feat(test-support): Use test name for dir when running tests

[Muscraft]

In #11738, I made a test's "sandbox" folder based on its name instead of a generated number, which can change from run to run. #11812 reverted this change due to problems with path length limits on Windows. After getting frustrated with the generated folders while trying to debug a test recently, I decided that it would be a good idea to bring back name-based folders to platforms that support them.

[rustbot]

r? @ehuss

rustbot has assigned @ehuss.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

[weihanglo]

I'd like to call out @ehuss's comment #11738 (comment), though personally fine with this PR.

[Muscraft]

Not sure how I missed this when I was going back through that PR. Given that concern, it might make sense for me to bring this up in the next team meeting.

[ehuss]

There is some documentation in https://github.com/rust-lang/cargo/blob/HEAD/src/doc/contrib/src/tests/writing.md that would need to be updated.

[ehuss]

@rustbot author

I think @Muscraft indicated he could add some kind of symlink to the first test.
There's also the docs to be updated.

[weihanglo]

Ran this twice and and it failed

$ cargo t --test testsuite config_include::bad_format
    Finished `test` profile [unoptimized + debuginfo] target(s) in 0.80s
     Running tests/testsuite/main.rs (/Users/user/.cargo/build-cache/debug/deps/testsuite-da0240c1196e03cf)

running 1 test
test config_include::bad_format ... ok

test result: ok. 1 passed; 0 failed; 0 ignored; 0 measured; 3900 filtered out; finished in 0.11s
$ cargo t --test testsuite config_include::bad_format
    Finished `test` profile [unoptimized + debuginfo] target(s) in 0.74s
     Running tests/testsuite/main.rs (/Users/user/.cargo/build-cache/debug/deps/testsuite-da0240c1196e03cf)

running 1 test
test config_include::bad_format ... FAILED

failures:

---- config_include::bad_format stdout ----

thread 'config_include::bad_format' (91574537) panicked at crates/cargo-test-support/src/lib.rs:200:9:

failed running os::unix::fs::symlink(&self.dst, &self.src)
error: File exists (os error 17)
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace


failures:
    config_include::bad_format

test result: FAILED. 0 passed; 1 failed; 0 ignored; 0 measured; 3900 filtered out; finished in 0.00s

I am on Mac OS 15.7.1 [64-bit]

[Muscraft]

I figured out the issue, cargo normally rm -rf's the test directory and I wasn't doing that before I tried to create the symlink.

[rustbot]

This PR was rebased onto a different master commit. Here's a range-diff highlighting what actually changed.

Rebasing is a normal part of keeping PRs up to date, so no action is needed—this note is just to help reviewers.