b2sum: report malformed status input

[wondr-wclabs]

What changed

b2sum -c --status now still reports the structural error when the checksum input contains no properly formatted checksum lines.

The change is in the shared checksum validator: the no properly formatted checksum lines found diagnostic is no longer hidden by ChecksumVerbose::Status. Regular checksum mismatches under --status remain silent; only the “there was nothing valid to check” diagnostic is surfaced.

Why

GNU b2sum -c --status reports this case because it is not a per-file mismatch result, it is an invalid checksum-list input. The previous logic treated it like other status-mode output and suppressed the diagnostic, leaving only the exit code.

Fixes #12590.

[github-actions]

GNU testsuite comparison:

Skipping an intermittent issue tests/pr/bounded-memory (passes in this run but fails in the 'main' branch)
Note: The gnu test tests/rm/many-dir-entries-vs-OOM is now being skipped but was previously passing.

[wondr-wclabs]

I inspected the failing macOS localization job. The failure is in test_touch::test_touch_changes_time_of_file_in_stdout:

touch: setting times of '-': Bad file descriptor

This PR only changes src/uucore/src/lib/features/checksum/validate.rs and tests/by-util/test_b2sum.rs, so that touch - failure is outside the changed code path. The native GNU comparison comment also only reports unrelated intermittent tail skips plus unrelated improvements (dd/no-allocate, seq-epipe).

I do not see a b2sum/checksum regression to fix here. If the macOS touch - failure repeats on main/other PRs, it looks like it should be handled separately from this checksum-status-input change.

[sylvestre]

The Validation section is useless, please don't add it in the future

[wondr-wclabs]

Agreed. I removed the Validation section from the PR body.

For uutils PR openings I will keep the description focused on the behavior change, compatibility rationale, and linked issue. I will only mention local commands later when they are directly useful for review, for example explaining a CI discrepancy or documenting a targeted check after a requested change.

[sylvestre]

@wondr-wclabs I would also appreciate you don't use a llm to reply to me...