| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
| |
The generated file declares (provide 'gallery-tokens), but `require` resolves a feature by filename, so as tokens.el it could never load from a load-path. The name now matches the feature. gen_tokens.py, the renderer's load, and the README follow.
|
| |
|
|
|
|
|
|
|
|
| |
gallery-widget.el is the proof-of-concept Emacs renderer: it reads the generated tokens.el and renders the needle gauge (card 10) as SVG via svg.el. Rasterized through librsvg (the renderer Emacs itself uses), it matches the web card side by side: same arc, ticks, glowing amber needle, half-dome hub, and readout from the same tokens. That closes the riskiest unknown in the three-target plan, so the component pattern can now scale widget by widget.
The comparison surfaced a real gallery bug: the web gauge's ticks were invisible. Their transform-origin put the rotation center 40px below the dial, so the rotated ticks landed outside the clipped area. I moved the ticks to the arc's top edge with the origin at the pivot, and all three now show at -60/0/+60.
The parseability test caught a second bug: the mono token's font stack carries double quotes, which broke the font-family XML attribute. The renderer swaps them to single quotes.
Tests are ERT (tests/gallery-widgets/), covering token resolution, angle math (normal/boundary/error), and the rendered document's structure. make test-unit now runs the elisp suites alongside the python ones, and make test-elisp runs them alone.
|
| |
|
|
|
|
|
|
|
|
|
|
| |
The colors, glows, gradient ramp, and pulse rate were hardcoded literals scattered across ~40 sites in the gallery. Retuning the amber meant a find-and-replace through the whole file. I pulled them into tokens.json as the single source.
gen_tokens.py reads that source and emits three targets: web CSS custom properties (into the gallery :root, between markers), waybar GTK @define-color declarations, and an elisp alist for the future svg.el renderer. The three genuinely differ (CSS --vars with rgb-triple glows, GTK @define-color with underscore names, elisp hex alist), which is why one generator beats three hand-maintained copies.
The amber hue is defined once, and its glow triples are derived from the hex, so retuning it means editing one line and rerunning the generator.
The gallery render is unchanged. I verified it pixel-identical against the prior commit with reduced-motion forced and the live clock masked (diff of 0).
Tests cover hex conversion, the three emitters, marker replacement, and idempotency: 27 tests, 100% line coverage, wired into make test-unit.
|
| |
|
|
| |
One touch buys exactly one transaction. The maintenance console's forced live update sets the sentinel and clears it afterward, but a caller that dies mid-update would leave the file behind and keep the guard disarmed until reboot. The hook now deletes the sentinel at the moment it honors it, so a missed cleanup costs nothing. The env override is unchanged, and the caller's clear step stays as the belt-and-suspenders for transactions the hook never fires on.
|
| |
|
|
|
|
|
|
| |
The hypr-live-update-guard hook blocked on target names alone, so a same-version reinstall of hyprland (the maintenance console's integrity REINSTALL, right after a full update) aborted identically to a real upgrade. A pure reinstall writes identical bytes over identical bytes and can't crash the live compositor: the SIGABRT hazard needs the on-disk version to actually change.
The guard now compares each target's installed version (pacman -Q) against the sync-db candidate (expac -S %v). Both are read-only queries and safe under the transaction lock, pinned live with db.lck present. Targets that match pass as reinstalls, and the block message lists only the version-changing packages. Unresolvable versions (AUR targets, -U transactions, expac missing) still block conservatively.
Verified live on ratio: the previously blocked six-package reinstall ran clean with Hyprland up. The block path is pinned by the test suite (13 tests, four new). I deployed the script to /usr/local/bin on both hosts by hand since the installer only copies it on fresh systems.
|
| |
|
|
|
|
| |
pacman --needed skips a package that is already on the system as a dependency and leaves its install reason alone. A declared package can then sit as asdeps, surface as an orphan once its accidental dependent leaves, and get swept away by an orphan cleanup. expac and lm_sensors, both maintenance-console runtime deps, nearly went this way: they were asdeps on both hosts and only survived today's orphan sweep because something still depended on them.
pacman_install now marks the package explicit after a successful install. The mark never runs on a failed install, and a mark failure never fails the install. I flipped expac and lm_sensors to explicit on ratio and velox by hand since the installer only runs on fresh systems.
|
| |
|
|
|
|
|
|
| |
The maint console and the system-health-check workflow grade against ~/.config/archsetup/maintenance-thresholds.toml, but nothing installed it. A fresh install got a working maint CLI and no thresholds file.
user_customizations now installs the shipped TOML from the cloned repo. A re-run refreshes it, and the user layer in ~/.config/maint/ is never touched, so a reinstall can't eat curation. It also enables maint-scan.timer and maint-net-scan.timer through wants-symlinks (systemctl --user has no session bus during install, the syncthing idiom). Timer enablement is hyprland-only because the units live in that stow tier.
The install lists gain the runtime tools the probes call but nothing carried: expac, lm_sensors, fwupd.
|
| |
|
|
|
|
| |
Nine break/fix/assert scenarios run the real maint fix inside the test VM. The runner batches non-conflicting scenarios into one VM boot, restores the snapshot only between groups, and leaves the base image pristine afterwards. A systemd-nspawn fast lane runs the pacman-level group against a cached pacstrap rootfs in seconds. The plan layer validates the scenario contract without a VM and carries a 19-test unit suite. make test-maint wires the VM lane in.
The zfs lane is filtered but unexercised: the FS_PROFILE=zfs base image fails to build (ZFS DKMS module not found on linux-lts 6.18). The failure is tracked in the todo.
|
| |
|
|
| |
run_network_diagnostics tested HTTP before DNS and returned on the first failure, so a DNS failure surfaced as a generic "no internet" and the IP/route/resolver evidence was never reached. It now collects read-only facts first (ip -brief addr, default route, resolv.conf) and prints them regardless of outcome, runs every check, and reports all failures together in a summary. Generic checks (DNS, egress, TLS) are split from Arch-specific ones (mirror, AUR) so a DNS failure is named as DNS, not a mirror problem. Raw fact output is saved to the results dir when one is set.
|
| |
|
|
| |
scripts/import-wireguard-configs.sh imports the seven Proton configs into NetworkManager with autoconnect forced off. Each config stages through a wgpvpn.conf temp copy (NM's import name must be a valid interface name; several config names exceed the 15-char limit) and is renamed by the UUID parsed from the import output, so a stray same-named connection can't be hit. A leftover wgpvpn connection — a run that died between import and rename, autoconnect still armed — makes the script refuse to run. 10 tests over a fake nmcli; velox migration verified (all seven wireguard, autoconnect no). The tunnels spec is implemented: all six phases shipped.
|
| |
|
|
| |
Installing Hyprland on an NVIDIA box silently produced a rough Wayland session. nvidia_preflight_report detects the card via modalias (DRM, then PCI display-class only), prints the required env vars and the pre-Turing AUR caveat, and checks the repo's nvidia-utils major against 535. preflight_checks aborts when the requirement can't be met and asks before continuing on a detected card. 9 unit tests over fake modalias trees and a fake pacman.
|
| |
|
|
|
|
| |
archsetup took sanoid from install-archzfs but never ported the pre-pacman snapshot hook, so a ZFS-root install had no transaction-triggered rollback point — the working setup only existed as a hand-placed script on velox, lost on reinstall. Add configure_pre_pacman_snapshots(): a PreTransaction pacman hook plus a self-pruning script that keeps the 10 most recent pre-pacman snapshots (sanoid ignores them — they aren't autosnap_ names). It's gated to ZFS-root and runs late in boot_ux, so the hook doesn't fire during the install's own package operations and the first snapshot is the fresh system.
The script ships as scripts/zfs-pre-snapshot, made ZFS_PRE_* env-overridable so the pruning logic is unit-testable. Unit tests drive it against a fake zfs (creates a snapshot, prunes the oldest past KEEP, ignores non-pre-pacman snapshots, honors the lockfile interval, warns on failure); a Testinfra test asserts the hook and script land on a ZFS install; the orchestrator test pins the new boot_ux substep.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
A pacman -Syu that swaps mesa/hyprland/wayland runtime libs out from under
a running Hyprland session crashes the compositor: the next GPU-lib call
hits a now-"(deleted)" library and SIGABRTs, taking the Wayland clients
with it (hit ratio 2026-06-07, mesa + hyprland upgraded live). It's a
likely driver of ratio's high unsafe-shutdown ratio.
I added a pacman PreTransaction hook (hypr-live-update-guard) on the
GPU/compositor runtime set. When such an upgrade is pending and Hyprland is
running, it aborts before any package is swapped and tells the user to
re-run from a TTY with the session stopped. Aborting at PreTransaction is
safe: nothing is replaced yet, so the live session is untouched. With no
Hyprland running (the from-a-TTY path) the guard stays quiet and the
upgrade proceeds. Override with HYPR_ALLOW_LIVE_UPDATE=1 or by touching the
sentinel file named in the abort message.
archsetup installs the guard and hook in the hyprland path. The decision
logic is covered by tests/hypr-live-update-guard (running/not, override,
multi-package, empty-target). The hook firing against a real pacman
transaction needs a live Hyprland session, filed as a manual test.
|
| |
|
|
|
|
|
|
| |
The decomposition left each big step function as a thin list of sub-step
calls with no runtime coverage. These tests sed-extract each orchestrator,
stub its sub-functions as recorders, and assert the exact call order, so a
dropped or reordered step fails the suite. configure_snapshots also gets a
per-filesystem dispatch check (zfs / btrfs / other).
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The installer announced, ran, and warned on each operation with a
hand-written two-line pair, repeated ~35 times:
action="enabling rngd service" && display "task" "$action"
systemctl enable rngd >> "$logfile" 2>&1 || error_warn "$action" "$?"
I added a run_task "desc" cmd... helper that does this in one line, plus an
enable_service wrapper for the canonical "enabling <unit> service" case. The
35 single-command sites now call run_task. The three exact-wording service
enables (rngd, upower, fail2ban) use enable_service. Multi-line sites
(heredocs, subshells, intervening logic) keep the explicit form.
Behavior is unchanged: same descriptions, same commands, same logfile
redirection, same non-fatal warning on the real exit code. tests/run-task
covers the helper across Normal/Boundary/Error including exit-code
propagation, and the full unit suite stays green.
|
| |
|
|
|
|
|
|
| |
Add a backup_system_file helper that snapshots a pre-existing file to <path>.archsetup.bak before archsetup edits it in place, so a botched edit to fstab, mkinitcpio.conf, or sudoers is recoverable. It is idempotent: it never overwrites an existing backup, so the pristine original survives repeated edits within a run and across re-runs. It uses cp -p to preserve mode and ownership.
Only the in-place sed and append edits to pre-existing files route through it (locale.gen, makepkg.conf, pacman.conf, sudoers, wireless-regdom, geoclue.conf, pacman-contrib, fstab, mkinitcpio.conf, vconsole.conf). The brand-new drop-in files archsetup fully owns are skipped: there is no prior state to save, and recovery is just deleting them.
Covered by tests/backup-system-file/ (Normal, Boundary, Error cases, including mode preservation and the no-overwrite guarantee).
|
| |
|
|
|
|
|
|
|
|
| |
Arch's rolling repo ships zig 0.16+, but ghostel's native-module compile fallback needs exactly 0.15.2: ghostel pins ghostty 1.3.2-dev, whose build does requireZig(0.15.2), and 0.16's build-API changes break the dependency build scripts. So a plain pacman -S zig produces a zig that can't build ghostel.
install_zig_pin downloads zig-x86_64-linux-0.15.2.tar.xz from ziglang.org, verifies the sha256, extracts to /opt/zig-0.15.2, and symlinks /usr/local/bin/zig ahead of /usr/bin on PATH, where pacman -Syu can't bump it. I split the verify-and-install core (zig_install_from_tarball) out so it stays network-free and unit-testable: it refuses on a sha256 mismatch, a missing tarball, or a tree with no zig binary, and short-circuits when a correct install already exists.
ghostel's default path downloads a prebuilt module and needs no zig, so this only matters for the offline compile fallback. The pin needs a one-line bump (ZIG_VERSION + ZIG_SHA256) whenever ghostel moves to a newer ghostty.
Tests live in tests/zig-pin/: 7 cases covering extract+symlink, idempotency, sha256-mismatch refusal, missing tarball, and no-binary cleanup, run against the real function extracted from the script.
|
| |
|
|
|
|
|
|
| |
package-inventory compares archsetup's declared packages against the live system but had no tests, so a future archsetup edit (a new for-loop shape, a renamed install helper) could silently break the extraction.
I added two env seams so the script is testable without the real system. PKGINV_ARCHSETUP points the extractor at a fixture installer, PKGINV_PACMAN swaps in a fake pacman serving controlled query output. Both default to the real targets, so normal use is unchanged, and the seams match the env-override pattern audit-packages.sh already uses.
The 7 tests pin the extraction (direct calls, for-loop lists, variable-arg skip) and both diff directions against the fixture, with no network or real pacman db. I also added a make package-diff target so the tool is reachable alongside the test targets.
|
| |
|
|
|
|
| |
scripts/audit-packages.sh extracts every pacman_install/aur_install package (loop lists included) and verifies each against its declared source — sync dbs for official, one batched RPC query for AUR — flagging movers in both directions. Unit-tested against fixture installers with fake pacman/curl.
First real run over 420 packages found four that vanished from both sources, each now fixed: libva-mesa-driver folded into mesa (line dropped), nvidia-dkms replaced by nvidia-open-dkms (Turing+; legacy cards are the preflight task's problem), swww replaced by awww (its successor, already what both machines run), and libappindicator-gtk3 replaced by libayatana-appindicator. Fifteen AUR entries that graduated to official repos still install fine via yay and are left as-is.
|
| |
|
|
|
|
|
|
| |
Since the installer clones DOTFILES_REPO into ~/.dotfiles and stows from there, the in-repo dotfiles/ tree was dead weight. Nothing reads it at install time. I removed it (831 files) now that both machines are migrated.
The Makefile's stow / restow / reset / unstow / import targets and the dotfile-script unit suites moved to the dotfiles repo. They sit alongside the scripts they manage and run standalone (cd ~/.dotfiles && make ...). This Makefile keeps the VM-integration targets and the installer-helper suite (safe-rm-rf).
I updated CLAUDE.md and README.md so stow operations run from ~/.dotfiles, and the dotfile-management, theme, and unit-test sections point at the standalone repo. The README was already describing the old in-repo model from before the installer switched to cloning. This brings it in line.
|
| |
|
|
|
|
|
|
| |
The touchpad toggle's notification was too loud, and the eight notify sounds varied by ~13 dB in RMS loudness — bug and fail came out two to three times louder than info or security.
I added a --silent flag to notify (shows the popup, plays no sound) and a NOTIFY_VOLUME knob (paplay scale, default 65536) so the master level can drop without re-encoding. toggle-touchpad now passes --silent on both enable and disable. normalize-notify-sounds.sh measures each .ogg and shifts it to a uniform -31 dB mean. It writes through the file instead of mv-ing over it, so the stow symlinks survive when the script runs against the live sound dir. I re-encoded all eight sounds to the new level.
Tests: a new tests/notify suite (12 tests) covers --silent, the volume knob, flag composition, and the error paths.
|
| |
|
|
|
|
|
|
| |
I added a laptop-only waybar button that drops the machine into a low-power state and restores it on a second click. Engaging turns wifi off, sets the CPU energy-performance preference to power, dims the backlight to 35%, and stops network-only services (tailscale, proton-vpn, avahi, cups, wsdd, geoclue, sshd, fail2ban, syncthing). Bluetooth is left alone so earbuds keep working.
Disengaging replays the state recorded when airplane mode was engaged rather than writing hardcoded defaults. A lever already in its low-power position is left untouched: wifi that was already off stays off, and a service that was already stopped isn't restarted.
The indicator hides itself on machines with no battery, so desktops never show the button. State lives in $XDG_RUNTIME_DIR/airplane-state, and the bar refreshes the moment the toggle fires via a realtime signal.
|
| |
|
|
|
|
|
|
| |
The $mod+F9 toggle and the toggle-touchpad / touchpad-auto scripts already worked, but the scripts lived only in ~/.local/bin and were never committed, and there was no way to see the touchpad's state at a glance.
I committed both scripts into the repo so stow installs them, and added a waybar indicator: a waybar-touchpad status script and a custom/touchpad module that shows a mouse glyph when the touchpad is on and a mouse-off glyph (in the dupre orange) when it's off. The scripts signal the module with pkill -RTMIN+9 waybar after each state change, so the icon updates the moment the touchpad toggles. touchpad-auto now runs at login via exec-once.
The waybar-touchpad script has unit tests under tests/waybar-touchpad/ covering the enabled, disabled, and missing-state-file cases.
|
| |
|
|
|
|
|
|
| |
Three rm -rf sites in archsetup delete paths built from variables: $state_dir for --fresh, and $source_dir/$prog_name for the git and AUR clone-retry cleanups. If a path variable were empty or malformed (preflight skipped, a degenerate git URL), the delete could expand to a top-level or otherwise unintended directory.
I added a safe_rm_rf <path> <allowed_prefix> helper that refuses to run unless the target is absolute, free of '..', deeper than a bare top-level dir, strictly inside the allowed prefix, and a real directory rather than a symlink. On the happy path it delegates to rm -rf, so successful installs are unchanged. The helper is self-contained and defined before the top-level --fresh handler, which runs before the logging helpers exist.
I covered the guard with unit tests under tests/safe-rm-rf/ that source the real function and exercise normal, boundary, and error cases against temp directories.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
tmux-util rename closes out the original six-subcommand plan. The flow:
1. fzf-pick a session from the list.
2. Prompt for a new name on stdin.
3. Bail with a useful message on empty input, same-as-old, or conflict with an existing session.
4. Otherwise `tmux rename-session -t <old> <new>` and confirm.
The conflict check uses `tmux has-session -t =<new>` with the same `=`-prefix exact-match guard as the go subcommand. Without it, tmux's default prefix matching would let `rename foo` succeed even when a session named `foobar` already exists, then surprise the user later.
5 new tests cover Normal cases (pick + rename happy path) and Boundary cases (no sessions, fzf cancel, empty new name, same-as-old no-op, conflict with existing session). The test harness's run_script grew an `stdin=` param so tests can feed the prompt input. fake-tmux picked up a rename-session handler that mutates the state file. Total suite: 48 tests, all green.
Six subcommands shipped: go, pick, ls, find, reap, rename. The original "no args prints help" requirement still holds, and the stub-test for unimplemented subcommands got removed since everything's wired now.
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
tmux-util pick lists every session ("attached"/"detached" plus name), pipes it through fzf, and attaches or switch-clients to the chosen one (matching the go subcommand's inside-vs-outside-tmux discipline).
If the user cancels fzf (Ctrl-C, Esc, empty selection), the pipeline returns empty and pick exits 0 without touching tmux state.
The new fake-fzf testing fake is driven by env vars:
- FAKE_FZF_CHOICE_LINE=<N>: return the Nth line of stdin as the selection
- FAKE_FZF_CHOICE=<string>: return the literal string (ignores stdin)
- (neither): exit 130 to simulate cancel
4 new tests cover Normal cases (pick second line, inside/outside tmux behavior) and Boundary cases (no sessions, user cancellation). Total suite: 43 tests, all green.
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
tmux-util find <pattern> walks every pane across every session, queries each pane's foreground command (`#{pane_current_command}`), and prints the location of any pane whose command matches the pattern. Output format: `<session>:<window>.<pane> <command>`, one per line.
Pattern is a basic ERE (passed through `grep -E`), so anchors and alternation work. Substring matching is the common case.
Exit code:
- 0 with output: matches found
- 1 with no output: no matches (lets you script around it)
- 2 with usage on stderr: missing or empty pattern
7 new tests cover Normal cases (single match, multi-match across sessions, format verification) and Boundary cases (no matches, no sessions, missing pattern, empty pattern). fake-tmux now parses pid:cmd entries in the state file so panes can carry a synthetic command name. Total suite: 39 tests, all green.
|
| |
|
|
|
|
|
|
|
|
|
| |
tmux-util go <name> attaches to a session named <name> if it exists, creates it otherwise. Behavior depends on whether the caller is already inside tmux:
- Outside tmux: `tmux attach-session -t <name>` (existing) or `tmux new-session -s <name> -c $PWD` (new).
- Inside tmux (TMUX env set): `tmux switch-client -t <name>` (existing) or `tmux new-session -d -s <name> -c $PWD` followed by `switch-client` (new). Attaching from inside tmux would nest sessions and break the outer view, so the inside path uses switch-client instead.
The existence check uses `tmux has-session -t =<name>` with the leading `=` to force exact-match. Without it, tmux does prefix matching, which would let `go foo` resolve to a session named `foobar`.
I added 6 new tests covering both inside/outside-tmux paths, both create/attach paths, plus error handling for missing or empty name arguments. fake-tmux picked up handlers for new-session (mutates state), attach-session and switch-client (record-only), and the `=`-prefix form of has-session. Total suite: 32 tests, all green.
|
| |
|
|
|
|
|
|
|
|
|
|
| |
tmux-util ls is an opinionated replacement for `tmux ls` with columns for state (attached/detached), name, idle time (humanized), window count, and the current pane's cwd (tilde-fied if it sits under $HOME).
The cwd query goes through `tmux display -p -t <session> '#{pane_current_path}'`, which returns the cwd of the active pane of the active window. That's close enough to "what the session is about" for a one-line summary.
Idle calculation reads `date +%s` by default and accepts an override via the TMUX_UTIL_NOW env var so tests can pin "now" to a known epoch.
12 new tests cover Normal cases (attached / detached, multiple sessions) and Boundary cases (no sessions, idle exactly at minute / hour / day boundaries, $HOME tilde). One existing dispatch test got reworked because the original stub target (`ls`) is no longer unimplemented. Total suite is 26 tests, all green.
The fake-tmux harness picked up two things along the way: real format-string parsing for `list-sessions -F` and a new handler for `display -p`. The state file format extended to include activity epoch, window count, and cwd, with sensible defaults for older 3-tuple test inputs so the reap tests keep passing untouched.
|
| |
|
|
|
|
|
|
|
|
|
|
| |
A new utility in dotfiles/common/.local/bin/ for managing tmux sessions. The eventual plan covers six subcommands (go, pick, ls, find, reap, rename). This commit ships the skeleton, the dispatch + help, and the first subcommand: reap.
reap walks every unattached tmux session whose name doesn't match $TMUX_UTIL_REAP_SKIP (default `^aiv-`), sends SIGHUP to each pane's PID (the same signal that fires when you close a terminal window), waits up to three seconds for the session to wind down, and falls back to `tmux kill-session` if anything's still alive.
Tests live under tests/tmux-util/ with the same fake-binary-on-PATH pattern layout-navigate uses. fake-tmux reads canned session state from a file and records every invocation. fake-kill records signal calls without sending them. fake-sleep is a no-op so tests don't actually wait. 14 tests cover Normal / Boundary cases for dispatch + reap. Run them with:
cd tests && python3 -m unittest tmux-util.test_tmux_util
The other five subcommands stub out for now and exit non-zero with "not implemented yet" so future TDD turns can drop them in one at a time.
|
|
|
When focus is inside a special workspace (e.g. special:stash),
layoutmsg cyclenext/cycleprev only operates within that workspace,
trapping $mod+J inside the scratchpad overlay.
Detect workspace name starting with "special:" on focus navigation
(not move), toggle the overlay off first, re-read active window
state, then fall through to the normal layout/floating branches.
Add unit tests with a fake hyprctl harness in
tests/layout-navigate/. ```
|
