; reportAllUnsolved unsolved  dimitris committed Sep 06, 2012 165 166 167  -- Postcondition of solveWantedsTcM is that returned -- constraints are zonked. So Precondition of reportUnsolved -- is true.  Simon Peyton Jones committed Sep 03, 2012 168 169  ; traceTc "reportUnsolved }" empty  simonpj@microsoft.com committed Sep 13, 2010 170 171  ; return () } \end{code}  simonpj committed Jan 25, 2001 172   simonpj committed May 03, 2001 173   dimitris committed Nov 16, 2011 174 ***********************************************************************************  simonpj@microsoft.com committed Sep 13, 2010 175 * *  dimitris committed Nov 16, 2011 176 * Deriving *  simonpj@microsoft.com committed Sep 13, 2010 177 178 * * ***********************************************************************************  simonpj committed May 03, 2001 179   simonpj@microsoft.com committed Sep 13, 2010 180 181 \begin{code} simplifyDeriv :: CtOrigin  simonpj committed Apr 19, 2011 182 183 184 185  -> PredType -> [TyVar] -> ThetaType -- Wanted -> TcM ThetaType -- Needed  simonpj@microsoft.com committed Sep 13, 2010 186 187 -- Given instance (wanted) => C inst_ty -- Simplify 'wanted' as much as possibles  simonpj@microsoft.com committed Jan 12, 2011 188 -- Fail if not possible  simonpj committed Apr 19, 2011 189 simplifyDeriv orig pred tvs theta  Simon Peyton Jones committed Mar 02, 2012 190  = do { (skol_subst, tvs_skols) <- tcInstSkolTyVars tvs -- Skolemize  simonpj@microsoft.com committed Feb 17, 2011 191 192 193 194  -- The constraint solving machinery -- expects *TcTyVars* not TyVars. -- We use *non-overlappable* (vanilla) skolems -- See Note [Overlap and deriving]  simonpj@microsoft.com committed Sep 13, 2010 195   Simon Peyton Jones committed Mar 02, 2012 196  ; let subst_skol = zipTopTvSubst tvs_skols$ map mkTyVarTy tvs  Simon Peyton Jones committed Jul 23, 2011 197  skol_set = mkVarSet tvs_skols  dimitris committed Jun 08, 2012 198  doc = ptext (sLit "deriving") <+> parens (ppr pred)  simonpj@microsoft.com committed Jan 12, 2011 199 200 201  ; wanted <- newFlatWanteds orig (substTheta skol_subst theta)  dimitris committed Jun 08, 2012 202 203  ; traceTc "simplifyDeriv" $vcat [ pprTvBndrs tvs $$ppr theta$$ ppr wanted, doc ]  Simon Peyton Jones committed Jan 12, 2012 204  ; (residual_wanted, _ev_binds1)  Simon Peyton Jones committed Sep 01, 2012 205  <- solveWantedsTcM (mkFlatWC wanted)  dimitris committed Sep 06, 2012 206  -- Post: residual_wanted are already zonked  simonpj@microsoft.com committed Sep 13, 2010 207   simonpj@microsoft.com committed Jan 12, 2011 208 209  ; let (good, bad) = partitionBagWith get_good (wc_flat residual_wanted) -- See Note [Exotic derived instance contexts]  dimitris committed Nov 16, 2011 210  get_good :: Ct -> Either PredType Ct  dimitris committed Jul 19, 2012 211 212 213 214 215 216  get_good ct | validDerivPred skol_set p , isWantedCt ct = Left p -- NB: residual_wanted may contain unsolved -- Derived and we stick them into the bad set -- so that reportUnsolved may decide what to do with them | otherwise = Right ct  dimitris committed Nov 29, 2011 217  where p = ctPred ct  simonpj@microsoft.com committed Sep 13, 2010 218   Simon Peyton Jones committed Jan 12, 2012 219 220  -- We never want to defer these errors because they are errors in the -- compiler! Hence the False below  Simon Peyton Jones committed Sep 17, 2012 221  ; reportAllUnsolved (residual_wanted { wc_flat = bad })  simonpj@microsoft.com committed Sep 13, 2010 222   simonpj@microsoft.com committed Jan 12, 2011 223 224  ; let min_theta = mkMinimalBySCs (bagToList good) ; return (substTheta subst_skol min_theta) }  simonpj@microsoft.com committed Sep 13, 2010 225 \end{code}  simonpj committed May 03, 2001 226   simonpj@microsoft.com committed Feb 17, 2011 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 Note [Overlap and deriving] ~~~~~~~~~~~~~~~~~~~~~~~~~~~ Consider some overlapping instances: data Show a => Show [a] where .. data Show [Char] where ... Now a data type with deriving: data T a = MkT [a] deriving( Show ) We want to get the derived instance instance Show [a] => Show (T a) where... and NOT instance Show a => Show (T a) where... so that the (Show (T Char)) instance does the Right Thing It's very like the situation when we're inferring the type of a function f x = show [x] and we want to infer f :: Show [a] => a -> String BOTTOM LINE: use vanilla, non-overlappable skolems when inferring the context for the derived instance. Hence tcInstSkolTyVars not tcInstSuperSkolTyVars  simonpj@microsoft.com committed Sep 13, 2010 252 253 254 255 256 257 258 Note [Exotic derived instance contexts] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ In a 'derived' instance declaration, we *infer* the context. It's a bit unclear what rules we should apply for this; the Haskell report is silent. Obviously, constraints like (Eq a) are fine, but what about data T f a = MkT (f a) deriving( Eq ) where we'd get an Eq (f a) constraint. That's probably fine too.  simonpj committed May 03, 2001 259   simonpj@microsoft.com committed Sep 13, 2010 260 261 262 One could go further: consider data T a b c = MkT (Foo a b c) deriving( Eq ) instance (C Int a, Eq b, Eq c) => Eq (Foo a b c)  simonpj committed Feb 10, 2004 263   simonpj@microsoft.com committed Sep 13, 2010 264 265 Notice that this instance (just) satisfies the Paterson termination conditions. Then we *could* derive an instance decl like this:  simonpj committed Feb 10, 2004 266   simonpj@microsoft.com committed Sep 13, 2010 267 268 269 270  instance (C Int a, Eq b, Eq c) => Eq (T a b c) even though there is no instance for (C Int a), because there just *might* be an instance for, say, (C Int Bool) at a site where we need the equality instance for T's.  simonpj committed Feb 10, 2004 271   simonpj@microsoft.com committed Sep 13, 2010 272 273 274 However, this seems pretty exotic, and it's quite tricky to allow this, and yet give sensible error messages in the (much more common) case where we really want that instance decl for C.  simonpj committed Feb 10, 2004 275   simonpj@microsoft.com committed Sep 13, 2010 276 277 So for now we simply require that the derived instance context should have only type-variable constraints.  simonpj committed Feb 10, 2004 278   simonpj@microsoft.com committed Sep 13, 2010 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 Here is another example: data Fix f = In (f (Fix f)) deriving( Eq ) Here, if we are prepared to allow -XUndecidableInstances we could derive the instance instance Eq (f (Fix f)) => Eq (Fix f) but this is so delicate that I don't think it should happen inside 'deriving'. If you want this, write it yourself! NB: if you want to lift this condition, make sure you still meet the termination conditions! If not, the deriving mechanism generates larger and larger constraints. Example: data Succ a = S a data Seq a = Cons a (Seq (Succ a)) | Nil deriving Show Note the lack of a Show instance for Succ. First we'll generate instance (Show (Succ a), Show a) => Show (Seq a) and then instance (Show (Succ (Succ a)), Show (Succ a), Show a) => Show (Seq a) and so on. Instead we want to complain of no instance for (Show (Succ a)). The bottom line ~~~~~~~~~~~~~~~ Allow constraints which consist only of type variables, with no repeats. ********************************************************************************* * * * Inference * * ***********************************************************************************  simonpj committed Feb 10, 2004 308   dreixel committed Nov 11, 2011 309 310 311 312 313 314 315 316 317 318 319 320 Note [Which variables to quantify] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Suppose the inferred type of a function is T kappa (alpha:kappa) -> Int where alpha is a type unification variable and kappa is a kind unification variable Then we want to quantify over *both* alpha and kappa. But notice that kappa appears "at top level" of the type, as well as inside the kind of alpha. So it should be fine to just look for the "top level" kind/type variables of the type, without looking transitively into the kinds of those type variables.  simonpj@microsoft.com committed Sep 13, 2010 321 \begin{code}  Simon Peyton Jones committed Aug 16, 2011 322 simplifyInfer :: Bool  simonpj@microsoft.com committed Jan 12, 2011 323 324 325  -> Bool -- Apply monomorphism restriction -> [(Name, TcTauType)] -- Variables to be generalised, -- and their tau-types  Simon Peyton Jones committed Aug 29, 2012 326  -> WantedConstraints  simonpj@microsoft.com committed Sep 13, 2010 327 328  -> TcM ([TcTyVar], -- Quantify over these type variables [EvVar], -- ... and these constraints  Simon Peyton Jones committed Aug 16, 2011 329 330 331  Bool, -- The monomorphism restriction did something -- so the results type is not as general as -- it could be  simonpj@microsoft.com committed Sep 13, 2010 332  TcEvBinds) -- ... binding these evidence variables  Simon Peyton Jones committed Aug 29, 2012 333 simplifyInfer _top_lvl apply_mr name_taus wanteds  simonpj@microsoft.com committed Jan 12, 2011 334 335 336  | isEmptyWC wanteds = do { gbl_tvs <- tcGetGlobalTyVars -- Already zonked ; zonked_taus <- zonkTcTypes (map snd name_taus)  Simon Peyton Jones committed Feb 16, 2012 337  ; let tvs_to_quantify = varSetElems (tyVarsOfTypes zonked_taus minusVarSet gbl_tvs)  dreixel committed Nov 11, 2011 338 339 340  -- tvs_to_quantify can contain both kind and type vars -- See Note [Which variables to quantify] ; qtvs <- zonkQuantifiedTyVars tvs_to_quantify  Simon Peyton Jones committed Aug 16, 2011 341  ; return (qtvs, [], False, emptyTcEvBinds) }  simonpj committed Feb 10, 2004 342   simonpj@microsoft.com committed Sep 13, 2010 343  | otherwise  Simon Peyton Jones committed Sep 17, 2012 344  = do { zonked_tau_tvs <- zonkTyVarsAndFV (tyVarsOfTypes (map snd name_taus))  simonpj@microsoft.com committed Jan 12, 2011 345   dimitris committed Sep 06, 2012 346  ; ev_binds_var <- newTcEvBinds  simonpj@microsoft.com committed Sep 13, 2010 347  ; traceTc "simplifyInfer {"$ vcat  Simon Peyton Jones committed Aug 16, 2011 348  [ ptext (sLit "names =") <+> ppr (map fst name_taus)  Simon Peyton Jones committed Mar 02, 2012 349 350  , ptext (sLit "taus =") <+> ppr (map snd name_taus) , ptext (sLit "tau_tvs (zonked) =") <+> ppr zonked_tau_tvs  Simon Peyton Jones committed Aug 16, 2011 351 352  , ptext (sLit "closed =") <+> ppr _top_lvl , ptext (sLit "apply_mr =") <+> ppr apply_mr  dimitris committed Sep 06, 2012 353  , ptext (sLit "(unzonked) wanted =") <+> ppr wanteds  simonpj@microsoft.com committed Sep 13, 2010 354 355  ]  Simon Peyton Jones committed Jun 21, 2012 356 357 358 359 360  -- Historical note: Before step 2 we used to have a -- HORRIBLE HACK described in Note [Avoid unecessary -- constraint simplification] but, as described in Trac -- #4361, we have taken in out now. That's why we start -- with step 2!  simonpj@microsoft.com committed Jan 12, 2011 361   Simon Peyton Jones committed Jun 21, 2012 362 363 364 365 366 367 368 369  -- Step 2) First try full-blown solving -- NB: we must gather up all the bindings from doing -- this solving; hence (runTcSWithEvBinds ev_binds_var). -- And note that since there are nested implications, -- calling solveWanteds will side-effect their evidence -- bindings, so we can't just revert to the input -- constraint.  dimitris committed Sep 06, 2012 370 371 372  ; wanted_transformed <- solveWantedsTcMWithEvBinds ev_binds_var wanteds $solve_wanteds_and_drop -- Post: wanted_transformed are zonked  Simon Peyton Jones committed Jun 21, 2012 373 374  -- Step 4) Candidates for quantification are an approximation of wanted_transformed  dimitris committed Jun 08, 2012 375 376 377  -- NB: Already the fixpoint of any unifications that may have happened -- NB: We do not do any defaulting when inferring a type, this can lead -- to less polymorphic types, see Note [Default while Inferring]  dimitris committed Jul 19, 2012 378   Simon Peyton Jones committed Jun 21, 2012 379 380  -- Step 5) Minimize the quantification candidates -- Step 6) Final candidates for quantification  Simon Peyton Jones committed Sep 01, 2012 381  -- We discard bindings, insolubles etc, because all we are  dimitris committed Sep 06, 2012 382 383  -- care aout it  Simon Peyton Jones committed Sep 01, 2012 384  ; (quant_pred_candidates, _extra_binds)  Simon Peyton Jones committed Sep 17, 2012 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399  <- if insolubleWC wanted_transformed then return ([], emptyBag) -- See Note [Quantification with errors] else runTcS$ do { let quant_candidates = approximateWC wanted_transformed ; traceTcS "simplifyWithApprox" $text "quant_candidates = " <+> ppr quant_candidates ; promoteTyVars quant_candidates ; _implics <- solveInteract quant_candidates ; (flats, _insols) <- getInertUnsolved -- NB: Dimitrios is slightly worried that we will get -- family equalities (F Int ~ alpha) in the quantification -- candidates, as we have performed no further unflattening -- at this point. Nothing bad, but inferred contexts might -- look complicated. ; return (map ctPred$ filter isWantedCt (bagToList flats)) }  Simon Peyton Jones committed Sep 01, 2012 400 401 402  -- NB: quant_pred_candidates is already the fixpoint of any -- unifications that may have happened  dimitris committed Jun 08, 2012 403 404  ; gbl_tvs <- tcGetGlobalTyVars -- TODO: can we just use untch instead of gbl_tvs?  Simon Peyton Jones committed Mar 02, 2012 405  ; zonked_tau_tvs <- zonkTyVarsAndFV zonked_tau_tvs  dimitris committed Jun 08, 2012 406   407  ; let init_tvs = zonked_tau_tvs minusVarSet gbl_tvs  Simon Peyton Jones committed Sep 01, 2012 408  poly_qtvs = growThetaTyVars quant_pred_candidates init_tvs  409  minusVarSet gbl_tvs  Simon Peyton Jones committed Sep 01, 2012 410  pbound = filter (quantifyPred poly_qtvs) quant_pred_candidates  dimitris committed Jun 08, 2012 411   Simon Peyton Jones committed Aug 16, 2011 412  -- Monomorphism restriction  Simon Peyton Jones committed Sep 01, 2012 413 414  mr_qtvs = init_tvs minusVarSet constrained_tvs constrained_tvs = tyVarsOfTypes quant_pred_candidates  415  mr_bites = apply_mr && not (null pbound)  Simon Peyton Jones committed Aug 16, 2011 416   Simon Peyton Jones committed Sep 01, 2012 417 418  (qtvs, bound) | mr_bites = (mr_qtvs, []) | otherwise = (poly_qtvs, pbound)  dimitris committed Jun 08, 2012 419   Simon Peyton Jones committed Sep 01, 2012 420 421 422 423 424 425 426  ; traceTc "simplifyWithApprox" $vcat [ ptext (sLit "quant_pred_candidates =") <+> ppr quant_pred_candidates , ptext (sLit "gbl_tvs=") <+> ppr gbl_tvs , ptext (sLit "zonked_tau_tvs=") <+> ppr zonked_tau_tvs , ptext (sLit "pbound =") <+> ppr pbound , ptext (sLit "init_qtvs =") <+> ppr init_tvs , ptext (sLit "poly_qtvs =") <+> ppr poly_qtvs ]  simonpj@microsoft.com committed Jan 12, 2011 427   428  ; if isEmptyVarSet qtvs && null bound  Simon Peyton Jones committed Jul 14, 2012 429 430 431 432  then do { traceTc "} simplifyInfer/no quantification" empty ; emitConstraints wanted_transformed -- Includes insolubles (if -fdefer-type-errors) -- as well as flats and implications  dimitris committed Jun 08, 2012 433  ; return ([], [], mr_bites, TcEvBinds ev_binds_var) }  simonpj@microsoft.com committed Jan 12, 2011 434 435  else do  dimitris committed Jun 08, 2012 436 437 438  { traceTc "simplifyApprox"$ ptext (sLit "bound are =") <+> ppr bound  simonpj@microsoft.com committed Jan 12, 2011 439  -- Step 4, zonk quantified variables  440  ; let minimal_flat_preds = mkMinimalBySCs bound  Simon Peyton Jones committed Aug 16, 2011 441 442  skol_info = InferSkol [ (name, mkSigmaTy [] minimal_flat_preds ty) | (name, ty) <- name_taus ]  simonpj@microsoft.com committed Jan 12, 2011 443 444 445 446  -- Don't add the quantified variables here, because -- they are also bound in ic_skols and we want them to be -- tidied uniformly  Simon Peyton Jones committed Feb 16, 2012 447  ; qtvs_to_return <- zonkQuantifiedTyVars (varSetElems qtvs)  simonpj@microsoft.com committed Jan 12, 2011 448   Simon Peyton Jones committed Jun 21, 2012 449  -- Step 7) Emit an implication  simonpj@microsoft.com committed Jan 12, 2011 450  ; minimal_bound_ev_vars <- mapM TcMType.newEvVar minimal_flat_preds  Simon Peyton Jones committed Sep 17, 2012 451 452  ; lcl_env <- TcRnMonad.getLclEnv ; let implic = Implic { ic_untch = pushUntouchables (tcl_untch lcl_env)  Simon Peyton Jones committed Jan 12, 2012 453  , ic_skols = qtvs_to_return  Simon Peyton Jones committed Aug 28, 2012 454 455  , ic_fsks = [] -- wanted_tansformed arose only from solveWanteds -- hence no flatten-skolems (which come from givens)  simonpj@microsoft.com committed Jan 12, 2011 456  , ic_given = minimal_bound_ev_vars  dimitris committed Jun 08, 2012 457  , ic_wanted = wanted_transformed  simonpj@microsoft.com committed Jan 12, 2011 458 459  , ic_insol = False , ic_binds = ev_binds_var  Simon Peyton Jones committed Sep 17, 2012 460 461  , ic_info = skol_info , ic_env = lcl_env }  simonpj@microsoft.com committed Jan 12, 2011 462  ; emitImplication implic  dimitris committed Jun 08, 2012 463   simonpj@microsoft.com committed Jan 12, 2011 464 465 466  ; traceTc "} simplifyInfer/produced residual implication for quantification" $vcat [ ptext (sLit "implic =") <+> ppr implic -- ic_skols, ic_given give rest of result  Simon Peyton Jones committed Aug 16, 2011 467  , ptext (sLit "qtvs =") <+> ppr qtvs_to_return  Simon Peyton Jones committed Sep 01, 2012 468  , ptext (sLit "spb =") <+> ppr quant_pred_candidates  simonpj@microsoft.com committed Jan 12, 2011 469 470  , ptext (sLit "bound =") <+> ppr bound ]  Simon Peyton Jones committed Aug 16, 2011 471 472  ; return ( qtvs_to_return, minimal_bound_ev_vars , mr_bites, TcEvBinds ev_binds_var) } }  simonpj@microsoft.com committed Jan 12, 2011 473 \end{code}  simonpj@microsoft.com committed Sep 13, 2010 474   Simon Peyton Jones committed Sep 17, 2012 475 476 477 478 479 480 481 482 483 Note [Quantification with errors] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ If we find that the RHS of the definition has some absolutely-insoluble constraints, we abandon all attempts to find a context to quantify over, and instead make the function fully-polymorphic in whatever type we have found. For two reasons a) Minimise downstream errors b) Avoid spurious errors from this function  simonpj@microsoft.com committed Sep 13, 2010 484   485 486 Note [Default while Inferring] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~  dimitris committed Jun 08, 2012 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 Our current plan is that defaulting only happens at simplifyTop and not simplifyInfer. This may lead to some insoluble deferred constraints Example: instance D g => C g Int b constraint inferred = (forall b. 0 => C gamma alpha b) /\ Num alpha type inferred = gamma -> gamma Now, if we try to default (alpha := Int) we will be able to refine the implication to (forall b. 0 => C gamma Int b) which can then be simplified further to (forall b. 0 => D gamma) Finally we /can/ approximate this implication with (D gamma) and infer the quantified type: forall g. D g => g -> g Instead what will currently happen is that we will get a quantified type (forall g. g -> g) and an implication: forall g. 0 => (forall b. 0 => C g alpha b) /\ Num alpha which, even if the simplifyTop defaults (alpha := Int) we will still be left with an unsolvable implication: forall g. 0 => (forall b. 0 => D g) The concrete example would be: h :: C g a s => g -> a -> ST s a f (x::gamma) = (\_ -> x) (runST (h x (undefined::alpha)) + 1) But it is quite tedious to do defaulting and resolve the implication constraints and we have not observed code breaking because of the lack of defaulting in inference so we don't do it for now.  simonpj@microsoft.com committed Jan 12, 2011 521 522 523 524 525 526 527 528 529 Note [Minimize by Superclasses] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ When we quantify over a constraint, in simplifyInfer we need to quantify over a constraint that is minimal in some sense: For instance, if the final wanted constraint is (Eq alpha, Ord alpha), we'd like to quantify over Ord alpha, because we can just get Eq alpha from superclass selection from Ord alpha. This minimization is what mkMinimalBySCs does. Then, simplifyInfer uses the minimal constraint to check the original wanted.  simonpj@microsoft.com committed Sep 13, 2010 530   simonpj committed Jan 25, 2001 531   simonpj@microsoft.com committed Oct 08, 2010 532 533 Note [Avoid unecessary constraint simplification] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~  Simon Peyton Jones committed Jun 21, 2012 534 535 536 537  -------- NB NB NB (Jun 12) ------------- This note not longer applies; see the notes with Trac #4361. But I'm leaving it in here so we remember the issue.) ----------------------------------------  simonpj@microsoft.com committed Oct 08, 2010 538 When inferring the type of a let-binding, with simplifyInfer,  Simon Peyton Jones committed Jun 21, 2012 539 try to avoid unnecessarily simplifying class constraints.  simonpj@microsoft.com committed Oct 08, 2010 540 541 Doing so aids sharing, but it also helps with delicate situations like  dimitris committed Jun 08, 2012 542   simonpj@microsoft.com committed Oct 08, 2010 543  instance C t => C [t] where ..  dimitris committed Jun 08, 2012 544   simonpj@microsoft.com committed Oct 08, 2010 545 546 547 548 549 550 551 552 553 554 555  f :: C [t] => .... f x = let g y = ...(constraint C [t])... in ... When inferring a type for 'g', we don't want to apply the instance decl, because then we can't satisfy (C t). So we just notice that g isn't quantified over 't' and partition the contraints before simplifying. This only half-works, but then let-generalisation only half-works.  simonpj@microsoft.com committed Sep 13, 2010 556 557 558 559 560 ********************************************************************************* * * * RULES * * * ***********************************************************************************  simonpj committed May 12, 2004 561   Simon Peyton Jones committed Apr 16, 2012 562 See note [Simplifying RULE consraints] in TcRule  simonpj committed Jun 28, 1999 563   Simon Peyton Jones committed Apr 16, 2012 564 565 566 567 568 569 570 571 572 573 574 575 576 577 Note [RULE quanfification over equalities] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Decideing which equalities to quantify over is tricky: * We do not want to quantify over insoluble equalities (Int ~ Bool) (a) because we prefer to report a LHS type error (b) because if such things end up in 'givens' we get a bogus "inaccessible code" error * But we do want to quantify over things like (a ~ F b), where F is a type function. The difficulty is that it's hard to tell what is insoluble! So we see whether the simplificaiotn step yielded any type errors, and if so refrain from quantifying over *any* equalites.  simonpj committed Sep 13, 2002 578 579  \begin{code}  simonpj@microsoft.com committed Sep 13, 2010 580 581 582 simplifyRule :: RuleName -> WantedConstraints -- Constraints from LHS -> WantedConstraints -- Constraints from RHS  Simon Peyton Jones committed Apr 16, 2012 583 584 585  -> TcM ([EvVar], WantedConstraints) -- LHS evidence varaibles -- See Note [Simplifying RULE constraints] in TcRule simplifyRule name lhs_wanted rhs_wanted  Simon Peyton Jones committed Sep 03, 2012 586  = do { -- We allow ourselves to unify environment  dimitris committed Jun 08, 2012 587  -- variables: runTcS runs with NoUntouchables  Simon Peyton Jones committed Sep 03, 2012 588  (resid_wanted, _) <- solveWantedsTcM (lhs_wanted andWC rhs_wanted)  dimitris committed Sep 06, 2012 589  -- Post: these are zonked and unflattened  Simon Peyton Jones committed Apr 16, 2012 590   Simon Peyton Jones committed Sep 17, 2012 591 592  ; zonked_lhs_flats <- zonkCts (wc_flat lhs_wanted) ; let (q_cts, non_q_cts) = partitionBag quantify_me zonked_lhs_flats  Simon Peyton Jones committed Apr 16, 2012 593 594 595 596 597 598 599 600 601 602 603 604  quantify_me -- Note [RULE quantification over equalities] | insolubleWC resid_wanted = quantify_insol | otherwise = quantify_normal quantify_insol ct = not (isEqPred (ctPred ct)) quantify_normal ct | EqPred t1 t2 <- classifyPredType (ctPred ct) = not (t1 eqType t2) | otherwise = True  simonpj@microsoft.com committed Jan 12, 2011 605  ; traceTc "simplifyRule"$  Simon Peyton Jones committed Sep 03, 2012 606  vcat [ ptext (sLit "LHS of rule") <+> doubleQuotes (ftext name)  Simon Peyton Jones committed Sep 17, 2012 607  , text "zonked_lhs_flats" <+> ppr zonked_lhs_flats  Simon Peyton Jones committed Apr 16, 2012 608 609  , text "q_cts" <+> ppr q_cts ]  Simon Peyton Jones committed May 07, 2012 610  ; return ( map (ctEvId . ctEvidence) (bagToList q_cts)  Simon Peyton Jones committed Sep 17, 2012 611  , lhs_wanted { wc_flat = non_q_cts }) }  simonpj committed Sep 13, 2002 612 613 614 \end{code}  simonpj@microsoft.com committed Sep 13, 2010 615 616 617 618 619 ********************************************************************************* * * * Main Simplifier * * * ***********************************************************************************  lewie committed Mar 02, 2000 620   Simon Peyton Jones committed Jan 12, 2012 621 622 623 624 625 626 Note [Deferring coercion errors to runtime] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ While developing, sometimes it is desirable to allow compilation to succeed even if there are type errors in the code. Consider the following case: module Main where  simonpj@microsoft.com committed Sep 13, 2010 627   Simon Peyton Jones committed Jan 12, 2012 628 629  a :: Int a = 'a'  simonpj@microsoft.com committed Sep 13, 2010 630   Simon Peyton Jones committed Jan 12, 2012 631  main = print "b"  simonpj@microsoft.com committed Sep 13, 2010 632   Simon Peyton Jones committed Jan 12, 2012 633 634 Even though a is ill-typed, it is not used in the end, so if all that we're interested in is main it is handy to be able to ignore the problems in a.  simonpj@microsoft.com committed Sep 13, 2010 635   Simon Peyton Jones committed Jan 12, 2012 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 Since we treat type equalities as evidence, this is relatively simple. Whenever we run into a type mismatch in TcUnify, we normally just emit an error. But it is always safe to defer the mismatch to the main constraint solver. If we do that, a will get transformed into co :: Int ~ Char co = ... a :: Int a = 'a' cast co The constraint solver would realize that co is an insoluble constraint, and emit an error with reportUnsolved. But we can also replace the right-hand side of co with error "Deferred type error: Int ~ Char". This allows the program to compile, and it will run fine unless we evaluate a. This is what deferErrorsToRuntime does. It does this by keeping track of which errors correspond to which coercion in TcErrors (with ErrEnv). TcErrors.reportTidyWanteds does not print the errors and does not fail if -fwarn-type-errors is on, so that we can continue compilation. The errors are turned into warnings in reportUnsolved.  Simon Peyton Jones committed Sep 17, 2012 658 659 660 661 662 663 664 665 666 667 Note [Zonk after solving] ~~~~~~~~~~~~~~~~~~~~~~~~~ We zonk the result immediately after constraint solving, for two reasons: a) because zonkWC generates evidence, and this is the moment when we have a suitable evidence variable to hand. Note that *after* solving the constraints are typically small, so the overhead is not great.  Simon Peyton Jones committed Jan 12, 2012 668 \begin{code}  dimitris committed Sep 06, 2012 669 670 671 672 solveWantedsTcMWithEvBinds :: EvBindsVar -> WantedConstraints -> (WantedConstraints -> TcS WantedConstraints) -> TcM WantedConstraints  Simon Peyton Jones committed Sep 17, 2012 673 674 675 676 677 678 -- Returns a *zonked* result -- We zonk when we finish primarily to un-flatten out any -- flatten-skolems etc introduced by canonicalisation of -- types involving type funuctions. Happily the result -- is typically much smaller than the input, indeed it is -- often empty.  dimitris committed Sep 06, 2012 679 solveWantedsTcMWithEvBinds ev_binds_var wc tcs_action  Simon Peyton Jones committed Sep 17, 2012 680 681  = do { traceTc "solveWantedsTcMWithEvBinds" $text "wanted=" <+> ppr wc ; wc2 <- runTcSWithEvBinds ev_binds_var (tcs_action wc)  dimitris committed Sep 06, 2012 682  ; zonkWC ev_binds_var wc2 }  Simon Peyton Jones committed Sep 17, 2012 683  -- See Note [Zonk after solving]  dimitris committed Sep 06, 2012 684   Simon Peyton Jones committed Sep 01, 2012 685 solveWantedsTcM :: WantedConstraints -> TcM (WantedConstraints, Bag EvBind)  Simon Peyton Jones committed Sep 03, 2012 686 -- Zonk the input constraints, and simplify them  dimitris committed Jul 19, 2012 687 -- Return the evidence binds in the BagEvBinds result  Simon Peyton Jones committed Aug 28, 2012 688 -- Discards all Derived stuff in result  dimitris committed Sep 06, 2012 689 -- Postcondition: fully zonked and unflattened constraints  Simon Peyton Jones committed Sep 03, 2012 690 solveWantedsTcM wanted  dimitris committed Sep 06, 2012 691 692 693  = do { ev_binds_var <- newTcEvBinds ; wanteds' <- solveWantedsTcMWithEvBinds ev_binds_var wanted solve_wanteds_and_drop ; binds <- TcRnMonad.getTcEvBinds ev_binds_var  Simon Peyton Jones committed Sep 03, 2012 694  ; return (wanteds', binds) }  dimitris committed Jul 19, 2012 695   Simon Peyton Jones committed Sep 01, 2012 696 697 698 699 700 solve_wanteds_and_drop :: WantedConstraints -> TcS (WantedConstraints) -- Since solve_wanteds returns the residual WantedConstraints, -- it should alway be called within a runTcS or something similar, solve_wanteds_and_drop wanted = do { wc <- solve_wanteds wanted ; return (dropDerivedWC wc) }  dimitris committed Jul 19, 2012 701 702  solve_wanteds :: WantedConstraints -> TcS WantedConstraints  Simon Peyton Jones committed Sep 01, 2012 703 -- so that the inert set doesn't mindlessly propagate.  dimitris committed Jul 19, 2012 704 -- NB: wc_flats may be wanted /or/ derived now  dimitris committed Nov 16, 2011 705 solve_wanteds wanted@(WC { wc_flat = flats, wc_impl = implics, wc_insol = insols })  simonpj@microsoft.com committed Jan 12, 2011 706 707  = do { traceTcS "solveWanteds {" (ppr wanted)  dimitris committed Jul 19, 2012 708  -- Try the flat bit, including insolubles. Solving insolubles a  Simon Peyton Jones committed Sep 28, 2012 709  -- second time round is a bit of a waste; but the code is simple  Simon Peyton Jones committed Jul 23, 2012 710 711 712  -- and the program is wrong anyway, and we don't run the danger -- of adding Derived insolubles twice; see -- TcSMonad Note [Do not add duplicate derived insolubles]  Simon Peyton Jones committed Aug 28, 2012 713  ; traceTcS "solveFlats {" empty  Simon Peyton Jones committed Jul 23, 2012 714  ; let all_flats = flats unionBags insols  Simon Peyton Jones committed Aug 28, 2012 715 716  ; impls_from_flats <- solveInteract all_flats ; traceTcS "solveFlats end }" (ppr impls_from_flats)  simonpj@microsoft.com committed Jan 12, 2011 717   dimitris committed Nov 16, 2011 718 719  -- solve_wanteds iterates when it is able to float equalities -- out of one or more of the implications.  dimitris committed Apr 10, 2012 720  ; unsolved_implics <- simpl_loop 1 (implics unionBags impls_from_flats)  simonpj@microsoft.com committed Jan 12, 2011 721   Simon Peyton Jones committed Aug 28, 2012 722 723  ; (unsolved_flats, insoluble_flats) <- getInertUnsolved  dimitris committed Sep 06, 2012 724 725 726 727 728 729  -- We used to unflatten here but now we only do it once at top-level -- during zonking -- see Note [Unflattening while zonking] in TcMType ; let wc = WC { wc_flat = unsolved_flats , wc_impl = unsolved_implics , wc_insol = insoluble_flats }  dimitris committed Nov 16, 2011 730  ; bb <- getTcEvBindsMap  simonpj@microsoft.com committed Nov 12, 2010 731  ; tb <- getTcSTyBindsMap  simonpj@microsoft.com committed Sep 13, 2010 732  ; traceTcS "solveWanteds }"$  simonpj@microsoft.com committed Nov 12, 2010 733  vcat [ text "unsolved_flats =" <+> ppr unsolved_flats  simonpj@microsoft.com committed Jan 12, 2011 734  , text "unsolved_implics =" <+> ppr unsolved_implics  dimitris committed Nov 16, 2011 735  , text "current evbinds =" <+> ppr (evBindMapBinds bb)  simonpj@microsoft.com committed Nov 12, 2010 736  , text "current tybinds =" <+> vcat (map ppr (varEnvElts tb))  Simon Peyton Jones committed Aug 28, 2012 737  , text "final wc =" <+> ppr wc ]  Simon Peyton Jones committed Mar 02, 2012 738   Simon Peyton Jones committed Aug 28, 2012 739  ; return wc }  dimitris committed Nov 16, 2011 740 741 742 743 744 745 746 747  simpl_loop :: Int -> Bag Implication -> TcS (Bag Implication) simpl_loop n implics | n > 10 = traceTcS "solveWanteds: loop!" empty >> return implics | otherwise  Simon Peyton Jones committed Aug 28, 2012 748 749 750 751 752 753 754  = do { (floated_eqs, unsolved_implics) <- solveNestedImplications implics ; if isEmptyBag floated_eqs then return unsolved_implics else do { -- Put floated_eqs into the current inert set before looping impls_from_eqs <- solveInteract floated_eqs ; simpl_loop (n+1) (unsolved_implics unionBags impls_from_eqs)} }  755   dimitris committed Jul 19, 2012 756   dimitris committed Nov 16, 2011 757 758 759 760 761 762 763 764 765 solveNestedImplications :: Bag Implication -> TcS (Cts, Bag Implication) -- Precondition: the TcS inerts may contain unsolved flats which have -- to be converted to givens before we go inside a nested implication. solveNestedImplications implics | isEmptyBag implics = return (emptyBag, emptyBag) | otherwise = do { inerts <- getTcSInerts  Simon Peyton Jones committed Aug 28, 2012 766 767  ; let thinner_inerts = prepareInertsForImplications inerts -- See Note [Preparing inert set for implications]  dimitris committed Jul 19, 2012 768   Simon Peyton Jones committed Aug 28, 2012 769  ; traceTcS "solveNestedImplications starting {" $ dimitris committed Jul 19, 2012 770  vcat [ text "original inerts = " <+> ppr inerts  dimitris committed Jun 08, 2012 771 772  , text "thinner_inerts = " <+> ppr thinner_inerts ]  Simon Peyton Jones committed Aug 28, 2012 773  ; (floated_eqs, unsolved_implics)  Simon Peyton Jones committed Aug 28, 2012 774  <- flatMapBagPairM (solveImplication thinner_inerts) implics  dimitris committed Nov 16, 2011 775 776 777 778  -- ... and we are back in the original TcS inerts -- Notice that the original includes the _insoluble_flats so it was safe to ignore -- them in the beginning of this function.  Simon Peyton Jones committed Aug 28, 2012 779  ; traceTcS "solveNestedImplications end }"$  Simon Peyton Jones committed Aug 28, 2012 780  vcat [ text "all floated_eqs =" <+> ppr floated_eqs  781 782  , text "unsolved_implics =" <+> ppr unsolved_implics ]  Simon Peyton Jones committed Aug 28, 2012 783  ; return (floated_eqs, unsolved_implics) }  784   Simon Peyton Jones committed Aug 28, 2012 785 solveImplication :: InertSet  dimitris committed Nov 16, 2011 786 787 788 789 790  -> Implication -- Wanted -> TcS (Cts, -- All wanted or derived floated equalities: var = type Bag Implication) -- Unsolved rest (always empty or singleton) -- Precondition: The TcS monad contains an empty worklist and given-only inerts -- which after trying to solve this implication we must restore to their original value  Simon Peyton Jones committed Aug 28, 2012 791 solveImplication inerts  dimitris committed Nov 16, 2011 792  imp@(Implic { ic_untch = untch  simonpj@microsoft.com committed Oct 06, 2010 793 794  , ic_binds = ev_binds , ic_skols = skols  Simon Peyton Jones committed Aug 28, 2012 795  , ic_fsks = old_fsks  simonpj@microsoft.com committed Oct 06, 2010 796  , ic_given = givens  simonpj@microsoft.com committed Sep 13, 2010 797  , ic_wanted = wanteds  Simon Peyton Jones committed Sep 17, 2012 798 799  , ic_info = info , ic_env = env })  Simon Peyton Jones committed Sep 01, 2012 800  =  simonpj@microsoft.com committed Sep 13, 2010 801 802  do { traceTcS "solveImplication {" (ppr imp)  Simon Peyton Jones committed Aug 28, 2012 803  -- Solve the nested constraints  Simon Peyton Jones committed Sep 01, 2012 804 805 806  -- NB: 'inerts' has empty inert_fsks ; (new_fsks, residual_wanted) <- nestImplicTcS ev_binds untch inerts $ Simon Peyton Jones committed Sep 17, 2012 807  do { solveInteractGiven (mkGivenLoc info env) old_fsks givens  Simon Peyton Jones committed Sep 01, 2012 808  ; residual_wanted <- solve_wanteds wanteds  Simon Peyton Jones committed Sep 17, 2012 809 810 811  -- solve_wanteds, *not* solve_wanteds_and_drop, because -- we want to retain derived equalities so we can float -- them out in floatEqualities  Simon Peyton Jones committed Sep 01, 2012 812 813 814 815 816 817 818  ; more_fsks <- getFlattenSkols ; return (more_fsks ++ old_fsks, residual_wanted) } ; (floated_eqs, final_wanted) <- floatEqualities (skols ++ new_fsks) givens residual_wanted ; let res_implic | isEmptyWC final_wanted  Simon Peyton Jones committed Aug 28, 2012 819 820  = emptyBag | otherwise  Simon Peyton Jones committed Sep 01, 2012 821 822 823  = unitBag (imp { ic_fsks = new_fsks , ic_wanted = dropDerivedWC final_wanted , ic_insol = insolubleWC final_wanted })  simonpj@microsoft.com committed Sep 13, 2010 824   dimitris committed Nov 16, 2011 825  ; evbinds <- getTcEvBindsMap  simonpj@microsoft.com committed Sep 13, 2010 826  ; traceTcS "solveImplication end }"$ vcat  Simon Peyton Jones committed Aug 28, 2012 827  [ text "floated_eqs =" <+> ppr floated_eqs  Simon Peyton Jones committed Sep 01, 2012 828  , text "new_fsks =" <+> ppr new_fsks  Simon Peyton Jones committed Aug 28, 2012 829 830  , text "res_implic =" <+> ppr res_implic , text "implication evbinds = " <+> ppr (evBindMapBinds evbinds) ]  simonpj@microsoft.com committed Sep 13, 2010 831   Simon Peyton Jones committed Aug 28, 2012 832  ; return (floated_eqs, res_implic) }  dimitris committed Jul 19, 2012 833 834 835 836 \end{code} \begin{code}  Simon Peyton Jones committed Sep 01, 2012 837 838 floatEqualities :: [TcTyVar] -> [EvVar] -> WantedConstraints -> TcS (Cts, WantedConstraints)  simonpj@microsoft.com committed Jan 12, 2011 839 840 -- Post: The returned FlavoredEvVar's are only Wanted or Derived -- and come from the input wanted ev vars or deriveds  Simon Peyton Jones committed Sep 01, 2012 841 842 -- Also performs some unifications, adding to monadically-carried ty_binds -- These will be used when processing floated_eqs later  Simon Peyton Jones committed Aug 28, 2012 843 844 floatEqualities skols can_given wanteds@(WC { wc_flat = flats }) | hasEqualities can_given  Simon Peyton Jones committed Sep 01, 2012 845  = return (emptyBag, wanteds) -- Note [Float Equalities out of Implications]  Simon Peyton Jones committed Aug 28, 2012 846  | otherwise  Simon Peyton Jones committed Sep 01, 2012 847 848  = do { let (float_eqs, remaining_flats) = partitionBag is_floatable flats ; promoteTyVars float_eqs  Simon Peyton Jones committed Sep 01, 2012 849  ; ty_binds <- getTcSTyBindsMap  Simon Peyton Jones committed Sep 01, 2012 850  ; traceTcS "floatEqualities" (vcat [ text "Floated eqs =" <+> ppr float_eqs  Simon Peyton Jones committed Sep 01, 2012 851 852  , text "Ty binds =" <+> ppr ty_binds]) ; return (float_eqs, wanteds { wc_flat = remaining_flats }) }  Simon Peyton Jones committed Aug 28, 2012 853 854 855 856 857 858 859 860 861  where skol_set = growSkols wanteds (mkVarSet skols) is_floatable :: Ct -> Bool is_floatable ct = isEqPred pred && skol_set disjointVarSet tyVarsOfType pred where pred = ctPred ct  Simon Peyton Jones committed Sep 01, 2012 862 promoteTyVars :: Cts -> TcS ()  Simon Peyton Jones committed Sep 17, 2012 863 864 865 -- When we float a constraint out of an implication we -- must restore (MetaTvInv) in Note [Untouchable type variables] -- in TcType  Simon Peyton Jones committed Sep 01, 2012 866 867 868 869 promoteTyVars cts = do { untch <- TcSMonad.getUntouchables ; mapM_ (promote_tv untch) (varSetElems (tyVarsOfCts cts)) } where  Simon Peyton Jones committed Sep 01, 2012 870 871 872 873 874 875 876 877  promote_tv untch tv | isFloatedTouchableMetaTyVar untch tv = do { cloned_tv <- TcSMonad.cloneMetaTyVar tv ; let rhs_tv = setMetaTyVarUntouchables cloned_tv untch ; setWantedTyBind tv (mkTyVarTy rhs_tv) } | otherwise = return ()  Simon Peyton Jones committed Aug 28, 2012 878 879 880 881 882 883 884 885 886 growSkols :: WantedConstraints -> VarSet -> VarSet -- Find all the type variables that might possibly be unified -- with a type that mentions a skolem. This test is very conservative. -- I don't *think* we need look inside the implications, because any -- relevant unification variables in there are untouchable. growSkols (WC { wc_flat = flats }) skols = growThetaTyVars theta skols where theta = foldrBag ((:) . ctPred) [] flats  Simon Peyton Jones committed Sep 01, 2012 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911  approximateWC :: WantedConstraints -> Cts -- Postcondition: Wanted or Derived Cts approximateWC wc = float_wc emptyVarSet wc where float_wc :: TcTyVarSet -> WantedConstraints -> Cts float_wc skols (WC { wc_flat = flat, wc_impl = implic }) = floats1 unionBags floats2 where floats1 = do_bag (float_flat skols) flat floats2 = do_bag (float_implic skols) implic float_implic :: TcTyVarSet -> Implication -> Cts float_implic skols imp = float_wc skols' (ic_wanted imp) where skols' = skols extendVarSetList ic_skols imp extendVarSetList ic_fsks imp float_flat :: TcTyVarSet -> Ct -> Cts float_flat skols ct | tyVarsOfCt ct disjointVarSet skols = singleCt ct | otherwise = emptyCts do_bag :: (a -> Bag c) -> Bag a -> Bag c do_bag f = foldrBag (unionBags.f) emptyBag \end{code}  partain committed Jan 08, 1996 912   simonpj@microsoft.com committed Feb 14, 2011 913 914 Note [Float Equalities out of Implications] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~  dimitris committed Jun 08, 2012 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 For ordinary pattern matches (including existentials) we float equalities out of implications, for instance: data T where MkT :: Eq a => a -> T f x y = case x of MkT _ -> (y::Int) We get the implication constraint (x::T) (y::alpha): forall a. [untouchable=alpha] Eq a => alpha ~ Int We want to float out the equality into a scope where alpha is no longer untouchable, to solve the implication! But we cannot float equalities out of implications whose givens may yield or contain equalities: data T a where T1 :: T Int T2 :: T Bool T3 :: T a h :: T a -> a -> Int f x y = case x of T1 -> y::Int T2 -> y::Bool T3 -> h x y We generate constraint, for (x::T alpha) and (y :: beta): [untouchables = beta] (alpha ~ Int => beta ~ Int) -- From 1st branch [untouchables = beta] (alpha ~ Bool => beta ~ Bool) -- From 2nd branch (alpha ~ beta) -- From 3rd branch If we float the equality (beta ~ Int) outside of the first implication and the equality (beta ~ Bool) out of the second we get an insoluble constraint. But if we just leave them inside the implications we unify alpha := beta and solve everything. Principle: We do not want to float equalities out which may need the given *evidence* to become soluble. Consequence: classes with functional dependencies don't matter (since there is no evidence for a fundep equality), but equality superclasses do matter (since they carry evidence).  Simon Peyton Jones committed Sep 17, 2012 958 959 960 961 962 963 964 965 966 Note [Promoting unification variables] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ When we float an equality out of an implication we must "promote" free unification variables of the equality, in order to maintain Invariant (MetaTvInv) from Note [Untouchable type variables] in TcType. for the leftover implication. This is absolutely necessary. Consider the following example. We start with two implications and a class with a functional dependency.  dimitris committed Jun 08, 2012 967   Simon Peyton Jones committed Aug 28, 2012 968 969 970 971 972  class C x y | x -> y instance C [a] [a] (I1) [untch=beta]forall b. 0 => F Int ~ [beta] (I2) [untch=beta]forall c. 0 => F Int ~ [[alpha]] /\ C beta [c]  dimitris committed Jun 08, 2012 973 974 975 976 977 978 979  We float (F Int ~ [beta]) out of I1, and we float (F Int ~ [[alpha]]) out of I2. They may react to yield that (beta := [alpha]) which can then be pushed inwards the leftover of I2 to get (C [alpha] [a]) which, using the FunDep, will mean that (alpha := a). In the end we will have the skolem 'b' escaping in the untouchable beta! Concrete example is in indexed_types/should_fail/ExtraTcsUntch.hs:  Simon Peyton Jones committed Aug 28, 2012 980 981 982 983 984 985 986 987 988 989 990 991 992  class C x y | x -> y where op :: x -> y -> () instance C [a] [a] type family F a :: * h :: F Int -> () h = undefined data TEx where TEx :: a -> TEx  dimitris committed Jun 08, 2012 993   Simon Peyton Jones committed Aug 28, 2012 994 995 996 997 998 999  f (x::beta) = let g1 :: forall b. b -> () g1 _ = h [x] g2 z = case z of TEx y -> (h [[undefined]], op x [y]) in (g1 '3', g2 undefined)  simonpj@microsoft.com committed Sep 13, 2010 1000   simonpj@microsoft.com committed Nov 12, 2010 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019  Note [Solving Family Equations] ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ After we are done with simplification we may be left with constraints of the form: [Wanted] F xis ~ beta If 'beta' is a touchable unification variable not already bound in the TyBinds then we'd like to create a binding for it, effectively "defaulting" it to be 'F xis'. When is it ok to do so? 1) 'beta' must not already be defaulted to something. Example: [Wanted] F Int ~ beta <~ Will default [beta := F Int] [Wanted] F Char ~ beta <~ Already defaulted, can't default again. We have to report this as unsolved. 2) However, we must still do an occurs check when defaulting (F xis ~ beta), to set [beta := F xis] only if beta is not among the free variables of xis. 3) Notice that 'beta' can't be bound in ty binds already because we rewrite RHS  dimitris committed Sep 06, 2012 1020 1021  of type family equations. See Inert Set invariants in TcInteract.  Simon Peyton Jones committed Sep 17, 2012 1022 This solving is now happening during zonking, see Note [Unflattening while zonking]  dimitris committed Sep 06, 2012 1023 in TcMType.  simonpj@microsoft.com committed Nov 12, 2010 1024 1025   simonpj@microsoft.com committed Sep 13, 2010 1026 1027 1028 1029 1030 ********************************************************************************* * * * Defaulting and disamgiguation * * * *********************************************************************************  simonm committed Jan 08, 1998 1031 \begin{code}  Simon Peyton Jones committed Sep 01, 2012 1032 1033 1034 applyDefaultingRules :: Cts -> TcS Bool -- True <=> I did some defaulting, reflected in ty_binds  dimitris committed Jul 19, 2012 1035 1036 -- Return some extra derived equalities, which express the -- type-class default choice.