The proper ownership for "caching proxy" in Rust? - rust

I'd like to use Factory to build an object from the String and have multiple impls: 1) actual building and 2) caching (stores in-memory in HashMap). The problem is that in case #1 it have to pass the ownership and in case #2 HashMap owns the value and a reference can be returned only.
use std::collections::HashMap;
// product interface
pub trait TProduct {
fn get_title(&self) -> &String;
}
// and concrete impls
pub struct ConcreteProduct1 {
}
impl TProduct for ConcreteProduct1 {
// ...
}
pub struct ConcreteProduct2 {
}
impl TProduct for ConcreteProduct2 {
// ...
}
// factory interface
pub trait TProductFactory {
fn product_from_text(&mut self, text: String) -> Box<dyn TProduct>;
// QUESTION: should it be Box (required for ProductFactory) or &Box (required for ProductCachingProxy)?
}
// actual building factory
pub struct ProductFactory {
}
impl TProductFactory for ProductFactory {
fn product_from_text(&mut self, text: String) -> Box<dyn TProduct> {
//...
// depending on some conditions
Box::new(ConcreteProduct1::from_text(text)); // has to pass the ownership
// or
Box::new(ConcreteProduct2::from_text(text)); // has to pass the ownership
//...
}
}
// caching proxy
trait TProductCache {
fn put(&mut self, text: &String, product: Box<dyn TProduct>);
fn get(&self, text: &String) -> Option<&Box<dyn TProduct>>;
fn clear(&mut self);
}
struct InMemoryProductCache {
map: HashMap<String, Box<dyn TProduct>>
}
impl InMemoryProductCache {
fn new() -> Self {
return InMemoryProductCache {
map: HashMap::new()
}
}
}
impl TProductCache for InMemoryProductCache {
fn put(&mut self, text: &String, product: Box<dyn TProduct>) {
self.map.insert(text.to_string(), product);
}
fn get(&self, text: &String) -> Option<&Box<dyn TProduct>> {
return match self.map.get(text) {
Some(boxed_product) => Some(boxed_product), // have to pass a reference to let it still own the value
None => None
}
}
fn clear(&mut self) {
self.map.clear();
}
}
struct ProductCachingProxy {
product_factory: Box<dyn TProductFactory>,
cache: Box<dyn TProductCache>
}
impl ProductCachingProxy {
fn new_for_factory(product_factory: Box<dyn TProductFactory>, cache: Box<dyn TProductCache>) -> Self {
return ProductCachingProxy {
product_factory,
cache
}
}
}
impl TProductFactory for ProductCachingProxy {
fn product_from_text(&mut self, text: String) -> &Box<dyn TProduct> { // can't pass ownership
let boxed_product = match self.cache.get(&text) {
Some(found_boxed_product) => found_boxed_product,
_ => {
// delegate creation to wrapped TProductFactory impl (`product_factory`)
let boxed_product = self.product_factory.product_from_text(text.clone());
// ... and put to the cache
self.cache.put(&text, boxed_product);
&boxed_product
}
};
return boxed_product;
}
}
// QUESTION: should it be Box (required for ProductFactory) or &Box (required for ProductCachingProxy) to be returned from TProductFactory.fn product_from_text(&mut self, text: String) -> Box<dyn TProduct>; ?
If caching proxy to return a Box, how can it be created from a reference without copying/cloning (TProductCache.get(..))?

Replace Box with Rc (or Arc if you use threads). It provides shared ownership and suites both your cases with single signature. Another option is to use Cow that is a enum of owned and borrowed states.

Related

How to use the typestate pattern in other struct

I want to use the typestate pattern to define several states that allow some exclusive operations on each of them.
I'm using traits instead of an enum to allow further customizations.
So, I'm able to use this pattern until I try to include it inside a struct (the Session part) that is mutated when files are added, changed or removed.
trait IssueState {}
struct Open;
impl IssueState for Open {}
struct WIP {
elapsed_time: u32,
}
impl IssueState for WIP {}
struct Closed {
elapsed_time: u32,
}
impl IssueState for Closed {}
struct Issue<T: IssueState + ?Sized> {
state: Box<T>,
comments: Vec<String>,
}
impl<T: IssueState> Issue<T> {
pub fn comment<S: Into<String>>(&mut self, comment: S) -> &mut Self {
self.comments.push(comment.into());
self
}
}
impl Issue<Open> {
pub fn new() -> Self {
Self {
state: Box::new(Open),
comments: vec![],
}
}
pub fn start(self) -> Issue<WIP> {
Issue {
state: Box::new(WIP { elapsed_time: 0 }),
comments: self.comments,
}
}
}
impl Issue<WIP> {
pub fn work(&mut self, time: u32) -> &mut Self {
self.state.elapsed_time += time;
self
}
pub fn done(self) -> Issue<Closed> {
let elapsed_time = self.state.elapsed_time;
Issue {
state: Box::new(Closed { elapsed_time }),
comments: self.comments,
}
}
}
impl Issue<Closed> {
pub fn elapsed(&self) -> u32 {
self.state.elapsed_time
}
}
struct Session<T: IssueState> {
user: String,
current_issue: Issue<T>,
}
impl<T: IssueState> Session<T> {
pub fn new<S: Into<String>>(user: S, issue: Issue<T>) -> Self {
Self {
user: user.into(),
current_issue: issue,
}
}
pub fn comment<S: Into<String>>(&mut self, comment: S) {
self.current_issue.comment(comment);
}
}
impl Session<WIP> {
pub fn work(&mut self, time: u32) {
self.current_issue.work(time);
}
}
trait Watcher {
fn watch_file_create(&mut self);
fn watch_file_change(&mut self);
fn watch_file_delete(&mut self);
}
impl<T: IssueState> Watcher for Session<T> {
fn watch_file_create(&mut self) {
self.current_issue = Issue::<Open>::new();
}
fn watch_file_change(&mut self) {}
fn watch_file_delete(&mut self) {}
}
fn main() {
let open = Issue::<Open>::new();
let mut wip = open.start();
wip.work(10).work(30).work(60);
let closed = wip.done();
println!("Elapsed {}", closed.elapsed());
let mut session = Session::new("Reviewer", closed);
session.comment("It is OK");
session.watch_file_create();
}
Rust Playground (original)
Rust Playground (edited)
What can I do to fix the problems?
Is the typestate pattern limited to only some situations that do not depend a lot on external events? I mean, I'm trying to use it for processing events, but is it a dead end?, why?
Your Session has a Issue<dyn IssueState> member, but you want to implement its work method by calling Issue<WIP>'s work method. The compiler complains, because an Issue<dyn IssueState> is not (necessarily) a Issue<WIP> and so does not implement that method.

How do I dynamically create a struct from a string from input?

Given a situation where we receive inputs for some nodes type like 'nodeA' or 'nodeB', and we want to initialize structs with that same input. Is it possible without a gigantic switch block? These structs share similar behaviour (using a Trait for that) but with some differences.
pub trait Executable {
fn run(&self);
}
pub struct NodeA {}
impl Executable for NodeA {
fn run(&self) {}
}
pub struct NodeB {}
impl Executable for NodeB {
fn run(&self) {}
}
Flow:
User inputs 'nodeA'
Program initializes struct nodeA with some data
User inputs 'nodeB'
Program initializes struct nodeB with some data
...
To specify better, the final use case is reading a JSON file with all the nodes and respective params to be instantiated. Some of those nodes can come from external plugins, so the number of existing nodes can become very big.
For smaller, static number of nodes, I think a match - case construct is perfectly fine.
But if you have a larger number of nodes, or the available nodes is dynamically changing, I would implement something like this:
pub trait Executable {
fn run(&self);
}
pub struct NodeA {}
impl Executable for NodeA {
fn run(&self) {
println!("NodeA::run()");
}
}
pub struct NodeB {}
impl Executable for NodeB {
fn run(&self) {
println!("NodeB::run()");
}
}
pub trait Matcher {
fn try_match(&self, s: &str) -> Option<Box<dyn Executable>>;
}
pub struct NodeAMatcher;
pub struct NodeBMatcher;
impl Matcher for NodeAMatcher {
fn try_match(&self, s: &str) -> Option<Box<dyn Executable>> {
(s == "NodeA").then(|| Box::new(NodeA {}) as Box<dyn Executable>)
}
}
impl Matcher for NodeBMatcher {
fn try_match(&self, s: &str) -> Option<Box<dyn Executable>> {
(s == "NodeB").then(|| Box::new(NodeB {}) as Box<dyn Executable>)
}
}
struct MatcherRegistry {
matchers: Vec<Box<dyn Matcher>>,
}
impl MatcherRegistry {
fn new() -> Self {
Self { matchers: vec![] }
}
fn register_matcher(&mut self, matcher: impl Matcher + 'static) {
self.matchers.push(Box::new(matcher));
}
fn try_get_node(&self, s: &str) -> Option<Box<dyn Executable>> {
self.matchers
.iter()
.filter_map(|matcher| matcher.try_match(s))
.next()
}
fn try_execute(&self, s: &str) {
if let Some(node) = self.try_get_node(s) {
node.run();
} else {
println!("'{}' not found.", s);
}
}
}
fn main() {
let mut registry = MatcherRegistry::new();
registry.register_matcher(NodeAMatcher);
registry.register_matcher(NodeBMatcher);
registry.try_execute("NodeA");
registry.try_execute("NodeB");
registry.try_execute("NodeC");
}
NodeA::run()
NodeB::run()
'NodeC' not found.
Here, you have a factory pattern.
The structs NodeAMatcher and NodeBMatcher are factories for NodeA and NodeB. They can check if the input matches, and then create an Executable object.
Then, you collect all possible factories (or Matchers here) in a registry, here called MatcherRegistry. You can then, at runtime, add or remove matchers as you wish.
Of course, if you don't need to create a new object every time and the act of executing doesn't consume it, you can reduce the complexity a little by bypassing the factory pattern:
use std::collections::HashMap;
pub trait Executable {
fn run(&self);
}
pub struct NodeA {}
impl Executable for NodeA {
fn run(&self) {
println!("NodeA::run()");
}
}
pub struct NodeB {}
impl Executable for NodeB {
fn run(&self) {
println!("NodeB::run()");
}
}
struct ExecutableRegistry {
executables: HashMap<&'static str, Box<dyn Executable>>,
}
impl ExecutableRegistry {
fn new() -> Self {
Self {
executables: HashMap::new(),
}
}
fn register_executable(
&mut self,
command: &'static str,
executable: impl Executable + 'static,
) {
self.executables.insert(command, Box::new(executable));
}
fn try_execute(&self, s: &str) {
if let Some(node) = self.executables.get(s) {
node.run();
} else {
println!("'{}' not found.", s);
}
}
}
fn main() {
let mut registry = ExecutableRegistry::new();
registry.register_executable("NodeA", NodeA {});
registry.register_executable("NodeB", NodeB {});
registry.try_execute("NodeA");
registry.try_execute("NodeB");
registry.try_execute("NodeC");
}
Of course there exists a large mount of other variations of the same patterns. Which one you implement is up to you and your usecase.

Is this the right way to capture a string by value?

I'm making a functional builder that configures an object such as:
struct Person
{
name: String,
position: String
}
The builder itself keeps a list of boxed closures to be applied to the object when it needs to be constructed:
struct FunctionalBuilder<TSubject>
where TSubject : Default
{
actions: Vec<Box<dyn Fn(&mut TSubject) -> ()>>
}
impl<TSubject> FunctionalBuilder<TSubject>
where TSubject : Default
{
fn build(self) -> TSubject
{
let mut subj = TSubject::default();
for action in self.actions
{
(*action)(&mut subj);
}
subj
}
}
The idea being that one can aggregate this builder and then customize it for an object such as Person. Now, let's say I want to have a builder method called() that takes a name and saves the assignment of the name in the closure. I implement it as follows:
impl PersonBuilder
{
pub fn called(mut self, name: &str) -> PersonBuilder
{
let value = name.to_string();
self.builder.actions.push(Box::new(move |x| {
x.name = value.clone();
}));
self
}
}
Is this the right way of doing things? Is there a better way that avoids the temporary variable and clone() call?
Complete working example:
#[derive(Debug, Default)]
struct Person {
name: String,
position: String,
}
struct FunctionalBuilder<TSubject>
where
TSubject: Default,
{
actions: Vec<Box<dyn Fn(&mut TSubject) -> ()>>,
}
impl<TSubject> FunctionalBuilder<TSubject>
where
TSubject: Default,
{
fn build(self) -> TSubject {
let mut subj = TSubject::default();
for action in self.actions {
(*action)(&mut subj);
}
subj
}
fn new() -> FunctionalBuilder<TSubject> {
Self {
actions: Vec::new(),
}
}
}
struct PersonBuilder {
builder: FunctionalBuilder<Person>,
}
impl PersonBuilder {
pub fn new() -> Self {
PersonBuilder {
builder: FunctionalBuilder::<Person>::new(),
}
}
pub fn called(mut self, name: &str) -> PersonBuilder {
let value = name.to_string();
self.builder.actions.push(Box::new(move |x| {
x.name = value;
}));
self
}
pub fn build(self) -> Person {
self.builder.build()
}
}
pub fn main() {
let builder = PersonBuilder::new();
let me = builder.called("Dmitri").build();
println!("{:?}", me);
}
https://play.rust-lang.org/?version=stable&mode=debug&edition=2018&gist=27eb6283836a478d5c68aa025aa4698d
You already do it, value is owned by your closure, the problem is that you require the Fn trait. This mean that action (the function) need to be able to be called many times. This mean value need to be cloned to keep it valid inside the closure. The closure can't give away its ownership.
One way would be to use FnOnce, which would make it possible to remove the clone, but this means that the builder can only be used once. To do this, use actions: Vec<Box<dyn FnOnce(&mut TSubject) -> ()>>, and action(&mut subj);.
More:
"cannot move a value of type FnOnce" when moving a boxed function
Fn* closure traits implemented for Box<dyn Fn*>

Collection with Traits that return Self

I'm trying to have a collection of objects that implement a particular trait.
If I use a trait that returns a value, this works
use std::collections::BTreeMap;
struct World {
entities: Vec<usize>,
database: BTreeMap<usize, Box<ReadValue>>,
//database : BTreeMap<usize,Box<ReadEcs>>, // Doesn't work
}
struct SourceInputGateway {
entity_id: usize,
}
trait ReadValue {
fn read(&self) -> f32;
}
impl ReadValue for SourceInputGateway {
fn read(&self) -> f32 {
0.0
}
}
But if I want to return Self as a value then this doesn't work, either as a method template param or associated type
trait ReadEcs {
type T;
fn read(&self) -> &Self::T;
}
impl ReadEcs for SourceInputGateway {
type T = SourceInputGateway;
fn read(&self) -> &Self::T {
self
}
}
What I would like to do is have a map of types that implement ReadEcs, the concrete type of which is not important.
Further clarification edit
If I expand the example by adding
// Different sized type
struct ComputeCalculator {
entity_id : usize,
name : String,
}
impl ReadValue for ComputeCalculator {
fn read(&self) -> f32 {
1230.0
}
}
then I can do this
#[cfg(test)]
mod tests {
use super::*;
#[test]
fn read_write() {
let mut world = World::new();
world.database.insert(0,Box::new(SourceInputGateway{ entity_id : 1}));
world.database.insert(2,Box::new(ComputeCalculator{ entity_id : 2 , name : "foo".into() }));
for (k,ref v) in world.database {
let item : &Box<ReadValue> = v;
item.read();
}
}
}
but if I change or add a trait method that returns Self, I can't do this.
I'd like to understand a way to get around it without unsafe pointers.
I thought about this a little more and I think it's possible to solve this while retaining all the advantages of type safety and contiguous storage.
Defining a entity manager with pointers to storage
struct Entities {
entities: Vec<usize>,
containers: Vec<Box<Storage>>,
}
The components themselves that store data related to behaviour
struct Position {
entity_id: usize,
position: f32,
}
struct Velocity {
entity_id: usize,
velocity: f32,
}
We will have many instances of components, so we need a contiguous memory store, accessed by an index.
struct PositionStore {
storage: Vec<Position>,
}
struct VelocityStore {
storage: Vec<Velocity>,
}
trait Storage {
// Create and delete instances of the component
fn allocate(&mut self, entity_id: usize) -> usize;
// Interface methods that would correspond to a base class in C++
fn do_this(&self);
// fn do_that(&self);
}
The trait for the store implements an arena style storage plus the methods it would pass to to the components. This would likely be in the "System" part of an ECS but is left as an exercise for later.
I would like a hint as to how to pass a Fn() that constructs a custom object to the allocate() method. I've not figured that out yet.
impl Storage for PositionStore {
fn allocate(&mut self, entity_id: usize) -> usize {
self.storage.push(Position {
entity_id,
position: 0.0,
});
self.storage.len() - 1
}
fn run(&self) {
self.storage.iter().for_each(|item| { println!("{}",item.position); });
}
}
impl Storage for VelocityStore {
fn allocate(&mut self, entity_id: usize) -> usize {
self.storage.push(Velocity {
entity_id,
velocity: 0.0,
});
self.storage.len() - 1
}
fn do_this(&self) {
self.storage.iter().for_each(|item| { println!("{}",item.velocity); });
}
}
Some boiler plate.
impl Default for PositionStore {
fn default() -> PositionStore {
PositionStore {
storage: Vec::new(),
}
}
}
impl Default for VelocityStore {
fn default() -> VelocityStore {
VelocityStore {
storage: Vec::new(),
}
}
}
I think this could be thought about a little more. It stores the relationship between the storage of the components and their position
Instead of T::default(), you may want to pass a lambda function that has a specific initialisation for each of your components
impl Entities {
fn register<T>(&mut self) -> usize
where
T: Storage + Default + 'static,
{
self.containers.push(Box::new(T::default()));
self.containers.len() - 1
}
fn create<T>(&mut self, entity_id: usize, id: usize) -> usize {
self.containers[id].allocate(entity_id)
}
fn run_loop(&self) {
self.containers.iter().for_each(|x| x.do_this());
}
}
A test case to see if this works
#[cfg(test)]
mod tests {
use super::*;
#[test]
fn arbitary() {
let mut entities = Entities {
entities: Vec::new(),
containers: Vec::new(),
};
let velocity_store_id = entities.register::<VelocityStore>();
let position_store_id = entities.register::<PositionStore>();
let _ = entities.create::<Velocity>(123, velocity_store_id);
let _ = entities.create::<Velocity>(234, velocity_store_id);
let _ = entities.create::<Position>(234, position_store_id);
let _ = entities.create::<Position>(567, position_store_id);
entities.run_loop();
}
}

Is there an owned version of String::chars?

The following code does not compile:
use std::str::Chars;
struct Chunks {
remaining: Chars,
}
impl Chunks {
fn new(s: String) -> Self {
Chunks {
remaining: s.chars(),
}
}
}
The error is:
error[E0106]: missing lifetime specifier
--> src/main.rs:4:16
|
4 | remaining: Chars,
| ^^^^^ expected lifetime parameter
Chars doesn't own the characters it iterates over and it can't outlive the &str or String it was created from.
Is there an owned version of Chars that does not need a lifetime parameter or do I have to keep a Vec<char> and an index myself?
std::vec::IntoIter is an owned version of every iterator, in a sense.
use std::vec::IntoIter;
struct Chunks {
remaining: IntoIter<char>,
}
impl Chunks {
fn new(s: String) -> Self {
Chunks {
remaining: s.chars().collect::<Vec<_>>().into_iter(),
}
}
}
Playground link
Downside is additional allocation and a space overhead, but I am not aware of the iterator for your specific case.
Ouroboros
You can use the ouroboros crate to create a self-referential struct containing the String and a Chars iterator:
use ouroboros::self_referencing; // 0.4.1
use std::str::Chars;
#[self_referencing]
pub struct IntoChars {
string: String,
#[borrows(string)]
chars: Chars<'this>,
}
// All these implementations are based on what `Chars` implements itself
impl Iterator for IntoChars {
type Item = char;
#[inline]
fn next(&mut self) -> Option<Self::Item> {
self.with_mut(|me| me.chars.next())
}
#[inline]
fn count(mut self) -> usize {
self.with_mut(|me| me.chars.count())
}
#[inline]
fn size_hint(&self) -> (usize, Option<usize>) {
self.with(|me| me.chars.size_hint())
}
#[inline]
fn last(mut self) -> Option<Self::Item> {
self.with_mut(|me| me.chars.last())
}
}
impl DoubleEndedIterator for IntoChars {
#[inline]
fn next_back(&mut self) -> Option<Self::Item> {
self.with_mut(|me| me.chars.next_back())
}
}
impl std::iter::FusedIterator for IntoChars {}
// And an extension trait for convenience
trait IntoCharsExt {
fn into_chars(self) -> IntoChars;
}
impl IntoCharsExt for String {
fn into_chars(self) -> IntoChars {
IntoCharsBuilder {
string: self,
chars_builder: |s| s.chars(),
}
.build()
}
}
See also:
How can I store a Chars iterator in the same struct as the String it is iterating on?
Rental
You can use the rental crate to create a self-referential struct containing the String and a Chars iterator:
#[macro_use]
extern crate rental;
rental! {
mod into_chars {
pub use std::str::Chars;
#[rental]
pub struct IntoChars {
string: String,
chars: Chars<'string>,
}
}
}
use into_chars::IntoChars;
// All these implementations are based on what `Chars` implements itself
impl Iterator for IntoChars {
type Item = char;
#[inline]
fn next(&mut self) -> Option<Self::Item> {
self.rent_mut(|chars| chars.next())
}
#[inline]
fn count(mut self) -> usize {
self.rent_mut(|chars| chars.count())
}
#[inline]
fn size_hint(&self) -> (usize, Option<usize>) {
self.rent(|chars| chars.size_hint())
}
#[inline]
fn last(mut self) -> Option<Self::Item> {
self.rent_mut(|chars| chars.last())
}
}
impl DoubleEndedIterator for IntoChars {
#[inline]
fn next_back(&mut self) -> Option<Self::Item> {
self.rent_mut(|chars| chars.next_back())
}
}
impl std::iter::FusedIterator for IntoChars {}
// And an extension trait for convenience
trait IntoCharsExt {
fn into_chars(self) -> IntoChars;
}
impl IntoCharsExt for String {
fn into_chars(self) -> IntoChars {
IntoChars::new(self, |s| s.chars())
}
}
See also:
How can I store a Chars iterator in the same struct as the String it is iterating on?
There's also the owned-chars crate, which
provides an extension trait for String with two methods, into_chars and into_char_indices. These methods parallel String::chars and String::char_indices, but the iterators they create consume the String instead of borrowing it.
You could implement your own iterator, or wrap Chars like this (with just one small unsafe block):
// deriving Clone would be buggy. With Rc<>/Arc<> instead of Box<> it would work though.
struct OwnedChars {
// struct fields are dropped in order they are declared,
// see https://stackoverflow.com/a/41056727/1478356
// with `Chars` it probably doesn't matter, but for good style `inner`
// should be dropped before `storage`.
// 'static lifetime must not "escape" lifetime of the struct
inner: ::std::str::Chars<'static>,
// we need to box anyway to be sure the inner reference doesn't move when
// moving the storage, so we can erase the type as well.
// struct OwnedChar<S: AsRef<str>> { ..., storage: Box<S> } should work too
storage: Box<AsRef<str>>,
}
impl OwnedChars {
pub fn new<S: AsRef<str>+'static>(s: S) -> Self {
let storage = Box::new(s) as Box<AsRef<str>>;
let raw_ptr : *const str = storage.as_ref().as_ref();
let ptr : &'static str = unsafe { &*raw_ptr };
OwnedChars{
storage: storage,
inner: ptr.chars(),
}
}
pub fn as_str(&self) -> &str {
self.inner.as_str()
}
}
impl Iterator for OwnedChars {
// just `char` of course
type Item = <::std::str::Chars<'static> as Iterator>::Item;
fn next(&mut self) -> Option<Self::Item> {
self.inner.next()
}
}
impl DoubleEndedIterator for OwnedChars {
fn next_back(&mut self) -> Option<Self::Item> {
self.inner.next_back()
}
}
impl Clone for OwnedChars {
fn clone(&self) -> Self {
// need a new allocation anyway, so simply go for String, and just
// clone the remaining string
OwnedChars::new(String::from(self.inner.as_str()))
}
}
impl ::std::fmt::Debug for OwnedChars {
fn fmt(&self, f: &mut ::std::fmt::Formatter) -> ::std::fmt::Result {
let storage : &str = self.storage.as_ref().as_ref();
f.debug_struct("OwnedChars")
.field("storage", &storage)
.field("inner", &self.inner)
.finish()
}
}
// easy access
trait StringExt {
fn owned_chars(self) -> OwnedChars;
}
impl<S: AsRef<str>+'static> StringExt for S {
fn owned_chars(self) -> OwnedChars {
OwnedChars::new(self)
}
}
See playground
As copied from How can I store a Chars iterator in the same struct as the String it is iterating on?:
use std::mem;
use std::str::Chars;
/// I believe this struct to be safe because the String is
/// heap-allocated (stable address) and will never be modified
/// (stable address). `chars` will not outlive the struct, so
/// lying about the lifetime should be fine.
///
/// TODO: What about during destruction?
/// `Chars` shouldn't have a destructor...
struct OwningChars {
_s: String,
chars: Chars<'static>,
}
impl OwningChars {
fn new(s: String) -> Self {
let chars = unsafe { mem::transmute(s.chars()) };
OwningChars { _s: s, chars }
}
}
impl Iterator for OwningChars {
type Item = char;
fn next(&mut self) -> Option<Self::Item> {
self.chars.next()
}
}
Here is a solution without unsafe.
It provides the same effect as s.chars().collect::<Vec<_>>().into_iter(), but without the allocation overhead.
struct OwnedChars {
s: String,
index: usize,
}
impl OwnedChars {
pub fn new(s: String) -> Self {
Self { s, index: 0 }
}
}
impl Iterator for OwnedChars {
type Item = char;
fn next(&mut self) -> Option<Self::Item> {
// Slice of leftover characters
let slice = &self.s[self.index..];
// Iterator over leftover characters
let mut chars = slice.chars();
// Query the next char
let next_char = chars.next()?;
// Compute the new index by looking at how many bytes are left
// after querying the next char
self.index = self.s.len() - chars.as_str().len();
// Return next char
Some(next_char)
}
}
Together with a little bit of trait magic:
trait StringExt {
fn into_chars(self) -> OwnedChars;
}
impl StringExt for String {
fn into_chars(self) -> OwnedChars {
OwnedChars::new(self)
}
}
You can do:
struct Chunks {
remaining: OwnedChars,
}
impl Chunks {
fn new(s: String) -> Self {
Chunks {
remaining: s.into_chars(),
}
}
}

Resources