prop-check-rs

prop-check-rs is a property-based testing library written in Rust. It leverages functional programming concepts to efficiently generate and validate test data.

Read this in other languages: 日本語

What is Property-Based Testing?

Property-based testing is a testing methodology where instead of testing specific input values, you define properties that your program should satisfy and then verify these properties against a large number of randomly generated inputs. This approach helps discover edge cases that developers might not have anticipated.

Features

• Rich generators: Easily generate test data of various types
• Functional programming style: Composable API utilizing monads
• State-based testing: Support for state machine simulation
• Advanced customization: Define your own generators and properties

Installation

Add the following to your Cargo.toml:

[dependencies]
prop-check-rs = "0.0.862"

Basic Usage

1. Simple Property Test

Here's an example testing a property about list length:

use prop_check_rs::gen::Gens;
use prop_check_rs::prop::{for_all_gen, test_with_prop};
use prop_check_rs::rng::RNG;
use anyhow::Result;

#[test]
fn test_list_length_property() -> Result<()> {
    // Generator for a list of integers from 0 to 100
    let gen = Gens::list_of_n(10, Gens::choose_i32(0, 100));
    
    // Property: The list length is always 10
    let prop = for_all_gen(gen, |list| {
        list.len() == 10
    });
    
    // Test the property (max size 1, 100 test cases)
    test_with_prop(prop, 1, 100, RNG::new())
}

2. Choosing Values

use prop_check_rs::gen::Gens;
use prop_check_rs::prop::{for_all_gen, test_with_prop};
use prop_check_rs::rng::RNG;
use anyhow::Result;

#[test]
fn test_one_of() -> Result<()> {
    // Generator that selects one of the specified characters
    let gen = Gens::one_of_values(['a', 'b', 'c', 'x', 'y', 'z']);
    
    // Property: The selected character is always one of the specified characters
    let prop = for_all_gen(gen, move |value| {
        log::info!("value = {}", value);
        ['a', 'b', 'c', 'x', 'y', 'z'].contains(&value)
    });
    
    test_with_prop(prop, 1, 100, RNG::new())
}

3. Using Sized Generators

use prop_check_rs::gen::Gens;
use prop_check_rs::prop::{for_all_gen_for_size, test_with_prop};
use prop_check_rs::rng::RNG;
use anyhow::Result;

#[test]
fn test_sized_generator() -> Result<()> {
    let gen = Gens::one_of_values(['a', 'b', 'c', 'x', 'y', 'z']);
    
    // Generate lists based on size
    let prop = for_all_gen_for_size(
        move |size| Gens::list_of_n(size as usize, gen.clone()),
        move || {
            move |list| {
                // Verify that the list length matches the size
                log::info!("list = {:?}", list);
                true
            }
        },
    );
    
    // Max size 10, 100 test cases
    test_with_prop(prop, 10, 100, RNG::new())
}

Key Components

Gen

// Generator for integers from 1 to 100
let int_gen = Gens::choose_i32(1, 100);

// Generator that converts integers to strings
let string_gen = int_gen.map(|n| n.to_string());

Gens

Gens is a factory for creating various generators. It provides generators for:

• Basic types (integers, floating-point numbers, characters, booleans, etc.)
• Lists
• Optional values
• Choosing one from multiple options
• Probability-based selection

// Generators for basic types
let int_gen = Gens::one_i32();
let float_gen = Gens::one_f64();
let bool_gen = Gens::one_bool();

// Generator with a specified range
let range_gen = Gens::choose_i32(1, 100);

// Generator for lists
let list_gen = Gens::list_of_n(10, range_gen);

// Generator that chooses one from multiple options
let choice_gen = Gens::one_of_values(["apple", "banana", "orange"]);

// Generator based on probabilities
let weighted_gen = Gens::frequency_values([(1, "rare"), (5, "common"), (2, "uncommon")]);

Prop

Prop is a structure representing a property. A property defines a condition to verify against values generated by a generator.

// Property verifying that integers are always positive
let positive_prop = for_all_gen(Gens::choose_i32(1, 100), |n| n > 0);

// Combining multiple properties
let combined_prop = positive_prop.and(another_prop);

State<S, A>

State<S, A> is a monad representing a computation with state S that produces a value of type A. This allows composing computations while maintaining state.

// Get the state
let get_state = State::<i32, i32>::get();

// Set the state
let set_state = State::<i32, ()>::set(42);

// Modify the state
let modify_state = State::<i32, ()>::modify(|s| s + 1);

// Compose stateful computations
let computation = get_state.flat_map(|s| {
    if s > 0 {
        State::pure(s * 2)
    } else {
        State::pure(0)
    }
});

Advanced Usage Examples

Testing State Machines

The machine.rs module provides an example of simulating a state machine. Here's an example of a candy vending machine simulation:

// Input
enum Input {
    Coin,
    Turn,
}

// State machine
struct Machine {
    locked: bool,
    candies: i32,
    coins: i32,
}

// Simulating the state machine
let inputs = vec![Input::Coin, Input::Turn, Input::Coin, Input::Turn];
let simulation = Machine::simulate_machine(inputs);
let result = simulation.run(Machine { locked: true, candies: 5, coins: 10 });

Creating Custom Generators

You can create your own generators to generate domain-specific test data:

// Generator for valid email addresses
fn email_gen() -> Gen<String> {
    let username_gen = Gens::list_of_n(8, Gens::choose_char('a', 'z'))
        .map(|chars| chars.into_iter().collect::<String>());
    
    let domain_gen = Gens::one_of_values(["example.com", "test.org", "mail.net"]);
    
    username_gen.and_then(domain_gen, |username, domain| {
        format!("{}@{}", username, domain)
    })
}

// Usage example
let prop = for_all_gen(email_gen(), |email| {
    // Email validation logic
    email.contains('@')
});

Benchmarks

prop-check-rs includes optimizations for efficiently generating large amounts of test data. To run the benchmarks:

cargo bench

License

Licensed under either of

• Apache License, Version 2.0, (LICENSE-APACHE or https://www.apache.org/licenses/LICENSE-2.0)
• MIT license (LICENSE-MIT or https://opensource.org/licenses/MIT)

at your option.

Contribution

Unless you explicitly state otherwise, any contribution intentionally submitted for inclusion in the work by you, as defined in the Apache-2.0 license, shall be dual licensed as above, without any additional terms or conditions.