CheckedInt.h
/* -*- Mode: C++; tab-width: 8; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
/* vim: set ts=8 sts=2 et sw=2 tw=80: */
/* This Source Code Form is subject to the terms of the Mozilla Public
* License, v. 2.0. If a copy of the MPL was not distributed with this
* file, You can obtain one at http://mozilla.org/MPL/2.0/. */
/* Provides checked integers, detecting integer overflow and divide-by-0. */
#ifndef mozilla_CheckedInt_h
#define mozilla_CheckedInt_h
#include <stdint.h>
#include "mozilla/Assertions.h"
#include "mozilla/Attributes.h"
#include "mozilla/IntegerTypeTraits.h"
namespace mozilla {
template<typename T> class CheckedInt;
namespace detail {
/*
* Step 1: manually record supported types
*
* What's nontrivial here is that there are different families of integer
* types: basic integer types and stdint types. It is merrily undefined which
* types from one family may be just typedefs for a type from another family.
*
* For example, on GCC 4.6, aside from the basic integer types, the only other
* type that isn't just a typedef for some of them, is int8_t.
*/
struct UnsupportedType {};
template<typename IntegerType>
struct IsSupportedPass2
{
static const bool value = false;
};
template<typename IntegerType>
struct IsSupported
{
static const bool value = IsSupportedPass2<IntegerType>::value;
};
template<>
struct IsSupported<int8_t>
{ static const bool value = true; };
template<>
struct IsSupported<uint8_t>
{ static const bool value = true; };
template<>
struct IsSupported<int16_t>
{ static const bool value = true; };
template<>
struct IsSupported<uint16_t>
{ static const bool value = true; };
template<>
struct IsSupported<int32_t>
{ static const bool value = true; };
template<>
struct IsSupported<uint32_t>
{ static const bool value = true; };
template<>
struct IsSupported<int64_t>
{ static const bool value = true; };
template<>
struct IsSupported<uint64_t>
{ static const bool value = true; };
template<>
struct IsSupportedPass2<char>
{ static const bool value = true; };
template<>
struct IsSupportedPass2<signed char>
{ static const bool value = true; };
template<>
struct IsSupportedPass2<unsigned char>
{ static const bool value = true; };
template<>
struct IsSupportedPass2<short>
{ static const bool value = true; };
template<>
struct IsSupportedPass2<unsigned short>
{ static const bool value = true; };
template<>
struct IsSupportedPass2<int>
{ static const bool value = true; };
template<>
struct IsSupportedPass2<unsigned int>
{ static const bool value = true; };
template<>
struct IsSupportedPass2<long>
{ static const bool value = true; };
template<>
struct IsSupportedPass2<unsigned long>
{ static const bool value = true; };
template<>
struct IsSupportedPass2<long long>
{ static const bool value = true; };
template<>
struct IsSupportedPass2<unsigned long long>
{ static const bool value = true; };
/*
* Step 2: Implement the actual validity checks.
*
* Ideas taken from IntegerLib, code different.
*/
template<typename IntegerType, size_t Size = sizeof(IntegerType)>
struct TwiceBiggerType
{
typedef typename detail::StdintTypeForSizeAndSignedness<
sizeof(IntegerType) * 2,
IsSigned<IntegerType>::value
>::Type Type;
};
template<typename IntegerType>
struct TwiceBiggerType<IntegerType, 8>
{
typedef UnsupportedType Type;
};
template<typename T>
inline bool
HasSignBit(T aX)
{
// In C++, right bit shifts on negative values is undefined by the standard.
// Notice that signed-to-unsigned conversions are always well-defined in the
// standard, as the value congruent modulo 2**n as expected. By contrast,
// unsigned-to-signed is only well-defined if the value is representable.
return bool(typename MakeUnsigned<T>::Type(aX) >>
PositionOfSignBit<T>::value);
}
// Bitwise ops may return a larger type, so it's good to use this inline
// helper guaranteeing that the result is really of type T.
template<typename T>
inline T
BinaryComplement(T aX)
{
return ~aX;
}
template<typename T,
typename U,
bool IsTSigned = IsSigned<T>::value,
bool IsUSigned = IsSigned<U>::value>
struct DoesRangeContainRange
{
};
template<typename T, typename U, bool Signedness>
struct DoesRangeContainRange<T, U, Signedness, Signedness>
{
static const bool value = sizeof(T) >= sizeof(U);
};
template<typename T, typename U>
struct DoesRangeContainRange<T, U, true, false>
{
static const bool value = sizeof(T) > sizeof(U);
};
template<typename T, typename U>
struct DoesRangeContainRange<T, U, false, true>
{
static const bool value = false;
};
template<typename T,
typename U,
bool IsTSigned = IsSigned<T>::value,
bool IsUSigned = IsSigned<U>::value,
bool DoesTRangeContainURange = DoesRangeContainRange<T, U>::value>
struct IsInRangeImpl {};
template<typename T, typename U, bool IsTSigned, bool IsUSigned>
struct IsInRangeImpl<T, U, IsTSigned, IsUSigned, true>
{
static bool run(U)
{
return true;
}
};
template<typename T, typename U>
struct IsInRangeImpl<T, U, true, true, false>
{
static bool run(U aX)
{
return aX <= MaxValue<T>::value && aX >= MinValue<T>::value;
}
};
template<typename T, typename U>
struct IsInRangeImpl<T, U, false, false, false>
{
static bool run(U aX)
{
return aX <= MaxValue<T>::value;
}
};
template<typename T, typename U>
struct IsInRangeImpl<T, U, true, false, false>
{
static bool run(U aX)
{
return sizeof(T) > sizeof(U) || aX <= U(MaxValue<T>::value);
}
};
template<typename T, typename U>
struct IsInRangeImpl<T, U, false, true, false>
{
static bool run(U aX)
{
return sizeof(T) >= sizeof(U)
? aX >= 0
: aX >= 0 && aX <= U(MaxValue<T>::value);
}
};
template<typename T, typename U>
inline bool
IsInRange(U aX)
{
return IsInRangeImpl<T, U>::run(aX);
}
template<typename T>
inline bool
IsAddValid(T aX, T aY)
{
// Addition is valid if the sign of aX+aY is equal to either that of aX or
// that of aY. Since the value of aX+aY is undefined if we have a signed
// type, we compute it using the unsigned type of the same size. Beware!
// These bitwise operations can return a larger integer type, if T was a
// small type like int8_t, so we explicitly cast to T.
typename MakeUnsigned<T>::Type ux = aX;
typename MakeUnsigned<T>::Type uy = aY;
typename MakeUnsigned<T>::Type result = ux + uy;
return IsSigned<T>::value
? HasSignBit(BinaryComplement(T((result ^ aX) & (result ^ aY))))
: BinaryComplement(aX) >= aY;
}
template<typename T>
inline bool
IsSubValid(T aX, T aY)
{
// Subtraction is valid if either aX and aY have same sign, or aX-aY and aX
// have same sign. Since the value of aX-aY is undefined if we have a signed
// type, we compute it using the unsigned type of the same size.
typename MakeUnsigned<T>::Type ux = aX;
typename MakeUnsigned<T>::Type uy = aY;
typename MakeUnsigned<T>::Type result = ux - uy;
return IsSigned<T>::value
? HasSignBit(BinaryComplement(T((result ^ aX) & (aX ^ aY))))
: aX >= aY;
}
template<typename T,
bool IsTSigned = IsSigned<T>::value,
bool TwiceBiggerTypeIsSupported =
IsSupported<typename TwiceBiggerType<T>::Type>::value>
struct IsMulValidImpl {};
template<typename T, bool IsTSigned>
struct IsMulValidImpl<T, IsTSigned, true>
{
static bool run(T aX, T aY)
{
typedef typename TwiceBiggerType<T>::Type TwiceBiggerType;
TwiceBiggerType product = TwiceBiggerType(aX) * TwiceBiggerType(aY);
return IsInRange<T>(product);
}
};
template<typename T>
struct IsMulValidImpl<T, true, false>
{
static bool run(T aX, T aY)
{
const T max = MaxValue<T>::value;
const T min = MinValue<T>::value;
if (aX == 0 || aY == 0) {
return true;
}
if (aX > 0) {
return aY > 0
? aX <= max / aY
: aY >= min / aX;
}
// If we reach this point, we know that aX < 0.
return aY > 0
? aX >= min / aY
: aY >= max / aX;
}
};
template<typename T>
struct IsMulValidImpl<T, false, false>
{
static bool run(T aX, T aY)
{
return aY == 0 || aX <= MaxValue<T>::value / aY;
}
};
template<typename T>
inline bool
IsMulValid(T aX, T aY)
{
return IsMulValidImpl<T>::run(aX, aY);
}
template<typename T>
inline bool
IsDivValid(T aX, T aY)
{
// Keep in mind that in the signed case, min/-1 is invalid because
// abs(min)>max.
return aY != 0 &&
!(IsSigned<T>::value && aX == MinValue<T>::value && aY == T(-1));
}
template<typename T, bool IsTSigned = IsSigned<T>::value>
struct IsModValidImpl;
template<typename T>
inline bool
IsModValid(T aX, T aY)
{
return IsModValidImpl<T>::run(aX, aY);
}
/*
* Mod is pretty simple.
* For now, let's just use the ANSI C definition:
* If aX or aY are negative, the results are implementation defined.
* Consider these invalid.
* Undefined for aY=0.
* The result will never exceed either aX or aY.
*
* Checking that aX>=0 is a warning when T is unsigned.
*/
template<typename T>
struct IsModValidImpl<T, false>
{
static inline bool run(T aX, T aY)
{
return aY >= 1;
}
};
template<typename T>
struct IsModValidImpl<T, true>
{
static inline bool run(T aX, T aY)
{
if (aX < 0) {
return false;
}
return aY >= 1;
}
};
template<typename T, bool IsSigned = IsSigned<T>::value>
struct NegateImpl;
template<typename T>
struct NegateImpl<T, false>
{
static CheckedInt<T> negate(const CheckedInt<T>& aVal)
{
// Handle negation separately for signed/unsigned, for simpler code and to
// avoid an MSVC warning negating an unsigned value.
return CheckedInt<T>(0, aVal.isValid() && aVal.mValue == 0);
}
};
template<typename T>
struct NegateImpl<T, true>
{
static CheckedInt<T> negate(const CheckedInt<T>& aVal)
{
// Watch out for the min-value, which (with twos-complement) can't be
// negated as -min-value is then (max-value + 1).
if (!aVal.isValid() || aVal.mValue == MinValue<T>::value) {
return CheckedInt<T>(aVal.mValue, false);
}
return CheckedInt<T>(-aVal.mValue, true);
}
};
} // namespace detail
/*
* Step 3: Now define the CheckedInt class.
*/
/**
* @class CheckedInt
* @brief Integer wrapper class checking for integer overflow and other errors
* @param T the integer type to wrap. Can be any type among the following:
* - any basic integer type such as |int|
* - any stdint type such as |int8_t|
*
* This class implements guarded integer arithmetic. Do a computation, check
* that isValid() returns true, you then have a guarantee that no problem, such
* as integer overflow, happened during this computation, and you can call
* value() to get the plain integer value.
*
* The arithmetic operators in this class are guaranteed not to raise a signal
* (e.g. in case of a division by zero).
*
* For example, suppose that you want to implement a function that computes
* (aX+aY)/aZ, that doesn't crash if aZ==0, and that reports on error (divide by
* zero or integer overflow). You could code it as follows:
@code
bool computeXPlusYOverZ(int aX, int aY, int aZ, int* aResult)
{
CheckedInt<int> checkedResult = (CheckedInt<int>(aX) + aY) / aZ;
if (checkedResult.isValid()) {
*aResult = checkedResult.value();
return true;
} else {
return false;
}
}
@endcode
*
* Implicit conversion from plain integers to checked integers is allowed. The
* plain integer is checked to be in range before being casted to the
* destination type. This means that the following lines all compile, and the
* resulting CheckedInts are correctly detected as valid or invalid:
* @code
// 1 is of type int, is found to be in range for uint8_t, x is valid
CheckedInt<uint8_t> x(1);
// -1 is of type int, is found not to be in range for uint8_t, x is invalid
CheckedInt<uint8_t> x(-1);
// -1 is of type int, is found to be in range for int8_t, x is valid
CheckedInt<int8_t> x(-1);
// 1000 is of type int16_t, is found not to be in range for int8_t,
// x is invalid
CheckedInt<int8_t> x(int16_t(1000));
// 3123456789 is of type uint32_t, is found not to be in range for int32_t,
// x is invalid
CheckedInt<int32_t> x(uint32_t(3123456789));
* @endcode
* Implicit conversion from
* checked integers to plain integers is not allowed. As shown in the
* above example, to get the value of a checked integer as a normal integer,
* call value().
*
* Arithmetic operations between checked and plain integers is allowed; the
* result type is the type of the checked integer.
*
* Checked integers of different types cannot be used in the same arithmetic
* expression.
*
* There are convenience typedefs for all stdint types, of the following form
* (these are just 2 examples):
@code
typedef CheckedInt<int32_t> CheckedInt32;
typedef CheckedInt<uint16_t> CheckedUint16;
@endcode
*/
template<typename T>
class CheckedInt
{
protected:
T mValue;
bool mIsValid;
template<typename U>
CheckedInt(U aValue, bool aIsValid) : mValue(aValue), mIsValid(aIsValid)
{
static_assert(detail::IsSupported<T>::value &&
detail::IsSupported<U>::value,
"This type is not supported by CheckedInt");
}
friend struct detail::NegateImpl<T>;
public:
/**
* Constructs a checked integer with given @a value. The checked integer is
* initialized as valid or invalid depending on whether the @a value
* is in range.
*
* This constructor is not explicit. Instead, the type of its argument is a
* separate template parameter, ensuring that no conversion is performed
* before this constructor is actually called. As explained in the above
* documentation for class CheckedInt, this constructor checks that its
* argument is valid.
*/
template<typename U>
MOZ_IMPLICIT CheckedInt(U aValue) MOZ_NO_ARITHMETIC_EXPR_IN_ARGUMENT
: mValue(T(aValue)),
mIsValid(detail::IsInRange<T>(aValue))
{
static_assert(detail::IsSupported<T>::value &&
detail::IsSupported<U>::value,
"This type is not supported by CheckedInt");
}
template<typename U>
friend class CheckedInt;
template<typename U>
CheckedInt<U> toChecked() const
{
CheckedInt<U> ret(mValue);
ret.mIsValid = ret.mIsValid && mIsValid;
return ret;
}
/** Constructs a valid checked integer with initial value 0 */
CheckedInt() : mValue(0), mIsValid(true)
{
static_assert(detail::IsSupported<T>::value,
"This type is not supported by CheckedInt");
}
/** @returns the actual value */
T value() const
{
MOZ_ASSERT(mIsValid, "Invalid checked integer (division by zero or integer overflow)");
return mValue;
}
/**
* @returns true if the checked integer is valid, i.e. is not the result
* of an invalid operation or of an operation involving an invalid checked
* integer
*/
bool isValid() const
{
return mIsValid;
}
template<typename U>
friend CheckedInt<U> operator +(const CheckedInt<U>& aLhs,
const CheckedInt<U>& aRhs);
template<typename U>
CheckedInt& operator +=(U aRhs);
CheckedInt& operator +=(const CheckedInt<T>& aRhs);
template<typename U>
friend CheckedInt<U> operator -(const CheckedInt<U>& aLhs,
const CheckedInt<U>& aRhs);
template<typename U>
CheckedInt& operator -=(U aRhs);
CheckedInt& operator -=(const CheckedInt<T>& aRhs);
template<typename U>
friend CheckedInt<U> operator *(const CheckedInt<U>& aLhs,
const CheckedInt<U>& aRhs);
template<typename U>
CheckedInt& operator *=(U aRhs);
CheckedInt& operator *=(const CheckedInt<T>& aRhs);
template<typename U>
friend CheckedInt<U> operator /(const CheckedInt<U>& aLhs,
const CheckedInt<U>& aRhs);
template<typename U>
CheckedInt& operator /=(U aRhs);
CheckedInt& operator /=(const CheckedInt<T>& aRhs);
template<typename U>
friend CheckedInt<U> operator %(const CheckedInt<U>& aLhs,
const CheckedInt<U>& aRhs);
template<typename U>
CheckedInt& operator %=(U aRhs);
CheckedInt& operator %=(const CheckedInt<T>& aRhs);
CheckedInt operator -() const
{
return detail::NegateImpl<T>::negate(*this);
}
/**
* @returns true if the left and right hand sides are valid
* and have the same value.
*
* Note that these semantics are the reason why we don't offer
* a operator!=. Indeed, we'd want to have a!=b be equivalent to !(a==b)
* but that would mean that whenever a or b is invalid, a!=b
* is always true, which would be very confusing.
*
* For similar reasons, operators <, >, <=, >= would be very tricky to
* specify, so we just avoid offering them.
*
* Notice that these == semantics are made more reasonable by these facts:
* 1. a==b implies equality at the raw data level
* (the converse is false, as a==b is never true among invalids)
* 2. This is similar to the behavior of IEEE floats, where a==b
* means that a and b have the same value *and* neither is NaN.
*/
bool operator ==(const CheckedInt& aOther) const
{
return mIsValid && aOther.mIsValid && mValue == aOther.mValue;
}
/** prefix ++ */
CheckedInt& operator++()
{
*this += 1;
return *this;
}
/** postfix ++ */
CheckedInt operator++(int)
{
CheckedInt tmp = *this;
*this += 1;
return tmp;
}
/** prefix -- */
CheckedInt& operator--()
{
*this -= 1;
return *this;
}
/** postfix -- */
CheckedInt operator--(int)
{
CheckedInt tmp = *this;
*this -= 1;
return tmp;
}
private:
/**
* The !=, <, <=, >, >= operators are disabled:
* see the comment on operator==.
*/
template<typename U> bool operator !=(U aOther) const = delete;
template<typename U> bool operator < (U aOther) const = delete;
template<typename U> bool operator <=(U aOther) const = delete;
template<typename U> bool operator > (U aOther) const = delete;
template<typename U> bool operator >=(U aOther) const = delete;
};
#define MOZ_CHECKEDINT_BASIC_BINARY_OPERATOR(NAME, OP) \
template<typename T> \
inline CheckedInt<T> \
operator OP(const CheckedInt<T>& aLhs, const CheckedInt<T>& aRhs) \
{ \
if (!detail::Is##NAME##Valid(aLhs.mValue, aRhs.mValue)) { \
return CheckedInt<T>(0, false); \
} \
return CheckedInt<T>(aLhs.mValue OP aRhs.mValue, \
aLhs.mIsValid && aRhs.mIsValid); \
}
MOZ_CHECKEDINT_BASIC_BINARY_OPERATOR(Add, +)
MOZ_CHECKEDINT_BASIC_BINARY_OPERATOR(Sub, -)
MOZ_CHECKEDINT_BASIC_BINARY_OPERATOR(Mul, *)
MOZ_CHECKEDINT_BASIC_BINARY_OPERATOR(Div, /)
MOZ_CHECKEDINT_BASIC_BINARY_OPERATOR(Mod, %)
#undef MOZ_CHECKEDINT_BASIC_BINARY_OPERATOR
// Implement castToCheckedInt<T>(x), making sure that
// - it allows x to be either a CheckedInt<T> or any integer type
// that can be casted to T
// - if x is already a CheckedInt<T>, we just return a reference to it,
// instead of copying it (optimization)
namespace detail {
template<typename T, typename U>
struct CastToCheckedIntImpl
{
typedef CheckedInt<T> ReturnType;
static CheckedInt<T> run(U aU) { return aU; }
};
template<typename T>
struct CastToCheckedIntImpl<T, CheckedInt<T> >
{
typedef const CheckedInt<T>& ReturnType;
static const CheckedInt<T>& run(const CheckedInt<T>& aU) { return aU; }
};
} // namespace detail
template<typename T, typename U>
inline typename detail::CastToCheckedIntImpl<T, U>::ReturnType
castToCheckedInt(U aU)
{
static_assert(detail::IsSupported<T>::value &&
detail::IsSupported<U>::value,
"This type is not supported by CheckedInt");
return detail::CastToCheckedIntImpl<T, U>::run(aU);
}
#define MOZ_CHECKEDINT_CONVENIENCE_BINARY_OPERATORS(OP, COMPOUND_OP) \
template<typename T> \
template<typename U> \
CheckedInt<T>& CheckedInt<T>::operator COMPOUND_OP(U aRhs) \
{ \
*this = *this OP castToCheckedInt<T>(aRhs); \
return *this; \
} \
template<typename T> \
CheckedInt<T>& CheckedInt<T>::operator COMPOUND_OP(const CheckedInt<T>& aRhs) \
{ \
*this = *this OP aRhs; \
return *this; \
} \
template<typename T, typename U> \
inline CheckedInt<T> operator OP(const CheckedInt<T>& aLhs, U aRhs) \
{ \
return aLhs OP castToCheckedInt<T>(aRhs); \
} \
template<typename T, typename U> \
inline CheckedInt<T> operator OP(U aLhs, const CheckedInt<T>& aRhs) \
{ \
return castToCheckedInt<T>(aLhs) OP aRhs; \
}
MOZ_CHECKEDINT_CONVENIENCE_BINARY_OPERATORS(+, +=)
MOZ_CHECKEDINT_CONVENIENCE_BINARY_OPERATORS(*, *=)
MOZ_CHECKEDINT_CONVENIENCE_BINARY_OPERATORS(-, -=)
MOZ_CHECKEDINT_CONVENIENCE_BINARY_OPERATORS(/, /=)
MOZ_CHECKEDINT_CONVENIENCE_BINARY_OPERATORS(%, %=)
#undef MOZ_CHECKEDINT_CONVENIENCE_BINARY_OPERATORS
template<typename T, typename U>
inline bool
operator ==(const CheckedInt<T>& aLhs, U aRhs)
{
return aLhs == castToCheckedInt<T>(aRhs);
}
template<typename T, typename U>
inline bool
operator ==(U aLhs, const CheckedInt<T>& aRhs)
{
return castToCheckedInt<T>(aLhs) == aRhs;
}
// Convenience typedefs.
typedef CheckedInt<int8_t> CheckedInt8;
typedef CheckedInt<uint8_t> CheckedUint8;
typedef CheckedInt<int16_t> CheckedInt16;
typedef CheckedInt<uint16_t> CheckedUint16;
typedef CheckedInt<int32_t> CheckedInt32;
typedef CheckedInt<uint32_t> CheckedUint32;
typedef CheckedInt<int64_t> CheckedInt64;
typedef CheckedInt<uint64_t> CheckedUint64;
} // namespace mozilla
#endif /* mozilla_CheckedInt_h */
