Testing

Tests live next to the code in #[cfg(test)] mod tests. We exercise basic ops, iteration in both directions, clone, equality, and a stress loop. Then we run the same tests under Miri, which interprets the program and flags every undefined-behavior trigger we managed to write past the type system.

1. 01
   A handful of unit tests: round-trip push/pop on each end, drain via iteration, mutate through IterMut, clone-equals-original. Together these touch every method at least once and walk both directions.

   #[cfg(test)]
   mod tests {
       use super::*;
   
       #[test]
       fn push_pop_front() {
           let mut l = LinkedList::new();
           l.push_front(1);
           l.push_front(2);
           assert_eq!(l.pop_front(), Some(2));
           assert_eq!(l.pop_front(), Some(1));
           assert_eq!(l.pop_front(), None);
       }
   
       #[test]
       fn iter_mut_doubles() {
           let mut l: LinkedList<i32> = (1..=4).collect();
           for x in l.iter_mut() { *x *= 2; }
           assert!(l.iter().copied().eq([2, 4, 6, 8]));
       }
   
       #[test]
       fn clone_eq() {
           let a: LinkedList<String> = (0..50).map(|i| i.to_string()).collect();
           let b = a.clone();
           assert_eq!(a, b);
       }
   }

2. 02
   Run the unit tests under Miri: cargo +nightly miri test. Miri interprets every operation and validates against the stacked-borrows / tree-borrows model. Any aliasing or provenance violation we wrote earns a stack trace.

   $ rustup toolchain install nightly
   $ rustup +nightly component add miri
   $ cargo +nightly miri test
      Compiling sixth v0.1.0
        Running unittests src/lib.rs
   running 3 tests
   test sixth::tests::push_pop_front ... ok
   test sixth::tests::iter_mut_doubles ... ok
   test sixth::tests::clone_eq ... ok
   
   test result: ok. 3 passed; 0 failed