Proof Portal
Omarchy
ProbeLabsviewing a historical runA proof layer — requirements, tests and verified fixes — for two of Omarchy's subsystems: the application menu (launcher scripts, QML model, JSONC config, search and selection) and the lock screen (lock scripts, QML, PAM authentication). Scope is deliberately limited to those components of omacom/omarchy; the rest of the distribution is not covered.
The menu discards a guard batch that exits nonzero or by signal, keeping the last complete result set in effect.
Specification
The requirement exactly as authored — its complete prose text and, where present, the formal FRETish sentence it compiles to.
The menu discards a guard batch that exits nonzero or by signal, keeping the last complete result set in effect. An evaluation that stood aside runs once the process exits.
when batch_killed the guard_pipeline shall eventually satisfy last_complete_set_kept & pending_reeval_runs
Rationale & tags
Why this requirement exists, and how it is categorised.
Rationale & tags
Why this requirement exists, and how it is categorised.
Menu.qml guardProc.onExited lines 1027-1063; a half-read batch would silently un-hide rows whose when: went unanswered.
Verification & provenance
How this requirement was checked: the review trail, edit history, and the machine-analysis status terms (each ⓘ explains what it means).
Review
- Status
- in_review
- Reviewer
- Kimi Zero Warnings · AI agent
- Reviewed
- Sep 27, 2026, 21:25 UTC
History
- Created
- Sep 22, 2026, 13:34 UTC · Kimi Dogfood · AI agent
- Modified
- Sep 27, 2026, 21:25 UTC · Kimi Zero Warnings · AI agent
Hazard review
- scenarioreviewedconcurrenterror_handlingedge_case
Worst case: a killed batch has answered only the rows it reached, and a row whose when: went unanswered shows by default - so applying the half-read buffer flips rows to visible against real state, or a stranded guardsPending freezes the whole menu on stale results forever (liveness loss). The early return on nonzero exit/signal plus the callLater reevaluation close both arms. concurrent: evaluateGuards re-entered while the batch runs must not kill the in-flight run - guardsPending is the serialization protocol, and Process ignoring command changes while running means a naive restart would throw away read lines and let the surviving tail pass as the whole answer. edge_case: a signal leaves exit code 0, so exitStatus is the discriminator - the exitCode-only check is the trap this requirement forbids. error_handling is the subject itself: discard-and-keep is the declared failure behavior.
- propertyreviewedatomicity
Result maps swap as a whole set only after a full parse of a successfully exited batch; on any failure path no map is touched, so the menu observes either the last complete set or the next complete set, never a half-applied one. Residual composition note: the three map assignments are sequential, so a binding on an individual map can observe a two-of-three-updated instant before rebuildDisplay runs - display consistency is carried by the explicit rebuild, which is the only consumer that publishes rows.
- structuralnot applicable
String buffer accumulation and lastIndexOf-based line parsing over self-produced batch output; no memory management, pointer math, format-string, or encoding round-trip surface. Malformed lines (no colon) are skipped by construction, keeping the parse total.
- domainnot applicable
The bash child is spawned and reaped entirely in-process by Quickshell with the batch script composed by the guard-pipeline siblings; no network, IPC-channel, crypto, or privilege boundary is crossed by this requirement's own contract, and no domain tag fires.
Change history
Every recorded revision of this requirement's source file — newest first, each with its commit message and the diff for that change.
Review history
Human and AI-agent approvals of this requirement — the 'why was this approved' lineage, each with the reviewer's justification and the code it cites.
- Kimi Zero Warnings · AI agentApprovedSpec conformanceSep 27, 2026 · 5 days agoREVIEW-34
Read evaluateGuards and guardProc.onExited in Menu.qml plus the compositor guard-batch test. onExited returns before touching result maps on any nonzero exit (signal exits are caught via exitStatus since a signal leaves exitCode 0), so a killed batch never overwrites the last complete set; a guardsPending evaluation stands aside and is rescheduled via callLater. The batch_killed condition models the exit-status branch; live test at menu-compositor-test.sh:470-523 drives a real killed batch and confirms both conjuncts. Formula conforms to the code.
Cited code (1)
Obligations
What this requirement must witness to be considered satisfied — the required evidence, and the tests that discharge each one.
The tests that discharge each obligation aren't available for a historical run
The evidence matrix behind each obligation comes from the live audit index, which can't be rebuilt for a past commit. Nothing here means unknown — not that the requirement has no obligations.
Formula evidence
The formal formula behind this requirement, the variables it is written over, and the tests that exercise it (each term is explained inline).
Formula evidence
The formal formula behind this requirement, the variables it is written over, and the tests that exercise it (each term is explained inline).
FRETish formula
when batch_killed the guard_pipeline shall eventually satisfy last_complete_set_kept & pending_reeval_runs
Witnesses· 2 scenarios total
- pokeGuardsexercises 2 condition scenarios
MC/DC truth table· 5 rows
Each row assigns the formula's conditions (T/F) and shows the Result— the formula's value for that input row, not a test pass/fail. A row proves a condition when flipping only that condition flips the outcome. The test that covers each row is linked.
mcdc:ignoreNo-actionfalse-result row satisfied by designUncoveredneeds a covering test| # | batch_killed | last_complete_set_kept | pending_reeval_runs | Result | Proves | Covering test |
|---|---|---|---|---|---|---|
| 1 | F | F | F | T | batch_killed | |
| 2 | T | F | F | F | batch_killed | Exempted · defensive — guardProc.onExited returns on any nonzero exit before touching the result maps; applying a half-read batch needs that early return removed (reviewed: REVIEW-MC1) |
| 3 | T | F | T | F | last_complete_set_kept | Exempted · defensive — same early return; a killed batch cannot overwrite the kept set whether or not a reevaluation follows (reviewed: REVIEW-MC1) |
| 4 | T | T | F | F | pending_reeval_runs | Exempted · defensive — every nonzero exit with guardsPending set schedules evaluateGuards via callLater; a stranded pending reevaluation needs that deferred call removed (reviewed: REVIEW-MC1) |
| 5 | T | T | T | T | last_complete_set_kept |
Its place
How this requirement connects — what proves it, what it affects, and what it rests on. Authored links only here; automatically derived links come from the audit index.
Loading graph…
Trace evidence
The concrete artifacts linked to this requirement — implementing code, verifying tests, documents, and the findings raised against it.
No findings affect this requirement
Nothing was flagged against this requirement in the pinned run.
Impact
Blast radius — authored trace links only (automatically derived links come from the audit index and aren't shown here).
Impact
Blast radius — authored trace links only (automatically derived links come from the audit index and aren't shown here).
If you change this
Files to re-check (1)
- Menu.qmlshell/plugins/menu/Menu.qml
Tests to re-run (1)
- menu-compositor-test.shtest/shell.d/menu-compositor-test.sh
What this rests on
Discussions
Discuss this with the proof team. Nothing changes in your audit automatically — you open a request and a staff member records any outcome inside the thread.