Implicit conversions
When an expression is used in the context where a value of a different type is expected, conversion may occur:
int n = 1L; // expression 1L has type long, int is expected n = 2.1; // expression 2.1 has type double, int is expected char *p = malloc(10); // expression malloc(10) has type void*, char* is expected
Conversions take place in the following situations:
[edit] Conversion as if by assignment
- In the assignment operator, the value of the right-hand operand is converted to the unqualified type of the left-hand operand.
- In scalar initialization, the value of the initializer expression is converted to the unqualified type of the object being initialized
- In a function-call expression, to a function that has a prototype, the value of each argument expression is converted to the type of the unqualified declared types of the corresponding parameter
- In a return statement, the value of the operand of
return
is converted to an object having the return type of the function
Note that actual assignment, in addition to the conversion, also removes extra range and precision from floating-point types and prohibits overlaps; those characteristics do not apply to conversion as if by assignment.
[edit] Default argument promotions
In a function call expression when the call is made to
Each argument of integer type undergoes integer promotion (see below), and each argument of type float is implicitly converted to the type double
int add_nums(int count, ...); int sum = add_nums(2, 'c', true); // add_nums is called with three ints: (2, 99, 1)
Note that float complex and float imaginary are not promoted to double complex and double imaginary in this context. |
(since C99) |
[edit] Usual arithmetic conversions
The arguments of the following arithmetic operators undergo implicit conversions for the purpose of obtaining the common real type, which is the type in which the calculation is performed:
- binary arithmetic *, /, %, +, -
- relational operators <, >, <=, >=, ==, !=
- binary bitwise arithmetic &, ^, |,
- the conditional operator ?:
1) If one operand has decimal floating type, the other operand shall not have standard floating,
complex, or imaginary type.
|
(since C23) |
- integer or real floating type to long double
(since C99) |
- integer or real floating type to double
(since C99) |
- integer type to float (the only real type possible is float, which remains as-is)
(since C99) |
- If the types are the same, that type is the common type.
- Else, the types are different:
- If the types have the same signedness (both signed or both unsigned), the operand whose type has the lesser conversion rank1 is implicitly converted2 to the other type.
- Else, the operands have different signedness:
- If the unsigned type has conversion rank greater than or equal to the rank of the signed type, then the operand with the signed type is implicitly converted to the unsigned type.
- Else, the unsigned type has conversion rank less than the signed type:
- If the signed type can represent all values of the unsigned type, then the operand with the unsigned type is implicitly converted to the signed type.
- Else, both operands undergo implicit conversion to the unsigned type counterpart of the signed operand's type.
- 1. Refer to "integer promotions" below for the rules on ranking.
- 2. Refer to "integer conversions" under "implicit conversion semantics" below.
1.f + 20000001; // int is converted to float, giving 20000000.00 // addition and then rounding to float gives 20000000.00 (char)'a' + 1L; // first, char 'a', which is 97, is promoted to int // different types: int and long // same signedness: both signed // different rank: long is of greater rank than int // therefore, int 97 is converted to long 97 // the result is 97 + 1 = 98 of type signed long 2u - 10; // different types: unsigned int and signed int // different signedness // same rank // therefore, signed int 10 is converted to unsigned int 10 // since the arithmetic operation is performed for unsigned integers // (see "Arithmetic operators" topic), the calculation performed is (2 - 10) // modulo (2 raised to n), where n is the number of value bits of unsigned int // if unsigned int is 32-bit long and there is no padding bits in its object // representation, then the result is (-8) modulo (2 raised to 32) = 4294967288 // of type unsigned int 5UL - 2ULL; // different types: unsigned long and unsigned long long // same signedness // different rank: rank of unsigned long long is greater // therefore, unsigned long 5 is converted to unsigned long long 5 // since the arithmetic operation is performed for unsigned integers // (see "Arithmetic operators" topic), // if unsigned long long is 64-bit long, then // the result is (5 - 2) modulo (2 raised to 64) = 3 of type // unsigned long long 0UL - 1LL; // different types: unsigned long and signed long long // different signedness // different rank: rank of signed long long is greater. // if ULONG_MAX > LLONG_MAX, then signed long long cannot represent all // unsigned long therefore, this is the last case: both operands are converted // to unsigned long long unsigned long 0 is converted to unsigned long long 0 // long long 1 is converted to unsigned long long 1 since the arithmetic // operation is performed for unsigned integers // (see "Arithmetic operators" topic), // if unsigned long long is 64-bit long, then // the calculation is (0 - 1) modulo (2 raised to 64) // thus, the result is 18446744073709551615 (ULLONG_MAX) of type // unsigned long long
The result type is determined as follows:
double complex z = 1 + 2*I; double f = 3.0; z + f; // z remains as-is, f is converted to double, the result is double complex |
(since C99) |
As always, the result of a floating-point operator may have greater range and precision than is indicated by its type (see FLT_EVAL_METHOD).
Note: real and imaginary operands are not implicitly converted to complex because doing so would require extra computation, while producing undesirable results in certain cases involving infinities, NaNs and signed zeros. For example, if reals were converted to complex, 2.0×(3.0+i∞) would evaluate as (2.0+i0.0)×(3.0+i∞) ⇒ (2.0×3.0–0.0×∞) + i(2.0×∞+0.0×3.0) ⇒ NaN+i∞ rather than the correct 6.0+i∞. If imaginaries were converted to complex, i2.0×(∞+i3.0) would evaluate as (0.0+i2.0) × (∞+i3.0) ⇒ (0.0×∞ – 2.0×3.0) + i(0.0×3.0 + 2.0×∞) ⇒ NaN + i∞ instead of –6.0 + i∞. |
(since C99) |
Note: regardless of usual arithmetic conversions, the calculation may always be performed in a narrower type than specified by these rules under the as-if rule
[edit] Value transformations
[edit] Lvalue conversion
Any lvalue expression of any non-array type, when used in any context other than
- as the operand of the address-of operator (if allowed)
- as the operand of the pre/post increment and decrement operators.
- as the left-hand operand of the member access (dot) operator.
- as the left-hand operand of the assignment and compound assignment operators.
- as the operand of sizeof
undergoes lvalue conversion: the type remains the same, but loses const/volatile/restrict-qualifiers and atomic properties, if any. The value remains the same, but loses its lvalue properties (the address may no longer be taken).
If the lvalue has incomplete type, the behavior is undefined.
If the lvalue designates an object of automatic storage duration whose address was never taken and if that object was uninitialized (not declared with an initializer and no assignment to it has been performed prior to use), the behavior is undefined.
This conversion models the memory load of the value of the object from its location.
volatile int n = 1; int x = n; // lvalue conversion on n reads the value of n volatile int* p = &n; // no lvalue conversion: does not read the value of n
[edit] Array to pointer conversion
Any lvalue expression of array type, when used in any context other than
- as the operand of the address-of operator
- as the operand of sizeof
- as the operand of typeof and typeof_unqual (since C23)
- as the string literal used for array initialization
undergoes a conversion to the non-lvalue pointer to its first element.
If the array was declared register, the behavior is undefined.
int a[3], b[3][4]; int* p = a; // conversion to &a[0] int (*q)[4] = b; // conversion to &b[0]
[edit] Function to pointer conversion
Any function designator expression, when used in any context other than
- as the operand of the address-of operator
- as the operand of sizeof
- as the operand of typeof and typeof_unqual (since C23)
undergoes a conversion to the non-lvalue pointer to the function designated by the expression.
int f(int); int (*p)(int) = f; // conversion to &f (***p)(1); // repeated dereference to f and conversion back to &f
[edit] Implicit conversion semantics
Implicit conversion, whether as if by assignment or a usual arithmetic conversion, consists of two stages:
[edit] Compatible types
Conversion of a value of any type to any compatible type is always a no-op and does not change the representation.
uint8_t (*a)[10]; // if uint8_t is a typedef to unsigned char unsigned char (*b)[] = a; // then these pointer types are compatible
[edit] Integer promotions
Integer promotion is the implicit conversion of a value of any integer type with rank less or equal to rank of int or of a bit-field of type _Bool(until C23)bool(since C23), int, signed int, unsigned int, to the value of type int or unsigned int.
If int can represent the entire range of values of the original type (or the range of values of the original bit-field), the value is converted to type int. Otherwise the value is converted to unsigned int.
The value from a bit-field of a bit-precise integer type is converted to the corresponding bit-precise integer type. Otherwise, bit-precise integer types are exempt from the integer promotion rules. |
(since C23) |
Integer promotions preserve the value, including the sign:
int main(void) { void f(); // old-style function declaration // since C23, void f(...) has the same behavior wrt promotions char x = 'a'; // integer conversion from int to char f(x); // integer promotion from char back to int } void f(x) int x; {} // the function expects int
rank above is a property of every integer type and is defined as follows:
8) the rank of a bit-precise signed integer type shall be greater than the rank of any standard integer type with less width or any bit-precise integer type with less width.
9) the rank of any bit-precise integer type relative to an extended integer type of the same width is implementation-defined.
|
(since C23) |
Note: integer promotions are applied only
- as part of usual arithmetic conversions (see above)
- as part of default argument promotions (see above)
- to the operand of the unary arithmetic operators + and -
- to the operand of the unary bitwise operator ~
- to both operands of the shift operators << and >>
Boolean conversionA value of any scalar type can be implicitly converted to _Bool(until C23)bool(since C23). The values that compare equal to an integer constant expression of value zero(until C23)are a zero (for arithmetic types), null (for pointer types) or have a type of nullptr_t(since C23) are converted to 0(until C23)false(since C23), all other values are converted to 1(until C23)true(since C23). bool b1 = 0.5; // b1 == 1 (0.5 converted to int would be zero) bool b2 = 2.0*_Imaginary_I; // b2 == 1 (but converted to int would be zero) bool b3 = 0.0 + 3.0*I; // b3 == 1 (but converted to int would be zero) bool b4 = 0.0/0.0; // b4 == 1 (NaN does not compare equal to zero) bool b5 = nullptr; // b5 == 0 (since C23: nullptr is converted to false) |
(since C99) |
[edit] Integer conversions
A value of any integer type can be implicitly converted to any other integer type. Except where covered by promotions and boolean conversions above, the rules are:
- if the target type can represent the value, the value is unchanged
- otherwise, if the target type is unsigned, the value 2b
, where b is the number of value bits in the target type, is repeatedly subtracted or added to the source value until the result fits in the target type. In other words, unsigned integers implement modulo arithmetic. - otherwise, if the target type is signed, the behavior is implementation-defined (which may include raising a signal)
char x = 'a'; // int -> char, result unchanged unsigned char n = -123456; // target is unsigned, result is 192 (that is, -123456+483*256) signed char m = 123456; // target is signed, result is implementation-defined assert(sizeof(int) > -1); // assert fails: // operator > requests conversion of -1 to size_t, // target is unsigned, result is SIZE_MAX
[edit] Real floating-integer conversions
A finite value of any real floating type can be implicitly converted to any integer type. Except where covered by boolean conversion above, the rules are:
- The fractional part is discarded (truncated towards zero).
- If the resulting value can be represented by the target type, that value is used
- otherwise, the behavior is undefined
int n = 3.14; // n == 3 int x = 1e10; // undefined behavior for 32-bit int
A value of any integer type can be implicitly converted to any real floating type.
- if the value can be represented exactly by the target type, it is unchanged
- if the value can be represented, but cannot be represented exactly, the result is an implementation-defined choice of either the nearest higher or nearest lower value, although if IEEE arithmetic is supported, rounding is to nearest. It is unspecified whether FE_INEXACT is raised in this case.
- if the value cannot be represented, the behavior is undefined, although if IEEE arithmetic is supported, FE_INVALID is raised and the result value is unspecified.
The result of this conversion may have greater range and precision than its target type indicates (see FLT_EVAL_METHOD).
If control over FE_INEXACT is needed in floating-to-integer conversions, rint and nearbyint may be used.
double d = 10; // d = 10.00 float f = 20000001; // f = 20000000.00 (FE_INEXACT) float x = 1+(long long)FLT_MAX; // undefined behavior
[edit] Real floating-point conversions
A value of any real floating type can be implicitly converted to any other real floating type.
- If the value can be represented by the target type exactly, it is unchanged
- if the value can be represented, but cannot be represented exactly, the result is the nearest higher or the nearest lower value (in other words, rounding direction is implementation-defined), although if IEEE arithmetic is supported, rounding is to nearest
- if the value cannot be represented, the behavior is undefined
This section is incomplete
Reason: check IEEE if appropriately-signed infinity is required
The result of this conversion may have greater range and precision than its target type indicates (see FLT_EVAL_METHOD).
double d = 0.1; // d = 0.1000000000000000055511151231257827021181583404541015625 float f = d; // f = 0.100000001490116119384765625 float x = 2*(double)FLT_MAX; // undefined
Complex type conversionsA value of any complex type can be implicitly converted to any other complex type. The real part and the imaginary part individually follow the conversion rules for the real floating types. Imaginary type conversionsA value of any imaginary type can be implicitly converted to any other imaginary type. The imaginary part follows the conversion rules for the real floating types. double imaginary d = 0.1*_Imaginary_I; float imaginary f = d; // f is 0.100000001490116119384765625*I Real-complex conversionsA value of any real floating type can be implicitly converted to any complex type.
A value of any complex type can be implicitly converted to any real floating type
Note: in complex-to-real conversion, a NaN in the imaginary part will not propagate to the real result. double complex z = 0.5 + 3*I; float f = z; // the imaginary part is discarded, f is set to 0.5 z = f; // sets z to 0.5 + 0*I Real-imaginary conversionsA value of any imaginary type can be implicitly converted to any real type (integer or floating-point). The result is always a positive (or unsigned) zero, except when the target type is _Bool(until C23)bool(since C23), in which case boolean conversion rules apply. A value of any real type can be implicitly converted to any imaginary type. The result is always a positive imaginary zero. double imaginary z = 3*I; bool b = z; // Boolean conversion: sets b to true float f = z; // Real-imaginary conversion: sets f to 0.0 z = 3.14; // Imaginary-real conversion: sets z to 0*_Imaginary_I Complex-imaginary conversionsA value of any imaginary type can be implicitly converted to any complex type.
A value of any complex type can be implicitly converted to any imaginary type
double imaginary z = I * (3*I); // the complex result -3.0+0i loses real part // sets z to 0*_Imaginary_I |
(since C99) |
[edit] Pointer conversions
A pointer to void can be implicitly converted to and from any pointer to object type with the following semantics:
- If a pointer to object is converted to a pointer to void and back, its value compares equal to the original pointer.
- No other guarantees are offered
int* p = malloc(10 * sizeof(int)); // malloc returns void*
A pointer to an unqualified type may be implicitly converted to the pointer to qualified version of that type (in other words, const, volatile, and restrict qualifiers can be added). The original pointer and the result compare equal.
int n; const int* p = &n; // &n has type int*
Any integer constant expression with value 0 as well as integer pointer expression with value zero cast to the type void* can be implicitly converted to any pointer type (both pointer to object and pointer to function). The result is the null pointer value of its type, guaranteed to compare unequal to any non-null pointer value of that type. This integer or void* expression is known as null pointer constant and the standard library provides one definition of this constant as the macro NULL.
int* p = 0; double* q = NULL;
[edit] Notes
Although signed integer overflow in any arithmetic operator is undefined behavior, overflowing a signed integer type in an integer conversion is merely unspecified behavior.
On the other hand, although unsigned integer overflow in any arithmetic operator (and in integer conversion) is a well-defined operation and follows the rules of modulo arithmetic, overflowing an unsigned integer in a floating-to-integer conversion is undefined behavior: the values of real floating type that can be converted to unsigned integer are the values from the open interval (-1; Unnn_MAX+1).
unsigned int n = -1.0; // undefined behavior
Conversions between pointers and integers (except from pointer to _Bool(until C23)bool(since C23) and (since C99)from integer constant expression with the value zero to pointer), between pointers to objects (except where either to or from is a pointer to void) and conversions between pointers to functions (except when the functions have compatible types) are never implicit and require a cast operator.
There are no conversions (implicit or explicit) between pointers to functions and pointers to objects (including void*) or integers.
[edit] References
- C23 standard (ISO/IEC 9899:2024):
- 6.3 Conversions (p: TBD)
- C17 standard (ISO/IEC 9899:2018):
- 6.3 Conversions (p: 37-41)
- C11 standard (ISO/IEC 9899:2011):
- 6.3 Conversions (p: 50-56)
- C99 standard (ISO/IEC 9899:1999):
- 6.3 Conversions (p: 42-48)
- C89/C90 standard (ISO/IEC 9899:1990):
- 3.2 Conversions
[edit] See also
C++ documentation for Implicit conversions
|