Closed onkoe closed 1 month ago
It feels like the test harness is failing to detect the panic - you can see in stderr that the panic occurs as expected, but the test asserts that it didn't! Oddly enough, revision-bad.rs#bar is detected correctly.
It's actually because your username barrett
contains bar
, which the foo
revision is explicitly checking to make sure that it is not present in test run stderr. But that is present in the panic backtrace of test stderr because the panic backtrace contains a path including your username. In #131579 I'm fixing it by remapping the path prefix so the panic backtrace paths are predictable and does not include username info or intermediate folders.
Hey there! I'm experiencing a weird issue where the
tests/ui/meta/revision-bad.rs#foo
test fails on the latestmaster
branch commit when I run ./x test. The issue has occurred for me since at least1b6d674
and still occurs now (e6eb451
).Here are the logs:
It feels like the test harness is failing to detect the panic - you can see in
stderr
that the panic occurs as expected, but the test asserts that it didn't! Oddly enough,revision-bad.rs#bar
is detected correctly.rustc --version --verbose
set -L