----------------------------------------------------------------------------- -- | -- Module : Data.SBV.Provers.Prover -- Copyright : (c) Levent Erkok -- License : BSD3 -- Maintainer : erkokl@gmail.com -- Stability : experimental -- -- Provable abstraction and the connection to SMT solvers ----------------------------------------------------------------------------- {-# LANGUAGE CPP #-} {-# LANGUAGE BangPatterns #-} {-# LANGUAGE FlexibleInstances #-} {-# LANGUAGE LambdaCase #-} {-# LANGUAGE NamedFieldPuns #-} {-# LANGUAGE ScopedTypeVariables #-} {-# LANGUAGE TypeSynonymInstances #-} module Data.SBV.Provers.Prover ( SMTSolver(..), SMTConfig(..), Predicate, Provable(..), Goal , ThmResult(..), SatResult(..), AllSatResult(..), SafeResult(..), OptimizeResult(..), SMTResult(..) , isSatisfiable, isSatisfiableWith, isTheorem, isTheoremWith , prove, proveWith , sat, satWith , allSat, allSatWith , safe, safeWith, isSafe , optimize, optimizeWith , isVacuous, isVacuousWith , SatModel(..), Modelable(..), displayModels, extractModels , getModelDictionaries, getModelValues, getModelUninterpretedValues , boolector, cvc4, yices, z3, mathSAT, abc, defaultSMTCfg , compileToSMTLib, generateSMTBenchmarks , internalSATCheck ) where import Data.Char (isSpace) import Data.List (intercalate, nub) import Control.Monad (when, unless) import System.FilePath (addExtension, splitExtension) import System.Time (getClockTime) import System.IO (hGetBuffering, hSetBuffering, stdout, hFlush, BufferMode(..)) import System.IO.Unsafe (unsafeInterleaveIO) import Control.Concurrent.Async (async, wait, cancel, waitAny, Async) import GHC.Stack.Compat #if !MIN_VERSION_base(4,9,0) import GHC.SrcLoc.Compat #endif import qualified Data.Set as Set (toList) import Data.SBV.Core.Data import Data.SBV.Core.Symbolic import Data.SBV.SMT.SMT import Data.SBV.SMT.SMTLib import Data.SBV.Utils.TDiff import Control.DeepSeq (rnf) import Control.Exception (bracket) import qualified Data.SBV.Provers.Boolector as Boolector import qualified Data.SBV.Provers.CVC4 as CVC4 import qualified Data.SBV.Provers.Yices as Yices import qualified Data.SBV.Provers.Z3 as Z3 import qualified Data.SBV.Provers.MathSAT as MathSAT import qualified Data.SBV.Provers.ABC as ABC mkConfig :: SMTSolver -> SMTLibVersion -> [String] -> SMTConfig mkConfig s smtVersion tweaks = SMTConfig { verbose = False , timing = NoTiming , sBranchTimeOut = Nothing , timeOut = Nothing , printBase = 10 , printRealPrec = 16 , smtFile = Nothing , solver = s , solverTweaks = tweaks , smtLibVersion = smtVersion , optimizeArgs = [] , satCmd = "(check-sat)" , isNonModelVar = const False -- i.e., everything is a model-variable by default , roundingMode = RoundNearestTiesToEven , useLogic = Nothing , getUnsatCore = False } -- | Default configuration for the Boolector SMT solver boolector :: SMTConfig boolector = mkConfig Boolector.boolector SMTLib2 [] -- | Default configuration for the CVC4 SMT Solver. cvc4 :: SMTConfig cvc4 = mkConfig CVC4.cvc4 SMTLib2 [] -- | Default configuration for the Yices SMT Solver. yices :: SMTConfig yices = mkConfig Yices.yices SMTLib2 [] -- | Default configuration for the Z3 SMT solver z3 :: SMTConfig z3 = mkConfig Z3.z3 SMTLib2 ["(set-option :smt.mbqi true) ; use model based quantifier instantiation"] -- | Default configuration for the MathSAT SMT solver mathSAT :: SMTConfig mathSAT = mkConfig MathSAT.mathSAT SMTLib2 [] -- | Default configuration for the ABC synthesis and verification tool. abc :: SMTConfig abc = mkConfig ABC.abc SMTLib2 [] -- | The default solver used by SBV. This is currently set to z3. defaultSMTCfg :: SMTConfig defaultSMTCfg = z3 -- | A predicate is a symbolic program that returns a (symbolic) boolean value. For all intents and -- purposes, it can be treated as an n-ary function from symbolic-values to a boolean. The 'Symbolic' -- monad captures the underlying representation, and can/should be ignored by the users of the library, -- unless you are building further utilities on top of SBV itself. Instead, simply use the 'Predicate' -- type when necessary. type Predicate = Symbolic SBool -- | A goal is a symbolic program that returns no values. The idea is that the constraints/min-max -- goals will serve as appropriate directives for sat/prove calls. type Goal = Symbolic () -- | A type @a@ is provable if we can turn it into a predicate. -- Note that a predicate can be made from a curried function of arbitrary arity, where -- each element is either a symbolic type or up-to a 7-tuple of symbolic-types. So -- predicates can be constructed from almost arbitrary Haskell functions that have arbitrary -- shapes. (See the instance declarations below.) class Provable a where -- | Turns a value into a universally quantified predicate, internally naming the inputs. -- In this case the sbv library will use names of the form @s1, s2@, etc. to name these variables -- Example: -- -- > forAll_ $ \(x::SWord8) y -> x `shiftL` 2 .== y -- -- is a predicate with two arguments, captured using an ordinary Haskell function. Internally, -- @x@ will be named @s0@ and @y@ will be named @s1@. forAll_ :: a -> Predicate -- | Turns a value into a predicate, allowing users to provide names for the inputs. -- If the user does not provide enough number of names for the variables, the remaining ones -- will be internally generated. Note that the names are only used for printing models and has no -- other significance; in particular, we do not check that they are unique. Example: -- -- > forAll ["x", "y"] $ \(x::SWord8) y -> x `shiftL` 2 .== y -- -- This is the same as above, except the variables will be named @x@ and @y@ respectively, -- simplifying the counter-examples when they are printed. forAll :: [String] -> a -> Predicate -- | Turns a value into an existentially quantified predicate. (Indeed, 'exists' would have been -- a better choice here for the name, but alas it's already taken.) forSome_ :: a -> Predicate -- | Version of 'forSome' that allows user defined names forSome :: [String] -> a -> Predicate instance Provable Predicate where forAll_ = id forAll [] = id forAll xs = error $ "SBV.forAll: Extra unmapped name(s) in predicate construction: " ++ intercalate ", " xs forSome_ = id forSome [] = id forSome xs = error $ "SBV.forSome: Extra unmapped name(s) in predicate construction: " ++ intercalate ", " xs instance Provable SBool where forAll_ = return forAll _ = return forSome_ = return forSome _ = return {- -- The following works, but it lets us write properties that -- are not useful.. Such as: prove $ \x y -> (x::SInt8) == y -- Running that will throw an exception since Haskell's equality -- is not be supported by symbolic things. (Needs .==). instance Provable Bool where forAll_ x = forAll_ (if x then true else false :: SBool) forAll s x = forAll s (if x then true else false :: SBool) forSome_ x = forSome_ (if x then true else false :: SBool) forSome s x = forSome s (if x then true else false :: SBool) -} -- Functions instance (SymWord a, Provable p) => Provable (SBV a -> p) where forAll_ k = forall_ >>= \a -> forAll_ $ k a forAll (s:ss) k = forall s >>= \a -> forAll ss $ k a forAll [] k = forAll_ k forSome_ k = exists_ >>= \a -> forSome_ $ k a forSome (s:ss) k = exists s >>= \a -> forSome ss $ k a forSome [] k = forSome_ k -- SFunArrays (memory, functional representation), only supported universally for the time being instance (HasKind a, HasKind b, Provable p) => Provable (SArray a b -> p) where forAll_ k = declNewSArray (\t -> "array_" ++ show t) Nothing >>= \a -> forAll_ $ k a forAll (s:ss) k = declNewSArray (const s) Nothing >>= \a -> forAll ss $ k a forAll [] k = forAll_ k forSome_ _ = error "SBV.forSome: Existential arrays are not currently supported." forSome _ _ = error "SBV.forSome: Existential arrays are not currently supported." -- SArrays (memory, SMT-Lib notion of arrays), only supported universally for the time being instance (HasKind a, HasKind b, Provable p) => Provable (SFunArray a b -> p) where forAll_ k = declNewSFunArray Nothing >>= \a -> forAll_ $ k a forAll (_:ss) k = declNewSFunArray Nothing >>= \a -> forAll ss $ k a forAll [] k = forAll_ k forSome_ _ = error "SBV.forSome: Existential arrays are not currently supported." forSome _ _ = error "SBV.forSome: Existential arrays are not currently supported." -- 2 Tuple instance (SymWord a, SymWord b, Provable p) => Provable ((SBV a, SBV b) -> p) where forAll_ k = forall_ >>= \a -> forAll_ $ \b -> k (a, b) forAll (s:ss) k = forall s >>= \a -> forAll ss $ \b -> k (a, b) forAll [] k = forAll_ k forSome_ k = exists_ >>= \a -> forSome_ $ \b -> k (a, b) forSome (s:ss) k = exists s >>= \a -> forSome ss $ \b -> k (a, b) forSome [] k = forSome_ k -- 3 Tuple instance (SymWord a, SymWord b, SymWord c, Provable p) => Provable ((SBV a, SBV b, SBV c) -> p) where forAll_ k = forall_ >>= \a -> forAll_ $ \b c -> k (a, b, c) forAll (s:ss) k = forall s >>= \a -> forAll ss $ \b c -> k (a, b, c) forAll [] k = forAll_ k forSome_ k = exists_ >>= \a -> forSome_ $ \b c -> k (a, b, c) forSome (s:ss) k = exists s >>= \a -> forSome ss $ \b c -> k (a, b, c) forSome [] k = forSome_ k -- 4 Tuple instance (SymWord a, SymWord b, SymWord c, SymWord d, Provable p) => Provable ((SBV a, SBV b, SBV c, SBV d) -> p) where forAll_ k = forall_ >>= \a -> forAll_ $ \b c d -> k (a, b, c, d) forAll (s:ss) k = forall s >>= \a -> forAll ss $ \b c d -> k (a, b, c, d) forAll [] k = forAll_ k forSome_ k = exists_ >>= \a -> forSome_ $ \b c d -> k (a, b, c, d) forSome (s:ss) k = exists s >>= \a -> forSome ss $ \b c d -> k (a, b, c, d) forSome [] k = forSome_ k -- 5 Tuple instance (SymWord a, SymWord b, SymWord c, SymWord d, SymWord e, Provable p) => Provable ((SBV a, SBV b, SBV c, SBV d, SBV e) -> p) where forAll_ k = forall_ >>= \a -> forAll_ $ \b c d e -> k (a, b, c, d, e) forAll (s:ss) k = forall s >>= \a -> forAll ss $ \b c d e -> k (a, b, c, d, e) forAll [] k = forAll_ k forSome_ k = exists_ >>= \a -> forSome_ $ \b c d e -> k (a, b, c, d, e) forSome (s:ss) k = exists s >>= \a -> forSome ss $ \b c d e -> k (a, b, c, d, e) forSome [] k = forSome_ k -- 6 Tuple instance (SymWord a, SymWord b, SymWord c, SymWord d, SymWord e, SymWord f, Provable p) => Provable ((SBV a, SBV b, SBV c, SBV d, SBV e, SBV f) -> p) where forAll_ k = forall_ >>= \a -> forAll_ $ \b c d e f -> k (a, b, c, d, e, f) forAll (s:ss) k = forall s >>= \a -> forAll ss $ \b c d e f -> k (a, b, c, d, e, f) forAll [] k = forAll_ k forSome_ k = exists_ >>= \a -> forSome_ $ \b c d e f -> k (a, b, c, d, e, f) forSome (s:ss) k = exists s >>= \a -> forSome ss $ \b c d e f -> k (a, b, c, d, e, f) forSome [] k = forSome_ k -- 7 Tuple instance (SymWord a, SymWord b, SymWord c, SymWord d, SymWord e, SymWord f, SymWord g, Provable p) => Provable ((SBV a, SBV b, SBV c, SBV d, SBV e, SBV f, SBV g) -> p) where forAll_ k = forall_ >>= \a -> forAll_ $ \b c d e f g -> k (a, b, c, d, e, f, g) forAll (s:ss) k = forall s >>= \a -> forAll ss $ \b c d e f g -> k (a, b, c, d, e, f, g) forAll [] k = forAll_ k forSome_ k = exists_ >>= \a -> forSome_ $ \b c d e f g -> k (a, b, c, d, e, f, g) forSome (s:ss) k = exists s >>= \a -> forSome ss $ \b c d e f g -> k (a, b, c, d, e, f, g) forSome [] k = forSome_ k -- | Prove a predicate, equivalent to @'proveWith' 'defaultSMTCfg'@ prove :: Provable a => a -> IO ThmResult prove = proveWith defaultSMTCfg -- | Find a satisfying assignment for a predicate, equivalent to @'satWith' 'defaultSMTCfg'@ sat :: Provable a => a -> IO SatResult sat = satWith defaultSMTCfg -- | Return all satisfying assignments for a predicate, equivalent to @'allSatWith' 'defaultSMTCfg'@. -- Satisfying assignments are constructed lazily, so they will be available as returned by the solver -- and on demand. -- -- NB. Uninterpreted constant/function values and counter-examples for array values are ignored for -- the purposes of @'allSat'@. That is, only the satisfying assignments modulo uninterpreted functions and -- array inputs will be returned. This is due to the limitation of not having a robust means of getting a -- function counter-example back from the SMT solver. allSat :: Provable a => a -> IO AllSatResult allSat = allSatWith defaultSMTCfg -- | Optimize a given collection of `Objective`s optimize :: Provable a => a -> IO OptimizeResult optimize = optimizeWith defaultSMTCfg -- | Check that all the 'sAssert' calls are safe, equivalent to @'safeWith' 'defaultSMTCfg'@ safe :: SExecutable a => a -> IO [SafeResult] safe = safeWith defaultSMTCfg -- | Check if the given constraints are satisfiable, equivalent to @'isVacuousWith' 'defaultSMTCfg'@. -- See the function 'constrain' for an example use of 'isVacuous'. Also see the 'CheckConstrVacuity' -- tactic. isVacuous :: Provable a => a -> IO Bool isVacuous = isVacuousWith defaultSMTCfg -- Decision procedures (with optional timeout) -- | Check whether a given property is a theorem, with an optional time out and the given solver. -- Returns @Nothing@ if times out, or the result wrapped in a @Just@ otherwise. isTheoremWith :: Provable a => SMTConfig -> Maybe Int -> a -> IO (Maybe Bool) isTheoremWith cfg mbTo p = do r <- proveWith cfg{timeOut = mbTo} p case r of ThmResult Unsatisfiable{} -> return $ Just True ThmResult Satisfiable{} -> return $ Just False ThmResult TimeOut{} -> return Nothing _ -> error $ "SBV.isTheorem: Received:\n" ++ show r -- | Check whether a given property is satisfiable, with an optional time out and the given solver. -- Returns @Nothing@ if times out, or the result wrapped in a @Just@ otherwise. isSatisfiableWith :: Provable a => SMTConfig -> Maybe Int -> a -> IO (Maybe Bool) isSatisfiableWith cfg mbTo p = do r <- satWith cfg{timeOut = mbTo} p case r of SatResult Satisfiable{} -> return $ Just True SatResult Unsatisfiable{} -> return $ Just False SatResult TimeOut{} -> return Nothing _ -> error $ "SBV.isSatisfiable: Received: " ++ show r -- | Checks theoremhood within the given optional time limit of @i@ seconds. -- Returns @Nothing@ if times out, or the result wrapped in a @Just@ otherwise. isTheorem :: Provable a => Maybe Int -> a -> IO (Maybe Bool) isTheorem = isTheoremWith defaultSMTCfg -- | Checks satisfiability within the given optional time limit of @i@ seconds. -- Returns @Nothing@ if times out, or the result wrapped in a @Just@ otherwise. isSatisfiable :: Provable a => Maybe Int -> a -> IO (Maybe Bool) isSatisfiable = isSatisfiableWith defaultSMTCfg -- | Compiles to SMT-Lib and returns the resulting program as a string. Useful for saving -- the result to a file for off-line analysis, for instance if you have an SMT solver that's not natively -- supported out-of-the box by the SBV library. It takes two arguments: -- -- * version: The SMTLib-version to produce. Note that we currently only support SMTLib2. -- -- * isSat : If 'True', will translate it as a SAT query, i.e., in the positive. If 'False', will -- translate as a PROVE query, i.e., it will negate the result. (In this case, the check-sat -- call to the SMT solver will produce UNSAT if the input is a theorem, as usual.) compileToSMTLib :: Provable a => SMTLibVersion -- ^ Version of SMTLib to compile to. (Only SMTLib2 supported currently.) -> Bool -- ^ If True, translate directly, otherwise negate the goal. (Use True for SAT queries, False for PROVE queries.) -> a -> IO String compileToSMTLib version isSat a = do t <- getClockTime let comments = ["Created on " ++ show t] cvt = case version of SMTLib2 -> toSMTLib2 SMTProblem{smtLibPgm} <- simulate cvt defaultSMTCfg isSat comments a let out = show (smtLibPgm defaultSMTCfg NoCase) return $ out ++ "\n(check-sat)\n" -- | Create SMT-Lib benchmarks, for supported versions of SMTLib. The first argument is the basename of the file. -- The 'Bool' argument controls whether this is a SAT instance, i.e., translate the query -- directly, or a PROVE instance, i.e., translate the negated query. (See the second boolean argument to -- 'compileToSMTLib' for details.) generateSMTBenchmarks :: Provable a => Bool -> FilePath -> a -> IO () generateSMTBenchmarks isSat f a = mapM_ gen [minBound .. maxBound] where gen v = do s <- compileToSMTLib v isSat a let fn = f `addExtension` smtLibVersionExtension v writeFile fn s putStrLn $ "Generated " ++ show v ++ " benchmark " ++ show fn ++ "." -- | Make sure we're line-buffering if there's going to be parallel calls. bufferSanity :: Bool -> IO a -> IO a bufferSanity False a = a bufferSanity True a = bracket before after (const a) where before = do b <- hGetBuffering stdout hSetBuffering stdout LineBuffering return b after b = do hFlush stdout hSetBuffering stdout b hFlush stdout -- | Make sure sat/prove calls don't have objectives, and optimize does! objectiveCheck :: Bool -> [Objective a] -> String -> IO () objectiveCheck False [] _ = return () objectiveCheck False os w = error $ unlines $ ("\n*** Unsupported call to " ++ show w ++ " in the presence of objective(s):") : [ "***\t" ++ intercalate ", " (map objectiveName os) , "*** Use \"optimize\" to optimize for these objectives instead of " ++ show w ] objectiveCheck True [] w = error $ "*** Unsupported call to " ++ w ++ " when no objectives are present. Use \"sat\" for plain satisfaction" objectiveCheck True _ _ = return () -- | Pick the converter, based on the SMTLib version. Note that -- we no longer support SMTLib1, so the following is more or less a no-op, -- but it's good to use it since if we add some other target GHC's pattern-match -- warning will point us to here. getConverter :: SMTConfig -> SMTLibConverter getConverter SMTConfig{smtLibVersion} = case smtLibVersion of SMTLib2 -> toSMTLib2 -- | Proves the predicate using the given SMT-solver proveWith :: Provable a => SMTConfig -> a -> IO ThmResult proveWith config a = do simRes@SMTProblem{tactics, objectives} <- simulate (getConverter config) config False [] a objectiveCheck False objectives "prove" let hasPar = any isParallelCaseAnywhere tactics bufferSanity hasPar $ applyTactics config (False, hasPar) (wrap, unwrap) [] tactics objectives $ callSolver False "Checking Theoremhood.." [] mwrap simRes where wrap = ThmResult unwrap (ThmResult r) = r mwrap [r] = wrap r mwrap xs = error $ "SBV.proveWith: Backend solver returned a non-singleton answer:\n" ++ show (map ThmResult xs) -- | Find a satisfying assignment using the given SMT-solver satWith :: Provable a => SMTConfig -> a -> IO SatResult satWith config a = do simRes@SMTProblem{tactics, objectives} <- simulate (getConverter config) config True [] a objectiveCheck False objectives "sat" let hasPar = any isParallelCaseAnywhere tactics bufferSanity hasPar $ applyTactics config (True, hasPar) (wrap, unwrap) [] tactics objectives $ callSolver True "Checking Satisfiability.." [] mwrap simRes where wrap = SatResult unwrap (SatResult r) = r mwrap [r] = wrap r mwrap xs = error $ "SBV.satWith: Backend solver returned a non-singleton answer:\n" ++ show (map SatResult xs) -- | Optimizes the objectives using the given SMT-solver optimizeWith :: Provable a => SMTConfig -> a -> IO OptimizeResult optimizeWith config a = do msg "Optimizing.." sbvPgm@SMTProblem{objectives, tactics} <- simulate (getConverter config) config True [] a objectiveCheck True objectives "optimize" let hasPar = any isParallelCaseAnywhere tactics style = case nub [s | OptimizePriority s <- tactics] of [] -> Lexicographic [s] -> s ss -> error $ "SBV: Multiple optimization priorities found: " ++ intercalate ", " (map show ss) ++ ". Please use only one." optimizer = case style of Lexicographic -> optLexicographic Independent -> optIndependent Pareto -> optPareto optimizer hasPar config sbvPgm where msg = when (verbose config) . putStrLn . ("** " ++) -- | Construct a lexicographic optimization result optLexicographic :: Bool -> SMTConfig -> SMTProblem -> IO OptimizeResult optLexicographic hasPar config sbvPgm@SMTProblem{objectives, tactics} = do result <- bufferSanity hasPar $ applyTactics config (True, hasPar) (id, id) [] tactics objectives $ callSolver True "Lexicographically optimizing.." [] mwrap sbvPgm return $ LexicographicResult result where mwrap [r] = r mwrap xs = error $ "SBV.optLexicographic: Backend solver returned a non-singleton answer:\n" ++ show (map SatResult xs) -- | Construct an independent optimization result optIndependent :: Bool -> SMTConfig -> SMTProblem -> IO OptimizeResult optIndependent hasPar config sbvPgm@SMTProblem{objectives, tactics} = do let ns = map objectiveName objectives result <- bufferSanity hasPar $ applyTactics config (True, hasPar) (wrap ns, unwrap) [] tactics objectives $ callSolver True "Independently optimizing.." [] mwrap sbvPgm return $ IndependentResult result where wrap :: [String] -> SMTResult -> [(String, SMTResult)] wrap ns r = zip ns (repeat r) -- the role of unwrap here is to take the result with more info in case a case-split is -- performed and we need to decide in a SAT context. unwrap :: [(String, SMTResult)] -> SMTResult unwrap xs = case [r | (_, r@Satisfiable{}) <- xs] ++ [r | (_, r@SatExtField{}) <- xs] ++ map snd xs of (r:_) -> r [] -> error "SBV.optIndependent: Impossible happened: Received no results!" mwrap xs | lobs == lxs = zip (map objectiveName objectives) xs | True = error $ "SBV.optIndependent: Expected " ++ show lobs ++ " objective results, but received: " ++ show lxs ++ ":\n" ++ show (map SatResult xs) where lxs = length xs lobs = length objectives -- | Construct a pareto-front optimization result optPareto :: Bool -> SMTConfig -> SMTProblem -> IO OptimizeResult optPareto hasPar config sbvPgm@SMTProblem{objectives, tactics} = do result <- bufferSanity hasPar $ applyTactics config (True, hasPar) (wrap, unwrap) [] tactics objectives $ callSolver True "Pareto optimizing.." [] id sbvPgm return $ ParetoResult result where wrap :: SMTResult -> [SMTResult] wrap r = [r] -- the role of unwrap here is to take the result with more info in case a case-split is -- performed and we need to decide in a SAT context. unwrap :: [SMTResult] -> SMTResult unwrap xs = case [r | r@Satisfiable{} <- xs] ++ [r | r@SatExtField{} <- xs] ++ xs of (r:_) -> r [] -> error "SBV.optPareto: Impossible happened: Received no results!" -- | Apply the given tactics to a problem applyTactics :: SMTConfig -- ^ Solver configuration -> (Bool, Bool) -- ^ Are we a sat-problem? Do we have anything parallel going on? (Parallel-case split.) -> (SMTResult -> res, res -> SMTResult) -- ^ Wrapper/unwrapper pair from result to SMT answer -> [(String, (String, SW))] -- ^ Level at which we are called. (In case of a nested case-split) -> [Tactic SW] -- ^ Tactics active at this level -> [Objective (SW, SW)] -- ^ Optimization goals we have -> (SMTConfig -> Maybe (OptimizeStyle, Int) -> CaseCond -> IO res) -- ^ The actual continuation at this point -> IO res applyTactics cfgIn (isSat, hasPar) (wrap, unwrap) levels tactics objectives cont = do -- -- TODO: The management of tactics here is quite adhoc. We should have a better story -- Currently, we: -- -- - Perform optimization (which requires sat and no case-splitting) -- - Check for vacuity if asked -- - Do case-splitting -- -- If we have more interesting tactics, we'll have to come up with a better "proof manager." The current -- code is sufficient, however, for the use cases we have now. -- check that if we have objectives, then we must be sat and there must be no case-splits when (hasObjectives && not isSat) $ error "SBV: Optimization is only available for sat calls." when (hasObjectives && hasCaseSplits) $ error "SBV: Optimization and case-splits are not supported together." let mbOptInfo | not hasObjectives = Nothing | True = Just (optimizePriority, length objectives) if hasObjectives then cont (finalOptConfig objectives) mbOptInfo (Opt objectives) else do -- Check vacuity if asked. If result is Nothing, it means we're good to go. mbRes <- if not shouldCheckConstrVacuity then return Nothing else constraintVacuityCheck isSat finalConfig mbOptInfo (wrap, unwrap) cont -- Do case split, if vacuity said continue case mbRes of Just r -> return r Nothing -> if null caseSplits then cont finalConfig mbOptInfo (CasePath (map (snd . snd) levels)) else caseSplit finalConfig mbOptInfo shouldCheckCaseVacuity (parallelCase, hasPar) isSat (wrap, unwrap) levels chatty cases cont where (caseSplits, checkCaseVacuity, parallelCases, checkConstrVacuity, timeOuts, checkUsing, useLogics, useSolvers, optimizePriorities) = foldr (flip classifyTactics) ([], [], [], [], [], [], [], [], []) tactics classifyTactics (a, b, c, d, e, f, g, h, i) = \case t@CaseSplit{} -> (t:a, b, c, d, e, f, g, h, i) t@CheckCaseVacuity{} -> ( a, t:b, c, d, e, f, g, h, i) t@ParallelCase{} -> ( a, b, t:c, d, e, f, g, h, i) t@CheckConstrVacuity{} -> ( a, b, c, t:d, e, f, g, h, i) t@StopAfter{} -> ( a, b, c, d, t:e, f, g, h, i) t@CheckUsing{} -> ( a, b, c, d, e, t:f, g, h, i) t@UseLogic{} -> ( a, b, c, d, e, f, t:g, h, i) t@UseSolver{} -> ( a, b, c, d, e, f, g, t:h, i) t@OptimizePriority{} -> ( a, b, c, d, e, f, g, h, t:i) hasObjectives = not $ null objectives hasCaseSplits = not $ null cases parallelCase = not $ null parallelCases optimizePriority = case [s | OptimizePriority s <- optimizePriorities] of [] -> Lexicographic [s] -> s ss -> error $ "SBV.OptimizePriority: Multiple optimization priorities found, at most one is allowed: " ++ intercalate "," (map show ss) shouldCheckCaseVacuity = case [b | CheckCaseVacuity b <- checkCaseVacuity] of [] -> True -- default is to check-case-vacuity bs -> or bs -- otherwise check vacuity if we're asked to do so -- for constraint vacuity, default is *not* to check; so a simple or suffices shouldCheckConstrVacuity = or [b | CheckConstrVacuity b <- checkConstrVacuity] (chatty, cases) = let (vs, css) = unzip [(v, cs) | CaseSplit v cs <- caseSplits] in (or (verbose cfgIn : vs), concat css) grabStops c = case [i | StopAfter i <- timeOuts] of [] -> c xs -> c {timeOut = Just (maximum xs)} grabCheckUsing c = case [s | CheckUsing s <- checkUsing] of [] -> c [s] -> c {satCmd = "(check-sat-using " ++ s ++ ")"} ss -> c {satCmd = "(check-sat-using (then " ++ unwords ss ++ "))"} grabUseLogic c = case [l | UseLogic l <- useLogics] of [] -> c ss -> c { useLogic = Just (last ss) } configToUse = case [s | UseSolver s <- useSolvers] of [] -> cfgIn [s] -> s ss -> error $ "SBV.UseSolver: Multiple UseSolver tactics found, at most one is allowed: " ++ intercalate "," (map show ss) finalConfig = grabUseLogic . grabCheckUsing . grabStops $ configToUse finalOptConfig goals = finalConfig { optimizeArgs = optimizeArgs finalConfig ++ optimizerDirectives } where optimizerDirectives | hasObjectives = map minmax goals ++ style optimizePriority | True = [] minmax (Minimize _ (_, v)) = "(minimize " ++ show v ++ ")" minmax (Maximize _ (_, v)) = "(maximize " ++ show v ++ ")" minmax (AssertSoft nm (_, v) mbp) = "(assert-soft " ++ show v ++ penalize mbp ++ ")" where penalize DefaultPenalty = "" penalize (Penalty w mbGrp) | w <= 0 = error $ unlines [ "SBV.AssertSoft: Goal " ++ show nm ++ " is assigned a non-positive penalty: " ++ shw , "All soft goals must have > 0 penalties associated." ] | True = " :weight " ++ shw ++ maybe "" group mbGrp where shw = show (fromRational w :: Double) group g = " :id " ++ g style Lexicographic = [] -- default, no option needed style Independent = ["(set-option :opt.priority box)"] style Pareto = [ "(set-option :opt.priority pareto)" , "(set-option :opt.print_model true)" ] -- | Implements the "constraint vacuity check" tactic, making sure the calls to "constrain" -- describe a satisfiable condition. Returns: -- -- - Nothing if this is a SAT call, as that would be a weird thing to do (we only would care about constraint-vacuity in a proof context), -- - Nothing if satisfiable: The world is OK, just keep moving -- - ProofError if unsatisfiable. In this case we found that the constraints given are just bad! -- -- NB. We'll do a SAT call even if there are *no* constraints! This is OK, as the call will be cheap; and this is an opt-in call. (i.e., -- the user asked us to do it explicitly.) constraintVacuityCheck :: forall res. Bool -- ^ isSAT? -> SMTConfig -- ^ config -> Maybe (OptimizeStyle, Int) -- ^ optimization info -> (SMTResult -> res, res -> SMTResult) -- ^ wrappers back and forth from final result -> (SMTConfig -> Maybe (OptimizeStyle, Int) -> CaseCond -> IO res) -- ^ continuation -> IO (Maybe res) -- ^ result, wrapped in Maybe if vacuity fails constraintVacuityCheck True _ _ _ _ = return Nothing -- for a SAT check, vacuity is meaningless (what would be the point)? constraintVacuityCheck False config d (wrap, unwrap) f = do res <- f config d CstrVac case unwrap res of Satisfiable{} -> return Nothing _ -> return $ Just $ wrap vacuityFailResult where vacuityFailResult = ProofError config [ "Constraint vacuity check failed." , "User given constraints are not satisfiable." ] -- | Implements the case-split tactic. Works for both Sat and Proof, hence the quantification on @res@ caseSplit :: forall res. SMTConfig -- ^ Solver config -> Maybe (OptimizeStyle, Int) -- ^ Are we optimizing? -> Bool -- ^ Should we check vacuity of cases? -> (Bool, Bool) -- ^ Should we run the cases in parallel? Second bool: Is anything parallel going on? -> Bool -- ^ True if we're sat solving -> (SMTResult -> res, res -> SMTResult) -- ^ wrapper, unwrapper from sat/proof to the actual result -> [(String, (String, SW))] -- ^ Path condition as we reached here. (In a nested case split, First #, then actual name.) -> Bool -- ^ Should we be chatty on the case-splits? -> [(String, SW, [Tactic SW])] -- ^ List of cases. Case name, condition, plus further tactics for nested case-splitting etc. -> (SMTConfig -> Maybe (OptimizeStyle, Int) -> CaseCond -> IO res) -- ^ The "solver" once we provide it with a problem and a case -> IO res caseSplit config mbOptInfo checkVacuity (runParallel, hasPar) isSAT (wrap, unwrap) level chatty cases cont | runParallel = goParallel tasks | True = goSerial tasks where tasks = zip caseNos cases lids = map fst level noOfCases = length cases casePad = length (show noOfCases) tagLength = maximum $ map length $ "Coverage" : [s | (s, _, _) <- cases] showTag t = take tagLength (t ++ repeat ' ') shCaseId i = let si = show i in replicate (casePad - length si) ' ' ++ si caseNos = map shCaseId [(1::Int) .. ] tag tagChar = replicate 2 tagChar ++ replicate (2 * length level) tagChar mkCaseNameBase s i = "Case " ++ intercalate "." (lids ++ [i]) ++ ": " ++ showTag s mkCovNameBase = "Coverage " ++ replicate (casePad - 1) ' ' ++ "X" mkCaseName tagChar s i = tag tagChar ++ ' ' : mkCaseNameBase s i mkCovName tagChar = tag tagChar ++ ' ' : mkCovNameBase startCase :: Bool -> Maybe (String, String) -> IO () startCase multi mbis | not chatty = return () | Just (i, s) <- mbis = printer $ mkCaseName tagChar s i ++ start | True = printer $ mkCovName tagChar ++ start where line = multi || hasPar printer | line = putStrLn | True = putStr tagChar | line = '>' | True = '*' start = " [Started]" vacuityMsg :: Maybe Bool -> Bool -> (String, String) -> IO () vacuityMsg mbGood multi (i, s) | not chatty = return () | line = putStrLn $ mkCaseName '=' s i ++ msg | True = printer msg where line = multi || hasPar printer | failed = putStrLn | True = putStr (failed, msg) = case mbGood of Nothing -> (False, " [Vacuity Skipped]") Just True -> (False, " [Vacuity OK]") Just False -> (True, " [Vacuity Failed]") endCase :: Bool -> Maybe (String, String) -> String -> IO () endCase multi mbis msg | not chatty = return () | not line = putStrLn $ ' ' : msg | Just (i, s) <- mbis = putStrLn $ mkCaseName '<' s i ++ ' ' : msg | True = putStrLn $ mkCovName '<' ++ ' ' : msg where line = multi || hasPar ----------------------------------------------------------------------------------------------------------------- -- Serial case analysis ----------------------------------------------------------------------------------------------------------------- goSerial :: [(String, (String, SW, [Tactic SW]))] -> IO res goSerial [] -- At the end, we do a coverage call = do let multi = runParallel startCase multi Nothing res <- cont config mbOptInfo (CaseCov (map (snd . snd) level) [c | (_, c, _) <- cases]) decideSerial multi Nothing (unwrap res) (return res) goSerial ((i, (nm, cond, ts)):cs) -- Still going down, do a regular call = do let multi = not . null $ [() | CaseSplit{} <- ts] mbis = Just (i, nm) startCase multi mbis continue <- if isSAT -- for a SAT check, vacuity is meaningless (what would be the point)? then return True else if checkVacuity then do res <- cont config mbOptInfo (CaseVac (map (snd . snd) level) cond) case unwrap res of Satisfiable{} -> vacuityMsg (Just True) multi (i, nm) >> return True _ -> vacuityMsg (Just False) multi (i, nm) >> return False else vacuityMsg Nothing multi (i, nm) >> return True if continue then do res <- applyTactics config (isSAT, hasPar) (wrap, unwrap) (level ++ [(i, (nm, cond))]) ts [] cont decideSerial multi mbis (unwrap res) (goSerial cs) else return $ wrap $ vacuityFailResult (i, nm) vacuityFailResult cur = ProofError config $ [ "Vacuity check failed." , "Case constraint not satisfiable. Leading path:" ] ++ map (" " ++) (align ([(i, n) | (i, (n, _)) <- level] ++ [cur])) ++ ["HINT: Try \"CheckCaseVacuity False\" tactic to ignore case vacuity checks."] where align :: [(String, String)] -> [String] align path = map join cpath where len = maximum (0 : map (length . fst) cpath) join (c, n) = reverse (take len (reverse c ++ repeat ' ')) ++ ": " ++ n cpath = [(intercalate "." (reverse ls), j) | (ls, j) <- cascade [] path] trim = reverse . dropWhile isSpace . reverse . dropWhile isSpace cascade _ [] = [] cascade sofar ((i, j) : rest) = let new = trim i : sofar in (new, j) : cascade new rest decideSerial | isSAT = decideSerialSAT | True = decideSerialProof -- short name diag Unsatisfiable{} = "[Unsatisfiable]" diag Satisfiable {} = "[Satisfiable]" diag SatExtField {} = "[Satisfiable in Field Extension]" diag Unknown {} = "[Unknown]" diag ProofError {} = "[ProofError]" diag TimeOut {} = "[TimeOut]" -- If we're SAT, we stop at first satisfiable and report back. Otherwise continue. -- Note that we also stop if we get a ProofError, as that clearly is not OK decideSerialSAT :: Bool -> Maybe (String, String) -> SMTResult -> IO res -> IO res decideSerialSAT multi mbis r@Satisfiable{} _ = endCase multi mbis (diag r) >> return (wrap r) decideSerialSAT multi mbis r@ProofError{} _ = endCase multi mbis (diag r) >> return (wrap r) decideSerialSAT multi mbis r k = endCase multi mbis (diag r) >> k -- If we're Prove, we stop at first *not* unsatisfiable and report back. Otherwise continue. decideSerialProof :: Bool -> Maybe (String, String) -> SMTResult -> IO res -> IO res decideSerialProof multi mbis Unsatisfiable{} k = endCase multi mbis "[Proved]" >> k decideSerialProof multi mbis r _ = endCase multi mbis "[Failed]" >> return (wrap r) ----------------------------------------------------------------------------------------------------------------- -- Parallel case analysis ----------------------------------------------------------------------------------------------------------------- goParallel :: [(String, (String, SW, [Tactic SW]))] -> IO res goParallel cs = do when chatty $ putStrLn $ topName '>' "[Starting]" -- Create the case claim: let mkTask (i, (nm, cond, ts)) = let caseProof = do continue <- if isSAT -- for a SAT check, vacuity is meaningless (what would be the point)? then return True else if checkVacuity then do res <- cont config mbOptInfo (CaseVac (map (snd . snd) level) cond) case unwrap res of Satisfiable{} -> return True _ -> return False else return True if continue then unwrap `fmap` applyTactics config (isSAT, hasPar) (wrap, unwrap) (level ++ [(i, (nm, cond))]) ts [] cont else return $ vacuityFailResult (i, nm) in (mkCaseNameBase nm i, caseProof) -- Create the coverage claim let cov = unwrap `fmap` cont config mbOptInfo (CaseCov (map (snd . snd) level) [c | (_, c, _) <- cases]) (decidingTag, res) <- decideParallel $ map mkTask cs ++ [(mkCovNameBase, cov)] let trim = reverse . dropWhile isSpace . reverse . dropWhile isSpace let caseMsg | isSAT = satMsg | True = proofMsg where addTag x = x ++ " (at " ++ trim decidingTag ++ ")" (satMsg, proofMsg) = case res of Unsatisfiable{} -> ("[Unsatisfiable]", "[Proved]") Satisfiable{} -> (addTag "[Satisfiable]", addTag "[Failed]") _ -> let d = diag res in (addTag d, addTag d) when chatty $ putStrLn $ topName '<' caseMsg return $ wrap res where topName c w = tag c ++ topTag ++ " Parallel case split: " ++ range ++ ": " ++ w topTag = " Case" ++ s ++ intercalate "." lids ++ dot ++ "[1-" ++ show (length cs + 1) ++ "]:" where dot | null lids = "" | True = "." s | null cs = " " | True = "s " range = case cs of [] -> "Coverage" [_] -> "One case and coverage" xs -> show (length xs) ++ " cases and coverage" -- Parallel decision: -- - If SAT: Run all cases in parallel and return a SAT result from any. If none-of-them is SAT, then we return the last finishing -- - If Prove: Run all cases in parallel and return the last one if all return UNSAT. Otherwise return the first SAT one. decideParallel :: [(String, IO SMTResult)] -> IO (String, SMTResult) decideParallel caseTasks = mapM try caseTasks >>= pick where try (nm, task) = async $ task >>= \r -> return (nm, r) pick :: [Async (String, SMTResult)] -> IO (String, SMTResult) pick [] = error "SBV.caseSplit.decideParallel: Impossible happened, ran out of proofs!" pick [a] = wait a pick as = do (d, r) <- waitAny as let others = filter (/= d) as continue = pick others stop = mapM_ cancel others >> return r case snd r of Unsatisfiable{} -> continue Satisfiable{} -> stop SatExtField{} -> stop ProofError{} -> stop Unknown{} -> if isSAT then continue else stop TimeOut{} -> if isSAT then continue else stop -- | Check if any of the assertions can be violated safeWith :: SExecutable a => SMTConfig -> a -> IO [SafeResult] safeWith cfg a = do res@Result{resAssertions=asserts} <- runSymbolic (True, cfg) $ sName_ a >>= output mapM (verify res) asserts where locInfo (Just ps) = Just $ let loc (f, sl) = concat [srcLocFile sl, ":", show (srcLocStartLine sl), ":", show (srcLocStartCol sl), ":", f] in intercalate ",\n " (map loc ps) locInfo _ = Nothing verify res (msg, cs, cond) = do result <- runProofOn (getConverter cfg) cfg True [] pgm >>= \p -> callSolver True msg [] mwrap p cfg Nothing NoCase return $ SafeResult (locInfo (getCallStack `fmap` cs), msg, result) where pgm = res { resInputs = [(EX, n) | (_, n) <- resInputs res] -- make everything existential , resOutputs = [cond] } mwrap [r] = r mwrap xs = error $ "SBV.safeWith: Backend solver returned a non-singleton answer:\n" ++ show (map SatResult xs) -- | Check if a safe-call was safe or not, turning a 'SafeResult' to a Bool. isSafe :: SafeResult -> Bool isSafe (SafeResult (_, _, result)) = case result of Unsatisfiable{} -> True Satisfiable{} -> False SatExtField{} -> False -- conservative Unknown{} -> False -- conservative ProofError{} -> False -- conservative TimeOut{} -> False -- conservative -- | Determine if the constraints are vacuous using the given SMT-solver. Also see -- the 'CheckConstrVacuity' tactic. isVacuousWith :: Provable a => SMTConfig -> a -> IO Bool isVacuousWith config a = do Result ki tr uic is cs ts as uis ax asgn cstr tactics goals asserts _out <- runSymbolic (True, config) $ forAll_ a >>= output case cstr of [] -> return False -- no constraints, no need to check _ -> do let is' = [(EX, i) | (_, i) <- is] -- map all quantifiers to "exists" for the constraint check res' = Result ki tr uic is' cs ts as uis ax asgn cstr tactics goals asserts [trueSW] result <- runProofOn (getConverter config) config True [] res' >>= \p -> callSolver True "Checking Vacuity.." [] mwrap p config Nothing NoCase case result of Unsatisfiable{} -> return True -- constraints are unsatisfiable! Satisfiable{} -> return False -- constraints are satisfiable! SatExtField{} -> error "SBV: isVacuous: Solver returned a model in the extension field!" Unknown{} -> error "SBV: isVacuous: Solver returned unknown!" ProofError _ ls -> error $ "SBV: isVacuous: error encountered:\n" ++ unlines ls TimeOut _ -> error "SBV: isVacuous: time-out." where mwrap [r] = r mwrap xs = error $ "SBV.isVacuousWith: Backend solver returned a non-singleton answer:\n" ++ show (map SatResult xs) -- | Find all satisfying assignments using the given SMT-solver allSatWith :: Provable a => SMTConfig -> a -> IO AllSatResult allSatWith config p = do msg "Checking Satisfiability, all solutions.." sbvPgm@SMTProblem{smtInputs=qinps, kindsUsed=ki} <- simulate (getConverter config) config True [] p let usorts = [s | us@(KUserSort s _) <- Set.toList ki, isFree us] where isFree (KUserSort _ (Left _)) = True isFree _ = False unless (null usorts) $ msg $ "SBV.allSat: Uninterpreted sorts present: " ++ unwords usorts ++ "\n SBV will use equivalence classes to generate all-satisfying instances." results <- unsafeInterleaveIO $ go sbvPgm (1::Int) [] -- See if there are any existentials below any universals -- If such is the case, then the solutions are unique upto prefix existentials let w = ALL `elem` map fst qinps return $ AllSatResult (w, results) where msg = when (verbose config) . putStrLn . ("** " ++) go sbvPgm = loop where hasPar = any isParallelCaseAnywhere (tactics sbvPgm) loop !n nonEqConsts = do curResult <- invoke nonEqConsts hasPar n sbvPgm case curResult of Nothing -> return [] Just (SatResult r) -> let cont model = do let modelOnlyAssocs = [v | v@(x, _) <- modelAssocs model, not (isNonModelVar config x)] rest <- unsafeInterleaveIO $ loop (n+1) (modelOnlyAssocs : nonEqConsts) return (r : rest) in case r of -- We are done! This is really how we should always stop. Unsatisfiable{} -> return [] -- We have a model. If there are bindings, continue; otherwise stop Satisfiable _ (SMTModel _ []) -> return [r] Satisfiable _ model -> cont model -- Satisfied in an extension field. Stop if no new bindings, otherwise continue if all regular. -- If the model is in the extension, we also stop SatExtField _ (SMTModel _ []) -> return [r] SatExtField _ model@(SMTModel [] _) -> cont model SatExtField{} -> return [] -- Something bad happened, we stop here. Note that we treat Unknown as bad too in this context. Unknown{} -> return [r] ProofError{} -> return [r] TimeOut{} -> return [r] invoke nonEqConsts hasPar n simRes@SMTProblem{smtInputs, tactics, objectives} = do objectiveCheck False objectives "allSat" msg $ "Looking for solution " ++ show n case addNonEqConstraints (smtLibVersion config) (roundingMode config) smtInputs nonEqConsts of Nothing -> -- no new constraints refuted models, stop return Nothing Just refutedModels -> do let wrap = SatResult unwrap (SatResult r) = r mwrap [r] = wrap r mwrap xs = error $ "SBV.allSatWith: Backend solver returned a non-singleton answer:\n" ++ show (map SatResult xs) res <- bufferSanity hasPar $ applyTactics (updateName (n-1) config) (True, hasPar) (wrap, unwrap) [] tactics objectives $ callSolver True "Checking Satisfiability.." refutedModels mwrap simRes return $ Just res updateName i cfg = cfg{smtFile = upd `fmap` smtFile cfg} where upd nm = let (b, e) = splitExtension nm in b ++ "_allSat_" ++ show i ++ e callSolver :: Bool -> String -> [String] -> ([SMTResult] -> b) -> SMTProblem -> SMTConfig -> Maybe (OptimizeStyle, Int) -> CaseCond -> IO b callSolver isSat checkMsg refutedModels wrap SMTProblem{smtInputs, smtSkolemMap, smtLibPgm} config mbOptInfo caseCond = do let msg = when (verbose config) . putStrLn . ("** " ++) finalPgm = intercalate "\n" (pgm ++ refutedModels) where SMTLibPgm _ pgm = smtLibPgm config caseCond msg checkMsg msg $ "Generated SMTLib program:\n" ++ (finalPgm ++ intercalate "\n" ("" : optimizeArgs config ++ [satCmd config])) smtAnswer <- engine (solver config) config isSat mbOptInfo smtInputs smtSkolemMap finalPgm msg "Done.." return $ wrap smtAnswer simulate :: Provable a => SMTLibConverter -> SMTConfig -> Bool -> [String] -> a -> IO SMTProblem simulate converter config isSat comments predicate = do let msg = when (verbose config) . putStrLn . ("** " ++) isTiming = timing config msg "Starting symbolic simulation.." res <- timeIf isTiming ProblemConstruction $ runSymbolic (isSat, config) $ (if isSat then forSome_ else forAll_) predicate >>= output msg $ "Generated symbolic trace:\n" ++ show res msg "Translating to SMT-Lib.." runProofOn converter config isSat comments res runProofOn :: SMTLibConverter -> SMTConfig -> Bool -> [String] -> Result -> IO SMTProblem runProofOn converter config isSat comments res = let isTiming = timing config in case res of Result ki _qcInfo _codeSegs is consts tbls arrs uis axs pgm cstrs tacs goals assertions [o@(SW KBool _)] -> timeIf isTiming Translation $ let skolemMap = skolemize (if isSat then is else map flipQ is) where flipQ (ALL, x) = (EX, x) flipQ (EX, x) = (ALL, x) skolemize :: [(Quantifier, NamedSymVar)] -> [Either SW (SW, [SW])] skolemize qinps = go qinps ([], []) where go [] (_, sofar) = reverse sofar go ((ALL, (v, _)):rest) (us, sofar) = go rest (v:us, Left v : sofar) go ((EX, (v, _)):rest) (us, sofar) = go rest (us, Right (v, reverse us) : sofar) smtScript = converter ki isSat comments is skolemMap consts tbls arrs uis axs pgm cstrs o result = SMTProblem {smtInputs=is, smtSkolemMap=skolemMap, kindsUsed=ki, smtAsserts=assertions, tactics=tacs, objectives=goals, smtLibPgm=smtScript} in rnf smtScript `seq` return result Result{resOutputs = os} -> case length os of 0 -> error $ "Impossible happened, unexpected non-outputting result\n" ++ show res 1 -> error $ "Impossible happened, non-boolean output in " ++ show os ++ "\nDetected while generating the trace:\n" ++ show res _ -> error $ "User error: Multiple output values detected: " ++ show os ++ "\nDetected while generating the trace:\n" ++ show res ++ "\n*** Check calls to \"output\", they are typically not needed!" -- | Run an external proof on the given condition to see if it is satisfiable. internalSATCheck :: SMTConfig -> SBool -> State -> String -> IO SatResult internalSATCheck cfg condInPath st msg = do sw <- sbvToSW st condInPath () <- forceSWArg sw Result ki tr uic is cs ts as uis ax asgn cstr tactics goals assertions _ <- extractSymbolicSimulationState st let -- Construct the corresponding sat-checker for the branch. Note that we need to -- forget about the quantifiers and just use an "exist", as we're looking for a -- point-satisfiability check here; whatever the original program was. pgm = Result ki tr uic [(EX, n) | (_, n) <- is] cs ts as uis ax asgn cstr tactics goals assertions [sw] mwrap [r] = SatResult r mwrap xs = error $ "SBV.internalSATCheck: Backend solver returned a non-singleton answer:\n" ++ show (map SatResult xs) runProofOn (getConverter cfg) cfg True [] pgm >>= \p -> callSolver True msg [] mwrap p cfg Nothing NoCase {-# ANN module ("HLint: ignore Reduce duplication" :: String) #-}