Stephen Chung ef6dd9414a Allow if expressions.

2020-03-27 23:47:23 +08:00

59 KiB

Raw Blame History

Rhai - Embedded Scripting for Rust

Rhai is an embedded scripting language and evaluation engine for Rust that gives a safe and easy way to add scripting to any application.

Rhai's current features set:

no-std support
Easy integration with Rust functions and data types, supporting getter/setter methods
Easily call a script-defined function from Rust
Fairly efficient (1 million iterations in 0.75 sec on my 5 year old laptop)
Low compile-time overhead (~0.6 sec debug/~3 sec release for script runner app)
Easy-to-use language similar to JS+Rust
Support for overloaded functions
Compiled script is optimized for repeat evaluations
Very few additional dependencies (right now only num-traits to do checked arithmetic operations); for no_std builds, a number of additional dependencies are pulled in to provide for functionalities that used to be in std.

Note: Currently, the version is 0.11.0, so the language and API's may change before they stabilize.

Installation

Install the Rhai crate by adding this line to dependencies:

[dependencies]
rhai = "0.11.0"

Use the latest released crate version on crates.io:

[dependencies]
rhai = "*"

Crate versions are released on crates.io infrequently, so if you want to track the latest features, enhancements and bug fixes, pull directly from GitHub:

[dependencies]
rhai = { git = "https://github.com/jonathandturner/rhai" }

Beware that in order to use pre-releases (e.g. alpha and beta), the exact version must be specified in the Cargo.toml.

Optional features

Feature	Description
`no_stdlib`	Exclude the standard library of utility functions in the build, and only include the minimum necessary functionalities. Standard types are not affected.
`unchecked`	Exclude arithmetic checking (such as overflows and division by zero). Beware that a bad script may panic the entire system!
`no_function`	Disable script-defined functions if not needed.
`no_index`	Disable arrays and indexing features if not needed.
`no_float`	Disable floating-point numbers and math if not needed.
`no_optimize`	Disable the script optimizer.
`only_i32`	Set the system integer type to `i32` and disable all other integer types. `INT` is set to `i32`.
`only_i64`	Set the system integer type to `i64` and disable all other integer types. `INT` is set to `i64`.
`no_std`	Build for `no-std`. Notice that additional dependencies will be pulled in to replace `std` features.

By default, Rhai includes all the standard functionalities in a small, tight package. Most features are here to opt-out of certain functionalities that are not needed. Excluding unneeded functionalities can result in smaller, faster builds as well as less bugs due to a more restricted language.

Other cool projects to check out:

ChaiScript - A strong inspiration for Rhai. An embedded scripting language for C++ that I helped created many moons ago, now being led by my cousin.
Check out the list of scripting languages for Rust on awesome-rust

Examples

A number of examples can be found in the examples folder:

Example	Description
`arrays_and_structs`	demonstrates registering a new type to Rhai and the usage of arrays on it
`custom_types_and_methods`	shows how to register a type and methods for it
`hello`	simple example that evaluates an expression and prints the result
`no_std`	example to test out `no-std` builds
`reuse_scope`	evaluates two pieces of code in separate runs, but using a common `Scope`
`rhai_runner`	runs each filename passed to it as a Rhai script
`simple_fn`	shows how to register a Rust function to a Rhai `Engine`
`repl`	a simple REPL, interactively evaluate statements from stdin

Examples can be run with the following command:

cargo run --example name

The repl example is a particularly good one as it allows you to interactively try out Rhai's language features in a standard REPL (Read-Eval-Print Loop).

Example Scripts

There are also a number of examples scripts that showcase Rhai's features, all in the scripts folder:

Language feature scripts	Description
`array.rhai`	arrays in Rhai
`assignment.rhai`	variable declarations
`comments.rhai`	just comments
`for1.rhai`	for loops
`function_decl1.rhai`	a function without parameters
`function_decl2.rhai`	a function with two parameters
`function_decl3.rhai`	a function with many parameters
`if1.rhai`	if example
`loop.rhai`	endless loop in Rhai, this example emulates a do..while cycle
`op1.rhai`	just a simple addition
`op2.rhai`	simple addition and multiplication
`op3.rhai`	change evaluation order with parenthesis
`string.rhai`	string operations
`while.rhai`	while loop

Example scripts	Description
`speed_test.rhai`	a simple program to measure the speed of Rhai's interpreter (1 million iterations)
`primes.rhai`	use Sieve of Eratosthenes to find all primes smaller than a limit

To run the scripts, either make a tiny program or use of the rhai_runner example:

cargo run --example rhai_runner scripts/any_script.rhai

Hello world

To get going with Rhai, create an instance of the scripting engine and then call eval:

use rhai::{Engine, EvalAltResult};

fn main() -> Result<(), EvalAltResult>
{
    let mut engine = Engine::new();

    let result = engine.eval::<i64>("40 + 2")?;

    println!("Answer: {}", result);  // prints 42

    Ok(())
}

The type parameter is used to specify the type of the return value, which must match the actual type or an error is returned. Rhai is very strict here. There are two ways to specify the return type - turbofish notation, or type inference.

let result = engine.eval::<i64>("40 + 2")?;     // return type is i64, specified using 'turbofish' notation

let result: i64 = engine.eval("40 + 2")?;       // return type is inferred to be i64

let result = engine.eval<String>("40 + 2")?;    // returns an error because the actual return type is i64, not String

Evaluate a script file directly:

let result = engine.eval_file::<i64>("hello_world.rhai".into())?;       // 'eval_file' takes a 'PathBuf'

To repeatedly evaluate a script, compile it first into an AST (abstract syntax tree) form:

// Compile to an AST and store it for later evaluations
let ast = engine.compile("40 + 2")?;

for _ in 0..42 {
    let result: i64 = engine.eval_ast(&ast)?;

    println!("Answer #{}: {}", i, result);  // prints 42
}

Compiling a script file is also supported:

let ast = engine.compile_file("hello_world.rhai".into())?;

Rhai also allows working backwards from the other direction - i.e. calling a Rhai-scripted function from Rust - via call_fn:

// Define a function in a script and load it into the Engine.
// Pass true to 'retain_functions' otherwise these functions will be cleared at the end of consume()
engine.consume(true,
    r"
        // a function with two parameters: String and i64
        fn hello(x, y) {
            x.len() + y
        }

        // functions can be overloaded: this one takes only one parameter
        fn hello(x) {
            x * 2
        }
    ")?;

// Evaluate the function in the AST, passing arguments into the script as a tuple
// if there are more than one. Beware, arguments must be of the correct types because
// Rhai does not have built-in type conversions. If arguments of the wrong types are passed,
// the Engine will not find the function.

let result: i64 = engine.call_fn("hello", &ast, ( String::from("abc"), 123_i64 ) )?;
//                                              ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ put arguments in a tuple

let result: i64 = engine.call_fn("hello", 123_i64)?
//                                        ^^^^^^^ calls 'hello' with one parameter (no need for tuple)

Evaluate expressions only

Sometimes a use case does not require a full-blown scripting language, but only needs to evaluate expressions. In these cases, use the compile_expression and eval_expression methods or their _with_scope variants.

let result = engine.eval_expression::<i64>("2 + (10 + 10) * 2")?;

When evaluation expressions, no control-flow statement (e.g. if, while, for) is not supported and will be parse errors when encountered - not even variable assignments.

// The following are all syntax errors because the script is not an expression.
engine.eval_expression::<()>("x = 42")?;
let ast = engine.compile_expression("let x = 42")?;
let result = engine.eval_expression_with_scope::<i64>(&mut scope, "if x { 42 } else { 123 }")?;

Values and types

The following primitive types are supported natively:

Category	Equivalent Rust types	`type_of()` name
Integer number	`u8`, `i8`, `u16`, `i16`, `u32`, `i32` (default for `only_i32`), `u64`, `i64` (default)	same as type
Floating-point number (disabled with `no_float`)	`f32`, `f64` (default)	same as type
Boolean value	`bool`	`"bool"`
Unicode character	`char`	`"char"`
Unicode string	`String` (not `&str`)	`"string"`
Array (disabled with `no_index`)	`rhai::Array`	`"array"`
Dynamic value (i.e. can be anything)	`rhai::Dynamic`	the actual type
System number (current configuration)	`rhai::INT` (`i32` or `i64`), `rhai::FLOAT` (`f32` or `f64`)	same as type
Nothing/void/nil/null (or whatever you want to call it)	`()`	`"()"`

All types are treated strictly separate by Rhai, meaning that i32 and i64 and u32 are completely different - they even cannot be added together. This is very similar to Rust.

The default integer type is i64. If other integer types are not needed, it is possible to exclude them and make a smaller build with the only_i64 feature.

If only 32-bit integers are needed, enabling the only_i32 feature will remove support for all integer types other than i32, including i64. This is useful on some 32-bit systems where using 64-bit integers incurs a performance penalty.

If no floating-point is needed or supported, use the no_float feature to remove it.

There is a type_of function to detect the actual type of a value. This is useful because all variables are Dynamic.

// Use 'type_of()' to get the actual types of values
type_of('c') == "char";
type_of(42) == "i64";

let x = 123;
x.type_of();                // error - 'type_of' cannot use postfix notation
type_of(x) == "i64";

x = 99.999;
type_of(x) == "f64";

x = "hello";
if type_of(x) == "string" {
    do_something_with_string(x);
}

Value conversions

There is a to_float function to convert a supported number to an f64, and a to_int function to convert a supported number to i64 and that's about it. For other conversions, register custom conversion functions.

let x = 42;
let y = x * 100.0;              // error: cannot multiply i64 with f64
let y = x.to_float() * 100.0;   // works
let z = y.to_int() + x;         // works

let c = 'X';                    // character
print("c is '" + c + "' and its code is " + c.to_int());    // prints "c is 'X' and its code is 88"

Working with functions

Rhai's scripting engine is very lightweight. It gets most of its abilities from functions. To call these functions, they need to be registered with the Engine.

use rhai::{Engine, EvalAltResult};
use rhai::RegisterFn;                       // use `RegisterFn` trait for `register_fn`
use rhai::{Dynamic, RegisterDynamicFn};     // use `RegisterDynamicFn` trait for `register_dynamic_fn`

// Normal function
fn add(x: i64, y: i64) -> i64 {
    x + y
}

// Function that returns a Dynamic value
fn get_an_any() -> Dynamic {
    Box::new(42_i64)
}

fn main() -> Result<(), EvalAltResult>
{
    let mut engine = Engine::new();

    engine.register_fn("add", add);

    let result = engine.eval::<i64>("add(40, 2)")?;

    println!("Answer: {}", result);  // prints 42

    // Functions that return Dynamic values must use register_dynamic_fn()
    engine.register_dynamic_fn("get_an_any", get_an_any);

    let result = engine.eval::<i64>("get_an_any()")?;

    println!("Answer: {}", result);  // prints 42

    Ok(())
}

To return a Dynamic value from a Rust function, simply Box it and return it.

fn decide(yes_no: bool) -> Dynamic {
    if yes_no {
        Box::new(42_i64)
    } else {
        Box::new("hello world!".to_string())    // remember &str is not supported
    }
}

Generic functions

Generic functions can be used in Rhai, but separate instances for each concrete type must be registered separately:

use std::fmt::Display;

use rhai::{Engine, RegisterFn};

fn show_it<T: Display>(x: &mut T) -> () {
    println!("put up a good show: {}!", x)
}

fn main()
{
    let mut engine = Engine::new();

    engine.register_fn("print", show_it as fn(x: &mut i64)->());
    engine.register_fn("print", show_it as fn(x: &mut bool)->());
    engine.register_fn("print", show_it as fn(x: &mut String)->());
}

This example shows how to register multiple functions (or, in this case, multiple instances of the same function) to the same name in script. This enables function overloading based on the number and types of parameters.

Fallible functions

If a function is fallible (i.e. it returns a Result<_, Error>), it can be registered with register_result_fn (using the RegisterResultFn trait).

The function must return Result<_, EvalAltResult>. EvalAltResult implements From<&str> and From<String> etc. and the error text gets converted into EvalAltResult::ErrorRuntime.

use rhai::{Engine, EvalAltResult, Position};
use rhai::RegisterResultFn;     // use `RegisterResultFn` trait for `register_result_fn`

// Function that may fail
fn safe_divide(x: i64, y: i64) -> Result<i64, EvalAltResult> {
    if y == 0 {
        // Return an error if y is zero
        Err("Division by zero detected!".into())  // short-cut to create EvalAltResult
    } else {
        Ok(x / y)
    }
}

fn main()
{
    let mut engine = Engine::new();

    // Fallible functions that return Result values must use register_result_fn()
    engine.register_result_fn("divide", safe_divide);

    if let Err(error) = engine.eval::<i64>("divide(40, 0)") {
       println!("Error: {:?}", error);  // prints ErrorRuntime("Division by zero detected!", (1, 1)")
    }
}

Overriding built-in functions

Any similarly-named function defined in a script overrides any built-in function.

// Override the built-in function 'to_int'
fn to_int(num) {
    print("Ha! Gotcha! " + num);
}

print(to_int(123));     // what happens?

Custom types and methods

Here's an more complete example of working with Rust. First the example, then we'll break it into parts:

use rhai::{Engine, EvalAltResult};
use rhai::RegisterFn;

#[derive(Clone)]
struct TestStruct {
    field: i64
}

impl TestStruct {
    fn update(&mut self) {
        self.field += 41;
    }

    fn new() -> Self {
        TestStruct { field: 1 }
    }
}

fn main() -> Result<(), EvalAltResult>
{
    let mut engine = Engine::new();

    engine.register_type::<TestStruct>();

    engine.register_fn("update", TestStruct::update);
    engine.register_fn("new_ts", TestStruct::new);

    let result = engine.eval::<TestStruct>("let x = new_ts(); x.update(); x")?;

    println!("result: {}", result.field);           // prints 42

    Ok(())
}

All custom types must implement Clone. This allows the Engine to pass by value.

#[derive(Clone)]
struct TestStruct {
    field: i64
}

Next, we create a few methods that we'll later use in our scripts. Notice that we register our custom type with the Engine.

impl TestStruct {
    fn update(&mut self) {
        self.field += 41;
    }

    fn new() -> Self {
        TestStruct { field: 1 }
    }
}

let mut engine = Engine::new();

engine.register_type::<TestStruct>();

To use methods and functions with the Engine, we need to register them. There are some convenience functions to help with this. Below I register update and new with the Engine.

Note: Engine follows the convention that methods use a &mut first parameter so that invoking methods can update the value in memory.

engine.register_fn("update", TestStruct::update);   // registers 'update(&mut ts)'
engine.register_fn("new_ts", TestStruct::new);      // registers 'new'

Finally, we call our script. The script can see the function and method we registered earlier. We need to get the result back out from script land just as before, this time casting to our custom struct type.

let result = engine.eval::<TestStruct>("let x = new_ts(); x.update(); x")?;

println!("result: {}", result.field);               // prints 42

In fact, any function with a first argument (either by copy or via a &mut reference) can be used as a method-call on that type because internally they are the same thing: methods on a type is implemented as a functions taking an first argument.

fn foo(ts: &mut TestStruct) -> i64 {
    ts.field
}

engine.register_fn("foo", foo);

let result = engine.eval::<i64>("let x = new_ts(); x.foo()")?;

println!("result: {}", result);                     // prints 1

type_of() works fine with custom types and returns the name of the type. If register_type_with_name is used to register the custom type with a special "pretty-print" name, type_of() will return that name instead.

engine.register_type::<TestStruct>();
engine.register_fn("new_ts", TestStruct::new);
let x = new_ts();
print(type_of(x));                                 // prints "path::to::module::TestStruct"

engine.register_type_with_name::<TestStruct>("Hello");
engine.register_fn("new_ts", TestStruct::new);
let x = new_ts();
print(type_of(x));                                 // prints "Hello"

Getters and setters

Similarly, custom types can expose members by registering a get and/or set function.

#[derive(Clone)]
struct TestStruct {
    field: i64
}

impl TestStruct {
    fn get_field(&mut self) -> i64 {
        self.field
    }

    fn set_field(&mut self, new_val: i64) {
        self.field = new_val;
    }

    fn new() -> Self {
        TestStruct { field: 1 }
    }
}

let mut engine = Engine::new();

engine.register_type::<TestStruct>();

engine.register_get_set("xyz", TestStruct::get_field, TestStruct::set_field);
engine.register_fn("new_ts", TestStruct::new);

let result = engine.eval::<i64>("let a = new_ts(); a.xyz = 42; a.xyz")?;

println!("Answer: {}", result);                     // prints 42

Initializing and maintaining state

By default, Rhai treats each Engine invocation as a fresh one, persisting only the functions that have been defined but no global state. This gives each evaluation a clean starting slate. In order to continue using the same global state from one invocation to the next, such a state must be manually created and passed in.

In this example, a global state object (a Scope) is created with a few initialized variables, then the same state is threaded through multiple invocations:

use rhai::{Engine, Scope, EvalAltResult};

fn main() -> Result<(), EvalAltResult>
{
    let mut engine = Engine::new();

    // First create the state
    let mut scope = Scope::new();

    // Then push some initialized variables into the state
    // NOTE: Remember the system number types in Rhai are i64 (i32 if 'only_i32') ond f64.
    //       Better stick to them or it gets hard working with the script.
    scope.push("y", 42_i64);
    scope.push("z", 999_i64);
    scope.push("s", "hello, world!".to_string());   // remember to use 'String', not '&str'

    // First invocation
    engine.eval_with_scope::<()>(&mut scope, r"
        let x = 4 + 5 - y + z + s.len();
        y = 1;
    ")?;

    // Second invocation using the same state
    let result = engine.eval_with_scope::<i64>(&mut scope, "x")?;

    println!("result: {}", result);                 // prints 979

    // Variable y is changed in the script
    assert_eq!(scope.get_value::<i64>("y").expect("variable x should exist"), 1);

    Ok(())
}

Engine configuration options

Method	Description
`set_optimization_level`	Set the amount of script optimizations performed. See `script optimization`.
`set_max_call_levels`	Set the maximum number of function call levels (default 50) to avoid infinite recursion.

Rhai Language Guide

Comments

Comments are C-style, including '/* ... */' pairs and '//' for comments to the end of the line.

let /* intruder comment */ name = "Bob";

// This is a very important comment

/* This comment spans
   multiple lines, so it
   only makes sense that
   it is even more important */

/* Fear not, Rhai satisfies all nesting needs with nested comments:
   /*/*/*/*/**/*/*/*/*/
*/

Statements

Statements are terminated by semicolons ';' - they are mandatory, except for the last statement where it can be omitted.

A statement can be used anywhere where an expression is expected. The last statement of a statement block (enclosed by '{' .. '}' pairs) is always the return value of the statement. If a statement has no return value (e.g. variable definitions, assignments) then the value will be ().

let a = 42;             // normal assignment statement
let a = foo(42);        // normal function call statement
foo < 42;               // normal expression as statement

let a = { 40 + 2 };     // 'a' is set to the value of the statement block, which is the value of the last statement
//              ^ notice that the last statement does not require a terminating semicolon (although it also works with it)
//                ^ notice that a semicolon is required here to terminate the assignment statement; it is syntax error without it

4 * 10 + 2              // this is also a statement, which is an expression, with no ending semicolon because
                        // it is the last statement of the whole block

Variables

Variables in Rhai follow normal C naming rules (i.e. must contain only ASCII letters, digits and underscores '_').

Variable names must start with an ASCII letter or an underscore '_', must contain at least one ASCII letter, and must start with an ASCII letter before a digit. Therefore, names like '_', '_42', '3a' etc. are not legal variable names, but '_c3po' and 'r2d2' are. Variable names are also case sensitive.

Variables are defined using the let keyword. A variable defined within a statement block is local to that block.

let x = 3;          // ok
let _x = 42;        // ok
let x_ = 42;        // also ok
let _x_ = 42;       // still ok

let _ = 123;        // syntax error - illegal variable name
let _9 = 9;         // syntax error - illegal variable name

let x = 42;         // variable is 'x', lower case
let X = 123;        // variable is 'X', upper case
x == 42;
X == 123;

{
    let x = 999;    // local variable 'x' shadows the 'x' in parent block
    x == 999;       // access to local 'x'
}
x == 42;            // the parent block's 'x' is not changed

Constants

Constants can be defined using the const keyword and are immutable. Constants follow the same naming rules as variables.

const x = 42;
print(x * 2);       // prints 84
x = 123;            // syntax error - cannot assign to constant

Constants must be assigned a value, not an expression.

const x = 40 + 2;   // syntax error - cannot assign expression to constant

Numbers

Integer numbers follow C-style format with support for decimal, binary ('0b'), octal ('0o') and hex ('0x') notations.

The default system integer type (also aliased to INT) is i64. It can be turned into i32 via the only_i32 feature.

Floating-point numbers are also supported if not disabled with no_float. The default system floating-point type is i64 (also aliased to FLOAT).

'_' separators can be added freely and are ignored within a number.

Format	Type
`123_345`, `-42`	`i64` in decimal
`0o07_76`	`i64` in octal
`0xabcd_ef`	`i64` in hex
`0b0101_1001`	`i64` in binary
`123_456.789`	`f64`

Numeric operators

Numeric operators generally follow C styles.

Operator	Description	Integers only
`+`	Plus
`-`	Minus
`*`	Multiply
`/`	Divide (integer division if acting on integer types)
`%`	Modulo (remainder)
`~`	Power
`&`	Binary And bit-mask	Yes
`\|`	Binary Or bit-mask	Yes
`^`	Binary Xor bit-mask	Yes
`<<`	Left bit-shift	Yes
`>>`	Right bit-shift	Yes

let x = (1 + 2) * (6 - 4) / 2;  // arithmetic, with parentheses
let reminder = 42 % 10;         // modulo
let power = 42 ~ 2;             // power (i64 and f64 only)
let left_shifted = 42 << 3;     // left shift
let right_shifted = 42 >> 3;    // right shift
let bit_op = 42 | 99;           // bit masking

Unary operators

Operator	Description
`+`	Plus
`-`	Negative

let number = -5;
number = -5 - +5;

Numeric functions

The following standard functions (defined in the standard library but excluded if no_stdlib) operate on i8, i16, i32, i64, f32 and f64 only:

Function	Description
`abs`	absolute value
`to_float`	converts an integer type to `f64`

Floating-point functions

The following standard functions (defined in the standard library but excluded if no_stdlib) operate on f64 only:

Category	Functions
Trigonometry	`sin`, `cos`, `tan`, `sinh`, `cosh`, `tanh` in degrees
Arc-trigonometry	`asin`, `acos`, `atan`, `asinh`, `acosh`, `atanh` in degrees
Square root	`sqrt`
Exponential	`exp` (base e)
Logarithmic	`ln` (base e), `log10` (base 10), `log` (any base)
Rounding	`floor`, `ceiling`, `round`, `int`, `fraction`
Conversion	`to_int`
Testing	`is_nan`, `is_finite`, `is_infinite`

Strings and Chars

String and char literals follow C-style formatting, with support for Unicode ('\uxxxx' or '\Uxxxxxxxx') and hex ('\xxx') escape sequences.

Hex sequences map to ASCII characters, while '\u' maps to 16-bit common Unicode code points and '\U' maps the full, 32-bit extended Unicode code points.

Internally Rhai strings are stored as UTF-8 just like Rust (they are Rust Strings!), but there are major differences. In Rhai a string is the same as an array of Unicode characters and can be directly indexed (unlike Rust). This is similar to most other languages where strings are internally represented not as UTF-8 but as arrays of multi-byte Unicode characters. Individual characters within a Rhai string can also be replaced just as if the string is an array of Unicode characters. In Rhai, there is also no separate concepts of String and &str as in Rust.

Strings can be built up from other strings and types via the + operator (provided by the standard library but excluded if no_stdlib). This is particularly useful when printing output.

type_of() a string returns "string".

let name = "Bob";
let middle_initial = 'C';
let last = "Davis";

let full_name = name + " " + middle_initial + ". " + last;
full_name == "Bob C. Davis";

// String building with different types
let age = 42;
let record = full_name + ": age " + age;
record == "Bob C. Davis: age 42";

// Unlike Rust, Rhai strings can be indexed to get a character
// (disabled with 'no_index')
let c = record[4];
c == 'C';

ts.s = record;                          // custom type properties can take strings

let c = ts.s[4];
c == 'C';

let c = "foo"[0];                       // indexing also works on string literals...
c == 'f';

let c = ("foo" + "bar")[5];             // ... and expressions returning strings
c == 'r';

// Escape sequences in strings
record += " \u2764\n";                  // escape sequence of '❤' in Unicode
record == "Bob C. Davis: age 42 ❤\n";   // '\n' = new-line

// Unlike Rust, Rhai strings can be directly modified character-by-character
// (disabled with 'no_index')
record[4] = '\x58'; // 0x58 = 'X'
record == "Bob X. Davis: age 42 ❤\n";

The following standard functions (defined in the standard library but excluded if no_stdlib) operate on strings:

Function	Description
`len`	returns the number of characters (not number of bytes) in the string
`pad`	pads the string with an character until a specified number of characters
`append`	Adds a character or a string to the end of another string
`clear`	empties the string
`truncate`	cuts off the string at exactly a specified number of characters
`contains`	checks if a certain character or sub-string occurs in the string
`replace`	replaces a substring with another
`trim`	trims the string

Examples:

let full_name == " Bob C. Davis ";
full_name.len() == 14;

full_name.trim();
full_name.len() == 12;
full_name == "Bob C. Davis";

full_name.pad(15, '$');
full_name.len() == 15;
full_name == "Bob C. Davis$$$";

full_name.truncate(6);
full_name.len() == 6;
full_name == "Bob C.";

full_name.replace("Bob", "John");
full_name.len() == 7;
full_name = "John C.";

full_name.contains('C') == true;
full_name.contains("John") == true;

full_name.clear();
full_name.len() == 0;

Arrays

Arrays are first-class citizens in Rhai. Like C, arrays are accessed with zero-based, non-negative integer indices. Array literals are built within square brackets '[' ... ']' and separated by commas ','.

The Rust type of a Rhai array is rhai::Array.

type_of() an array returns "array".

Arrays are disabled via the no_index feature.

The following functions (defined in the standard library but excluded if no_stdlib) operate on arrays:

Function	Description
`push`	inserts an element at the end
`pop`	removes the last element and returns it (`()` if empty)
`shift`	removes the first element and returns it (`()` if empty)
`len`	returns the number of elements
`pad`	pads the array with an element until a specified length
`clear`	empties the array
`truncate`	cuts off the array at exactly a specified length (discarding all subsequent elements)

Examples:

let y = [1, 2, 3];      // array literal with 3 elements
y[1] = 42;

print(y[1]);            // prints 42

ts.list = y;            // arrays can be assigned completely (by value copy)
let foo = ts.list[1];
foo == 42;

let foo = [1, 2, 3][0];
foo == 1;

fn abc() {
    [42, 43, 44]        // a function returning an array literal
}

let foo = abc()[0];
foo == 42;

let foo = y[0];
foo == 1;

y.push(4);              // 4 elements
y.push(5);              // 5 elements

print(y.len());         // prints 5

let first = y.shift();  // remove the first element, 4 elements remaining
first == 1;

let last = y.pop();     // remove the last element, 3 elements remaining
last == 5;

print(y.len());         // prints 3

y.pad(10, "hello");     // pad the array up to 10 elements

print(y.len());         // prints 10

y.truncate(5);          // truncate the array to 5 elements

print(y.len());         // prints 5

y.clear();              // empty the array

print(y.len());         // prints 0

push and pad are only defined for standard built-in types. For custom types, type-specific versions must be registered:

engine.register_fn("push", |list: &mut Array, item: MyType| list.push(Box::new(item)) );

Comparison operators

Comparing most values of the same data type work out-of-the-box for standard types supported by the system.

However, if the no_stdlib feature is turned on, comparisons can only be made between restricted system types - INT (i64 or i32 depending on only_i32 and only_i64), f64 (if not no_float), string, array, bool, char.

42 == 42;           // true
42 > 42;            // false
"hello" > "foo";    // true
"42" == 42;         // false

Comparing two values of different data types, or of unknown data types, always results in false.

42 == 42.0;         // false - i64 is different from f64
42 > "42";          // false - i64 is different from string
42 <= "42";         // false again

let ts = new_ts();  // custom type
ts == 42;           // false - types are not the same

Boolean operators

Operator	Description
`!`	Boolean Not
`&&`	Boolean And (short-circuits)
`\|\|`	Boolean Or (short-circuits)
`&`	Boolean And (full evaluation)
`\|`	Boolean Or (full evaluation)

Double boolean operators && and || short-circuit, meaning that the second operand will not be evaluated if the first one already proves the condition wrong.

Single boolean operators & and | always evaluate both operands.

this() || that();   // that() is not evaluated if this() is true
this() && that();   // that() is not evaluated if this() is false

this() | that();    // both this() and that() are evaluated
this() & that();    // both this() and that() are evaluated

Compound assignment operators

let number = 5;
number += 4;        // number = number + 4
number -= 3;        // number = number - 3
number *= 2;        // number = number * 2
number /= 1;        // number = number / 1
number %= 3;        // number = number % 3
number <<= 2;       // number = number << 2
number >>= 1;       // number = number >> 1

The += operator can also be used to build strings:

let my_str = "abc";
my_str += "ABC";
my_str += 12345;

my_str == "abcABC12345"

`if` statements

if foo(x) {
    print("It's true!");
} else if bar == baz {
    print("It's true again!");
} else if ... {
        :
} else if ... {
        :
} else {
    print("It's finally false!");
}

All branches of an if statement must be enclosed within braces '{' .. '}', even when there is only one statement. Like Rust, there is no ambiguity regarding which if clause a statement belongs to.

if (decision) print("I've decided!");
//            ^ syntax error, expecting '{' in statement block

Like Rust, if statements can also be used as expressions, replacing the ? : conditional operators in other C-like languages.

let x = 1 + if true { 42 } else { 123 } / 2;
x == 22;

let x = if false { 42 };    // No else branch defaults to '()'
x == ();

`while` loops

let x = 10;

while x > 0 {
    print(x);
    if x == 5 { break; }    // break out of while loop
    x = x - 1;
}

Infinite `loop`

let x = 10;

loop {
    print(x);
    x = x - 1;
    if x == 0 { break; }    // break out of loop
}

`for` loops

Iterating through a range or an array is provided by the for ... in loop.

let array = [1, 3, 5, 7, 9, 42];

// Iterate through array
for x in array {
    print(x);
    if x == 42 { break; }
}

// The 'range' function allows iterating from first to last-1
for x in range(0, 50) {
    print(x);
    if x == 42 { break; }
}

`return`-ing values

return;             // equivalent to return ();

return 123 + 456;   // returns 579

Errors and `throw`-ing exceptions

All of Engine's evaluation/consuming methods return Result<T, rhai::EvalAltResult> with EvalAltResult holding error information. To deliberately return an error during an evaluation, use the throw keyword.

if some_bad_condition_has_happened {
    throw error;  // 'throw' takes a string to form the exception text
}

throw;            // empty exception text: ""

Exceptions thrown via throw in the script can be captured by matching Err(EvalAltResult::ErrorRuntime(reason, position)) with the exception text captured by the first parameter.

let result = engine.eval::<i64>(r#"
    let x = 42;

    if x > 0 {
        throw x + " is too large!";
    }
"#);

println!(result);   // prints "Runtime error: 42 is too large! (line 5, position 15)"

Functions

Rhai supports defining functions in script (unless disabled with no_function):

fn add(x, y) {
    return x + y;
}

print(add(2, 3));

Implicit return

Just like in Rust, an implicit return can be used. In fact, the last statement of a block is always the block's return value regardless of whether it is terminated with a semicolon ';'. This is different from Rust.

fn add(x, y) {
    x + y;          // value of the last statement (no need for ending semicolon) is used as the return value
}

fn add2(x) {
    return x + 2;   // explicit return
}

print(add(2, 3));   // prints 5
print(add2(42));    // prints 44

No access to external scope

Functions can only access their parameters. They cannot access external variables (even global variables).

let x = 42;

fn foo() { x }      // syntax error - variable 'x' doesn't exist

Passing arguments by value

Functions defined in script always take Dynamic parameters (i.e. the parameter can be of any type). It is important to remember that all arguments are passed by value, so all functions are pure (i.e. they never modify their arguments). Any update to an argument will not be reflected back to the caller. This can introduce subtle bugs, if not careful.

fn change(s) {  // 's' is passed by value
    s = 42;     // only a COPY of 's' is changed
}

let x = 500;
x.change();     // de-sugars to change(x)
x == 500;       // 'x' is NOT changed!

Global definitions only

Functions can only be defined at the global level, never inside a block or another function.

// Global level is OK
fn add(x, y) {
    x + y
}

// The following will not compile
fn do_addition(x) {
    fn add_y(n) {   // functions cannot be defined inside another function
        n + y
    }

    add_y(x)
}

Functions overloading

Functions can be overloaded and are resolved purely upon the function's name and the number of parameters (but not parameter types, since all parameters are the same type - Dynamic). New definitions overwrite previous definitions of the same name and number of parameters.

fn foo(x,y,z) { print("Three!!! " + x + "," + y + "," + z) }
fn foo(x) { print("One! " + x) }
fn foo(x,y) { print("Two! " + x + "," + y) }
fn foo() { print("None.") }
fn foo(x) { print("HA! NEW ONE! " + x) }    // overwrites previous definition

foo(1,2,3);     // prints "Three!!! 1,2,3"
foo(42);        // prints "HA! NEW ONE! 42"
foo(1,2);       // prints "Two!! 1,2"
foo();          // prints "None."

Members and methods

Properties and methods in a Rust custom type registered with the Engine can be called just like in Rust.

let a = new_ts();       // constructor function
a.field = 500;          // property access
a.update();             // method call

update(a);              // this works, but 'a' is unchanged because only a COPY of 'a' is passed to 'update' by VALUE

`print` and `debug`

The print and debug functions default to printing to stdout, with debug using standard debug formatting.

print("hello");         // prints hello to stdout
print(1 + 2 + 3);       // prints 6 to stdout
print("hello" + 42);    // prints hello42 to stdout
debug("world!");        // prints "world!" to stdout using debug formatting

Overriding `print` and `debug` with callback functions

When embedding Rhai into an application, it is usually necessary to trap print and debug output (for logging into a tracking log, for example).

// Any function or closure that takes an &str argument can be used to override
// print and debug
engine.on_print(|x| println!("hello: {}", x));
engine.on_debug(|x| println!("DEBUG: {}", x));

// Example: quick-'n-dirty logging
let mut log: Vec<String> = Vec::new();

// Redirect print/debug output to 'log'
engine.on_print(|s| log.push(format!("entry: {}", s)));
engine.on_debug(|s| log.push(format!("DEBUG: {}", s)));

// Evaluate script
engine.eval::<()>(script)?;

// 'log' captures all the 'print' and 'debug' output
for entry in log {
    println!("{}", entry);
}

Script optimization

Rhai includes an optimizer that tries to optimize a script after parsing. This can reduce resource utilization and increase execution speed. Script optimization can be turned off via the no_optimize feature.

For example, in the following:

{
    let x = 999;        // NOT eliminated - Rhai doesn't check yet whether a variable is used later on
    123;                // eliminated - no effect
    "hello";            // eliminated - no effect
    [1, 2, x, x*2, 5];  // eliminated - no effect
    foo(42);            // NOT eliminated - the function 'foo' may have side effects
    666                 // NOT eliminated - this is the return value of the block,
                        //                  and the block is the last one
                        //                  so this is the return value of the whole script
}

Rhai attempts to eliminate dead code (i.e. code that does nothing, for example an expression by itself as a statement, which is allowed in Rhai). The above script optimizes to:

{
    let x = 999;
    foo(42);
    666
}

Constants propagation is used to remove dead code:

const ABC = true;
if ABC || some_work() { print("done!"); }   // 'ABC' is constant so it is replaced by 'true'...
if true || some_work() { print("done!"); }  // since '||' short-circuits, 'some_work' is never called
if true { print("done!"); }                 // <- the line above is equivalent to this
print("done!");                             // <- the line above is further simplified to this
                                            //     because the condition is always true

These are quite effective for template-based machine-generated scripts where certain constant values are spliced into the script text in order to turn on/off certain sections. For fixed script texts, the constant values can be provided in a user-defined Scope object to the Engine for use in compilation and evaluation.

Beware, however, that most operators are actually function calls, and those functions can be overridden, so they are not optimized away:

const DECISION = 1;

if DECISION == 1 {          // NOT optimized away because you can define
    :                       //   your own '==' function to override the built-in default!
    :
} else if DECISION == 2 {   // same here, NOT optimized away
    :
} else if DECISION == 3 {   // same here, NOT optimized away
    :
} else {
    :
}

because no operator functions will be run (in order not to trigger side effects) during the optimization process (unless the optimization level is set to OptimizationLevel::Full). So, instead, do this:

const DECISION_1 = true;
const DECISION_2 = false;
const DECISION_3 = false;

if DECISION_1 {
    :                       // this branch is kept and promoted to the parent level
} else if DECISION_2 {
    :                       // this branch is eliminated
} else if DECISION_3 {
    :                       // this branch is eliminated
} else {
    :                       // this branch is eliminated
}

In general, boolean constants are most effective for the optimizer to automatically prune large if-else branches because they do not depend on operators.

Alternatively, turn the optimizer to OptimizationLevel::Full

Here be dragons!

Optimization levels

There are actually three levels of optimizations: None, Simple and Full.

None is obvious - no optimization on the AST is performed.
Simple (default) performs relatively safe optimizations without causing side effects (i.e. it only relies on static analysis and will not actually perform any function calls).
Full is much more aggressive, including running functions on constant arguments to determine their result. One benefit to this is that many more optimization opportunities arise, especially with regards to comparison operators.

An Engine's optimization level is set via a call to set_optimization_level:

// Turn on aggressive optimizations
engine.set_optimization_level(rhai::OptimizationLevel::Full);

When the optimization level is OptimizationLevel::Full, the Engine assumes all functions to be pure and will eagerly evaluated all function calls with constant arguments, using the result to replace the call. This also applies to all operators (which are implemented as functions). For instance, the same example above:

// When compiling the following with OptimizationLevel::Full...

const DECISION = 1;
                        // this condition is now eliminated because 'DECISION == 1'
if DECISION == 1 {      // is a function call to the '==' function, and it returns 'true'
    print("hello!");    // this block is promoted to the parent level
} else {
    print("boo!");      // this block is eliminated because it is never reached
}

print("hello!");        // <- the above is equivalent to this ('print' and 'debug' are handled specially)

Because of the eager evaluation of functions, many constant expressions will be evaluated and replaced by the result. This does not happen with OptimizationLevel::Simple which doesn't assume all functions to be pure.

// When compiling the following with OptimizationLevel::Full...

let x = (1 + 2) * 3 - 4 / 5 % 6;    // <- will be replaced by 'let x = 9'
let y = (1 > 2) || (3 <= 4);        // <- will be replaced by 'let y = true'

Function side effect considerations

All of Rhai's built-in functions (and operators which are implemented as functions) are pure (i.e. they do not mutate state nor cause side any effects, with the exception of print and debug which are handled specially) so using OptimizationLevel::Full is usually quite safe unless you register your own types and functions.

If custom functions are registered, they may be called (or maybe not, if the calls happen to lie within a pruned code block). If custom functions are registered to replace built-in operators, they will also be called when the operators are used (in an if statement, for example) and cause side-effects.

Function volatility considerations

Even if a custom function does not mutate state nor cause side effects, it may still be volatile, i.e. it depends on the external environment and not pure. A perfect example is a function that gets the current time - obviously each run will return a different value! The optimizer, when using OptimizationLevel::Full, assumes that all functions are pure, so when it finds constant arguments it will eagerly run execute the function call. This causes the script to behave differently from the intended semantics because essentially the result of the function call will always be the same value.

Therefore, avoid using OptimizationLevel::Full if you intend to register non-pure custom types and/or functions.

Subtle semantic changes

Some optimizations can alter subtle semantics of the script. For example:

if true {       // condition always true
    123.456;    // eliminated
    hello;      // eliminated, EVEN THOUGH the variable doesn't exist!
    foo(42)     // promoted up-level
}

foo(42)         // <- the above optimizes to this

Nevertheless, if the original script were evaluated instead, it would have been an error - the variable hello doesn't exist, so the script would have been terminated at that point with an error return.

In fact, any errors inside a statement that has been eliminated will silently disappear:

print("start!");
if my_decision { /* do nothing... */ }  // eliminated due to no effect
print("end!");

// The above optimizes to:

print("start!");
print("end!");

In the script above, if my_decision holds anything other than a boolean value, the script should have been terminated due to a type error. However, after optimization, the entire if statement is removed (because an access to my_decision produces no side effects), thus the script silently runs to completion without errors.

Turning off optimizations

It is usually a bad idea to depend on a script failing or such kind of subtleties, but if it turns out to be necessary (why? I would never guess), turn it off by setting the optimization level to OptimizationLevel::None.

let engine = rhai::Engine::new();

// Turn off the optimizer
engine.set_optimization_level(rhai::OptimizationLevel::None);

`eval` - or "How to Shoot Yourself in the Foot even Easier"

Saving the best for last: in addition to script optimizations, there is the ever-dreaded... eval function!

let x = 10;

fn foo(x) { x += 12; x }

let script = "let y = x;";      // build a script
script +=    "y += foo(y);";
script +=    "x + y";

let result = eval(script);      // <- look, JS, we can also do this!

print("Answer: " + result);     // prints 42

print("x = " + x);              // prints 10 - functions call arguments are passed by value
print("y = " + y);              // prints 32 - variables defined in 'eval' persist!

eval("{ let z = y }");          // to keep a variable local, use a statement block

print("z = " + z);              // error - variable 'z' not found

"print(42)".eval();             // nope - just like 'type_of' postfix notation doesn't work

Script segments passed to eval execute inside the current Scope, so they can access and modify everything, including all variables that are visible at that position in code! It is almost as if the script segments were physically pasted in at the position of the eval call.

let script = "x += 32";
let x = 10;
eval(script);                   // variable 'x' in the current scope is visible!
print(x);                       // prints 42

// The above is equivalent to:
let script = "x += 32";
let x = 10;
x += 32;
print(x);

For those who subscribe to the (very sensible) motto of "eval is evil", disable eval by overriding it, probably with something that throws.

fn eval(script) { throw "eval is evil! I refuse to run " + script }

let x = eval("40 + 2");         // 'eval' here throws "eval is evil! I refuse to run 40 + 2"

Or override it from Rust:

fn alt_eval(script: String) -> Result<(), EvalAltResult> {
    Err(format!("eval is evil! I refuse to run {}", script).into())
}

engine.register_result_fn("eval", alt_eval);

59 KiB Raw Blame History