Files
BT411/tools/btconsole.py
arcattackandClaude Fable 5 8674bca2d8 btconsole: survive socket errors + flush output (the mid-session sync-freeze lead)
The console emulator's recv loop only caught socket.timeout -- any reset/abort killed
the thread UNHANDLED, its buffered stdout died with it, and with both threads gone the
process exited silently.  The 2-node session then froze replication BOTH ways (each pod
holding a dead ConsoleHost socket; the engine never logged a console disconnect).  Now:
OSErrors are caught + logged and the thread idles alive; prints flush.  Run with
python -u and > console.log to capture the death reason on any recurrence.  The engine-
side question (why a dead console freezes peer replication + the disconnect handler's
gameListenerSocket-vs-consoleListenerSocket close) is logged in open-questions.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-14 09:30:47 -05:00

167 lines
7.0 KiB
Python

#!/usr/bin/env python
"""btconsole.py -- minimal Tesla/WinTesla CONSOLE emulator (the operator station).
The pod-side engine (MUNGA_L4/L4NET.CPP) boots `-net <port>` into ConsoleOnly state:
it listens on TCP <port> and waits for a console to connect and stream the mission
egg. The real console software is absent from every archive, so this tool speaks
the console's wire protocol:
packet := NetworkPacketHeader + NetworkManager::ReceiveEggFileMessage
off 0 int32 clientID = 0 (NetworkClient::NetworkManagerClientID)
off 4 int32 gameID = 0 (NETWORK.cpp:121 -- always 0)
off 8 int32 fromHost = 1 (FirstLegalHostID == the console)
off 12 int32 timeStamp = 0 (Time; unused by the egg handler)
off 16 uint32 messageLength = 1024 (sizeof(ReceiveEggFileMessage))
off 20 int32 messageID = 3 (NetworkManager::ReceiveEggFileMessageID)
off 24 uint32 messageFlags = 1 (Receiver::Message::ReliableFlag)
off 28 int32 sequenceNumber (0..n-1; -1 is the solo-local path only)
off 32 int32 notationFileLength (total egg bytes)
off 36 int32 thisMessageLength (bytes used in this chunk, <= 1000)
off 40 char notationData[1000]
total 1040 bytes per chunk
(constants verified from the port build via `btl4.exe` BT_NET_PROBE=1)
The pod reassembles chunks sequentially (L4NET.CPP:773-816: seq 0 allocates,
append until notationFileLength reached), sets NormalState, CreateMission ->
StartConnecting (the egg's [pilots] list forms the pod mesh), and sends an
AcknowledgeEggFile back on this socket. The console must STAY CONNECTED (a
disconnect trips the pod's console-loss path, which also closes the game
listener -- an engine bug).
Usage:
python btconsole.py <egg-file> <host:port> [<host:port> ...]
Example (two instances on one box, -net 1501 / -net 1601):
python btconsole.py MP.EGG 127.0.0.1:1501 127.0.0.1:1601
"""
import socket
import struct
import sys
import threading
import time
CHUNK = 1000
PKT_FMT_HDR = "<iiii" # clientID, gameID, fromHost, timeStamp
PKT_FMT_MSG = "<IiIiii" # messageLength, messageID, messageFlags, seq, fileLen, thisLen
MESSAGE_LENGTH = 1024 # sizeof(ReceiveEggFileMessage)
EGG_MESSAGE_ID = 3 # NetworkManager::ReceiveEggFileMessageID
RELIABLE_FLAG = 1
CONSOLE_HOST_ID = 1 # FirstLegalHostID
# The LAUNCH: pods load to WaitingForLaunch and sit until the console sends
# Application::RunMissionMessage (the operator's launch button). The handler
# ladder needs it twice: WaitingForLaunch->LaunchingMission (dispatches the
# player MissionStarting/translocation), then LaunchingMission->RunningMission.
# Sending it early (state==CreatingMission) hits the handler's Fail() -- hence
# the settle delay. Constants from BT_NET_PROBE=1.
APPLICATION_CLIENT_ID = 4 # NetworkClient::ApplicationClientID
RUN_MISSION_MESSAGE_ID = 5 # Application::RunMissionMessageID
LAUNCH_SETTLE_SECONDS = 20.0 # egg -> first launch (pods must reach WaitingForLaunch)
LAUNCH_STEP_SECONDS = 4.0 # first launch -> second (Launching -> Running)
def run_mission_packet():
pkt = struct.pack(PKT_FMT_HDR, APPLICATION_CLIENT_ID, 0, CONSOLE_HOST_ID, 0)
pkt += struct.pack("<IiI", 12, RUN_MISSION_MESSAGE_ID, RELIABLE_FLAG)
assert len(pkt) == 16 + 12
return pkt
def egg_wire_bytes(egg_text_bytes):
"""Convert raw egg TEXT to the wire form NotationFile::ReadText expects:
NUL-separated lines (NOTATION.cpp:1043 walks `strchr(buffer,'\\0')+1` per
line). The real 1996 console pre-processed the egg the same way; sending
the raw file text makes the whole egg parse as one garbage line ("ERROR:
no map in egg!")."""
lines = egg_text_bytes.replace(b"\r\n", b"\n").split(b"\n")
return b"".join(line + b"\0" for line in lines)
def egg_packets(egg_bytes):
total = len(egg_bytes)
seq = 0
off = 0
while off < total:
chunk = egg_bytes[off:off + CHUNK]
pkt = struct.pack(PKT_FMT_HDR, 0, 0, CONSOLE_HOST_ID, 0)
pkt += struct.pack(PKT_FMT_MSG, MESSAGE_LENGTH, EGG_MESSAGE_ID, RELIABLE_FLAG,
seq, total, len(chunk))
pkt += chunk.ljust(CHUNK, b"\0")
assert len(pkt) == 16 + MESSAGE_LENGTH
yield pkt
off += len(chunk)
seq += 1
def serve_pod(hostport, egg_bytes):
host, port = hostport.rsplit(":", 1)
port = int(port)
name = f"{host}:{port}"
# The pod may not be listening yet -- retry like the pods do to each other.
while True:
try:
s = socket.create_connection((host, port), timeout=5)
break
except OSError as e:
print(f"[{name}] waiting for pod ({e})")
time.sleep(1)
print(f"[{name}] connected; streaming egg ({len(egg_bytes)} bytes)")
n = 0
for pkt in egg_packets(egg_bytes):
s.sendall(pkt)
n += 1
print(f"[{name}] egg sent in {n} chunk(s); launch in {LAUNCH_SETTLE_SECONDS}s")
s.settimeout(2.0)
launch_at = time.time() + LAUNCH_SETTLE_SECONDS
launches_sent = 0
while True:
if launches_sent < 2 and time.time() >= launch_at:
s.sendall(run_mission_packet())
launches_sent += 1
launch_at = time.time() + LAUNCH_STEP_SECONDS
print(f"[{name}] RunMission #{launches_sent} sent")
try:
data = s.recv(4096)
if not data:
print(f"[{name}] pod closed the console socket", flush=True)
return
# Expect the AcknowledgeEggFile (and possibly other console traffic).
if len(data) >= 24:
(clid, gid, fh, ts) = struct.unpack_from("<iiii", data, 0)
(mlen, mid) = struct.unpack_from("<Ii", data, 16)
print(f"[{name}] pod->console packet: clientID={clid} fromHost={fh} "
f"msgID={mid} len={mlen} ({len(data)} bytes)")
else:
print(f"[{name}] pod->console {len(data)} bytes")
except socket.timeout:
continue
except OSError as e:
# Previously UNHANDLED: a reset/abort here killed the thread silently
# (buffered stdout lost) and the console process exited -- leaving the
# pods holding dead console sockets. Log it and keep the thread alive
# (idle) so the process + diagnosis survive.
print(f"[{name}] console socket error: {e!r} -- idling", flush=True)
while True:
time.sleep(60)
def main():
if len(sys.argv) < 3:
print(__doc__)
return 2
egg_bytes = egg_wire_bytes(open(sys.argv[1], "rb").read())
pods = sys.argv[2:]
threads = [threading.Thread(target=serve_pod, args=(p, egg_bytes), daemon=True)
for p in pods]
for t in threads:
t.start()
try:
while any(t.is_alive() for t in threads):
time.sleep(0.5)
except KeyboardInterrupt:
print("console shutting down")
return 0
if __name__ == "__main__":
sys.exit(main())