Non-unique test names yield incorrect test output

[skogseth]

The following code

#[libtest2::main]
fn main() {}

#[libtest2::test]
fn foo(_: &libtest2::TestContext) {}

mod inner {
    #[libtest2::test]
    fn foo(_: &libtest2::TestContext) {}
}

gives the output

running 2 tests
test foo ... ok
test foo ... ok

test result: ok. 1 passed; 0 failed; 0 ignored; 0 filtered out; finished in 0.000s

The number of tests that passed is incorrectly listed as "1".

The changes in #165 will make this impossible to reproduce using the macros, as they will always give unique test names, but the issue will still remain with the "normal API".

[epage]

I'm less concerned about the test output than I am for the json output which it is based on as the primary focus is on stabilizing json output atm.

In this case though, I think I would consider this a development time error though we should

• make sure this is documented
• call this out in our json stabilization report