|  | /// The addition operator `+`. | 
|  | /// | 
|  | /// Note that `Rhs` is `Self` by default, but this is not mandatory. For | 
|  | /// example, [`std::time::SystemTime`] implements `Add<Duration>`, which permits | 
|  | /// operations of the form `SystemTime = SystemTime + Duration`. | 
|  | /// | 
|  | /// [`std::time::SystemTime`]: ../../std/time/struct.SystemTime.html | 
|  | /// | 
|  | /// # Examples | 
|  | /// | 
|  | /// ## `Add`able points | 
|  | /// | 
|  | /// ``` | 
|  | /// use std::ops::Add; | 
|  | /// | 
|  | /// #[derive(Debug, Copy, Clone, PartialEq)] | 
|  | /// struct Point { | 
|  | ///     x: i32, | 
|  | ///     y: i32, | 
|  | /// } | 
|  | /// | 
|  | /// impl Add for Point { | 
|  | ///     type Output = Self; | 
|  | /// | 
|  | ///     fn add(self, other: Self) -> Self { | 
|  | ///         Self { | 
|  | ///             x: self.x + other.x, | 
|  | ///             y: self.y + other.y, | 
|  | ///         } | 
|  | ///     } | 
|  | /// } | 
|  | /// | 
|  | /// assert_eq!(Point { x: 1, y: 0 } + Point { x: 2, y: 3 }, | 
|  | ///            Point { x: 3, y: 3 }); | 
|  | /// ``` | 
|  | /// | 
|  | /// ## Implementing `Add` with generics | 
|  | /// | 
|  | /// Here is an example of the same `Point` struct implementing the `Add` trait | 
|  | /// using generics. | 
|  | /// | 
|  | /// ``` | 
|  | /// use std::ops::Add; | 
|  | /// | 
|  | /// #[derive(Debug, Copy, Clone, PartialEq)] | 
|  | /// struct Point<T> { | 
|  | ///     x: T, | 
|  | ///     y: T, | 
|  | /// } | 
|  | /// | 
|  | /// // Notice that the implementation uses the associated type `Output`. | 
|  | /// impl<T: Add<Output = T>> Add for Point<T> { | 
|  | ///     type Output = Self; | 
|  | /// | 
|  | ///     fn add(self, other: Self) -> Self::Output { | 
|  | ///         Self { | 
|  | ///             x: self.x + other.x, | 
|  | ///             y: self.y + other.y, | 
|  | ///         } | 
|  | ///     } | 
|  | /// } | 
|  | /// | 
|  | /// assert_eq!(Point { x: 1, y: 0 } + Point { x: 2, y: 3 }, | 
|  | ///            Point { x: 3, y: 3 }); | 
|  | /// ``` | 
|  | #[lang = "add"] | 
|  | #[stable(feature = "rust1", since = "1.0.0")] | 
|  | #[rustc_const_unstable(feature = "const_ops", issue = "143802")] | 
|  | #[rustc_on_unimplemented( | 
|  | on(all(Self = "{integer}", Rhs = "{float}"), message = "cannot add a float to an integer",), | 
|  | on(all(Self = "{float}", Rhs = "{integer}"), message = "cannot add an integer to a float",), | 
|  | message = "cannot add `{Rhs}` to `{Self}`", | 
|  | label = "no implementation for `{Self} + {Rhs}`", | 
|  | append_const_msg | 
|  | )] | 
|  | #[doc(alias = "+")] | 
|  | #[const_trait] | 
|  | pub trait Add<Rhs = Self> { | 
|  | /// The resulting type after applying the `+` operator. | 
|  | #[stable(feature = "rust1", since = "1.0.0")] | 
|  | type Output; | 
|  |  | 
|  | /// Performs the `+` operation. | 
|  | /// | 
|  | /// # Example | 
|  | /// | 
|  | /// ``` | 
|  | /// assert_eq!(12 + 1, 13); | 
|  | /// ``` | 
|  | #[must_use = "this returns the result of the operation, without modifying the original"] | 
|  | #[rustc_diagnostic_item = "add"] | 
|  | #[stable(feature = "rust1", since = "1.0.0")] | 
|  | fn add(self, rhs: Rhs) -> Self::Output; | 
|  | } | 
|  |  | 
|  | macro_rules! add_impl { | 
|  | ($($t:ty)*) => ($( | 
|  | #[stable(feature = "rust1", since = "1.0.0")] | 
|  | #[rustc_const_unstable(feature = "const_ops", issue = "143802")] | 
|  | impl const Add for $t { | 
|  | type Output = $t; | 
|  |  | 
|  | #[inline] | 
|  | #[track_caller] | 
|  | #[rustc_inherit_overflow_checks] | 
|  | fn add(self, other: $t) -> $t { self + other } | 
|  | } | 
|  |  | 
|  | forward_ref_binop! { impl Add, add for $t, $t } | 
|  | )*) | 
|  | } | 
|  |  | 
|  | add_impl! { usize u8 u16 u32 u64 u128 isize i8 i16 i32 i64 i128 f16 f32 f64 f128 } | 
|  |  | 
|  | /// The subtraction operator `-`. | 
|  | /// | 
|  | /// Note that `Rhs` is `Self` by default, but this is not mandatory. For | 
|  | /// example, [`std::time::SystemTime`] implements `Sub<Duration>`, which permits | 
|  | /// operations of the form `SystemTime = SystemTime - Duration`. | 
|  | /// | 
|  | /// [`std::time::SystemTime`]: ../../std/time/struct.SystemTime.html | 
|  | /// | 
|  | /// # Examples | 
|  | /// | 
|  | /// ## `Sub`tractable points | 
|  | /// | 
|  | /// ``` | 
|  | /// use std::ops::Sub; | 
|  | /// | 
|  | /// #[derive(Debug, Copy, Clone, PartialEq)] | 
|  | /// struct Point { | 
|  | ///     x: i32, | 
|  | ///     y: i32, | 
|  | /// } | 
|  | /// | 
|  | /// impl Sub for Point { | 
|  | ///     type Output = Self; | 
|  | /// | 
|  | ///     fn sub(self, other: Self) -> Self::Output { | 
|  | ///         Self { | 
|  | ///             x: self.x - other.x, | 
|  | ///             y: self.y - other.y, | 
|  | ///         } | 
|  | ///     } | 
|  | /// } | 
|  | /// | 
|  | /// assert_eq!(Point { x: 3, y: 3 } - Point { x: 2, y: 3 }, | 
|  | ///            Point { x: 1, y: 0 }); | 
|  | /// ``` | 
|  | /// | 
|  | /// ## Implementing `Sub` with generics | 
|  | /// | 
|  | /// Here is an example of the same `Point` struct implementing the `Sub` trait | 
|  | /// using generics. | 
|  | /// | 
|  | /// ``` | 
|  | /// use std::ops::Sub; | 
|  | /// | 
|  | /// #[derive(Debug, PartialEq)] | 
|  | /// struct Point<T> { | 
|  | ///     x: T, | 
|  | ///     y: T, | 
|  | /// } | 
|  | /// | 
|  | /// // Notice that the implementation uses the associated type `Output`. | 
|  | /// impl<T: Sub<Output = T>> Sub for Point<T> { | 
|  | ///     type Output = Self; | 
|  | /// | 
|  | ///     fn sub(self, other: Self) -> Self::Output { | 
|  | ///         Point { | 
|  | ///             x: self.x - other.x, | 
|  | ///             y: self.y - other.y, | 
|  | ///         } | 
|  | ///     } | 
|  | /// } | 
|  | /// | 
|  | /// assert_eq!(Point { x: 2, y: 3 } - Point { x: 1, y: 0 }, | 
|  | ///            Point { x: 1, y: 3 }); | 
|  | /// ``` | 
|  | #[lang = "sub"] | 
|  | #[stable(feature = "rust1", since = "1.0.0")] | 
|  | #[rustc_const_unstable(feature = "const_ops", issue = "143802")] | 
|  | #[rustc_on_unimplemented( | 
|  | message = "cannot subtract `{Rhs}` from `{Self}`", | 
|  | label = "no implementation for `{Self} - {Rhs}`", | 
|  | append_const_msg | 
|  | )] | 
|  | #[doc(alias = "-")] | 
|  | #[const_trait] | 
|  | pub trait Sub<Rhs = Self> { | 
|  | /// The resulting type after applying the `-` operator. | 
|  | #[stable(feature = "rust1", since = "1.0.0")] | 
|  | type Output; | 
|  |  | 
|  | /// Performs the `-` operation. | 
|  | /// | 
|  | /// # Example | 
|  | /// | 
|  | /// ``` | 
|  | /// assert_eq!(12 - 1, 11); | 
|  | /// ``` | 
|  | #[must_use = "this returns the result of the operation, without modifying the original"] | 
|  | #[rustc_diagnostic_item = "sub"] | 
|  | #[stable(feature = "rust1", since = "1.0.0")] | 
|  | fn sub(self, rhs: Rhs) -> Self::Output; | 
|  | } | 
|  |  | 
|  | macro_rules! sub_impl { | 
|  | ($($t:ty)*) => ($( | 
|  | #[stable(feature = "rust1", since = "1.0.0")] | 
|  | #[rustc_const_unstable(feature = "const_ops", issue = "143802")] | 
|  | impl const Sub for $t { | 
|  | type Output = $t; | 
|  |  | 
|  | #[inline] | 
|  | #[track_caller] | 
|  | #[rustc_inherit_overflow_checks] | 
|  | fn sub(self, other: $t) -> $t { self - other } | 
|  | } | 
|  |  | 
|  | forward_ref_binop! { impl Sub, sub for $t, $t } | 
|  | )*) | 
|  | } | 
|  |  | 
|  | sub_impl! { usize u8 u16 u32 u64 u128 isize i8 i16 i32 i64 i128 f16 f32 f64 f128 } | 
|  |  | 
|  | /// The multiplication operator `*`. | 
|  | /// | 
|  | /// Note that `Rhs` is `Self` by default, but this is not mandatory. | 
|  | /// | 
|  | /// # Examples | 
|  | /// | 
|  | /// ## `Mul`tipliable rational numbers | 
|  | /// | 
|  | /// ``` | 
|  | /// use std::ops::Mul; | 
|  | /// | 
|  | /// // By the fundamental theorem of arithmetic, rational numbers in lowest | 
|  | /// // terms are unique. So, by keeping `Rational`s in reduced form, we can | 
|  | /// // derive `Eq` and `PartialEq`. | 
|  | /// #[derive(Debug, Eq, PartialEq)] | 
|  | /// struct Rational { | 
|  | ///     numerator: usize, | 
|  | ///     denominator: usize, | 
|  | /// } | 
|  | /// | 
|  | /// impl Rational { | 
|  | ///     fn new(numerator: usize, denominator: usize) -> Self { | 
|  | ///         if denominator == 0 { | 
|  | ///             panic!("Zero is an invalid denominator!"); | 
|  | ///         } | 
|  | /// | 
|  | ///         // Reduce to lowest terms by dividing by the greatest common | 
|  | ///         // divisor. | 
|  | ///         let gcd = gcd(numerator, denominator); | 
|  | ///         Self { | 
|  | ///             numerator: numerator / gcd, | 
|  | ///             denominator: denominator / gcd, | 
|  | ///         } | 
|  | ///     } | 
|  | /// } | 
|  | /// | 
|  | /// impl Mul for Rational { | 
|  | ///     // The multiplication of rational numbers is a closed operation. | 
|  | ///     type Output = Self; | 
|  | /// | 
|  | ///     fn mul(self, rhs: Self) -> Self { | 
|  | ///         let numerator = self.numerator * rhs.numerator; | 
|  | ///         let denominator = self.denominator * rhs.denominator; | 
|  | ///         Self::new(numerator, denominator) | 
|  | ///     } | 
|  | /// } | 
|  | /// | 
|  | /// // Euclid's two-thousand-year-old algorithm for finding the greatest common | 
|  | /// // divisor. | 
|  | /// fn gcd(x: usize, y: usize) -> usize { | 
|  | ///     let mut x = x; | 
|  | ///     let mut y = y; | 
|  | ///     while y != 0 { | 
|  | ///         let t = y; | 
|  | ///         y = x % y; | 
|  | ///         x = t; | 
|  | ///     } | 
|  | ///     x | 
|  | /// } | 
|  | /// | 
|  | /// assert_eq!(Rational::new(1, 2), Rational::new(2, 4)); | 
|  | /// assert_eq!(Rational::new(2, 3) * Rational::new(3, 4), | 
|  | ///            Rational::new(1, 2)); | 
|  | /// ``` | 
|  | /// | 
|  | /// ## Multiplying vectors by scalars as in linear algebra | 
|  | /// | 
|  | /// ``` | 
|  | /// use std::ops::Mul; | 
|  | /// | 
|  | /// struct Scalar { value: usize } | 
|  | /// | 
|  | /// #[derive(Debug, PartialEq)] | 
|  | /// struct Vector { value: Vec<usize> } | 
|  | /// | 
|  | /// impl Mul<Scalar> for Vector { | 
|  | ///     type Output = Self; | 
|  | /// | 
|  | ///     fn mul(self, rhs: Scalar) -> Self::Output { | 
|  | ///         Self { value: self.value.iter().map(|v| v * rhs.value).collect() } | 
|  | ///     } | 
|  | /// } | 
|  | /// | 
|  | /// let vector = Vector { value: vec![2, 4, 6] }; | 
|  | /// let scalar = Scalar { value: 3 }; | 
|  | /// assert_eq!(vector * scalar, Vector { value: vec![6, 12, 18] }); | 
|  | /// ``` | 
|  | #[lang = "mul"] | 
|  | #[stable(feature = "rust1", since = "1.0.0")] | 
|  | #[rustc_const_unstable(feature = "const_ops", issue = "143802")] | 
|  | #[diagnostic::on_unimplemented( | 
|  | message = "cannot multiply `{Self}` by `{Rhs}`", | 
|  | label = "no implementation for `{Self} * {Rhs}`" | 
|  | )] | 
|  | #[doc(alias = "*")] | 
|  | #[const_trait] | 
|  | pub trait Mul<Rhs = Self> { | 
|  | /// The resulting type after applying the `*` operator. | 
|  | #[stable(feature = "rust1", since = "1.0.0")] | 
|  | type Output; | 
|  |  | 
|  | /// Performs the `*` operation. | 
|  | /// | 
|  | /// # Example | 
|  | /// | 
|  | /// ``` | 
|  | /// assert_eq!(12 * 2, 24); | 
|  | /// ``` | 
|  | #[must_use = "this returns the result of the operation, without modifying the original"] | 
|  | #[rustc_diagnostic_item = "mul"] | 
|  | #[stable(feature = "rust1", since = "1.0.0")] | 
|  | fn mul(self, rhs: Rhs) -> Self::Output; | 
|  | } | 
|  |  | 
|  | macro_rules! mul_impl { | 
|  | ($($t:ty)*) => ($( | 
|  | #[stable(feature = "rust1", since = "1.0.0")] | 
|  | #[rustc_const_unstable(feature = "const_ops", issue = "143802")] | 
|  | impl const Mul for $t { | 
|  | type Output = $t; | 
|  |  | 
|  | #[inline] | 
|  | #[track_caller] | 
|  | #[rustc_inherit_overflow_checks] | 
|  | fn mul(self, other: $t) -> $t { self * other } | 
|  | } | 
|  |  | 
|  | forward_ref_binop! { impl Mul, mul for $t, $t } | 
|  | )*) | 
|  | } | 
|  |  | 
|  | mul_impl! { usize u8 u16 u32 u64 u128 isize i8 i16 i32 i64 i128 f16 f32 f64 f128 } | 
|  |  | 
|  | /// The division operator `/`. | 
|  | /// | 
|  | /// Note that `Rhs` is `Self` by default, but this is not mandatory. | 
|  | /// | 
|  | /// # Examples | 
|  | /// | 
|  | /// ## `Div`idable rational numbers | 
|  | /// | 
|  | /// ``` | 
|  | /// use std::ops::Div; | 
|  | /// | 
|  | /// // By the fundamental theorem of arithmetic, rational numbers in lowest | 
|  | /// // terms are unique. So, by keeping `Rational`s in reduced form, we can | 
|  | /// // derive `Eq` and `PartialEq`. | 
|  | /// #[derive(Debug, Eq, PartialEq)] | 
|  | /// struct Rational { | 
|  | ///     numerator: usize, | 
|  | ///     denominator: usize, | 
|  | /// } | 
|  | /// | 
|  | /// impl Rational { | 
|  | ///     fn new(numerator: usize, denominator: usize) -> Self { | 
|  | ///         if denominator == 0 { | 
|  | ///             panic!("Zero is an invalid denominator!"); | 
|  | ///         } | 
|  | /// | 
|  | ///         // Reduce to lowest terms by dividing by the greatest common | 
|  | ///         // divisor. | 
|  | ///         let gcd = gcd(numerator, denominator); | 
|  | ///         Self { | 
|  | ///             numerator: numerator / gcd, | 
|  | ///             denominator: denominator / gcd, | 
|  | ///         } | 
|  | ///     } | 
|  | /// } | 
|  | /// | 
|  | /// impl Div for Rational { | 
|  | ///     // The division of rational numbers is a closed operation. | 
|  | ///     type Output = Self; | 
|  | /// | 
|  | ///     fn div(self, rhs: Self) -> Self::Output { | 
|  | ///         if rhs.numerator == 0 { | 
|  | ///             panic!("Cannot divide by zero-valued `Rational`!"); | 
|  | ///         } | 
|  | /// | 
|  | ///         let numerator = self.numerator * rhs.denominator; | 
|  | ///         let denominator = self.denominator * rhs.numerator; | 
|  | ///         Self::new(numerator, denominator) | 
|  | ///     } | 
|  | /// } | 
|  | /// | 
|  | /// // Euclid's two-thousand-year-old algorithm for finding the greatest common | 
|  | /// // divisor. | 
|  | /// fn gcd(x: usize, y: usize) -> usize { | 
|  | ///     let mut x = x; | 
|  | ///     let mut y = y; | 
|  | ///     while y != 0 { | 
|  | ///         let t = y; | 
|  | ///         y = x % y; | 
|  | ///         x = t; | 
|  | ///     } | 
|  | ///     x | 
|  | /// } | 
|  | /// | 
|  | /// assert_eq!(Rational::new(1, 2), Rational::new(2, 4)); | 
|  | /// assert_eq!(Rational::new(1, 2) / Rational::new(3, 4), | 
|  | ///            Rational::new(2, 3)); | 
|  | /// ``` | 
|  | /// | 
|  | /// ## Dividing vectors by scalars as in linear algebra | 
|  | /// | 
|  | /// ``` | 
|  | /// use std::ops::Div; | 
|  | /// | 
|  | /// struct Scalar { value: f32 } | 
|  | /// | 
|  | /// #[derive(Debug, PartialEq)] | 
|  | /// struct Vector { value: Vec<f32> } | 
|  | /// | 
|  | /// impl Div<Scalar> for Vector { | 
|  | ///     type Output = Self; | 
|  | /// | 
|  | ///     fn div(self, rhs: Scalar) -> Self::Output { | 
|  | ///         Self { value: self.value.iter().map(|v| v / rhs.value).collect() } | 
|  | ///     } | 
|  | /// } | 
|  | /// | 
|  | /// let scalar = Scalar { value: 2f32 }; | 
|  | /// let vector = Vector { value: vec![2f32, 4f32, 6f32] }; | 
|  | /// assert_eq!(vector / scalar, Vector { value: vec![1f32, 2f32, 3f32] }); | 
|  | /// ``` | 
|  | #[lang = "div"] | 
|  | #[stable(feature = "rust1", since = "1.0.0")] | 
|  | #[rustc_const_unstable(feature = "const_ops", issue = "143802")] | 
|  | #[diagnostic::on_unimplemented( | 
|  | message = "cannot divide `{Self}` by `{Rhs}`", | 
|  | label = "no implementation for `{Self} / {Rhs}`" | 
|  | )] | 
|  | #[doc(alias = "/")] | 
|  | #[const_trait] | 
|  | pub trait Div<Rhs = Self> { | 
|  | /// The resulting type after applying the `/` operator. | 
|  | #[stable(feature = "rust1", since = "1.0.0")] | 
|  | type Output; | 
|  |  | 
|  | /// Performs the `/` operation. | 
|  | /// | 
|  | /// # Example | 
|  | /// | 
|  | /// ``` | 
|  | /// assert_eq!(12 / 2, 6); | 
|  | /// ``` | 
|  | #[must_use = "this returns the result of the operation, without modifying the original"] | 
|  | #[rustc_diagnostic_item = "div"] | 
|  | #[stable(feature = "rust1", since = "1.0.0")] | 
|  | fn div(self, rhs: Rhs) -> Self::Output; | 
|  | } | 
|  |  | 
|  | macro_rules! div_impl_integer { | 
|  | ($(($($t:ty)*) => $panic:expr),*) => ($($( | 
|  | /// This operation rounds towards zero, truncating any | 
|  | /// fractional part of the exact result. | 
|  | /// | 
|  | /// # Panics | 
|  | /// | 
|  | #[doc = $panic] | 
|  | #[stable(feature = "rust1", since = "1.0.0")] | 
|  | #[rustc_const_unstable(feature = "const_ops", issue = "143802")] | 
|  | impl const Div for $t { | 
|  | type Output = $t; | 
|  |  | 
|  | #[inline] | 
|  | #[track_caller] | 
|  | fn div(self, other: $t) -> $t { self / other } | 
|  | } | 
|  |  | 
|  | forward_ref_binop! { impl Div, div for $t, $t } | 
|  | )*)*) | 
|  | } | 
|  |  | 
|  | div_impl_integer! { | 
|  | (usize u8 u16 u32 u64 u128) => "This operation will panic if `other == 0`.", | 
|  | (isize i8 i16 i32 i64 i128) => "This operation will panic if `other == 0` or the division results in overflow." | 
|  | } | 
|  |  | 
|  | macro_rules! div_impl_float { | 
|  | ($($t:ty)*) => ($( | 
|  | #[stable(feature = "rust1", since = "1.0.0")] | 
|  | #[rustc_const_unstable(feature = "const_ops", issue = "143802")] | 
|  | impl const Div for $t { | 
|  | type Output = $t; | 
|  |  | 
|  | #[inline] | 
|  | fn div(self, other: $t) -> $t { self / other } | 
|  | } | 
|  |  | 
|  | forward_ref_binop! { impl Div, div for $t, $t } | 
|  | )*) | 
|  | } | 
|  |  | 
|  | div_impl_float! { f16 f32 f64 f128 } | 
|  |  | 
|  | /// The remainder operator `%`. | 
|  | /// | 
|  | /// Note that `Rhs` is `Self` by default, but this is not mandatory. | 
|  | /// | 
|  | /// # Examples | 
|  | /// | 
|  | /// This example implements `Rem` on a `SplitSlice` object. After `Rem` is | 
|  | /// implemented, one can use the `%` operator to find out what the remaining | 
|  | /// elements of the slice would be after splitting it into equal slices of a | 
|  | /// given length. | 
|  | /// | 
|  | /// ``` | 
|  | /// use std::ops::Rem; | 
|  | /// | 
|  | /// #[derive(PartialEq, Debug)] | 
|  | /// struct SplitSlice<'a, T> { | 
|  | ///     slice: &'a [T], | 
|  | /// } | 
|  | /// | 
|  | /// impl<'a, T> Rem<usize> for SplitSlice<'a, T> { | 
|  | ///     type Output = Self; | 
|  | /// | 
|  | ///     fn rem(self, modulus: usize) -> Self::Output { | 
|  | ///         let len = self.slice.len(); | 
|  | ///         let rem = len % modulus; | 
|  | ///         let start = len - rem; | 
|  | ///         Self {slice: &self.slice[start..]} | 
|  | ///     } | 
|  | /// } | 
|  | /// | 
|  | /// // If we were to divide &[0, 1, 2, 3, 4, 5, 6, 7] into slices of size 3, | 
|  | /// // the remainder would be &[6, 7]. | 
|  | /// assert_eq!(SplitSlice { slice: &[0, 1, 2, 3, 4, 5, 6, 7] } % 3, | 
|  | ///            SplitSlice { slice: &[6, 7] }); | 
|  | /// ``` | 
|  | #[lang = "rem"] | 
|  | #[stable(feature = "rust1", since = "1.0.0")] | 
|  | #[rustc_const_unstable(feature = "const_ops", issue = "143802")] | 
|  | #[diagnostic::on_unimplemented( | 
|  | message = "cannot calculate the remainder of `{Self}` divided by `{Rhs}`", | 
|  | label = "no implementation for `{Self} % {Rhs}`" | 
|  | )] | 
|  | #[doc(alias = "%")] | 
|  | #[const_trait] | 
|  | pub trait Rem<Rhs = Self> { | 
|  | /// The resulting type after applying the `%` operator. | 
|  | #[stable(feature = "rust1", since = "1.0.0")] | 
|  | type Output; | 
|  |  | 
|  | /// Performs the `%` operation. | 
|  | /// | 
|  | /// # Example | 
|  | /// | 
|  | /// ``` | 
|  | /// assert_eq!(12 % 10, 2); | 
|  | /// ``` | 
|  | #[must_use = "this returns the result of the operation, without modifying the original"] | 
|  | #[rustc_diagnostic_item = "rem"] | 
|  | #[stable(feature = "rust1", since = "1.0.0")] | 
|  | fn rem(self, rhs: Rhs) -> Self::Output; | 
|  | } | 
|  |  | 
|  | macro_rules! rem_impl_integer { | 
|  | ($(($($t:ty)*) => $panic:expr),*) => ($($( | 
|  | /// This operation satisfies `n % d == n - (n / d) * d`. The | 
|  | /// result has the same sign as the left operand. | 
|  | /// | 
|  | /// # Panics | 
|  | /// | 
|  | #[doc = $panic] | 
|  | #[stable(feature = "rust1", since = "1.0.0")] | 
|  | #[rustc_const_unstable(feature = "const_ops", issue = "143802")] | 
|  | impl const Rem for $t { | 
|  | type Output = $t; | 
|  |  | 
|  | #[inline] | 
|  | #[track_caller] | 
|  | fn rem(self, other: $t) -> $t { self % other } | 
|  | } | 
|  |  | 
|  | forward_ref_binop! { impl Rem, rem for $t, $t } | 
|  | )*)*) | 
|  | } | 
|  |  | 
|  | rem_impl_integer! { | 
|  | (usize u8 u16 u32 u64 u128) => "This operation will panic if `other == 0`.", | 
|  | (isize i8 i16 i32 i64 i128) => "This operation will panic if `other == 0` or if `self / other` results in overflow." | 
|  | } | 
|  |  | 
|  | macro_rules! rem_impl_float { | 
|  | ($($t:ty)*) => ($( | 
|  |  | 
|  | /// The remainder from the division of two floats. | 
|  | /// | 
|  | /// The remainder has the same sign as the dividend and is computed as: | 
|  | /// `x - (x / y).trunc() * y`. | 
|  | /// | 
|  | /// # Examples | 
|  | /// ``` | 
|  | /// let x: f32 = 50.50; | 
|  | /// let y: f32 = 8.125; | 
|  | /// let remainder = x - (x / y).trunc() * y; | 
|  | /// | 
|  | /// // The answer to both operations is 1.75 | 
|  | /// assert_eq!(x % y, remainder); | 
|  | /// ``` | 
|  | #[stable(feature = "rust1", since = "1.0.0")] | 
|  | #[rustc_const_unstable(feature = "const_ops", issue = "143802")] | 
|  | impl const Rem for $t { | 
|  | type Output = $t; | 
|  |  | 
|  | #[inline] | 
|  | fn rem(self, other: $t) -> $t { self % other } | 
|  | } | 
|  |  | 
|  | forward_ref_binop! { impl Rem, rem for $t, $t } | 
|  | )*) | 
|  | } | 
|  |  | 
|  | rem_impl_float! { f16 f32 f64 f128 } | 
|  |  | 
|  | /// The unary negation operator `-`. | 
|  | /// | 
|  | /// # Examples | 
|  | /// | 
|  | /// An implementation of `Neg` for `Sign`, which allows the use of `-` to | 
|  | /// negate its value. | 
|  | /// | 
|  | /// ``` | 
|  | /// use std::ops::Neg; | 
|  | /// | 
|  | /// #[derive(Debug, PartialEq)] | 
|  | /// enum Sign { | 
|  | ///     Negative, | 
|  | ///     Zero, | 
|  | ///     Positive, | 
|  | /// } | 
|  | /// | 
|  | /// impl Neg for Sign { | 
|  | ///     type Output = Self; | 
|  | /// | 
|  | ///     fn neg(self) -> Self::Output { | 
|  | ///         match self { | 
|  | ///             Sign::Negative => Sign::Positive, | 
|  | ///             Sign::Zero => Sign::Zero, | 
|  | ///             Sign::Positive => Sign::Negative, | 
|  | ///         } | 
|  | ///     } | 
|  | /// } | 
|  | /// | 
|  | /// // A negative positive is a negative. | 
|  | /// assert_eq!(-Sign::Positive, Sign::Negative); | 
|  | /// // A double negative is a positive. | 
|  | /// assert_eq!(-Sign::Negative, Sign::Positive); | 
|  | /// // Zero is its own negation. | 
|  | /// assert_eq!(-Sign::Zero, Sign::Zero); | 
|  | /// ``` | 
|  | #[lang = "neg"] | 
|  | #[stable(feature = "rust1", since = "1.0.0")] | 
|  | #[doc(alias = "-")] | 
|  | pub trait Neg { | 
|  | /// The resulting type after applying the `-` operator. | 
|  | #[stable(feature = "rust1", since = "1.0.0")] | 
|  | type Output; | 
|  |  | 
|  | /// Performs the unary `-` operation. | 
|  | /// | 
|  | /// # Example | 
|  | /// | 
|  | /// ``` | 
|  | /// let x: i32 = 12; | 
|  | /// assert_eq!(-x, -12); | 
|  | /// ``` | 
|  | #[must_use = "this returns the result of the operation, without modifying the original"] | 
|  | #[rustc_diagnostic_item = "neg"] | 
|  | #[stable(feature = "rust1", since = "1.0.0")] | 
|  | fn neg(self) -> Self::Output; | 
|  | } | 
|  |  | 
|  | macro_rules! neg_impl { | 
|  | ($($t:ty)*) => ($( | 
|  | #[stable(feature = "rust1", since = "1.0.0")] | 
|  | impl Neg for $t { | 
|  | type Output = $t; | 
|  |  | 
|  | #[inline] | 
|  | #[rustc_inherit_overflow_checks] | 
|  | fn neg(self) -> $t { -self } | 
|  | } | 
|  |  | 
|  | forward_ref_unop! { impl Neg, neg for $t } | 
|  | )*) | 
|  | } | 
|  |  | 
|  | neg_impl! { isize i8 i16 i32 i64 i128 f16 f32 f64 f128 } | 
|  |  | 
|  | /// The addition assignment operator `+=`. | 
|  | /// | 
|  | /// # Examples | 
|  | /// | 
|  | /// This example creates a `Point` struct that implements the `AddAssign` | 
|  | /// trait, and then demonstrates add-assigning to a mutable `Point`. | 
|  | /// | 
|  | /// ``` | 
|  | /// use std::ops::AddAssign; | 
|  | /// | 
|  | /// #[derive(Debug, Copy, Clone, PartialEq)] | 
|  | /// struct Point { | 
|  | ///     x: i32, | 
|  | ///     y: i32, | 
|  | /// } | 
|  | /// | 
|  | /// impl AddAssign for Point { | 
|  | ///     fn add_assign(&mut self, other: Self) { | 
|  | ///         *self = Self { | 
|  | ///             x: self.x + other.x, | 
|  | ///             y: self.y + other.y, | 
|  | ///         }; | 
|  | ///     } | 
|  | /// } | 
|  | /// | 
|  | /// let mut point = Point { x: 1, y: 0 }; | 
|  | /// point += Point { x: 2, y: 3 }; | 
|  | /// assert_eq!(point, Point { x: 3, y: 3 }); | 
|  | /// ``` | 
|  | #[lang = "add_assign"] | 
|  | #[stable(feature = "op_assign_traits", since = "1.8.0")] | 
|  | #[diagnostic::on_unimplemented( | 
|  | message = "cannot add-assign `{Rhs}` to `{Self}`", | 
|  | label = "no implementation for `{Self} += {Rhs}`" | 
|  | )] | 
|  | #[doc(alias = "+")] | 
|  | #[doc(alias = "+=")] | 
|  | pub trait AddAssign<Rhs = Self> { | 
|  | /// Performs the `+=` operation. | 
|  | /// | 
|  | /// # Example | 
|  | /// | 
|  | /// ``` | 
|  | /// let mut x: u32 = 12; | 
|  | /// x += 1; | 
|  | /// assert_eq!(x, 13); | 
|  | /// ``` | 
|  | #[stable(feature = "op_assign_traits", since = "1.8.0")] | 
|  | fn add_assign(&mut self, rhs: Rhs); | 
|  | } | 
|  |  | 
|  | macro_rules! add_assign_impl { | 
|  | ($($t:ty)+) => ($( | 
|  | #[stable(feature = "op_assign_traits", since = "1.8.0")] | 
|  | impl AddAssign for $t { | 
|  | #[inline] | 
|  | #[track_caller] | 
|  | #[rustc_inherit_overflow_checks] | 
|  | fn add_assign(&mut self, other: $t) { *self += other } | 
|  | } | 
|  |  | 
|  | forward_ref_op_assign! { impl AddAssign, add_assign for $t, $t } | 
|  | )+) | 
|  | } | 
|  |  | 
|  | add_assign_impl! { usize u8 u16 u32 u64 u128 isize i8 i16 i32 i64 i128 f16 f32 f64 f128 } | 
|  |  | 
|  | /// The subtraction assignment operator `-=`. | 
|  | /// | 
|  | /// # Examples | 
|  | /// | 
|  | /// This example creates a `Point` struct that implements the `SubAssign` | 
|  | /// trait, and then demonstrates sub-assigning to a mutable `Point`. | 
|  | /// | 
|  | /// ``` | 
|  | /// use std::ops::SubAssign; | 
|  | /// | 
|  | /// #[derive(Debug, Copy, Clone, PartialEq)] | 
|  | /// struct Point { | 
|  | ///     x: i32, | 
|  | ///     y: i32, | 
|  | /// } | 
|  | /// | 
|  | /// impl SubAssign for Point { | 
|  | ///     fn sub_assign(&mut self, other: Self) { | 
|  | ///         *self = Self { | 
|  | ///             x: self.x - other.x, | 
|  | ///             y: self.y - other.y, | 
|  | ///         }; | 
|  | ///     } | 
|  | /// } | 
|  | /// | 
|  | /// let mut point = Point { x: 3, y: 3 }; | 
|  | /// point -= Point { x: 2, y: 3 }; | 
|  | /// assert_eq!(point, Point {x: 1, y: 0}); | 
|  | /// ``` | 
|  | #[lang = "sub_assign"] | 
|  | #[stable(feature = "op_assign_traits", since = "1.8.0")] | 
|  | #[diagnostic::on_unimplemented( | 
|  | message = "cannot subtract-assign `{Rhs}` from `{Self}`", | 
|  | label = "no implementation for `{Self} -= {Rhs}`" | 
|  | )] | 
|  | #[doc(alias = "-")] | 
|  | #[doc(alias = "-=")] | 
|  | pub trait SubAssign<Rhs = Self> { | 
|  | /// Performs the `-=` operation. | 
|  | /// | 
|  | /// # Example | 
|  | /// | 
|  | /// ``` | 
|  | /// let mut x: u32 = 12; | 
|  | /// x -= 1; | 
|  | /// assert_eq!(x, 11); | 
|  | /// ``` | 
|  | #[stable(feature = "op_assign_traits", since = "1.8.0")] | 
|  | fn sub_assign(&mut self, rhs: Rhs); | 
|  | } | 
|  |  | 
|  | macro_rules! sub_assign_impl { | 
|  | ($($t:ty)+) => ($( | 
|  | #[stable(feature = "op_assign_traits", since = "1.8.0")] | 
|  | impl SubAssign for $t { | 
|  | #[inline] | 
|  | #[track_caller] | 
|  | #[rustc_inherit_overflow_checks] | 
|  | fn sub_assign(&mut self, other: $t) { *self -= other } | 
|  | } | 
|  |  | 
|  | forward_ref_op_assign! { impl SubAssign, sub_assign for $t, $t } | 
|  | )+) | 
|  | } | 
|  |  | 
|  | sub_assign_impl! { usize u8 u16 u32 u64 u128 isize i8 i16 i32 i64 i128 f16 f32 f64 f128 } | 
|  |  | 
|  | /// The multiplication assignment operator `*=`. | 
|  | /// | 
|  | /// # Examples | 
|  | /// | 
|  | /// ``` | 
|  | /// use std::ops::MulAssign; | 
|  | /// | 
|  | /// #[derive(Debug, PartialEq)] | 
|  | /// struct Frequency { hertz: f64 } | 
|  | /// | 
|  | /// impl MulAssign<f64> for Frequency { | 
|  | ///     fn mul_assign(&mut self, rhs: f64) { | 
|  | ///         self.hertz *= rhs; | 
|  | ///     } | 
|  | /// } | 
|  | /// | 
|  | /// let mut frequency = Frequency { hertz: 50.0 }; | 
|  | /// frequency *= 4.0; | 
|  | /// assert_eq!(Frequency { hertz: 200.0 }, frequency); | 
|  | /// ``` | 
|  | #[lang = "mul_assign"] | 
|  | #[stable(feature = "op_assign_traits", since = "1.8.0")] | 
|  | #[diagnostic::on_unimplemented( | 
|  | message = "cannot multiply-assign `{Self}` by `{Rhs}`", | 
|  | label = "no implementation for `{Self} *= {Rhs}`" | 
|  | )] | 
|  | #[doc(alias = "*")] | 
|  | #[doc(alias = "*=")] | 
|  | pub trait MulAssign<Rhs = Self> { | 
|  | /// Performs the `*=` operation. | 
|  | /// | 
|  | /// # Example | 
|  | /// | 
|  | /// ``` | 
|  | /// let mut x: u32 = 12; | 
|  | /// x *= 2; | 
|  | /// assert_eq!(x, 24); | 
|  | /// ``` | 
|  | #[stable(feature = "op_assign_traits", since = "1.8.0")] | 
|  | fn mul_assign(&mut self, rhs: Rhs); | 
|  | } | 
|  |  | 
|  | macro_rules! mul_assign_impl { | 
|  | ($($t:ty)+) => ($( | 
|  | #[stable(feature = "op_assign_traits", since = "1.8.0")] | 
|  | impl MulAssign for $t { | 
|  | #[inline] | 
|  | #[track_caller] | 
|  | #[rustc_inherit_overflow_checks] | 
|  | fn mul_assign(&mut self, other: $t) { *self *= other } | 
|  | } | 
|  |  | 
|  | forward_ref_op_assign! { impl MulAssign, mul_assign for $t, $t } | 
|  | )+) | 
|  | } | 
|  |  | 
|  | mul_assign_impl! { usize u8 u16 u32 u64 u128 isize i8 i16 i32 i64 i128 f16 f32 f64 f128 } | 
|  |  | 
|  | /// The division assignment operator `/=`. | 
|  | /// | 
|  | /// # Examples | 
|  | /// | 
|  | /// ``` | 
|  | /// use std::ops::DivAssign; | 
|  | /// | 
|  | /// #[derive(Debug, PartialEq)] | 
|  | /// struct Frequency { hertz: f64 } | 
|  | /// | 
|  | /// impl DivAssign<f64> for Frequency { | 
|  | ///     fn div_assign(&mut self, rhs: f64) { | 
|  | ///         self.hertz /= rhs; | 
|  | ///     } | 
|  | /// } | 
|  | /// | 
|  | /// let mut frequency = Frequency { hertz: 200.0 }; | 
|  | /// frequency /= 4.0; | 
|  | /// assert_eq!(Frequency { hertz: 50.0 }, frequency); | 
|  | /// ``` | 
|  | #[lang = "div_assign"] | 
|  | #[stable(feature = "op_assign_traits", since = "1.8.0")] | 
|  | #[diagnostic::on_unimplemented( | 
|  | message = "cannot divide-assign `{Self}` by `{Rhs}`", | 
|  | label = "no implementation for `{Self} /= {Rhs}`" | 
|  | )] | 
|  | #[doc(alias = "/")] | 
|  | #[doc(alias = "/=")] | 
|  | pub trait DivAssign<Rhs = Self> { | 
|  | /// Performs the `/=` operation. | 
|  | /// | 
|  | /// # Example | 
|  | /// | 
|  | /// ``` | 
|  | /// let mut x: u32 = 12; | 
|  | /// x /= 2; | 
|  | /// assert_eq!(x, 6); | 
|  | /// ``` | 
|  | #[stable(feature = "op_assign_traits", since = "1.8.0")] | 
|  | fn div_assign(&mut self, rhs: Rhs); | 
|  | } | 
|  |  | 
|  | macro_rules! div_assign_impl { | 
|  | ($($t:ty)+) => ($( | 
|  | #[stable(feature = "op_assign_traits", since = "1.8.0")] | 
|  | impl DivAssign for $t { | 
|  | #[inline] | 
|  | #[track_caller] | 
|  | fn div_assign(&mut self, other: $t) { *self /= other } | 
|  | } | 
|  |  | 
|  | forward_ref_op_assign! { impl DivAssign, div_assign for $t, $t } | 
|  | )+) | 
|  | } | 
|  |  | 
|  | div_assign_impl! { usize u8 u16 u32 u64 u128 isize i8 i16 i32 i64 i128 f16 f32 f64 f128 } | 
|  |  | 
|  | /// The remainder assignment operator `%=`. | 
|  | /// | 
|  | /// # Examples | 
|  | /// | 
|  | /// ``` | 
|  | /// use std::ops::RemAssign; | 
|  | /// | 
|  | /// struct CookieJar { cookies: u32 } | 
|  | /// | 
|  | /// impl RemAssign<u32> for CookieJar { | 
|  | ///     fn rem_assign(&mut self, piles: u32) { | 
|  | ///         self.cookies %= piles; | 
|  | ///     } | 
|  | /// } | 
|  | /// | 
|  | /// let mut jar = CookieJar { cookies: 31 }; | 
|  | /// let piles = 4; | 
|  | /// | 
|  | /// println!("Splitting up {} cookies into {} even piles!", jar.cookies, piles); | 
|  | /// | 
|  | /// jar %= piles; | 
|  | /// | 
|  | /// println!("{} cookies remain in the cookie jar!", jar.cookies); | 
|  | /// ``` | 
|  | #[lang = "rem_assign"] | 
|  | #[stable(feature = "op_assign_traits", since = "1.8.0")] | 
|  | #[diagnostic::on_unimplemented( | 
|  | message = "cannot calculate and assign the remainder of `{Self}` divided by `{Rhs}`", | 
|  | label = "no implementation for `{Self} %= {Rhs}`" | 
|  | )] | 
|  | #[doc(alias = "%")] | 
|  | #[doc(alias = "%=")] | 
|  | pub trait RemAssign<Rhs = Self> { | 
|  | /// Performs the `%=` operation. | 
|  | /// | 
|  | /// # Example | 
|  | /// | 
|  | /// ``` | 
|  | /// let mut x: u32 = 12; | 
|  | /// x %= 10; | 
|  | /// assert_eq!(x, 2); | 
|  | /// ``` | 
|  | #[stable(feature = "op_assign_traits", since = "1.8.0")] | 
|  | fn rem_assign(&mut self, rhs: Rhs); | 
|  | } | 
|  |  | 
|  | macro_rules! rem_assign_impl { | 
|  | ($($t:ty)+) => ($( | 
|  | #[stable(feature = "op_assign_traits", since = "1.8.0")] | 
|  | impl RemAssign for $t { | 
|  | #[inline] | 
|  | #[track_caller] | 
|  | fn rem_assign(&mut self, other: $t) { *self %= other } | 
|  | } | 
|  |  | 
|  | forward_ref_op_assign! { impl RemAssign, rem_assign for $t, $t } | 
|  | )+) | 
|  | } | 
|  |  | 
|  | rem_assign_impl! { usize u8 u16 u32 u64 u128 isize i8 i16 i32 i64 i128 f16 f32 f64 f128 } |