Testing

Rust has built-in capabilities to write tests.

Test Functions

A function is a test function when #[test] attribute is applied to it.

#[cfg(test)]
mod tests {
  #[test]
  fn it_works() {
    assert_eq!(2 + 2, 4);
  }
}

A function fails when it panics. Otherwise, it passes. Each test is run in a new thread.

Macros and attributes

Some useful testing macros:

#[cfg(test)] - placed on a module. Thanks to it, the test module is compiled only when cargo test is used. It saves time and space when building the code.
#[test] - required for a function to be runnable with cargo test
assert!(should_be_true)
assert_eq!(expected, actual)
assert_ne!(something, not_something)
#[should_panic] - it accepts an optional string that the panic should be thrown with. This helps to verify that the expected panic was thrown
#[ignore] - test will not run. We can still force cargo to run it using cargo test -- --ignored. It’ll run only the ignored tests.

Controlling Tests

cargo test parameters

--test-threads=n - allows to specify how many threads run the tests. We could set it to 1 to disable multithreading.
--show-output - by default print! does not show anything for passing tests. This setting changes that
passing single test function name will run just that function

Organization

Unit Tests

Unit tests should be placed in the same file as the code they’re testing. A separate testing module (with #[cfg(test)]) should be created to include all the tests.

Integration Tests

Integration tests should be separate from the tested code. There should be a tests directory next to src. Each file in that directory will become a separate crate.

These tests will be run with cargo test.

← Concurrency

Tips →