such as
#[test]
#[should_err]
fn test() -> Result<(), i32> {
Err(-1)
}
such as
#[test]
#[should_err]
fn test() -> Result<(), i32> {
Err(-1)
}
Perhaps it would be more logical to do the following:
#[test]
#[should_panic]
fn test(){
let mut val: Result<(), i32> = Err(-1);
//Run test and set `val`
val.unwrap();
}
I'm not a big fan of should_panic
, since it's really easy to have the test succeed because of the wrong panic.
Why no just do assert!(foo().is_err())
or assert_matches!(foo(), Err(-1));
? (The latter using matches::assert_matches - Rust)