Portability  portable 

Stability  experimental 
Maintainer  erkokl@gmail.com 
Safe Haskell  SafeInfered 
 Programming with symbolic values
 Uninterpreted constants and functions
 Properties, proofs, and satisfiability
 Optimization
 Computing expected values
 Model extraction
 SMT Interface: Configurations and solvers
 Symbolic computations
 Getting SMTLib output (for offline analysis)
 Test case generation
 Code generation from symbolic programs
 Module exports
(The sbv library is hosted at http://github.com/LeventErkok/sbv. Comments, bug reports, and patches are always welcome.)
SBV: Symbolic Bit Vectors in Haskell
Express properties about bitprecise Haskell programs and automatically prove them using SMT solvers.
>>>
prove $ \x > x `shiftL` 2 .== 4 * (x :: SWord8)
Q.E.D.
>>>
prove $ forAll ["x"] $ \x > x `shiftL` 2 .== (x :: SWord8)
Falsifiable. Counterexample: x = 51 :: SWord8
The function prove
has the following type:
prove
::Provable
a => a >IO
ThmResult
The class Provable
comes with instances for nary predicates, for arbitrary n.
The predicates are just regular Haskell functions over symbolic signed and unsigned
bitvectors. Functions for checking satisfiability (sat
and allSat
) are also
provided.
In particular, the sbv library introduces the types:

SBool
: Symbolic Booleans (bits) 
SWord8
,SWord16
,SWord32
,SWord64
: Symbolic Words (unsigned) 
SInt8
,SInt16
,SInt32
,SInt64
: Symbolic Ints (signed) 
SArray
,SFunArray
: Flat arrays of symbolic values  Symbolic polynomials over GF(2^n), polynomial arithmetic, and CRCs
 Uninterpreted constants and functions over symbolic values, with user defined SMTLib axioms
The user can construct ordinary Haskell programs using these types, which behave
very similar to their concrete counterparts. In particular these types belong to the
standard classes Num
, Bits
, custom versions of Eq
(EqSymbolic
)
and Ord
(OrdSymbolic
), along with several other custom classes for simplifying
bitprecise programming with symbolic values. The framework takes full advantage
of Haskell's type inference to avoid many common mistakes.
Furthermore, predicates (i.e., functions that return SBool
) built out of
these types can also be:
 proven correct via an external SMT solver (the
prove
function)  checked for satisfiability (the
sat
,allSat
functions)  used in synthesis (the
sat
function with existentials)  quickchecked
If a predicate is not valid, prove
will return a counterexample: An
assignment to inputs such that the predicate fails. The sat
function will
return a satisfying assignment, if there is one. The allSat
function returns
all satisfying assignments, lazily.
The sbv library uses thirdparty SMT solvers via the standard SMTLib interface: http://goedel.cs.uiowa.edu/smtlib/.
The SBV library is designed to work with any SMTLib compliant SMTsolver. Currently, we support the Yices SMT solver from SRI: http://yices.csl.sri.com/, and the Z3 SMT solver from Microsoft: http://research.microsoft.com/enus/um/redmond/projects/z3/.
You should download and install Yices on your machine, and make sure the
yices
executable is in your path before using the sbv library, as it is the
current default solver. Alternatively, you can specify the location of yices
executable in the environment variable SBV_YICES
and the options to yices
in SBV_YICES_OPTIONS
.
Use of quantified variables require an installation of z3. Again,
z3 must be in your path. Or, you can use the SBV_Z3
and SBV_Z3_OPTIONS
environment variables to set the executable and the options.
 type SBool = SBV Bool
 type SWord8 = SBV Word8
 type SWord16 = SBV Word16
 type SWord32 = SBV Word32
 type SWord64 = SBV Word64
 type SInt8 = SBV Int8
 type SInt16 = SBV Int16
 type SInt32 = SBV Int32
 type SInt64 = SBV Int64
 type SInteger = SBV Integer
 data SBV a
 class SymArray array where
 newArray_ :: (HasSignAndSize a, HasSignAndSize b) => Maybe (SBV b) > Symbolic (array a b)
 newArray :: (HasSignAndSize a, HasSignAndSize b) => String > Maybe (SBV b) > Symbolic (array a b)
 readArray :: array a b > SBV a > SBV b
 resetArray :: SymWord b => array a b > SBV b > array a b
 writeArray :: SymWord b => array a b > SBV a > SBV b > array a b
 mergeArrays :: SymWord b => SBV Bool > array a b > array a b > array a b
 data SArray a b
 data SFunArray a b
 mkSFunArray :: (SBV a > SBV b) > SFunArray a b
 type STree i e = STreeInternal (SBV i) (SBV e)
 readSTree :: (Bits i, SymWord i, SymWord e) => STree i e > SBV i > SBV e
 writeSTree :: (Mergeable (SBV e), Bits i, SymWord i, SymWord e) => STree i e > SBV i > SBV e > STree i e
 mkSTree :: forall i e. HasSignAndSize i => [SBV e] > STree i e
 sbvTestBit :: (Bits a, SymWord a) => SBV a > Int > SBool
 sbvPopCount :: (Bits a, SymWord a) => SBV a > SWord8
 setBitTo :: (Bits a, SymWord a) => SBV a > Int > SBool > SBV a
 oneIf :: (Num a, SymWord a) => SBool > SBV a
 lsb :: (Bits a, SymWord a) => SBV a > SBool
 msb :: (Bits a, SymWord a) => SBV a > SBool
 allEqual :: (Eq a, SymWord a) => [SBV a] > SBool
 allDifferent :: (Eq a, SymWord a) => [SBV a] > SBool
 blastBE :: (Bits a, SymWord a) => SBV a > [SBool]
 blastLE :: (Bits a, SymWord a) => SBV a > [SBool]
 class FromBits a where
 fromBitsLE, fromBitsBE :: [SBool] > a
 class Splittable a b  b > a where
 class SignCast a b  a > b, b > a where
 signCast :: a > b
 unsignCast :: b > a
 class Bits a => Polynomial a where
 crcBV :: Int > [SBool] > [SBool] > [SBool]
 crc :: (FromBits (SBV a), FromBits (SBV b), Bits a, Bits b, SymWord a, SymWord b) => Int > SBV a > SBV b > SBV b
 class Mergeable a where
 class EqSymbolic a where
 class (Mergeable a, EqSymbolic a) => OrdSymbolic a where
 class BVDivisible a where
 bvQuotRem :: a > a > (a, a)
 class Boolean b where
 bAnd :: Boolean b => [b] > b
 bOr :: Boolean b => [b] > b
 bAny :: Boolean b => (a > b) > [a] > b
 bAll :: Boolean b => (a > b) > [a] > b
 class PrettyNum a where
 readBin :: Num a => String > a
 class Uninterpreted a where
 uninterpret :: String > a
 uninterpretWithHandle :: String > (SBVUF, a)
 cgUninterpret :: String > [String] > a > a
 sbvUninterpret :: Maybe ([String], a) > String > (SBVUF, a)
 data SBVUF
 sbvUFName :: SBVUF > String
 addAxiom :: String > [String] > Symbolic ()
 type Predicate = Symbolic SBool
 class Provable a where
 class Equality a where
 prove :: Provable a => a > IO ThmResult
 proveWith :: Provable a => SMTConfig > a > IO ThmResult
 isTheorem :: Provable a => a > IO Bool
 isTheoremWithin :: Provable a => Int > a > IO (Maybe Bool)
 sat :: Provable a => a > IO SatResult
 satWith :: Provable a => SMTConfig > a > IO SatResult
 isSatisfiable :: Provable a => a > IO Bool
 isSatisfiableWithin :: Provable a => Int > a > IO (Maybe Bool)
 allSat :: Provable a => a > IO AllSatResult
 allSatWith :: Provable a => SMTConfig > a > IO AllSatResult
 numberOfModels :: Provable a => a > IO Int
 constrain :: SBool > Symbolic ()
 pConstrain :: Double > SBool > Symbolic ()
 isVacuous :: Provable a => a > IO Bool
 isVacuousWith :: Provable a => SMTConfig > a > IO Bool
 minimize :: (SatModel a, SymWord a, Show a, SymWord c, Show c) => OptimizeOpts > ([SBV a] > SBV c) > Int > ([SBV a] > SBool) > IO (Maybe [a])
 maximize :: (SatModel a, SymWord a, Show a, SymWord c, Show c) => OptimizeOpts > ([SBV a] > SBV c) > Int > ([SBV a] > SBool) > IO (Maybe [a])
 optimize :: (SatModel a, SymWord a, Show a, SymWord c, Show c) => OptimizeOpts > (SBV c > SBV c > SBool) > ([SBV a] > SBV c) > Int > ([SBV a] > SBool) > IO (Maybe [a])
 minimizeWith :: (SatModel a, SymWord a, Show a, SymWord c, Show c) => SMTConfig > OptimizeOpts > ([SBV a] > SBV c) > Int > ([SBV a] > SBool) > IO (Maybe [a])
 maximizeWith :: (SatModel a, SymWord a, Show a, SymWord c, Show c) => SMTConfig > OptimizeOpts > ([SBV a] > SBV c) > Int > ([SBV a] > SBool) > IO (Maybe [a])
 optimizeWith :: (SatModel a, SymWord a, Show a, SymWord c, Show c) => SMTConfig > OptimizeOpts > (SBV c > SBV c > SBool) > ([SBV a] > SBV c) > Int > ([SBV a] > SBool) > IO (Maybe [a])
 expectedValue :: Outputtable a => Symbolic a > IO [Double]
 expectedValueWith :: Outputtable a => Bool > Int > Maybe Int > Double > Symbolic a > IO [Double]
 newtype ThmResult = ThmResult SMTResult
 newtype SatResult = SatResult SMTResult
 newtype AllSatResult = AllSatResult (Bool, [SMTResult])
 data SMTResult
 = Unsatisfiable SMTConfig
  Satisfiable SMTConfig SMTModel
  Unknown SMTConfig SMTModel
  ProofError SMTConfig [String]
  TimeOut SMTConfig
 class SatModel a where
 class Modelable a where
 modelExists :: a > Bool
 getModel :: SatModel b => a > Either String (Bool, b)
 extractModel :: SatModel b => a > Maybe b
 displayModels :: SatModel a => (Int > (Bool, a) > IO ()) > AllSatResult > IO Int
 extractModels :: SatModel a => AllSatResult > [a]
 data SMTConfig = SMTConfig {}
 data OptimizeOpts
 = Iterative Bool
  Quantified
 data SMTSolver = SMTSolver {}
 yices :: SMTConfig
 z3 :: SMTConfig
 defaultSMTCfg :: SMTConfig
 sbvCheckSolverInstallation :: SMTConfig > IO Bool
 data Symbolic a
 output :: Outputtable a => a > Symbolic a
 class (HasSignAndSize a, Ord a) => SymWord a where
 forall :: String > Symbolic (SBV a)
 forall_ :: Symbolic (SBV a)
 mkForallVars :: Int > Symbolic [SBV a]
 exists :: String > Symbolic (SBV a)
 exists_ :: Symbolic (SBV a)
 mkExistVars :: Int > Symbolic [SBV a]
 free :: String > Symbolic (SBV a)
 free_ :: Symbolic (SBV a)
 mkFreeVars :: Int > Symbolic [SBV a]
 literal :: a > SBV a
 unliteral :: SBV a > Maybe a
 fromCW :: CW > a
 isConcrete :: SBV a > Bool
 isSymbolic :: SBV a > Bool
 isConcretely :: SBV a > (a > Bool) > Bool
 mbMaxBound, mbMinBound :: Maybe a
 compileToSMTLib :: Provable a => Bool > a > IO String
 generateSMTBenchmarks :: Provable a => FilePath > a > IO ()
 genTest :: Outputtable a => Int > Symbolic a > IO TestVectors
 getTestValues :: TestVectors > [([CW], [CW])]
 data TestVectors
 data TestStyle
 renderTest :: TestStyle > TestVectors > String
 data CW = CW {}
 newtype Size = Size {}
 cwToBool :: CW > Bool
 data SBVCodeGen a
 cgPerformRTCs :: Bool > SBVCodeGen ()
 cgSetDriverValues :: [Integer] > SBVCodeGen ()
 cgGenerateDriver :: Bool > SBVCodeGen ()
 cgGenerateMakefile :: Bool > SBVCodeGen ()
 cgInput :: SymWord a => String > SBVCodeGen (SBV a)
 cgInputArr :: SymWord a => Int > String > SBVCodeGen [SBV a]
 cgOutput :: SymWord a => String > SBV a > SBVCodeGen ()
 cgOutputArr :: SymWord a => String > [SBV a] > SBVCodeGen ()
 cgReturn :: SymWord a => SBV a > SBVCodeGen ()
 cgReturnArr :: SymWord a => [SBV a] > SBVCodeGen ()
 cgAddPrototype :: [String] > SBVCodeGen ()
 cgAddDecl :: [String] > SBVCodeGen ()
 cgAddLDFlags :: [String] > SBVCodeGen ()
 cgIntegerSize :: Int > SBVCodeGen ()
 compileToC :: Maybe FilePath > String > SBVCodeGen () > IO ()
 compileToCLib :: Maybe FilePath > String > [(String, SBVCodeGen ())] > IO ()
 module Data.Bits
 module Data.Word
 module Data.Int
Programming with symbolic values
The SBV library is really two things:
 A framework for writing bitprecise programs in Haskell
 A framework for proving properties of such programs using SMT solvers
In this first section we will look at the constructs that will let us construct such
programs in Haskell. The goal is to have a seamless experience, i.e., program in
the usual Haskell style without distractions of symbolic coding. While Haskell helps
in some aspects (the Num
and Bits
classes simplify coding), it makes life harder
in others. For instance, ifthenelse
only takes Bool
as a test in Haskell, and
comparisons (>
etc.) only return Bool
s. Clearly we would like these values to be
symbolic (i.e., SBool
), thus stopping us from using some native Haskell constructs.
When symbolic versions of operators are needed, they are typically obtained by prepending a dot,
for instance ==
becomes .==
. Care has been taken to make the transition painless. In
particular, any Haskell program you build out of symbolic components is fully concretely
executable within Haskell, without the need for any custom interpreters. (They are truly
Haskell programs, not AST's built out of pieces of syntax.) This provides for an integrated
feel of the system, one of the original design goals for SBV.
Symbolic types
Symbolic bit
Unsigned symbolic bitvectors
Signed symbolic bitvectors
Signed unbounded integers
The SBV library supports unbounded signed integers with the type SInteger
, which are not subject to
overflow/underflow as it is the case with the bounded types, such as SWord8
, SInt16
, etc. However,
some bitvector based operations are not supported for the SInteger
type while in the verification mode. That
is, you can use these operations on SInteger
values during normal programming/simulation.
but the SMT translation will not support these operations since there corresponding operations are not supported in SMTLib.
Note that this should rarely be a problem in practice, as these operations are mostly meaningful on fixedsize
bitvectors. The operations that are restricted to bounded word/int sizes are:
 Rotations and shifts:
rotateL
,rotateR
,shiftL
,shiftR
 Bitwise logical ops:
.&.
,..
,xor
,complement
 Extraction and concatenation:
split
, '#', andextend
(see theSplittable
class)
Usual arithmetic (+
, 
, *
, bvQuotRem
) and logical operations (.<
, .<=
, .>
, .>=
, .==
, ./=
) operations are
supported for SInteger
fully, both in programming and verification modes.
Abstract SBV type
The Symbolic value. Either a constant (Left
) or a symbolic
value (Right Cached
). Note that caching is essential for making
sure sharing is preserved. The parameter a
is phantom, but is
extremely important in keeping the user interface strongly typed.
Arrays of symbolic values
class SymArray array whereSource
Flat arrays of symbolic values
An array a b
is an array indexed by the type
, with elements of type SBV
a
If an initial value is not provided in SBV
bnewArray_
and newArray
methods, then the elements
are left unspecified, i.e., the solver is free to choose any value. This is the right thing
to do if arrays are used as inputs to functions to be verified, typically.
While it's certainly possible for user to create instances of SymArray
, the
SArray
and SFunArray
instances already provided should cover most use cases
in practice. (There are some differences between these models, however, see the corresponding
declaration.)
Minimal complete definition: All methods are required, no defaults.
newArray_ :: (HasSignAndSize a, HasSignAndSize b) => Maybe (SBV b) > Symbolic (array a b)Source
Create a new array, with an optional initial value
newArray :: (HasSignAndSize a, HasSignAndSize b) => String > Maybe (SBV b) > Symbolic (array a b)Source
Create a named new array, with an optional initial value
readArray :: array a b > SBV a > SBV bSource
Read the array element at a
resetArray :: SymWord b => array a b > SBV b > array a bSource
Reset all the elements of the array to the value b
writeArray :: SymWord b => array a b > SBV a > SBV b > array a bSource
Update the element at a
to be b
mergeArrays :: SymWord b => SBV Bool > array a b > array a b > array a bSource
Merge two given arrays on the symbolic condition
Intuitively: mergeArrays cond a b = if cond then a else b
.
Merging pushes the ifthenelse choice down on to elements
Arrays implemented in terms of SMTarrays: http://goedel.cs.uiowa.edu/smtlib/theories/ArraysEx.smt2
 Maps directly to SMTlib arrays
 Reading from an unintialized value is OK and yields an uninterpreted result
 Can check for equality of these arrays
 Cannot quickcheck theorems using
SArray
values  Typically slower as it heavily relies on SMTsolving for the array theory
SymArray SArray  
(HasSignAndSize a, HasSignAndSize b) => Show (SArray a b)  
SymWord b => Mergeable (SArray a b)  
EqSymbolic (SArray a b) 
Arrays implemented internally as functions
 Internally handled by the library and not mapped to SMTLib
 Reading an uninitialized value is considered an error (will throw exception)
 Cannot check for equality (internally represented as functions)
 Can quickcheck
 Typically faster as it gets compiled away during translation
mkSFunArray :: (SBV a > SBV b) > SFunArray a bSource
Lift a function to an array. Useful for creating arrays in a pure context. (Otherwise use newArray
.)
Full binary trees
type STree i e = STreeInternal (SBV i) (SBV e)Source
A symbolic tree containing values of type e, indexed by
elements of type i. Note that these are fulltrees, and their
their shapes remain constant. There is no API provided that
can change the shape of the tree. These structures are useful
when dealing with datastructures that are indexed with symbolic
values where access time is important. STree
structures provide
logarithmic time reads and writes.
readSTree :: (Bits i, SymWord i, SymWord e) => STree i e > SBV i > SBV eSource
Reading a value. We bitblast the index and descend down the full tree according to bitvalues.
writeSTree :: (Mergeable (SBV e), Bits i, SymWord i, SymWord e) => STree i e > SBV i > SBV e > STree i eSource
Writing a value, similar to how reads are done. The important thing is that the tree representation keeps updates to a minimum.
mkSTree :: forall i e. HasSignAndSize i => [SBV e] > STree i eSource
Construct the fully balanced initial tree using the given values
Operations on symbolic words
Word level
sbvPopCount :: (Bits a, SymWord a) => SBV a > SWord8Source
Replacement for popCount
. Since popCount
returns an Int
, we cannot implement
it for symbolic words. Here, we return an SWord8
, which can overflow when used on
quantities that have more than 255 bits. Currently, that's only the SInteger
type
that SBV supports, all other types are safe. Even with SInteger
, this will only
overflow if there are at least 256bits set in the number, and the smallest such
number is 2^2561, which is a pretty darn big number to worry about for practical
purposes. In any case, we do not support sbvPopCount
for unbounded symbolic integers,
as the only possible implementation wouldn't symbolically terminate. So the only overflow
issue is with reallyreally large concrete SInteger
values
lsb :: (Bits a, SymWord a) => SBV a > SBoolSource
Least significant bit of a word, always stored at index 0
msb :: (Bits a, SymWord a) => SBV a > SBoolSource
Most significant bit of a word, always stored at the last position
List level
allEqual :: (Eq a, SymWord a) => [SBV a] > SBoolSource
Returns (symbolic) true if all the elements of the given list are the same
allDifferent :: (Eq a, SymWord a) => [SBV a] > SBoolSource
Returns (symbolic) true if all the elements of the given list are different
Blasting/Unblasting
blastBE :: (Bits a, SymWord a) => SBV a > [SBool]Source
Bigendian blasting of a word into its bits. Also see the FromBits
class
blastLE :: (Bits a, SymWord a) => SBV a > [SBool]Source
Littleendian blasting of a word into its bits. Also see the FromBits
class
Unblasting a value from symbolicbits. The bits can be given littleendian or bigendian. For a signed number in littleendian, we assume the very last bit is the sign digit. This is a bit awkward, but it is more consistent with the reverse view of littlebigendian representations
Minimal complete definiton: fromBitsLE
fromBitsLE, fromBitsBE :: [SBool] > aSource
Splitting, joining, and extending
class Splittable a b  b > a whereSource
Splitting an a
into two b
's and joining back.
Intuitively, a
is a larger bitsize word than b
, typically double.
The extend
operation captures embedding of a b
value into an a
without changing its semantic value.
Minimal complete definition: All, no defaults.
Signcasting
class SignCast a b  a > b, b > a whereSource
Sign casting a value into another. This essentially
means forgetting the sign bit and reinterpreting the bits
accordingly when converting a signed value to an unsigned
one. Similarly, when an unsigned quantity is converted to
a signed one, the most significant bit is interpreted
as the sign. We only define instances when the source
and target types are precisely the same size.
The idea is that signCast
and unsignCast
must form
an isomorphism pair between the types a
and b
, i.e., we
expect the following two properties to hold:
signCast . unsignCast = id unsingCast . signCast = id
Note that one naive way to implement both these operations
is simply to compute fromBitsLE . blastLE
, i.e., first
get all the bits of the word and then reconstruct in the target
type. While this is semantically correct, it generates a lot
of code (both during proofs via SMTLib, and when compiled to C).
The goal of this class is to avoid that cost, so these operations
can be compiled very efficiently, they will essentially become noop's.
Minimal complete definition: All, no defaults.
Polynomial arithmetic and CRCs
class Bits a => Polynomial a whereSource
Implements polynomial addition, multiplication, division, and modulus operations
over GF(2^n). NB. Similar to bvQuotRem
, division by 0
is interpreted as follows:
x pDivMod
0 = (0, x)
for all x
(including 0
)
Minimal complete definiton: pMult
, pDivMod
, showPolynomial
polynomial :: [Int] > aSource
Given bitpositions to be set, create a polynomial For instance
polynomial [0, 1, 3] :: SWord8
will evaluate to 11
, since it sets the bits 0
, 1
, and 3
. Mathematicans would write this polynomial
as x^3 + x + 1
. And in fact, showPoly
will show it like that.
Add two polynomials in GF(2^n)
pMult :: (a, a, [Int]) > aSource
Multiply two polynomials in GF(2^n), and reduce it by the irreducible specified by the polynomial as specified by coefficients of the third argument. Note that the third argument is specifically left in this form as it is usally in GF(2^(n+1)), which is not available in our formalism. (That is, we would need SWord9 for SWord8 multiplication, etc.) Also note that we do not support symbolic irreducibles, which is a minor shortcoming. (Most GF's will come with fixed irreducibles, so this should not be a problem in practice.)
Passing [] for the third argument will multiply the polynomials and then ignore the higher bits that won't fit into the resulting size.
Divide two polynomials in GF(2^n), see above note for division by 0
Compute modulus of two polynomials in GF(2^n), see above note for modulus by 0
pDivMod :: a > a > (a, a)Source
Division and modulus packed together
Display a polynomial like a mathematician would (over the monomial x
), with a type
showPolynomial :: Bool > a > StringSource
Display a polynomial like a mathematician would (over the monomial x
), the first argument
controls if the final type is shown as well.
crcBV :: Int > [SBool] > [SBool] > [SBool]Source
Compute CRCs over bitvectors. The call crcBV n m p
computes
the CRC of the message m
with respect to polynomial p
. The
inputs are assumed to be blasted bigendian. The number
n
specifies how many bits of CRC is needed. Note that n
is actually the degree of the polynomial p
, and thus it seems
redundant to pass it in. However, in a typical proof context,
the polynomial can be symbolic, so we cannot compute the degree
easily. While this can be workedaround by generating code that
accounts for all possible degrees, the resulting code would
be unnecessarily big and complicated, and much harder to reason
with. (Also note that a CRC is just the remainder from the
polynomial division, but this routine is much faster in practice.)
NB. The n
th bit of the polynomial p
must be set for the CRC
to be computed correctly. Note that the polynomial argument p
will
not even have this bit present most of the time, as it will typically
contain bits 0
through n1
as usual in the CRC literature. The higher
order n
th bit is simply assumed to be set, as it does not make
sense to use a polynomial of a lesser degree. This is usually not a problem
since CRC polynomials are designed and expressed this way.
NB. The literature on CRC's has many variants on how CRC's are computed. We follow the painless guide (http://www.ross.net/crc/download/crc_v3.txt) and compute the CRC as follows:
 Extend the message
m
by addingn
0 bits on the right  Divide the polynomial thus obtained by the
p
 The remainder is the CRC value.
There are many variants on final XOR's, reversed polynomials etc., so it is essential to double check you use the correct algorithm.
crc :: (FromBits (SBV a), FromBits (SBV b), Bits a, Bits b, SymWord a, SymWord b) => Int > SBV a > SBV b > SBV bSource
Conditionals: Mergeable values
Symbolic choice operator, parameterized via a class
select
is a totalindexing function, with the default.
Minimal complete definition: symbolicMerge
symbolicMerge :: SBool > a > a > aSource
Merge two values based on the condition
ite :: SBool > a > a > aSource
Choose one or the other element, based on the condition.
This is similar to symbolicMerge
, but it has a default
implementation that makes sure it's shortcut if the condition is concrete
select :: (Bits b, SymWord b, Integral b) => [a] > a > SBV b > aSource
Total indexing operation. select xs default index
is intuitively
the same as xs !! index
, except it evaluates to default
if index
overflows
Mergeable ()  
Mergeable Mostek  
Mergeable Status  
Mergeable a => Mergeable [a]  
Mergeable a => Mergeable (Maybe a)  
SymWord a => Mergeable (SBV a)  
Mergeable a => Mergeable (Move a)  
Mergeable b => Mergeable (a > b)  
(Mergeable a, Mergeable b) => Mergeable (Either a b)  
(Mergeable a, Mergeable b) => Mergeable (a, b)  
(Ix a, Mergeable b) => Mergeable (Array a b)  
SymWord b => Mergeable (SFunArray a b)  
SymWord b => Mergeable (SArray a b)  
(SymWord e, Mergeable (SBV e)) => Mergeable (STree i e)  
(Mergeable a, Mergeable b, Mergeable c) => Mergeable (a, b, c)  
(Mergeable a, Mergeable b, Mergeable c, Mergeable d) => Mergeable (a, b, c, d)  
(Mergeable a, Mergeable b, Mergeable c, Mergeable d, Mergeable e) => Mergeable (a, b, c, d, e)  
(Mergeable a, Mergeable b, Mergeable c, Mergeable d, Mergeable e, Mergeable f) => Mergeable (a, b, c, d, e, f)  
(Mergeable a, Mergeable b, Mergeable c, Mergeable d, Mergeable e, Mergeable f, Mergeable g) => Mergeable (a, b, c, d, e, f, g) 
Symbolic equality
class EqSymbolic a whereSource
Symbolic Equality. Note that we can't use Haskell's Eq
class since Haskell insists on returning Bool
Comparing symbolic values will necessarily return a symbolic value.
Minimal complete definition: .==
EqSymbolic Bool  
EqSymbolic a => EqSymbolic [a]  
EqSymbolic a => EqSymbolic (Maybe a)  
EqSymbolic (SBV a)  
(EqSymbolic a, EqSymbolic b) => EqSymbolic (Either a b)  
(EqSymbolic a, EqSymbolic b) => EqSymbolic (a, b)  
EqSymbolic (SArray a b)  
(EqSymbolic a, EqSymbolic b, EqSymbolic c) => EqSymbolic (a, b, c)  
(EqSymbolic a, EqSymbolic b, EqSymbolic c, EqSymbolic d) => EqSymbolic (a, b, c, d)  
(EqSymbolic a, EqSymbolic b, EqSymbolic c, EqSymbolic d, EqSymbolic e) => EqSymbolic (a, b, c, d, e)  
(EqSymbolic a, EqSymbolic b, EqSymbolic c, EqSymbolic d, EqSymbolic e, EqSymbolic f) => EqSymbolic (a, b, c, d, e, f)  
(EqSymbolic a, EqSymbolic b, EqSymbolic c, EqSymbolic d, EqSymbolic e, EqSymbolic f, EqSymbolic g) => EqSymbolic (a, b, c, d, e, f, g) 
Symbolic ordering
class (Mergeable a, EqSymbolic a) => OrdSymbolic a whereSource
Symbolic Comparisons. Similar to Eq
, we cannot implement Haskell's Ord
class
since there is no way to return an Ordering
value from a symbolic comparison.
Furthermore, OrdSymbolic
requires Mergeable
to implement ifthenelse, for the
benefit of implementing symbolic versions of max
and min
functions.
Minimal complete definition: .<
OrdSymbolic a => OrdSymbolic [a]  
OrdSymbolic a => OrdSymbolic (Maybe a)  
SymWord a => OrdSymbolic (SBV a)  
(OrdSymbolic a, OrdSymbolic b) => OrdSymbolic (Either a b)  
(OrdSymbolic a, OrdSymbolic b) => OrdSymbolic (a, b)  
(OrdSymbolic a, OrdSymbolic b, OrdSymbolic c) => OrdSymbolic (a, b, c)  
(OrdSymbolic a, OrdSymbolic b, OrdSymbolic c, OrdSymbolic d) => OrdSymbolic (a, b, c, d)  
(OrdSymbolic a, OrdSymbolic b, OrdSymbolic c, OrdSymbolic d, OrdSymbolic e) => OrdSymbolic (a, b, c, d, e)  
(OrdSymbolic a, OrdSymbolic b, OrdSymbolic c, OrdSymbolic d, OrdSymbolic e, OrdSymbolic f) => OrdSymbolic (a, b, c, d, e, f)  
(OrdSymbolic a, OrdSymbolic b, OrdSymbolic c, OrdSymbolic d, OrdSymbolic e, OrdSymbolic f, OrdSymbolic g) => OrdSymbolic (a, b, c, d, e, f, g) 
Division
class BVDivisible a whereSource
The BVDivisible
class captures the essence of division of words.
Unfortunately we cannot use Haskell's Integral
class since the Real
and Enum
superclasses are not implementable for symbolic bitvectors.
However, quotRem
makes perfect sense, and the BVDivisible
class captures
this operation. One issue is how division by 0 behaves. The verification
technology requires total functions, and there are several design choices
here. We follow Isabelle/HOL approach of assigning the value 0 for division
by 0. Therefore, we impose the following law:
x bvQuotRem
0 = (0, x)
Note that our instances implement this law even when x
is 0
itself.
Minimal complete definition: bvQuotRem
The Boolean class
The Boolean
class: a generalization of Haskell's Bool
type
Haskell Bool
and SBV's SBool
are instances of this class, unifying the treatment of boolean values.
Minimal complete definition: true
, bnot
, &&&
However, it's advisable to define false
, and 
as well (typically), for clarity.
logical true
logical false
complement
and
or
nand
nor
xor
implies
equivalence
cast from Bool
Generalizations of boolean operations
Prettyprinting and reading numbers in Hex & Binary
PrettyNum class captures printing of numbers in hex and binary formats; also supporting negative numbers.
Show a number in hexadecimal (starting with 0x
and type.)
Show a number in binary (starting with 0b
and type.)
Show a number in hex, without prefix, or types.
Show a number in bin, without prefix, or types.
readBin :: Num a => String > aSource
A more convenient interface for reading binary numbers, also supports negative numbers
Uninterpreted constants and functions
class Uninterpreted a whereSource
Uninterpreted constants and functions. An uninterpreted constant is a value that is indexed by its name. The only property the prover assumes about these values are that they are equivalent to themselves; i.e., (for functions) they return the same results when applied to same arguments. We support uninterpretedfunctions as a general means of blackbox'ing operations that are irrelevant for the purposes of the proof; i.e., when the proofs can be performed without any knowledge about the function itself.
Minimal complete definition: uninterpretWithHandle
. However, most instances in
practice are already provided by SBV, so endusers should not need to define their
own instances.
uninterpret :: String > aSource
Uninterpret a value, receiving an object that can be used instead. Use this version when you do not need to add an axiom about this value.
uninterpretWithHandle :: String > (SBVUF, a)Source
Uninterpret a value, but also get a handle to the resulting object. This handle
can be used to add axioms for this object. (See addAxiom
.)
cgUninterpret :: String > [String] > a > aSource
Uninterpret a value, only for the purposes of codegeneration. For execution and verification the value is used as is. For codegeneration, the alternate definition is used. This is useful when we want to take advantage of native libraries on the target languages.
sbvUninterpret :: Maybe ([String], a) > String > (SBVUF, a)Source
Most generalized form of uninterpretation, this function should not be needed by endusercode, but is rather useful for the library development.
Accessing the handle
An uninterpreted function handle. This is the handle to be used for adding axioms about uninterpreted constants/functions. Note that we will leave this abstract for safety purposes
sbvUFName :: SBVUF > StringSource
Get the name associated with the uninterpretedvalue; useful when constructing axioms about this UI.
Adding axioms
addAxiom :: String > [String] > Symbolic ()Source
Add a user specified axiom to the generated SMTLib file. Note that the input is a mere string; we perform no checking on the input that it's wellformed or is sensical. A separate formalization of SMTLib would be very useful here.
Properties, proofs, and satisfiability
The SBV library provides a pushbutton verification system via automated SMT solving. The design goal is to let SMT solvers be used without any knowledge of how SMT solvers work or how different logics operate. The details are hidden behind the SBV framework, providing Haskell programmers with a clean API that is unencumbered by the details of individual solvers. To that end, we use the SMTLib standard (http://goedel.cs.uiowa.edu/smtlib/) to communicate with arbitrary SMT solvers. Unfortunately, the SMTLib version 1.X does not standardize how models are communicated back from solvers, so there is some work in parsing individual SMT solver output. The 2.X version of the SMTLib standard (not yet implemented by SMT solvers widely, unfortunately) will bring new standard features for getting models; at which time the SBV framework can be modified into a truly plugandplay system where arbitrary SMT solvers can be used.
Predicates
type Predicate = Symbolic SBoolSource
A predicate is a symbolic program that returns a (symbolic) boolean value. For all intents and
purposes, it can be treated as an nary function from symbolicvalues 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.
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 upto a 7tuple of symbolictypes. So
predicates can be constructed from almost arbitrary Haskell functions that have arbitrary
shapes. (See the instance declarations below.)
forAll_ :: a > PredicateSource
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 :: [String] > a > PredicateSource
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 counterexamples when they are printed.
forSome_ :: a > PredicateSource
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 :: [String] > a > PredicateSource
Version of forSome
that allows user defined names
Provable SBool  
Provable Predicate  
(SymWord a, SymWord b, Provable p) => Provable ((SBV a, SBV b) > p)  
(SymWord a, SymWord b, SymWord c, Provable p) => Provable ((SBV a, SBV b, SBV c) > p)  
(SymWord a, SymWord b, SymWord c, SymWord d, Provable p) => Provable ((SBV a, SBV b, SBV c, SBV d) > p)  
(SymWord a, SymWord b, SymWord c, SymWord d, SymWord e, Provable p) => Provable ((SBV a, SBV b, SBV c, SBV d, SBV e) > p)  
(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)  
(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)  
(HasSignAndSize a, HasSignAndSize b, SymArray array, Provable p) => Provable (array a b > p)  
(SymWord a, Provable p) => Provable (SBV a > p) 
Equality as a proof method. Allows for very concise construction of equivalence proofs, which is very typical in bitprecise proofs.
Proving properties
prove :: Provable a => a > IO ThmResultSource
Prove a predicate, equivalent to proveWith
defaultSMTCfg
proveWith :: Provable a => SMTConfig > a > IO ThmResultSource
Proves the predicate using the given SMTsolver
isTheoremWithin :: Provable a => Int > a > IO (Maybe Bool)Source
Checks theoremhood within the given time limit of i
seconds.
Returns Nothing
if times out, or the result wrapped in a Just
otherwise.
Checking satisfiability
sat :: Provable a => a > IO SatResultSource
Find a satisfying assignment for a predicate, equivalent to satWith
defaultSMTCfg
satWith :: Provable a => SMTConfig > a > IO SatResultSource
Find a satisfying assignment using the given SMTsolver
isSatisfiable :: Provable a => a > IO BoolSource
Checks satisfiability
isSatisfiableWithin :: Provable a => Int > a > IO (Maybe Bool)Source
Checks satisfiability within the given time limit of i
seconds.
Returns Nothing
if times out, or the result wrapped in a Just
otherwise.
Finding all satisfying assignments
allSat :: Provable a => a > IO AllSatResultSource
Return all satisfying assignments for a predicate, equivalent to
.
Satisfying assignments are constructed lazily, so they will be available as returned by the solver
and on demand.
allSatWith
defaultSMTCfg
NB. Uninterpreted constant/function values and counterexamples for array values are ignored for
the purposes of
. 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 counterexample back from the SMT solver.
allSat
allSatWith :: Provable a => SMTConfig > a > IO AllSatResultSource
Find all satisfying assignments using the given SMTsolver
numberOfModels :: Provable a => a > IO IntSource
Returns the number of models that satisfy the predicate, as it would
be returned by allSat
. Note that the number of models is always a
finite number, and hence this will always return a result. Of course,
computing it might take quite long, as it literally generates and counts
the number of satisfying models.
Adding constraints
A constraint is a means for restricting the input domain of a formula. Here's a simple example:
do x <exists
"x" y <exists
"y"constrain
$ x .> yconstrain
$ x + y .>= 12constrain
$ y .>= 3 ...
The first constraint requires x
to be larger than y
. The scond one says that
sum of x
and y
must be at least 12
, and the final one says that y
to be at least 3
.
Constraints provide an easy way to assert additional properties on the input domain, right at the point of
the introduction of variables.
Note that the proper reading of a constraint depends on the context:
 In a
sat
(orallSat
) call: The constraint added is asserted conjunctively. That is, the resulting satisfying model (if any) will always satisfy all the constraints given.  In a
prove
call: In this case, the constraint acts as an implication. The property is proved under the assumption that the constraint holds. In other words, the constraint says that we only care about the input space that satisfies the constraint.  In a
quickCheck
call: The constraint acts as a filter forquickCheck
; if the constraint does not hold, then the input value is considered to be irrelevant and is skipped. Note that this is similar toprove
, but is stronger: We do not accept a test case to be valid just because the constraints fail on them, although semantically the implication does hold. We simply skip that test case as a bad test vector.  In a
genTest
call: Similar toquickCheck
andprove
: If a constraint does not hold, the input value is ignored and is not included in the test set.
A good use case (in fact the motivating use case) for constrain
is attaching a
constraint to a forall
or exists
variable at the time of its creation.
Also, the conjunctive semantics for sat
and the implicative
semantics for prove
simplify programming by choosing the correct interpretation
automatically. However, one should be aware of the semantic difference. For instance, in
the presence of constraints, formulas that are provable are not necessarily
satisfiable. To wit, consider:
do x <exists
"x"constrain
$ x .< x return $ x .< (x ::SWord8
)
This predicate is unsatisfiable since no element of SWord8
is less than itself. But
it's (vacuously) true, since it excludes the entire domain of values, thus making the proof
trivial. Hence, this predicate is provable, but is not satisfiable. To make sure the given
constraints are not vacuous, the functions isVacuous
(and isVacuousWith
) can be used.
Also note that this semantics imply that test case generation (genTest
) and quickcheck
can take arbitrarily long in the presence of constraints, if the random input values generated
rarely satisfy the constraints. (As an extreme case, consider
.)
constrain
false
A probabilistic constraint (see pConstrain
) attaches a probability threshold for the
constraint to be considered. For instance:
pConstrain
0.8 c
will make sure that the condition c
is satisfied 80% of the time (and correspondingly, falsified 20%
of the time), in expectation. This variant is useful for genTest
and quickCheck
functions, where we
want to filter the test cases according to some probability distribution, to make sure that the testvectors
are drawn from interesting subsets of the input space. For instance, if we were to generate 100 test cases
with the above constraint, we'd expect about 80 of them to satisfy the condition c
, while about 20 of them
will fail it.
The following properties hold:
constrain
=pConstrain
1pConstrain
t c =pConstrain
(1t) (not c)
Note that while constrain
can be used freely, pConstrain
is only allowed in the contexts of
genTest
or quickCheck
. Calls to pConstrain
in a prove/sat call will be rejected as SBV does not
deal with probabilistic constraints when it comes to satisfiability and proofs.
Also, both constrain
and pConstrain
calls during codegeneration will also be rejected, for similar reasons.
pConstrain :: Double > SBool > Symbolic ()Source
Adding a probabilistic constraint. The Double
argument is the probability
threshold. Probabilistic constraints are useful for genTest
and quickCheck
calls where we restrict our attention to interesting parts of the input domain.
Checking constraint vacuity
isVacuous :: Provable a => a > IO BoolSource
Check if the given constraints are satisfiable, equivalent to
. This
call can be used to ensure that the specified constraints (via isVacuousWith
defaultSMTCfg
constrain
) are satisfiable, i.e., that
the proof involving these constraints is not passing vacuously. Here is an example. Consider the following
predicate:
>>>
let pred = do { x < forall "x"; constrain $ x .< x; return $ x .>= (5 :: SWord8) }
This predicate asserts that all 8bit values are larger than 5, subject to the constraint that the
values considered satisfy x .< x
, i.e., they are less than themselves. Since there are no values that
satisfy this constraint, the proof will pass vacuously:
>>>
prove pred
Q.E.D.
We can use isVacuous
to make sure to see that the pass was vacuous:
>>>
isVacuous pred
True
While the above example is trivial, things can get complicated if there are multiple constraints with nonstraightforward relations; so if constraints are used one should make sure to check the predicate is not vacuously true. Here's an example that is not vacuous:
>>>
let pred' = do { x < forall "x"; constrain $ x .> 6; return $ x .>= (5 :: SWord8) }
This time the proof passes as expected:
>>>
prove pred'
Q.E.D.
And the proof is not vacuous:
>>>
isVacuous pred'
False
isVacuousWith :: Provable a => SMTConfig > a > IO BoolSource
Determine if the constraints are vacuous using the given SMTsolver
Optimization
Symbolic optimization. A call of the form:
minimize Quantified cost n valid
returns Just xs
, such that:

xs
has preciselyn
elements 
valid xs
holds 
cost xs
is minimal. That is, for all sequencesys
that satisfy the first two criteria above,cost xs .<= cost ys
holds.
If there is no such sequence, then minimize
will return Nothing
.
The function maximize
is similar, except the comparator is .>=
. So the value returned has the largest cost (or value, in that case).
The function optimize
allows the user to give a custom comparison function.
The OptimizeOpts
argument controls how the optimization is done. If Quantified
is used, then the SBV optimization engine satisfies the following predicate:
exists xs. forall ys. valid xs && (valid ys `implies
` (cost xs `cmp
` cost ys))
Note that this may cause efficiency problems as it involves alternating quantifiers.
If OptimizeOpts
is set to Iterative
True
, then SBV will programmatically
search for an optimal solution, by repeatedly calling the solver appropriately. (The boolean argument controls whether progress reports are given. Use
False
for quiet operation.) Note that the quantified and iterative versions are two different optimization approaches and may not necessarily yield the same
results. In particular, the quantified version can find solutions where there is no global optimum value, while the iterative version would simply loop forever
in such cases. On the other hand, the iterative version might be more suitable if the quantified version of the problem is too hard to deal with by the SMT solver.
minimize :: (SatModel a, SymWord a, Show a, SymWord c, Show c) => OptimizeOpts > ([SBV a] > SBV c) > Int > ([SBV a] > SBool) > IO (Maybe [a])Source
Minimizes a cost function with respect to a constraint. Examples:
>>>
minimize Quantified sum 3 (bAll (.> (10 :: SInteger)))
Just [11,11,11]
maximize :: (SatModel a, SymWord a, Show a, SymWord c, Show c) => OptimizeOpts > ([SBV a] > SBV c) > Int > ([SBV a] > SBool) > IO (Maybe [a])Source
Maximizes a cost function with respect to a constraint. Examples:
>>>
maximize Quantified sum 3 (bAll (.< (10 :: SInteger)))
Just [9,9,9]
optimize :: (SatModel a, SymWord a, Show a, SymWord c, Show c) => OptimizeOpts > (SBV c > SBV c > SBool) > ([SBV a] > SBV c) > Int > ([SBV a] > SBool) > IO (Maybe [a])Source
Variant of optimizeWith
using z3
minimizeWith :: (SatModel a, SymWord a, Show a, SymWord c, Show c) => SMTConfig > OptimizeOpts > ([SBV a] > SBV c) > Int > ([SBV a] > SBool) > IO (Maybe [a])Source
Variant of minimize
allowing the use of a user specified solver.
maximizeWith :: (SatModel a, SymWord a, Show a, SymWord c, Show c) => SMTConfig > OptimizeOpts > ([SBV a] > SBV c) > Int > ([SBV a] > SBool) > IO (Maybe [a])Source
Variant of maximize
allowing the use of a user specified solver.
Computing expected values
expectedValue :: Outputtable a => Symbolic a > IO [Double]Source
Given a symbolic computation that produces a value, compute the
expected value that value would take if this computation is run
with its free variables drawn from uniform distributions of its
respective values, satisfying the given constraints specified by
constrain
and pConstrain
calls. This is equivalent to calling
expectedValueWith
the following parameters: verbose, warmup
round count of 10000
, no maximum iteration count, and with
convergence margin 0.0001
.
expectedValueWith :: Outputtable a => Bool > Int > Maybe Int > Double > Symbolic a > IO [Double]Source
Generalized version of expectedValue
, allowing the user to specify the
warmup count and the convergence factor. Maximum iteration count can also
be specified, at which point convergence won't be sought. The boolean controls verbosity.
Model extraction
The default Show
instances for prover calls provide all the counterexample information in a
humanreadable form and should be sufficient for most casual uses of sbv. However, tools built
on top of sbv will inevitably need to look into the constructed models more deeply, programmatically
extracting their results and performing actions based on them. The API provided in this section
aims at simplifying this task.
Inspecting proof results
ThmResult
, SatResult
, and AllSatResult
are simple newtype wrappers over SMTResult
. Their
main purpose is so that we can provide custom Show
instances to print results accordingly.
A prove
call results in a ThmResult
newtype AllSatResult Source
An allSat
call results in a AllSatResult
. The boolean says whether
we should warn the user about prefixexistentials.
AllSatResult (Bool, [SMTResult]) 
The result of an SMT solver call. Each constructor is tagged with
the SMTConfig
that created it so that further tools can inspect it
and build layers of results, if needed. For ordinary uses of the library,
this type should not be needed, instead use the accessor functions on
it. (Custom Show instances and model extractors.)
Unsatisfiable SMTConfig  Unsatisfiable 
Satisfiable SMTConfig SMTModel  Satisfiable with model 
Unknown SMTConfig SMTModel  Prover returned unknown, with a potential (possibly bogus) model 
ProofError SMTConfig [String]  Prover errored out 
TimeOut SMTConfig  Computation timed out (see the 
Programmable model extraction
While default Show
instances are sufficient for most use cases, it is sometimes desirable (especially
for library construction) that the SMTmodels are reinterpreted in terms of domain types. Programmable
extraction allows getting arbitrarily typed models out of SMT models.
Instances of SatModel
can be automatically extracted from models returned by the
solvers. The idea is that the sbv infrastructure provides a stream of CW'
s (constantwords)
coming from the solver, and the type a
is interpreted based on these constants. Many typical
instances are already provided, so new instances can be declared with relative ease.
Minimum complete definition: parseCWs
parseCWs :: [CW] > Maybe (a, [CW])Source
Given a sequence of constantwords, extract one instance of the type a
, returning
the remaining elements untouched. If the next element is not what's expected for this
type you should return Nothing
cvtModel :: (a > Maybe b) > Maybe (a, [CW]) > Maybe (b, [CW])Source
Given a parsed model instance, transform it using f
, and return the result.
The default definition for this method should be sufficient in most use cases.
SatModel Bool  
SatModel Int8  
SatModel Int16  
SatModel Int32  
SatModel Int64  
SatModel Integer  
SatModel Word8  
SatModel Word16  
SatModel Word32  
SatModel Word64  
SatModel ()  
SatModel U2Member  
SatModel a => SatModel [a]  
(SatModel a, SatModel b) => SatModel (a, b)  
(SatModel a, SatModel b, SatModel c) => SatModel (a, b, c)  
(SatModel a, SatModel b, SatModel c, SatModel d) => SatModel (a, b, c, d)  
(SatModel a, SatModel b, SatModel c, SatModel d, SatModel e) => SatModel (a, b, c, d, e)  
(SatModel a, SatModel b, SatModel c, SatModel d, SatModel e, SatModel f) => SatModel (a, b, c, d, e, f)  
(SatModel a, SatModel b, SatModel c, SatModel d, SatModel e, SatModel f, SatModel g) => SatModel (a, b, c, d, e, f, g) 
Various SMT results that we can extract models out of.
modelExists :: a > BoolSource
Is there a model?
getModel :: SatModel b => a > Either String (Bool, b)Source
Extract a model, the result is a tuple where the first argument (if True) indicates whether the model was probable. (i.e., if the solver returned unknown.)
extractModel :: SatModel b => a > Maybe bSource
A simpler variant of getModel
to get a model out without the fuss.
displayModels :: SatModel a => (Int > (Bool, a) > IO ()) > AllSatResult > IO IntSource
Given an allSat
call, we typically want to iterate over it and print the results in sequence. The
displayModels
function automates this task by calling disp
on each result, consecutively. The first
Int
argument to disp
'is the current model number. The second argument is a tuple, where the first
element indicates whether the model is alleged (i.e., if the solver is not sure, returing Unknown)
extractModels :: SatModel a => AllSatResult > [a]Source
Return all the models from an allSat
call, similar to extractModel
but
is suitable for the case of multiple results.
SMT Interface: Configurations and solvers
Solver configuration
SMTConfig  

data OptimizeOpts Source
Optimizer configuration. Note that iterative and quantified approaches are in general not interchangeable. For instance, iterative solutions will loop infinitely when there is no optimal value, but quantified solutions can handle such problems. Of course, quantified problems are harder for SMT solvers, naturally.
Iterative Bool  Iteratively search. if True, it will be reporting progress 
Quantified  Use quantifiers 
An SMT solver
defaultSMTCfg :: SMTConfigSource
The default solver used by SBV. This is currently set to z3.
sbvCheckSolverInstallation :: SMTConfig > IO BoolSource
Check whether the given solver is installed and is ready to go. This call does a simple call to the solver to ensure all is well.
Symbolic computations
A Symbolic computation. Represented by a reader monad carrying the state of the computation, layered on top of IO for creating unique references to hold onto intermediate results.
class (HasSignAndSize a, Ord a) => SymWord a whereSource
A SymWord
is a potential symbolic bitvector that can be created instances of
to be fed to a symbolic program. Note that these methods are typically not needed
in casual uses with prove
, sat
, allSat
etc, as default instances automatically
provide the necessary bits.
Minimal complete definiton: forall, forall_, exists, exists_, literal, fromCW
forall :: String > Symbolic (SBV a)Source
Create a user named input (universal)
forall_ :: Symbolic (SBV a)Source
Create an automatically named input
mkForallVars :: Int > Symbolic [SBV a]Source
Get a bunch of new words
exists :: String > Symbolic (SBV a)Source
Create an existential variable
exists_ :: Symbolic (SBV a)Source
Create an automatically named existential variable
mkExistVars :: Int > Symbolic [SBV a]Source
Create a bunch of existentials
free :: String > Symbolic (SBV a)Source
Create a free variable, universal in a proof, existential in sat
free_ :: Symbolic (SBV a)Source
Create an unnamed free variable, universal in proof, existential in sat
mkFreeVars :: Int > Symbolic [SBV a]Source
Create a bunch of free vars
Turn a literal constant to symbolic
unliteral :: SBV a > Maybe aSource
Extract a literal, if the value is concrete
Extract a literal, from a CW representation
isConcrete :: SBV a > BoolSource
Is the symbolic word concrete?
isSymbolic :: SBV a > BoolSource
Is the symbolic word really symbolic?
isConcretely :: SBV a > (a > Bool) > BoolSource
Does it concretely satisfy the given predicate?
mbMaxBound, mbMinBound :: Maybe aSource
max/minbounds, if available. Note that we don't want to impose Bounded on our class as Integer is not Bounded but it is a SymWord
Getting SMTLib output (for offline analysis)
compileToSMTLib :: Provable a => Bool > a > IO StringSource
Compiles to SMTLib and returns the resulting program as a string. Useful for saving
the result to a file for offline analysis, for instance if you have an SMT solver that's not natively
supported outofthe box by the SBV library. If smtLib2
parameter is False, then we will generate
SMTLib1 output, otherwise we will generate SMTLib2 output
Test case generation
genTest :: Outputtable a => Int > Symbolic a > IO TestVectorsSource
Generate a set of concrete test values from a symbolic program. The output
can be rendered as test vectors in different languages as necessary. Use the
function output
call to indicate what fields should be in the test result.
(Also see constrain
and pConstrain
for filtering acceptable test values.)
getTestValues :: TestVectors > [([CW], [CW])]Source
Retrieve the test vectors for further processing. This function
is useful in cases where renderTest
is not sufficient and custom
output (or further preprocessing) is needed.
data TestVectors Source
Type of test vectors (abstract)
Test output style
Haskell String  As a Haskell value with given name 
C String  As a C array of structs with given name 
Forte String Bool ([Int], [Int])  As a Forte/Verilog value with given name. If the boolean is True then vectors are blasted bigendian, otherwise littleendian The indices are the split points on bitvectors for input and output values 
renderTest :: TestStyle > TestVectors > StringSource
Render the test as a Haskell value with the given name n
.
CW
represents a concrete word of a fixed size:
Endianness is mostly irrelevant (see the FromBits
class).
For signed words, the most significant digit is considered to be the sign.
Code generation from symbolic programs
The SBV library can generate straightline executable code in C. (While other target languages are
certainly possible, currently only C is supported.) The generated code will perform no runtime memoryallocations,
(no calls to malloc
), so its memory usage can be predicted ahead of time. Also, the functions will execute precisely the
same instructions in all calls, so they have predictable timing properties as well. The generated code
has no loops or jumps, and is typically quite fast. While the generated code can be large due to complete unrolling,
these characteristics make them suitable for use in hard realtime systems, as well as in traditional computing.
data SBVCodeGen a Source
The codegeneration monad. Allows for precise layout of input values reference parameters (for returning composite values in languages such as C), and return values.
Monad SBVCodeGen  
MonadIO SBVCodeGen  
MonadState CgState SBVCodeGen 
Setting codegeneration options
cgPerformRTCs :: Bool > SBVCodeGen ()Source
Sets RTC (runtimechecks) for indexoutofbounds, shiftwithlarge value etc. on/off. Default: False
.
cgSetDriverValues :: [Integer] > SBVCodeGen ()Source
Sets driver program run time values, useful for generating programs with fixed drivers for testing. Default: None, i.e., use random values.
cgGenerateDriver :: Bool > SBVCodeGen ()Source
Should we generate a driver program? Default: True
. When a library is generated, it will have
a driver if any of the contituent functions has a driver. (See compileToCLib
.)
cgGenerateMakefile :: Bool > SBVCodeGen ()Source
Should we generate a Makefile? Default: True
.
Designating inputs
cgInput :: SymWord a => String > SBVCodeGen (SBV a)Source
Creates an atomic input in the generated code.
cgInputArr :: SymWord a => Int > String > SBVCodeGen [SBV a]Source
Creates an array input in the generated code.
Designating outputs
cgOutput :: SymWord a => String > SBV a > SBVCodeGen ()Source
Creates an atomic output in the generated code.
cgOutputArr :: SymWord a => String > [SBV a] > SBVCodeGen ()Source
Creates an array output in the generated code.
Designating return values
cgReturn :: SymWord a => SBV a > SBVCodeGen ()Source
Creates a returned (unnamed) value in the generated code.
cgReturnArr :: SymWord a => [SBV a] > SBVCodeGen ()Source
Creates a returned (unnamed) array value in the generated code.
Code generation with uninterpreted functions
cgAddPrototype :: [String] > SBVCodeGen ()Source
Adds the given lines to the header file generated, useful for generating programs with uninterpreted functions.
cgAddDecl :: [String] > SBVCodeGen ()Source
Adds the given lines to the program file generated, useful for generating programs with uninterpreted functions.
cgAddLDFlags :: [String] > SBVCodeGen ()Source
Adds the given words to the compiler options in the generated Makefile, useful for linking extra stuff in
cgIntegerSize :: Int > SBVCodeGen ()Source
Sets number of bits to be used for representing the SInteger
type in the generated C code.
The argument must be one of 8
, 16
, 32
, or 64
. Note that this is essentially unsafe as
the semantics of unbounded Haskell integers becomes reduced to the corresponding bit size, as
typical in most C implementations.
Compilation to C
compileToC :: Maybe FilePath > String > SBVCodeGen () > IO ()Source
Given a symbolic computation, render it as an equivalent collection of files that make up a C program:
 The first argument is the directory name under which the files will be saved. To save
files in the current directory pass
. UseJust
"."Nothing
for printing to stdout.  The second argument is the name of the C function to generate.
 The final argument is the function to be compiled.
Compilation will also generate a Makefile
, a header file, and a driver (test) program, etc.
compileToCLib :: Maybe FilePath > String > [(String, SBVCodeGen ())] > IO ()Source
Create code to generate a library archive (.a) from given symbolic functions. Useful when generating code from multiple functions that work together as a library.
 The first argument is the directory name under which the files will be saved. To save
files in the current directory pass
. UseJust
"."Nothing
for printing to stdout.  The second argument is the name of the archive to generate.
 The third argument is the list of functions to include, in the form of functionname/code pairs, similar
to the second and third arguments of
compileToC
, except in a list.
Module exports
The SBV library exports the following modules wholesale, as user programs will have to import these three modules to make any sensible use of the SBV functionality.
module Data.Bits
module Data.Word
module Data.Int