24 KiB
Data Types
Constants
constants can be delared anywhere, convension to use all caps in const need const keyword, not always eval at compile time variables can only be assigned once (needs mut to assign more than once (need to be same type))
const SECONDS_PER_HOUR: i32 = 60 * 60;
Variables
variables are immutable by default variables can be inferred but sometimes needs explicit typing
let foo = 5;
need to add mut keyword to enable rewriting, generally avoid unless actually used
let mut bar = 6;
SHADOWING
Cannot have mutable shadows
allows for reuse of namespace instead of spaces_str and spaces_num
let spaces = " _ _ ";
let spaces = spaces.len();
// will output 5 instead of " _ _ " beacuse that is how long it is
// the shadow of spaces (first) wont be printed until the overshadow of spaces goes out of scope
println!("{spaces}"); // output: 5
not allowed shadow
let mut spaces = " _ _ ";
spaces = spaces.len();
cannot change type of variable once declared
Primitive Data Types
Scalars
Integers
u is for usigned integers i is for signed integers
number indicated how many bits it takes in memory
let z: i8; // takes up 8 bits, can store values from -128 to 127
let c: i16; // takes up 16 bits
let d: i32; // takes up 32 bits (default for integers)
let e: i64; // takes up 64 bits
let f: i128; // takes up 128 bits
let g: isize; // takes up x bits, depends on the system's architecture/cpu
let h: u8; // takes up 8 bits, unsigned version (only positive)
// can store values from 0 to 255
Integer Overflow
will reset to the lowest value ie i8 129 -> -126
let example_over_flow: i8 = 129;
behavor only in production mode dev mode will cause a panic and error out/tell you
Floats
better to use double point due to modern cpus where there is not much difference in speed
Single Point Float
takes up 32 bits
let a: f32 = 4.0;
Double Point Float
takes up 64 bits
let b: f64 = 2.01;
Integers Represented Differently
can represent values in hex, oct, bin or dec can hover with rust-analyzer extension to see value in dec
Dec with Reading Aid
value stored 1000 _ used to make easier to read
let i = 1_000;
Hexidecimal
value stored 255
let j = 0xff;
Octal
value stored 63
let k = 0o77;
Binary
value stored 13
let l = 0b1101;
Bytes
u8 only value stored 0x41 or 65
let m = b'A';
Numeric Operators / Basic Math
Numbers for reference
let x: i16 = 8;
let y: i16 = 5;
Addition
let sum = x + y; // result: 13
Subtraction
let difference = x - y; //result: 3
Multiplication
let product: i16;
product = x * y;
Division
let quotent = 45.1 / 54.2;
let truncated = x / y; // results in 1 (always rounds down)
Remainder
let remainder = x % y;
Booleans
must be explicity typed to true or false 0 or 1 not allowed even with let var: bool
let n = false;
let o = true;
&& is and operator
Char
must use single quotes and not "" otherwise will be inferred as string literal is stored as Unicode Scalar Value allowing for emoji, japanse char and other languages not supported by ASCII takes 4 bytes in size or 32 bits
let p = 'a';
Compound Types
multiple values into one type
Tuple
A general way of grouping multiple a number of values into one compound type types do not need to be the same in every position
let tup: (i32, f64, u8) = (500, 6.4, 1);
The variable tup has values written to it at initialization but it is not requried, order does not matter similar to a struct in c
Vaules must be destructed out of a tuple to be accessed inidivually, can use a pattern matching to the tuple
let (q, r, s) = tup;
This is called destructing becasue it breaks it into 3 parts
Can also be accessed with a .
INDEX STARTS AT 0
let t = tup.0; // t = 500
let u = tup.1; // u = 6.4
let v = tup.2; // v = 1
A Unit
This is a special value where a tuple has no values
let w: () = ();
This represents an empty type or an empty return type
Expressions will implicitly return a unit if they dont return anything else
Array
A collection of multiple values Must have every value be the same type, cannot mix and match Arrays must be a fixed length at initialization useful when you want a set number of values or is static
Values are in [] and seperated by ,
let xa = [1, 2, 3, 4, 5, 6];
Array located in stack same with above types
If you need your array/list to grow or shrink use a vector If unsure weather to use an Array or Vector pick a vector
Times where using an array is better
let months = ["January", "February", "March", "April", "May", "June", "July", "August", "September", "October", "November", "December"];
Accessing items in an array
let ya = xa[0]; //value is 1
Initializing an Array
let za: [i32; 5]; // allows for 5 32 bit signed integers inside
let aa = [i8; 6]; // allow for 6 8 bit signed integers inside
Invalid Array Elements
use std::io;
fn main() {
let a = [1, 2, 3, 4, 5];
// Input of a number
println!("Please enter an array index.");
let mut index = String::new();
io::stdin()
.read_line(&mut index)
.expect("Failed to read line");
// change into a integer
let index: usize = index
.trim()
.parse()
.expect("Index entered was not a number");
// access elemetn in array
let element = a[index];
println!("The value of the element at index {index} is: {element}");
}
this program would compile with not problems for example inputting 7 into the program this would cause a runtime error the program would output an error because it didnt get to the final line println! before exiting it casue the program to exit before attempting to access the invalid space this is a form of safe memory management that rust name
Complex Data Type
String Literal
This is a string literal it is hardcoded into a program Always immutable Fast and efficient, stored on the stack, property of being immuatable not of any real value
let s: &str = "hello";
String
This is a string that is stored on the heap, this can store data unkown (size, char, etc) to you at compile time Can be mutable, but must request space on the heap then return that memory to the heap, will be returned as soon as it is no longer valid (it calls the drop method from String) not as fast and efficient Example of a string being created form a string literal
let ab:String = String::from("hello");
String concatinization example
let mut s = String::from("hello");
s.push_str(", world!"); // push_str() appends a literal to a String
println!("{s}"); // This will print `hello, world!`
Structures
Custom data type that packages up multiple data types into a meaningful manner and call the collection something More similar to an object, can define related methods to them
Similar to tuples but have to name and define everything inside a struct, like a dictionary but with set names and order Dont need to know order just know key
Defining
Need struct keywork then name of struct, which should describe the significance of the gropued data
All values are seperated by commas these are called fields
general definition of the type created
struct User {
active: bool,
username: String,
email: String,
sign_in_count: u64,
}
Initiating
To use give a owning var name thne concretely define what each value is define the key: vaule pairs dont need to initate in the same order they were defined in
let mut user1 = User {
active: true,
username: String::from("someusername123"),
email: String::from("someone@example.com"),
sign_in_count: 1,
};
to access values from the struct the dotnotation is used
note the WHOLE struct must be mutable, rust does not allow for partial mutability
// user1 email field now is equal to the string example@mail.com
user1.email = String::from("example@mail.com");
Can build a struct with implicit values input by default
fn build_user (email: String, username: String) {
User {
active: true,
email: email,
username: username,
sign_in_count: 1,
}
}
Init Field Shorthand
This is useful when the param and the struct definition share the same name. This reduces the amount of time spent on repeating key:value pairs
fn build_user(email: String, username: String) -> User {
User {
active: true,
username,
email,
sign_in_count: 1,
}
}
Only works beacuse param share same name as field key
this is equivalent to username:username or email:email
Creating Instances from Other Instances with Struct Update Syntax
Often useful to do so, only need to change 1 value
Slow method
let user2 = User {
active: user1.active,
username: user1.username,
email: String::from("another@emial.com"),
sign_in_count: user.sign_in_count,
};
Using Update syntax .. this can be done a LOT Quicker
let user2 = User {
email: String::from("another@email.com"),
..user1
}
// user1 no longer completely valid
// can still use user1.email, .active and .sign_in_count
This specifies htat the fields not explicity set should be the same as the given instance This uses the = assignment operator and therefore a ownership move occurs with the ../update syntax
user1 would still be valid if both of the String types in were given new values
Tuple Structs
This is also allowed but not key:value pairs This still holds values in the field
struct RGBColour (i32, i32, i32);
struct Point (i32, i32, i32);
let black = RGBColour(0, 0, 0);
let origin = Point(0, 0, 0);
Unit Like Struct
This is similar to a unit () This holds no data in itself
Useful for when you need to implement a trait on some type but dont want to store data in the type itself
Delcaration
struct unit_like_type;
let using_unit_like_struct = unit_like_type; // instance of unit_like_type
No need for () in the delcaration
Structure Ownership
Want each instance of a struct to own the values inside so that the values inside are always valid unless specified
Can use references but need to take advantage of lifetimes which ensures that the reference is valid whilst the structure is valid
This is valid but compiler will ask for lifetime specifiers
struct User {
active: bool,
username: &str,
email: &str,
sign_in_count: u64,
}
In short use data times that are owned rather than references
Adding Increased Functionality of Structs with derived traits
print can do many different types of formatting Cant print out structs by default because there are so many options with or without braces, commas, should all fields be shown This will cause an error
struct Rectangle {
length: u32,
width: u32,
}
let rect1 = Rectangle {
length: 8,
width: 4,
};
println!("rect1 contains {}", rect1);
{} tell println to use Display by default because there is only one way to show many primitive data types
{var_name:?} this is for the format Debug {var_name:#?} this is for pretty printing in Debug format, good for larger structs
Debug is also not implemented for the struct and therefore not supported
#[derive(Debug)]
struct Rectangle {
length: u32,
width: u32,
}
// snip
println!("rect1 contains {rect1:?}"); // single line print, in debug format, output: rect1 contains Rectangle { length: 8, width: 4 }
println!("rect1 contains {rect1:#?}"); // pretty print in debug format, output: rect1 contains Rectangle {
// length: 8,
// width: 4,
// }
Another way to output pretty debug format by default is dbg! macro this prints out the file and line number as well of where it was called and returns the ownership of the value this prints to the stderr output stream this takes ownership of values compaired to println! prints to stdout output stream
example of using dbg
fn main() {
let scale = 2;
let rect1 = Rectangle {
width: dbg!(30 * scale),
height: 50,
};
dbg!(&rect1); // because it takes ownership need to pass in a reference
}
output [src/main.rs:10:16] 30 * scale = 60 [src/main.rs:14:5] &rect1 = Rectangle { width: 60, height: 50, }
Methods
Fucntions that are more closely related to structs
similar to functions decalred, param and output are all the same
run the code when the method is declared elsewhere
unlike functions they are defined in the context of a struct, an enum or a trait
first parameter is always self, which represents the instance of the struct that is is being called upon just like python methods
definition
struct Rectangle {
length: u32,
width: u32,
}
// implementation block for Rectangle used to define fn related to the struct
// put in this blcok so that the use case doesnt need to be searched
impl Rectangle {
// fn moved to here to that it has access to the instance with the self reference
// fn now closely related to the rect struct
// first param must be the type self: Self or &self which rust lets you shorthand
// self can be borrowed, mutably borrowed, or take ownership of self
// should always borrow unless need to transferownership or mutate the stored var
// &mut self for mutable version of selfs
// use self when you want to transform the self into something else
fn area (&self) -> u32 {
self.length * self.width
}
// this is an example of a mthod with exterior params requried
fn fit_into (&self, other: &Rectangle) {
self.length > other.length && self.width > other.width
}
}
useage
let rect1 = Rectangle {
length: 8,
width: 4,
}
println!("The area of the reactangle is {} square units",
// method syntax to call the area func
// notice dont need any param in, already has access to self reference
rect1.area()
);
this provides method syntax and dont have to repeat the structure they effect one impl can house all the methods for a struct, we can have tthe same method name as field name just differenitate with the use of () this is for a mthod not the field
this is used in getters where you want read only access to a struct, you can make the field pravate but the method public
Associated Functions
functions in the impl block are associated with the struct all functions within tthe impl block is associated functions
not all associated methods have a self reference and therefore arent methods
can be used in constructors or destructors, often constructors are what they are used for
new is often used for constructors, not a protected keyword
example
impl Rectangle {
// Self is only allowed within the impl scope, referes to what the scope is for, this determines the return type
fn square(side: i32) -> Self {
Self {
length = side,
width = side,
}
}
}
to call these types of functions use the :: operator, this is also used in namespaces
ex
let square1 = Rectangle::square(3);
you have have multiple impl blocks associated with a struct, is the same as hvaing one monolithic one (better for readibility) some use cases for multiple impl blocks
Enums
allows for defining a type determined by its possible variants can only be one variant at a time
definition
enum IpAddrKind {
V4, // possible variant value
V6, // possible variant value
}
initiation/storing enum value
let six = IpAddrKind::V6;
let four = IpAddrKind::V4;
these need to be namespaced using the :: operator in order to tell the compiler which value you mean
other useage
fn route (ip_kind: IpAddrKind) {}
// can be called by using either varaint
route(IpAddrKind::V4);
route(IpAddrKind::V6);
we only know the kind of addr using the enum, to store the value can be done by using a struct ex
struct IpAddr {
kind: IpAddrKind,
address: String,
}
let homeIp = IpAddr {
kind: IpAddrKind::V4,
address: String::from("127.0.0.1"),
};
let loopback = IpAddr {
kind: IpAddrKind::V6,
address: String::from("::1"),
}
now each type of ip addr as an assocatiated value
with just an enum and no struct, where the values are directly attached this is more concise but communicates the same thing
enum IpAddr {
V4(String),
V6(String),
}
let homeIp = IpAddr::V4(String::from("127.0.0.1"));
let loopback = IpAddr::V6(String::from("::1"));
enums can have different types from another ipv4 will always have 4 different values from 0-255
enum IpAddr {
V4(u8, u8, u8, u8),
V6(String),
}
let homeIp = IpAddr::V4(127, 0, 0, 1);
let loopback = IpAddr::V6(String::from("::1"));
This encoding and differentiating of Ip adresses is so common that the standard library has an enum named IpAddr, this needs to be brought into scope in order to work
This stores two structs for V4 and V6
like this
struct Ipv4Addr {
// --snip--
}
struct Ipv6Addr {
// --snip--
}
enum IpAddr {
V4(Ipv4Addr),
V6(Ipv6Addr),
}
can put anything in a enum and enums can have unlimited types assocaited with them
could be nothing, i32, string, tuple for example
can implement behavior in relation to an enum
The Option Enum and Advantages over Null types
this is a specail case where the variants are nothing nas something this is part of the standard library can can be included
this should be handled so that the compiler can check for handling all types this then handle the case of what if it is empty
a feature of rust is excluding null references
if you try to use a null value as a not null value you get an error this is due to null or not null as pervasive and extrememly easy to make this kind of error
null is still useful for expressing a vlaue that is not present or not valid for some reason this is a problem of implementation
rust doesnt have nulls but can be expressed as the enum Option which is defined by the standard library as
enum Option<T> {
None,
Some(T),
}
don't need to explicitly bring into scope but can
can also call it by Some or None
is the generic type parameter
all of these replace the T generic type in the Option enum
let some_number = Some(5); // can be inferred due to the value being stored
let some_char = Some('e'); // can be inferred due to the value being stored
let absent_number: Option<i32> = None; // needs generic type specification for the None option
all valid T is any type
these are not the same and therefore cannot be added i8 and Option
the compiler will always ensure that i8 is a valid value the complier will not ensure that Option stores a valid value will have to check if it has a non-null value in the enum then convert it into a T type from Option type in order to use it
this eliminates the issue of it being assumed that is it a non-null type by being forced to handle the null variant everywhere where the value isn't a Option can be safely assumed to be a non null type
This has lots of useful values attached to it
In general you want code that only runs when you have some T and another code that runs when you have a None value The match expression is a control flow construct that an handle this, it is suited for enums, it will run different code depending on the enum state/value that it has that code can then be used inside the matching value
Match and Control Flow
A construct that allows you to compare against a series of patterns then execute code based on which pattern matches
Patterns can be made up of literal values, variable names, wildcards and many other things
The power of matches comes from the expressiveness in patterns and the fact that the compiler handles all possible cases
This is like a coin storer where the coin goes into the hole that fits first Match does this same concept
Coin Example This takes in a coin enum and returns the value in cents
enum Coin {
Penny,
Nickel,
Dime,
Quarter,
}
fn value_in_cents (coin: Coin) -> u8 {
match coin {
Coin::Penny => 1,
Coin::Nickel => 5,
Coin::Dime => 10,
Coin::Quarter => 25,
}
}
It is used by first saying match then an expression (like a variable without ;) This is like a if statement but if needs to evaluate to a Boolean for the condition
Match arms has a pattern then the code to run separated by a => , or a => {}, (optional comma) (for multiline/longer expressions) If the pattern doesn't match then the next arm is tried until a arm fits (compiler will check that all possibilities are accounted for)
Patterns that Bind to Values
Arms can bind to parts of values that match the pattern This can allow us to extract values out of enum variants
example quarters now have a state in which they were minted in
#[derive(Debug)] // so we can inspect the state in a minute
enum UsState {
Alabama,
Alaska,
// --snip--
}
enum Coin {
Penny,
Nickel,
Dime,
Quarter(UsState),
}
fn value_in_cents(coin: Coin) -> u8 {
match coin {
Coin::Penny => 1,
Coin::Nickel => 5,
Coin::Dime => 10,
// state is a bind
Coin::Quarter(state) => { // the binding for state will be the value UsState::Alaska
println!("State quarter from {state:?}!");
25
}
}
}
if the value to match was coin::Quarter(UsState::Alaska) then it would print out State quarter from Alaska because that is the only pattern that matches
Matching with Option T
this can be done
here is an example where there is a function that adds 1 to a non-null type and returns a null type if it is null
fn plus_one(x: Option<i32>) -> Option<i32> {
match x {
None => None, // the null part
// i binds to the value inside Option i32
Some(i) => Some(i + 1), // the non-null part
}
}
let five = Some(5);
let six = plus_one(five);
let none = plus_one(None);
this handles both the None case and the not-null case
very common pattern in rust to combine enums and matches then binding to the value inside
very useful for when a range of data types are included
Matches are Exhaustive
Matches must cover all possibilities there must be an arm for each pattern that could possibly come into contact with it if you don't he compiler will give an error
all matches must exist therefore making matches exhaustive by default
Catch All Patterns and _ Placeholder
for example lets say you have a game where if you roll a 3 you add a fancy hat and if you roll a 7 you remove a fancy hat and all other rolls result in the player moving
it would look something like this
let dice_roll = 9;
match dice_roll {
// can do specific numbers as the pattern
3 => add_fancy_hat(),
7 => remove_fancy_hat(),
// this is allowed as a catch all this matches against all patterns not listed
// should go last in the case that it matches all patterns (including 3 and 7)
other => move_player(other),
}
fn add_fancy_hat() {}
fn remove_fancy_hat() {}
fn move_player(num_spaces: u8) {}
this uses a variable named other not a keyword this will match this catch all of other will satisfy the exhaustive requirements of match catch all must be last, rust will give a warning if this the wrong order and it will never run
if we want to catch all but not use the value do _ this matches to all values and does not blind to that value this is like ignoring all other values not found in this arm
if you want nothing to happen from the ignore all then return a unit type from the match
for example the arm _ => () does this, this should not be used in the case of a match trying to return something