Skip to content

omamobile-shell 0.2.124, omamobile-omarchy 0.1.43, omamobile-firstboot 0.1.4: what Omarchy 4.5 renamed or dropped under us - #662

Merged
derekross merged 2 commits into
mainfrom
omarchy-45-breaks
Nov 11, 2026
Merged

derekross merged 2 commits into
mainfrom
omarchy-45-breaks

Conversation

@derekross

@derekross derekross commented Nov 11, 2026 •

Copy link
Copy Markdown
Owner

Preparation for the move to Omarchy 4.5 (#434). Not run on a phone. Every fix works on both Omarchy versions, since phones stay on 4.0.6 until the pin moves.

An audit of our packages against the real omarchy 4.0.6-1 and 4.5.0-3 packages found six places that break on 4.5. Two are the notification store and Clear all, which #482 covers. This is the other four.

What

Where On 4.5 today Now
Restart buttons (update sheet, mobile network page) They run omarchy-system-reboot, which 4.5 replaced with omarchy-system reboot: nothing happens omamobile-reboot (new, in omamobile-shell) runs whichever the installed Omarchy has, and systemctl reboot with neither
Name and email in onboarding It runs install/user/git.sh, which is gone, prints "Saved." and sets nothing Hands both to 4.5's omarchy-setup-git; without that command sets git itself and runs Omarchy's Compose script, as on 4.0; says so when the save failed
Name and email at first boot 4.5's user setup no longer writes git's identity from the installer's answers omamobile-firstboot writes it after Omarchy's setup, under either version
The update's restart step 4.5 calls it twice (--services-only, --reboot-only); our stand-in ignored the argument and did everything twice Each call does its half; a service that failed to restart in the first is still in the second's warnings
The update's boot verification (new in 4.5) No stand-in, though patch 0012's header names one A stand-in that passes: slot B's images are checked by omamobile-boot-update, not by a computer's boot package

omamobile-update's terminal hint now says omarchy reboot, an alias both versions have.

Second commit: behaviour changes from the same audit

  • Menu rows. 4.5 moved Style to Appearance, so the rows the phone hides came back under new ids; they are hidden under both sets. New in 4.5 and hidden too: the dock for floating workspaces and Update > Kernel.
  • Dictation keys. 4.5 binds Alt + Space to its dictation command and its migration keeps F9 for earlier dictation users. phone.lua takes both away from the on-screen keyboard and no longer adds 4.0's Voxtype F9 bind on 4.5; an F9 the owner bound to something else is left alone. The Mic key and the Dictation tile are unchanged.
  • Dictation default. Voxtype is recorded as Omarchy's default backend once, for a user who has none.
  • Migrations. 1790042972 (a world clock on Omarchy's bar) is held. Read and left to run: 1791554699 (Omarchy's power-profiles-daemon build), 1791123731 (comment lines in looknfeel.lua), 1788129995 (omasnap replaces Tensaku). Of the 98 migrations new in 4.5, the phone's patterns hold 22.
  • Screenshots: no change yet. 4.5 routes them through omasnap and no longer reads the variable that pointed a tap at Loupe. Our override stays for 4.0.6 and is inert on 4.5. Whether omasnap's Open and Edit work by touch goes on the phone checklist.

Tested

Stand-ins only, on the laptop:

  • packages/omamobile-omarchy/test/all on omarchy 4.5.0-3 (219 checks) and on 4.0.6-1 (269)
  • packages/omamobile-shell/test/all (322), with the new test/reboot-command.sh, test/onboarding-identity.sh and test/phone-dictation.lua (phone.lua under a stub of 4.5's helpers, four cases)
  • packages/omamobile-firstboot/test/all (13)
  • tests/grep-q-pipes, tools/pr-check

Not tested, for the phone checklist

  • Both Restart buttons on a phone (4.0.6 now, 4.5 after the move).
  • One full update on 4.5: the shell restarts once at the end, plus once before pending migrations, which is Omarchy's own step and unchanged here. Whether the update sheet reopens after that earlier restart is not known.

Versions

Other open pull requests claim omamobile-shell 0.2.119 to 0.2.123 and omamobile-omarchy 0.1.42, so these take the next free numbers.

🤖 Generated with Claude Code

derekross and others added 2 commits November 11, 2026 12:53
…t 0.1.4: what Omarchy 4.5 renamed or dropped under us

An audit of our packages against the real Omarchy 4.0.6 and 4.5.0-3
packages found places where we call something 4.5 no longer has. Each fix
works on both versions, since phones stay on 4.0.6 until the pin moves.

Restart buttons (shell): the update sheet and the mobile network page ran
omarchy-system-reboot, which 4.5 replaced with "omarchy-system reboot", so
both buttons would have done nothing. They now run omamobile-reboot, which
uses whichever the installed Omarchy has, and systemd's reboot with
neither. omamobile-update's terminal hint says "omarchy reboot", an alias
both versions have.

Name and email in onboarding (shell): it ran install/user/git.sh, which
4.5 removed, printed "Saved." and set nothing. It now hands both to 4.5's
omarchy-setup-git, or without that command sets git itself and runs
Omarchy's Compose script as before, and says so when the save failed.

Name and email at first boot (firstboot): 4.5's user setup no longer
writes git's identity from the installer's answers. Firstboot writes it
itself after Omarchy's setup, under either version.

The update's restart step (omarchy): 4.5 runs it twice, --services-only
and --reboot-only. The phone's stand-in ignored the argument and did
everything twice, restarting the shell twice. Each call now does its half,
and a service that failed to restart in the first is still in the second's
warnings. 4.5's new boot verification gets the stand-in patch 0012's
header already named: a phone has no UKI or initramfs for it to check, and
slot B's images are checked by omamobile-boot-update.

Tested with stand-ins only, on the laptop: omamobile-omarchy's suite on
omarchy 4.5.0-3 (219 checks) and 4.0.6-1 (269), omamobile-shell's (319)
and omamobile-firstboot's (13). Not run on a phone.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
…its dock and kernel rows, dictation keys and default, and one more held migration

More from the 4.0.6 to 4.5.0-3 audit; each works on both versions.

Menu rows: Omarchy 4.5 moved Style to Appearance, so the rows the phone
hides (bar position, hide/show bar, one-window ratio, workspace layout,
the disk-unlock screen) came back under new ids. They are hidden under
both sets. New in 4.5 and hidden too: the dock for floating workspaces
and Update > Kernel, which would install a kernel the phone can't boot.

Dictation: on 4.5 Omarchy binds Alt + Space to its dictation command, and
a migration keeps F9 for someone who had dictation before. phone.lua takes
both away from the on-screen keyboard, as it did 4.0's F9, and no longer
adds 4.0's Voxtype F9 bind on 4.5; an F9 the owner bound to something
else is left alone. The phone's own way in stays the keyboard's Mic key
and the Dictation tile, which drive Voxtype directly. Voxtype is recorded
as Omarchy's dictation default once, for a user who has none (4.5 ships
two backends and a new user on a 4.5 image has no answer).

Migrations: 1790042972, which puts a world clock on Omarchy's bar, is
held (the phone has its own status bar). Read and left to run: 1791554699
(Omarchy's power-profiles-daemon build, which 4.5's package list installs
on a fresh image anyway), 1791123731 (four comment lines appended to
looknfeel.lua) and 1788129995 (omasnap replaces Tensaku). Of the 98
migrations new in 4.5, the phone's patterns hold 22, all boot loader,
kernel, PAM, input-method or titlebar ones.

Screenshots need no change yet: 4.5's omarchy-capture-screenshot runs
omasnap (in [omarchy] for aarch64) and no longer reads the variable that
pointed a tap at Loupe; our override stays for 4.0.6 and is inert on 4.5.
Whether omasnap's Open and Edit work by touch is a look on a 4.5 phone.

Tested with stand-ins: omamobile-shell's suite (322 checks, with
phone.lua under a stub of 4.5's helpers in four cases) and
omamobile-omarchy's on 4.5.0-3 (219) and 4.0.6-1 (269). Not run on a
phone.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
@derekross
derekross merged commit 41ad322 into main Nov 11, 2026
2 checks passed
@derekross
derekross deleted the omarchy-45-breaks branch November 11, 2026 18:01
derekross added a commit that referenced this pull request Nov 11, 2026
…was checked against (#670)

[omamobile] held omarchy and omarchy-settings at 4.0.6 while the patch
series didn't apply to 4.5 (#433, #434). The 4.5 family, the update's
stand-ins and the held migrations are in (#661, #662, #665, #482), a
rehearsal image updated from 4.0.6 in one and in two steps, and the
7 Pro took the update through its Updates sheet on 2026-11-11: exit 0,
28 patches verified, 98 migrations with 24 held and none failed,
phone-smoke passing after the restart.

The pin moves rather than goes: a phone takes a newer Omarchy only after
the series was checked against it. Phones on 4.0.6 move to 4.5.0-3 at
their next omarchy update; legacy-4.0.6/ stays for them until then.

Both files were checked against upstream's signatures.

Co-authored-by: Claude Fable 5.2 <noreply@anthropic.com>
derekross added a commit that referenced this pull request Nov 12, 2026
…hes user setup (#674)

0.1.4 (#662) wrote git's name and email in the user-setup step with a
bare ${C[FULL_NAME]} and ${C[EMAIL]}. The installer writes those keys
only when the owner gave them, so under set -u a blank one ended the step
with "unbound variable", before omamobile-setup-user, the credential
record and the lock toggles. The step is optional, so firstboot still
marked itself done and never retried. Phones that updated are unaffected:
firstboot ran on them long ago. The next root image would have carried it.

Both keys are now read with a default. The test's no-name case checked
only the exit status and that git wasn't called; it now also fails on an
unset-key message or a run that ends with warnings, and fails that way
against 0.1.4's script.

Found by the round 52 review of #662.

Co-authored-by: Claude Fable 5.2 <noreply@anthropic.com>
robertzas pushed a commit to robertzas/omarchy-mobile that referenced this pull request Nov 14, 2026
…t 0.1.4: what Omarchy 4.5 renamed or dropped under us (derekross#662)

* omamobile-shell 0.2.124, omamobile-omarchy 0.1.43, omamobile-firstboot 0.1.4: what Omarchy 4.5 renamed or dropped under us

An audit of our packages against the real Omarchy 4.0.6 and 4.5.0-3
packages found places where we call something 4.5 no longer has. Each fix
works on both versions, since phones stay on 4.0.6 until the pin moves.

Restart buttons (shell): the update sheet and the mobile network page ran
omarchy-system-reboot, which 4.5 replaced with "omarchy-system reboot", so
both buttons would have done nothing. They now run omamobile-reboot, which
uses whichever the installed Omarchy has, and systemd's reboot with
neither. omamobile-update's terminal hint says "omarchy reboot", an alias
both versions have.

Name and email in onboarding (shell): it ran install/user/git.sh, which
4.5 removed, printed "Saved." and set nothing. It now hands both to 4.5's
omarchy-setup-git, or without that command sets git itself and runs
Omarchy's Compose script as before, and says so when the save failed.

Name and email at first boot (firstboot): 4.5's user setup no longer
writes git's identity from the installer's answers. Firstboot writes it
itself after Omarchy's setup, under either version.

The update's restart step (omarchy): 4.5 runs it twice, --services-only
and --reboot-only. The phone's stand-in ignored the argument and did
everything twice, restarting the shell twice. Each call now does its half,
and a service that failed to restart in the first is still in the second's
warnings. 4.5's new boot verification gets the stand-in patch 0012's
header already named: a phone has no UKI or initramfs for it to check, and
slot B's images are checked by omamobile-boot-update.

Tested with stand-ins only, on the laptop: omamobile-omarchy's suite on
omarchy 4.5.0-3 (219 checks) and 4.0.6-1 (269), omamobile-shell's (319)
and omamobile-firstboot's (13). Not run on a phone.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>

* omamobile-shell, omamobile-omarchy: Omarchy 4.5's renamed menu rows, its dock and kernel rows, dictation keys and default, and one more held migration

More from the 4.0.6 to 4.5.0-3 audit; each works on both versions.

Menu rows: Omarchy 4.5 moved Style to Appearance, so the rows the phone
hides (bar position, hide/show bar, one-window ratio, workspace layout,
the disk-unlock screen) came back under new ids. They are hidden under
both sets. New in 4.5 and hidden too: the dock for floating workspaces
and Update > Kernel, which would install a kernel the phone can't boot.

Dictation: on 4.5 Omarchy binds Alt + Space to its dictation command, and
a migration keeps F9 for someone who had dictation before. phone.lua takes
both away from the on-screen keyboard, as it did 4.0's F9, and no longer
adds 4.0's Voxtype F9 bind on 4.5; an F9 the owner bound to something
else is left alone. The phone's own way in stays the keyboard's Mic key
and the Dictation tile, which drive Voxtype directly. Voxtype is recorded
as Omarchy's dictation default once, for a user who has none (4.5 ships
two backends and a new user on a 4.5 image has no answer).

Migrations: 1790042972, which puts a world clock on Omarchy's bar, is
held (the phone has its own status bar). Read and left to run: 1791554699
(Omarchy's power-profiles-daemon build, which 4.5's package list installs
on a fresh image anyway), 1791123731 (four comment lines appended to
looknfeel.lua) and 1788129995 (omasnap replaces Tensaku). Of the 98
migrations new in 4.5, the phone's patterns hold 22, all boot loader,
kernel, PAM, input-method or titlebar ones.

Screenshots need no change yet: 4.5's omarchy-capture-screenshot runs
omasnap (in [omarchy] for aarch64) and no longer reads the variable that
pointed a tap at Loupe; our override stays for 4.0.6 and is inert on 4.5.
Whether omasnap's Open and Edit work by touch is a look on a 4.5 phone.

Tested with stand-ins: omamobile-shell's suite (322 checks, with
phone.lua under a stub of 4.5's helpers in four cases) and
omamobile-omarchy's on 4.5.0-3 (219) and 4.0.6-1 (269). Not run on a
phone.

Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>

---------

Co-authored-by: Claude Fable 5.2 <noreply@anthropic.com>
robertzas pushed a commit to robertzas/omarchy-mobile that referenced this pull request Nov 14, 2026
…was checked against (derekross#670)

[omamobile] held omarchy and omarchy-settings at 4.0.6 while the patch
series didn't apply to 4.5 (derekross#433, derekross#434). The 4.5 family, the update's
stand-ins and the held migrations are in (derekross#661, derekross#662, derekross#665, derekross#482), a
rehearsal image updated from 4.0.6 in one and in two steps, and the
7 Pro took the update through its Updates sheet on 2026-11-11: exit 0,
28 patches verified, 98 migrations with 24 held and none failed,
phone-smoke passing after the restart.

The pin moves rather than goes: a phone takes a newer Omarchy only after
the series was checked against it. Phones on 4.0.6 move to 4.5.0-3 at
their next omarchy update; legacy-4.0.6/ stays for them until then.

Both files were checked against upstream's signatures.

Co-authored-by: Claude Fable 5.2 <noreply@anthropic.com>
robertzas pushed a commit to robertzas/omarchy-mobile that referenced this pull request Nov 14, 2026
…hes user setup (derekross#674)

0.1.4 (derekross#662) wrote git's name and email in the user-setup step with a
bare ${C[FULL_NAME]} and ${C[EMAIL]}. The installer writes those keys
only when the owner gave them, so under set -u a blank one ended the step
with "unbound variable", before omamobile-setup-user, the credential
record and the lock toggles. The step is optional, so firstboot still
marked itself done and never retried. Phones that updated are unaffected:
firstboot ran on them long ago. The next root image would have carried it.

Both keys are now read with a default. The test's no-name case checked
only the exit status and that git wasn't called; it now also fails on an
unset-key message or a run that ends with warnings, and fails that way
against 0.1.4's script.

Found by the round 52 review of derekross#662.

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.

1 participant