What #[derive(Mul)] generates

NOTE: Div, Rem, Shr and Shl derives are fully equivalent to the Mul derive described below.

In contrast with deriving Add, there are two ways to derive Mul: scalar and structural.

Scalar implementation

By default, deriving Mul is quite different from deriving Add. It is not used to multiply two structs together. Instead, it will normally multiply a struct, which can have multiple fields, with a single primitive type (e.g. a u64). A new struct is then created with all the fields from the previous struct multiplied by this other value.

A simple way of explaining the reasoning behind this difference between Add and Mul deriving, is looking at arithmetic on meters. One meter can be added to one meter, to get two meters. Also, one meter times two would be two meters, but one meter times one meter would be one square meter. As this second case clearly requires more knowledge about the meaning of the type in question deriving for this is not implemented.

Structs

Deriving Mul for a struct with a single field multiplies its field with anything Multiplicable, producing the resulting struct.

# use derive_more::Mul;
#
#[derive(Mul)]
struct MyInt(i32);

#[derive(Mul)]
struct Point1D {
    x: i32,
}

This generates code equivalent to:

# use std::ops::Mul;
#
# struct MyInt(i32);
#
# struct Point1D {
#     x: i32,
# }
#
impl<Rhs> Mul<Rhs> for MyInt
where
    i32: Mul<Rhs, Output = i32>
{
    type Output = Self;

    fn mul(self, rhs: Rhs) -> Self::Output {
        match self {
            Self(self_0) => Self(Mul::mul(self_0, rhs)),
        }
    }
}

impl<Rhs> Mul<Rhs> for Point1D
where
    i32: Mul<Rhs, Output = i32>
{
    type Output = Self;

    fn mul(self, rhs: Rhs) -> Self::Output {
        match self {
            Self { x: self_0 } => Self { x: Mul::mul(self_0, rhs) },
        }
    }
}

The behaviour is slightly different for multiple fields, since the right hand side of the multiplication now needs the Copy trait.

# use derive_more::Mul;
#
#[derive(Mul)]
struct MyInts(i32, i32);

#[derive(Mul)]
struct Point2D {
    x: i32,
    y: i32,
}

This generates code equivalent to:

# use std::ops::Mul;
#
# struct MyInts(i32, i32);
#
# struct Point2D {
#     x: i32,
#     y: i32,
# }
#
impl<Rhs: Copy> Mul<Rhs> for MyInts
where
    i32: Mul<Rhs, Output = i32>
{
    type Output = Self;

    fn mul(self, rhs: Rhs) -> Self::Output {
        match self {
            Self(self_0, self_1) => {
                Self(Mul::mul(self_0, rhs), Mul::mul(self_1, rhs))
            }
        }
    }
}

impl<Rhs: Copy> Mul<Rhs> for Point2D
where
    i32: Mul<Rhs, Output = i32>
{
    type Output = Self;

    fn mul(self, rhs: Rhs) -> Self::Output {
        match self {
            Self { x: self_0, y: self_1 } => Self {
                x: Mul::mul(self_0, rhs),
                y: Mul::mul(self_1, rhs),
            },
        }
    }
}

Ignoring

Sometimes a struct needs to hold a field (most commonly PhantomData) that doesn't participate in a scalar Mul implementation. Such field could be ignored using the #[mul(skip)] attribute.

# use core::marker::PhantomData;
# use derive_more::Mul;
#
#[derive(Mul)]
struct TupleWithZst<T>(i32, #[mul(skip)] PhantomData<T>);

#[derive(Mul)]
struct StructWithZst<T> {
    x: i32,
    #[mul(skip)] // or #[mul(ignore)]
    _marker: PhantomData<T>,
}

Enums

Deriving scalar Mul implementation for enums is not (yet) supported.

Although it shouldn't be impossible no effort has been put into this yet.

Structural implementation

Specifying the #[mul(forward)] attribute generates a structural Mul implementation with the same semantics as Add: Multiplying respective fields together and producing a new value with these fields.

Structs

# use derive_more::Mul;
#
#[derive(Mul)]
#[mul(forward)]
struct MyInts(i32, i32);

#[derive(Mul)]
#[mul(forward)]
struct Point2D {
    x: i32,
    y: i32,
}

This generates code equivalent to:

# use std::ops::Mul;
#
# struct MyInts(i32, i32);
#
# struct Point2D {
#     x: i32,
#     y: i32,
# }
#
impl Mul for MyInts {
    type Output = Self;

    fn mul(self, rhs: Self) -> Self::Output {
        match (self, rhs) {
            (Self(self_0, self_1), Self(rhs_0, rhs_1)) => {
                Self(Mul::mul(self_0, rhs_0), Mul::mul(self_1, rhs_1))
            }
        }
    }
}

impl Mul for Point2D {
    type Output = Self;

    fn mul(self, rhs: Self) -> Self::Output {
        match (self, rhs) {
            (Self { x: self_0, y: self_1 }, Self { x: rhs_0, y: rhs_1 }) => {
                Self { x: Mul::mul(self_0, rhs_0), y: Mul::mul(self_1, rhs_1) }
            }
        }
    }
}

The behaviour is similar with more or less fields.

Ignoring

Sometimes a struct needs to hold a field (most commonly PhantomData) that doesn't participate in a structural Mul implementation. Such field could be ignored using the #[mul(skip)] attribute.

# use core::marker::PhantomData;
# use derive_more::Mul;
#
#[derive(Mul)]
#[mul(forward)]
struct TupleWithZst<T>(i32, #[mul(skip)] PhantomData<T>);

#[derive(Mul)]
#[mul(forward)]
struct StructWithZst<T> {
    x: i32,
    #[mul(skip)] // or #[mul(ignore)]
    _marker: PhantomData<T>,
}

Enums

For enums each variant can be Multiplied structurally in a similar way to another instance of the same variant. There's one big difference however: it returns a Result<EnumType>, because an error is returned when two different variants are Multiplied together.

# use derive_more::Mul;
#
#[derive(Mul)]
#[mul(forward)]
enum MixedInts {
    BigInt(i64),
    NamedSmallInts { x: i32, y: i32 },
    Unit,
}

This generates code equivalent to:

# use std::ops::Mul;
#
# enum MixedInts {
#     BigInt(i64),
#     NamedSmallInts { x: i32, y: i32 },
#     Unit,
# }
#
impl Mul for MixedInts {
    type Output = Result<Self, derive_more::BinaryError>;

    fn mul(self, rhs: Self) -> Self::Output {
        match (self, rhs) {
            (Self::BigInt(self_0), Self::BigInt(rhs_0)) => {
                Ok(Self::BigInt(Mul::mul(self_0, rhs_0)))
            }
            (Self::NamedSmallInts { x: self_0, y: self_1 },
             Self::NamedSmallInts { x: rhs_0, y: rhs_1 }) => {
                Ok(Self::NamedSmallInts {
                    x: Mul::mul(self_0, rhs_0),
                    y: Mul::mul(self_1, rhs_1),
                })
            }
            (Self::Unit, Self::Unit) => Err(derive_more::BinaryError::Unit(
                derive_more::UnitError::new("add"),
            )),
            _ => Err(derive_more::BinaryError::Mismatch(
                derive_more::WrongVariantError::new("add"),
            )),
        }
    }
}

Also note the Unit variant that throws a derive_more::UnitError when Multiplying it to itself.

Ignoring

Similarly to structs, enum fields could be ignored using the #[add(skip)] attribute.

# use derive_more::Mul;
#
struct I32(i32); // doesn't implement `Mul`

#[derive(Mul)]
#[mul(forward)]
enum MixedInts {
    TwoSmallInts(i32, #[mul(skip)] I32),
    NamedSmallInts {
        #[mul(skip)] // or #[mul(ignore)]
        x: I32,
        y: i32,
    },
}

NOTE: Ignoring all the fields of a variant or ignoring the variant itself is not allowed (results in a compilation error).