Package org.sosy_lab.java_smt.api

Interface BitvectorFormulaManager

    • Method Detail

      • makeBitvector

        BitvectorFormula makeBitvector​(int length,
                               long pI)
        Convert a number into a bitvector with given size.
        Throws:
        IllegalArgumentException - if the number is out of range for the given length.

      • makeBitvector

        BitvectorFormula makeBitvector​(int length,
                               NumeralFormula.IntegerFormula pI)
        Convert/Cast/Interpret a numeral formula into a bitvector with given size.

        If the numeral formula is too large for the given length, we cut off the largest bits and only use the lest significant bits.

      • makeVariable

        BitvectorFormula makeVariable​(int length,
                              String pVar)
        Creates a variable with exactly the given name and bitwidth.

        Please make sure that the given name is valid in SMT-LIB2. Take a look at FormulaManager.isValidName(java.lang.String) for further information.

        This method does not quote or unquote the given name, but uses the plain name "AS IS". Formula.toString() can return a different String than the given one.

      • getLength

        int getLength​(BitvectorFormula number)
        This method returns the length of a bitvector, also denoted as bit-size.

      • negate

        BitvectorFormula negate​(BitvectorFormula number)
        This method returns the negated number, i.e., it is multiplied by "-1". The given number is interpreted as signed bitvector and corresponds to "2^BITSIZE - number". The result has the same length as the given number.

      • add

        BitvectorFormula add​(BitvectorFormula number1,
                     BitvectorFormula number2)
        This method returns the addition of the given bitvectors. The result has the same length as the given parameters. There can be an overflow, i.e., as one would expect from bitvector logic. There is no difference in signed and unsigned numbers.
        Parameters:
        number1 - a Formula
        number2 - a Formula
        Returns:
        number1 + number2

      • subtract

        BitvectorFormula subtract​(BitvectorFormula number1,
                          BitvectorFormula number2)
        This method returns the subtraction of the given bitvectors. The result has the same length as the given parameters. There can be an overflow, i.e., as one would expect from bitvector logic. There is no difference in signed and unsigned numbers.
        Parameters:
        number1 - a Formula
        number2 - a Formula
        Returns:
        number1 - number2

      • divide

        BitvectorFormula divide​(BitvectorFormula dividend,
                        BitvectorFormula divisor,
                        boolean signed)
        This method returns the division for two bitvector formulas.

        For signed bitvectors, the result is rounded towards zero (also called "truncated integer division", similar to the division in the C99 standard), e.g., a user can assume the following equations:

        • 10 / 5 = 2
        • 10 / 3 = 3
        • 10 / (-3) = -3
        • -10 / 5 = -2
        • -10 / 3 = -3
        • -10 / (-3) = 3

        If the divisor evaluates to zero (division-by-zero), either directly as value or indirectly via an additional constraint, then the result of the division is defined as:

        • "-1" interpreted as bitvector (i.e., all bits set to "1"), if the dividend is non-negative, and
        • "1" interpreted as bitvector (i.e., all bits set to "0", except the last bit), if the dividend is negative.

        We refer to the SMTLIB standard version 2.6 for the division and remainder operators in BV theory.

        Parameters:
        dividend - dividend of the operation.
        divisor - divisor of the operation.
        signed - whether to interpret all operands as signed or as unsigned numbers.

      • modulo

        @Deprecated(forRemoval=true)
default BitvectorFormula modulo​(BitvectorFormula dividend,
                                BitvectorFormula divisor,
                                boolean signed)
        Deprecated, for removal: This API element is subject to removal in a future version.
        Deprecated and unsupported operation.

        Returns the remainder of the given bitvectors and behaves equally to remainder(BitvectorFormula, BitvectorFormula, boolean).

        Deprecated in favor of remainder() and smodulo() due to confusing method naming and inconsistent behavior (for signed modulo, the sign of the result follows the divisor, but for signed remainder() it follows the dividend). Unsigned remainder() is equivalent to unsigned modulo().

      • smodulo

        BitvectorFormula smodulo​(BitvectorFormula dividend,
                         BitvectorFormula divisor)
        This method returns the two complement signed remainder for the Euclidean division (modulo) of two bitvector formulas.

        The sign of the result follows the sign of the divisor, i.e. the quotient calculated in the modulo operation is rounded in such a way that the result of the smodulo operation follows the sign of the divisor, e.g., a user can assume the following equations, with bitvectors interpreted as signed decimal numbers and % representing signed modulo, to hold:

        • 10 % 5 == 0
        • 10 % 3 == 1
        • 10 % (-3) == -2
        • -10 % 5 == 0
        • -10 % 3 == 2
        • -10 % (-3) == -1

        If the divisor evaluates to zero (modulo-by-zero), either directly as value or indirectly via an additional constraint, then the result of the modulo operation is defined as the dividend itself. We refer to the SMTLIB standard version 2.6 for the division and remainder operators in BV theory.

        For unsigned modulo, we refer to the unsigned remainder method.

        Parameters:
        dividend - dividend of the operation.
        divisor - divisor of the operation.

      • remainder

        BitvectorFormula remainder​(BitvectorFormula dividend,
                           BitvectorFormula divisor,
                           boolean signed)
        This method returns the remainder for two bitvector formulas using the divide(BitvectorFormula, BitvectorFormula, boolean) operation.

        For unsigned bitvectors, this returns the remainder of unsigned bitvector division.

        For signed bitvectors, the sign of the result follows the sign of the dividend, i.e. the quotient of the division is rounded in such a way that the sign of the result of the remainder operation follows the sign of the dividend, e.g., a user can assume the following equations, with bitvectors interpreted as signed decimal numbers and % representing signed remainder (similar to the C programming language), to hold:

        • 10 % 5 == 0
        • 10 % 3 == 1
        • 10 % (-3) == 1
        • -10 % 5 == 0
        • -10 % 3 == -1
        • -10 % (-3) == -1

        If the divisor evaluates to zero (modulo-by-zero), either directly as value or indirectly via an additional constraint, then the result of the modulo operation is defined as the dividend itself. We refer to the SMTLIB standard version 2.6 for the division and remainder operators in BV theory.

        Parameters:
        dividend - dividend of the operation. The sign bit is carried over from this bitvector for signed operations.
        divisor - divisor of the operation.
        signed - whether to interpret all operands as signed or as unsigned numbers.

      • multiply

        BitvectorFormula multiply​(BitvectorFormula number1,
                          BitvectorFormula number2)
        This method returns the multiplication of the given bitvectors. The result has the same length as the given parameters. There can be an overflow, i.e., as one would expect from bitvector logic. There is no difference in signed and unsigned numbers.
        Parameters:
        number1 - a Formula
        number2 - a Formula
        Returns:
        number1 - number2

      • equal

        BooleanFormula equal​(BitvectorFormula number1,
                     BitvectorFormula number2)
        This method returns the bit-wise equality of the given bitvectors.
        Parameters:
        number1 - a Formula
        number2 - a Formula
        Returns:
        number1 == number2

      • greaterThan

        BooleanFormula greaterThan​(BitvectorFormula number1,
                           BitvectorFormula number2,
                           boolean signed)
        Compare the given bitvectors.
        Parameters:
        number1 - a Formula
        number2 - a Formula
        signed - interpret the bitvectors as signed numbers instead of unsigned numbers
        Returns:
        number1 > number2

      • greaterOrEquals

        BooleanFormula greaterOrEquals​(BitvectorFormula number1,
                               BitvectorFormula number2,
                               boolean signed)
        Compare the given bitvectors.
        Parameters:
        number1 - a Formula
        number2 - a Formula
        signed - interpret the bitvectors as signed numbers instead of unsigned numbers
        Returns:
        number1 >= number2

      • lessThan

        BooleanFormula lessThan​(BitvectorFormula number1,
                        BitvectorFormula number2,
                        boolean signed)
        Compare the given bitvectors.
        Parameters:
        number1 - a Formula
        number2 - a Formula
        signed - interpret the bitvectors as signed numbers instead of unsigned numbers
        Returns:
        number1 < number2

      • lessOrEquals

        BooleanFormula lessOrEquals​(BitvectorFormula number1,
                            BitvectorFormula number2,
                            boolean signed)
        Compare the given bitvectors.
        Parameters:
        number1 - a Formula
        number2 - a Formula
        signed - interpret the bitvectors as signed numbers instead of unsigned numbers
        Returns:
        number1 <= number2

      • not

        BitvectorFormula not​(BitvectorFormula bits)
        This method returns the bit-wise complement of the given bitvector.
        Parameters:
        bits - Formula
        Returns:
        ~bits

      • shiftRight

        BitvectorFormula shiftRight​(BitvectorFormula number,
                            BitvectorFormula toShift,
                            boolean signed)
        This method returns a term representing the right shift (towards least-significant bit) of number by toShift. The result has the same length as the given number. On the left side, we fill up the most significant bits with ones (i.e., arithmetic shift), if the number is signed and negative, else we fill up with zeroes. For "toShift >= bitsize", we return a bitvector with value zero, if number was zero or positive, or all bits set to one, if negative.

      • shiftLeft

        BitvectorFormula shiftLeft​(BitvectorFormula number,
                           BitvectorFormula toShift)
        This method returns a term representing the left shift (towards most-significant bit) of number by toShift. The result has the same length as the given number. On the right side, we fill up the least significant bits with zeroes. For "toShift >= bitsize", we return a bitvector with value zero.

      • rotateLeft

        BitvectorFormula rotateLeft​(BitvectorFormula number,
                            int toRotate)
        This method returns a term representing the left rotation (towards most-significant bit) of number by toRotate. The result has the same length as the given number. For "toRotate % bitsize == 0", we return the number itself.
        Parameters:
        toRotate - the number of bits to be moved

      • rotateLeft

        BitvectorFormula rotateLeft​(BitvectorFormula number,
                            BitvectorFormula toRotate)
        This method returns a term representing the left rotation (towards most-significant bit) of number by toRotate. The result has the same length as the given number. For "toRotate % bitsize == 0", we return the number itself.
        Parameters:
        toRotate - unsigned bitvector, represents the number of bits to be moved

      • rotateRight

        BitvectorFormula rotateRight​(BitvectorFormula number,
                             int toRotate)
        This method returns a term representing the right rotation (towards least-significant bit) of number by toRotate. The result has the same length as the given number. For "toRotate % bitsize == 0", we return the number itself.
        Parameters:
        toRotate - the number of bits to be moved

      • rotateRight

        BitvectorFormula rotateRight​(BitvectorFormula number,
                             BitvectorFormula toRotate)
        This method returns a term representing the right rotation (towards least-significant bit) of number by toRotate. The result has the same length as the given number. For "toRotate % bitsize == 0", we return the number itself.
        Parameters:
        toRotate - unsigned bitvector, represents the number of bits to be moved

      • extract

        BitvectorFormula extract​(BitvectorFormula number,
                         int msb,
                         int lsb)
        Extract a range of bits from a bitvector. We require 0 <= lsb <= msb < bitsize.

        If msb equals lsb, then a single bit will be returned, i.e., the bit from the given position. If lsb equals 0 and msb equals bitsize-1, then the complete bitvector will be returned.

        Parameters:
        number - from where the bits are extracted.
        msb - Upper index for the most significant bit. Must be in interval from lsb to bitsize-1.
        lsb - Lower index for the least significant bit. Must be in interval from 0 to msb.

      • extend

        BitvectorFormula extend​(BitvectorFormula number,
                        int extensionBits,
                        boolean signed)
        Extend a bitvector to the left (add most significant bits). If signed is set and the given number is negative, then the bit "1" will be added several times, else "0".
        Parameters:
        number - The bitvector to extend.
        extensionBits - How many bits to add.
        signed - Whether the extension should depend on the sign bit.