tmp/tmp910n4727/{from.md → to.md}
RENAMED
|
@@ -0,0 +1,76 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
### Saturation arithmetic <a id="numeric.sat">[[numeric.sat]]</a>
|
| 2 |
+
|
| 3 |
+
#### Arithmetic functions <a id="numeric.sat.func">[[numeric.sat.func]]</a>
|
| 4 |
+
|
| 5 |
+
In the following descriptions, an arithmetic operation is performed as a
|
| 6 |
+
mathematical operation with infinite range and then it is determined
|
| 7 |
+
whether the mathematical result fits into the result type.
|
| 8 |
+
|
| 9 |
+
``` cpp
|
| 10 |
+
template<class T>
|
| 11 |
+
constexpr T add_sat(T x, T y) noexcept;
|
| 12 |
+
```
|
| 13 |
+
|
| 14 |
+
*Constraints:* `T` is a signed or unsigned integer
|
| 15 |
+
type [[basic.fundamental]].
|
| 16 |
+
|
| 17 |
+
*Returns:* If `x` + `y` is representable as a value of type `T`,
|
| 18 |
+
`x` + `y`; otherwise, either the largest or smallest representable value
|
| 19 |
+
of type `T`, whichever is closer to the value of `x` + `y`.
|
| 20 |
+
|
| 21 |
+
``` cpp
|
| 22 |
+
template<class T>
|
| 23 |
+
constexpr T sub_sat(T x, T y) noexcept;
|
| 24 |
+
```
|
| 25 |
+
|
| 26 |
+
*Constraints:* `T` is a signed or unsigned integer
|
| 27 |
+
type [[basic.fundamental]].
|
| 28 |
+
|
| 29 |
+
*Returns:* If `x` - `y` is representable as a value of type `T`,
|
| 30 |
+
`x` - `y`; otherwise, either the largest or smallest representable value
|
| 31 |
+
of type `T`, whichever is closer to the value of `x` - `y`.
|
| 32 |
+
|
| 33 |
+
``` cpp
|
| 34 |
+
template<class T>
|
| 35 |
+
constexpr T mul_sat(T x, T y) noexcept;
|
| 36 |
+
```
|
| 37 |
+
|
| 38 |
+
*Constraints:* `T` is a signed or unsigned integer
|
| 39 |
+
type [[basic.fundamental]].
|
| 40 |
+
|
| 41 |
+
*Returns:* If `x` × `y` is representable as a value of type `T`,
|
| 42 |
+
`x` × `y`; otherwise, either the largest or smallest representable value
|
| 43 |
+
of type `T`, whichever is closer to the value of `x` × `y`.
|
| 44 |
+
|
| 45 |
+
``` cpp
|
| 46 |
+
template<class T>
|
| 47 |
+
constexpr T div_sat(T x, T y) noexcept;
|
| 48 |
+
```
|
| 49 |
+
|
| 50 |
+
*Constraints:* `T` is a signed or unsigned integer
|
| 51 |
+
type [[basic.fundamental]].
|
| 52 |
+
|
| 53 |
+
*Preconditions:* `y != 0` is `true`.
|
| 54 |
+
|
| 55 |
+
*Returns:* If `T` is a signed integer type and
|
| 56 |
+
`x == numeric_limits<T>::min() && y == -1` is `true`,
|
| 57 |
+
`numeric_limits<T>::max()`, otherwise, `x / y`.
|
| 58 |
+
|
| 59 |
+
*Remarks:* A function call expression that violates the precondition in
|
| 60 |
+
the *Preconditions* element is not a core constant
|
| 61 |
+
expression [[expr.const]].
|
| 62 |
+
|
| 63 |
+
#### Casting <a id="numeric.sat.cast">[[numeric.sat.cast]]</a>
|
| 64 |
+
|
| 65 |
+
``` cpp
|
| 66 |
+
template<class R, class T>
|
| 67 |
+
constexpr R saturate_cast(T x) noexcept;
|
| 68 |
+
```
|
| 69 |
+
|
| 70 |
+
*Constraints:* `R` and `T` are signed or unsigned integer
|
| 71 |
+
types [[basic.fundamental]].
|
| 72 |
+
|
| 73 |
+
*Returns:* If `x` is representable as a value of type `R`, `x`;
|
| 74 |
+
otherwise, either the largest or smallest representable value of type
|
| 75 |
+
`R`, whichever is closer to the value of `x`.
|
| 76 |
+
|