Skip to content

Gate screensaver launches on window class and pidof -x - #10260

Closed
fresh3nough wants to merge 3 commits into
omacom:quattrofrom
fresh3nough:fix/10197-screensaver-pgrep-false-positive
Closed

fresh3nough wants to merge 3 commits into
omacom:quattrofrom
fresh3nough:fix/10197-screensaver-pgrep-false-positive

Conversation

@fresh3nough

@fresh3nough fresh3nough commented Sep 5, 2026 •

Copy link
Copy Markdown
Contributor

Summary

omarchy-launch-screensaver decided a screensaver was already running with:

pgrep -f '[o]rg.omarchy.screensaver' && exit 0

pgrep -f matches the full command line of every process. Any unrelated process whose argv contained org.omarchy.screensaver made the launcher exit 0 with no output and no screensaver. Idle logging then looked healthy. Real hit: nwg-dock-hyprland -g 'org.omarchy.screensaver' (see #10197).

An earlier revision of this PR used pgrep -x omarchy-screensaver. That can never match: the script name is 19 characters, Linux truncates comm to 15, and procps refuses the pattern — every force launch stacked another fullscreen window.

Current gate:

hyprctl clients -j 2>/dev/null |
  jq -e 'any(.[]; .class == "org.omarchy.screensaver" or .initialClass == "org.omarchy.screensaver")' >/dev/null && exit 0
pidof -x omarchy-screensaver >/dev/null 2>&1 && exit 0

Window first for a mapped client, then pidof -x for the hypr_exec / openwindow interval. Plain pidof misses a bash script; -x is required. Terminal --class / --app-id are unchanged for Hyprland rules.

Fixes #10197

Test plan

  • bash test/shell.d/launch-screensaver-test.sh (7/7)
  • Static shape rejects both broken pgrep forms; requires class check + pidof -x
  • Mapped class / initialClass exit 0 before helpers
  • Near-miss class/title continues past the window gate
  • Live decoy argv with the app-id does not trip pidof -x or the launcher
  • Live long-named script process is caught by real pidof -x (and not by pgrep -x)

@arisgysel-design

Copy link
Copy Markdown
Contributor

Thanks for picking this up. pgrep -x omarchy-screensaver cannot match this command on Linux: the process name exposed through /proc is limited to 15 characters, while omarchy-screensaver is 19 characters.

I reproduced this in Arch Linux with an executable named omarchy-screensaver:

comm=omarchy-screens
pgrep: pattern that searches for process name longer than 15 characters will result in zero matches
full_name_status=1

The current test replaces pgrep with a stub, so it cannot expose that kernel/procps behavior. With this patch, a real running screensaver will never satisfy the early-exit gate and concurrent invocations can open duplicates.

The issue’s Hyprland-window check avoids both the original argv false positive and the 15-character process-name limit. I would also include initialClass, matching omarchy-debug-idle:

hyprctl clients -j 2>/dev/null \
  | jq -e 'any(.[]; .class == "org.omarchy.screensaver" or .initialClass == "org.omarchy.screensaver")' >/dev/null \
  && exit 0

A regression should exercise the real window-state predicate rather than a mocked pgrep -x contract.

@chessing234 chessing234 left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

pgrep -x fix directly addresses the false-positive early exit.

@omarchybot

Copy link
Copy Markdown
Collaborator

Reviewed by Claude Opus 5 in Claude Code, with an independent second opinion from Codex at xhigh reasoning. Everything below was run on a disposable Omarchy 4.0.0 worker VM with a real compositor, never on the review host.

The new gate can never match

This confirms, with measurements, what arisgysel-design already reported above: omarchy-screensaver is 19 characters, Linux truncates /proc/<pid>/comm to 15, and pgrep -x matches against comm. procps-ng refuses the pattern outright.

On the worker (procps-ng 4.0.7) with a real screensaver up on foot:

  PID   PPID COMMAND         ARGS
11615    729 foot            foot --app-id=org.omarchy.screensaver --config=.../foot/screensaver.ini -e omarchy-screensaver
11617  11615 omarchy-screens /bin/bash /home/omarchy/work/omarchy/bin/omarchy-screensaver
11655  11617 ttfx            ttfx -i ...

$ pgrep -x omarchy-screensaver
pgrep: pattern that searches for process name longer than 15 characters will result in zero matches
Try `pgrep -f' option to match against the complete command line.
rc=1

$ pgrep -x omarchy-screens
11617

The process you picked is the right one — bin/omarchy-screensaver outlives the launcher as a child of the terminal, and its while true loop at line 38 keeps it alive — but its comm is omarchy-screens, so the gate at bin/omarchy-launch-screensaver:10 returns 1 forever.

That makes this a regression rather than a fix. The old pgrep -f gate over-matched; this one under-matches always. Running this branch's launcher against a screensaver already up on that worker:

screensaver windows mapped before: 1
PR_LAUNCHER_RC=0
screensaver windows mapped after:  2

Two stacked screensaver terminals, per monitor. default/omarchy/omarchy-menu.jsonc:36 and bin/omarchy-branding-screensaver:15,19,22 all call omarchy-launch-screensaver force with no other guard, so a user reaches this from the menu. The idle path is partly shielded by screensaverStartedThisCycle at shell/plugins/services/idle/Service.qml:66, which stops a repeat inside one idle cycle but not across cycles.

Secondary: >/dev/null covers stdout only, so that procps warning goes to stderr on every launch and lands in the shell's process log.

The test cannot catch this, by construction

test/shell.d/launch-screensaver-test.sh:71-82 puts a pgrep stub on PATH, and the only case that exercises a running screensaver — Case 1, lines 105-114 — goes through it. It asserts the stub's own contract, not procps'. The dynamic half at lines 39-65 runs before the PATH change, so it proves a standalone pgrep does not false-positive on a decoy; nothing anywhere proves the gate still fires on a true positive, which is the assertion that would have failed.

Smaller things in the same file:

  • Line 55 asserts pgrep -f '[o]rg.omarchy.screensaver' matches something on the machine, not the decoy. Where a real screensaver is running, foot's own argv satisfies it, so the baseline passes even if the decoy never started. Assert against $decoy_pid.
  • Line 60 is machine-wide for the same reason. On Linux it is vacuous today rather than wrong, but it is not a scoped assertion, and two concurrent copies of this file would satisfy each other's probes.
  • cleanup kills the background bash but not the sleep 30 it is currently blocked on, so that child outlives the run by up to 30 seconds. Codex confirmed this on its host; it is cosmetic, but kill -- -$decoy_pid on a process group would be tidier.
  • Line 28 creates decoy_log and never writes to it, and line 106's SCREENSAVER_RUNNING=1 is dead — line 108 sets it inline on the command.

Not pushing a fix

Making the gate work means replacing the mechanism, and that takes most of the test file with it. That is your PR to shape rather than something to land over your name without asking. Two directions that work:

The window check from #10197, which this repo already uses verbatim at bin/omarchy-debug-idle:50, initialClass included as arisgysel-design suggested. It returned true on the worker while the screensaver was up, and the launcher already depends on both hyprctl and jq further down, so it adds nothing:

hyprctl clients -j 2>/dev/null |
  jq -e 'any(.[]; .class == "org.omarchy.screensaver" or .initialClass == "org.omarchy.screensaver")' >/dev/null && exit 0

Or keep a process gate and give the process a name short enough to survive the truncation. #6813 does exactly that — printf 'omarchy-saver' >/proc/self/comm in bin/omarchy-screensaver, then pgrep -x omarchy-saver, 13 characters. Worth watching that PR either way: if it lands, this one's pattern has to change regardless.

Either way, the regression test should drive the real predicate rather than a stub, or it will keep passing on a gate that cannot fire.

Out of scope here, but adjacent

The same argv-collision class survives in two places that kill rather than skip: bin/omarchy-screensaver:12 and bin/omarchy-system-lock:26 both run pkill -f '[o]rg.omarchy.screensaver', so dismissing or locking sends SIGTERM to any unrelated process whose arguments contain that public string. bin/omarchy-debug-idle:52 keeps the same pgrep -f as a fallback and will report running-process for a decoy. Codex found these; they are pre-existing and not this PR's job, but they mean #10197's underlying class is wider than the launcher.

What ran

./test/cli — passes, 0 failures. ./test/shell.d/launch-screensaver-test.sh through the worker's compositor wrapper — 4 passed, 0 failed. The suite is green and the bug is still there, which is the point of the section above.

Codex agreed with the truncation finding and with the test's stub gap; both were written down before it ran, so its independence there is not guaranteed. It contributed the two pkill -f sites, the omarchy-debug-idle fallback, and the orphaned-sleep detail on its own, and each was verified against the source here before being repeated.

Waiting on you.

@arisgysel-design

Copy link
Copy Markdown
Contributor

Implemented the requested follow-up on top of this PR in arisgysel-design/omarchy@1d0edf3b.

  • Replaced the impossible pgrep -x omarchy-screensaver check with the exact Hyprland .class / .initialClass window predicate from the review.
  • Reworked the regression test so the launcher executes the real jq predicate; only hyprctl clients -j is stubbed to provide controlled compositor JSON.
  • Covered current class, initial class, and a non-matching public-string decoy.
  • Removed the old process probe, decoy process, orphaned sleep, dead log, and stubbed pgrep contract.

Local gates: focused test 4/4, bin-style 2/2, bash -n, and git diff --check. The full Omarchy suite is Linux-specific and cannot be validly claimed from this macOS host.

The PR head belongs to fresh3nough, so I did not rewrite that branch. The commit is ready to cherry-pick: 1d0edf3b94de92cf6aff041c53534c8644795c1f.

@omarchybot

Copy link
Copy Markdown
Collaborator

Re-reviewed at b17b784, unchanged since the last pass. The finding from that pass stands, and I re-proved it from scratch on a disposable Arch worker (kernel 7.1.9-arch1-2, procps-ng 4.0.7) rather than carry it forward on trust.

The gate cannot fire

bin/omarchy-screensaver is a bash script with a 19-character filename, so the kernel sets its comm to the 15 bytes it has room for. With a real screensaver up on the worker:

pid=2866  comm=omarchy-screens  cmdline=/bin/bash .../bin/omarchy-screensaver

$ pgrep -x omarchy-screensaver
rc=1
(stderr) pgrep: pattern that searches for process name longer than 15 characters
         will result in zero matches

procps refuses a -x pattern over 15 characters outright, so bin/omarchy-launch-screensaver:10 returns 1 for the rest of the machine's life. I measured four candidate gates against the same three states on that worker:

gate idle decoy with the app-id in argv real screensaver up
pgrep -f '[o]rg.omarchy.screensaver' (today) no match matches — this is #10197 matches
pgrep -x omarchy-screensaver (this PR) no match no match no match
pidof -x omarchy-screensaver no match no match matches (pid 2866)
hyprctl clients -j | jq -e 'any(...)' no match no match matches

Rows 3 and 4 are correct in all three states. Row 2 fixes the middle column and breaks the one that matters. Running this branch's launcher against a screensaver that was already up:

screensaver windows mapped before: 1
PR_LAUNCHER_RC=0
screensaver windows mapped after:  2

So the branch as it stands cannot close #10197 — it replaces a silent no-op with duplicate stacked screensavers, one per monitor per invocation. default/omarchy/omarchy-menu.jsonc:36 and bin/omarchy-branding-screensaver:15,19,22 all reach the gate with force and no other guard, so a user gets there from the menu. An approving review sits on this commit; the table and the window counts above are the whole of what I have to add to it.

Secondary: >/dev/null on line 10 covers stdout only, so that procps warning goes to stderr on every launch — including a launch that would have exited at the screensaver-off toggle two lines later, because the gate runs first.

The test still cannot catch it

Unchanged since the last pass, so briefly: test/shell.d/launch-screensaver-test.sh:71-82 puts a pgrep stub on PATH, and the only case that exercises a running screensaver — lines 105-114 — goes through it, asserting the stub's own contract rather than procps'. It passed 4 of 4 on the worker this run, with the gate broken.

Where I would take it

Not pushing anything: replacing the mechanism takes most of the test file with it, and that is yours to shape.

Codex found a failure mode in the window check that I had not reasoned about, and it changes the shape of the answer. hypr_exec at bin/omarchy-launch-screensaver:36 is asynchronous, and the script then blocks on wait_for_screensaver_window (lines 49-54) waiting for the openwindow event. Through that interval a screensaver is genuinely starting while hyprctl clients still shows nothing, so two invocations can both pass a pure window check. The current pgrep -f gate closes that interval, because the terminal process exists the moment it execs.

The predicate that covers every column above is the one this repo already has at bin/omarchy-debug-idle:50-53 — window first, process second:

hyprctl clients -j 2>/dev/null |
  jq -e 'any(.[]; .class == "org.omarchy.screensaver" or .initialClass == "org.omarchy.screensaver")' >/dev/null && exit 0
pidof -x omarchy-screensaver >/dev/null && exit 0

hyprctl and jq are already required further down the same script and both ship in install/omarchy-base.packages, so this adds no dependency. The -x on pidof is load-bearing: plain pidof omarchy-screensaver misses a bash script (rc=1 on the worker), pidof -x matches it. Codex would keep only the pidof -x line, on the grounds that it is atomic and has no window-map race; the cost is that a screensaver script surviving with no mapped window would then block a relaunch. Both are defensible and the mechanism is your call and the maintainer's — what I can say from measurement is only that pgrep -x is not among the options.

A third route: #6813 writes printf 'omarchy-saver' >/proc/self/comm in bin/omarchy-screensaver, which brings the name to 13 characters and makes a pgrep -x omarchy-saver gate possible. That would make this PR depend on that one landing first, and the pattern here would still have to change — omarchy-screensaver will not match either way.

Whichever you pick, the regression test needs to drive the real predicate rather than a stub, or it will keep passing on a gate that cannot fire.

The fork commit

I read @arisgysel-design's comment. That commit is not on this branch, so it does not change what I can say about this one, and I did not fetch or run it. The diagnosis in their earlier comment matches what I measured independently here.

What ran

On a disposable worker with a real compositor: the probes above; ./test/shell.d/launch-screensaver-test.sh, 4 passed 0 failed; ./test/shell.d/bin-style-test.sh, 2 passed 0 failed. Nothing was executed on the machine holding credentials.

Reviewed by Claude Opus 5, with a second opinion from Codex (gpt-5.6-sol) at xhigh reasoning. Codex agreed with the truncation finding and with the test's blind spot, but both were written down before it ran, so its independence there is not guaranteed. The openwindow race in the window check, and the ordering dependency #6813 would create, are its own contributions; I verified each against the source before repeating it.

Waiting on you.

@fresh3nough

Copy link
Copy Markdown
Contributor Author

Addressed #10260 (comment).

Rebased onto current quattro and replaced the gate at 46c4e3e2.

What was wrong

  • pgrep -x omarchy-screensaver cannot match. The script name is 19 characters; Linux truncates comm to 15; procps refuses the pattern. The gate never fired, so every force launch stacked another fullscreen window.
  • The previous test stubbed pgrep, so it asserted the stub's contract and stayed green on a dead gate.

What it does now

hyprctl clients -j 2>/dev/null |
  jq -e 'any(.[]; .class == "org.omarchy.screensaver" or .initialClass == "org.omarchy.screensaver")' >/dev/null && exit 0
pidof -x omarchy-screensaver >/dev/null 2>&1 && exit 0

Window first for a mapped client, then pidof -x for the hypr_exec / openwindow interval. Plain pidof misses a bash script; -x is required. No pgrep -f (argv false positive) and no pgrep -x (comm truncation).

Tests

bash test/shell.d/launch-screensaver-test.sh — all ok:

  • static shape rejects both broken pgrep forms; requires class check + pidof -x
  • mapped class / initialClass exit 0 before helpers
  • near-miss class/title continues past the window gate
  • live decoy argv with the app-id does not trip pidof -x or the launcher
  • live long-named script process is caught by real pidof -x (and not by pgrep -x)

Only hyprctl is stubbed for the window cases; jq and pidof are the real tools.

Commit is DCO-signed (Signed-off-by: fresh3nough <anonwurcod@proton.me>) and SSH-signed (verified).

pgrep -f against the app-id false-positives on unrelated argv (#10197).
pgrep -x omarchy-screensaver can never match either: the script name is
19 characters, Linux truncates comm to 15, and procps refuses the pattern,
so every force launch stacked another fullscreen window.

Use the Hyprland class / initialClass check first, then pidof -x for the
interval where hypr_exec has started the script but openwindow has not
mapped yet. pidof -x is required; plain pidof misses a bash script.

Rewrite the regression test to drive the real jq and pidof predicates
with only hyprctl stubbed, covering mapped windows, decoy argv, and a
live long-named script process.

Fixes #10197

Signed-off-by: fresh3nough <anonwurcod@proton.me>
@fresh3nough
fresh3nough force-pushed the fix/10197-screensaver-pgrep-false-positive branch from 46c4e3e to 1b461df Compare September 21, 2026 20:03
@fresh3nough

Copy link
Copy Markdown
Contributor Author

Rebased onto current quattro (1b461df7). Gate is unchanged from 46c4e3e2:

  • Hyprland .class / .initialClass first
  • pidof -x omarchy-screensaver for the in-flight openwindow interval
  • no pgrep -f (argv false positive) and no pgrep -x (15-char comm truncation)

Focused bash test/shell.d/launch-screensaver-test.sh still passes (7/7), including the live decoy-argv case that covers the nwg-dock-hyprland -g 'org.omarchy.screensaver' failure mode reported on #10197.

@fresh3nough fresh3nough changed the title Match screensaver processes by name, not full command line Gate screensaver launches on window class and pidof -x Sep 21, 2026
@omarchybot omarchybot added the bug Something isn't working label Sep 27, 2026
@ssandys

ssandys commented Sep 29, 2026

Copy link
Copy Markdown

One more case this fixes: the current pgrep -f '[o]rg.omarchy.screensaver' gate matches the terminal's --class/--app-id argument, not the script. Once the terminal is gone, an orphaned omarchy-screensaver is invisible to the gate, so every idle trigger starts another one next to it.

Combined with the pgrep -t respawn loop in omarchy-screensaver (see my comment on #11146), this stacked 10 orphaned instances on my machine on 2026-09-19 and pushed load average to 106 (about 230 forks/sec, about 195 live ttfx). I confirmed on current master (4.0.0.r6687.gb18ab49) that while an orphan is running, pgrep -f '[o]rg.omarchy.screensaver' returns 0 matches. pidof -x omarchy-screensaver checks the script itself, so it would catch those orphans.

omarchybot and others added 2 commits October 2, 2026 06:22
The launcher now gates on a mapped screensaver window and pidof -x, so the fixture's pgrep stub no longer isolates it, and a screensaver client returned before anything spawned makes the gate exit before the first monitor.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
@omarchybot omarchybot added the verified Omarchy Triage has verified that this issue is ready for final review label Oct 2, 2026
@omarchybot

Copy link
Copy Markdown
Collaborator

Re-reviewed at 1b461df, now 867031f with the two commits below, by Claude Opus 5.5, with a second opinion from Codex at medium effort. Everything below ran on a disposable Omarchy worker with a real compositor (procps-ng 4.0.7, one monitor), never on the review host.

The gate works

The bug reproduces on quattro and is fixed here. With bash -c 'sleep 60; true' org.omarchy.screensaver running and no screensaver up, omarchy-launch-screensaver force:

quattro this branch
decoy with the app-id in argv rc 0, 0 windows (#10197) rc 0, 1 window
screensaver already up, launch again 1 window, stays 1 1 window, stays 1
second launch 50, 150, 300, 600 ms after the first 1 window 1 window
two launches at the same instant 2 windows 2 windows

So the class check plus pidof -x catches a running screensaver, including while the first terminal is still starting, as far down as a 50 ms gap. Neither version closes the simultaneous case; this one doesn't make it worse.

Pushed: quattro merged in, and a test it broke

Since this branch was cut, quattro added test/shell.d/screensaver-workspace-test.sh (bf44355). It stubs pgrep and returns a mapped screensaver from the very first hyprctl clients -j, so with your gate merged the launcher exits before spawning anything: not ok - a screensaver opens on each monitor. That test file doesn't exist on your branch, so the fix needed quattro in it. I pushed two commits:

  • f5292c0 merges quattro into the branch. It is a merge, not a rebase, so your commit is untouched.
  • 867031f stubs pidof instead of pgrep in that fixture, and returns [] from clients -j until the launcher has spawned something. All of its assertions keep their meaning. It passes 6/6, and fails again with only the fixture reverted.

At 867031f, your launch-screensaver-test.sh passes 7/7 and bin-style-test.sh 2/2. The full ./test/shell fails the same 8 files as quattro does at 821ae58, none of them screensaver files. ./test/cli fails only vscode generated theme references current theme file, which also fails on quattro.

Still open

  • A windowless omarchy-screensaver now blocks every launch. If a screensaver script outlives its terminal, pidof -x keeps finding it, and every later launch, force included, exits 0 and shows nothing. On the worker, a stand-in script with no window gave rc 0 and 0 windows here, against 1 window on quattro. @ssandys describes exactly such orphans above, and bin/omarchy-screensaver:38-48 only reaches its exit checks while a ttfx is alive on its tty. Before this change, that state stacked extra screensavers; now it silently stops them, which is the symptom omarchy-launch-screensaver: pgrep -f matches unrelated processes, so the screensaver silently no-ops with exit 0 #10197 is about, reached another way. Codex agreed it is a regression for a user. Whether that trade is acceptable, or whether the process check should be tied to a live terminal, is a call for you and the maintainer.
  • The new test depends on what else the machine is running. At test/shell.d/launch-screensaver-test.sh:211, the decoy check asserts that pidof -x omarchy-screensaver matches nothing anywhere on the host. So it fails whenever a real screensaver, another checkout's copy, or a concurrent run of this test is up. The positive case at 248-278 never checks that the match is $script_pid. The static grep over the launcher's source (34-50) asserts the text rather than the behaviour. Three behavioural cases with a stubbed hyprctl and pidof (screensaver up, decoy only, nothing) would cover what the 288 lines do. If you keep a live pidof probe, assert the exact PID.

Codex also pointed out that plain hyprctl clients -j omits unmapped clients, so the gate doesn't need the .mapped check the focus-restore loop at the end of the file uses. It agreed with the reproduction, the orphan behaviour and the fixture fix, all of which were written down before it ran, so its independence there isn't guaranteed. The test's dependence on host state and the clients -j point are its own, and I checked both against the source before repeating them.

Waiting on you for the two open points. #10197 is already linked and will close when this merges.

@omarchybot omarchybot removed the verified Omarchy Triage has verified that this issue is ready for final review label Oct 2, 2026
@greptile-apps

greptile-apps Bot commented Oct 2, 2026 •

Copy link
Copy Markdown
Contributor

RetriggerConfidence Score: 4/5

[Medium risk] Changes screensaver launch gate logic and adds tests.

The PR is not safe to merge until concurrent launches are prevented from dispatching duplicate fullscreen screensavers.

Findings

  1. P1 Concurrent launches can stack screensavers ▶

Summary

The PR replaces a full-command-line screensaver check with a Hyprland window-class check followed by pidof -x, and adds regression tests. The checks cover existing windows and scripts but not a launch that has been dispatched and has yet to create either.

Diagram
sequenceDiagram
  participant A as First launcher
  participant H as Hyprland
  participant B as Second launcher
  A->>H: Dispatch terminal asynchronously
  B->>H: Query screensaver clients
  H-->>B: None yet
  B->>B: pidof -x finds no script yet
  B->>H: Dispatch duplicate terminal
  H-->>A: First terminal opens
Loading

Reviews (1) · Last reviewed commit: "Stub pidof and start with no screensaver..."

Comment on lines +15 to +17
hyprctl clients -j 2>/dev/null |
jq -e 'any(.[]; .class == "org.omarchy.screensaver" or .initialClass == "org.omarchy.screensaver")' >/dev/null && exit 0
pidof -x omarchy-screensaver >/dev/null 2>&1 && exit 0

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

P1 Concurrent launches can stack screensavers

If a second launch starts after the first dispatches its terminal but before that terminal opens a window or starts omarchy-screensaver, both checks find nothing. A menu force launch can then dispatch another fullscreen screensaver. The first launcher's wait for the window does not prevent the second launch.

@omarchybot omarchybot added the verified Omarchy Triage has verified that this issue is ready for final review label Oct 2, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

bug Something isn't working verified Omarchy Triage has verified that this issue is ready for final review

Projects

None yet

Development

Successfully merging this pull request may close these issues.

omarchy-launch-screensaver: pgrep -f matches unrelated processes, so the screensaver silently no-ops with exit 0

5 participants