While preparing a potential update of util-linux (LP#2142862) and doing a test build,
I noticed that one test case started to fail (on all platforms).
"
1 tests of 337 FAILED
setpriv/landlock
make[5]: *** [Makefile:19529: check-local-tests] Error 1
...
dh_auto_test: error: make -j1 check ... returned exit code 2
"
The The failing sub-test is setpriv/landlock → partial-access-fail, and the "failure" is a string mismatch (stonking_s390x.txt:12227):
The test case seems to expect cp output from GNU coreutils, since it works with it.
It looks for:
cp: cannot create regular file '/dev/zero': Permission denied
But the coreutils-from-uutils cp produces in the same test case this:
cp: '/dev/null' -> '/dev/zero': Permission denied
Which leads to a failed test.
The chroot that is used while doing the test buildi (in PPA) for Ubuntu 26.10 has coreutils-from-uutils 0.0.0~ubuntu27 / rust-coreutils 0.8.0-0ubuntu4.
While rebuilding the same in PPA with -proposed disabled, the test is successful (now with rust-coreutils 0.8.0-0ubuntu3).
Since rust-coreutils is supposed to be a drop-in replacement, and esp. since this works on the previous version, I think this is a regression and should be solved in coreutils.
I'm cross referencing the Ubuntu Launchpad bug:
https://bugs.launchpad.net/bugs/2157011
While preparing a potential update of util-linux (LP#2142862) and doing a test build,
I noticed that one test case started to fail (on all platforms).
"
1 tests of 337 FAILED
setpriv/landlock
make[5]: *** [Makefile:19529: check-local-tests] Error 1
...
dh_auto_test: error: make -j1 check ... returned exit code 2
"
The The failing sub-test is setpriv/landlock → partial-access-fail, and the "failure" is a string mismatch (stonking_s390x.txt:12227):
The test case seems to expect cp output from GNU coreutils, since it works with it.
It looks for:
cp: cannot create regular file '/dev/zero': Permission denied
But the coreutils-from-uutils cp produces in the same test case this:
cp: '/dev/null' -> '/dev/zero': Permission denied
Which leads to a failed test.
The chroot that is used while doing the test buildi (in PPA) for Ubuntu 26.10 has coreutils-from-uutils 0.0.0~ubuntu27 / rust-coreutils 0.8.0-0ubuntu4.
While rebuilding the same in PPA with -proposed disabled, the test is successful (now with rust-coreutils 0.8.0-0ubuntu3).
Since rust-coreutils is supposed to be a drop-in replacement, and esp. since this works on the previous version, I think this is a regression and should be solved in coreutils.
I'm cross referencing the Ubuntu Launchpad bug:
https://bugs.launchpad.net/bugs/2157011