Generates builder methods of every field of a struct. It is meant to be used on structs that
implement Default
. There is no separate builder struct generated and no need to call a
build()
method at the end or .unwrap()
.
This crate is used by the crate leptos-use
for the option structs that
can be passed to the various functions.
In your project folder run
cargo add default-struct-builder
It is very easy to use:
use default_struct_builder::DefaultBuilder;
#[derive(DefaultBuilder, Default)]
pub struct SomeOptions {
throttle: f64,
#[builder(into)]
offset: Option<f64>,
#[builder(skip)]
not_included: u32,
}
you can then use the struct like this:
let options = SomeOptions::default().offset(4.0);
assert_eq!(options.offset, Some(4.0));
assert_eq!(options.throttle, 0.0);
assert_eq!(options.not_included, 0);
The macro is ready to be used on generic structs.
use default_struct_builder::DefaultBuilder;
#[derive(DefaultBuilder, Default)]
pub struct SomeOptions<T>
where
T: Default,
{
some_field: T,
}
All doc comments on fields are directly passed on to their generated setter methods.
The derive macro generates the following code:
impl SomeOptions {
// setter methods are given that consume `self` and return a new `Self` with the field value changed
pub fn throttle(self, value: f64) -> Self {
Self {
throttle: value,
..self
}
}
// because `into` was specified this method is generic and calls `.into()` when setting the value
pub fn offset<T>(self, value: T) -> Self
where
T: Into<Option<f64>>,
{
Self {
offset: value.into(),
..self
}
}
// no method for field `not_included` because `skip` was specified
}
In the case of a generic field the generated method is a bit more complex because by calling the method the type of the type parameter can be different than before.
Let's look at the following example.
use default_struct_builder::DefaultBuilder;
#[derive(DefaultBuilder, Default)]
pub struct SomeOptions<T>
where
T: Default,
{
some_field: T,
other_field: i16,
}
impl SomeOptions<f32> {
pub fn new() -> Self {
Self {
some_field: 42.0,
other_field: 0,
}
}
}
This generates the setter method below.
impl<T> SomeOptions<T>
where
T: Default,
{
pub fn some_field<NewT>(self, value: NewT) -> SomeOptions<NewT>
where
NewT: Default,
{
SomeOptions::<NewT> {
some_field: value,
other_field: self.other_field,
}
}
}
fn main() {
let options = SomeOptions::new() // at first SomeOptions<f32>
.some_field("string"); // changed to SomeOptions<&str>
}
In cases where you don't want a generic field to be able to change the generic type you
can annotate it with keep_type
.
#[derive(DefaultBuilder)]
struct SomeOptions<T> {
#[builder(keep_type)]
the_field: T,
}
this will generate a standard builder method as if T
wasn't generic.
The macro detects if a field is a Box
(or Rc
or Arc
) and generates a builder method that
accepts the inner type (without Box
or Rc
or Arc
) and adds the outer type in the body.
In case it's a Box<dyn Trait>
the builder method will have an argument of type
impl Trait
. The same goes for Rc
and Arc
.
If you want to prevent this auto un-wrapping you can use the #[builder(keep_outer)]
attribute.
trait Test {}
#[derive(DefaultBuilder)]
struct SomeOptions {
the_field: Box<dyn Test>,
other_field: Rc<String>,
#[builder(keep_outer)]
keep: Box<String>,
}
This will generate the following code:
impl SomeOptions {
pub fn the_field(self, value: impl Test + 'static) -> Self {
Self {
the_field: Box::new(value),
..self
}
}
pub fn other_field(self, value: String) -> Self {
Self {
other_field: Rc::new(value),
..self
}
}
pub fn keep(self, value: Box<String>) -> Self {
Self {
keep: value,
..self
}
}
}
For more general purposes please check out the much more powerful
derive_builder
crate.