Dropped-from-Arch candidates
Tracks findings from https://buildmaster.archlinux32.org/checker/summary.html's
consistency checker, which periodically flags packages with a “missing state
file” – its heuristic for “this package left official Arch’s extra/ repo
and might be worth picking up as an AUR package instead, to keep 32-bit builds
of it alive.” The checker’s heuristic is noisy (see below), so every entry
here has been verified against the live Arch package database before being
categorized – don’t trust the checker’s flag alone.
Last checked: 2026-11-05.
Summary
| Package | Status | Why | Worth maintaining for 32-bit? |
|---|---|---|---|
| libclc | False positive | Still in extra (23.1.3-4); checker caught an LLVM-toolchain staging sync mid-flight. |
N/A – nothing to do |
| polly | False positive | Same LLVM sync blip as libclc, still in extra. |
N/A – nothing to do |
| openmp | False positive | Same LLVM sync blip, still in extra. |
N/A – nothing to do |
| dgop | Renamed | Redirects to dux-gop, still in extra under the new name. |
N/A – just a rename |
| dms-shell | Renamed | extra/dux-shell explicitly replaces/conflicts with dms-shell (and dms-shell-hyprland/-niri/dux-shell-bin), providing it as an alias. |
N/A – just a rename |
| gedit-plugins | Merged | Redirects to gedit – plugins folded into the main package. |
N/A – nothing to do |
| intellij-idea-community-edition | Renamed | Redirects to intellij-idea-oss, still in extra. |
N/A – just a rename |
| chatblade | Already on AUR | Package exists on AUR (chatblade, 0.7.0-3) but is orphaned (no maintainer). Checker just confirms it left extra for AUR as intended. |
N/A – already there; orphan adoption is a manual call, not automated ([[project_orphaned_category]]) |
| python-pew | Genuinely dropped | Superseded by pipenv/stdlib venv. |
No |
| kgamma | Genuinely dropped | Plasma5-era KDE display-gamma applet, dropped with Plasma5. | Marginal |
| wacomtablet | Genuinely dropped | Plasma5-era KDE Wacom config GUI, same Plasma5 wave. | No |
| colord-kde | Genuinely dropped | Plasma5-era KDE color-management integration, same wave. | No |
| kwin-x11 | Genuinely dropped | Arch dropped KWin’s X11 backend as Plasma defaults to Wayland-only. | Maybe – only if X11-on-old-hardware is in scope |
| mutter46 | Genuinely dropped | Stale version-pinned shim, superseded by current mutter. |
No |
| qt5-feedback | Genuinely dropped | Qt5 haptic-feedback module, same Qt5/Plasma5 EOL wave. | Yes – adopted, see below |
| tea | Genuinely dropped, adoption attempted and abandoned | CLI tool for Gitea servers; dropped from extra very recently (2026-11-04). Both candidate AUR names (tea, gitea-tea) are real, already-occupied packages – see detail below. |
No – deliberately not adopted, see memory/tea.md |
| python-nose | Already on AUR | Package exists on AUR (python-nose, 1.3.7-19) but orphaned (no maintainer), itself only recently (re)submitted there. |
N/A – already there; orphan adoption is a manual call, not automated ([[project_orphaned_category]]) |
Action taken
- qt5-feedback: adopted as
maintained/qt5-feedback, full git history preserved from Arch’s own packaging git (never existed on AUR before). Seememory/qt5-feedback.mdfor the setup detail andadapted/README/maintained/category reasoning. Pending: the user pushing the first commit tossh://aur@aur.archlinux.org/qt5-feedback.git(repo doesn’t exist until that happens). - tea: adoption attempted, then abandoned entirely. Tried
teafirst – rejected on push, turned out AUR already has a real, unrelated, actively- maintained Qt6 text editor at that exact name (also installing/usr/bin/tea, a genuine binary collision, not just a name coincidence). Renamed the attempt togitea-tea– turned out that’s also real: the original packaging of this same Gitea CLI tool by a different maintainer (John K. Luebs), stale at0.7.0, not a clean adoption target either. User decision: drop it entirely rather than take over either name. Reverted all local changes and removed the erroneously-published packages fromarchlinuxaba. Seememory/tea.mdfor the full story and the “don’t retry this without re-reading” note. - Everything else: no action – either a false positive, a plain rename
with nothing to transfer, or judged not worth the maintenance burden for
a 32-bit build target. Revisit
kwin-x11if X11-on-old-hardware desktop use becomes an actual requirement for this project. - dms-shell and chatblade (new in the 2026-11-02 check): neither
needs anything –
dms-shellis just a rename (see detail below), andchatbladeis already packaged on AUR, just orphaned. Orphan adoption is a deliberate human staging step, not something this update automates. - tea and python-nose (new in the 2026-11-05 check):
teawas genuinely dropped fromextra, but turned out to collide with two real AUR packages under both candidate names – adoption attempted and then abandoned, see detail below.python-noseis already on AUR, just orphaned – same manual-staging reasoning aschatblade.
Per-package detail
Sourced findings for the 7 genuine drops – the specific commit/news post/ removal event behind each, not just an inference from the HTTP 410.
kwin-x11, kgamma, wacomtablet, colord-kde – shared cause: Plasma’s X11 drop
All four were X11-session-dependent KDE components, and all four stop at the same point in Arch’s packaging history, right before Plasma 6.8 (shipped 2026-10-14):
- Arch news,
Plasma 6.4 will need manual intervention if you are on X11 –
kwin-x11was first split out from the mainkwinpackage here, making the X11 session opt-in rather than default. - Writeup on the Plasma 6.8 X11 removal
and a CachyOS forum thread confirming it –
Plasma 6.8 dropped the X11 session entirely;
kwin-x11was never rebuilt past6.7.5-3. kgammaandwacomtabletare both X11-only KDE System Settings modules (display gamma, Wacom tablet config) that became functionally dead once there was no X11 session left to configure – Arch’s packaging history for both just stops at6.7.5-1, consistent with the same wave, no separate announcement found.colord-kde(KDE’s colord color-management integration) is X11-session- dependent the same way – packaging history stops at26.08.2-1.
dms-shell
extra/dux-shell’s package page
states it directly: “replaces several previous versions: dms-shell,
dms-shell-hyprland, dms-shell-niri, and dux-shell-bin. It also provides
dms-shell=2.0.0 as an alias and conflicts with dms-shell.” Same
d* -> dux-* rename pattern already seen with dgop -> dux-gop.
Nothing to transfer.
chatblade
Not actually dropped from Arch’s perspective the checker cares about –
it’s a Python CLI tool (A CLI Swiss Army Knife for ChatGPT) that was
never in extra as far as this check goes, and the AUR RPC confirms it
already exists there (chatblade 0.7.0-3), just with Maintainer: null.
Per [[project_orphaned_category]], orphan adoption is a manual staging
decision – not acted on here.
mutter46
wingpanel’s current PKGBUILD
now depends on mutter49. mutter46 existed only as a version-pinned
compat shim to satisfy Pantheon desktop’s wingpanel/
lightdm-pantheon-greeter, which needed an older Mutter ABI; once
wingpanel itself moved to mutter49, the pin became dead weight and was
removed.
python-pew
No active maintainer (listed “Orphan” before removal), and its functionality
is fully covered by pipenv or the stdlib venv module – pruned as
unmaintained dead weight, not a technical failure.
tea (not adopted)
Arch’s own packaging history
shows the project (“A command line tool to interact with Gitea servers”) was
last touched 2026-11-04 – a very recent drop, caught by this cycle’s scan
essentially right after it happened. pacman -Si tea confirms it’s gone
from extra.
AUR RPC claimed both tea and gitea-tea were free (resultcount: 0 for
each), but both turned out wrong – confirmed via git ls-remote/git
fetch against the real AUR SSH host:
teais a real, actively-maintained, unrelated package: a Qt6 text editor (tea.ourproject.org, maintainer Fabio ‘Lolix’ Loli), which also installs/usr/bin/tea– a genuine binary collision, not just a name clash.gitea-teais a real, legitimately-related but stale package: the original packaging of this same CLI tool by John K. Luebs, stuck at0.7.0with a bare-bones build. This explains the official PKGBUILD’sreplaces=(gitea-tea)line – it exists for the upgrade path from this package, not as a hint that the name is free.
Decision: don’t adopt under either name. See memory/tea.md for the full
investigation and cleanup.
python-nose
pacman -Si python-nose confirms it’s gone from extra, but AUR RPC shows
it already exists there (python-nose 1.3.7-19, Maintainer: null,
submitted very recently per FirstSubmitted) – someone else already did
the AUR-side work, it’s just sitting unmaintained. Per
[[project_orphaned_category]], orphan adoption is a manual staging
decision, not acted on here.
qt5-feedback
No dedicated announcement found – the only record is
Arch’s own packaging git history
(8 commits, 2015-2025, still reachable even though gone from the live
extra/ tree). A Qt5 Tactile Feedback module, dormant since a single 2015
upstream commit it was pinned to, swept up in the general Qt5/Plasma5-era
cleanup. Now adopted – see “Action taken” above.