Nomarchy/agent/BACKLOG.md

# Backlog — the prioritized task queue

**This is the only executable work list for agents.** Product themes and
v1.0 intent live in [`docs/VISION.md`](../docs/VISION.md); design history
in [`docs/ROADMAP.md`](../docs/ROADMAP.md); map in
[`docs/README.md`](../docs/README.md) and [`agent/README.md`](README.md).

**Rules:**
- Agents take the topmost actionable item (see LOOP.md). Finished items
  are **deleted** here — the journal + git log are the record; durable
  design notes get a ✓-entry in docs/ROADMAP.md (and/or a note in VISION)
  if worth keeping.
- Item numbers are **stable IDs** — never renumbered or reused. A gap in
  the sequence means shipped (or dropped) work; new items take the next
  free number regardless of tier.
- Tags: `[blocked:hw]` needs real hardware (see HARDWARE-QUEUE.md) ·
  `[human]` needs Bernardo · `[stuck]` two failed attempts, needs help ·
  `[big]` must be split before starting.
- Agents may append to **PROPOSED** and **Decisions** freely (include
  `VISION § …` or `ROADMAP § …` when relevant); only Bernardo moves items
  *out* of PROPOSED into the tiers.

---

## NOW

### Live ISO / install hardware findings — Acer Aspire M5-481T + Dell XPS 9350

Bernardo, real installs 2026-07-13–14 (photos of the install end screens and
post-boot sessions). Preserve separation: the installer bake failure and the
flake pin are different root causes even when they show up on the same machine.
(Terminal / Ghostty-on-Acer → shipped as Kitty-only, #95.)

### 127. `[blocked:hw]` Docked idle: black screens, no wake, undock does not restore panel

Bernardo 2026-07-15 (AMD dev box, docked): left the laptop docked; after a
few minutes of idle it appeared to "sleep". Then:

1. External keyboard/mouse did **not** wake any display.
2. Unplugging the external monitor did **not** turn the laptop panel back
   on (dock mode has eDP disabled — undock *should* re-enable it).
3. The machine was still alive: Caps Lock LED toggled on the keyboard.

This also killed a long agent session (looked like a "crash" from the
outside). Separate from awake undock recovery (HARDWARE-QUEUE round 8):
the failure is **idle while docked**, then no path back to a usable panel.

**Likely shape (code-side, unconfirmed on the incident):** on AC, hypridle
does **not** call `systemctl suspend` (`onAc || suspend` — only battery
suspends at 15 min). Idle path is lock @5 min + `dpms off` @10 min
(`modules/home/idle.nix`). Caps Lock working fits **DPMS/lock blackout**
better than deep S3. Dock mode has already disabled eDP; recovery then
needs either (a) input → hypridle `on-resume` / `dpms on` on the external,
or (b) undock → `nomarchy-display-transition undock` re-enabling the
panel. Both failed. `after_sleep_cmd` only does `hyprctl dispatch dpms on`
and never re-enables a disabled internal.

**Progress 2026-07-15:**
- Recovery tooling: undock ends with `dpms on`; `nomarchy-display-wake` /
  `nomarchy-display-dump` (SUPER+SHIFT+D; TTY if it works).
- **Mitigation (shipped):** do **not** DPMS-off when docked — if any
  laptop internal (eDP/LVDS/DSI) exists in `monitors all` but is not
  enabled, the 10 min listener skips blanking. Lock at 5 min still runs;
  undocked laptops still DPMS-off as before. Bernardo: Ctrl+Alt+F3 did
  not yield a usable TTY on the original brick, so prevention > dump path.
- **Revisit later (LATER):** intentional DPMS-when-docked once wake is
  proven trustworthy (monitor power-save without a brick). Tracked as
  **#135** (state key + the surface to delete this by) and **#136** (the
  gate is clamshell, not "docked" — lid-open docked still blanks). Both
  are downstream of the hardware test below.
- **Caveat the test must settle:** the mitigation is a cure for an
  **unconfirmed cause**. "Blanking the only live output caused the brick"
  is inference from one incident, shipped at V1. If DPMS-off is not the
  trigger, we are paying a permanent behaviour cost (docked externals never
  sleep) and the brick is still live — so confirming the cause outranks
  #135/#136.

**Pass (current mitigation):** docked on AC, idle past 10+ min → external
stays lit under hyprlock (journal: `dpms-off skipped: docked`); undocked
still blanks at 10 min and wakes on input. Full original pass (wake after
DPMS-while-docked) deferred to the LATER revisit.

**Fallback if the cause resists (Bernardo 2026-07-16):** diagnosing stays the
preference — DPMS when docked is the outcome we want. But the mitigation as
shipped holds a near-static lock screen on the externals indefinitely, so if
the cause cannot be pinned down, a **screensaver while docked** is the
consolation prize: it protects the panels (burn-in — the same cost #135
names) without re-entering the blanking path that bricked the session. Not a
new item: it is the "if diagnosis fails" branch of this one, and #135's state
key is where it would land.

## NEXT

### 135. Give the #127 DPMS mitigation a state key (and a way back)

The shipped mitigation is hardcoded policy: docked (clamshell) externals now
**never** blank, and hyprlock has no timeout of its own, so they stay lit
indefinitely. Two real costs land on the user — a couple of 27" panels at
~60–100W all night, and **OLED burn-in** from a near-static lock screen held
for hours. Today there is no way to opt back in, which also breaks the
in-flake-state rule that settings are menu-writable.

Add `settings.idle.dpmsWhenDocked` (default **false** = today's safe
behaviour), read by `modules/home/idle.nix`'s `dpmsOff` gate, with a
Preferences row. The real point is not the option: it gives the mitigation a
**named surface to delete** once wake is proven, instead of an unowned LATER
note that quietly becomes the product. Whoever removes the mitigation removes
the key in the same commit.

**Order:** after the #127 hardware test. If DPMS-off turns out not to be the
trigger, the mitigation is reverted wholesale and this key never exists —
building it first risks shipping an option for a policy we withdraw.

### 136. `[human]` #127's gate is clamshell, not "docked"

The code skips DPMS-off when a laptop internal is present in `monitors all`
but absent from the enabled list — i.e. **clamshell**. Commit 060bf52, the
#127 entry and the journal all say "docked", which is broader: **docked with
the lid open** (internal + externals all enabled) falls straight through to
`hyprctl dispatch dpms off` and still blanks. So either

- the narrow scope is deliberate (the brick needs the internal disabled, and
  lid-open docked is genuinely safe) → fix the wording in 060bf52's
  descendants so the docs stop claiming more than the code does; or
- it is an accidental gap → the mitigation misses half the docked cases and
  the brick can still fire with the lid open.

Bernardo's call, and the #127 hardware test is what settles it: if the brick
reproduces lid-open, this is a gap. Cheap either way — one `jq` predicate or
one docs sweep. Filed because a doc that overclaims its own fix is how the
next session mis-diagnoses the recurrence.

### 129. Permanent guard for the GTK button-colour trap (#98 follow-up)

The #98 regression (destructive-action at 1.03:1, invisible) was caught only
by rendering a dialog — every scripted check was green, because the bug lives
in adw-gtk3's `mix(@destructive_color, alpha(currentColor,…))` background, not
in our CSS. `tools/theme-shot.nix` is deliberately not a CI gate (full-desktop
VM), so add the cheap static half: assert that for every theme, any rule that
pins a button label colour also pins that button's `background-color` — i.e.
no pinned label may sit on a `currentColor`-derived background. Guards the
class, not the instance. The scratch render harness that found it is worth
promoting to `tools/dialog-shot.nix` as a maintainer tool alongside theme-shot.

### 130. GTK accent buttons sit at ~2.7:1 in light themes

Fallout noted while fixing #98: in light palettes the suggested/destructive
labels are base00 (cream) on saturated accent/bad — summer-day measures
**2.72:1** (suggested) and **2.74:1** (destructive), under AA 4.5 and even
under the 3:1 large-text floor. Not a regression: upstream adw-gtk3 does the
same (`color: white` on the accent), so we currently match GNOME. Decide
whether Nomarchy holds itself higher — darker label, or a darker accent mix
for light palettes — and note `tools/check-theme-contrast.py` covers palette
pairings only, never GTK widget surfaces. `[human]` for the aesthetic call.

### 132. VM cannot judge rofi menu geometry

The guest renders the picker with icons at roughly a quarter of `ui.iconSize`
(44 → ~10px) and rows short enough that a 6-row root clips its last entry
behind a scrollbar, though `lines = 8` and `dynamic = true` should fit it.
Root is unchanged since before #105, so this is the guest's font/icon
environment, not a regression — but it means `theme-shot.nix`/`menu-shot`
screenshots are trustworthy for **menu content** and not for spacing, icon
size, or truncation. Either fix the guest (fontconfig + icon cache in the
test node) or document the gap in docs/TESTING.md §5 so the next agent does
not chase it — right now nothing warns them.

### 133. No check pins the #107 legacy-name compat shim

`lib.mkFlake`, doctor and lifecycle all still accept `theme-state.json`, and
`nomarchy-state-sync` migrates it on write — but nothing in `checks.*`
evaluates a checkout that has *only* the legacy name, so the shim can rot
silently while every check stays green. It is load-bearing until every
existing machine has taken one menu write. Add a cheap eval check: a fixture
dir with `theme-state.json` only → `mkFlake` resolves it; plus the write-side
migration (state.json created + git-staged, legacy `git rm`'d). Proven by
hand 2026-07-15 (drvPaths identical across the rename) — this just keeps it
proven. Delete together with the shim + the `nomarchy-theme-sync` alias.

### 137. Plymouth splash draws on head 0's geometry, so it is off-centre when docked

Bernardo 2026-07-16 (T14s, docked): the boot **and** shutdown logo is not
centred on the external monitor.

**The first diagnosis here was wrong — do not act on it.** It claimed
`Window.GetWidth()` returns head 0 and that the fix was per-head sprites. Both
are false, and per-head sprites would have been a *regression* (duplicate
logos). Read from plymouth's own source (`plymouth.src`, the build nixpkgs
ships) rather than from a summary:

- `Window.GetWidth()` **without an index returns `data->max_width`**
  (`script-lib-sprite.c:227`) — the widest head, not head 0.
- The script plugin builds a virtual canvas of `max_width × max_height` and
  **centres every display inside it**: `display->x = (max_width - width) / 2`
  (`:536`). Sprites are drawn per display at `sprite->x - display->x` (`:452`).
- Therefore a canvas-centred sprite **is** centred on every head, and today's
  `Window.GetWidth()/2 - w/2` is *correct at the instant it runs*. Note the DRM
  renderer sets every head's `area.x = 0` (`renderers/drm/plugin.c:612`) — the
  origins that matter are the plugin's computed ones, not the crtc's.

**The real bug is time, not geometry.** `nomarchy.script` computes every
position as a **top-level statement at parse time** (`:9`, `:15–16`, `:129`,
`:210`, `:219`), i.e. once, against whatever canvas existed at that moment.
When a display arrives or leaves *during* the splash — which is exactly
"booting up / shutting down **with** an external" — `max_width/max_height`
change, the plugin re-centres each display (`:532-537`), but the sprite's frozen
canvas coordinate does not move with it. Every element then sits off by
`(new_max - old_max) / 2` **on every head**. A single monitor never changes the
canvas, which is precisely why this has never been seen undocked.

**Fix shape:** hoist the position math into a `reposition()` and call it from
the existing `refresh_callback` (already running at 50 fps) whenever
`Window.GetWidth()/GetHeight()` differ from the last canvas — logo (scale
included), entry, lock, any live bullets, progress box/bar. Keep it a
canvas-change check, not an every-frame recompute.

**The rig works, needs no root, and is in `scratchpad/ply-rig.sh`** (recipe
below — rebuild it, it is ~20 lines). Rendered on the dev box 2026-07-16:

- `unshare -rm` gives a user + mount namespace: root *in there* only, every
  mount invisible to the real system and gone when it exits.
- `/run/plymouth` is BOTH the daemon's socket dir **and** its compiled-in
  plugin/theme path (`strings plymouthd` → `/run/plymouth/plugins/`). Bind a
  writable dir over it **and recreate its three symlinks** (`plugins`,
  `themes`, `plymouthd.defaults` → `/etc/plymouth/…`) or plymouth cannot find
  `x11.so`, silently skips X11 and reaches for a tty.
- **Do not pass `--tty`**: `main.c:2133` only takes the X11 path while
  `default_tty` is unset. Bind an empty dir over `/dev/dri` too — with X11
  first in the renderer order (`ply-renderer.c:283`) DRM is never used, and the
  mask makes that a guarantee rather than a hope.
- `PLY_CREATE_FAKE_MULTI_HEAD_SETUP=1` → 800×600 + 640×480 windows.
  `plymouth show-splash`, then `plymouth ask-for-password` renders the LUKS
  dialog. Debug: `--kernel-command-line="plymouth.debug=stream:/path splash"`
  (**`stream:`**, not `file:` — `file:` buffers and you get nothing).
- Fast loop: copy the theme into the rig's `themes/` and sed its `.plymouth`
  `ScriptFile=`/`ImageDir=` to the copy — then editing the script is a 10s
  render instead of a 3min `nixos-rebuild`.

**What the rig already proved (2026-07-16):** with a *stable* two-head canvas
the **current** theme centres correctly on both heads — logo and progress bar
each in their own head's middle. So the geometry is not the bug, which is the
evidence for the timing diagnosis above. The fake path builds both heads up
front, so it cannot reproduce a canvas *change*: for that, plug/unplug while
the splash is up, or compare a boot with and without the external.

**Attempt 1 failed, and the rig caught it — that is the point.** Hoisting the
positions into a `layout()` called from `refresh_callback` rendered the
background on both heads and **nothing else**: no logo, no progress bar. The
script plugin logs `starting boot animation` and **no parse error**, so it
fails *silently* — reverted rather than shipped (that script draws the LUKS box
on an encrypted machine). Next session: bisect with the fast loop. Prime
suspect is scope — plymouth script assigns into `local` unless the name already
exists, so a `logo.image = …` inside a function may not be the `logo` the rest
of the script sees, even though the same pattern works in `update_progress_bar`.
Test that hypothesis *first*, in isolation, before rewriting anything.

Pass = with the rig, a two-head render centred on both **plus the password
dialog visible**; then a real docked boot + shutdown, undocked unchanged.

### 143. Rofi keeps the selected *row number* when the search changes, so the highlight lands on an unrelated entry

Bernardo 2026-07-16, and his repro is exact: open the app menu, type `ca`, move
the selection two items right (onto ChatGPT), press backspace once (query is
now `c`). The list re-expands — first entry is now Slack — but the highlight
stays on the **third row**, which is now GHex, an app that has nothing to do
with what he typed. Enter at that moment launches the wrong thing.

**Cause — upstream rofi (2.0.0, `source/widgets/listview.c`), not our config.**
A filter change ends in `listview_set_num_elements()`, which does not reset the
selection; it re-asserts the old one:

```c
lv->req_elements = rows;
listview_set_selected(lv, lv->selected);   /* the *previous* index */
```

and `listview_set_selected()` only clamps it to the new bounds:

```c
lv->selected = MIN(selected, lv->req_elements - 1);
```

So what is preserved is the **row number**, not the entry — nothing checks that
row 2 still holds the thing that was selected, or that it matches the query at
all. Typing *more* hides this (the list shrinks, the clamp usually drags the
highlight along); **backspacing exposes it**, because the list grows back
underneath a stationary index. There is no option for it: `-dump-config` has
nothing about selection-on-refilter, and `auto-select` is a different feature
(it fires only when exactly one row survives).

**One correction to the report:** it is not the app menu. `listview` is the
widget behind **every** rofi surface we ship, so our text menus (Recovery,
System, Preferences…) do the same. It is merely far more visible in drun, where
the list is long and — because we set `sort = true` — reorders on every
keystroke, so a stationary index points somewhere new each time. In a short
text menu you would probably see it and correct without noticing why.

**Reported upstream 2026-07-16: https://github.com/davatorium/rofi/issues/2317**
(open, filed as `bemagri`, with the 2.0.0 line numbers and the observation that
`rofi_view_set_selected_line` already maps a stable id through `line_map` — so
the fix reuses an existing primitive rather than inventing one). No patch was
sent: the scoping call is the maintainer's, since keeping the index across a
*reload* looks deliberate (#1064 asked for exactly that, and got it in 1.7.4).
No duplicate found before filing.

**Options.** (a) Wait for upstream — done, now their move. (b) Carry the patch
ourselves in an overlay — a few lines, but it means **building rofi from source
on every user's machine**: our derivations are in no binary cache (#120's
gotcha), so this trades a real install/rebuild cost for it. (c) Live with it
and say nothing — the status quo, and the reason Bernardo hit it cold.
Recommend (a) + (b) together if the patch stays small: file it upstream, carry
it until it lands, delete on the release that fixes it. `[human]` for whether
a source-built rofi is a price worth paying.

### 120. A netinstall ISO, next to the fat offline one

Bernardo 2026-07-14, after seeing the measured size: **keep the current ISO
exactly as it is** — the guaranteed offline install is the feature it buys —
and ship a **much lighter netinstall variant alongside it**. Two products, one
distro: "works on a plane" and "8 GiB is absurd to download" are both true, and
a second target settles them without compromising either.

**Measured facts (2026-07-14), so this starts from numbers, not vibes.**
*(These stand as measured: #121 would have cut ~0.67 GiB of duplicate chromium
from them, but it was **reverted** — decided against, ROADMAP § one chromium,
not two. If a netinstall ships, revisit it: the duplicate is worth ~195 MiB of
**download**, which is this item's whole currency, even though it is worth
almost nothing on disk or on the ISO.)*

> **Read this before using the numbers below.** They are **closure arithmetic**,
> and #121 proved the hard way that closure size is neither disk size nor image
> size: removing a 687 MiB path shrank the ISO by **8 KiB**, because
> **mksquashfs dedupes duplicate files** and **`auto-optimise-store` hardlinks**
> them on disk. So a change that looks like it sheds gigabytes of closure can
> shed nothing off the actual image. **Measure the artifact — build the ISO and
> `stat` it.** The corollary cuts the other way and is the good news for this
> item: what dedupe cannot help is the **wire**, so a netinstall's download is
> the one figure closure/NAR size predicts honestly (`nix path-info --store
> https://cache.nixos.org --json` gives the real `downloadSize`).
- Current ISO **8.078 GiB** compressed; **18.03 GiB** of store uncompressed
  (`zstd -19`, 2.23:1 — compression is already near-max, not the lever).
- The offline pin (`system.extraDependencies`, 60 roots: a representative
  installed system + the template HM closure + all flake inputs) is **4.02 GiB
  uncompressed of that — only ~22%**. Dropping it entirely still leaves a
  **~13.3 GiB** desktop → roughly **6 GiB** compressed at the same ratio.
  **So "no pin" alone is NOT the lighter ISO** — this is the trap to avoid.
- The desktop's own top weights: libreoffice 1457 MiB, initrd 1369,
  linux-firmware 770, chromium 1391 (two builds — #121, left in), llvm-lib 540,
  bibata-cursors 322, mesa 264, mbrola-voices 259, nerd-fonts ~420 combined.
  Note what that list implies: no single lever gets a desktop ISO under ~4 GB —
  which is the case for (b) below.

**So the real decision is what a netinstall ISO IS**, and it should be settled
first (`[human]`): (a) the full try-before-install desktop minus the pin
(~6.3 GiB — barely lighter, probably not worth a second target); (b) a **TUI
installer only, no desktop** (~1 GiB, the actual "netinstall" in the Debian
sense) which drops "try before install" from that medium — the fat ISO still
offers it; (c) a middle desktop (no libreoffice/chromium — but note #103 just
put those there deliberately, and a *demo* desktop that can't browse is the
bug #103 fixed).

**The gotcha that decides feasibility:** without the pin, a netinstall target
fetches from `cache.nixos.org` for stock nixpkgs paths — but **Nomarchy's own
derivations are in no binary cache**, so they would build *from source on the
user's machine* during install. That is the same failure `tools/vm/gap-analysis.py`
exists to diagnose (and #113 is a live instance of). So this item probably
depends on a public binary cache (cachix) for the flake's own outputs, or it
trades an 8 GiB download for a 40-minute install. Establish that before
building the target.

Pass = a second, documented ISO target that is *substantially* smaller (state
the measured number, both ISOs built from one tree), installs successfully with
a network in a QEMU run, says clearly at boot that it needs one, and leaves the
offline ISO's behaviour untouched (`checks.*` for the offline path stay green).

## LATER

- **DPMS when docked (revisit #127):** today we **skip** display blanking
  while any laptop internal is present-but-disabled (dock/clamshell), so
  idle only locks. That avoids blanking the sole external and the brick
  where wake/TTY failed (2026-07-15). Revisit when we want monitor
  power-save on dock again: re-enable DPMS-off for that mode only if
  `nomarchy-display-wake` (or better) reliably restores the external and
  undock-while-black restores eDP on the AMD dock setup — without
  depending on SSH. Until then, leave the skip in `modules/home/idle.nix`.
- **Wallpapers artifact split** (ROADMAP § Faster switches — decided,
  deferred): pinned `Nomarchy-wallpapers` input so a state write stops
  re-copying 86 MB. Follow-on: pre-built theme variants if switches are
  still slow after.
- **Installer round 2** (ROADMAP § Installer): multi-disk BTRFS RAID,
  impermanence, BIOS/legacy boot.
- **Boot-from-snapshot**: a systemd-boot equivalent of grub-btrfs.
- **MIPI/IPU software-ISP camera** support (no-UVC machines).
- **NixOS release bump → v2** `[human]`: deliberate, hand-edited, never
  automated; the previous attempt was discarded (2026-06-22) over a
  Hyprland OOM blocker — see MEMORY.md before retrying.

## FUTURE (decided deferred — not the agent queue head)

Work we **intend** someday but explicitly **not** NEXT. Agents do not
pick these unless Bernardo promotes one into NEXT/NOW.

### 20. KVM runner → VM suite in CI `[human]`
**Status (2026-07-10):** keep **eval-only** CI on the current Gitea
stack (act_runner in docker-compose on the 4c/4 GB IONOS VPS). Nested
KVM + RAM headroom on that host are a poor fit next to Gitea; full
`checks.*` VMs stay local / promotion-time until a **separate**
KVM-capable machine exists.

**When ready:** register a second runner (host-mode nix + `/dev/kvm`,
label `nix-kvm` — not the existing docker eval runner), then uncomment
the `vm-checks` job in `.gitea/workflows/check.yml` (`runs-on: nix-kvm`,
`nix flake check` + toplevel/HM builds). Do not enable the job until
that label is online (Gitea queues forever otherwise).

### Formatter — adopt later `[human]`
**Intent:** add a Nix formatter (likely `nixfmt-rfc-style`) in a dedicated
pass: reformat the tree once, document in CONVENTIONS, optional CI
check. **Not** the queue head — no drive-by reformats until that pass.

## PROPOSED (agent suggestions — await human triage)

*Agents: append here with a one-paragraph pitch (what/why/cost). Do not
implement. Bernardo moves accepted items into a tier.*

*Open work only. Shipped exam/A–C items (#47–#63, #14, #52 theme
high-ROI, etc.) live in the journal + ROADMAP — not here.*

### Product / day-2

### 134. `unstable.<pkg>` in the downstream — a newer app without a second flake

Bernardo 2026-07-15, wanting a newer LM Studio (pinned 0.4.15-2 vs unstable
0.4.19-2) and disliking the only path that works today. There is **no seam**:
`mkFlake` takes `src`/`username`/`hardwareProfile`/`system` only, and
`homeConfigurations` is built from mkFlake's own `pkgs` (`inherit pkgs`), so a
`nixpkgs.overlays` in system.nix cannot reach a home package. The user is left
hand-pinning a `fetchTarball` rev+sha256 in home.nix (verified working) — a
second pin with no lock, which is exactly the Nix expertise the distro exists
to spare them. Sketch: `home.packages = [ unstable.lmstudio ];`, and maybe
`stable.` as an explicit synonym for today's bare `pkgs`.

The shape that keeps the promises: Nomarchy carries the `nixpkgs-unstable`
input **itself** and exposes it through `overlays.default` as an `unstable`
attrset, so downstream stays **one input**, stays locked, and `nomarchy-pull`
still moves it. Cost/questions to settle before building:
- Lock churn: every lock bump now moves two channels, and the full-checklist
  rule (VERIFICATION §5) applies to both.
- Duplication: measured — pinned vs unstable lmstudio build closures share
  only 690 of ~3.7–4k paths (separate stdenv bootstrap). Opt-in per package,
  but a user who reaches for it pays a second glibc/stack in the closure.
- "Tested together upstream" weakens by construction: nothing validates
  pinned-Nomarchy + unstable-app. Needs an explicit, documented "you own this
  combination" line, not silence.
- Does `unstable` belong in the eval by default, or only when referenced?
  (An unused input still locks + fetches.)
- Scope: home packages only, or system.nix too?

Alternative if that is too much surface: give `mkFlake` an `overlays ? []`
param (one-line seam, user still owns their own input/pin — cheaper for us,
still expert-only for them). Related: #133.

### 114. Greeter ignores per-device keyboard layouts

Found by Bernardo 2026-07-14: logging out while docked lands on tuigreet,
where his external keyboard (remembered as `us` via
`settings.keyboard.devices`) types the session layout `gb` — so the password
prompt fights him. Not a regression and not docking-related: per-device
layouts are applied with `hyprctl keyword device[<name>]:kb_layout`, which
only exists inside a running Hyprland session, while tuigreet draws on a
kernel VT whose single keymap comes from `console.useXkbConfig` ←
`services.xserver.xkb.layout`. A VT structurally cannot do per-device
layouts, so this is a design gap, not a bug to patch. Options, cheapest
first: (a) document it and stop there; (b) a greeter layout-cycle key
(tuigreet has no such feature — would need the keymap swapped under it);
(c) host tuigreet inside a small Wayland compositor (cage/labwc), which
*does* get per-device XKB and would let the greeter honour the same
in-flake state the session uses — real work, and it changes the greeter's
whole rendering path (`modules/nixos/greeter.nix` themes tuigreet through
the 16 ANSI console slots, so (c) is not a drop-in). Worth deciding whether
the greeter is meant to be layout-aware at all before costing it.

- **NVIDIA first-class options** — **deferred past v1** (Bernardo
  2026-07-10). Keep #59 commented install guidance; no
  `nomarchy.hardware.nvidia.*` until a hybrid maintainer + queue.

- **Post-install hardware hints** (`VISION § B`) — After the general
  “you're set” card (#81), optionally fire **one** additional
  self-gated notify when the machine actually has the hardware:
  (a) `fwupdmgr` on PATH → “System › Firmware to check LVFS updates”;
  (b) `fprintd-list` on PATH → “System › Fingerprint to enroll”.
  One-shot markers in `settings.*` (same in-checkout discipline as
  `firstBootShown`); never a permanent MOTD nag. Cost: small — extend
  `nomarchy-first-boot` or a sibling oneshot + `checks.first-boot`
  fixture. Control-center / MOTD already mention these; the gap is the
  silent first *graphical* session for people who never open those.

_(#80–#83 + #85–#88 shipped 2026-07-11. Theme A day-2 + neon-glass finish
  shipped — VISION ✓. Dock/hibernate V3 → HARDWARE-QUEUE. Parallel
  fingerprint-or-password shipped 2026-07-12 (Bernardo promoted it live;
  `fingerprint.parallel`, pam-fprint-grosshack) — reader V3 →
  HARDWARE-QUEUE.)_

### v1.0 pointer

See **VISION**. Open PROPOSED: post-install hardware hints; NVIDIA
deferred past v1; IR portal (b)/(c) need T14s (HARDWARE-QUEUE § T14s).
Standing calls: browser = Chromium; power = PPD.

## Decisions `[human]`

Open calls only Bernardo can make; agents add options/evidence but never
decide. **Resolved** entries stay for history; agents treat them as closed.

### Resolved (2026-07-10)

- **Docs site vs Markdown-in-repo** — **markdown in-repo for now**
  (`docs/`, README). A rendered docs site is FUTURE if wanted.
- **Default browser** — **ship Chromium** in
  `templates/downstream/home.nix`; mime → `chromium-browser.desktop`.
  Opt out: delete the line / override mime.
- **Default power backend** — **keep PPD** (`nomarchy.system.power.backend`
  default). TLP remains the one-line opt-in. Rationale: stability + live
  profile API for menu/Waybar; Omarchy’s TLP experiment reverted.

### Resolved (2026-07-10, more)

- **Formatter adoption** — **yes, but not now.** Tracked as FUTURE
  (below). Nix-source style only (`nixfmt-rfc-style` or similar); one
  bulk reformat + CI/check when promoted. Until then: hand-aligned
  style per CONVENTIONS.

- **Hibernation** — **want by default** (product intent). Needs a
  disk-backed swap (file or partition) sized for resume; not zram alone.
  **Shipped as #76**; V3 power-cycle PASSED on TuringMachine 2026-07-12
  (ROADMAP § Hibernation + zram by default).

### Resolved (2026-07-10, #76 design)

- **Swap sizing** — **exactly RAM** (installer default, unchanged). Hibernate
  image ≤ RAM; zram takes day-to-day paging. **`swapSize=0`** stays no-swap.
- **Migration** — **docs runbook** (`docs/MIGRATION.md`), not a tool.
- **No-swap Hibernate** — keep the menu row; **notify on failure**.