{-
Author: George Karachalias
Pattern Matching Coverage Checking.
-}
{-# LANGUAGE CPP, GADTs, DataKinds, KindSignatures #-}
{-# LANGUAGE TupleSections #-}
{-# LANGUAGE ViewPatterns #-}
module Check (
-- Checking and printing
checkSingle, checkMatches, checkGuardMatches, isAnyPmCheckEnabled,
-- See Note [Type and Term Equality Propagation]
genCaseTmCs1, genCaseTmCs2
) where
#include "HsVersions.h"
import GhcPrelude
import TmOracle
import Unify( tcMatchTy )
import DynFlags
import HsSyn
import TcHsSyn
import Id
import ConLike
import Name
import FamInstEnv
import TysPrim (tYPETyCon)
import TysWiredIn
import TyCon
import SrcLoc
import Util
import Outputable
import FastString
import DataCon
import PatSyn
import HscTypes (CompleteMatch(..))
import BasicTypes (Boxity(..))
import DsMonad
import TcSimplify (tcCheckSatisfiability)
import TcType (isStringTy)
import Bag
import ErrUtils
import Var (EvVar)
import TyCoRep
import Type
import UniqSupply
import DsUtils (isTrueLHsExpr)
import qualified GHC.LanguageExtensions as LangExt
import Data.List (find)
import Data.Maybe (catMaybes, isJust, fromMaybe)
import Control.Monad (forM, when, forM_, zipWithM)
import Coercion
import TcEvidence
import TcSimplify (tcNormalise)
import IOEnv
import qualified Data.Semigroup as Semi
import ListT (ListT(..), fold, select)
{-
This module checks pattern matches for:
\begin{enumerate}
\item Equations that are redundant
\item Equations with inaccessible right-hand-side
\item Exhaustiveness
\end{enumerate}
The algorithm is based on the paper:
"GADTs Meet Their Match:
Pattern-matching Warnings That Account for GADTs, Guards, and Laziness"
http://people.cs.kuleuven.be/~george.karachalias/papers/p424-karachalias.pdf
%************************************************************************
%* *
Pattern Match Check Types
%* *
%************************************************************************
-}
-- We use the non-determinism monad to apply the algorithm to several
-- possible sets of constructors. Users can specify complete sets of
-- constructors by using COMPLETE pragmas.
-- The algorithm only picks out constructor
-- sets deep in the bowels which makes a simpler `mapM` more difficult to
-- implement. The non-determinism is only used in one place, see the ConVar
-- case in `pmCheckHd`.
type PmM a = ListT DsM a
liftD :: DsM a -> PmM a
liftD m = ListT $ \sk fk -> m >>= \a -> sk a fk
-- Pick the first match complete covered match or otherwise the "best" match.
-- The best match is the one with the least uncovered clauses, ties broken
-- by the number of inaccessible clauses followed by number of redundant
-- clauses.
--
-- This is specified in the
-- "Disambiguating between multiple ``COMPLETE`` pragmas" section of the
-- users' guide. If you update the implementation of this function, make sure
-- to update that section of the users' guide as well.
getResult :: PmM PmResult -> DsM PmResult
getResult ls
= do { res <- fold ls goM (pure Nothing)
; case res of
Nothing -> panic "getResult is empty"
Just a -> return a }
where
goM :: PmResult -> DsM (Maybe PmResult) -> DsM (Maybe PmResult)
goM mpm dpm = do { pmr <- dpm
; return $ Just $ go pmr mpm }
-- Careful not to force unecessary results
go :: Maybe PmResult -> PmResult -> PmResult
go Nothing rs = rs
go (Just old@(PmResult prov rs (UncoveredPatterns us) is)) new
| null us && null rs && null is = old
| otherwise =
let PmResult prov' rs' (UncoveredPatterns us') is' = new
in case compareLength us us'
`mappend` (compareLength is is')
`mappend` (compareLength rs rs')
`mappend` (compare prov prov') of
GT -> new
EQ -> new
LT -> old
go (Just (PmResult _ _ (TypeOfUncovered _) _)) _new
= panic "getResult: No inhabitation candidates"
data PatTy = PAT | VA -- Used only as a kind, to index PmPat
-- The *arity* of a PatVec [p1,..,pn] is
-- the number of p1..pn that are not Guards
data PmPat :: PatTy -> * where
PmCon :: { pm_con_con :: ConLike
, pm_con_arg_tys :: [Type]
, pm_con_tvs :: [TyVar]
, pm_con_dicts :: [EvVar]
, pm_con_args :: [PmPat t] } -> PmPat t
-- For PmCon arguments' meaning see @ConPatOut@ in hsSyn/HsPat.hs
PmVar :: { pm_var_id :: Id } -> PmPat t
PmLit :: { pm_lit_lit :: PmLit } -> PmPat t -- See Note [Literals in PmPat]
PmNLit :: { pm_lit_id :: Id
, pm_lit_not :: [PmLit] } -> PmPat 'VA
PmGrd :: { pm_grd_pv :: PatVec
, pm_grd_expr :: PmExpr } -> PmPat 'PAT
instance Outputable (PmPat a) where
ppr = pprPmPatDebug
-- data T a where
-- MkT :: forall p q. (Eq p, Ord q) => p -> q -> T [p]
-- or MkT :: forall p q r. (Eq p, Ord q, [p] ~ r) => p -> q -> T r
type Pattern = PmPat 'PAT -- ^ Patterns
type ValAbs = PmPat 'VA -- ^ Value Abstractions
type PatVec = [Pattern] -- ^ Pattern Vectors
data ValVec = ValVec [ValAbs] Delta -- ^ Value Vector Abstractions
-- | Term and type constraints to accompany each value vector abstraction.
-- For efficiency, we store the term oracle state instead of the term
-- constraints. TODO: Do the same for the type constraints?
data Delta = MkDelta { delta_ty_cs :: Bag EvVar
, delta_tm_cs :: TmState }
type ValSetAbs = [ValVec] -- ^ Value Set Abstractions
type Uncovered = ValSetAbs
-- Instead of keeping the whole sets in memory, we keep a boolean for both the
-- covered and the divergent set (we store the uncovered set though, since we
-- want to print it). For both the covered and the divergent we have:
--
-- True <=> The set is non-empty
--
-- hence:
-- C = True ==> Useful clause (no warning)
-- C = False, D = True ==> Clause with inaccessible RHS
-- C = False, D = False ==> Redundant clause
data Covered = Covered | NotCovered
deriving Show
instance Outputable Covered where
ppr (Covered) = text "Covered"
ppr (NotCovered) = text "NotCovered"
-- Like the or monoid for booleans
-- Covered = True, Uncovered = False
instance Semi.Semigroup Covered where
Covered <> _ = Covered
_ <> Covered = Covered
NotCovered <> NotCovered = NotCovered
instance Monoid Covered where
mempty = NotCovered
mappend = (Semi.<>)
data Diverged = Diverged | NotDiverged
deriving Show
instance Outputable Diverged where
ppr Diverged = text "Diverged"
ppr NotDiverged = text "NotDiverged"
instance Semi.Semigroup Diverged where
Diverged <> _ = Diverged
_ <> Diverged = Diverged
NotDiverged <> NotDiverged = NotDiverged
instance Monoid Diverged where
mempty = NotDiverged
mappend = (Semi.<>)
-- | When we learned that a given match group is complete
data Provenance =
FromBuiltin -- ^ From the original definition of the type
-- constructor.
| FromComplete -- ^ From a user-provided @COMPLETE@ pragma
deriving (Show, Eq, Ord)
instance Outputable Provenance where
ppr = text . show
instance Semi.Semigroup Provenance where
FromComplete <> _ = FromComplete
_ <> FromComplete = FromComplete
_ <> _ = FromBuiltin
instance Monoid Provenance where
mempty = FromBuiltin
mappend = (Semi.<>)
data PartialResult = PartialResult {
presultProvenance :: Provenance
-- keep track of provenance because we don't want
-- to warn about redundant matches if the result
-- is contaminated with a COMPLETE pragma
, presultCovered :: Covered
, presultUncovered :: Uncovered
, presultDivergent :: Diverged }
instance Outputable PartialResult where
ppr (PartialResult prov c vsa d)
= text "PartialResult" <+> ppr prov <+> ppr c
<+> ppr d <+> ppr vsa
instance Semi.Semigroup PartialResult where
(PartialResult prov1 cs1 vsa1 ds1)
<> (PartialResult prov2 cs2 vsa2 ds2)
= PartialResult (prov1 Semi.<> prov2)
(cs1 Semi.<> cs2)
(vsa1 Semi.<> vsa2)
(ds1 Semi.<> ds2)
instance Monoid PartialResult where
mempty = PartialResult mempty mempty [] mempty
mappend = (Semi.<>)
-- newtype ChoiceOf a = ChoiceOf [a]
-- | Pattern check result
--
-- * Redundant clauses
-- * Not-covered clauses (or their type, if no pattern is available)
-- * Clauses with inaccessible RHS
--
-- More details about the classification of clauses into useful, redundant
-- and with inaccessible right hand side can be found here:
--
-- https://ghc.haskell.org/trac/ghc/wiki/PatternMatchCheck
--
data PmResult =
PmResult {
pmresultProvenance :: Provenance
, pmresultRedundant :: [Located [LPat GhcTc]]
, pmresultUncovered :: UncoveredCandidates
, pmresultInaccessible :: [Located [LPat GhcTc]] }
-- | Either a list of patterns that are not covered, or their type, in case we
-- have no patterns at hand. Not having patterns at hand can arise when
-- handling EmptyCase expressions, in two cases:
--
-- * The type of the scrutinee is a trivially inhabited type (like Int or Char)
-- * The type of the scrutinee cannot be reduced to WHNF.
--
-- In both these cases we have no inhabitation candidates for the type at hand,
-- but we don't want to issue just a wildcard as missing. Instead, we print a
-- type annotated wildcard, so that the user knows what kind of patterns is
-- expected (e.g. (_ :: Int), or (_ :: F Int), where F Int does not reduce).
data UncoveredCandidates = UncoveredPatterns Uncovered
| TypeOfUncovered Type
-- | The empty pattern check result
emptyPmResult :: PmResult
emptyPmResult = PmResult FromBuiltin [] (UncoveredPatterns []) []
-- | Non-exhaustive empty case with unknown/trivial inhabitants
uncoveredWithTy :: Type -> PmResult
uncoveredWithTy ty = PmResult FromBuiltin [] (TypeOfUncovered ty) []
{-
%************************************************************************
%* *
Entry points to the checker: checkSingle and checkMatches
%* *
%************************************************************************
-}
-- | Check a single pattern binding (let)
checkSingle :: DynFlags -> DsMatchContext -> Id -> Pat GhcTc -> DsM ()
checkSingle dflags ctxt@(DsMatchContext _ locn) var p = do
tracePmD "checkSingle" (vcat [ppr ctxt, ppr var, ppr p])
mb_pm_res <- tryM (getResult (checkSingle' locn var p))
case mb_pm_res of
Left _ -> warnPmIters dflags ctxt
Right res -> dsPmWarn dflags ctxt res
-- | Check a single pattern binding (let)
checkSingle' :: SrcSpan -> Id -> Pat GhcTc -> PmM PmResult
checkSingle' locn var p = do
liftD resetPmIterDs -- set the iter-no to zero
fam_insts <- liftD dsGetFamInstEnvs
clause <- liftD $ translatePat fam_insts p
missing <- mkInitialUncovered [var]
tracePm "checkSingle: missing" (vcat (map pprValVecDebug missing))
-- no guards
PartialResult prov cs us ds <- runMany (pmcheckI clause []) missing
let us' = UncoveredPatterns us
return $ case (cs,ds) of
(Covered, _ ) -> PmResult prov [] us' [] -- useful
(NotCovered, NotDiverged) -> PmResult prov m us' [] -- redundant
(NotCovered, Diverged ) -> PmResult prov [] us' m -- inaccessible rhs
where m = [cL locn [cL locn p]]
-- | Exhaustive for guard matches, is used for guards in pattern bindings and
-- in @MultiIf@ expressions.
checkGuardMatches :: HsMatchContext Name -- Match context
-> GRHSs GhcTc (LHsExpr GhcTc) -- Guarded RHSs
-> DsM ()
checkGuardMatches hs_ctx guards@(GRHSs _ grhss _) = do
dflags <- getDynFlags
let combinedLoc = foldl1 combineSrcSpans (map getLoc grhss)
dsMatchContext = DsMatchContext hs_ctx combinedLoc
match = cL combinedLoc $
Match { m_ext = noExt
, m_ctxt = hs_ctx
, m_pats = []
, m_grhss = guards }
checkMatches dflags dsMatchContext [] [match]
checkGuardMatches _ (XGRHSs _) = panic "checkGuardMatches"
-- | Check a matchgroup (case, functions, etc.)
checkMatches :: DynFlags -> DsMatchContext
-> [Id] -> [LMatch GhcTc (LHsExpr GhcTc)] -> DsM ()
checkMatches dflags ctxt vars matches = do
tracePmD "checkMatches" (hang (vcat [ppr ctxt
, ppr vars
, text "Matches:"])
2
(vcat (map ppr matches)))
mb_pm_res <- tryM $ getResult $ case matches of
-- Check EmptyCase separately
-- See Note [Checking EmptyCase Expressions]
[] | [var] <- vars -> checkEmptyCase' var
_normal_match -> checkMatches' vars matches
case mb_pm_res of
Left _ -> warnPmIters dflags ctxt
Right res -> dsPmWarn dflags ctxt res
-- | Check a matchgroup (case, functions, etc.). To be called on a non-empty
-- list of matches. For empty case expressions, use checkEmptyCase' instead.
checkMatches' :: [Id] -> [LMatch GhcTc (LHsExpr GhcTc)] -> PmM PmResult
checkMatches' vars matches
| null matches = panic "checkMatches': EmptyCase"
| otherwise = do
liftD resetPmIterDs -- set the iter-no to zero
missing <- mkInitialUncovered vars
tracePm "checkMatches': missing" (vcat (map pprValVecDebug missing))
(prov, rs,us,ds) <- go matches missing
return $ PmResult {
pmresultProvenance = prov
, pmresultRedundant = map hsLMatchToLPats rs
, pmresultUncovered = UncoveredPatterns us
, pmresultInaccessible = map hsLMatchToLPats ds }
where
go :: [LMatch GhcTc (LHsExpr GhcTc)] -> Uncovered
-> PmM (Provenance
, [LMatch GhcTc (LHsExpr GhcTc)]
, Uncovered
, [LMatch GhcTc (LHsExpr GhcTc)])
go [] missing = return (mempty, [], missing, [])
go (m:ms) missing = do
tracePm "checMatches': go" (ppr m $$ ppr missing)
fam_insts <- liftD dsGetFamInstEnvs
(clause, guards) <- liftD $ translateMatch fam_insts m
r@(PartialResult prov cs missing' ds)
<- runMany (pmcheckI clause guards) missing
tracePm "checMatches': go: res" (ppr r)
(ms_prov, rs, final_u, is) <- go ms missing'
let final_prov = prov `mappend` ms_prov
return $ case (cs, ds) of
-- useful
(Covered, _ ) -> (final_prov, rs, final_u, is)
-- redundant
(NotCovered, NotDiverged) -> (final_prov, m:rs, final_u,is)
-- inaccessible
(NotCovered, Diverged ) -> (final_prov, rs, final_u, m:is)
hsLMatchToLPats :: LMatch id body -> Located [LPat id]
hsLMatchToLPats (dL->L l (Match { m_pats = pats })) = cL l pats
hsLMatchToLPats _ = panic "checMatches'"
-- | Check an empty case expression. Since there are no clauses to process, we
-- only compute the uncovered set. See Note [Checking EmptyCase Expressions]
-- for details.
checkEmptyCase' :: Id -> PmM PmResult
checkEmptyCase' var = do
tm_ty_css <- pmInitialTmTyCs
mb_candidates <- inhabitationCandidates (delta_ty_cs tm_ty_css) (idType var)
case mb_candidates of
-- Inhabitation checking failed / the type is trivially inhabited
Left ty -> return (uncoveredWithTy ty)
-- A list of inhabitant candidates is available: Check for each
-- one for the satisfiability of the constraints it gives rise to.
Right (_, candidates) -> do
missing_m <- flip mapMaybeM candidates $
\InhabitationCandidate{ ic_val_abs = va, ic_tm_ct = tm_ct
, ic_ty_cs = ty_cs
, ic_strict_arg_tys = strict_arg_tys } -> do
mb_sat <- pmIsSatisfiable tm_ty_css tm_ct ty_cs strict_arg_tys
pure $ fmap (ValVec [va]) mb_sat
return $ if null missing_m
then emptyPmResult
else PmResult FromBuiltin [] (UncoveredPatterns missing_m) []
-- | Returns 'True' if the argument 'Type' is a fully saturated application of
-- a closed type constructor.
--
-- Closed type constructors are those with a fixed right hand side, as
-- opposed to e.g. associated types. These are of particular interest for
-- pattern-match coverage checking, because GHC can exhaustively consider all
-- possible forms that values of a closed type can take on.
--
-- Note that this function is intended to be used to check types of value-level
-- patterns, so as a consequence, the 'Type' supplied as an argument to this
-- function should be of kind @Type@.
pmIsClosedType :: Type -> Bool
pmIsClosedType ty
= case splitTyConApp_maybe ty of
Just (tc, ty_args)
| is_algebraic_like tc && not (isFamilyTyCon tc)
-> ASSERT2( ty_args `lengthIs` tyConArity tc, ppr ty ) True
_other -> False
where
-- This returns True for TyCons which /act like/ algebraic types.
-- (See "Type#type_classification" for what an algebraic type is.)
--
-- This is qualified with \"like\" because of a particular special
-- case: TYPE (the underlyind kind behind Type, among others). TYPE
-- is conceptually a datatype (and thus algebraic), but in practice it is
-- a primitive builtin type, so we must check for it specially.
--
-- NB: it makes sense to think of TYPE as a closed type in a value-level,
-- pattern-matching context. However, at the kind level, TYPE is certainly
-- not closed! Since this function is specifically tailored towards pattern
-- matching, however, it's OK to label TYPE as closed.
is_algebraic_like :: TyCon -> Bool
is_algebraic_like tc = isAlgTyCon tc || tc == tYPETyCon
pmTopNormaliseType_maybe :: FamInstEnvs -> Bag EvVar -> Type
-> PmM (Maybe (Type, [DataCon], Type))
-- ^ Get rid of *outermost* (or toplevel)
-- * type function redex
-- * data family redex
-- * newtypes
--
-- Behaves exactly like `topNormaliseType_maybe`, but instead of returning a
-- coercion, it returns useful information for issuing pattern matching
-- warnings. See Note [Type normalisation for EmptyCase] for details.
--
-- NB: Normalisation can potentially change kinds, if the head of the type
-- is a type family with a variable result kind. I (Richard E) can't think
-- of a way to cause trouble here, though.
pmTopNormaliseType_maybe env ty_cs typ
= do (_, mb_typ') <- liftD $ initTcDsForSolver $ tcNormalise ty_cs typ
-- Before proceeding, we chuck typ into the constraint solver, in case
-- solving for given equalities may reduce typ some. See
-- "Wrinkle: local equalities" in
-- Note [Type normalisation for EmptyCase].
pure $ do typ' <- mb_typ'
((ty_f,tm_f), ty) <- topNormaliseTypeX stepper comb typ'
-- We need to do topNormaliseTypeX in addition to tcNormalise,
-- since topNormaliseX looks through newtypes, which
-- tcNormalise does not do.
Just (eq_src_ty ty (typ' : ty_f [ty]), tm_f [], ty)
where
-- Find the first type in the sequence of rewrites that is a data type,
-- newtype, or a data family application (not the representation tycon!).
-- This is the one that is equal (in source Haskell) to the initial type.
-- If none is found in the list, then all of them are type family
-- applications, so we simply return the last one, which is the *simplest*.
eq_src_ty :: Type -> [Type] -> Type
eq_src_ty ty tys = maybe ty id (find is_closed_or_data_family tys)
is_closed_or_data_family :: Type -> Bool
is_closed_or_data_family ty = pmIsClosedType ty || isDataFamilyAppType ty
-- For efficiency, represent both lists as difference lists.
-- comb performs the concatenation, for both lists.
comb (tyf1, tmf1) (tyf2, tmf2) = (tyf1 . tyf2, tmf1 . tmf2)
stepper = newTypeStepper `composeSteppers` tyFamStepper
-- A 'NormaliseStepper' that unwraps newtypes, careful not to fall into
-- a loop. If it would fall into a loop, it produces 'NS_Abort'.
newTypeStepper :: NormaliseStepper ([Type] -> [Type],[DataCon] -> [DataCon])
newTypeStepper rec_nts tc tys
| Just (ty', _co) <- instNewTyCon_maybe tc tys
= case checkRecTc rec_nts tc of
Just rec_nts' -> let tyf = ((TyConApp tc tys):)
tmf = ((tyConSingleDataCon tc):)
in NS_Step rec_nts' ty' (tyf, tmf)
Nothing -> NS_Abort
| otherwise
= NS_Done
tyFamStepper :: NormaliseStepper ([Type] -> [Type], [DataCon] -> [DataCon])
tyFamStepper rec_nts tc tys -- Try to step a type/data family
= let (_args_co, ntys, _res_co) = normaliseTcArgs env Representational tc tys in
-- NB: It's OK to use normaliseTcArgs here instead of
-- normalise_tc_args (which takes the LiftingContext described
-- in Note [Normalising types]) because the reduceTyFamApp below
-- works only at top level. We'll never recur in this function
-- after reducing the kind of a bound tyvar.
case reduceTyFamApp_maybe env Representational tc ntys of
Just (_co, rhs) -> NS_Step rec_nts rhs ((rhs:), id)
_ -> NS_Done
-- | Determine suitable constraints to use at the beginning of pattern-match
-- coverage checking by consulting the sets of term and type constraints
-- currently in scope. If one of these sets of constraints is unsatisfiable,
-- use an empty set in its place. (See
-- @Note [Recovering from unsatisfiable pattern-matching constraints]@
-- for why this is done.)
pmInitialTmTyCs :: PmM Delta
pmInitialTmTyCs = do
ty_cs <- liftD getDictsDs
tm_cs <- map toComplex . bagToList <$> liftD getTmCsDs
sat_ty <- tyOracle ty_cs
let initTyCs = if sat_ty then ty_cs else emptyBag
initTmState = fromMaybe initialTmState (tmOracle initialTmState tm_cs)
pure $ MkDelta{ delta_tm_cs = initTmState, delta_ty_cs = initTyCs }
{-
Note [Recovering from unsatisfiable pattern-matching constraints]
~~~~~~~~~~~~~~~~
Consider the following code (see #12957 and #15450):
f :: Int ~ Bool => ()
f = case True of { False -> () }
We want to warn that the pattern-matching in `f` is non-exhaustive. But GHC
used not to do this; in fact, it would warn that the match was /redundant/!
This is because the constraint (Int ~ Bool) in `f` is unsatisfiable, and the
coverage checker deems any matches with unsatifiable constraint sets to be
unreachable.
We decide to better than this. When beginning coverage checking, we first
check if the constraints in scope are unsatisfiable, and if so, we start
afresh with an empty set of constraints. This way, we'll get the warnings
that we expect.
-}
-- | Given a conlike's term constraints, type constraints, and strict argument
-- types, check if they are satisfiable.
-- (In other words, this is the ⊢_Sat oracle judgment from the GADTs Meet
-- Their Match paper.)
--
-- For the purposes of efficiency, this takes as separate arguments the
-- ambient term and type constraints (which are known beforehand to be
-- satisfiable), as well as the new term and type constraints (which may not
-- be satisfiable). This lets us implement two mini-optimizations:
--
-- * If there are no new type constraints, then don't bother initializing
-- the type oracle, since it's redundant to do so.
-- * Since the new term constraint is a separate argument, we only need to
-- execute one iteration of the term oracle (instead of traversing the
-- entire set of term constraints).
--
-- Taking strict argument types into account is something which was not
-- discussed in GADTs Meet Their Match. For an explanation of what role they
-- serve, see @Note [Extensions to GADTs Meet Their Match]@.
pmIsSatisfiable
:: Delta -- ^ The ambient term and type constraints
-- (known to be satisfiable).
-> ComplexEq -- ^ The new term constraint.
-> Bag EvVar -- ^ The new type constraints.
-> [Type] -- ^ The strict argument types.
-> PmM (Maybe Delta)
-- ^ @'Just' delta@ if the constraints (@delta@) are
-- satisfiable, and each strict argument type is inhabitable.
-- 'Nothing' otherwise.
pmIsSatisfiable amb_cs new_tm_c new_ty_cs strict_arg_tys = do
mb_sat <- tmTyCsAreSatisfiable amb_cs new_tm_c new_ty_cs
case mb_sat of
Nothing -> pure Nothing
Just delta -> do
-- We know that the term and type constraints are inhabitable, so now
-- check if each strict argument type is inhabitable.
all_non_void <- checkAllNonVoid initRecTc delta strict_arg_tys
pure $ if all_non_void -- Check if each strict argument type
-- is inhabitable
then Just delta
else Nothing
-- | Like 'pmIsSatisfiable', but only checks if term and type constraints are
-- satisfiable, and doesn't bother checking anything related to strict argument
-- types.
tmTyCsAreSatisfiable
:: Delta -- ^ The ambient term and type constraints
-- (known to be satisfiable).
-> ComplexEq -- ^ The new term constraint.
-> Bag EvVar -- ^ The new type constraints.
-> PmM (Maybe Delta)
-- ^ @'Just' delta@ if the constraints (@delta@) are
-- satisfiable. 'Nothing' otherwise.
tmTyCsAreSatisfiable
(MkDelta{ delta_tm_cs = amb_tm_cs, delta_ty_cs = amb_ty_cs })
new_tm_c new_ty_cs = do
let ty_cs = new_ty_cs `unionBags` amb_ty_cs
sat_ty <- if isEmptyBag new_ty_cs
then pure True
else tyOracle ty_cs
pure $ case (sat_ty, solveOneEq amb_tm_cs new_tm_c) of
(True, Just term_cs) -> Just $ MkDelta{ delta_ty_cs = ty_cs
, delta_tm_cs = term_cs }
_unsat -> Nothing
-- | Implements two performance optimizations, as described in the
-- \"Strict argument type constraints\" section of
-- @Note [Extensions to GADTs Meet Their Match]@.
checkAllNonVoid :: RecTcChecker -> Delta -> [Type] -> PmM Bool
checkAllNonVoid rec_ts amb_cs strict_arg_tys = do
fam_insts <- liftD dsGetFamInstEnvs
let definitely_inhabited =
definitelyInhabitedType fam_insts (delta_ty_cs amb_cs)
tys_to_check <- filterOutM definitely_inhabited strict_arg_tys
let rec_max_bound | tys_to_check `lengthExceeds` 1
= 1
| otherwise
= defaultRecTcMaxBound
rec_ts' = setRecTcMaxBound rec_max_bound rec_ts
allM (nonVoid rec_ts' amb_cs) tys_to_check
-- | Checks if a strict argument type of a conlike is inhabitable by a
-- terminating value (i.e, an 'InhabitationCandidate').
-- See @Note [Extensions to GADTs Meet Their Match]@.
nonVoid
:: RecTcChecker -- ^ The per-'TyCon' recursion depth limit.
-> Delta -- ^ The ambient term/type constraints (known to be
-- satisfiable).
-> Type -- ^ The strict argument type.
-> PmM Bool -- ^ 'True' if the strict argument type might be inhabited by
-- a terminating value (i.e., an 'InhabitationCandidate').
-- 'False' if it is definitely uninhabitable by anything
-- (except bottom).
nonVoid rec_ts amb_cs strict_arg_ty = do
mb_cands <- inhabitationCandidates (delta_ty_cs amb_cs) strict_arg_ty
case mb_cands of
Right (tc, cands)
| Just rec_ts' <- checkRecTc rec_ts tc
-> anyM (cand_is_inhabitable rec_ts' amb_cs) cands
-- A strict argument type is inhabitable by a terminating value if
-- at least one InhabitationCandidate is inhabitable.
_ -> pure True
-- Either the type is trivially inhabited or we have exceeded the
-- recursion depth for some TyCon (so bail out and conservatively
-- claim the type is inhabited).
where
-- Checks if an InhabitationCandidate for a strict argument type:
--
-- (1) Has satisfiable term and type constraints.
-- (2) Has 'nonVoid' strict argument types (we bail out of this
-- check if recursion is detected).
--
-- See Note [Extensions to GADTs Meet Their Match]
cand_is_inhabitable :: RecTcChecker -> Delta
-> InhabitationCandidate -> PmM Bool
cand_is_inhabitable rec_ts amb_cs
(InhabitationCandidate{ ic_tm_ct = new_term_c
, ic_ty_cs = new_ty_cs
, ic_strict_arg_tys = new_strict_arg_tys }) = do
mb_sat <- tmTyCsAreSatisfiable amb_cs new_term_c new_ty_cs
case mb_sat of
Nothing -> pure False
Just new_delta -> do
checkAllNonVoid rec_ts new_delta new_strict_arg_tys
-- | @'definitelyInhabitedType' ty@ returns 'True' if @ty@ has at least one
-- constructor @C@ such that:
--
-- 1. @C@ has no equality constraints.
-- 2. @C@ has no strict argument types.
--
-- See the \"Strict argument type constraints\" section of
-- @Note [Extensions to GADTs Meet Their Match]@.
definitelyInhabitedType :: FamInstEnvs -> Bag EvVar -> Type -> PmM Bool
definitelyInhabitedType env ty_cs ty = do
mb_res <- pmTopNormaliseType_maybe env ty_cs ty
pure $ case mb_res of
Just (_, cons, _) -> any meets_criteria cons
Nothing -> False
where
meets_criteria :: DataCon -> Bool
meets_criteria con =
null (dataConEqSpec con) && -- (1)
null (dataConImplBangs con) -- (2)
{- Note [Type normalisation for EmptyCase]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
EmptyCase is an exception for pattern matching, since it is strict. This means
that it boils down to checking whether the type of the scrutinee is inhabited.
Function pmTopNormaliseType_maybe gets rid of the outermost type function/data
family redex and newtypes, in search of an algebraic type constructor, which is
easier to check for inhabitation.
It returns 3 results instead of one, because there are 2 subtle points:
1. Newtypes are isomorphic to the underlying type in core but not in the source
language,
2. The representational data family tycon is used internally but should not be
shown to the user
Hence, if pmTopNormaliseType_maybe env ty_cs ty = Just (src_ty, dcs, core_ty),
then
(a) src_ty is the rewritten type which we can show to the user. That is, the
type we get if we rewrite type families but not data families or
newtypes.
(b) dcs is the list of data constructors "skipped", every time we normalise a
newtype to its core representation, we keep track of the source data
constructor.
(c) core_ty is the rewritten type. That is,
pmTopNormaliseType_maybe env ty_cs ty = Just (src_ty, dcs, core_ty)
implies
topNormaliseType_maybe env ty = Just (co, core_ty)
for some coercion co.
To see how all cases come into play, consider the following example:
data family T a :: *
data instance T Int = T1 | T2 Bool
-- Which gives rise to FC:
-- data T a
-- data R:TInt = T1 | T2 Bool
-- axiom ax_ti : T Int ~R R:TInt
newtype G1 = MkG1 (T Int)
newtype G2 = MkG2 G1
type instance F Int = F Char
type instance F Char = G2
In this case pmTopNormaliseType_maybe env ty_cs (F Int) results in
Just (G2, [MkG2,MkG1], R:TInt)
Which means that in source Haskell:
- G2 is equivalent to F Int (in contrast, G1 isn't).
- if (x : R:TInt) then (MkG2 (MkG1 x) : F Int).
-----
-- Wrinkle: Local equalities
-----
Given the following type family:
type family F a
type instance F Int = Void
Should the following program (from #14813) be considered exhaustive?
f :: (i ~ Int) => F i -> a
f x = case x of {}
You might think "of course, since `x` is obviously of type Void". But the
idType of `x` is technically F i, not Void, so if we pass F i to
inhabitationCandidates, we'll mistakenly conclude that `f` is non-exhaustive.
In order to avoid this pitfall, we need to normalise the type passed to
pmTopNormaliseType_maybe, using the constraint solver to solve for any local
equalities (such as i ~ Int) that may be in scope.
-}
-- | Generate all 'InhabitationCandidate's for a given type. The result is
-- either @'Left' ty@, if the type cannot be reduced to a closed algebraic type
-- (or if it's one trivially inhabited, like 'Int'), or @'Right' candidates@,
-- if it can. In this case, the candidates are the signature of the tycon, each
-- one accompanied by the term- and type- constraints it gives rise to.
-- See also Note [Checking EmptyCase Expressions]
inhabitationCandidates :: Bag EvVar -> Type
-> PmM (Either Type (TyCon, [InhabitationCandidate]))
inhabitationCandidates ty_cs ty = do
fam_insts <- liftD dsGetFamInstEnvs
mb_norm_res <- pmTopNormaliseType_maybe fam_insts ty_cs ty
case mb_norm_res of
Just (src_ty, dcs, core_ty) -> alts_to_check src_ty core_ty dcs
Nothing -> alts_to_check ty ty []
where
-- All these types are trivially inhabited
trivially_inhabited = [ charTyCon, doubleTyCon, floatTyCon
, intTyCon, wordTyCon, word8TyCon ]
-- Note: At the moment we leave all the typing and constraint fields of
-- PmCon empty, since we know that they are not gonna be used. Is the
-- right-thing-to-do to actually create them, even if they are never used?
build_tm :: ValAbs -> [DataCon] -> ValAbs
build_tm = foldr (\dc e -> PmCon (RealDataCon dc) [] [] [] [e])
-- Inhabitation candidates, using the result of pmTopNormaliseType_maybe
alts_to_check :: Type -> Type -> [DataCon]
-> PmM (Either Type (TyCon, [InhabitationCandidate]))
alts_to_check src_ty core_ty dcs = case splitTyConApp_maybe core_ty of
Just (tc, tc_args)
| tc `elem` trivially_inhabited
-> case dcs of
[] -> return (Left src_ty)
(_:_) -> do var <- liftD $ mkPmId core_ty
let va = build_tm (PmVar var) dcs
return $ Right (tc, [InhabitationCandidate
{ ic_val_abs = va, ic_tm_ct = mkIdEq var
, ic_ty_cs = emptyBag, ic_strict_arg_tys = [] }])
| pmIsClosedType core_ty && not (isAbstractTyCon tc)
-- Don't consider abstract tycons since we don't know what their
-- constructors are, which makes the results of coverage checking
-- them extremely misleading.
-> liftD $ do
var <- mkPmId core_ty -- it would be wrong to unify x
alts <- mapM (mkOneConFull var tc_args . RealDataCon) (tyConDataCons tc)
return $ Right
(tc, [ alt{ic_val_abs = build_tm (ic_val_abs alt) dcs}
| alt <- alts ])
-- For other types conservatively assume that they are inhabited.
_other -> return (Left src_ty)
{- Note [Checking EmptyCase Expressions]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Empty case expressions are strict on the scrutinee. That is, `case x of {}`
will force argument `x`. Hence, `checkMatches` is not sufficient for checking
empty cases, because it assumes that the match is not strict (which is true
for all other cases, apart from EmptyCase). This gave rise to #10746. Instead,
we do the following:
1. We normalise the outermost type family redex, data family redex or newtype,
using pmTopNormaliseType_maybe (in types/FamInstEnv.hs). This computes 3
things:
(a) A normalised type src_ty, which is equal to the type of the scrutinee in
source Haskell (does not normalise newtypes or data families)
(b) The actual normalised type core_ty, which coincides with the result
topNormaliseType_maybe. This type is not necessarily equal to the input
type in source Haskell. And this is precicely the reason we compute (a)
and (c): the reasoning happens with the underlying types, but both the
patterns and types we print should respect newtypes and also show the
family type constructors and not the representation constructors.
(c) A list of all newtype data constructors dcs, each one corresponding to a
newtype rewrite performed in (b).
For an example see also Note [Type normalisation for EmptyCase]
in types/FamInstEnv.hs.
2. Function checkEmptyCase' performs the check:
- If core_ty is not an algebraic type, then we cannot check for
inhabitation, so we emit (_ :: src_ty) as missing, conservatively assuming
that the type is inhabited.
- If core_ty is an algebraic type, then we unfold the scrutinee to all
possible constructor patterns, using inhabitationCandidates, and then
check each one for constraint satisfiability, same as we for normal
pattern match checking.
%************************************************************************
%* *
Transform source syntax to *our* syntax
%* *
%************************************************************************
-}
-- -----------------------------------------------------------------------
-- * Utilities
nullaryConPattern :: ConLike -> Pattern
-- Nullary data constructor and nullary type constructor
nullaryConPattern con =
PmCon { pm_con_con = con, pm_con_arg_tys = []
, pm_con_tvs = [], pm_con_dicts = [], pm_con_args = [] }
{-# INLINE nullaryConPattern #-}
truePattern :: Pattern
truePattern = nullaryConPattern (RealDataCon trueDataCon)
{-# INLINE truePattern #-}
-- | A fake guard pattern (True <- _) used to represent cases we cannot handle
fake_pat :: Pattern
fake_pat = PmGrd { pm_grd_pv = [truePattern]
, pm_grd_expr = PmExprOther (EWildPat noExt) }
{-# INLINE fake_pat #-}
-- | Check whether a guard pattern is generated by the checker (unhandled)
isFakeGuard :: [Pattern] -> PmExpr -> Bool
isFakeGuard [PmCon { pm_con_con = RealDataCon c }] (PmExprOther (EWildPat _))
| c == trueDataCon = True
| otherwise = False
isFakeGuard _pats _e = False
-- | Generate a `canFail` pattern vector of a specific type
mkCanFailPmPat :: Type -> DsM PatVec
mkCanFailPmPat ty = do
var <- mkPmVar ty
return [var, fake_pat]
vanillaConPattern :: ConLike -> [Type] -> PatVec -> Pattern
-- ADT constructor pattern => no existentials, no local constraints
vanillaConPattern con arg_tys args =
PmCon { pm_con_con = con, pm_con_arg_tys = arg_tys
, pm_con_tvs = [], pm_con_dicts = [], pm_con_args = args }
{-# INLINE vanillaConPattern #-}
-- | Create an empty list pattern of a given type
nilPattern :: Type -> Pattern
nilPattern ty =
PmCon { pm_con_con = RealDataCon nilDataCon, pm_con_arg_tys = [ty]
, pm_con_tvs = [], pm_con_dicts = []
, pm_con_args = [] }
{-# INLINE nilPattern #-}
mkListPatVec :: Type -> PatVec -> PatVec -> PatVec
mkListPatVec ty xs ys = [PmCon { pm_con_con = RealDataCon consDataCon
, pm_con_arg_tys = [ty]
, pm_con_tvs = [], pm_con_dicts = []
, pm_con_args = xs++ys }]
{-# INLINE mkListPatVec #-}
-- | Create a (non-overloaded) literal pattern
mkLitPattern :: HsLit GhcTc -> Pattern
mkLitPattern lit = PmLit { pm_lit_lit = PmSLit lit }
{-# INLINE mkLitPattern #-}
-- -----------------------------------------------------------------------
-- * Transform (Pat Id) into of (PmPat Id)
translatePat :: FamInstEnvs -> Pat GhcTc -> DsM PatVec
translatePat fam_insts pat = case pat of
WildPat ty -> mkPmVars [ty]
VarPat _ id -> return [PmVar (unLoc id)]
ParPat _ p -> translatePat fam_insts (unLoc p)
LazyPat _ _ -> mkPmVars [hsPatType pat] -- like a variable
-- ignore strictness annotations for now
BangPat _ p -> translatePat fam_insts (unLoc p)
AsPat _ lid p -> do
-- Note [Translating As Patterns]
ps <- translatePat fam_insts (unLoc p)
let [e] = map vaToPmExpr (coercePatVec ps)
g = PmGrd [PmVar (unLoc lid)] e
return (ps ++ [g])
SigPat _ p _ty -> translatePat fam_insts (unLoc p)
-- See Note [Translate CoPats]
CoPat _ wrapper p ty
| isIdHsWrapper wrapper -> translatePat fam_insts p
| WpCast co <- wrapper, isReflexiveCo co -> translatePat fam_insts p
| otherwise -> do
ps <- translatePat fam_insts p
(xp,xe) <- mkPmId2Forms ty
let g = mkGuard ps (mkHsWrap wrapper (unLoc xe))
return [xp,g]
-- (n + k) ===> x (True <- x >= k) (n <- x-k)
NPlusKPat ty (dL->L _ _n) _k1 _k2 _ge _minus -> mkCanFailPmPat ty
-- (fun -> pat) ===> x (pat <- fun x)
ViewPat arg_ty lexpr lpat -> do
ps <- translatePat fam_insts (unLoc lpat)
-- See Note [Guards and Approximation]
case all cantFailPattern ps of
True -> do
(xp,xe) <- mkPmId2Forms arg_ty
let g = mkGuard ps (HsApp noExt lexpr xe)
return [xp,g]
False -> mkCanFailPmPat arg_ty
-- list
ListPat (ListPatTc ty Nothing) ps -> do
foldr (mkListPatVec ty) [nilPattern ty]
<$> translatePatVec fam_insts (map unLoc ps)
-- overloaded list
ListPat (ListPatTc _elem_ty (Just (pat_ty, _to_list))) lpats -> do
dflags <- getDynFlags
if xopt LangExt.RebindableSyntax dflags
then mkCanFailPmPat pat_ty
else case splitListTyConApp_maybe pat_ty of
Just e_ty -> translatePat fam_insts
(ListPat (ListPatTc e_ty Nothing) lpats)
Nothing -> mkCanFailPmPat pat_ty
-- (a) In the presence of RebindableSyntax, we don't know anything about
-- `toList`, we should treat `ListPat` as any other view pattern.
--
-- (b) In the absence of RebindableSyntax,
-- - If the pat_ty is `[a]`, then we treat the overloaded list pattern
-- as ordinary list pattern. Although we can give an instance
-- `IsList [Int]` (more specific than the default `IsList [a]`), in
-- practice, we almost never do that. We assume the `_to_list` is
-- the `toList` from `instance IsList [a]`.
--
-- - Otherwise, we treat the `ListPat` as ordinary view pattern.
--
-- See Trac #14547, especially comment#9 and comment#10.
--
-- Here we construct CanFailPmPat directly, rather can construct a view
-- pattern and do further translation as an optimization, for the reason,
-- see Note [Guards and Approximation].
ConPatOut { pat_con = (dL->L _ con)
, pat_arg_tys = arg_tys
, pat_tvs = ex_tvs
, pat_dicts = dicts
, pat_args = ps } -> do
groups <- allCompleteMatches con arg_tys
case groups of
[] -> mkCanFailPmPat (conLikeResTy con arg_tys)
_ -> do
args <- translateConPatVec fam_insts arg_tys ex_tvs con ps
return [PmCon { pm_con_con = con
, pm_con_arg_tys = arg_tys
, pm_con_tvs = ex_tvs
, pm_con_dicts = dicts
, pm_con_args = args }]
-- See Note [Translate Overloaded Literal for Exhaustiveness Checking]
NPat _ (dL->L _ olit) mb_neg _
| OverLit (OverLitTc False ty) (HsIsString src s) _ <- olit
, isStringTy ty ->
foldr (mkListPatVec charTy) [nilPattern charTy] <$>
translatePatVec fam_insts
(map (LitPat noExt . HsChar src) (unpackFS s))
| otherwise -> return [PmLit { pm_lit_lit = PmOLit (isJust mb_neg) olit }]
-- See Note [Translate Overloaded Literal for Exhaustiveness Checking]
LitPat _ lit
| HsString src s <- lit ->
foldr (mkListPatVec charTy) [nilPattern charTy] <$>
translatePatVec fam_insts
(map (LitPat noExt . HsChar src) (unpackFS s))
| otherwise -> return [mkLitPattern lit]
TuplePat tys ps boxity -> do
tidy_ps <- translatePatVec fam_insts (map unLoc ps)
let tuple_con = RealDataCon (tupleDataCon boxity (length ps))
tys' = case boxity of
Boxed -> tys
-- See Note [Unboxed tuple RuntimeRep vars] in TyCon
Unboxed -> map getRuntimeRep tys ++ tys
return [vanillaConPattern tuple_con tys' (concat tidy_ps)]
SumPat ty p alt arity -> do
tidy_p <- translatePat fam_insts (unLoc p)
let sum_con = RealDataCon (sumDataCon alt arity)
-- See Note [Unboxed tuple RuntimeRep vars] in TyCon
return [vanillaConPattern sum_con (map getRuntimeRep ty ++ ty) tidy_p]
-- --------------------------------------------------------------------------
-- Not supposed to happen
ConPatIn {} -> panic "Check.translatePat: ConPatIn"
SplicePat {} -> panic "Check.translatePat: SplicePat"
XPat {} -> panic "Check.translatePat: XPat"
{- Note [Translate Overloaded Literal for Exhaustiveness Checking]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
The translation of @NPat@ in exhaustiveness checker is a bit different
from translation in pattern matcher.
* In pattern matcher (see `tidyNPat' in deSugar/MatchLit.hs), we
translate integral literals to HsIntPrim or HsWordPrim and translate
overloaded strings to HsString.
* In exhaustiveness checker, in `genCaseTmCs1/genCaseTmCs2`, we use
`lhsExprToPmExpr` to generate uncovered set. In `hsExprToPmExpr`,
however we generate `PmOLit` for HsOverLit, rather than refine
`HsOverLit` inside `NPat` to HsIntPrim/HsWordPrim. If we do
the same thing in `translatePat` as in `tidyNPat`, the exhaustiveness
checker will fail to match the literals patterns correctly. See
Trac #14546.
In Note [Undecidable Equality for Overloaded Literals], we say: "treat
overloaded literals that look different as different", but previously we
didn't do such things.
Now, we translate the literal value to match and the literal patterns
consistently:
* For integral literals, we parse both the integral literal value and
the patterns as OverLit HsIntegral. For example:
case 0::Int of
0 -> putStrLn "A"
1 -> putStrLn "B"
_ -> putStrLn "C"
When checking the exhaustiveness of pattern matching, we translate the 0
in value position as PmOLit, but translate the 0 and 1 in pattern position
as PmSLit. The inconsistency leads to the failure of eqPmLit to detect the
equality and report warning of "Pattern match is redundant" on pattern 0,
as reported in Trac #14546. In this patch we remove the specialization of
OverLit patterns, and keep the overloaded number literal in pattern as it
is to maintain the consistency. We know nothing about the `fromInteger`
method (see Note [Undecidable Equality for Overloaded Literals]). Now we
can capture the exhaustiveness of pattern 0 and the redundancy of pattern
1 and _.
* For string literals, we parse the string literals as HsString. When
OverloadedStrings is enabled, it further be turned as HsOverLit HsIsString.
For example:
case "foo" of
"foo" -> putStrLn "A"
"bar" -> putStrLn "B"
"baz" -> putStrLn "C"
Previously, the overloaded string values are translated to PmOLit and the
non-overloaded string values are translated to PmSLit. However the string
patterns, both overloaded and non-overloaded, are translated to list of
characters. The inconsistency leads to wrong warnings about redundant and
non-exhaustive pattern matching warnings, as reported in Trac #14546.
In order to catch the redundant pattern in following case:
case "foo" of
('f':_) -> putStrLn "A"
"bar" -> putStrLn "B"
in this patch, we translate non-overloaded string literals, both in value
position and pattern position, as list of characters. For overloaded string
literals, we only translate it to list of characters only when it's type
is stringTy, since we know nothing about the toString methods. But we know
that if two overloaded strings are syntax equal, then they are equal. Then
if it's type is not stringTy, we just translate it to PmOLit. We can still
capture the exhaustiveness of pattern "foo" and the redundancy of pattern
"bar" and "baz" in the following code:
{-# LANGUAGE OverloadedStrings #-}
main = do
case "foo" of
"foo" -> putStrLn "A"
"bar" -> putStrLn "B"
"baz" -> putStrLn "C"
We must ensure that doing the same translation to literal values and patterns
in `translatePat` and `hsExprToPmExpr`. The previous inconsistent work led to
Trac #14546.
-}
-- | Translate a list of patterns (Note: each pattern is translated
-- to a pattern vector but we do not concatenate the results).
translatePatVec :: FamInstEnvs -> [Pat GhcTc] -> DsM [PatVec]
translatePatVec fam_insts pats = mapM (translatePat fam_insts) pats
-- | Translate a constructor pattern
translateConPatVec :: FamInstEnvs -> [Type] -> [TyVar]
-> ConLike -> HsConPatDetails GhcTc -> DsM PatVec
translateConPatVec fam_insts _univ_tys _ex_tvs _ (PrefixCon ps)
= concat <$> translatePatVec fam_insts (map unLoc ps)
translateConPatVec fam_insts _univ_tys _ex_tvs _ (InfixCon p1 p2)
= concat <$> translatePatVec fam_insts (map unLoc [p1,p2])
translateConPatVec fam_insts univ_tys ex_tvs c (RecCon (HsRecFields fs _))
-- Nothing matched. Make up some fresh term variables
| null fs = mkPmVars arg_tys
-- The data constructor was not defined using record syntax. For the
-- pattern to be in record syntax it should be empty (e.g. Just {}).
-- So just like the previous case.
| null orig_lbls = ASSERT(null matched_lbls) mkPmVars arg_tys
-- Some of the fields appear, in the original order (there may be holes).
-- Generate a simple constructor pattern and make up fresh variables for
-- the rest of the fields
| matched_lbls `subsetOf` orig_lbls
= ASSERT(orig_lbls `equalLength` arg_tys)
let translateOne (lbl, ty) = case lookup lbl matched_pats of
Just p -> translatePat fam_insts p
Nothing -> mkPmVars [ty]
in concatMapM translateOne (zip orig_lbls arg_tys)
-- The fields that appear are not in the correct order. Make up fresh
-- variables for all fields and add guards after matching, to force the
-- evaluation in the correct order.
| otherwise = do
arg_var_pats <- mkPmVars arg_tys
translated_pats <- forM matched_pats $ \(x,pat) -> do
pvec <- translatePat fam_insts pat
return (x, pvec)
let zipped = zip orig_lbls [ x | PmVar x <- arg_var_pats ]
guards = map (\(name,pvec) -> case lookup name zipped of
Just x -> PmGrd pvec (PmExprVar (idName x))
Nothing -> panic "translateConPatVec: lookup")
translated_pats
return (arg_var_pats ++ guards)
where
-- The actual argument types (instantiated)
arg_tys = conLikeInstOrigArgTys c (univ_tys ++ mkTyVarTys ex_tvs)
-- Some label information
orig_lbls = map flSelector $ conLikeFieldLabels c
matched_pats = [ (getName (unLoc (hsRecFieldId x)), unLoc (hsRecFieldArg x))
| (dL->L _ x) <- fs]
matched_lbls = [ name | (name, _pat) <- matched_pats ]
subsetOf :: Eq a => [a] -> [a] -> Bool
subsetOf [] _ = True
subsetOf (_:_) [] = False
subsetOf (x:xs) (y:ys)
| x == y = subsetOf xs ys
| otherwise = subsetOf (x:xs) ys
-- Translate a single match
translateMatch :: FamInstEnvs -> LMatch GhcTc (LHsExpr GhcTc)
-> DsM (PatVec,[PatVec])
translateMatch fam_insts (dL->L _ (Match { m_pats = lpats, m_grhss = grhss })) =
do
pats' <- concat <$> translatePatVec fam_insts pats
guards' <- mapM (translateGuards fam_insts) guards
return (pats', guards')
where
extractGuards :: LGRHS GhcTc (LHsExpr GhcTc) -> [GuardStmt GhcTc]
extractGuards (dL->L _ (GRHS _ gs _)) = map unLoc gs
extractGuards _ = panic "translateMatch"
pats = map unLoc lpats
guards = map extractGuards (grhssGRHSs grhss)
translateMatch _ _ = panic "translateMatch"
-- -----------------------------------------------------------------------
-- * Transform source guards (GuardStmt Id) to PmPats (Pattern)
-- | Translate a list of guard statements to a pattern vector
translateGuards :: FamInstEnvs -> [GuardStmt GhcTc] -> DsM PatVec
translateGuards fam_insts guards = do
all_guards <- concat <$> mapM (translateGuard fam_insts) guards
return (replace_unhandled all_guards)
-- It should have been (return all_guards) but it is too expressive.
-- Since the term oracle does not handle all constraints we generate,
-- we (hackily) replace all constraints the oracle cannot handle with a
-- single one (we need to know if there is a possibility of falure).
-- See Note [Guards and Approximation] for all guard-related approximations
-- we implement.
where
replace_unhandled :: PatVec -> PatVec
replace_unhandled gv
| any_unhandled gv = fake_pat : [ p | p <- gv, shouldKeep p ]
| otherwise = gv
any_unhandled :: PatVec -> Bool
any_unhandled gv = any (not . shouldKeep) gv
shouldKeep :: Pattern -> Bool
shouldKeep p
| PmVar {} <- p = True
| PmCon {} <- p = singleConstructor (pm_con_con p)
&& all shouldKeep (pm_con_args p)
shouldKeep (PmGrd pv e)
| all shouldKeep pv = True
| isNotPmExprOther e = True -- expensive but we want it
shouldKeep _other_pat = False -- let the rest..
-- | Check whether a pattern can fail to match
cantFailPattern :: Pattern -> Bool
cantFailPattern p
| PmVar {} <- p = True
| PmCon {} <- p = singleConstructor (pm_con_con p)
&& all cantFailPattern (pm_con_args p)
cantFailPattern (PmGrd pv _e)
= all cantFailPattern pv
cantFailPattern _ = False
-- | Translate a guard statement to Pattern
translateGuard :: FamInstEnvs -> GuardStmt GhcTc -> DsM PatVec
translateGuard fam_insts guard = case guard of
BodyStmt _ e _ _ -> translateBoolGuard e
LetStmt _ binds -> translateLet (unLoc binds)
BindStmt _ p e _ _ -> translateBind fam_insts p e
LastStmt {} -> panic "translateGuard LastStmt"
ParStmt {} -> panic "translateGuard ParStmt"
TransStmt {} -> panic "translateGuard TransStmt"
RecStmt {} -> panic "translateGuard RecStmt"
ApplicativeStmt {} -> panic "translateGuard ApplicativeLastStmt"
XStmtLR {} -> panic "translateGuard RecStmt"
-- | Translate let-bindings
translateLet :: HsLocalBinds GhcTc -> DsM PatVec
translateLet _binds = return []
-- | Translate a pattern guard
translateBind :: FamInstEnvs -> LPat GhcTc -> LHsExpr GhcTc -> DsM PatVec
translateBind fam_insts (dL->L _ p) e = do
ps <- translatePat fam_insts p
return [mkGuard ps (unLoc e)]
-- | Translate a boolean guard
translateBoolGuard :: LHsExpr GhcTc -> DsM PatVec
translateBoolGuard e
| isJust (isTrueLHsExpr e) = return []
-- The formal thing to do would be to generate (True <- True)
-- but it is trivial to solve so instead we give back an empty
-- PatVec for efficiency
| otherwise = return [mkGuard [truePattern] (unLoc e)]
{- Note [Guards and Approximation]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Even if the algorithm is really expressive, the term oracle we use is not.
Hence, several features are not translated *properly* but we approximate.
The list includes:
1. View Patterns
----------------
A view pattern @(f -> p)@ should be translated to @x (p <- f x)@. The term
oracle does not handle function applications so we know that the generated
constraints will not be handled at the end. Hence, we distinguish between two
cases:
a) Pattern @p@ cannot fail. Then this is just a binding and we do the *right
thing*.
b) Pattern @p@ can fail. This means that when checking the guard, we will
generate several cases, with no useful information. E.g.:
h (f -> [a,b]) = ...
h x ([a,b] <- f x) = ...
uncovered set = { [x |> { False ~ (f x ~ []) }]
, [x |> { False ~ (f x ~ (t1:[])) }]
, [x |> { False ~ (f x ~ (t1:t2:t3:t4)) }] }
So we have two problems:
1) Since we do not print the constraints in the general case (they may
be too many), the warning will look like this:
Pattern match(es) are non-exhaustive
In an equation for `h':
Patterns not matched:
_
_
_
Which is not short and not more useful than a single underscore.
2) The size of the uncovered set increases a lot, without gaining more
expressivity in our warnings.
Hence, in this case, we replace the guard @([a,b] <- f x)@ with a *dummy*
@fake_pat@: @True <- _@. That is, we record that there is a possibility
of failure but we minimize it to a True/False. This generates a single
warning and much smaller uncovered sets.
2. Overloaded Lists
-------------------
An overloaded list @[...]@ should be translated to @x ([...] <- toList x)@. The
problem is exactly like above, as its solution. For future reference, the code
below is the *right thing to do*:
ListPat (ListPatTc elem_ty (Just (pat_ty, _to_list))) lpats
otherwise -> do
(xp, xe) <- mkPmId2Forms pat_ty
ps <- translatePatVec (map unLoc lpats)
let pats = foldr (mkListPatVec elem_ty) [nilPattern elem_ty] ps
g = mkGuard pats (HsApp (noLoc to_list) xe)
return [xp,g]
3. Overloaded Literals
----------------------
The case with literals is a bit different. a literal @l@ should be translated
to @x (True <- x == from l)@. Since we want to have better warnings for
overloaded literals as it is a very common feature, we treat them differently.
They are mainly covered in Note [Undecidable Equality for Overloaded Literals]
in PmExpr.
4. N+K Patterns & Pattern Synonyms
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
An n+k pattern (n+k) should be translated to @x (True <- x >= k) (n <- x-k)@.
Since the only pattern of the three that causes failure is guard @(n <- x-k)@,
and has two possible outcomes. Hence, there is no benefit in using a dummy and
we implement the proper thing. Pattern synonyms are simply not implemented yet.
Hence, to be conservative, we generate a dummy pattern, assuming that the
pattern can fail.
5. Actual Guards
----------------
During translation, boolean guards and pattern guards are translated properly.
Let bindings though are omitted by function @translateLet@. Since they are lazy
bindings, we do not actually want to generate a (strict) equality (like we do
in the pattern bind case). Hence, we safely drop them.
Additionally, top-level guard translation (performed by @translateGuards@)
replaces guards that cannot be reasoned about (like the ones we described in
1-4) with a single @fake_pat@ to record the possibility of failure to match.
Note [Translate CoPats]
~~~~~~~~~~~~~~~~~~~~~~~
The pattern match checker did not know how to handle coerced patterns `CoPat`
efficiently, which gave rise to #11276. The original approach translated
`CoPat`s:
pat |> co ===> x (pat <- (e |> co))
Instead, we now check whether the coercion is a hole or if it is just refl, in
which case we can drop it. Unfortunately, data families generate useful
coercions so guards are still generated in these cases and checking data
families is not really efficient.
%************************************************************************
%* *
Utilities for Pattern Match Checking
%* *
%************************************************************************
-}
-- ----------------------------------------------------------------------------
-- * Basic utilities
-- | Get the type out of a PmPat. For guard patterns (ps <- e) we use the type
-- of the first (or the single -WHEREVER IT IS- valid to use?) pattern
pmPatType :: PmPat p -> Type
pmPatType (PmCon { pm_con_con = con, pm_con_arg_tys = tys })
= conLikeResTy con tys
pmPatType (PmVar { pm_var_id = x }) = idType x
pmPatType (PmLit { pm_lit_lit = l }) = pmLitType l
pmPatType (PmNLit { pm_lit_id = x }) = idType x
pmPatType (PmGrd { pm_grd_pv = pv })
= ASSERT(patVecArity pv == 1) (pmPatType p)
where Just p = find ((==1) . patternArity) pv
-- | Information about a conlike that is relevant to coverage checking.
-- It is called an \"inhabitation candidate\" since it is a value which may
-- possibly inhabit some type, but only if its term constraint ('ic_tm_ct')
-- and type constraints ('ic_ty_cs') are permitting, and if all of its strict
-- argument types ('ic_strict_arg_tys') are inhabitable.
-- See @Note [Extensions to GADTs Meet Their Match]@.
data InhabitationCandidate =
InhabitationCandidate
{ ic_val_abs :: ValAbs
, ic_tm_ct :: ComplexEq
, ic_ty_cs :: Bag EvVar
, ic_strict_arg_tys :: [Type]
}
{-
Note [Extensions to GADTs Meet Their Match]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
The GADTs Meet Their Match paper presents the formalism that GHC's coverage
checker adheres to. Since the paper's publication, there have been some
additional features added to the coverage checker which are not described in
the paper. This Note serves as a reference for these new features.
-----
-- Strict argument type constraints
-----
In the ConVar case of clause processing, each conlike K traditionally
generates two different forms of constraints:
* A term constraint (e.g., x ~ K y1 ... yn)
* Type constraints from the conlike's context (e.g., if K has type
forall bs. Q => s1 .. sn -> T tys, then Q would be its type constraints)
As it turns out, these alone are not enough to detect a certain class of
unreachable code. Consider the following example (adapted from #15305):
data K = K1 | K2 !Void
f :: K -> ()
f K1 = ()
Even though `f` doesn't match on `K2`, `f` is exhaustive in its patterns. Why?
Because it's impossible to construct a terminating value of type `K` using the
`K2` constructor, and thus it's impossible for `f` to ever successfully match
on `K2`.
The reason is because `K2`'s field of type `Void` is //strict//. Because there
are no terminating values of type `Void`, any attempt to construct something
using `K2` will immediately loop infinitely or throw an exception due to the
strictness annotation. (If the field were not strict, then `f` could match on,
say, `K2 undefined` or `K2 (let x = x in x)`.)
Since neither the term nor type constraints mentioned above take strict
argument types into account, we make use of the `nonVoid` function to
determine whether a strict type is inhabitable by a terminating value or not.
`nonVoid ty` returns True when either:
1. `ty` has at least one InhabitationCandidate for which both its term and type
constraints are satifiable, and `nonVoid` returns `True` for all of the
strict argument types in that InhabitationCandidate.
2. We're unsure if it's inhabited by a terminating value.
`nonVoid ty` returns False when `ty` is definitely uninhabited by anything
(except bottom). Some examples:
* `nonVoid Void` returns False, since Void has no InhabitationCandidates.
(This is what lets us discard the `K2` constructor in the earlier example.)
* `nonVoid (Int :~: Int)` returns True, since it has an InhabitationCandidate
(through the Refl constructor), and its term constraint (x ~ Refl) and
type constraint (Int ~ Int) are satisfiable.
* `nonVoid (Int :~: Bool)` returns False. Although it has an
InhabitationCandidate (by way of Refl), its type constraint (Int ~ Bool) is
not satisfiable.
* Given the following definition of `MyVoid`:
data MyVoid = MkMyVoid !Void
`nonVoid MyVoid` returns False. The InhabitationCandidate for the MkMyVoid
constructor contains Void as a strict argument type, and since `nonVoid Void`
returns False, that InhabitationCandidate is discarded, leaving no others.
* Performance considerations
We must be careful when recursively calling `nonVoid` on the strict argument
types of an InhabitationCandidate, because doing so naïvely can cause GHC to
fall into an infinite loop. Consider the following example:
data Abyss = MkAbyss !Abyss
stareIntoTheAbyss :: Abyss -> a
stareIntoTheAbyss x = case x of {}
In principle, stareIntoTheAbyss is exhaustive, since there is no way to
construct a terminating value using MkAbyss. However, both the term and type
constraints for MkAbyss are satisfiable, so the only way one could determine
that MkAbyss is unreachable is to check if `nonVoid Abyss` returns False.
There is only one InhabitationCandidate for Abyss—MkAbyss—and both its term
and type constraints are satisfiable, so we'd need to check if `nonVoid Abyss`
returns False... and now we've entered an infinite loop!
To avoid this sort of conundrum, `nonVoid` uses a simple test to detect the
presence of recursive types (through `checkRecTc`), and if recursion is
detected, we bail out and conservatively assume that the type is inhabited by
some terminating value. This avoids infinite loops at the expense of making
the coverage checker incomplete with respect to functions like
stareIntoTheAbyss above. Then again, the same problem occurs with recursive
newtypes, like in the following code:
newtype Chasm = MkChasm Chasm
gazeIntoTheChasm :: Chasm -> a
gazeIntoTheChasm x = case x of {} -- Erroneously warned as non-exhaustive
So this limitation is somewhat understandable.
Note that even with this recursion detection, there is still a possibility that
`nonVoid` can run in exponential time. Consider the following data type:
data T = MkT !T !T !T
If we call `nonVoid` on each of its fields, that will require us to once again
check if `MkT` is inhabitable in each of those three fields, which in turn will
require us to check if `MkT` is inhabitable again... As you can see, the
branching factor adds up quickly, and if the recursion depth limit is, say,
100, then `nonVoid T` will effectively take forever.
To mitigate this, we check the branching factor every time we are about to call
`nonVoid` on a list of strict argument types. If the branching factor exceeds 1
(i.e., if there is potential for exponential runtime), then we limit the
maximum recursion depth to 1 to mitigate the problem. If the branching factor
is exactly 1 (i.e., we have a linear chain instead of a tree), then it's okay
to stick with a larger maximum recursion depth.
Another microoptimization applies to data types like this one:
data S a = ![a] !T
Even though there is a strict field of type [a], it's quite silly to call
nonVoid on it, since it's "obvious" that it is inhabitable. To make this
intuition formal, we say that a type is definitely inhabitable (DI) if:
* It has at least one constructor C such that:
1. C has no equality constraints (since they might be unsatisfiable)
2. C has no strict argument types (since they might be uninhabitable)
It's relatively cheap to cheap if a type is DI, so before we call `nonVoid`
on a list of strict argument types, we filter out all of the DI ones.
-}
instance Outputable InhabitationCandidate where
ppr (InhabitationCandidate { ic_val_abs = va, ic_tm_ct = tm_ct
, ic_ty_cs = ty_cs
, ic_strict_arg_tys = strict_arg_tys }) =
text "InhabitationCandidate" <+>
vcat [ text "ic_val_abs =" <+> ppr va
, text "ic_tm_ct =" <+> ppr tm_ct
, text "ic_ty_cs =" <+> ppr ty_cs
, text "ic_strict_arg_tys =" <+> ppr strict_arg_tys ]
-- | Generate an 'InhabitationCandidate' for a given conlike (generate
-- fresh variables of the appropriate type for arguments)
mkOneConFull :: Id -> [Type] -> ConLike -> DsM InhabitationCandidate
-- * 'con' K is a conlike of algebraic data type 'T tys'
-- * 'tc_args' are the type arguments of the 'con's TyCon T
--
-- * 'x' is the variable for which we encode an equality constraint
-- in the term oracle
--
-- After instantiating the universal tyvars of K to tc_args we get
-- K @tys :: forall bs. Q => s1 .. sn -> T tys
--
-- Suppose y1 is a strict field. Then we get
-- Results: ic_val_abs: K (y1::s1) .. (yn::sn)
-- ic_tm_ct: x ~ K y1..yn
-- ic_ty_cs: Q
-- ic_strict_arg_tys: [s1]
mkOneConFull x tc_args con = do
let (univ_tvs, ex_tvs, eq_spec, thetas, _req_theta , arg_tys, _con_res_ty)
= conLikeFullSig con
arg_is_banged = map isBanged $ conLikeImplBangs con
subst1 = zipTvSubst univ_tvs tc_args
(subst, ex_tvs') <- cloneTyVarBndrs subst1 ex_tvs <$> getUniqueSupplyM
-- Field types
let arg_tys' = substTys subst arg_tys
-- Fresh term variables (VAs) as arguments to the constructor
arguments <- mapM mkPmVar arg_tys'
-- All constraints bound by the constructor (alpha-renamed)
let theta_cs = substTheta subst (eqSpecPreds eq_spec ++ thetas)
evvars <- mapM (nameType "pm") theta_cs
let con_abs = PmCon { pm_con_con = con
, pm_con_arg_tys = tc_args
, pm_con_tvs = ex_tvs'
, pm_con_dicts = evvars
, pm_con_args = arguments }
strict_arg_tys = filterByList arg_is_banged arg_tys'
return $ InhabitationCandidate
{ ic_val_abs = con_abs
, ic_tm_ct = (PmExprVar (idName x), vaToPmExpr con_abs)
, ic_ty_cs = listToBag evvars
, ic_strict_arg_tys = strict_arg_tys
}
-- ----------------------------------------------------------------------------
-- * More smart constructors and fresh variable generation
-- | Create a guard pattern
mkGuard :: PatVec -> HsExpr GhcTc -> Pattern
mkGuard pv e
| all cantFailPattern pv = PmGrd pv expr
| PmExprOther {} <- expr = fake_pat
| otherwise = PmGrd pv expr
where
expr = hsExprToPmExpr e
-- | Create a term equality of the form: `(False ~ (x ~ lit))`
mkNegEq :: Id -> PmLit -> ComplexEq
mkNegEq x l = (falsePmExpr, PmExprVar (idName x) `PmExprEq` PmExprLit l)
{-# INLINE mkNegEq #-}
-- | Create a term equality of the form: `(x ~ lit)`
mkPosEq :: Id -> PmLit -> ComplexEq
mkPosEq x l = (PmExprVar (idName x), PmExprLit l)
{-# INLINE mkPosEq #-}
-- | Create a term equality of the form: `(x ~ x)`
-- (always discharged by the term oracle)
mkIdEq :: Id -> ComplexEq
mkIdEq x = (PmExprVar name, PmExprVar name)
where name = idName x
{-# INLINE mkIdEq #-}
-- | Generate a variable pattern of a given type
mkPmVar :: Type -> DsM (PmPat p)
mkPmVar ty = PmVar <$> mkPmId ty
{-# INLINE mkPmVar #-}
-- | Generate many variable patterns, given a list of types
mkPmVars :: [Type] -> DsM PatVec
mkPmVars tys = mapM mkPmVar tys
{-# INLINE mkPmVars #-}
-- | Generate a fresh `Id` of a given type
mkPmId :: Type -> DsM Id
mkPmId ty = getUniqueM >>= \unique ->
let occname = mkVarOccFS $ fsLit "$pm"
name = mkInternalName unique occname noSrcSpan
in return (mkLocalId name ty)
-- | Generate a fresh term variable of a given and return it in two forms:
-- * A variable pattern
-- * A variable expression
mkPmId2Forms :: Type -> DsM (Pattern, LHsExpr GhcTc)
mkPmId2Forms ty = do
x <- mkPmId ty
return (PmVar x, noLoc (HsVar noExt (noLoc x)))
-- ----------------------------------------------------------------------------
-- * Converting between Value Abstractions, Patterns and PmExpr
-- | Convert a value abstraction an expression
vaToPmExpr :: ValAbs -> PmExpr
vaToPmExpr (PmCon { pm_con_con = c, pm_con_args = ps })
= PmExprCon c (map vaToPmExpr ps)
vaToPmExpr (PmVar { pm_var_id = x }) = PmExprVar (idName x)
vaToPmExpr (PmLit { pm_lit_lit = l }) = PmExprLit l
vaToPmExpr (PmNLit { pm_lit_id = x }) = PmExprVar (idName x)
-- | Convert a pattern vector to a list of value abstractions by dropping the
-- guards (See Note [Translating As Patterns])
coercePatVec :: PatVec -> [ValAbs]
coercePatVec pv = concatMap coercePmPat pv
-- | Convert a pattern to a list of value abstractions (will be either an empty
-- list if the pattern is a guard pattern, or a singleton list in all other
-- cases) by dropping the guards (See Note [Translating As Patterns])
coercePmPat :: Pattern -> [ValAbs]
coercePmPat (PmVar { pm_var_id = x }) = [PmVar { pm_var_id = x }]
coercePmPat (PmLit { pm_lit_lit = l }) = [PmLit { pm_lit_lit = l }]
coercePmPat (PmCon { pm_con_con = con, pm_con_arg_tys = arg_tys
, pm_con_tvs = tvs, pm_con_dicts = dicts
, pm_con_args = args })
= [PmCon { pm_con_con = con, pm_con_arg_tys = arg_tys
, pm_con_tvs = tvs, pm_con_dicts = dicts
, pm_con_args = coercePatVec args }]
coercePmPat (PmGrd {}) = [] -- drop the guards
-- | Check whether a data constructor is the only way to construct
-- a data type.
singleConstructor :: ConLike -> Bool
singleConstructor (RealDataCon dc) =
case tyConDataCons (dataConTyCon dc) of
[_] -> True
_ -> False
singleConstructor _ = False
-- | For a given conlike, finds all the sets of patterns which could
-- be relevant to that conlike by consulting the result type.
--
-- These come from two places.
-- 1. From data constructors defined with the result type constructor.
-- 2. From `COMPLETE` pragmas which have the same type as the result
-- type constructor. Note that we only use `COMPLETE` pragmas
-- *all* of whose pattern types match. See #14135
allCompleteMatches :: ConLike -> [Type] -> DsM [(Provenance, [ConLike])]
allCompleteMatches cl tys = do
let fam = case cl of
RealDataCon dc ->
[(FromBuiltin, map RealDataCon (tyConDataCons (dataConTyCon dc)))]
PatSynCon _ -> []
ty = conLikeResTy cl tys
pragmas <- case splitTyConApp_maybe ty of
Just (tc, _) -> dsGetCompleteMatches tc
Nothing -> return []
let fams cm = (FromComplete,) <$>
mapM dsLookupConLike (completeMatchConLikes cm)
from_pragma <- filter (\(_,m) -> isValidCompleteMatch ty m) <$>
mapM fams pragmas
let final_groups = fam ++ from_pragma
return final_groups
where
-- Check that all the pattern synonym return types in a `COMPLETE`
-- pragma subsume the type we're matching.
-- See Note [Filtering out non-matching COMPLETE sets]
isValidCompleteMatch :: Type -> [ConLike] -> Bool
isValidCompleteMatch ty = all go
where
go (RealDataCon {}) = True
go (PatSynCon psc) = isJust $ flip tcMatchTy ty $ patSynResTy
$ patSynSig psc
patSynResTy (_, _, _, _, _, res_ty) = res_ty
{-
Note [Filtering out non-matching COMPLETE sets]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Currently, conlikes in a COMPLETE set are simply grouped by the
type constructor heading the return type. This is nice and simple, but it does
mean that there are scenarios when a COMPLETE set might be incompatible with
the type of a scrutinee. For instance, consider (from #14135):
data Foo a = Foo1 a | Foo2 a
pattern MyFoo2 :: Int -> Foo Int
pattern MyFoo2 i = Foo2 i
{-# COMPLETE Foo1, MyFoo2 #-}
f :: Foo a -> a
f (Foo1 x) = x
`f` has an incomplete pattern-match, so when choosing which constructors to
report as unmatched in a warning, GHC must choose between the original set of
data constructors {Foo1, Foo2} and the COMPLETE set {Foo1, MyFoo2}. But observe
that GHC shouldn't even consider the COMPLETE set as a possibility: the return
type of MyFoo2, Foo Int, does not match the type of the scrutinee, Foo a, since
there's no substitution `s` such that s(Foo Int) = Foo a.
To ensure that GHC doesn't pick this COMPLETE set, it checks each pattern
synonym constructor's return type matches the type of the scrutinee, and if one
doesn't, then we remove the whole COMPLETE set from consideration.
One might wonder why GHC only checks /pattern synonym/ constructors, and not
/data/ constructors as well. The reason is because that the type of a
GADT constructor very well may not match the type of a scrutinee, and that's
OK. Consider this example (from #14059):
data SBool (z :: Bool) where
SFalse :: SBool False
STrue :: SBool True
pattern STooGoodToBeTrue :: forall (z :: Bool). ()
=> z ~ True
=> SBool z
pattern STooGoodToBeTrue = STrue
{-# COMPLETE SFalse, STooGoodToBeTrue #-}
wobble :: SBool z -> Bool
wobble STooGoodToBeTrue = True
In the incomplete pattern match for `wobble`, we /do/ want to warn that SFalse
should be matched against, even though its type, SBool False, does not match
the scrutinee type, SBool z.
-}
-- -----------------------------------------------------------------------
-- * Types and constraints
newEvVar :: Name -> Type -> EvVar
newEvVar name ty = mkLocalId name ty
nameType :: String -> Type -> DsM EvVar
nameType name ty = do
unique <- getUniqueM
let occname = mkVarOccFS (fsLit (name++"_"++show unique))
idname = mkInternalName unique occname noSrcSpan
return (newEvVar idname ty)
{-
%************************************************************************
%* *
The type oracle
%* *
%************************************************************************
-}
-- | Check whether a set of type constraints is satisfiable.
tyOracle :: Bag EvVar -> PmM Bool
tyOracle evs
= liftD $
do { ((_warns, errs), res) <- initTcDsForSolver $ tcCheckSatisfiability evs
; case res of
Just sat -> return sat
Nothing -> pprPanic "tyOracle" (vcat $ pprErrMsgBagWithLoc errs) }
{-
%************************************************************************
%* *
Sanity Checks
%* *
%************************************************************************
-}
-- | The arity of a pattern/pattern vector is the
-- number of top-level patterns that are not guards
type PmArity = Int
-- | Compute the arity of a pattern vector
patVecArity :: PatVec -> PmArity
patVecArity = sum . map patternArity
-- | Compute the arity of a pattern
patternArity :: Pattern -> PmArity
patternArity (PmGrd {}) = 0
patternArity _other_pat = 1
{-
%************************************************************************
%* *
Heart of the algorithm: Function pmcheck
%* *
%************************************************************************
Main functions are:
* mkInitialUncovered :: [Id] -> PmM Uncovered
Generates the initial uncovered set. Term and type constraints in scope
are checked, if they are inconsistent, the set is empty, otherwise, the
set contains only a vector of variables with the constraints in scope.
* pmcheck :: PatVec -> [PatVec] -> ValVec -> PmM PartialResult
Checks redundancy, coverage and inaccessibility, using auxilary functions
`pmcheckGuards` and `pmcheckHd`. Mainly handles the guard case which is
common in all three checks (see paper) and calls `pmcheckGuards` when the
whole clause is checked, or `pmcheckHd` when the pattern vector does not
start with a guard.
* pmcheckGuards :: [PatVec] -> ValVec -> PmM PartialResult
Processes the guards.
* pmcheckHd :: Pattern -> PatVec -> [PatVec]
-> ValAbs -> ValVec -> PmM PartialResult
Worker: This function implements functions `covered`, `uncovered` and
`divergent` from the paper at once. Slightly different from the paper because
it does not even produce the covered and uncovered sets. Since we only care
about whether a clause covers SOMETHING or if it may forces ANY argument, we
only store a boolean in both cases, for efficiency.
-}
-- | Lift a pattern matching action from a single value vector abstration to a
-- value set abstraction, but calling it on every vector and the combining the
-- results.
runMany :: (ValVec -> PmM PartialResult) -> (Uncovered -> PmM PartialResult)
runMany _ [] = return mempty
runMany pm (m:ms) = mappend <$> pm m <*> runMany pm ms
-- | Generate the initial uncovered set. It initializes the
-- delta with all term and type constraints in scope.
mkInitialUncovered :: [Id] -> PmM Uncovered
mkInitialUncovered vars = do
delta <- pmInitialTmTyCs
let patterns = map PmVar vars
return [ValVec patterns delta]
-- | Increase the counter for elapsed algorithm iterations, check that the
-- limit is not exceeded and call `pmcheck`
pmcheckI :: PatVec -> [PatVec] -> ValVec -> PmM PartialResult
pmcheckI ps guards vva = do
n <- liftD incrCheckPmIterDs
tracePm "pmCheck" (ppr n <> colon <+> pprPatVec ps
$$ hang (text "guards:") 2 (vcat (map pprPatVec guards))
$$ pprValVecDebug vva)
res <- pmcheck ps guards vva
tracePm "pmCheckResult:" (ppr res)
return res
{-# INLINE pmcheckI #-}
-- | Increase the counter for elapsed algorithm iterations, check that the
-- limit is not exceeded and call `pmcheckGuards`
pmcheckGuardsI :: [PatVec] -> ValVec -> PmM PartialResult
pmcheckGuardsI gvs vva = liftD incrCheckPmIterDs >> pmcheckGuards gvs vva
{-# INLINE pmcheckGuardsI #-}
-- | Increase the counter for elapsed algorithm iterations, check that the
-- limit is not exceeded and call `pmcheckHd`
pmcheckHdI :: Pattern -> PatVec -> [PatVec] -> ValAbs -> ValVec
-> PmM PartialResult
pmcheckHdI p ps guards va vva = do
n <- liftD incrCheckPmIterDs
tracePm "pmCheckHdI" (ppr n <> colon <+> pprPmPatDebug p
$$ pprPatVec ps
$$ hang (text "guards:") 2 (vcat (map pprPatVec guards))
$$ pprPmPatDebug va
$$ pprValVecDebug vva)
res <- pmcheckHd p ps guards va vva
tracePm "pmCheckHdI: res" (ppr res)
return res
{-# INLINE pmcheckHdI #-}
-- | Matching function: Check simultaneously a clause (takes separately the
-- patterns and the list of guards) for exhaustiveness, redundancy and
-- inaccessibility.
pmcheck :: PatVec -> [PatVec] -> ValVec -> PmM PartialResult
pmcheck [] guards vva@(ValVec [] _)
| null guards = return $ mempty { presultCovered = Covered }
| otherwise = pmcheckGuardsI guards vva
-- Guard
pmcheck (p@(PmGrd pv e) : ps) guards vva@(ValVec vas delta)
-- short-circuit if the guard pattern is useless.
-- we just have two possible outcomes: fail here or match and recurse
-- none of the two contains any useful information about the failure
-- though. So just have these two cases but do not do all the boilerplate
| isFakeGuard pv e = forces . mkCons vva <$> pmcheckI ps guards vva
| otherwise = do
y <- liftD $ mkPmId (pmPatType p)
let tm_state = extendSubst y e (delta_tm_cs delta)
delta' = delta { delta_tm_cs = tm_state }
utail <$> pmcheckI (pv ++ ps) guards (ValVec (PmVar y : vas) delta')
pmcheck [] _ (ValVec (_:_) _) = panic "pmcheck: nil-cons"
pmcheck (_:_) _ (ValVec [] _) = panic "pmcheck: cons-nil"
pmcheck (p:ps) guards (ValVec (va:vva) delta)
= pmcheckHdI p ps guards va (ValVec vva delta)
-- | Check the list of guards
pmcheckGuards :: [PatVec] -> ValVec -> PmM PartialResult
pmcheckGuards [] vva = return (usimple [vva])
pmcheckGuards (gv:gvs) vva = do
(PartialResult prov1 cs vsa ds) <- pmcheckI gv [] vva
(PartialResult prov2 css vsas dss) <- runMany (pmcheckGuardsI gvs) vsa
return $ PartialResult (prov1 `mappend` prov2)
(cs `mappend` css)
vsas
(ds `mappend` dss)
-- | Worker function: Implements all cases described in the paper for all three
-- functions (`covered`, `uncovered` and `divergent`) apart from the `Guard`
-- cases which are handled by `pmcheck`
pmcheckHd :: Pattern -> PatVec -> [PatVec] -> ValAbs -> ValVec
-> PmM PartialResult
-- Var
pmcheckHd (PmVar x) ps guards va (ValVec vva delta)
| Just tm_state <- solveOneEq (delta_tm_cs delta)
(PmExprVar (idName x), vaToPmExpr va)
= ucon va <$> pmcheckI ps guards (ValVec vva (delta {delta_tm_cs = tm_state}))
| otherwise = return mempty
-- ConCon
pmcheckHd ( p@(PmCon { pm_con_con = c1, pm_con_tvs = ex_tvs1
, pm_con_args = args1})) ps guards
(va@(PmCon { pm_con_con = c2, pm_con_tvs = ex_tvs2
, pm_con_args = args2})) (ValVec vva delta)
| c1 /= c2 =
return (usimple [ValVec (va:vva) delta])
| otherwise = do
let to_evvar tv1 tv2 = nameType "pmConCon" $
mkPrimEqPred (mkTyVarTy tv1) (mkTyVarTy tv2)
mb_to_evvar tv1 tv2
-- If we have identical constructors but different existential
-- tyvars, then generate extra equality constraints to ensure the
-- existential tyvars.
-- See Note [Coverage checking and existential tyvars].
| tv1 == tv2 = pure Nothing
| otherwise = Just <$> to_evvar tv1 tv2
evvars <- (listToBag . catMaybes) <$>
ASSERT(ex_tvs1 `equalLength` ex_tvs2)
liftD (zipWithM mb_to_evvar ex_tvs1 ex_tvs2)
let delta' = delta { delta_ty_cs = evvars `unionBags` delta_ty_cs delta }
kcon c1 (pm_con_arg_tys p) (pm_con_tvs p) (pm_con_dicts p)
<$> pmcheckI (args1 ++ ps) guards (ValVec (args2 ++ vva) delta')
-- LitLit
pmcheckHd (PmLit l1) ps guards (va@(PmLit l2)) vva =
case eqPmLit l1 l2 of
True -> ucon va <$> pmcheckI ps guards vva
False -> return $ ucon va (usimple [vva])
-- ConVar
pmcheckHd (p@(PmCon { pm_con_con = con, pm_con_arg_tys = tys }))
ps guards
(PmVar x) (ValVec vva delta) = do
(prov, complete_match) <- select =<< liftD (allCompleteMatches con tys)
cons_cs <- mapM (liftD . mkOneConFull x tys) complete_match
inst_vsa <- flip mapMaybeM cons_cs $
\InhabitationCandidate{ ic_val_abs = va, ic_tm_ct = tm_ct
, ic_ty_cs = ty_cs
, ic_strict_arg_tys = strict_arg_tys } -> do
mb_sat <- pmIsSatisfiable delta tm_ct ty_cs strict_arg_tys
pure $ fmap (ValVec (va:vva)) mb_sat
set_provenance prov .
force_if (canDiverge (idName x) (delta_tm_cs delta)) <$>
runMany (pmcheckI (p:ps) guards) inst_vsa
-- LitVar
pmcheckHd (p@(PmLit l)) ps guards (PmVar x) (ValVec vva delta)
= force_if (canDiverge (idName x) (delta_tm_cs delta)) <$>
mkUnion non_matched <$>
case solveOneEq (delta_tm_cs delta) (mkPosEq x l) of
Just tm_state -> pmcheckHdI p ps guards (PmLit l) $
ValVec vva (delta {delta_tm_cs = tm_state})
Nothing -> return mempty
where
us | Just tm_state <- solveOneEq (delta_tm_cs delta) (mkNegEq x l)
= [ValVec (PmNLit x [l] : vva) (delta { delta_tm_cs = tm_state })]
| otherwise = []
non_matched = usimple us
-- LitNLit
pmcheckHd (p@(PmLit l)) ps guards
(PmNLit { pm_lit_id = x, pm_lit_not = lits }) (ValVec vva delta)
| all (not . eqPmLit l) lits
, Just tm_state <- solveOneEq (delta_tm_cs delta) (mkPosEq x l)
-- Both guards check the same so it would be sufficient to have only
-- the second one. Nevertheless, it is much cheaper to check whether
-- the literal is in the list so we check it first, to avoid calling
-- the term oracle (`solveOneEq`) if possible
= mkUnion non_matched <$>
pmcheckHdI p ps guards (PmLit l)
(ValVec vva (delta { delta_tm_cs = tm_state }))
| otherwise = return non_matched
where
us | Just tm_state <- solveOneEq (delta_tm_cs delta) (mkNegEq x l)
= [ValVec (PmNLit x (l:lits) : vva) (delta { delta_tm_cs = tm_state })]
| otherwise = []
non_matched = usimple us
-- ----------------------------------------------------------------------------
-- The following three can happen only in cases like #322 where constructors
-- and overloaded literals appear in the same match. The general strategy is
-- to replace the literal (positive/negative) by a variable and recurse. The
-- fact that the variable is equal to the literal is recorded in `delta` so
-- no information is lost
-- LitCon
pmcheckHd (PmLit l) ps guards (va@(PmCon {})) (ValVec vva delta)
= do y <- liftD $ mkPmId (pmPatType va)
let tm_state = extendSubst y (PmExprLit l) (delta_tm_cs delta)
delta' = delta { delta_tm_cs = tm_state }
pmcheckHdI (PmVar y) ps guards va (ValVec vva delta')
-- ConLit
pmcheckHd (p@(PmCon {})) ps guards (PmLit l) (ValVec vva delta)
= do y <- liftD $ mkPmId (pmPatType p)
let tm_state = extendSubst y (PmExprLit l) (delta_tm_cs delta)
delta' = delta { delta_tm_cs = tm_state }
pmcheckHdI p ps guards (PmVar y) (ValVec vva delta')
-- ConNLit
pmcheckHd (p@(PmCon {})) ps guards (PmNLit { pm_lit_id = x }) vva
= pmcheckHdI p ps guards (PmVar x) vva
-- Impossible: handled by pmcheck
pmcheckHd (PmGrd {}) _ _ _ _ = panic "pmcheckHd: Guard"
{-
Note [Coverage checking and existential tyvars]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
GHC's implementation of the pattern-match coverage algorithm (as described in
the GADTs Meet Their Match paper) must take some care to emit enough type
constraints when handling data constructors with exisentially quantified type
variables. To better explain what the challenge is, consider a constructor K
of the form:
K @e_1 ... @e_m ev_1 ... ev_v ty_1 ... ty_n :: T u_1 ... u_p
Where:
* e_1, ..., e_m are the existentially bound type variables.
* ev_1, ..., ev_v are evidence variables, which may inhabit a dictionary type
(e.g., Eq) or an equality constraint (e.g., e_1 ~ Int).
* ty_1, ..., ty_n are the types of K's fields.
* T u_1 ... u_p is the return type, where T is the data type constructor, and
u_1, ..., u_p are the universally quantified type variables.
In the ConVar case, the coverage algorithm will have in hand the constructor
K as well as a list of type arguments [t_1, ..., t_n] to substitute T's
universally quantified type variables u_1, ..., u_n for. It's crucial to take
these in as arguments, as it is non-trivial to derive them just from the result
type of a pattern synonym and the ambient type of the match (#11336, #17112).
The type checker already did the hard work, so we should just make use of it.
The presence of existentially quantified type variables adds a significant
wrinkle. We always grab e_1, ..., e_m from the definition of K to begin with,
but we don't want them to appear in the final PmCon, because then
calling (mkOneConFull K) for other pattern variables might reuse the same
existential tyvars, which is certainly wrong.
Previously, GHC's solution to this wrinkle was to always create fresh names
for the existential tyvars and put them into the PmCon. This works well for
many cases, but it can break down if you nest GADT pattern matches in just
the right way. For instance, consider the following program:
data App f a where
App :: f a -> App f (Maybe a)
data Ty a where
TBool :: Ty Bool
TInt :: Ty Int
data T f a where
C :: T Ty (Maybe Bool)
foo :: T f a -> App f a -> ()
foo C (App TBool) = ()
foo is a total program, but with the previous approach to handling existential
tyvars, GHC would mark foo's patterns as non-exhaustive.
When foo is desugared to Core, it looks roughly like so:
foo @f @a (C co1 _co2) (App @a1 _co3 (TBool |> co1)) = ()
(Where `a1` is an existential tyvar.)
That, in turn, is processed by the coverage checker to become:
foo @f @a (C co1 _co2) (App @a1 _co3 (pmvar123 :: f a1))
| TBool <- pmvar123 |> co1
= ()
Note that the type of pmvar123 is `f a1`—this will be important later.
Now, we proceed with coverage-checking as usual. When we come to the
ConVar case for App, we create a fresh variable `a2` to represent its
existential tyvar. At this point, we have the equality constraints
`(a ~ Maybe a2, a ~ Maybe Bool, f ~ Ty)` in scope.
However, when we check the guard, it will use the type of pmvar123, which is
`f a1`. Thus, when considering if pmvar123 can match the constructor TInt,
it will generate the constraint `a1 ~ Int`. This means our final set of
equality constraints would be:
f ~ Ty
a ~ Maybe Bool
a ~ Maybe a2
a1 ~ Int
Which is satisfiable! Freshening the existential tyvar `a` to `a2` doomed us,
because GHC is unable to relate `a2` to `a1`, which really should be the same
tyvar.
Luckily, we can avoid this pitfall. Recall that the ConVar case was where we
generated a PmCon with too-fresh existentials. But after ConVar, we have the
ConCon case, which considers whether each constructor of a particular data type
can be matched on in a particular spot.
In the case of App, when we get to the ConCon case, we will compare our
original App PmCon (from the source program) to the App PmCon created from the
ConVar case. In the former PmCon, we have `a1` in hand, which is exactly the
existential tyvar we want! Thus, we can force `a1` to be the same as `a2` here
by emitting an additional `a1 ~ a2` constraint. Now our final set of equality
constraints will be:
f ~ Ty
a ~ Maybe Bool
a ~ Maybe a2
a1 ~ Int
a1 ~ a2
Which is unsatisfiable, as we desired, since we now have that
Int ~ a1 ~ a2 ~ Bool.
In general, App might have more than one constructor, in which case we
couldn't reuse the existential tyvar for App for a different constructor. This
means that we can only use this trick in ConCon when the constructors are the
same. But this is fine, since this is the only scenario where this situation
arises in the first place!
-}
-- ----------------------------------------------------------------------------
-- * Utilities for main checking
updateVsa :: (ValSetAbs -> ValSetAbs) -> (PartialResult -> PartialResult)
updateVsa f p@(PartialResult { presultUncovered = old })
= p { presultUncovered = f old }
-- | Initialise with default values for covering and divergent information.
usimple :: ValSetAbs -> PartialResult
usimple vsa = mempty { presultUncovered = vsa }
-- | Take the tail of all value vector abstractions in the uncovered set
utail :: PartialResult -> PartialResult
utail = updateVsa upd
where upd vsa = [ ValVec vva delta | ValVec (_:vva) delta <- vsa ]
-- | Prepend a value abstraction to all value vector abstractions in the
-- uncovered set
ucon :: ValAbs -> PartialResult -> PartialResult
ucon va = updateVsa upd
where
upd vsa = [ ValVec (va:vva) delta | ValVec vva delta <- vsa ]
-- | Given a data constructor of arity `a` and an uncovered set containing
-- value vector abstractions of length `(a+n)`, pass the first `n` value
-- abstractions to the constructor (Hence, the resulting value vector
-- abstractions will have length `n+1`)
kcon :: ConLike -> [Type] -> [TyVar] -> [EvVar]
-> PartialResult -> PartialResult
kcon con arg_tys ex_tvs dicts
= let n = conLikeArity con
upd vsa =
[ ValVec (va:vva) delta
| ValVec vva' delta <- vsa
, let (args, vva) = splitAt n vva'
, let va = PmCon { pm_con_con = con
, pm_con_arg_tys = arg_tys
, pm_con_tvs = ex_tvs
, pm_con_dicts = dicts
, pm_con_args = args } ]
in updateVsa upd
-- | Get the union of two covered, uncovered and divergent value set
-- abstractions. Since the covered and divergent sets are represented by a
-- boolean, union means computing the logical or (at least one of the two is
-- non-empty).
mkUnion :: PartialResult -> PartialResult -> PartialResult
mkUnion = mappend
-- | Add a value vector abstraction to a value set abstraction (uncovered).
mkCons :: ValVec -> PartialResult -> PartialResult
mkCons vva = updateVsa (vva:)
-- | Set the divergent set to not empty
forces :: PartialResult -> PartialResult
forces pres = pres { presultDivergent = Diverged }
-- | Set the divergent set to non-empty if the flag is `True`
force_if :: Bool -> PartialResult -> PartialResult
force_if True pres = forces pres
force_if False pres = pres
set_provenance :: Provenance -> PartialResult -> PartialResult
set_provenance prov pr = pr { presultProvenance = prov }
-- ----------------------------------------------------------------------------
-- * Propagation of term constraints inwards when checking nested matches
{- Note [Type and Term Equality Propagation]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
When checking a match it would be great to have all type and term information
available so we can get more precise results. For this reason we have functions
`addDictsDs' and `addTmCsDs' in PmMonad that store in the environment type and
term constraints (respectively) as we go deeper.
The type constraints we propagate inwards are collected by `collectEvVarsPats'
in HsPat.hs. This handles bug #4139 ( see example
https://ghc.haskell.org/trac/ghc/attachment/ticket/4139/GADTbug.hs )
where this is needed.
For term equalities we do less, we just generate equalities for HsCase. For
example we accurately give 2 redundancy warnings for the marked cases:
f :: [a] -> Bool
f x = case x of
[] -> case x of -- brings (x ~ []) in scope
[] -> True
(_:_) -> False -- can't happen
(_:_) -> case x of -- brings (x ~ (_:_)) in scope
(_:_) -> True
[] -> False -- can't happen
Functions `genCaseTmCs1' and `genCaseTmCs2' are responsible for generating
these constraints.
-}
-- | Generate equalities when checking a case expression:
-- case x of { p1 -> e1; ... pn -> en }
-- When we go deeper to check e.g. e1 we record two equalities:
-- (x ~ y), where y is the initial uncovered when checking (p1; .. ; pn)
-- and (x ~ p1).
genCaseTmCs2 :: Maybe (LHsExpr GhcTc) -- Scrutinee
-> [Pat GhcTc] -- LHS (should have length 1)
-> [Id] -- MatchVars (should have length 1)
-> DsM (Bag SimpleEq)
genCaseTmCs2 Nothing _ _ = return emptyBag
genCaseTmCs2 (Just scr) [p] [var] = do
fam_insts <- dsGetFamInstEnvs
[e] <- map vaToPmExpr . coercePatVec <$> translatePat fam_insts p
let scr_e = lhsExprToPmExpr scr
return $ listToBag [(var, e), (var, scr_e)]
genCaseTmCs2 _ _ _ = panic "genCaseTmCs2: HsCase"
-- | Generate a simple equality when checking a case expression:
-- case x of { matches }
-- When checking matches we record that (x ~ y) where y is the initial
-- uncovered. All matches will have to satisfy this equality.
genCaseTmCs1 :: Maybe (LHsExpr GhcTc) -> [Id] -> Bag SimpleEq
genCaseTmCs1 Nothing _ = emptyBag
genCaseTmCs1 (Just scr) [var] = unitBag (var, lhsExprToPmExpr scr)
genCaseTmCs1 _ _ = panic "genCaseTmCs1: HsCase"
{- Note [Literals in PmPat]
~~~~~~~~~~~~~~~~~~~~~~~~~~~
Instead of translating a literal to a variable accompanied with a guard, we
treat them like constructor patterns. The following example from
"./libraries/base/GHC/IO/Encoding.hs" shows why:
mkTextEncoding' :: CodingFailureMode -> String -> IO TextEncoding
mkTextEncoding' cfm enc = case [toUpper c | c <- enc, c /= '-'] of
"UTF8" -> return $ UTF8.mkUTF8 cfm
"UTF16" -> return $ UTF16.mkUTF16 cfm
"UTF16LE" -> return $ UTF16.mkUTF16le cfm
...
Each clause gets translated to a list of variables with an equal number of
guards. For every guard we generate two cases (equals True/equals False) which
means that we generate 2^n cases to feed the oracle with, where n is the sum of
the length of all strings that appear in the patterns. For this particular
example this means over 2^40 cases. Instead, by representing them like with
constructor we get the following:
1. We exploit the common prefix with our representation of VSAs
2. We prune immediately non-reachable cases
(e.g. False == (x == "U"), True == (x == "U"))
Note [Translating As Patterns]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Instead of translating x@p as: x (p <- x)
we instead translate it as: p (x <- coercePattern p)
for performance reasons. For example:
f x@True = 1
f y@False = 2
Gives the following with the first translation:
x |> {x == False, x == y, y == True}
If we use the second translation we get an empty set, independently of the
oracle. Since the pattern `p' may contain guard patterns though, it cannot be
used as an expression. That's why we call `coercePatVec' to drop the guard and
`vaToPmExpr' to transform the value abstraction to an expression in the
guard pattern (value abstractions are a subset of expressions). We keep the
guards in the first pattern `p' though.
%************************************************************************
%* *
Pretty printing of exhaustiveness/redundancy check warnings
%* *
%************************************************************************
-}
-- | Check whether any part of pattern match checking is enabled (does not
-- matter whether it is the redundancy check or the exhaustiveness check).
isAnyPmCheckEnabled :: DynFlags -> DsMatchContext -> Bool
isAnyPmCheckEnabled dflags (DsMatchContext kind _loc)
= wopt Opt_WarnOverlappingPatterns dflags || exhaustive dflags kind
instance Outputable ValVec where
ppr (ValVec vva delta)
= let (residual_eqs, subst) = wrapUpTmState (delta_tm_cs delta)
vector = substInValAbs subst vva
in ppr_uncovered (vector, residual_eqs)
-- | Apply a term substitution to a value vector abstraction. All VAs are
-- transformed to PmExpr (used only before pretty printing).
substInValAbs :: PmVarEnv -> [ValAbs] -> [PmExpr]
substInValAbs subst = map (exprDeepLookup subst . vaToPmExpr)
-- | Wrap up the term oracle's state once solving is complete. Drop any
-- information about unhandled constraints (involving HsExprs) and flatten
-- (height 1) the substitution.
wrapUpTmState :: TmState -> ([ComplexEq], PmVarEnv)
wrapUpTmState (residual, (_, subst)) = (residual, flattenPmVarEnv subst)
-- | Issue all the warnings (coverage, exhaustiveness, inaccessibility)
dsPmWarn :: DynFlags -> DsMatchContext -> PmResult -> DsM ()
dsPmWarn dflags ctx@(DsMatchContext kind loc) pm_result
= when (flag_i || flag_u) $ do
let exists_r = flag_i && notNull redundant && onlyBuiltin
exists_i = flag_i && notNull inaccessible && onlyBuiltin && not is_rec_upd
exists_u = flag_u && (case uncovered of
TypeOfUncovered _ -> True
UncoveredPatterns u -> notNull u)
when exists_r $ forM_ redundant $ \(dL->L l q) -> do
putSrcSpanDs l (warnDs (Reason Opt_WarnOverlappingPatterns)
(pprEqn q "is redundant"))
when exists_i $ forM_ inaccessible $ \(dL->L l q) -> do
putSrcSpanDs l (warnDs (Reason Opt_WarnOverlappingPatterns)
(pprEqn q "has inaccessible right hand side"))
when exists_u $ putSrcSpanDs loc $ warnDs flag_u_reason $
case uncovered of
TypeOfUncovered ty -> warnEmptyCase ty
UncoveredPatterns candidates -> pprEqns candidates
where
PmResult
{ pmresultProvenance = prov
, pmresultRedundant = redundant
, pmresultUncovered = uncovered
, pmresultInaccessible = inaccessible } = pm_result
flag_i = wopt Opt_WarnOverlappingPatterns dflags
flag_u = exhaustive dflags kind
flag_u_reason = maybe NoReason Reason (exhaustiveWarningFlag kind)
is_rec_upd = case kind of { RecUpd -> True; _ -> False }
-- See Note [Inaccessible warnings for record updates]
onlyBuiltin = prov == FromBuiltin
maxPatterns = maxUncoveredPatterns dflags
-- Print a single clause (for redundant/with-inaccessible-rhs)
pprEqn q txt = pp_context True ctx (text txt) $ \f -> ppr_eqn f kind q
-- Print several clauses (for uncovered clauses)
pprEqns qs = pp_context False ctx (text "are non-exhaustive") $ \_ ->
case qs of -- See #11245
[ValVec [] _]
-> text "Guards do not cover entire pattern space"
_missing -> let us = map ppr qs
in hang (text "Patterns not matched:") 4
(vcat (take maxPatterns us)
$$ dots maxPatterns us)
-- Print a type-annotated wildcard (for non-exhaustive `EmptyCase`s for
-- which we only know the type and have no inhabitants at hand)
warnEmptyCase ty = pp_context False ctx (text "are non-exhaustive") $ \_ ->
hang (text "Patterns not matched:") 4 (underscore <+> dcolon <+> ppr ty)
{- Note [Inaccessible warnings for record updates]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Consider (Trac #12957)
data T a where
T1 :: { x :: Int } -> T Bool
T2 :: { x :: Int } -> T a
T3 :: T a
f :: T Char -> T a
f r = r { x = 3 }
The desugarer will (conservatively generate a case for T1 even though
it's impossible:
f r = case r of
T1 x -> T1 3 -- Inaccessible branch
T2 x -> T2 3
_ -> error "Missing"
We don't want to warn about the inaccessible branch because the programmer
didn't put it there! So we filter out the warning here.
-}
-- | Issue a warning when the predefined number of iterations is exceeded
-- for the pattern match checker
warnPmIters :: DynFlags -> DsMatchContext -> DsM ()
warnPmIters dflags (DsMatchContext kind loc)
= when (flag_i || flag_u) $ do
iters <- maxPmCheckIterations <$> getDynFlags
putSrcSpanDs loc (warnDs NoReason (msg iters))
where
ctxt = pprMatchContext kind
msg is = fsep [ text "Pattern match checker exceeded"
, parens (ppr is), text "iterations in", ctxt <> dot
, text "(Use -fmax-pmcheck-iterations=n"
, text "to set the maximum number of iterations to n)" ]
flag_i = wopt Opt_WarnOverlappingPatterns dflags
flag_u = exhaustive dflags kind
dots :: Int -> [a] -> SDoc
dots maxPatterns qs
| qs `lengthExceeds` maxPatterns = text "..."
| otherwise = empty
-- | Check whether the exhaustiveness checker should run (exhaustiveness only)
exhaustive :: DynFlags -> HsMatchContext id -> Bool
exhaustive dflags = maybe False (`wopt` dflags) . exhaustiveWarningFlag
-- | Denotes whether an exhaustiveness check is supported, and if so,
-- via which 'WarningFlag' it's controlled.
-- Returns 'Nothing' if check is not supported.
exhaustiveWarningFlag :: HsMatchContext id -> Maybe WarningFlag
exhaustiveWarningFlag (FunRhs {}) = Just Opt_WarnIncompletePatterns
exhaustiveWarningFlag CaseAlt = Just Opt_WarnIncompletePatterns
exhaustiveWarningFlag IfAlt = Just Opt_WarnIncompletePatterns
exhaustiveWarningFlag LambdaExpr = Just Opt_WarnIncompleteUniPatterns
exhaustiveWarningFlag PatBindRhs = Just Opt_WarnIncompleteUniPatterns
exhaustiveWarningFlag PatBindGuards = Just Opt_WarnIncompletePatterns
exhaustiveWarningFlag ProcExpr = Just Opt_WarnIncompleteUniPatterns
exhaustiveWarningFlag RecUpd = Just Opt_WarnIncompletePatternsRecUpd
exhaustiveWarningFlag ThPatSplice = Nothing
exhaustiveWarningFlag PatSyn = Nothing
exhaustiveWarningFlag ThPatQuote = Nothing
exhaustiveWarningFlag (StmtCtxt {}) = Nothing -- Don't warn about incomplete patterns
-- in list comprehensions, pattern guards
-- etc. They are often *supposed* to be
-- incomplete
-- True <==> singular
pp_context :: Bool -> DsMatchContext -> SDoc -> ((SDoc -> SDoc) -> SDoc) -> SDoc
pp_context singular (DsMatchContext kind _loc) msg rest_of_msg_fun
= vcat [text txt <+> msg,
sep [ text "In" <+> ppr_match <> char ':'
, nest 4 (rest_of_msg_fun pref)]]
where
txt | singular = "Pattern match"
| otherwise = "Pattern match(es)"
(ppr_match, pref)
= case kind of
FunRhs { mc_fun = (dL->L _ fun) }
-> (pprMatchContext kind, \ pp -> ppr fun <+> pp)
_ -> (pprMatchContext kind, \ pp -> pp)
ppr_pats :: HsMatchContext Name -> [Pat GhcTc] -> SDoc
ppr_pats kind pats
= sep [sep (map ppr pats), matchSeparator kind, text "..."]
ppr_eqn :: (SDoc -> SDoc) -> HsMatchContext Name -> [LPat GhcTc] -> SDoc
ppr_eqn prefixF kind eqn = prefixF (ppr_pats kind (map unLoc eqn))
ppr_constraint :: (SDoc,[PmLit]) -> SDoc
ppr_constraint (var, lits) = var <+> text "is not one of"
<+> braces (pprWithCommas ppr lits)
ppr_uncovered :: ([PmExpr], [ComplexEq]) -> SDoc
ppr_uncovered (expr_vec, complex)
| null cs = fsep vec -- there are no literal constraints
| otherwise = hang (fsep vec) 4 $
text "where" <+> vcat (map ppr_constraint cs)
where
sdoc_vec = mapM pprPmExprWithParens expr_vec
(vec,cs) = runPmPprM sdoc_vec (filterComplex complex)
{- Note [Representation of Term Equalities]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
In the paper, term constraints always take the form (x ~ e). Of course, a more
general constraint of the form (e1 ~ e1) can always be transformed to an
equivalent set of the former constraints, by introducing a fresh, intermediate
variable: { y ~ e1, y ~ e1 }. Yet, implementing this representation gave rise
to #11160 (incredibly bad performance for literal pattern matching). Two are
the main sources of this problem (the actual problem is how these two interact
with each other):
1. Pattern matching on literals generates twice as many constraints as needed.
Consider the following (tests/ghci/should_run/ghcirun004):
foo :: Int -> Int
foo 1 = 0
...
foo 5000 = 4999
The covered and uncovered set *should* look like:
U0 = { x |> {} }
C1 = { 1 |> { x ~ 1 } }
U1 = { x |> { False ~ (x ~ 1) } }
...
C10 = { 10 |> { False ~ (x ~ 1), .., False ~ (x ~ 9), x ~ 10 } }
U10 = { x |> { False ~ (x ~ 1), .., False ~ (x ~ 9), False ~ (x ~ 10) } }
...
If we replace { False ~ (x ~ 1) } with { y ~ False, y ~ (x ~ 1) }
we get twice as many constraints. Also note that half of them are just the
substitution [x |-> False].
2. The term oracle (`tmOracle` in deSugar/TmOracle) uses equalities of the form
(x ~ e) as substitutions [x |-> e]. More specifically, function
`extendSubstAndSolve` applies such substitutions in the residual constraints
and partitions them in the affected and non-affected ones, which are the new
worklist. Essentially, this gives quadradic behaviour on the number of the
residual constraints. (This would not be the case if the term oracle used
mutable variables but, since we use it to handle disjunctions on value set
abstractions (`Union` case), we chose a pure, incremental interface).
Now the problem becomes apparent (e.g. for clause 300):
* Set U300 contains 300 substituting constraints [y_i |-> False] and 300
constraints that we know that will not reduce (stay in the worklist).
* To check for consistency, we apply the substituting constraints ONE BY ONE
(since `tmOracle` is called incrementally, it does not have all of them
available at once). Hence, we go through the (non-progressing) constraints
over and over, achieving over-quadradic behaviour.
If instead we allow constraints of the form (e ~ e),
* All uncovered sets Ui contain no substituting constraints and i
non-progressing constraints of the form (False ~ (x ~ lit)) so the oracle
behaves linearly.
* All covered sets Ci contain exactly (i-1) non-progressing constraints and
a single substituting constraint. So the term oracle goes through the
constraints only once.
The performance improvement becomes even more important when more arguments are
involved.
-}
-- Debugging Infrastructre
tracePm :: String -> SDoc -> PmM ()
tracePm herald doc = liftD $ tracePmD herald doc
tracePmD :: String -> SDoc -> DsM ()
tracePmD herald doc = do
dflags <- getDynFlags
printer <- mkPrintUnqualifiedDs
liftIO $ dumpIfSet_dyn_printer printer dflags
Opt_D_dump_ec_trace (text herald $$ (nest 2 doc))
pprPmPatDebug :: PmPat a -> SDoc
pprPmPatDebug (PmCon cc _arg_tys _con_tvs _con_dicts con_args)
= hsep [text "PmCon", ppr cc, hsep (map pprPmPatDebug con_args)]
pprPmPatDebug (PmVar vid) = text "PmVar" <+> ppr vid
pprPmPatDebug (PmLit li) = text "PmLit" <+> ppr li
pprPmPatDebug (PmNLit i nl) = text "PmNLit" <+> ppr i <+> ppr nl
pprPmPatDebug (PmGrd pv ge) = text "PmGrd" <+> hsep (map pprPmPatDebug pv)
<+> ppr ge
pprPatVec :: PatVec -> SDoc
pprPatVec ps = hang (text "Pattern:") 2
(brackets $ sep
$ punctuate (comma <> char '\n') (map pprPmPatDebug ps))
pprValAbs :: [ValAbs] -> SDoc
pprValAbs ps = hang (text "ValAbs:") 2
(brackets $ sep
$ punctuate (comma) (map pprPmPatDebug ps))
pprValVecDebug :: ValVec -> SDoc
pprValVecDebug (ValVec vas _d) = text "ValVec" <+>
parens (pprValAbs vas)