[Rust Guide] 11.4. Using should_panic to Check Panics

If you find this helpful, please like, bookmark, and follow. To keep learning along, follow this series.

11.4.1 Verifying Error-Handling Cases

In addition to verifying whether code returns the correct value, tests also need to verify whether code handles error cases as expected. For example, you can write a test to verify whether code panics under a specific condition.

Such tests need the extra should_panic attribute. For functions marked with it, if a panic occurs inside the function, the test passes; otherwise it fails.

For example:

pub struct Guess {
    value: i32,
}

impl Guess {
    pub fn new(value: i32) -> Guess {
        if value < 1 || value > 100 {
            panic!("Guess value must be between 1 and 100, got {value}.");
        }

        Guess { value }
    }
}

#[cfg(test)]
mod tests {
    use super::*;

    #[test]
    #[should_panic]
    fn greater_than_100() {
        Guess::new(200);
    }
}

• The Guess struct has a field named value of type i32. It provides an associated function new for creating a Guess instance, but only if the argument passed to new is between 1 and 100; otherwise it panics.
• The greater_than_100 test function passes a value greater than 100 to new. A panic should occur, so the test function is marked with the should_panic attribute, written as #[should_panic].

Test result:

$ cargo test
   Compiling guessing_game v0.1.0 (file:///projects/guessing_game)
    Finished `test` profile [unoptimized + debuginfo] target(s) in 0.58s
     Running unittests src/lib.rs (target/debug/deps/guessing_game-57d70c3acb738f4d)

running 1 test
test tests::greater_than_100 - should panic ... ok

test result: ok. 1 passed; 0 failed; 0 ignored; 0 measured; 0 filtered out; finished in 0.00s

   Doc-tests guessing_game

running 0 tests

test result: ok. 0 passed; 0 failed; 0 ignored; 0 measured; 0 filtered out; finished in 0.00s

Now let’s intentionally introduce a bug by removing the value > 100 check from new:

pub struct Guess {
    value: i32,
}

impl Guess {
    pub fn new(value: i32) -> Guess {
        if value < 1 || value > 100 {
            panic!("Guess value must be between 1 and 100, got {value}.");
        }

        Guess { value }
    }
}

#[cfg(test)]
mod tests {
    use super::*;

    #[test]
    #[should_panic]
    fn greater_than_100() {
        Guess::new(200);
    }
}

Now Guess::new(200); inside the test function will not panic. But because the function is marked with should_panic, a test that should have panicked but did not will fail:

$ cargo test
   Compiling guessing_game v0.1.0 (file:///projects/guessing_game)
    Finished `test` profile [unoptimized + debuginfo] target(s) in 0.62s
     Running unittests src/lib.rs (target/debug/deps/guessing_game-57d70c3acb738f4d)

running 1 test
test tests::greater_than_100 - should panic ... FAILED

failures:

---- tests::greater_than_100 stdout ----
note: test did not panic as expected

failures:
    tests::greater_than_100

test result: FAILED. 0 passed; 1 failed; 0 ignored; 0 measured; 0 filtered out; finished in 0.00s

error: test failed, to rerun pass `--lib`

11.4.2 Making should_panic More Precise

Sometimes tests that use should_panic can be a bit vague, because they only tell you whether the code panicked, even if the panic was not the one the programmer expected.

To make the test more precise, you can add an optional expected argument to should_panic. Then the program checks whether the failure message contains the specified text.

For example:

pub struct Guess {
    value: i32,
}

impl Guess {
    pub fn new(value: i32) -> Guess {
        if value < 1 {
            panic!(
                "Guess value must be greater than or equal to 1, got {value}."
            );
        } else if value > 100 {
            panic!(
                "Guess value must be less than or equal to 100, got {value}."
            );
        }

        Guess { value }
    }
}

#[cfg(test)]
mod tests {
    use super::*;

    #[test]
    #[should_panic(expected = "less than or equal to 100")]
    fn greater_than_100() {
        Guess::new(200);
    }
}

• The struct above has been slightly changed: the value < 1 and value > 100 cases in new now use two different panic messages.
• An expected argument is added to should_panic, and the text after = is the expected error message. The test passes only if the function panics and the panic message contains the expected text; otherwise it fails.

This program will definitely pass.

Using the same pattern, let’s manually introduce an error. For example, swap the panic messages for values less than 1 and greater than 100 in new:

pub struct Guess {
    value: i32,
}

impl Guess {
    pub fn new(value: i32) -> Guess {
        if value < 1 {
            panic!(
                "Guess value must be less than or equal to 100, got {value}."
            );
        } else if value > 100 {
            panic!(
                "Guess value must be greater than or equal to 1, got {value}."
            );
        }

        Guess { value }
    }
}

#[cfg(test)]
mod tests {
    use super::*;

    #[test]
    #[should_panic(expected = "less than or equal to 100")]
    fn greater_than_100() {
        Guess::new(200);
    }
}

Test result:

$ cargo test
   Compiling guessing_game v0.1.0 (file:///projects/guessing_game)
    Finished `test` profile [unoptimized + debuginfo] target(s) in 0.66s
     Running unittests src/lib.rs (target/debug/deps/guessing_game-57d70c3acb738f4d)

running 1 test
test tests::greater_than_100 - should panic ... FAILED

failures:

---- tests::greater_than_100 stdout ----
thread 'tests::greater_than_100' panicked at src/lib.rs:12:13:
Guess value must be greater than or equal to 1, got 200.
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace
note: panic did not contain expected string
      panic message: `"Guess value must be greater than or equal to 1, got 200."`,
 expected substring: `"less than or equal to 100"`

failures:
    tests::greater_than_100

test result: FAILED. 0 passed; 1 failed; 0 ignored; 0 measured; 0 filtered out; finished in 0.00s

error: test failed, to rerun pass `--lib`

The failure message shows that the test did panic, but the panic message did not contain the expected string less than or equal to 100. In this case, the panic message we actually received was Guess value must be greater than or equal to 1, got 200.. That gives us enough information to fix the bug.