Files
Panama/tests/quickshell/connectivity-contract.sh
Gabriel Brown 3c521cf5fa Handle Wi-Fi and Bluetooth in Settings
Network & Devices was 92 lines and two buttons that opened GNOME. It now
scans, joins, and pairs directly through Quickshell.Networking and
Quickshell.Bluetooth -- NetworkManager and BlueZ over DBus, no shelling
out to nmcli or bluetoothctl. That was the founding requirement for this
desktop: never having to drop to a terminal to join a network.

Scanning follows the page being visible. Wi-Fi scanning and especially
Bluetooth discovery hold the radio, and running either for a list nobody
is looking at spends airtime on nothing.

Joining a secured network gets a real password field, not the clipboard
popover's search box with different placeholder text: a Wi-Fi key typed
into a field that echoes it is readable by anyone behind you, and a
search glyph in front of a password prompt is simply wrong.

Two bugs found by looking at the rendered page, both silent:

The device lookups used enum names that do not exist --
NetworkDeviceType.Wifi rather than DeviceType.Wifi -- so both returned
null and the page reported "No Wi-Fi adapter" on a machine whose Wi-Fi
was connected. Nothing was logged; QML resolves an unknown enum member
to undefined and compares happily.

signalStrength is 0.0-1.0, not a percentage, so thresholds written for
0-100 put every network including the connected one in the bottom
bucket. The labels now use the same buckets as the icons in
quicksettings/WifiList.qml so the two cannot disagree.

The contract compares what the service resolves against what nmcli
reports, rather than only checking that nothing crashed.

Also makes the Home Assistant bridge hermetic: resolve_config read the
user's private env file even when a caller supplied an explicit
environment, so adding a real PANAMA_HOME_ASSISTANT_ENTITIES to that
file silently overrode a fixture asserting the legacy fallback. An
explicit environment is now the whole environment; production still
reads the file. Its live contract skips when no token is configured --
an absent credential is not a defect, and a suite expected to be red
stops being read -- while a configured-but-broken bridge still fails.

Claude-Session: https://claude.ai/code/session_01BRvzt4H8XXLPVH5MyYdk9L
2026-08-18 05:50:07 -04:00

100 lines
3.9 KiB
Bash
Executable File

#!/usr/bin/env bash
# Network & Devices reads real NetworkManager and BlueZ state.
#
# Both of the bugs this contract exists to prevent were silent. Neither logged
# anything; both produced a page that looked fine and told the user something
# false:
#
# * the device lookups used enum names that do not exist
# (NetworkDeviceType.Wifi rather than DeviceType.Wifi), so they returned
# null and the page reported "No Wi-Fi adapter" on a machine whose Wi-Fi was
# connected;
# * signalStrength is 0.0-1.0, not a percentage, so thresholds written for
# 0-100 put every network including the connected one in the bottom bucket.
#
# So this compares what the service resolves against what NetworkManager itself
# reports, rather than merely checking the service does not crash.
set -euo pipefail
repo_dir="$(cd "$(dirname "${BASH_SOURCE[0]}")/../.." && pwd)"
harness="$repo_dir/config/dot/quickshell/connectivity-harness.qml"
fail() {
printf 'connectivity contract: %s\n' "$1" >&2
exit 1
}
command -v nmcli >/dev/null || fail 'nmcli is needed to check the service against reality'
run() { qs -p "$harness" "$@"; }
harness_pid=""
cleanup() {
run ipc call connectivity-test setActive false >/dev/null 2>&1 || true
# By PID: never `pkill -f connectivity-harness`, which also matches the
# shell running this script.
[[ -n "$harness_pid" ]] && kill "$harness_pid" >/dev/null 2>&1 || true
}
trap cleanup EXIT
qs -p "$harness" --daemonize >/dev/null
for _ in $(seq 1 40); do
run ipc show 2>/dev/null | rg -q '^target connectivity-test$' && break
sleep 0.1
done
run ipc show 2>/dev/null | rg -q '^target connectivity-test$' || fail 'test IPC target did not start'
harness_pid="$(run list | awk '/Process ID:/ { print $3; exit }')"
# Scanning only runs while the page says it is visible.
run ipc call connectivity-test setActive true >/dev/null
sleep 3
state="$(run ipc call connectivity-test status)"
# ── Devices the service finds must match the ones NetworkManager reports ─────
nm_wifi="$(nmcli -t -f DEVICE,TYPE device | awk -F: '$2 == "wifi" { print $1; exit }')"
nm_wired="$(nmcli -t -f DEVICE,TYPE,STATE device | awk -F: '$2 == "ethernet" && $3 == "connected" { print $1; exit }')"
if [[ -n "$nm_wifi" ]]; then
[[ "$(jq -r .wifiDevice <<<"$state")" == "$nm_wifi" ]] \
|| fail "NetworkManager reports Wi-Fi device '$nm_wifi' but the service found '$(jq -r .wifiDevice <<<"$state")'"
fi
if [[ -n "$nm_wired" ]]; then
[[ "$(jq -r .wiredConnected <<<"$state")" == "true" ]] \
|| fail "NetworkManager reports '$nm_wired' connected but the service says it is not"
fi
# ── Signal strength is a ratio, and the labels must reflect that ─────────────
while IFS='|' read -r value expect; do
got="$(run ipc call connectivity-test labelFor "$value")"
[[ "$got" == "$expect" ]] || fail "signal $value labelled '$got', expected '$expect'"
done <<'CASES'
1.0|Excellent
0.85|Excellent
0.6|Good
0.4|Fair
0.1|Weak
0.0|No signal
CASES
# If a network is connected, it must not be described as the weakest possible
# thing -- that was the visible symptom of reading the ratio as a percentage.
active_ssid="$(jq -r .activeSsid <<<"$state")"
if [[ -n "$active_ssid" ]]; then
strength="$(jq -r .activeStrength <<<"$state")"
awk -v s="$strength" 'BEGIN { exit !(s >= 0 && s <= 1) }' \
|| fail "signalStrength $strength is outside 0.0-1.0; the label buckets assume a ratio"
fi
# ── Bluetooth ────────────────────────────────────────────────────────────────
if [[ "$(bluetoothctl list 2>/dev/null | wc -l)" -gt 0 ]]; then
[[ "$(jq -r .adapter <<<"$state")" == "true" ]] \
|| fail 'an adapter is present but the service did not find it'
fi
trap - EXIT
cleanup
printf 'connectivity contract: PASS\n'