Nomarchy/pkgs/nomarchy-doctor/nomarchy-doctor.sh

# nomarchy-doctor — one-shot, READ-ONLY health check: the things that
# actually break user machines, as a pass/fail sheet where every
# failure prints the one command that fixes it. It never changes
# anything itself. Exit 0 = no failures (warnings allowed), 1 = at
# least one ✖.

flake="${NOMARCHY_PATH:-$HOME/.nomarchy}"
fails=0
warns=0

grn=$'\033[32m'; red=$'\033[31m'; ylw=$'\033[33m'; dim=$'\033[2m'; rst=$'\033[0m'
ok()   { printf '  %b✔%b %s\n' "$grn" "$rst" "$1"; }
bad()  { printf '  %b✖%b %s\n' "$red" "$rst" "$1"
         printf '      %bfix: %s%b\n' "$dim" "$2" "$rst"
         fails=$((fails + 1)); }
warn() { printf '  %b●%b %s\n' "$ylw" "$rst" "$1"
         if [ $# -gt 1 ]; then printf '      %b%s%b\n' "$dim" "$2" "$rst"; fi
         warns=$((warns + 1)); }
skip() { printf '  %b– %s%b\n' "$dim" "$1" "$rst"; }

printf 'nomarchy-doctor — %s\n\n' "$(date '+%Y-%m-%d %H:%M')"

# ── systemd units ────────────────────────────────────────────────────
mapfile -t sysfailed < <(systemctl --failed --no-legend --plain 2>/dev/null | awk '{print $1}')
if [ "${#sysfailed[@]}" -eq 0 ]; then
  ok "no failed system units"
else
  bad "failed system unit(s): ${sysfailed[*]}" \
      "journalctl -b -u <unit>  (read why; sys-rebuild after fixing)"
fi

# A user bus only exists inside a session (not over bare SSH/root).
if [ -n "$(systemctl --user is-system-running 2>/dev/null || true)" ]; then
  mapfile -t usrfailed < <(systemctl --user --failed --no-legend --plain 2>/dev/null | awk '{print $1}')
  if [ "${#usrfailed[@]}" -eq 0 ]; then
    ok "no failed user units"
  else
    bad "failed user unit(s): ${usrfailed[*]}" \
        "journalctl --user -b -u <unit>"
  fi
else
  skip "user units (no user session bus here)"
fi

# ── disk space ───────────────────────────────────────────────────────
# Only real on-disk filesystems; skips the tmpfs/9p/overlay mounts of
# live systems and test VMs (where usage numbers mean nothing).
seen=""
for mp in / /boot /nix; do
  [ -d "$mp" ] || continue
  line=$(df -PT "$mp" 2>/dev/null | awk 'NR==2 {gsub("%","",$6); print $1, $2, $6, $5}')
  [ -n "$line" ] || continue
  read -r dev fstype use availkb <<<"$line"
  case "$fstype" in
    ext2|ext3|ext4|btrfs|xfs|vfat|f2fs|zfs) ;;
    *) continue ;;
  esac
  case " $seen " in *" $dev "*) continue ;; esac
  seen="$seen $dev"
  avail=$(numfmt --to=iec $((availkb * 1024)))
  if [ "$use" -ge 90 ]; then
    bad "$mp is ${use}% full (${avail} free)" \
        "sudo nomarchy-gen-prune  (14d + keep ≥3 past system/HM gens; then reclaims store)"
  else
    ok "$mp has space (${use}% used, ${avail} free)"
  fi
done

# ── the flake + state file ───────────────────────────────────────────
# #107: prefer state.json; theme-state.json still counts until migrated.
state_file=
if [ -f "$flake/state.json" ]; then
  state_file="$flake/state.json"
elif [ -f "$flake/theme-state.json" ]; then
  state_file="$flake/theme-state.json"
fi
if [ -n "$state_file" ]; then
  state_base=$(basename "$state_file")
  if jq empty "$state_file" 2>/dev/null; then
    ok "$state_base parses"
  else
    bad "$state_base is not valid JSON (rebuilds will fail)" \
        "nomarchy-state-sync validate  (names the spot; fix it, or re-apply a theme)"
  fi
  if git -C "$flake" ls-files --error-unmatch "$state_base" >/dev/null 2>&1; then
    ok "$state_base is git-tracked"
  else
    bad "$state_base is NOT git-tracked (flake evaluation can't see it)" \
        "git -C $flake add $state_base"
  fi
  if [ "$state_base" = "theme-state.json" ]; then
    warn "state file still named theme-state.json — run any menu write or" \
         "nomarchy-state-sync set … to migrate to state.json"
  fi
else
  skip "state.json (no flake checkout at $flake)"
fi

if [ -d "$flake/.git" ]; then
  dirty=$(git -C "$flake" status --porcelain 2>/dev/null | wc -l)
  if [ "$dirty" -gt 0 ]; then
    warn "$dirty uncommitted change(s) in $flake (normal — theme writes land there)" \
         "commit when happy: git -C $flake add -A && git -C $flake commit -m settings"
  else
    ok "flake checkout is clean"
  fi
  behind=$(git -C "$flake" rev-list --count 'HEAD..@{u}' 2>/dev/null || echo 0)
  if [ "$behind" -gt 0 ]; then
    warn "flake is $behind commit(s) behind its upstream" "git -C $flake pull, then sys-rebuild + home-update"
  fi
fi

# ── machine-state drift (survives migration, unseen by the flake) ────
# Worked example: systemd linger. The flake never sets it anywhere, so
# a "yes" here is leftover state from a previous OS/install, not
# something Nomarchy configured — it can cause months of divergent
# session behavior that no amount of flake auditing will explain.
linger=$(loginctl show-user "${USER:-$(id -un)}" --property=Linger --value 2>/dev/null || true)
case "$linger" in
  '') skip "user linger (user unknown to logind — no session)" ;;
  yes) warn "user linger is enabled but Nomarchy never sets it (machine state the flake can't see)" \
            "loginctl disable-linger ${USER:-$(id -un)}" ;;
  *) ok "no user linger (matches the flake)" ;;
esac

# ── generation age ───────────────────────────────────────────────────
# The profile SYMLINK's own mtime is the generation's creation time
# (store paths themselves are all epoch-1).
link=$(readlink /nix/var/nix/profiles/system 2>/dev/null || true)
if [ -n "$link" ] && [ -e "/nix/var/nix/profiles/$link" ]; then
  ts=$(stat -c %Y "/nix/var/nix/profiles/$link" 2>/dev/null || echo 0)
  days=$(( ($(date +%s) - ts) / 86400 ))
  if [ "$days" -gt 30 ]; then
    warn "last system rebuild was $days days ago" \
         "sys-update when convenient (security fixes arrive with input bumps)"
  else
    ok "system generation is $days day(s) old"
  fi
else
  skip "system generation age (no system profile)"
fi

# ── generation prune (#128: 14d + keep ≥3 past, system + HM) ─────────
if [ -n "$(systemctl list-unit-files nomarchy-gen-prune.timer --no-legend --plain 2>/dev/null)" ]; then
  if systemctl is-enabled --quiet nomarchy-gen-prune.timer 2>/dev/null; then
    ok "generation prune timer is enabled (14d, keep ≥3 past)"
  else
    bad "nomarchy-gen-prune.timer is installed but not enabled" \
        "systemctl enable --now nomarchy-gen-prune.timer"
  fi
elif command -v nomarchy-gen-prune >/dev/null 2>&1; then
  warn "nomarchy-gen-prune is on PATH but no timer unit" \
       "rebuild the system so gen-prune.nix ships the weekly timer"
else
  skip "generation prune (package not on this machine)"
fi

# ── snapper timeline (when enabled) ──────────────────────────────────
if [ -n "$(systemctl list-unit-files snapper-timeline.timer --no-legend --plain 2>/dev/null)" ]; then
  if systemctl is-active --quiet snapper-timeline.timer; then
    ok "snapper timeline snapshots are running"
  else
    bad "snapper is enabled but the timeline timer is not active" \
        "systemctl status snapper-timeline.timer  (then journalctl -u snapper-timeline.service)"
  fi
else
  skip "snapper (not enabled on this machine)"
fi

# ── first-boot HM pre-activate (installer fail flag, BACKLOG #83) ────
# Installer writes /var/log/nomarchy-hm-preactivate.log on the target.
# If that log exists and there is still no Home Manager generation, the
# desktop never baked — print the recovery one-liner. Override log path
# with NOMARCHY_HM_PREACTIVATE_LOG for checks.doctor.
pre_log="${NOMARCHY_HM_PREACTIVATE_LOG:-/var/log/nomarchy-hm-preactivate.log}"
hm_gen_present=0
uid_name="${USER:-$(id -un 2>/dev/null || echo)}"
for d in \
  "${XDG_STATE_HOME:-${HOME:-}/.local/state}/nix/profiles" \
  "/nix/var/nix/profiles/per-user/${uid_name}"; do
  [ -n "$d" ] || continue
  if [ -e "$d/home-manager" ] \
     || ls -d "$d"/home-manager-[0-9]*-link >/dev/null 2>&1; then
    hm_gen_present=1
    break
  fi
done
if [ ! -r "$pre_log" ]; then
  skip "first-boot pre-activate (no installer log — not a failed bake)"
elif [ "$hm_gen_present" -eq 1 ]; then
  ok "first-boot pre-activate: desktop generation is present"
else
  bad "desktop was not pre-activated at install (no Home Manager generation)" \
      "home-manager switch --flake ~/.nomarchy -b bak  (details: $pre_log)"
fi

# ══ hardware ═════════════════════════════════════════════════════════
# Every check below self-gates: it skips cleanly when the tool, service,
# or device isn't present, so the section shrinks to fit the machine and
# never fails just because a feature is absent. Still read-only.

# ── network (NetworkManager) ─────────────────────────────────────────
if command -v nmcli >/dev/null 2>&1; then
  nmstate=$(nmcli -t -f STATE general status 2>/dev/null || true)
  case "$nmstate" in
    connected*) ok "NetworkManager: $nmstate" ;;
    "")         skip "NetworkManager (no state reported)" ;;
    *)          warn "NetworkManager state: $nmstate" \
                     "connect in System › Network (or: nmcli device)" ;;
  esac
else
  skip "NetworkManager (nmcli not present)"
fi

# ── default audio sink (PipeWire) ────────────────────────────────────
# wpctl/pactl talk to the user's PipeWire session — only meaningful with
# a user bus (not over bare SSH/root).
if [ -n "$(systemctl --user is-system-running 2>/dev/null || true)" ]; then
  if command -v wpctl >/dev/null 2>&1; then
    if wpctl inspect @DEFAULT_AUDIO_SINK@ >/dev/null 2>&1; then
      ok "default audio sink present"
    else
      warn "no default PipeWire sink" "pick one in System › Audio (or: wpctl status)"
    fi
  elif command -v pactl >/dev/null 2>&1; then
    defsink=$(pactl get-default-sink 2>/dev/null || true)
    if [ -n "$defsink" ] && [ "$defsink" != "@DEFAULT_SINK@" ]; then
      ok "default audio sink: $defsink"
    else
      warn "no default audio sink" "pick one in System › Audio"
    fi
  else
    skip "audio sink (no wpctl/pactl)"
  fi
else
  skip "audio sink (no user session bus here)"
fi

# ── GPU acceleration smoke (only if the probes are installed) ────────
if command -v vainfo >/dev/null 2>&1; then
  if vainfo >/dev/null 2>&1; then
    ok "VA-API acceleration works (vainfo)"
  else
    warn "vainfo present but VA-API probe failed" \
         "check mkFlake.hardwareProfile / GPU drivers (or run: vainfo)"
  fi
else
  skip "VA-API smoke (vainfo not installed)"
fi
if command -v glxinfo >/dev/null 2>&1; then
  if [ -n "${WAYLAND_DISPLAY:-}${DISPLAY:-}" ]; then
    if glxinfo -B >/dev/null 2>&1; then
      ok "OpenGL renderer responds (glxinfo)"
    else
      warn "glxinfo present but GL probe failed" \
           "check mkFlake.hardwareProfile / GPU drivers (or run: glxinfo -B)"
    fi
  else
    skip "OpenGL smoke (no display attached)"
  fi
else
  skip "OpenGL smoke (glxinfo not installed)"
fi

# ── fingerprint (only when fprintd is enabled) ───────────────────────
if [ -n "$(systemctl list-unit-files fprintd.service --no-legend --plain 2>/dev/null || true)" ]; then
  enrolled=""
  if command -v fprintd-list >/dev/null 2>&1 && [ -n "${USER:-}" ]; then
    enrolled=$(fprintd-list "$USER" 2>/dev/null || true)
  fi
  if [ -n "$enrolled" ] && ! printf '%s' "$enrolled" | grep -qi 'no fingers'; then
    ok "fprintd enabled, fingerprint(s) enrolled for ${USER:-user}"
  elif [ -n "$enrolled" ]; then
    warn "fprintd enabled but no fingerprints enrolled" "enroll one: fprintd-enroll"
  else
    ok "fprintd unit installed (fingerprint enabled)"
  fi
else
  skip "fingerprint (fprintd not enabled)"
fi

# ── firmware updates (fwupd) ─────────────────────────────────────────
# fwupd is D-Bus-activated, so an idle daemon is normal, not a fault.
# get-updates reads local metadata (no flashing); a hit is a soft warn.
if command -v fwupdmgr >/dev/null 2>&1 \
   && [ -n "$(systemctl list-unit-files fwupd.service --no-legend --plain 2>/dev/null || true)" ]; then
  if systemctl is-active --quiet fwupd.service; then
    ok "fwupd daemon active"
  else
    skip "fwupd daemon (idle — D-Bus-activated on demand)"
  fi
  if timeout 20 fwupdmgr get-updates >/dev/null 2>&1; then
    warn "firmware updates are available" "review, then apply: fwupdmgr update"
  else
    ok "firmware up to date (no pending updates)"
  fi
else
  skip "fwupd (not enabled / fwupdmgr absent)"
fi

# ── laptop battery charge threshold (only when a limit is set) ───────
# Name-agnostic (BAT0, CMB0, …) — same type/scope filter as notify (#60).
# Dell: Adaptive mode ignores end_threshold while still reporting the
# written value — warn if type is not Custom when a limit is active.
bat_seen=0; bat_limited=0
for bat in /sys/class/power_supply/*/; do
  [ "$(cat "$bat/type" 2>/dev/null)" = Battery ] || continue
  [ "$(cat "$bat/scope" 2>/dev/null || echo System)" = Device ] && continue
  f="$bat/charge_control_end_threshold"
  [ -r "$f" ] || continue
  bat_seen=1
  lim=$(cat "$f" 2>/dev/null || true)
  case "$lim" in
    ''|*[!0-9]*) continue ;;
  esac
  if [ "$lim" -lt 100 ]; then
    bat_limited=1
    ctype=$(cat "$bat/charge_types" 2>/dev/null || cat "$bat/charge_type" 2>/dev/null || true)
    active=$(printf '%s' "$ctype" | sed -n 's/.*\[\([^]]*\)\].*/\1/p')
    name=$(basename "$bat")
    if [ -n "$active" ] && [ "$active" != Custom ]; then
      warn "$name charge limit ${lim}% but type is $active (not Custom)" \
        "thresholds only apply in Custom — run: systemctl restart nomarchy-battery-charge-limit"
    else
      ok "$name charge limit active at ${lim}%${active:+ (type $active)}"
    fi
  fi
done
if [ "$bat_seen" -eq 1 ] && [ "$bat_limited" -eq 0 ]; then
  skip "battery charge limit (none set — charges to 100%)"
fi

# ── battery health (report-only, BACKLOG #80) ────────────────────────
# cycle_count + retained capacity (charge_* µAh or energy_* µWh). Same
# system-battery filter as charge-limit/notify. Self-gates when no
# battery or the firmware omits the attrs (desktops, bare VMs). Override
# the sysfs root with NOMARCHY_POWER_SUPPLY_ROOT for the checks.doctor
# fixture (test_power has no cycle/design attrs).
ps_root="${NOMARCHY_POWER_SUPPLY_ROOT:-/sys/class/power_supply}"
bat_health_seen=0
for bat in "$ps_root"/*/; do
  [ -d "$bat" ] || continue
  [ "$(cat "$bat/type" 2>/dev/null)" = Battery ] || continue
  [ "$(cat "$bat/scope" 2>/dev/null || echo System)" = Device ] && continue
  name=$(basename "$bat")

  cycles=$(cat "$bat/cycle_count" 2>/dev/null || true)
  case "$cycles" in ''|*[!0-9]*) cycles= ;; esac
  # Some firmwares export 0 forever — treat as unknown, not "brand new".
  if [ -n "$cycles" ] && [ "$cycles" -eq 0 ]; then cycles=; fi

  full=""; design=""
  if [ -r "$bat/charge_full" ] && [ -r "$bat/charge_full_design" ]; then
    full=$(cat "$bat/charge_full" 2>/dev/null || true)
    design=$(cat "$bat/charge_full_design" 2>/dev/null || true)
  elif [ -r "$bat/energy_full" ] && [ -r "$bat/energy_full_design" ]; then
    full=$(cat "$bat/energy_full" 2>/dev/null || true)
    design=$(cat "$bat/energy_full_design" 2>/dev/null || true)
  fi
  case "$full" in ''|*[!0-9]*) full= ;; esac
  case "$design" in ''|*[!0-9]*) design= ;; esac
  pct=""
  if [ -n "$full" ] && [ -n "$design" ] && [ "$design" -gt 0 ]; then
    pct=$(( full * 100 / design ))
  fi

  [ -n "$cycles" ] || [ -n "$pct" ] || continue
  bat_health_seen=1

  detail=""
  [ -n "$cycles" ] && detail="${cycles} cycles"
  if [ -n "$pct" ]; then
    [ -n "$detail" ] && detail="$detail, "
    detail="${detail}${pct}% of design capacity"
  fi
  # Soft warn only — wear isn't a doctor "fix", just a heads-up.
  if [ -n "$pct" ] && [ "$pct" -lt 70 ]; then
    warn "$name health: $detail" \
         "battery wear is normal over years — replace when runtime suffers"
  else
    ok "$name health: $detail"
  fi
done
if [ "$bat_health_seen" -eq 0 ]; then
  skip "battery health (no system battery / no cycle or design capacity attrs)"
fi

# ── hibernate / sleep (BACKLOG #76) ──────────────────────────────────
# Read-only: is there a working hibernate path? zram is RAM-only and can't
# hold a resume image, so hibernation needs a disk swap (partition or file)
# PLUS resume= wiring; a swapfile additionally needs resume_offset. A
# swap=0 machine opts out cleanly and this section skips. Never fails the
# sheet — an absent or under-sized swap is advisory, not a running fault.
# (|| echo 0 keeps set -euo pipefail happy if a proc file is unreadable.)
disk_swap_kb=$(awk 'NR>1 && $1 !~ /zram/ {s+=$3} END{print s+0}' /proc/swaps 2>/dev/null || echo 0)
zram_kb=$(awk 'NR>1 && $1 ~ /zram/ {s+=$3} END{print s+0}' /proc/swaps 2>/dev/null || echo 0)
ram_kb=$(awk '/^MemTotal:/ {print $2}' /proc/meminfo 2>/dev/null || echo 0)

if [ "${disk_swap_kb:-0}" -gt 0 ]; then
  if grep -q 'resume=' /proc/cmdline 2>/dev/null; then
    ok "hibernate: resume device set on the kernel cmdline"
    # A swapfile also needs resume_offset to locate the image within it.
    if awk 'NR>1 && $1 !~ /zram/ && $2=="file"{f=1} END{exit !f}' /proc/swaps 2>/dev/null \
       && ! grep -q 'resume_offset=' /proc/cmdline 2>/dev/null; then
      warn "swapfile in use but no resume_offset= on the cmdline — resume will fail" \
           "add boot.kernelParams resume_offset (docs/MIGRATION.md → Enabling hibernation)"
    fi
  else
    warn "disk swap present but no resume= on the kernel cmdline — hibernate won't resume" \
         "set boot.resumeDevice (docs/MIGRATION.md → Enabling hibernation)"
  fi
  if [ "${ram_kb:-0}" -gt 0 ] && [ "$disk_swap_kb" -lt "$ram_kb" ]; then
    warn "disk swap ($(numfmt --to=iec $((disk_swap_kb * 1024)))) is smaller than RAM ($(numfmt --to=iec $((ram_kb * 1024)))) — a full hibernate image may not fit" \
         "size swap ≥ RAM (docs/MIGRATION.md → Enabling hibernation)"
  else
    ok "hibernate: disk swap ≥ RAM ($(numfmt --to=iec $((disk_swap_kb * 1024))))"
  fi
else
  skip "hibernate (no disk swap — swap=0 opt-out; zram alone can't resume)"
fi

# zram compressed-RAM swap — the memory-pressure layer oom.nix ships on
# by default (#76). Its absence on a Nomarchy box means something is off.
if [ "${zram_kb:-0}" -gt 0 ]; then
  ok "zram compressed-RAM swap active ($(numfmt --to=iec $((zram_kb * 1024))))"
else
  warn "zram swap not active (the default memory-pressure layer)" \
       "expected on by default — check modules/nixos/oom.nix, then sys-rebuild"
fi

# Best-effort: a suspend/hibernate failure recorded in the previous boot's
# kernel log. Needs journal read access; silently no-ops without it. The
# `|| true` keeps a no-match grep / missing -1 boot from tripping set -e.
if command -v journalctl >/dev/null 2>&1; then
  pmerr=$(journalctl -b -1 -k --no-pager 2>/dev/null \
          | grep -iE 'PM: .*(hibernat|suspend).*(fail|error)|Failed to (hibernate|suspend)' \
          | tail -n1 || true)
  if [ -n "$pmerr" ]; then
    warn "a suspend/hibernate error is in the previous boot's log" \
         "review: journalctl -b -1 -k -g 'hibernat|suspend'"
  fi
fi

# ── verdict ──────────────────────────────────────────────────────────
echo
if [ "$fails" -eq 0 ]; then
  printf '%b✔ healthy%b — %d warning(s)\n' "$grn" "$rst" "$warns"
  exit 0
else
  printf '%b✖ %d problem(s)%b, %d warning(s) — each ✖ above shows its fix\n' "$red" "$fails" "$rst" "$warns"
  exit 1
fi