Relocate tests in tests/ui

[reddevilmidzy]

Part of #133895

Moved tests from a top-level directory into a more appropriate subdirectory.
If there is anything else that could be improved, please let me know!

r? jieyouxu

[jieyouxu]

Thanks for the PR! So, the new locations and names for these tests look good -- however, I'd like to see some short doc comments on the test intentions (they were not very obvious to me).

Please also squash the commits into one.

[jieyouxu]

Discussion: I honestly have no idea what this test is actually trying to exercise. This goes all the way back to 2010 d6b7c96.

I assume this is trying to exercise some short-circuiting behavior of || and && but the test itself makes it hard to tell.

EDIT: the OG version of this Rust was

fn incr(mutable &int x) -> bool {
  x += 1;
  check (false);
  ret false;
}

fn main() {

  auto x = (1 == 2) || (3 == 3);
  check (x);

  let int y = 10;
  log x || incr(y);
  check (y == 10);

  if (true && x) {
    check (true);
  } else {
    check (false);
  }

}

so yeah, I believe this is indeed a || and && short-circuiting test. It has a no-arg closure, presumably because back then there must've been a case where the parser confused || and the no-arg closure form || EXPR or sth.

[jieyouxu]

What's amusing is that this is one of the listings of the archive of OG rust syntax listed at https://brson.github.io/archaea/lazy-and-or.html/.

EDIT: I realized you can actually see the dynamic evolution of this test on that listing!

[reddevilmidzy]

Thanks for the analysis and sharing that interesting link! Now I understand that this test verifies short-circuiting behavior of logical operators.

I'll update the test with clearer naming and comments to better express its purpose while preserving the original functionality.

[reddevilmidzy]

And thanks for the historical context! That's super helpful 😁

[jieyouxu]

Discussion: so this one is also introduced in the 2010 populate tree commit, and it kinda has been preserved all these years modulo some language changes. I believe this is some kind of "positive" smoke test of a tree of owned memory...?

Originally (like back in Rust 0.8), this was

tag list {
  cons(int,@list);
  nil;
}

fn main() {
  cons(10, @cons(11, @cons(12, @nil)));
}

This is also one of the OG rust syntax examples at https://brson.github.io/archaea/list.html.

[jieyouxu]

Remark: yeah, this test is because unary negation on an owned string didn't use to properly typeck, back in de4b383.

[jieyouxu]

One nit, otherwise looks good

@rustbot author

(Sorry oli, I had to do some more archaeology to figure out what the original tests were doing)

[rustbot]

Reminder, once the PR becomes ready for a review, use @rustbot ready.

[reddevilmidzy]

@rustbot ready

[jieyouxu]

Actually on another review pass I realized I have a few more nits.

[jieyouxu]

@rustbot author

[reddevilmidzy]

Thank you for the thorough review!

[jieyouxu]

Thanks

[jieyouxu]

@bors r+ rollup

[bors]

📌 Commit dd2d6b2 has been approved by jieyouxu

It is now in the queue for this repository.