x.py test: compile-fail-fulldeps failures are reported as if they are in compile-fail

[pnkfelix]

When I naively run x.py test --stage 1, I will get failures from tests that require a stage 2 compiler. This is expected behavior.

However, I also expect for the report of the failure to reflect the fact that the failing tests were in one of the subdirectories dedicated to such tests, such as compile-fail-fulldeps or run-pass-fulldeps.

Unfortunately, today it seems that if you do the naive run of x.py test --stage 1, the summary of the failing tests presents a very misleading picture:

failures:
    [compile-fail] compile-fail/proc-macro/derive-bad.rs
    [compile-fail] compile-fail/proc-macro/derive-still-gated.rs
    [compile-fail] compile-fail/proc-macro/expand-to-unstable-2.rs
    [compile-fail] compile-fail/proc-macro/expand-to-unstable.rs
    [compile-fail] compile-fail/proc-macro/import.rs
    [compile-fail] compile-fail/proc-macro/issue-37788.rs
    [compile-fail] compile-fail/proc-macro/issue-38586.rs
    [compile-fail] compile-fail/proc-macro/item-error.rs
    [compile-fail] compile-fail/proc-macro/load-panic.rs
    [compile-fail] compile-fail/proc-macro/no-macro-use-attr.rs
    [compile-fail] compile-fail/proc-macro/proc-macro-attributes.rs
    [compile-fail] compile-fail/proc-macro/resolve-error.rs
    [compile-fail] compile-fail/proc-macro/shadow.rs

test result: FAILED. 28 passed; 13 failed; 8 ignored; 0 measured

I can live with the fact that its using the prefix [compile-fail]; I don't know why it is, but if people want a short category there, fine.

But it really should also present the correct file path after that prefix (e.g. compile-fail-fulldeps/proc-macro/derive-bad.rs), both so that

• I get my reminder that I need to be more discerning when testing a stage 1 compiler, and
• Someone trying to find the test actually knows where to look for it.