feat: give every room the broker, one scale, and its own panel back
Little Waders took the Plasma build and reported what a container cannot. A disc player on a machine with no optical drive, a clock nobody at the back of the room could read, a browser at half the size of the panel around it, and Parent Call with no broker to call. The broker is the same for every classroom, so it is a default now and not something to fill in room by room. classroom-migrate brings an existing room to the same place without touching its name, its username or its password, and Parent Call distinguishes a broker it cannot reach from one that was never set up: those send somebody to look at completely different things. The disc player's detection was right and ran too late — alongside the session rather than before it, and the panel is built from the layout the moment a session starts. It runs before the display manager now, and again from udev when a drive is plugged in. UI_SCALE had a ceiling of 1.5, which cannot express a 4K television; a clock pinned at 15 points, which is the same size at every scale; and a browser that was never told the number at all. All three are fixed and the room doubles together at 200%. This also brings 0.1.47's work into the repository, where it had only ever existed in a build directory under /tmp: the hidden display dialog, its authenticated writer, the polkit action, and the church's own wallpaper. One piece of it is deliberately left out. The custom volume display needed a shell package of our own, and with one selected plasmashell builds no panel at all — measured twice, with that build's own files, on a classroom that deletes its panel configuration at every login by design. The lab connected to the real church broker the moment the package defaulted to it. Nothing was left behind, which was checked, but the guard that should have stopped it had never worked: it used iptables, which the lab image does not have, and probed with /dev/tcp under dash, where that always fails. Both reported success. Co-Authored-By: Claude Opus 5 <[email protected]>
This commit is contained in:
co-authored by
Claude Opus 5
parent
926668dc9b
commit
34fd02e08f
@@ -13,7 +13,8 @@ SRC="${1:-/srv/classroom}"
|
||||
# XDG_CONFIG_DIRS. plasma-workspace owns files by these names in /etc/xdg
|
||||
# and two packages cannot own one path.
|
||||
for setting in kwinrc kwinrulesrc kscreenlockerrc ksmserverrc \
|
||||
krunnerrc plasmanotifyrc powermanagementprofilesrc; do
|
||||
krunnerrc plasmanotifyrc powermanagementprofilesrc plasmarc \
|
||||
plasmaparc; do
|
||||
install -Dm644 "$SRC/config/plasma/xdg/$setting" \
|
||||
"/etc/ekklesia-classroom/xdg/$setting"
|
||||
rm -f "/etc/xdg/$setting.ekklesia-old"
|
||||
@@ -48,6 +49,20 @@ install -Dm644 "$SRC/config/plasma/lookandfeel/contents/defaults" \
|
||||
install -Dm644 \
|
||||
"$SRC/config/plasma/lookandfeel/contents/layouts/org.kde.plasma.desktop-layout.js" \
|
||||
"$LNF/contents/layouts/org.kde.plasma.desktop-layout.js"
|
||||
install -Dm644 "$SRC/config/polkit/org.ekklesia.classroom.display.policy" \
|
||||
/usr/share/polkit-1/actions/org.ekklesia.classroom.display.policy
|
||||
install -Dm644 "$SRC/config/plasma/autostart/ekklesia-classroom-scale.desktop" \
|
||||
/etc/xdg/autostart/ekklesia-classroom-scale.desktop
|
||||
|
||||
install -Dm644 "$SRC/config/plasma/desktoptheme/metadata.json" \
|
||||
/usr/share/plasma/desktoptheme/ekklesia/metadata.json
|
||||
install -Dm644 "$SRC/config/plasma/desktoptheme/dialogs/background.svg" \
|
||||
/usr/share/plasma/desktoptheme/ekklesia/dialogs/background.svg
|
||||
|
||||
for script in classroom-ui-scale classroom-scale classroom-desktop-layout \
|
||||
classroom-display-settings classroom-set-scale; do
|
||||
install -Dm755 "$SRC/scripts/$script" "/usr/local/bin/$script"
|
||||
done
|
||||
|
||||
for entry in "$SRC"/config/plasma/applications/*.desktop; do
|
||||
install -Dm644 "$entry" "/usr/share/applications/$(basename "$entry")"
|
||||
|
||||
@@ -77,6 +77,38 @@ up)
|
||||
done
|
||||
# startplasma looks for `qdbus`; Debian ships it as qdbus6.
|
||||
docker exec "$NAME" ln -sf /usr/bin/qdbus6 /usr/local/bin/qdbus
|
||||
|
||||
# The church's own networks, refused from inside the container.
|
||||
#
|
||||
# This is not theoretical. The moment the package began defaulting
|
||||
# MQTT_HOST to the church broker, the lab connected to it: a test
|
||||
# machine appearing on the real broker under a container id. Nothing
|
||||
# was left behind — the broker held no retained status afterwards,
|
||||
# which was checked — but the lab had no business being there at all,
|
||||
# and the Parent Call coordinator is on the same networks.
|
||||
#
|
||||
# Container-local: this netns belongs to the container alone and
|
||||
# nothing on the host or the church network is changed by it. Debian's
|
||||
# own mirrors are untouched, so installs still work.
|
||||
# A blackhole route rather than a firewall rule: the lab image has no
|
||||
# iptables, and the rig's guard had been quietly doing nothing because
|
||||
# of it. Routes need no netfilter modules and cannot silently fail.
|
||||
docker exec "$NAME" sh -c '
|
||||
for net in 10.30.0.0/16 10.40.0.0/16; do
|
||||
ip route replace blackhole "$net" || exit 1
|
||||
done' || echo " WARNING: could not fence the church networks"
|
||||
|
||||
# And checked with bash, not sh. The probe uses /dev/tcp, which is a
|
||||
# bash feature: under dash it fails whatever the network is doing, so
|
||||
# the check passed on a lab that could reach the broker perfectly
|
||||
# well. Measured.
|
||||
for target in 10.40.0.2/1883 10.30.0.2/1025; do
|
||||
if docker exec "$NAME" timeout 4 bash -c "exec 3<>/dev/tcp/${target}" 2>/dev/null; then
|
||||
echo "REFUSING TO CONTINUE: ${target} is reachable from the lab" >&2
|
||||
exit 1
|
||||
fi
|
||||
done
|
||||
echo " church networks fenced out of the lab"
|
||||
echo "lab $NAME is up ($(docker exec "$NAME" systemctl is-system-running 2>/dev/null))"
|
||||
;;
|
||||
|
||||
|
||||
@@ -24,14 +24,19 @@ inlab() { docker exec "$NAME" "$@"; }
|
||||
guard() {
|
||||
# Container-local. This netns belongs to the container alone, so
|
||||
# nothing on the host or the church network is changed by it.
|
||||
# Blackhole routes, not iptables: the lab image has no iptables at
|
||||
# all, so these two lines had been failing silently since they were
|
||||
# written. `tools/lab/lab up` now does the same thing at start-up, so
|
||||
# this is a second statement of it rather than the only one.
|
||||
inlab sh -c '
|
||||
iptables -C OUTPUT -d 10.30.0.0/16 -j REJECT 2>/dev/null ||
|
||||
iptables -I OUTPUT -d 10.30.0.0/16 -j REJECT
|
||||
iptables -C OUTPUT -d 10.40.0.0/16 -j REJECT 2>/dev/null ||
|
||||
iptables -I OUTPUT -d 10.40.0.0/16 -j REJECT'
|
||||
for net in 10.30.0.0/16 10.40.0.0/16; do
|
||||
ip route replace blackhole "$net" || exit 1
|
||||
done'
|
||||
|
||||
# bash, not sh. /dev/tcp is a bash feature and this probe reported
|
||||
# every target unreachable under dash, whatever the network was doing.
|
||||
for target in 10.40.0.2/1883 10.30.0.2/1025; do
|
||||
if inlab timeout 4 sh -c "exec 3<>/dev/tcp/${target}" 2>/dev/null; then
|
||||
if inlab timeout 4 bash -c "exec 3<>/dev/tcp/${target}" 2>/dev/null; then
|
||||
echo "REFUSING TO CONTINUE: ${target} is still reachable" >&2
|
||||
exit 1
|
||||
fi
|
||||
|
||||
Reference in New Issue
Block a user