21 releases (5 breaking)

0.6.0 May 27, 2024
0.5.6-alpha May 27, 2024
0.4.0 Apr 28, 2024
0.3.2 Apr 26, 2024
0.1.2 Apr 24, 2024

#558 in Rust patterns

Download history 432/week @ 2024-04-19 488/week @ 2024-04-26 134/week @ 2024-05-03 287/week @ 2024-05-10 178/week @ 2024-05-17 560/week @ 2024-05-24 37/week @ 2024-05-31 5/week @ 2024-06-07

62 downloads per month


192 lines


Anonymous struct implementation in rust.



cargo add structz

Create & access

use structz::*;

let age = 26;
let mut person = stru! {
    name: "John Doe",
    age,    // captures the value of the variable with the same name
    tags: vec!["developer", "rustacean"],

// immutable borrow
assert_eq!(field!(&person.name), &"John Doe");

// mutable borrow
*field!(&mut person.age) += 1;
assert_eq!(field!(&person.age), &27);

// consume the struct and get the field value
let tags = field!(person.tags);
assert_eq!(tags, vec!["developer", "rustacean"]);

// `person` cannot be used anymore.

NOTE: anonymous structs with the same fields but different field orders are considered structs of the same type.

use structz::*;

let rect1 = stru! { width: 1920, height: 1080 };
let rect2 = stru! { height: 1080, width: 1920 };
assert_eq!(rect1, rect2);

As argument

use structz::*;

fn print_person(person: stru_t! {name: &str, age: u8, tags: Vec<&str> }) {
        "{} is {} years old and has tags {:?}",

let person = stru! {
    tags: vec!["programmer", "artist"],
    name: "Alice",
    age: 30,

A better way is to use the named_args macro which helps you unpack the struct:

use structz::*;

fn print_person(name: &str, age: u8, tags: Vec<&str>) {
    println!("{} is {} years old and has tags {:?}", name, age, tags);

let person = stru! {
    tags: vec!["programmer", "artist"],
    name: "Alice",
    age: 30,

With the subseq() method by tuplez, you can get a sub-struct of the anonymous struct:

use structz::*;
use tuplez::TupleLike;

fn print_person(name: &str, age: u8) {
    println!("{} is {} years old", name, age);

let alice = stru! {
    jobs: "programmer",
    name: "Alice",
    age: 30,
    children: vec!["Bob"],

let bob = stru! {
    name: "Bob",
    parent: vec!["Alice", "John"],
    age: 7,
    grade: 1,

let empty = stru! {
    name: "**Empty**",
    age: 0,
print_person(empty.subseq());   // Of course it is a sub-struct of itself

As generic type

use stringz::ident;
use structz::*;

// `R1` and `R2` are "magic", used to indicate the position of the field in the structs,
// and these magic generic types will be automatically inferred by Rust.
// You should introduce a magic generic type for each field.
fn print_name_id<T, R1, R2>(any: &T)
    T: HasField<ident!(name), &'static str, R1>,
    T: HasField<ident!(id), usize, R2>,
    println!("{}", field!(&any.name));
    println!("{}", field!(&any.id));

let person = stru! {
    name: "John",
    age: 15,
    id: 1006,
    jobs: "Programmer",
let earth = stru! {
    name: "Earth",
    id: 3,
    galaxy: "Sol",
    satellites: vec!["Moon"],


The implementation of structz is based on stringz and tuplez.

First, macros sort the input fields in lexicographic order, which ensures that anonymous structs with the same fields but different field orders are of the same type.

Second, convert the field names into a specialized type consisting of a sequence of zero-sized types via stringz. Let's call them "field name types".

Finally, pack the field name type and the data type of each field, combine them into tuplez's Tuple.

Since the field names is effectively replaced with a zero-sized type, it cost you nothing:

use structz::*;

        stru_t! {
            age: u8,
            name: &'static str,
            tags: Vec<&'static str>,
    std::mem::size_of::<(u8, &'static str, Vec<&'static str>)>()


~21K SLoC