Regressions compared to GNU coreutils 9.2, 9.3 and 9.4

[sylvestre]

Diffing the results from before and after https://github.com/uutils/coreutils/pull/4601 and https://github.com/uutils/coreutils/pull/4754 Some of them should be easy.

I noticed some regressions:

PASS -> FAIL:

• [x] tests/cp/cp-i.sh (https://github.com/uutils/coreutils/pull/4630)
• [ ] tests/misc/tee.sh
• [x] tests/mv/i-1.pl (https://github.com/uutils/coreutils/pull/4630)
• [x] tests/mv/i-5.sh (https://github.com/uutils/coreutils/pull/4823)
• [X] tests/misc/date-f
• [x] tests/mv/update (https://github.com/uutils/coreutils/pull/4796)
• [x] tests/misc/read-errors.sh
• [ ] tests/misc/write-errors.sh
• [x] tests/dd/bytes.sh
• [x] tests/install/strip-program.sh
• [x] tests/mv/mv-n.sh (https://github.com/uutils/coreutils/pull/5347)
• [x] tests/cp/cp-i.sh (https://github.com/uutils/coreutils/pull/5348)
• [x] tests/tail/inotify-dir-recreate.sh
• [x] tests/cp/debug.sh
• [x] tests/mv/dir2dir.sh (https://github.com/uutils/coreutils/pull/5104)
• [x] tests/cksum/sm3sum.pl

{PASS,FAIL} -> SKIP

• [x] tests/cp/sparse-2.sh, tests/cp/sparse-extents-2.sh, tests/cp/sparse-extents.sh, tests/cp/sparse-perf.sh
• [ ] tests/ls/ls-time.sh

New tests - PASS directly

• [x] tests/misc/runcon-compute.sh
• [x] tests/mv/no-copy.sh

New tests - FAIL

• [x] tests/cp/debug.sh (https://github.com/uutils/coreutils/pull/4909)
• [x] tests/du/apparent.sh (https://github.com/uutils/coreutils/pull/4991)
• [x] tests/cksum/cksum-base64.pl
• [x] tests/cksum/cksum-raw.sh
• [x] tests/misc/wc-total.sh (https://github.com/uutils/coreutils/pull/4695)

To run them: bash util/run-gnu-test.sh <test>

How i did: I downloaded the results gnu-full-result.json in the action. The one before the update to 9.2 and just did a diff -u

[tertsdiepraam]

These match up with the changelog, for example:

• ls can now use modification time ls-time
• cp, mv and install I think have debug output
• cksum can output in base64

[sylvestre]

cp sparse tests are SKIP because it is parsing the --debug output that we don't have:

$ cp --debug --reflink=never file.sparse_ file.sparse_.cp
'file.sparse_' -> 'file.sparse_.cp'
copy offload: unknown, reflink: no, sparse detection: SEEK_HOLE

[RenjiSann]

Updated test statuses as of todays' GNU test coverage Shall we create another issue for tracking added tests ?

[sylvestre]

sure :)

[mfontanaar]

As far as I can see, two out of the three remaining regressions in this issue have been fixed. Both of - tests/misc/tee.sh - tests/ls/ls-time.sh pass on my system (Linux).

Unfortunately, it seems like tests/misc/write-errors.sh is still broken.

[RenjiSann]

Since we are now trying to match the GNU coreutils 9.8, I don't think this issue is still relevant

[cakebaker]

Superseded by https://github.com/uutils/coreutils/issues/9127