Files
transcience/tools/smoke.sh
T
claude e0c1e0d5c6
ci / verify (push) Successful in 49s
Stage 5: bosses that move, attacks that warn, and a second boss
Boss movement is a property of the PHASE, not of the boss -- a fight that
stands still and then starts hunting you is one boss with two phases. Four
modes (STATIC, ORBIT, CHASE, WAYPOINTS) handled generically in
SimWorld._move_boss, so a boss that moves is still data. BossDef.stationary
is gone rather than kept beside the phases: a flag claiming the boss stood
still while a phase walked around would be a second source of truth and the
wrong one, so moves() is derived.

CHASE holds a distance instead of closing, because a boss standing on top of
you is a boss whose bullets cannot be read. Waypoints are fractions of the
arena so one phase works in rooms of different sizes. Every mode is speed
clamped in one place -- ORBIT computes an absolute destination and would
otherwise snap onto its circle on the first tick -- and movement slides
against geometry so a boss cannot walk through the pillars its own arena was
designed around.

The room clamp moved to after movement, where it is finally load-bearing. It
was a no-op while every boss stood still, which is exactly when an invariant
is cheapest to establish: boss rooms deliberately do not lock, so walking out
is always an escape, and that only holds if the boss cannot follow.

TelegraphedStrikeEmitter marks spots and fills them a moment later. The moment
between is the feature: a burst at your feet is a coin flip, the same burst
with a second of notice is a question. It stays stateless like every other
emitter -- they are shared resources and two bosses of the same kind must not
stomp each other -- so strike positions are derived from the volley number and
a test asserts the burst lands where the marker promised. Markers are drawn
through fog and through walls, unlike everything else in the view, because a
warning you cannot see is an unavoidable hit with extra steps.

The Cantor of the Vault fights in the choir vault: static, then a four-corner
circuit, then a chase, then orbiting while marking. It exists to prove the
format stretched, and a test asserts it uses both new mechanisms.

Which boss a run has now comes from its SEED rather than its depth. Depth is a
dev flag nothing in play raises, so the arena was keyed to something no player
can change and the second boss was unreachable in an actual game.

Two things found while finishing:

  - tools/export_content.gd had a hand-maintained boss list and had already
    gone stale, silently not writing the Cantor. Content.ALL_ENEMIES and
    ALL_BOSSES now feed the export tool, the renderer and five tests that each
    kept their own copy.
  - diag_loot failed intermittently after another diagnostic. Taking over from
    the bot cleared its input queue but not its HELD input, so a starved server
    coasted on the bot's last movement vector for half a second and walked the
    player off the item it had been placed on. The press arrived correctly,
    which is why "the press reached the simulation" passed while everything it
    should have caused failed.

check.sh clean, 409 tests, SMOKE PASS (19 assertions), all four diagnostics
green.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-06 16:32:12 +02:00

158 lines
6.8 KiB
Bash
Executable File

#!/usr/bin/env bash
# End-to-end integration test: boot a dedicated server, connect two scripted
# bot clients over a real ENet socket, and assert that the whole loop happened --
# handshake, lobby, portal into a dungeon, and the emergency escape back.
#
# This is the test that catches everything unit tests structurally cannot: RPC
# wiring, codec round-trips over the wire, instance transfers, and the client's
# reconciliation loop. Run it before calling any networking change done.
#
# tools/smoke.sh # ~35s
# KEEP=1 tools/smoke.sh # keep the logs and print where they are
set -uo pipefail
cd "$(dirname "$0")/.."
GODOT="${GODOT:-godot}"
PORT="${PORT:-27099}"
SERVER_TICKS="${SERVER_TICKS:-2000}"
CLIENT_TICKS="${CLIENT_TICKS:-1700}"
OUT="$(mktemp -d -t transcience-smoke-XXXXXX)"
PIDS=()
cleanup() {
for pid in "${PIDS[@]:-}"; do
kill "$pid" 2>/dev/null
done
wait 2>/dev/null
if [[ "${KEEP:-0}" == "1" ]]; then
echo "logs kept in $OUT"
else
rm -rf "$OUT"
fi
}
trap cleanup EXIT
echo "smoke: server on port $PORT, logs in $OUT"
# A scratch character store: without it a rerun resumes the characters the
# previous run created, and "a character was created" stops being true.
"$GODOT" --headless --path . -- --server --port "$PORT" --autoquit "$SERVER_TICKS" \
--store "$OUT/characters.json" \
> "$OUT/server.log" 2>&1 &
PIDS+=($!)
for _ in $(seq 1 60); do
grep -q "SERVER_READY" "$OUT/server.log" 2>/dev/null && break
sleep 0.25
done
if ! grep -q "SERVER_READY" "$OUT/server.log" 2>/dev/null; then
echo "FAIL: server never became ready"; cat "$OUT/server.log"; exit 1
fi
for n in 1 2; do
"$GODOT" --headless --path . -- --bot --host 127.0.0.1 --port "$PORT" \
--name "bot$n" --account "$((7000 + n))" \
--autoquit "$CLIENT_TICKS" > "$OUT/bot$n.log" 2>&1 &
PIDS+=($!)
sleep 0.4
done
# A fourth bot that leaves *politely* -- the same Net.shutdown() the in-game
# menu's "Disconnect" button calls. The escape channel is keyed on the socket
# closing, not on how it closed, so a clean exit must be caught exactly like the
# SIGKILL below. If someone ever adds a "clean leave" message that bypasses the
# channel, this is what catches it.
"$GODOT" --headless --path . -- --bot --host 127.0.0.1 --port "$PORT" \
--name "leavebot" --account 7101 --leave-after 120 --autoquit "$CLIENT_TICKS" \
> "$OUT/leavebot.log" 2>&1 &
PIDS+=($!)
sleep 0.4
# A third bot that gets SIGKILLed the moment it is inside a dungeon. This is the
# anti-disconnect-cheese path: the server must keep its body in the world,
# channel it out over the same one second the escape button costs, and only then
# forget the peer -- never delete it instantly on socket close.
"$GODOT" --headless --path . -- --bot --host 127.0.0.1 --port "$PORT" \
--name "dropbot" --account 7102 --autoquit "$CLIENT_TICKS" > "$OUT/dropbot.log" 2>&1 &
DROP_PID=$!
for _ in $(seq 1 80); do
grep -q "entered instance .*DUNGEON" "$OUT/dropbot.log" 2>/dev/null && break
sleep 0.25
done
if grep -q "entered instance .*DUNGEON" "$OUT/dropbot.log" 2>/dev/null; then
# Reaped here so bash does not print its own "Killed" job-control line; the
# kill is the point of the test, not a failure.
{ kill -9 "$DROP_PID"; wait "$DROP_PID"; } 2>/dev/null || true
else
echo " WARN dropbot never reached a dungeon; drop assertions will fail"
PIDS+=("$DROP_PID")
fi
wait 2>/dev/null
fails=0
check() { # check <label> <file> <pattern>
if grep -qE "$3" "$2"; then
echo " ok $1"
else
echo " FAIL $1 (no match for /$3/ in $(basename "$2"))"
fails=$((fails + 1))
fi
}
refute() { # refute <label> <file> <pattern>
local hits
hits=$(grep -cE "$3" "$2" || true)
if [[ "$hits" == "0" ]]; then
echo " ok $1"
else
echo " FAIL $1 ($hits line(s) matched /$3/ in $(basename "$2"))"
grep -E "$3" "$2" | head -5 | sed 's/^/ /'
fails=$((fails + 1))
fi
}
echo "assertions:"
check "bot1 authenticated" "$OUT/server.log" "authenticated as account 7001"
check "bot2 authenticated" "$OUT/server.log" "authenticated as account 7002"
check "a character was created" "$OUT/server.log" "created 'bot1'"
check "and persisted to the store" "$OUT/characters.json" "bot1"
check "and are played" "$OUT/server.log" "playing 'bot1'"
# The bots press interact and the use key while in a dungeon, so the item half
# of the input frame crosses a real socket on every run. What they almost never
# do is produce a drop to act on -- that end of it is tools/diag_loot.tscn.
check "inventories reach the save file" "$OUT/characters.json" "\"inventory\""
check "a dungeon instance opened" "$OUT/server.log" "opened dungeon instance"
# Which arena a run gets is a coin flip on its seed, so this only asserts that
# SOME boss spawned and that it is one the content knows about -- which is what
# catches an arena wired to a boss id that does not resolve.
check "a boss spawned" "$OUT/server.log" "BOSS_SPAWNED (warden|cantor) in instance"
# Bots pick their entrance from their account id, so a run with several of them
# opens one of each kind. This is what catches a portal wired to the wrong
# dungeon, or an instance matcher that ignores which dungeon was asked for.
check "the real run opened" "$OUT/server.log" "opened dungeon instance [0-9]+ \(warden_descent\)"
check "the proving grounds opened" "$OUT/server.log" "opened dungeon instance [0-9]+ \(proving_grounds\)"
check "emergency escape completed" "$OUT/server.log" "escaped to lobby"
check "bot1 reached a dungeon" "$OUT/bot1.log" "entered instance .*DUNGEON"
check "bot1 returned to the lobby" "$OUT/bot1.log" "entered instance .*LOBBY"
check "a hard drop is channelled, not instant" \
"$OUT/server.log" "dropped in instance [0-9]+, channelling out"
check "the dropped body is released" "$OUT/server.log" "released from instance [0-9]+ after drop"
check "a polite disconnect leaves too" "$OUT/leavebot.log" "BOT_GRACEFUL_LEAVE"
check "a polite disconnect is also channelled" \
"$OUT/server.log" "'leavebot' dropped in instance [0-9]+, channelling out"
# No assertion for upgrades here on purpose. Bots never walk to the
# quartermaster, so anything written would either duplicate a check above or
# assert nothing at all. The upgrade-state message DOES cross the wire on every
# login, so a malformed one shows up in the script-error refutations below --
# and tools/diag_upgrades.tscn covers the loop itself.
refute "no server script errors" "$OUT/server.log" "SCRIPT ERROR|Parse Error|USER ERROR"
refute "no client script errors" "$OUT/bot1.log" "SCRIPT ERROR|Parse Error|USER ERROR"
echo
if [[ $fails -eq 0 ]]; then
echo "SMOKE PASS"
else
echo "SMOKE FAIL ($fails)"
echo "--- server.log (tail) ---"; tail -40 "$OUT/server.log"
echo "--- bot1.log (tail) ---"; tail -40 "$OUT/bot1.log"
fi
exit $fails