Skip to content

AGENTS: a plugin's link in ~/.config/omarchy/plugins outranks the package - #243

Closed
W4HO-ham wants to merge 2 commits into
derekross:mainfrom
W4HO-ham:agents-plugin-links
Closed

W4HO-ham wants to merge 2 commits into
derekross:mainfrom
W4HO-ham:agents-plugin-links

Conversation

@W4HO-ham

@W4HO-ham W4HO-ham commented Nov 1, 2026

Copy link
Copy Markdown
Contributor

On the palmtop (Pixel 9a), the notification shade from the website's third screenshot never appeared, whatever shell version was installed. The cause was ~/.config/omarchy/plugins/omamobile.quicksettings: a bring-up step on 10-17 had pointed it at a hand-edited copy under /var/lib, written before the shade existed. omamobile.switcher pointed at a 10-22 copy the same way. omamobile-setup-user links every packaged plugin there so updates reach the user, and a link to a private copy silently pins that copy through every upgrade.

After relinking with the setup script's own loop, the shade, the current switcher and the plugins that were missing (back, call, update, access) all came up on omamobile-shell 0.2.56.

This adds the gotcha to AGENTS' Userspace section: where to look first when a shell change doesn't show on a phone, and the one-line relink. Docs only; tools/pr-check passes.

@derekross

Copy link
Copy Markdown
Owner

Thanks for tracking this down. A shade that never appears whatever version is installed is maddening, and "check the plugin links" is a good thing to have written down. Before it goes into AGENTS, I'm not sure the cause is quite what the line says.

Since 0.2.15 the package enables omamobile-setup-user.service for every user, and it runs the setup script at every login. The script re-links every omamobile.* plugin each time (ln -sfn, lines 18-21), so on a normally set-up phone a hand-made link only lasts until the next login. The palmtop's links survived, and four plugins were never linked at all, which makes me think the service hasn't been running there. Its condition is the profile marker ~/.local/state/omarchy/done/omamobile-shell-profile, which a user set up during bring-up rather than by firstboot may not have. Could you check systemctl --user status omamobile-setup-user and ls ~/.local/state/omarchy/done/ on the palmtop?

If that's the cause, I'd like the gotcha to point there: if a shell change doesn't show, check that the setup service runs at login and that the marker exists. Your relink loop is fine as the immediate fix; please note why it's the loop and not the script (run without the marker, the script would re-seed shell.json over the user's). Please also use one date convention, and link the line to the NOTES entry in #244.

…kage

On the Pixel 9a, bring-up links to hand-edited copies of the quick settings and switcher plugins kept those copies running through every shell upgrade, so the notification shade never showed. The gotcha says where to look and how to relink.
…ervice and its marker

On the palmtop the user unit was skipped at every login for want of the profile marker, which is why the hand links survived. The line says how to check that, why to relink with the loop rather than the script, and links the NOTES entry.
@W4HO-ham
W4HO-ham force-pushed the agents-plugin-links branch from 90a8f14 to 1c83767 Compare November 1, 2026 14:13
@W4HO-ham

W4HO-ham commented Nov 1, 2026

Copy link
Copy Markdown
Contributor Author

You were right. On the palmtop:

omamobile-setup-user.service: disabled; preset: enabled
Active: inactive (dead)
Condition: start condition unmet
  ConditionPathExists=/home/mobile/.local/state/omarchy/done/omamobile-shell-profile was not met

~/.local/state/omarchy/done/ holds only omamobile-keyboard-pill-hint. The user was set up during bring-up, so it never got the profile marker, and the unit is skipped at every login.

The line now points there: check that the setup service runs at login and that the marker exists, with both commands. It also says why to use the loop and not the script. Without the markers, omarchy-done ensure omamobile-shell-profile succeeds and the package's shell.json is copied over the user's. Every other once-only step would run again too: the cursor block appended to looknfeel.lua again, squeekboard added to autostart.lua again, the monitor scale reseeded. Dates are ISO throughout, and the line links the leftovers entry in #244's NOTES.

So I haven't just touched the marker on the palmtop. The unit would then run all of those steps once over the bring-up setup. The safe repair is to mark every step that's already in effect (omarchy-done mark … for each), then let the unit run. If you'd like that as a small --adopt mode in the setup script for phones set up by hand, I can send it.

@derekross

Copy link
Copy Markdown
Owner

Thanks for checking. "Condition: start condition unmet … omamobile-shell-profile was not met" settles it, and the new line is much more useful: it sends the next person to the setup service and its marker first. The explanation of why to use the loop and not the script (the shell.json reseed and the repeated once-only steps) is exactly right, and so is your decision not to just touch the marker on the palmtop.

Yes please to the --adopt idea, as its own PR: a mode that marks the steps already in effect, so a phone set up by hand can join the normal login path. It would be good to have before the next person meets this.

Two things I'll handle at merge, nothing needed from you:

One tiny wording point, which I can fold in at merge: the marker is created by the first run of omamobile-setup-user, which firstboot does, rather than by firstboot itself. "which the first run of omamobile-setup-user (firstboot's) creates" would match the unit's own comment.

@derekross derekross closed this in 661de47 Nov 1, 2026
robertzas pushed a commit to robertzas/omarchy-mobile that referenced this pull request Nov 14, 2026
…e and its marker (derekross#243)

When a shell change doesn't show on a phone, the cause found on the 9a was omamobile-setup-user.service not running at login (its ConditionPathExists unmet), not a plugin link outranking the package. The gotcha now points at the setup service and its marker (W4HO-ham).

Closes derekross#243

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
gitColinX added a commit to gitColinX/omarchy-mobile that referenced this pull request Nov 16, 2026
…ekross#243 and derekross#244

- NOTES no longer calls SUSPEND-PLAN a draft (derekross#802).
- The ROADMAP Suspend row says mem_sleep read "s2idle [deep]" before
  kernel derekross#244 and reads "[s2idle] deep" on the published kernel (derekross#802).
- The ROADMAP Time row no longer says "No RTC": the S2MPG14's RTC is
  rtc0 since kernel derekross#243 (asked for on that PR).
- The 2026-11-15 NOTES entry records /proc/cmdline's
  mem_sleep_default=s2idle with mem_sleep (asked for on kernel derekross#244),
  and the RTC's probe lines, wakealarm and idle interrupt count (asked
  for on kernel derekross#243).

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
derekross pushed a commit that referenced this pull request Nov 18, 2026
…efore H2; NOTES and ROADMAP leftovers

H0b (the Wi-Fi and PCIe unload and reload, no sleep) and H1 (`pm_test` freezer, devices and platform through `systemctl suspend`: three successes, no failures) ran attended on Colin's 8 Pro on the published `omamobile` at `82f31d3a7` with device-husky 0.1.13; touch, the display and a played haptic effect after H1 are carried into H2, and H2 to H4 have not run. The plan now says that the UFS `dme-set: attr-id 0x4021` line at resume is expected and not a stop, lists the four brcmfmac noise lines, names the maintainer's go-ahead per stage and device-husky 0.1.14 or later for H2 to H4, and records the sleep configuration in effect. NOTES and ROADMAP pick up the leftovers from #802 and the RTC and `mem_sleep_default=s2idle` facts from kernel #243 and #244.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants