Arch dropped the prebuilt broadcom-wl on 2026-09-02 and rebuilt broadcom-wl-dkms with replaces=(broadcom-wl). That only helps upgrades of an installed package; as an explicit pacman -Syw target the old name fails with "target not found", which broke the offline mirror download. Published runtime packages that predate the rename still list it in omarchy-other.packages, so rewrite it at build time until every channel ships a runtime naming broadcom-wl-dkms itself.
Graceful guest poweroff over SSH first, escalating to TERM then KILL;
--kill skips straight to force. Finds the newest live run by its pidfile
and confirms before acting when run from a terminal, since a harness
still driving the VM dies with it.
A VNC viewer that resizes the guest framebuffer makes the console font
unreadable to OCR, which stranded the bootstrap at its login-prompt gate.
Getty is reliably on the spare TTY and SSH coming up is the real success
signal, so type the login blind and retry the whole sequence up to three
times; a failed pass just leaves getty re-prompting.
Also print the resize-safe viewer invocation whenever a VM starts: a
resize during early boot can wedge virtio-gpu before the system is up.
--rc appended a bare -rc suffix regardless of the version passed, so
every candidate of a train uploaded over the same omarchy-X.Y.Z-rc.iso
and nobody could tell rc1 from rc2 by name or URL. A version that
already numbers its candidate (4.0.2rc2) is the complete artifact name;
the suffix remains for unnumbered candidate builds.
--sync-omarchy now pushes all of test/ (the security coverage lives in
shell.d as well as acceptance.d), and the acceptance run gets
OMARCHY_ACCEPTANCE_SUDO_PASSWORD so the suite's opt-in sshd hardening
exercise can run against the throwaway VM.
--sync-all additionally runs the CLI and shell suites inside the guest
from the synced tree. The acceptance suite always runs against
/usr/share/omarchy either way: VM runs verify the finished product, and
qs matches shell instances by config path, so pointing the suite at the
synced tree just reads the session shell as "not running".
Omarchy now binds Super+Space to the Omarchy menu and Super+Alt+Space to
the apps menu, and the omarchy.launcher plugin is gone (the apps menu
replaced it). The smoke table still expected the old launcher binding, so
both chords failed against current builds. Both surfaces open under the
omarchy-menu layer namespace.
Tesseract cannot read the highlighted "Yes, install without encryption"
button under any preprocessing, so waiting for that text always timed out
and killed unencrypted installs at the disk warning. Capture the screen and
continue instead; a Ctrl+C toggle that failed to apply still surfaces when
the unexpectedly encrypted install stalls at the LUKS prompt on first boot.
The artifacts a release exists to hand out — the download URL and the
checksum to verify it with — were buried in rclone transfer noise or
absent entirely. Print them together as the final output, after the
upload has actually succeeded (set -e means reaching this line is the
success path).
The docker group is root-equivalent, so Omarchy operators stay out of
it on purpose — which left omarchy-iso-make unable to reach the daemon
at all. Fall back to sudo docker when a plain client call cannot reach
the socket; the script already prompts for interactive sudo elsewhere.
DOCKER_ARGS (including HOST_UID/HOST_GID) are assembled before the
escalation, so output ownership is unchanged.
omarchy-iso-release ran every step unconditionally: when docker denied
the build, it fell through to signing and uploading the newest stale
*-rc.iso from release/ under the new release's name — a six-month-old
ISO shipped as omarchy-4.0.2-rc.iso. set -e ends the release at the
first failed step, and also aborts when release/ holds no ISO for the
ref at all.
pacman-online-rc.conf's [omarchy] repo moves from pkgs.omarchy.org/edge to
pkgs.omarchy.org/rc: an RC ISO now bundles and installs the release-candidate
packages against the package set stable users will get, instead of edge
churn.
Pairs with omarchy-pkgs' fold of omarchy-limine into omarchy-settings.
Changes:
- EARLY_BOOTSTRAP_PACKAGES in phases_impl.py drops omarchy-limine.
The limine configs it shipped (mkinitcpio hooks, entry-tool drop-ins,
snapper template, default/limine/, default/snapper/) now come with
omarchy-settings, which is already in the early bootstrap.
- builder/build-iso.sh's offline-mirror seed list and post-mirror filter
no longer reference omarchy-limine.
- builder/build-omarchy-packages.sh's build queue drops omarchy-limine.
- Test manifests (fresh-4.json, fresh-4-semantic.json) no longer expect
omarchy-limine in the installed-packages list. The /etc/pacman.d/hooks/
99-omarchy-limine.hook entry stays — that hook file is generated by the
orchestrator regardless, and the filename is descriptive, not tied to
the package name.
Pairs with omarchy-installer's rename of bin/omarchy-setup-user to
bin/omarchy-finalize-user. The orchestrator now invokes
omarchy-finalize-user --force --first-install in the run_chroot_finalizer
phase, and the phase-ordering docstring is updated to match.
archinstall 4.3 maps the deprecated top-level 'bootloader': 'Limine'
to BootloaderConfiguration(bootloader=Limine, removable=True) (per
lib/args.py:185-190 and lib/models/bootloader.py:56-69). Our new
_install_limine_omarchy mirrors archinstall's removable handling: it
would copy BOOTX64.EFI to EFI/BOOT (fallback path) and SKIP
efibootmgr, neither of which is what we want.
Switch to the explicit bootloader_config form with removable=false
so we take the EFI/limine path + register the 'Limine' NVRAM entry
as intended.
archinstall's add_bootloader creates EFI/arch-limine/{BOOTX64.EFI, BOOTIA32.EFI}, writes a limine.conf with an Arch Linux entry, and registers an efibootmgr label "Arch Linux Limine Bootloader" — only for limine-snapper.sh to delete it all and rebuild with Omarchy branding via limine-mkinitcpio-hook.
Replace the add_bootloader call with _install_limine_omarchy, which does the same work archinstall's _add_limine_bootloader does (UEFI EFI binary copy + efibootmgr register, or BIOS limine-bios.sys + limine bios-install) but with Omarchy-owned paths and labels from the start:
- EFI subdir: EFI/limine (not EFI/arch-limine)
- efibootmgr: label "Limine" (not "Arch Linux Limine Bootloader")
- loader path: \EFI\limine\BOOTX64.EFI
Compute the kernel cmdline directly from disk_config instead of harvesting it from a file archinstall wrote (whose path keeps drifting across versions). Then write /etc/default/limine and /etc/kernel/cmdline, and stage /boot/limine.conf from our styling template — limine-mkinitcpio-hook will add the actual boot entries when it deploys.
Adds limine + efibootmgr to EARLY_PACKAGES (archinstall used to pacstrap these as a side effect of add_bootloader).
Does not modify limine-snapper.sh — its cleanup code will simply no-op because there are no arch-limine paths or "Arch Linux Limine Bootloader" entries to find. That code can be pruned in a follow-up.
archinstall 4.3 writes limine.conf to <target>/<esp_mount>/EFI/arch-limine/
limine.conf for UEFI non-removable installs (lib/installer.py:1487), not
to /mnt/boot/limine.conf as our code assumed. Older archinstall versions
used different paths.
Probe the 4.3 paths (arch-limine, BOOT, limine subdir) and the legacy
/boot/limine.conf in order; pick whichever exists. Error message now
lists everywhere it looked so future drift is easy to diagnose.
Restore the env + redirect setup that the previously-working ISO
installer used (matching /home/ryan/Downloads/ISO scripts/.automated_script.sh):
export COLUMNS=$(tput cols)
export LINES=$(tput lines)
exec > >(tee >(sed CSI-strip >> log) 2>/dev/null) 2>/dev/tty
export CLICOLOR_FORCE=1
export FORCE_COLOR=1
The cutover commit (064b5e2) silently dropped these. Subsequent
attempts to fix the TTY/gum rendering by reshaping the redirect were
all symptom-chasing — the proven shape was correct. The key insight:
- gum draws its TUI on stderr, not stdout. 2>/dev/tty keeps that
rendering on the real terminal.
- stdout-as-pipe (via tee) is *desired*: it lets gum detect a
non-TTY stdout and skip cursor-position probes (CSI 6n). Without
that detection, the DSR replies leak as visible ^[[13;1R bytes.
- CLICOLOR_FORCE/FORCE_COLOR force gum to still emit colour despite
stdout being a pipe.
Also drops the configurator's _drain_tty workaround — unnecessary
once gum stops sending DSR queries in the first place.
bash's `read -t 0` only TESTS whether input is available; it does not
consume the bytes. So the drain loop `while read -t 0 ...; do :; done`
spun forever the moment any bytes were pending (e.g., the DSR reply
sitting in the tty input buffer after gum style sent CSI 6n).
That hung the configurator at the first step() title screen — blank
prompt, blinking cursor, no keyboard form.
Use `read -t 0.05 -N 9999` once, which consumes up to 9999 bytes with
a 50ms timeout. Always returns after 50ms whether there's data or not.
Single call replaces the while loop.
configure_protected_boot used to arch-chroot pacman -S limine
efibootmgr, but at that phase the live ISO's offline mirror is not
bind-mounted into the target yet (run_chroot_finalizer does that
later) and the target's pacman.conf points at network mirrors that
aren't reachable during an offline install.
Pacstrap these packages alongside the runtime package list in
arch_install_base, where archinstall is still pacstrapping against
the live ISO's offline mirror. Drop the now-redundant arch-chroot
pacman call from configure_protected_boot.
gum's TUI startup queries the terminal (e.g., cursor position via
CSI 6n) and doesn't always read the reply. The reply bytes
("^[[13;1R") then surface visibly next to the next step title.
A one-time drain at script start (in .automated_script.sh) catches
the boot-time replies but not the gum-per-call ones, since the
artifact came back at every new step().
Add a small _drain_tty helper and call it from step() after the
title gum-style invocation so any pending DSR replies are consumed
before the next prompt runs.
archinstall 4.3's Installer.__init__ types target as Path and uses it
directly (self.target / 'boot', etc.). Passing str(mountpoint) makes
self.target a str, so every internal path operation raises
'unsupported operand type(s) for /: str and str' — most visibly during
set_mirrors, which constructs target / MIRRORLIST.relative_to_root().
Drop the str() cast; pass the Path through as-is.
ArchConfigHandler reads --config and --creds from sys.argv via argparse at
construction time; it ignores ARCHINSTALL_CONFIG/ARCHINSTALL_CREDS env
vars (confirmed in archinstall 4.3 lib/args.py:_parse_args). Setting those
env vars was a no-op, so handler.config.disk_config was always None when
the wrapper stripped our args off the real argv.
Fix: build a synthetic argv [prog, --config <path>, --creds <path>] and
swap it into sys.argv just for the ArchConfigHandler() call, then
restore. Drops the unused 'import os' now that the env var lines are
gone.
Fix the ^[[13;1R cursor-position-report artifacts visible next to the
first configurator step. archiso's boot init leaves DSR responses in the
kernel tty input buffer; the first gum invocation reads them as stray
input and echoes them as visible escape sequences.
- Drain stdin (with a short timeout) from /dev/tty before configurator
runs so any pending replies are consumed.
- Drop the explicit </dev/tty stdin redirect — the parent script's
stdin is already tty1 (inherited from archiso autostart), and the
explicit redirect opened a new fd with subtly different termios.
All install-phase logic now lives in the Python orchestrator and the
omarchy-iso-* bash helpers. .automated_script.sh becomes a thin shim that:
- sources install helpers (env for configurator)
- sets the Tokyo Night palette on the live VT
- tees stripped stdout/stderr to /var/log/omarchy-install.log
- runs the configurator wizard
- exec's omarchy-iso-install with absolute config paths
Absolute paths are required because the wrapper cd's into
/usr/share/omarchy-iso before invoking the orchestrator.
Removed install_base_system, install_omarchy, install_arch, install_disk,
cleanup_install_disk, configure_login_for_unencrypted_install, and
chroot_run_as_user — all replaced by Python phases or bash helpers.
Bind mounts set up in run_chroot_finalizer used to be torn down in finish,
but finish doesn't run when a phase between the two raises. Move the unmount
loop into a dedicated cleanup_bind_mounts function and call it from main()'s
finally so it runs on success, failure, and KeyboardInterrupt.
The orchestrator drives an offline install on the live ISO using
archinstall as a library subsystem. Lands under configs/airootfs/ so
mkarchiso bakes it into the live squashfs directly:
/usr/share/omarchy-iso/orchestrator/ (the Python package)
/usr/local/bin/omarchy-iso-install (thin shell wrapper that
execs python -m orchestrator.main)
Two corrections vs the previous omarchy-installer location:
1. omarchy-installer dropped from EARLY_PACKAGES \u2014 it's live-ISO-only
install tooling, must never land on the target. /etc/skel is still
populated pre-useradd via omarchy-settings.
2. run_chroot_finalizer now copies /usr/share/omarchy from the live
ISO into /mnt/tmp/omarchy-install/ before arch-chroot. finalize.sh
+ install/ scripts run from /tmp inside the chroot; /tmp is wiped
on first reboot, so the target boots with zero install detritus.