mirror of
https://github.com/kuhyx/testsAndMisc.git
synced 2026-11-11 21:14:01 +01:00
Krzysztof kuhy Rudnicki
08f3c764b2
Eight checks assigned status="warning" unconditionally, so a check running
later in the same function overwrote the status="error" an earlier one had
set. The five other sites in this family already guard the assignment with
`if [[ $status != "error" ]]`, so this was an inconsistency, not a design.
Only half of it was load-bearing, and the two halves fail differently:
services_units.sh (4 sites) reports through report_and_fix, which repairs
ONLY on "error". A downgrade there meant no repair ran at all. The is-active
check is the LAST status assignment in each of the three functions, so it
could mask any error above it -- in check_periodic_systems, three separate
is-enabled checks. A timer that was both disabled and inactive, which is the
ordinary shape of a broken timer, was reported and then left broken.
services_hosts.sh (4 sites) gates its own repair on `status != "ok"`, so the
repair still ran. There the bug printed a real error as a yellow [!] warning
and understated the summary row.
Demonstrated rather than argued, against a simulated broken machine (timer
disabled AND inactive, unit files present):
without the guards: status=warning, repair ran 0 times
with the guards: status=error, repair ran 1 time
The seven assertions that previously pinned the broken behaviour now assert
the correct behaviour; they were rewritten first and failed, then the guards
made them pass. Two further cases cover the masking shape directly: a broken
guard-lib instance must still be repaired when a later check warns.
`--status --dry-run` output is still byte-identical to rev 6691cb26, because
this machine has no disabled timer and no missing pacman hook -- the buggy
path is never taken here, which is precisely why it went unnoticed. The A/B
above is the evidence the live diff cannot provide.
222 assertions across 10 files. services_units.sh holds 100% coverage;
services_hosts.sh rises from 86.05% to 93.02%. test_services_hosts.sh needed
splitting twice to stay under the 250-line cap (hosts_fix, resolved), and
stage_hosts_ok moved into the harness so both new files can use it.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01F3PzGc75gBcdxP3PkHXGcV
62 lines
5.1 KiB
JSON
62 lines
5.1 KiB
JSON
{
|
|
"intent": "Fix the status-downgrade bug in check_and_enable_services, where a later check assigning status=\"warning\" unconditionally could overwrite the status=\"error\" an earlier check had set — so a genuinely broken service was reported and then silently never repaired.",
|
|
"scope": [
|
|
"linux_configuration/scripts/periodic_background/lib/services_units.sh — 4 unguarded sites, the load-bearing ones",
|
|
"linux_configuration/scripts/periodic_background/lib/services_hosts.sh — 4 unguarded sites, cosmetic but wrong",
|
|
"lib/tests/test_services_units.sh — 7 pinning assertions converted to regression assertions",
|
|
"lib/tests/test_services_hosts.sh — 2 new masking regression cases",
|
|
"lib/tests/ — test_services_hosts_fix.sh and test_services_hosts_resolved.sh split out for the 250-line cap; stage_hosts_ok moved into the harness so both files can use it",
|
|
"Non-goal: the 5 sites that were already guarded correctly"
|
|
],
|
|
"changes": [
|
|
"All 8 bare `status=\"warning\"` assignments now read `if [[ $status != \"error\" ]]; then status=\"warning\"; fi`, which is the idiom services_apps.sh and services_browser.sh already used — so this makes the family internally consistent rather than inventing a convention.",
|
|
"The blast radius differs between the two files and only one is load-bearing. services_units.sh reports through report_and_fix, which repairs ONLY on \"error\", so a downgrade there meant no repair ran at all. services_hosts.sh gates its own repair on `status != \"ok\"`, so repair still ran; there the bug printed a real error as a yellow [!] warning and understated the summary row.",
|
|
"The three unit checks were the worst case because the is-active check is the LAST status assignment in each function, so it could overwrite any error set above it. In check_periodic_systems that is hosts-file-monitor.service, which sits after three separate is-enabled checks."
|
|
],
|
|
"verification": [
|
|
{
|
|
"command": "lib/tests/test_services_units.sh (before the fix)",
|
|
"result": "7 failures, as intended",
|
|
"evidence": "The pinning assertions were rewritten to assert correct behaviour first, and failed: 'disabled+inactive stays an error (want error, got warning)' and 'so the disabled timer IS repaired (no call matching /ran setup_midnight_shutdown.sh enable/)' for all three unit checks."
|
|
},
|
|
{
|
|
"command": "A/B against a simulated broken machine: timer disabled AND inactive, unit files present",
|
|
"result": "pass",
|
|
"evidence": "Without the fix: status=warning, repair ran 0 times — the timer stays dead. With the fix: status=error, repair ran 1 time — the timer is re-enabled. Run by reverting only the 4 guards in services_units.sh and restoring them afterwards."
|
|
},
|
|
{
|
|
"command": "lib/tests/run_all.sh",
|
|
"result": "pass",
|
|
"evidence": "222 assertions across 10 test files, 0 failures."
|
|
},
|
|
{
|
|
"command": "diff <(sudo bash <worktree@6691cb26>/…/check_and_enable_services.sh --status --dry-run) <(sudo bash …/check_and_enable_services.sh --status --dry-run)",
|
|
"result": "identical",
|
|
"evidence": "Still byte-identical modulo Date:. This machine has no disabled timer and no missing pacman hook, so the buggy path is not taken here — which is exactly why the fault went unnoticed. The A/B above is what demonstrates the behaviour change, since the live diff cannot."
|
|
},
|
|
{
|
|
"command": "bash meta/scripts/shell_coverage.sh lib/tests/run_all.sh <lib> 100",
|
|
"result": "pass on 6 of 8",
|
|
"evidence": "services_units.sh holds 100.00% (72/72) through the change. services_hosts.sh rose from 86.05% to 93.02% as the new masking cases exercised more branches; the 6 remaining lines are the kcov attribution artifact diagnosed in check-and-enable-services-coverage-2026-08-21.json. services_hosts_fix.sh 97.30% (2 pipeline continuation lines)."
|
|
},
|
|
{
|
|
"command": "grep -n '^\\s*status=\"warning\"$' lib/services_*.sh",
|
|
"result": "pass",
|
|
"evidence": "No unguarded site remains."
|
|
},
|
|
{
|
|
"command": "shellcheck lib/tests/*.sh lib/services_*.sh check_and_enable_services.sh; bash meta/scripts/check_file_length.sh --all",
|
|
"result": "pass",
|
|
"evidence": "Zero shellcheck findings, no suppression. No file over 250 lines; the hosts test file needed splitting twice to stay under it."
|
|
}
|
|
],
|
|
"risks": [
|
|
"This deliberately changes live enforcement behaviour: services that previously sat in a warning state without being repaired will now be repaired on the next hourly run. That is the point of the fix, but it means the next run on a machine with a disabled timer will actually execute a setup script rather than doing nothing.",
|
|
"A service whose repair script is broken will now retry every run instead of staying quiet at warning. err_missing_script already logs that case at journal error priority, so it surfaces rather than looping silently."
|
|
],
|
|
"rollback": [
|
|
"git revert this commit restores the unguarded assignments and the pinning tests together, so the suite stays green either way.",
|
|
"After rollback, confirm with the A/B: a disabled+inactive timer should report warning and run no repair."
|
|
]
|
|
}
|