Closed
Description
If I modify a file in the rustc source code (src/librustc_c*
), ./x.py test
is smart enough to rerun all tests. However, if I modify something in src/tools/compiletest
, then ./x.py test
still ignores the tests that completed successfully in the previous run.
Steps to reproduce:
- Run tests:
git clone git@github.com:rust-lang/rust.git
./x.py test
- Modify
compiletest
. For example, replace theexpected_output_path
function inruntest.rs
with:
fn expected_output_path(&self, _kind: &str) -> PathBuf {
panic!();
}
- Run tests again:
./x.py test
Expected output: some tests fail.
Actual output:
Build completed successfully in 0:13:13