* Fix multiplayer standalone crash: client stats race + host accept code
Fixes a crash in multiplayer PXO/tracker games hosted on a standalone
server where clients crash at debrief close after scoring a kill:
pilotfile.cpp:295: Ship kills of '<ship class>' not found,
should have been added by pilotfile::update_stats
Two independent root causes were found, both of which must be fixed for
correct behavior:
--- Root cause 1: race condition in multi_sexp_end_mission() (sexp.cpp) ---
The previous code called send_debrief_event() before
multi_handle_end_mission_request() as a "hack" to skip the warp-out
sequence when an end-mission SEXP fires. On standalone servers this is
a race condition: scoring_eval_kill only runs on the master, and the
master hasn't broadcast m_okKills to clients yet at the point
multi_sexp_end_mission() is called. Entering debrief this early causes
scoring_level_close() to run with zeroed m_okKills, so
Pilot.update_stats() creates no kill entries for that session.
When debrief_close() later calls scoring_backout_accept() it cannot
find those entries and hits the UNREACHABLE.
Fix: Remove the send_debrief_event() call entirely. The normal flow
via multi_handle_end_mission_request() has the server broadcast stats
first and then send MISSION_END, which causes the client to enter
debrief with correct kill data already populated. The master enters
debrief through sexp_end_mission() -> send_debrief_event() as before.
The behavioral side-effect of removing this hack: clients that go
through the normal warp-out sequence will no longer skip it on
end-mission. Clients in a state where they cannot warp still receive
immediate debrief via multi_handle_sudden_mission_end().
--- Root cause 2: host accept code never set on PXO standalone (multiui.cpp) ---
In multi_debrief_accept_hit() and multi_debrief_esc_hit(), when the
game is a PXO/tracker game and the local player is the game host
(NETINFO_FLAG_GAME_HOST), the code calls send_store_stats_packet() to
tell clients to record their stats. It then sets
Multi_debrief_stats_accept_code = 1 via the per-client response path
inside the NETINFO_FLAG_AM_MASTER block.
On a standalone server the game host is NOT the master
(NETINFO_FLAG_AM_MASTER is false for the client-side host). The
NETINFO_FLAG_AM_MASTER block is skipped entirely, and
send_store_stats_packet() uses multi_io_send_to_all_reliable() which
skips the local player — so the host never receives its own packet.
Multi_debrief_stats_accept_code stays at -1 (unset).
When debrief_close() checks this value and sees it is not 1, it
assumes stats were tossed or the window was closed without accepting,
and runs scoring_backout_accept() unconditionally. With fix 1 in place
this no longer crashes (entries exist), but silently strips all session
kill credit from the host's pilot record.
Fix: After the send_store_stats_packet() block, call
multi_debrief_stats_accept() unconditionally for the game host so the
accept code is set locally regardless of whether the host is also the
standalone master.
Co-Authored-By: Claude Sonnet 4.6 <[email protected]>
* Preserve end-mission warp-out skip via deferred debrief entry
The previous fix removed send_debrief_event() from
multi_sexp_end_mission() entirely, which was correct for the stats
race condition but had a side effect: clients would no longer skip the
warp-out sequence when an end-mission SEXP fires. The warp-out skip
was an intentional feature — process_endgame_packet() checks whether
the client is already in debrief state and, if so, bypasses
multi_warpout_all_players().
Restore the original behavior by deferring the send_debrief_event()
call to process_endgame_packet() rather than eliminating it.
multi_sexp_end_mission() now sets Multi_sexp_end_mission_pending=true
before calling multi_handle_end_mission_request(). When the client
receives the MISSION_END packet, process_endgame_packet() checks the
flag, calls send_debrief_event() there, then falls through to the
debrief-state check which skips the warp-out exactly as before.
This is safe because send_endgame_packet() calls multi_broadcast_stats()
before sending MISSION_END. Both use reliable delivery, so stats are
guaranteed to be received and applied before MISSION_END is processed,
meaning scoring_level_close() now sees correct kill data when
send_debrief_event() fires.
Alt-J-initiated mission ends are unaffected: the flag is never set,
so the warp-out proceeds normally for those paths.
Co-Authored-By: Claude Sonnet 4.6 <[email protected]>
* Revert multiui.cpp accept code change — tracker master owns stats
The previous change added multi_debrief_stats_accept() calls for the
game host in multi_debrief_accept_hit() and multi_debrief_esc_hit() to
work around Multi_debrief_stats_accept_code staying -1 on a standalone
server. This was based on incorrect analysis.
In PXO tracker games the standalone master is intentionally the sole
authority for stat storage (via multi_fs_std_tracker_store_stats() in
multi_standalone_postgame_close()). The host and clients must not save
stats to their local pilot files independently. Multi_debrief_stats_
accept_code remaining -1 and triggering scoring_backout_accept() on the
host is correct behavior: kills are backed out of the local pilot file
because the tracker holds the authoritative record, not the local file.
The race condition fix (deferred send_debrief_event()) is sufficient to
prevent the crash: Pilot.update_stats() now creates kill entries with
correct m_okKills data, so scoring_backout_accept() can find and remove
them cleanly rather than hitting the UNREACHABLE.
Co-Authored-By: Claude Sonnet 4.6 <[email protected]>
* Move Multi_sexp_end_mission_pending from multiutil to multi and reset in multi_vars_init()
Addresses review feedback: the flag has no relation to multiutil, and
resetting it in multi_vars_init() ensures it is always cleared between
missions regardless of which code path was taken.
Co-Authored-By: Claude Opus 4.6 (1M context) <[email protected]>
---------
Co-authored-by: Claude Sonnet 4.6 <[email protected]>
* Enable clients to move turrets
This allows clients to user server data to move turrets according to the timing and target data on the server. Uses a small packet to do so.
A bit overkill for some of this but just covering the bases.
* fix issues with in-game join
* fix multi messaging and multi pause screen
* make sure maintenance stuff (pings, voice, file xfer, etc.) isn't
affected by pausing or time compression
Helps us by also dumping the amount of data that has been processed from that packet so far.
Also, dump how much data we need to get through to finish the packet correctly.
Previously, the game used a dead-reckoning system for measuring time, where several time counters were incremented as long as time was running. This unfortunately has no way to synchronize time in the long term, and therefore in-game time will tend to run faster or slower than real time by a few percent, depending on various factors.
This PR redesigns the timestamp system to use the SDL time counter which will always keep it in sync with real time. However this necessarily involves a bit of clever handling when the game is paused and unpaused. It also revealed several minor bugs in several places in the code where timestamps were handled.
This PR is currently in draft status until the design can be updated to handle time compression. In all other respects it should be working.
* refactor hud escort list to be less crazy
- don't keep rebuilding list
- add and remove ships more efficiently
- better handling of player ships in multi (fixes#1674)
* fix sorting issues
- incomplete priority sort
- reversed alpha sort
* fix additional sorting issue
* improve safety check in list sort function
* add safety checks to hud_add_ship_to_escort()
this wasn't meant to be used outside of the module so
it was not really safe to use with sexps previously
* remove entries properly in hud_add_remove_ship_escort()
* Limit instances of Client Collisions to Avoid Instant Death
* Set timestamps correctly (take 2)
* Handle server check the same everywhere
* Clearer Logic Here
* Multiplayer Dumbfire Rollback
This creates a packet that sends the relative position and orientation info for a primary or dumbfire missle shot. The client creates this packet when firing those two types of weapons and sends it to the server. The server goes back to the frame that the client fired the weapon and quicky resimulates all the frames in-between.
This includes collision detection, restoring positions, orientations and veloicties, but not subsystem orientations.
* Address Feedback
* allow clients to send subsystem info
* Check for object number not weapon number
* Readability Adjustment
* Object Update Refactor
1. Keeps the object update packet aligned, even if the ship is dying or has not yet arrived on the client.
2. Properly handles model point shields (yes, that's what was causing multi to crash the most!)
3. Reworks positional interpolation timing and adds rotational interpolation, by sending timing information with every object update packet in a header shared between objects.
4. Saves a lot of bandwidth by only sending data that is changing. (no more sending all subsystems once per second!)
5. Better handles out of order packets (but can still use their information if recent enough!)
6. Builds the basis for Rollback by beginning to record the positions and orientations and velocities of ships for the last 30 frames.
7. Adds subsystem rotation -- Yeah, you saw that right, subsystems will now spin in multi.
8. Waypoint movement is now enabled via sexp, and time to arrival of waypoints is updated on the target info gauge.
* Fix edge case crash and Incorporate feedback
* Clean Up Interpolate Angles Quick
* Restrain the client
* How was this still here???
* Make sure every part of OO uses OOC_INDEX_NULLPTR_SUBSYSTEM define
* Fix Way to Store Return Values
* Ok travis, we get it
* Ok travis.........
This introduces a new mechanism for handling bit flags. Instead of using an int (or two, or three) to write bits into, this uses the std::bitset in a wrapper class that accepts an enum class as input. This means that certain classes of errors we've had a few of in the engine, where a coder used the wrong set of flags or checked them against the wrong variable (SF_STEALTH vs SIF_STEALTH, for example) can now be detected at a compiler level rather than not at all.
Remove support for the IPX protocol. The IPX selector button in the multi
options menu is now hidden, and the TCP button is disabled since there is no other option.
