The one that seems to have been crashing most servers (from looking at the MDMPs), is that CAI_Pathfinder::IsLinkUsable() crash, which is part of a callstack from any NPC's think function, from either the "info_node_link" or "info_node_link_controller" entities, after being deleted and re-created via an autosave.
Sort version: the crash that was seemingly caused by the autosaves was fixed.
Toggle shoutbox Shoutbox
|
|||||||||||||||||||||||||||||||||||||||||
Synergy Update version 14
#62
Posted 16 September 2009 - 05:13 AM
If needed I can provide more MDMPs from a Win2k3 Synergy server. 
Edit
Also, mp_respawndelay doesn't give any good results. Set it to 1 and it still takes like 30 seconds to respawn.
Any possibility to set it to exactly 1 second or 5 or so?
Edit
Also, mp_respawndelay doesn't give any good results. Set it to 1 and it still takes like 30 seconds to respawn.
Any possibility to set it to exactly 1 second or 5 or so?
#63
Posted 16 September 2009 - 07:27 AM
mp_respawndelay only has a 0 and 1 setting.
0: Don't respawn anyone until everyone dies and a checkpoint is loaded
1: Respawn people after time
I really wish there would be some change, or even just an EXPLANATION why the possibility of a server var for shorter/instant respawn is an issue that the devs don't seem to like to discuss.
0: Don't respawn anyone until everyone dies and a checkpoint is loaded
1: Respawn people after time
I really wish there would be some change, or even just an EXPLANATION why the possibility of a server var for shorter/instant respawn is an issue that the devs don't seem to like to discuss.
#64
Posted 16 September 2009 - 12:47 PM
Instantly respawning players would be fine if we didn't have the map reset stuff.
We're trying to make an actual game with goals, including avoiding death. There needs to be *something* that prevents players from just dying.
We're trying to make an actual game with goals, including avoiding death. There needs to be *something* that prevents players from just dying.
#65
Posted 16 September 2009 - 02:17 PM
I'm just curious, what interpreter do you use to read those MDMPs? I'd like to help by actually providing the developer with the actual error rather than throwing files at them.
#66
Posted 16 September 2009 - 03:13 PM
QUOTE (VOSK @ Sep 16 2009, 10:47 AM) <{POST_SNAPBACK}>
Instantly respawning players would be fine if we didn't have the map reset stuff.
We're trying to make an actual game with goals, including avoiding death. There needs to be *something* that prevents players from just dying.
We're trying to make an actual game with goals, including avoiding death. There needs to be *something* that prevents players from just dying.
That's all well and good, but why not leave it up to server admins?
Team Fortress 2 is supposed to be a well balanced game, but Valve doesn't restrict people from using instant respawn and game-unbalancing features.
#67
Posted 16 September 2009 - 05:40 PM
QUOTE (crazydog @ Sep 17 2009, 07:13 AM) <{POST_SNAPBACK}>
That's all well and good, but why not leave it up to server admins?
Team Fortress 2 is supposed to be a well balanced game, but Valve doesn't restrict people from using instant respawn and game-unbalancing features.
Team Fortress 2 is supposed to be a well balanced game, but Valve doesn't restrict people from using instant respawn and game-unbalancing features.
Thanks for the update and the vote toggle VOSK will have to try it out.
#68
Posted 16 September 2009 - 07:39 PM
We really need to be able to disable everyone dying and the map resetting. Think of custom maps like js_build_puzzle, don't just think about the default ones. On the off-chance that everyone dies at the same time (It would have to be exactly at the same time due to the instant respawn) then the map gets reset and all the puzzles are reset and that is VERY annoying.
#69
Posted 17 September 2009 - 02:32 AM
Half-Life 2 actually loads some time, but Synergy seems to instantly reset like in Counter-Strike. Perhaps using "real" loading would fix the problems.
#70
Posted 17 September 2009 - 06:36 AM
I actually created a SourceMod plugin for fast respawn. It seems it works fine.
I'll publish it soon after I've tested it thoroughly.
I'll publish it soon after I've tested it thoroughly.
#71
Posted 17 September 2009 - 06:51 AM
QUOTE (Ywa @ Sep 17 2009, 10:36 PM) <{POST_SNAPBACK}>
I actually created a SourceMod plugin for fast respawn. It seems it works fine.
I'll publish it soon after I've tested it thoroughly.
I'll publish it soon after I've tested it thoroughly.
Disabling the vote system works, thanks.
Any word on what is going on with the player collision? Is this going to be updated with a disable cvar?
#73
Posted 17 September 2009 - 02:36 PM
QUOTE (TGP1994 @ Sep 16 2009, 03:17 PM) <{POST_SNAPBACK}>
I'm just curious, what interpreter do you use to read those MDMPs? I'd like to help by actually providing the developer with the actual error rather than throwing files at them.
VOSK?
#74
Posted 17 September 2009 - 05:30 PM
QUOTE (TGP1994 @ Sep 17 2009, 03:36 PM) <{POST_SNAPBACK}>
VOSK?
Microsoft Visual Studio 2008 Express. Sending the actual files is most helpful. You won't be able to see anything yourself except a callstack. (Most of the time, just a callstack is useless for me.)
#75
Posted 17 September 2009 - 09:01 PM
QUOTE (VOSK @ Sep 17 2009, 06:30 PM) <{POST_SNAPBACK}>
Microsoft Visual Studio 2008 Express. Sending the actual files is most helpful. You won't be able to see anything yourself except a callstack. (Most of the time, just a callstack is useless for me.)
Oh. What do you see?
#76
Posted 17 September 2009 - 10:52 PM
QUOTE (TGP1994 @ Sep 17 2009, 10:01 PM) <{POST_SNAPBACK}>
Oh. What do you see?
Having the source files, I can be taken directly to the lines of the functions in the call stack, beginning at the line that caused the crash.
However, most crashes are happening in the engine.dll or other engine binary file, which I do not have the sources to, therefore I don't always get to see what each crash was about.
#77
Posted 17 September 2009 - 10:59 PM
If you run puzzle maps and want to stop the map from resetting if everyone dies at the same time, try adding a bot or two. There's a couple of bot plugins over in the SourceMod forums, so you might want to check those out. Unfortunately, it also means you'll have to sacrifice a player slot.
#78
Posted 18 September 2009 - 02:27 PM
QUOTE (VOSK @ Sep 17 2009, 11:52 PM) <{POST_SNAPBACK}>
Having the source files, I can be taken directly to the lines of the functions in the call stack, beginning at the line that caused the crash.
However, most crashes are happening in the engine.dll or other engine binary file, which I do not have the sources to, therefore I don't always get to see what each crash was about.
However, most crashes are happening in the engine.dll or other engine binary file, which I do not have the sources to, therefore I don't always get to see what each crash was about.
Oh. That's weird, I thought when you choose to make a multiplayer mod, it gives you the source to everything?
#79
Posted 18 September 2009 - 11:48 PM
QUOTE (TGP1994 @ Sep 19 2009, 03:27 AM) <{POST_SNAPBACK}>
Oh. That's weird, I thought when you choose to make a multiplayer mod, it gives you the source to everything?
Not the engine source code though
#80
Posted 19 September 2009 - 01:47 AM
The OrangeBox engine doesn't seem as graceful as the older EP1 engine was. Which is why it seems to be more buggy than previous releases. The older engine would work around human errors and mistakes with a sane set of defaults, but the new engine just takes a dump when it finds something that it doesn't know what to do with or is missing something that it expected to be there.

Sign In »
Register Now!
Help


Back to top
















