Hard Light Productions Forums
Hosted Projects - Standalone => Diaspora => Diaspora Multiplayer Open Beta => Topic started by: karajorma on February 21, 2014, 02:54:52 am
-
As promised, here is the the first mission I've converted to multiplayer. It might be a little rough around the edges but it should be playable. The mission should support up to 12 players at one time (and I'd be very interested to see if the game can actually handle that many players correctly).
You'll need to have patch 1.1 installed. You'll need to download this 7zip file (http://diaspora.fs2downloads.com/Diaspora_Multiplayer_Beta.7z) and extract it to your Diaspora folder. Then run the launcher. Go to the mods tab and activate the Diaspora Multiplayer Beta option. Then go to the Basic Settings tab and make sure you select the FS2 Open 3.7.1 executable (Games will not work with the patch 1.1 exes!).
Once running a multiplayer game, you should now see a new mission called The Strike.
I need people to playtest the mission as much as possible. The kind of feedback I need include.
1) Did the tube launches work correctly? In order to make them work in multiplayer, the tube launch section of the mission was completely rewritten. The game now uses a scripting system designed by m|m and a new SEXP coded by myself in order to work in multiplayer. It's possible that one or the other has bugs still. Especially when it comes to the other players or the AI (it should work for Player 1).
2) Did combat landings work correctly? These have been reported as a source of error since patch 1.1 and again the events to handle combat landings have been rewritten somewhat.
3) Is the mission too hard or too easy for the number of players you played with. What difficulty did you play on? How skilled are the players at Diaspora? Did you change the number of respawns? What would be a good number of respawns to default to?
4) Was there any issues with lag or the underlying netcode?
5) Does the mission perform correctly when played on a standalone? Does it work when a player is also on the server?
6) Any other details you can give me will help.
I'm going to start work on converting the other missions, but the feedback from this one is very important in helping me balance things and get everything working correctly right from the start. More details on the plans for the Open Beta itself can be found here (http://www.hard-light.net/forums/index.php?topic=86896.new#new). So can we keep discussion on this thread to this particular mission please.
For any coders trying this out (or anyone on Linux who needs to make their own builds) here is a patch against r10467.
Index: code/parse/sexp.cpp
===================================================================
--- code/parse/sexp.cpp (revision 10465)
+++ code/parse/sexp.cpp (working copy)
@@ -677,6 +677,7 @@
{ "damaged-escort-priority-all", OP_DAMAGED_ESCORT_LIST_ALL, 1, MAX_COMPLETE_ESCORT_LIST, SEXP_ACTION_OPERATOR, }, // Goober5000
{ "set-support-ship", OP_SET_SUPPORT_SHIP, 6, 7, SEXP_ACTION_OPERATOR, }, // Goober5000
{ "script-eval", OP_SCRIPT_EVAL, 1, INT_MAX, SEXP_ACTION_OPERATOR, },
+ { "multi-eval", OP_SCRIPT_EVAL_MULTI, 2, INT_MAX, SEXP_ACTION_OPERATOR, },
{ "debug", OP_DEBUG, 2, 2, SEXP_ACTION_OPERATOR, }, // Karajorma
{ "do-nothing", OP_NOP, 0, 0, SEXP_ACTION_OPERATOR, },
@@ -21269,7 +21270,7 @@
}
//WMC - This is a bit of a hack, however, it's easier than
-//coding in a whole new SCript_system function.
+//coding in a whole new Script_system function.
int sexp_script_eval(int node, int return_type)
{
int n = node;
@@ -21304,6 +21305,97 @@
return r;
}
+void sexp_script_eval_multi(int node)
+{
+ char s[TOKEN_LENGTH];
+ bool success = true;
+ int execute_on_server;
+ int sindex;
+ player *p;
+
+ strcpy_s(s, CTEXT(node));
+
+ node = CDR(node);
+
+ execute_on_server = is_sexp_true(node);
+
+ node = CDR(node);
+
+ multi_start_callback();
+ multi_send_string(s);
+ // evalutate on all clients
+ if (node == -1) {
+ multi_send_bool(true);
+ execute_on_server = 1;
+ }
+ // we have to send to all clients but we need to send a list of ships so that they know if they evaluate or not
+ else {
+ multi_send_bool(false);
+
+ do {
+ p = get_player_from_ship_node(node, true);
+
+ // not a player ship so skip it
+ if (p == NULL ){
+ node = CDR(node);
+ continue;
+ }
+ else {
+ // if this is me, flag that we should execute the script
+ if (p == Player) {
+ execute_on_server = 1;
+ }
+ // otherwise notify the clients
+ else {
+ sindex = ship_name_lookup(CTEXT(node));
+ multi_send_ship(sindex);
+ }
+ }
+
+ node = CDR(node);
+ } while (node != -1);
+ }
+
+ multi_end_callback();
+
+ if (execute_on_server) {
+ success = Script_system.EvalString(s, NULL, NULL, s);
+ }
+
+ if(!success) {
+ Warning(LOCATION, "sexp-script-eval failed to evaluate string \"%s\"; check your syntax", s);
+ }
+}
+
+void multi_sexp_script_eval_multi()
+{
+ int sindex;
+ char s[TOKEN_LENGTH];
+ bool sent_to_all = false;
+ bool success = true;
+
+ multi_get_string(s);
+ multi_get_bool(sent_to_all);
+
+ if (sent_to_all) {
+ success = Script_system.EvalString(s, NULL, NULL, s);
+ }
+ // go through all the ships that were sent and see if any of them match this client.
+ else {
+ while (multi_get_ship(sindex)) {
+ Assertion(sindex >= 0, "Illegal value for the ship index sent in multi_sexp_script_eval_multi()! Ship %d does not exist!", sindex);
+ if (Player->objnum == Ships[sindex].objnum) {
+ success = Script_system.EvalString(s, NULL, NULL, s);
+ }
+ }
+ }
+
+ if(!success) {
+ Warning(LOCATION, "sexp-script-eval failed to evaluate string \"%s\"; check your syntax", s);
+ }
+}
+
+
void sexp_force_glide(int node)
{
ship *shipp;
@@ -23987,6 +24079,11 @@
case OP_SCRIPT_EVAL:
sexp_val = sexp_script_eval(node, OPR_NULL);
break;
+
+ case OP_SCRIPT_EVAL_MULTI:
+ sexp_script_eval_multi(node);
+ sexp_val = SEXP_TRUE;
+ break;
case OP_CHANGE_IFF_COLOR:
sexp_change_iff_color(node);
@@ -24407,6 +24504,10 @@
multi_sexp_set_ets_values();
break;
+ case OP_SCRIPT_EVAL_MULTI:
+ multi_sexp_script_eval_multi();
+ break;
+
// bad sexp in the packet
default:
// probably just a version error where the host supports a SEXP but a client does not
@@ -25010,6 +25111,7 @@
case OP_SET_SECONDARY_WEAPON:
case OP_SET_NUM_COUNTERMEASURES:
case OP_SCRIPT_EVAL:
+ case OP_SCRIPT_EVAL_MULTI:
case OP_ENABLE_BUILTIN_MESSAGES:
case OP_DISABLE_BUILTIN_MESSAGES:
case OP_LOCK_PRIMARY_WEAPON:
@@ -27037,6 +27139,14 @@
case OP_SCRIPT_EVAL:
return OPF_STRING;
+ case OP_SCRIPT_EVAL_MULTI:
+ if (argnum == 0)
+ return OPF_STRING;
+ else if (argnum == 1)
+ return OPF_BOOL;
+ else
+ return OPF_SHIP;
+
case OP_CHANGE_IFF_COLOR:
if ((argnum == 0) || (argnum == 1))
return OPF_IFF;
@@ -28796,6 +28906,7 @@
case OP_DAMAGED_ESCORT_LIST_ALL:
case OP_SET_SUPPORT_SHIP:
case OP_SCRIPT_EVAL:
+ case OP_SCRIPT_EVAL_MULTI:
return CHANGE_SUBCATEGORY_OTHER;
case OP_NUM_SHIPS_IN_BATTLE:
@@ -32392,17 +32503,25 @@
},
{OP_SCRIPT_EVAL_STRING, "script-eval-string\r\n"
- "\tEvaluates script to return a string"
+ "\tEvaluates script to return a string\r\n\r\n"
"Takes 1 argument...\r\n"
"\t1:\tScript\r\n"
},
{OP_SCRIPT_EVAL, "script-eval\r\n"
- "\tEvaluates script"
+ "\tEvaluates script\r\n\r\n"
"Takes at least 1 argument...\r\n"
"\t1:\tScript to evaluate\r\n"
},
+ {OP_SCRIPT_EVAL_MULTI, "multi-eval\r\n"
+ "\tEvaluates script\r\n\r\n"
+ "Takes at least 2 arguments...\r\n"
+ "\t1:\tScript to evaluate\r\n"
+ "\t2:\tTrue/False - Should the script evaluate on the server?\r\n"
+ "\t(rest):\tList of players who should evaluate this script. If no player is given, all clients will execute the script\r\n"
+ },
+
{OP_FORCE_GLIDE, "force-glide\r\n"
"\tForces a given ship into glide mode, provided it is capable of gliding. Note that the player will not be able to leave glide mode on his own,, and that a ship in glide mode cannot warp out or enter autopilot."
"Takes 2 Arguments...\r\n"
Index: code/parse/sexp.h
===================================================================
--- code/parse/sexp.h (revision 10465)
+++ code/parse/sexp.h (working copy)
@@ -719,6 +719,7 @@
#define OP_SET_ETS_VALUES (0x0023 | OP_CATEGORY_CHANGE2 | OP_NONCAMPAIGN_FLAG) // niffiwan
#define OP_CALL_SSM_STRIKE (0x0024 | OP_CATEGORY_CHANGE2 | OP_NONCAMPAIGN_FLAG) // X3N0-Life-Form
#define OP_SET_MOTION_DEBRIS (0x0025 | OP_CATEGORY_CHANGE2 | OP_NONCAMPAIGN_FLAG) // The E
+#define OP_SCRIPT_EVAL_MULTI (0x0026 | OP_CATEGORY_CHANGE2 | OP_NONCAMPAIGN_FLAG) // Karajorma
// defined for AI goals
#define OP_AI_CHASE (0x0000 | OP_CATEGORY_AI | OP_NONCAMPAIGN_FLAG)
[attachment deleted by an evil time traveler]
-
hi! with my friend up standalone server. starting dogfight mission(v3.7.1) - its OK! but if launch any coop-mission - server is dawn or game crash!
-
Try running the standalone in debug mode and tell me what kind of error message you get. You may get a few warnings you can click through by choosing No first.
-
Yes, we'd be very interested in any detailed error reports from the standalone server with this mission.
-
Try running the standalone in debug mode
you have little manual for this?
-
Basically go to the Diaspora folder and click on the file called fred2_open_Diaspora_R1_Debug. You will then find a file in Data called fs2_standalone.log or fs2_open.log. Get me a copy of that file. It will hopefully allow us to figure out what is causing your crashes.
-
logs -
[attachment deleted by an evil time traveler]
-
logs -
Is triggering from an Assertion() within the code base. See below for extract but I will take a look.
ASSERTION: "chunk_type == OP_DEFPOINTS" at modelcollide.cpp:735
-
I have ran into a problem myself.
The tube launch works when I put Alpha 1 into a tube, attached via pod. The launch goes smoothly, then my controls are frozen going at the speed of 0. I tried to fix this by putting a event chained under the Launch that set Alpha 1 mobile. However when I do this the ship exponentially increases its speed towards FTL standards and crashes the game.
No errors were encountered aside from the crash.
PS: multi mode was Coop
-
Is this when making your own missions or with Mission 2?
-
Its when I'm making my own missions from scratch. I just have a fade in, a message is sent. Then the Viper is launched.
Edit: I thought this topic appropriate since the Multi-beta DL-link is on this topic.
-
If you want, I can take a look at the mission. Or you can just post the SEXPs involved in the launch.
-
This is the SP version wherein I test stuff. So its only meant to accommodate one player. I didn't put any triggers/timers on it to trigger it yet, because I just wanted to see that it worked.
$Formula: ( when
( true )
( do-nothing )
( multi-eval
"tblParShip('BS Dalliance')"
( true )
)
( multi-eval
"tblSelDock('Stbd Tube1')"
( true )
)
( multi-eval
"tblSelShip('Alpha 1')"
( true )
)
( multi-eval
"tblLaunch('Alpha 1')"
( true )
)
)
+Name: Tube Launch Alpha 1
+Repeat Count: 1
+Interval: 1
+Chained: 5
-
Try having two events instead and putting a delay between the events with the docking and the launch.
-
I get an error now whenever I try to launch. I took your suggestion and put it up intwo different parts. First I chained the launch part with no delay to the first scipts. Then I unchained it. When I break into debugger mode I just have no sound or camera shake, but have the same results as when I didn't run into an error.
This was the same error as to when I didn't have the data needed. Also was the same for both failed script executions. Will put a DL link up soon. The mission itself isn't really filled to the brim with events as I've been working to make my first complete.
http://www.sendspace.com/file/ba0emr
Could not find index 'Player' in type 'HookVariables'
------------------------------------------------------------------
ADE Debug:
------------------------------------------------------------------
Name: (null)
Name of: (null)
Function type: (null)
Defined on: 0
Upvalues: 0
Source: (null)
Short source:
Current line: 0
- Function line: 0
------------------------------------------------------------------
------------------------------------------------------------------
stack traceback:
[C]: ?
[string "tubeLaunch-sct.tbm - On Gameplay Start"]:132: in function 'processLaunchStartDef'
[string "tubeLaunch-sct.tbm - On Gameplay Start"]:113: in function 'processTubeDef'
[string "tubeLaunch-sct.tbm - On Gameplay Start"]:100: in function 'onFrame'
[string "Simulation"]:2: in main chunk
------------------------------------------------------------------
1: Userdata [HookVariables]
2: String [Player]
------------------------------------------------------------------
-
I believe there's a scripting bug (feature?) where some HookVariables aren't initialised on the 1st gameplay frame. Try delaying the tubelaunch by 1 sec, maybe with ( has-time-elapsed 1 )?
-
Well now I'm having the problems I first had. It launches then player control isn't given back. Also really weird, even though my HUD tells me I am going 0/ms when I check my distance to the Dalliance it is increasing fast. I mean really fast, give it half a minute or so and you are over 20000 meters from the Dalliance.
-
That sounds a bit like a code bug that was fixed recently. Which version of Diaspora are you using?
-
1.1, using the data file provided by this link, and 3.7.1 provided to me by karajorma on a thread I posted.
Link to build I'm using: http://diaspora.fs2downloads.com/Multiplayer_Beta_Patch_1.1_Builds.7z
-
OK - that's me out of ideas for now. If I get a chance in the next few days I'll see if I can reproduce the issue on my PC.
-
Try the builds in the 1.1.1 patch instead.
-
So, jumping on this as I badly want to see the other missions converted to co-op (and from your recent comments (http://www.hard-light.net/forums/index.php?topic=88018.msg1759112#msg1759112) Karajorma it sounds like you're still badly in need of some testing for this).
Come on guys, this is co-op for one of the best free games ever we're talking about here! Let's get some more enthusiasm in place! :P
I gave The Strike a try tonight in my local testlab (an overpowered gaming PC and a slightly under-powered laptop, both on the same local network) , but hit some problems right out the door (literally); both player 1 and player 2 will tube-launch successfully. However, from player 2's perspective, player 1 is still "stuck" inside of the launch tube and hasn't actually launched, along with a number of other fighters. Upon inspection it looks like all of the fighters that are supposed to tube-launch simply don't from player 2's perspective - i.e. red, green, and blue squadrons stay stuck in the tubes, while yellow squadron (which I think starts the mission already in space, although I could be wrong?) appears fine.
This is all on the client end - from the host perspective, the fighters are in space and working fine, which means both players effectively go out of sync. Not completely so though - if the host player parks himself in front of the broken client player in space and shoots, the client player will see the projectiles but they just appear out of thin air (because on his screen the host player is still in the launch bay). As another side effect, the client player will also get random debug spam about docking bays already being occupied any time an AI ship in one of the playable squads dies and tries to respawn:
Network Killing off Delta 4
Network Killing off Red 3
Network Killing off Red 4
WARNING: "Docking bay "Stbd Tube5" is already occupied with ship ""!" at lua.cpp:11832
Frame 2850 too long!!: frametime = 3.741 (3.741)
Frame 2852 too long!!: frametime = 0.254 (0.254)
WARNING: "Docking bay "Stbd Tube7" is already occupied with ship ""!" at lua.cpp:11832
Yeah - basically all side effects of the tube-launch not seeming to work correctly from the client side :)
This was in a standard multiplayer game - I'll attach the logs from the client where the issues occurred (I can grab the host logs too if you need them).
As for the standalone server, I actually couldn't get far enough to test the mission properly. I could launch the standalone server, open a game client on my other system, and join it. But a few seconds after I started "The Strike", the server crashed (before my client even hit the mission-loading state) with the following pair of errors:
Desktop (debug):
---------------------------
Error!
---------------------------
Malloc Failed!
ntdll.dll! ZwWaitForSingleObject + 21 bytes
kernel32.dll! WaitForSingleObjectEx + 67 bytes
kernel32.dll! WaitForSingleObject + 18 bytes
fs2_open_Diaspora_R1_Debug.exe! <no symbol>
fs2_open_Diaspora_R1_Debug.exe! <no symbol>
fs2_open_Diaspora_R1_Debug.exe! <no symbol>
fs2_open_Diaspora_R1_Debug.exe! <no symbol>
fs2_open_Diaspora_R1_Debug.exe! <no symbol>
fs2_open_Diaspora_R1_Debug.exe! <no symbol>
fs2_open_Diaspora_R1_Debug.exe! <no symbol>
fs2_open_Diaspora_R1_Debug.exe! <no symbol>
fs2_open_Diaspora_R1_Debug.exe! <no symbol>
fs2_open_Diaspora_R1_Debug.exe! <no symbol>
fs2_open_Diaspora_R1_Debug.exe! <no symbol>
fs2_open_Diaspora_R1_Debug.exe! <no symbol>
fs2_open_Diaspora_R1_Debug.exe! <no symbol>
fs2_open_Diaspora_R1_Debug.exe! <no symbol>
fs2_open_Diaspora_R1_Debug.exe! <no symbol>
fs2_open_Diaspora_R1_Debug.exe! <no symbol>
fs2_open_Diaspora_R1_Debug.exe! <no symbol>
fs2_open_Diaspora_R1_Debug.exe! <no symbol>
fs2_open_Diaspora_R1_Debug.exe! <no symbol>
fs2_open_Diaspora_R1_Debug.exe! <no symbol>
fs2_open_Diaspora_R1_Debug.exe! <no symbol>
fs2_open_Diaspora_R1_Debug.exe! <no symbol>
fs2_open_Diaspora_R1_Debug.exe! <no symbol>
fs2_open_Diaspora_R1_Debug.exe! <no symbol>
fs2_open_Diaspora_R1_Debug.exe! <no symbol>
[...]
[ This info is in the clipboard so you can paste it somewhere now ]
Use Ok to break into Debugger, Cancel to exit.
---------------------------
OK Cancel
---------------------------
(Hit Cancel)
---------------------------
Microsoft Visual C++ Debug Library
---------------------------
Debug Error!
Program: D:\Diaspora\fs2_open_Diaspora_R1_Debug.exe
HEAP CORRUPTION DETECTED: after Normal block (#85165) at 0x0942B760.
CRT detected that the application wrote to memory after end of heap buffer.
Memory allocated at d:\code\freespace\diaspora_dev\code\globalincs\windebug.cpp(1784).
(Press Retry to debug the application)
---------------------------
Abort Retry Ignore
---------------------------
Other missions worked fine, it was only The Strike that caused it to shoot itself in the head. The corresponding server log is also included in my attachments.
I think once the tube launch issue is tackled, I can then move on to checking out the other bits in more detail (combat landings, latency, etc).
[attachment kidnapped by pirates]
-
That's rather strange. The tube launch aspect of The Strike is the part I tested the most heavily and it did work at one point.
I'm currently on holiday and it will be a couple of weeks until I'm in a position to have two PC's again so that I can check this. But I'll definitely look into it as soon as I get home.
-
Well, finally I played it. First, and most important: it works with recent nightlies! Tubes launch, Cylons arrive, goals fail - everything like it should be :)
Looking closely into it I found that notorious multiplayer (MP) problem is in full power here. It shows itself the most in mods like SWOT where:
1) speeds of ships and projectiles are much greater then in original FS;
2) ship sizes are smaller;
3) autoaim is used;
4) maybe some other factors.
It leads to your ability to hit target (which obviously drops in MP) drop almost to zero. If we express this in numbers it will be like this. I have following accuracy against fighters:
1) FS (no mods except MVPs) single player - ~25%
2) FS MP - ~12% (still playable I think)
3) SWOT or Diaspora - ~0-1%
It is all due to client-server data discrepancy, you can easily see this effect by watching your projectiles "hit" the enemy on your screen, and enemy ship could even start to burn, but its HUD damage stays like it was. I have a video how it is seen from client and server: https://www.youtube.com/watch?v=tinl9RIEBdY
BTW, -cap_object_update 0 wouldn't do any good here - enemy ships may appear to fly more smooth, but it is only on client side, on server side discrepancy may be even worse due to the low update rate.
But I hope we could overcome this problem in some time - there are several people on it.
Later I'll write down some unrelated issues with this mission, which I hope could be fixed without messing with the engine.
-
Well I'm glad the mission itself seems to be mostly working. I'll happily look into any remaining issues.
-
Here it goes.
I played as a client with no player as a server. I used 01-oct and 15-oct nightlies with no notable difference between them.
1) Strangely enough server doesn't react to "\" presses (while client do). That's just minor issue since "launch by time" performs just fine.
2) On the second respawn my client has "Glide" state locked (and I can't get rid of it pressing keys) and in this state he appears in the tube, and glides all the way to the left through the "Theseus" no matter of its hull structure (like if "no collision" is on). Third respawn is ok since it is outside the Theseus (but no "Glide" is on as well).
3) Theseus turrets animates and behaves somehow erratic (see the screen (https://s116vla.storage.yandex.net/rdisk/e0fe00797875daccb0bd023d5b0984ee361f273994ad9ccbfd9520b829855ad6/5bc78ba8/llEraxqXholA55PG7QR_T7kGZVM4PzDRJHYUSmAgoStPsm6Y5SmSYUttzU9KA6rhCu22W6cTdyz-26iAk-f8Tg==?uid=664943&filename=%D0%A1%D0%BA%D1%80%D0%B8%D0%BD%D1%88%D0%BE%D1%82%202018-10-14%2000.34.50.png&disposition=inline&hash=&limit=0&content_type=image%2Fpng&fsize=2659094&hid=981e9f91c517520771e6bb00c36b5e33&media_type=image&tknv=v2&etag=b8ff4f70d7ca4376ead8c4ba74a2f11b&rtoken=b9Vwt5JLUEet&force_default=yes&ycrid=na-98e79ce058849b559fa48a7d46846590-downloader6h&ts=5787192bc3a00&s=2fc216a8dd4e11571430d1abd8fd35a39218d63cdb1c1616ccedf46a07371075&pb=U2FsdGVkX1_uY3Dunke5oxq3id05W3EqMzwCQCiev0zZeKX3gMdqNmuP06T4Dq26ABgyXB18CpQeq1cRQe5-FJiq6ejkQ1h1NZIHJJogdwQ)).
4) I haven't investigate in details, but seems that AI wings doing nothing, at least I didn't feel any cover from them (which is supposed to be from the briefing).
That's all I remembered for now, maybe there will be more.
-
UPD By "AI doing nothing" I mean, for example, that nobody attacks missile launchers with me.
-
They don't make any attempt to cover you?