Proof Portal
Omarchy
ProbeLabs65 findings · 95 requirementsA 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.
omarchy-update-available is not concurrency-safe: overlapping runs report a false "up to date", hiding the bar's update icon on multi-monitor setups
Reported by jg-partisia · filed Sep 2, 2026Closed upstream on 2026-09-02 by its own reporter jg-partisia, as a duplicate of #9798 (R-1059). GitHub close reason: not_planned.
At a glance
- 1 · What is claimed?
- Only the report text below — no validator has restated it as checkable claims yet.
- 2 · What does it affect?
- Not yet mapped to any requirement or component.
- 3 · Is it reproducible?
- No validation has been run yet.
- 4 · What remains unclear?
- Not validated yet: no claims, no reproducer.
- 5 · What happens next?
- Our staff decide whether the report is valid. Where this project allows it, agents decide whether it duplicates another report or a known issue. No decision has been recorded yet.
Original report
System details
AMD Ryzen AI 7 350 / Radeon 860M, Omarchy 4.0.0-1, kernel 7.1.8-arch1-3, pacman-contrib 1.13.1-1
What's wrong?
bin/omarchy-update-available is not safe to run concurrently. A second
overlapping invocation reports "Omarchy is up to date" and exits 1 even
though an update is pending.
This matters because the bar calls it once per monitor: omarchy.system-update
is an ordinary bar widget, so each monitor's bar instantiates its own copy, and
each copy runs its own Process on the same triggeredOnStart timer. On a
multi-monitor machine the two checks fire together and one of them loses. Since
the widget is visible: updateAvailable, the loser silently renders nothing —
an available update shows no icon at all on that bar.
Reproduce (an update must actually be pending):
$ for i in 1 2; do (out=$(omarchy-update-available 2>&1); echo "run$i: exit=$? | $out") & done; wait
run2: exit=1 | Omarchy is up to date
run1: exit=0 | omarchy 4.0.0-1 -> 4.0.2-1Cause. checkupdates (pacman-contrib) syncs into one shared temp database,${TMPDIR:-/tmp}/checkup-db-${UID}, and removes db.lck from atrap ... INT TERM EXIT. Two concurrent runs share that directory, so whichever
finishes first drops the other's lock; the loser's pacman -Sy fails, or itspacman -Qu reads a half-synced database and prints nothing. Both pacman calls
are &> /dev/null / 2> /dev/null, so the error is invisible, andomarchy-update-available maps the empty result to "up to date":
update=$(checkupdates --nocolor 2>/dev/null | awk -v package="$package" '$1 == package { print; exit }' || true)A false "up to date" is the bad direction to fail in — it hides a pending
update rather than showing a spurious one.
Expected: concurrent callers agree, and a pending update is reported by all
of them.
Suggested fix: serialize the script under flock, so every caller benefits
rather than only the bar. Per-instance CHECKUPDATES_DB would also avoid the
shared directory, but it multiplies the pacman sync by the number of monitors.
Not reported as a display bug on purpose: the missing icon is the visible
symptom, but it is transient — a later refresh can bring it back on both bars —
and the script-level race above is the part that reproduces deterministically.
View raw text
### System details
AMD Ryzen AI 7 350 / Radeon 860M, Omarchy 4.0.0-1, kernel 7.1.8-arch1-3, pacman-contrib 1.13.1-1
### What's wrong?
`bin/omarchy-update-available` is not safe to run concurrently. A second
overlapping invocation reports **"Omarchy is up to date" and exits 1** even
though an update is pending.
This matters because the bar calls it once per monitor: `omarchy.system-update`
is an ordinary bar widget, so each monitor's bar instantiates its own copy, and
each copy runs its own `Process` on the same `triggeredOnStart` timer. On a
multi-monitor machine the two checks fire together and one of them loses. Since
the widget is `visible: updateAvailable`, the loser silently renders nothing —
an available update shows no icon at all on that bar.
**Reproduce** (an update must actually be pending):
```
$ for i in 1 2; do (out=$(omarchy-update-available 2>&1); echo "run$i: exit=$? | $out") & done; wait
run2: exit=1 | Omarchy is up to date
run1: exit=0 | omarchy 4.0.0-1 -> 4.0.2-1
```
**Cause.** `checkupdates` (pacman-contrib) syncs into one shared temp database,
`${TMPDIR:-/tmp}/checkup-db-${UID}`, and removes `db.lck` from a
`trap ... INT TERM EXIT`. Two concurrent runs share that directory, so whichever
finishes first drops the other's lock; the loser's `pacman -Sy` fails, or its
`pacman -Qu` reads a half-synced database and prints nothing. Both pacman calls
are `&> /dev/null` / `2> /dev/null`, so the error is invisible, and
`omarchy-update-available` maps the empty result to "up to date":
```bash
update=$(checkupdates --nocolor 2>/dev/null | awk -v package="$package" '$1 == package { print; exit }' || true)
```
A false "up to date" is the bad direction to fail in — it hides a pending
update rather than showing a spurious one.
**Expected:** concurrent callers agree, and a pending update is reported by all
of them.
**Suggested fix:** serialize the script under `flock`, so every caller benefits
rather than only the bar. Per-instance `CHECKUPDATES_DB` would also avoid the
shared directory, but it multiplies the pacman sync by the number of monitors.
Not reported as a display bug on purpose: the missing icon is the visible
symptom, but it is transient — a later refresh can bring it back on both bars —
and the script-level race above is the part that reproduces deterministically.
Validation
Not checked yet. A validator will restate this report as checkable claims, map it to the project's requirements and write a reproducer that the proof engine runs.
Linked
Not linked to any known issue, other report or change yet.
Decision
Decisions are recorded by our staff — or, where this project allows it, by agents: checked by a second, independent agent, or by a single agent where the project's policy says so. Any decision reached so far is shown below.
No decision has been recorded yet.