rustlings
rustlings copied to clipboard
Incorrect test case in /conversions/from_into.rs
The task is to implement From
trait for Person
. The steps are explicitly given:
// Steps:
// 1. If the length of the provided string is 0, then return the default of Person
// 2. Split the given string on the commas present in it
// 3. Extract the first element from the split operation and use it as the name
// 4. If the name is empty, then return the default of Person
// 5. Extract the other element from the split operation and parse it into a `usize` as the age
// If while parsing the age, something goes wrong, then return the default of Person
// Otherwise, then return an instantiated Person object with the results
There is a test:
#[test]
fn test_trailing_comma() {
let p: Person = Person::from("Mike,32,");
assert_eq!(p.name, "John");
assert_eq!(p.age, 30);
}
If the steps are followed exactly, Person::from("Mike,32,");
should return Person{ name: "Mike", age: 32 }
, not the default person.
- The length of the provided string is not zero
- Splitting on commas yields
["Mike", "32", ""]
- The first element is not empty, will become the name, "Mike"
- The second element is able to be parsed to
32usize
The instructions do not say to check the length of the Vec<&str>
created in step 2.
I would argue that the test is fine, but the problem description is incomplete. When I did this problem, I was also confused here, but eventually took the tests and the language "the other element" (step 5) to imply this was supposed to return a BadLen
error if we had anything other than two elements after splitting.
So I agree that there's a disconnect, but my inclination would be to modify the write-up instead of changing the test. Thoughts?