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.
A converter that refuses a video records an empty .failed marker keyed by the file content hash.
Specification
The requirement exactly as authored — its complete prose text and, where present, the formal FRETish sentence it compiles to.
A converter that refuses a video records an empty .failed marker keyed by the file content hash. Timeouts and kills write no marker, so a later open retries. A standing marker withholds the video row and keeps the rows uncached. Still images never produce markers, and a failed lazy generation leaves nothing behind.
when media_rejected the image_selector shall eventually satisfy rejection_marker_recorded
Rationale & tags
Why this requirement exists, and how it is categorised.
Rationale & tags
Why this requirement exists, and how it is categorised.
bin/omarchy-menu-images generate_thumbnail writes the marker, thumbnail_for hands it back, and is_video_path gates marker eligibility. Authored with SW-REQ-260929-THMB to resolve the family_under_modeled finding on the shared target.
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
- Opencode · AI agent
- Reviewed
- Sep 30, 2026, 01:05 UTC
History
- Created
- Sep 30, 2026, 00:23 UTC · Opencode · AI agent
- Modified
- Sep 30, 2026, 08:02 UTC · Leonid Bugaev
Hazard review
- scenarioreviewederror_handlingedge_caseconcurrent
Worst case: an environmental converter failure (ffmpegthumbnailer briefly absent, transient ENOSPC) writes a standing marker whose row stays suppressed until the file signature changes - the marker conflates file defect with environment defect, graded medium in obligation_hazards. error_handling: every refusal arm is dispositioned - non-timeout failure records the marker, timeout(124)/kill(137) deliberately record nothing so the next open retries, and success publishes the thumbnail; a failed lazy generation leaves nothing behind by the same failure arm. edge_case: still images never produce markers even on failure, so the marker asymmetry between lanes is contract, not accident. concurrent: two generators racing one file serialize on the content-hash flock and write byte-identical empty markers, so interleaving cannot produce a partial marker.
- propertyreviewedidempotency
idempotency: marker creation is a single creat of an empty file - repeating it (racing generators, repeated opens) converges on the same on-disk state, and re-running a failed open either reproduces the marker or clears the picture by succeeding; the marker key (path+size+mtime hash) makes retry-after-repair deterministic. determinism: marker decisions depend only on converter exit status and file identity, never on fan-out order. atomicity is suppressed with rationale (single creat, no partial window).
- structuralnot applicable
bash + one creat + one mv: no memory, pointer, or format-string surface; the marker carries no bytes to encode. The only structural question (empty-file create races) is dispositioned under scenario.concurrent.
- domainreviewed
external_call_timeout_bounded evaluated: the video converter IS bounded (timeout -k 5 10), and the marker policy deliberately distinguishes bounded-timeout (no marker, retry) from refusal (marker, suppress) - operation_outcome_recorded_faithfully holds because the record faithfully reflects the declared policy rather than merely the exit status. The still-lane timeout gap is owned by the sibling SW-REQ-260929-THMB deferral under KI-MENU-IMAGES-VIPS-NO-TIMEOUT; REJT contractually adds nothing there (stills never write markers). errors_propagated and atomic_write suppressed with standing rationales (marker is the propagation channel; single creat has no partial window). No trust boundary: markers live in the user cache dir keyed by content hash of user-owned files; a hostile co-tenant is outside this desktop threat model.
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.
- Opencode · AI agentApprovedSpec conformanceSep 30, 2026 · 3 days agoREVIEW-260930-VMF5
Read the rejection path in bin/omarchy-menu-images. generate_thumbnail (165-222) writes the empty .failed marker only for video paths (is_video_path at 157 gates eligibility) and only when the converter exits non-zero without a timeout (124) or kill (137); the tmp and thumbnail are removed on any failure. thumbnail_for (225-270) hands the standing marker back before any lane choice, the caller drops the row and marks the rows uncacheable, and the publication branch then removes the rows cache instead of writing it. Timeouts and kills leave no marker, so the next open retries. Formula: a refused video (fresh refusal or standing marker) ends recorded and honored -- marker present, row withheld, rows uncached.
Cited code (4)
Obligations
What this requirement must witness to be considered satisfied — the required evidence, and the tests that discharge each one.
2 obligations · 2 discharged
Browse the catalogueBehavior when operations fail or dependencies are unavailable.
a converter that fails for an environmental reason rather than a file reason (e.g. ffmpegthumbnailer briefly missing during a partial upgrade, or a transient ENOSPC) records a standing .failed marker, and the video row stays suppressed on every later open until the file signature (size+mtime) changes - the marker conflates file defect with environment defect
- negativerequiredpresentCovered by 1 test
- nominalrequiredpresentCovered by 1 test
Behavior when inputs are syntactically or structurally invalid.
a corrupt video that crashes ffmpegthumbnailer with a non-timeout status is marker-suppressed until edited; acceptable by design, but a file whose size+mtime are restored (touch -d) after a real repair re-inherits the stale marker
- negativerequiredpresentCovered by 1 test
- nominalrequiredpresentCovered by 1 test
Code signals
Static-analysis signals from external scanners that bear on this requirement's obligations — the source location, the obligation each touches, and its closure status.
- Coveredbuiltinstrict_mode_missingObligation: Error handling · Scenario
Behavior when operations fail or dependencies are unavailable.
script runs pipelines without `set -e`/`set -o pipefail`
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 media_rejected the image_selector shall eventually satisfy rejection_marker_recorded
Witnesses· 2 scenarios total
- menu-images-test.sh:1exercises 2 condition scenarios
MC/DC truth table· 3 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| # | media_rejected | rejection_marker_recorded | Result | Proves | Covering test |
|---|---|---|---|---|---|
| 1 | F | F | T | media_rejected | |
| 2 | T | F | F | media_rejected | Exempted · defensive — a refused video either writes its marker in the fan-out child or is honored through the standing marker; a rejection with neither needs a broken branch mcdc:witness-out-of-process (reviewed: REVIEW-260930-WJQK) |
| 3 | T | T | T | rejection_marker_recorded |
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)
- omarchy-menu-imagesbin/omarchy-menu-images
Tests to re-run (1)
- menu-images-test.shtest/shell.d/menu-images-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.