OOS troubleshooting

[Sephi]

That was fixed in patch H. I'm not sure whether the fix was a reversion to the previous code (which would have restored the bug where the Infernals always reduce the AC when destroying a city, even if it did not originally have AV present) or if it was new code that corrected both the OOS and the invalid AC decrease.

It was a reversion. This nice tweak from the UP fixes it

CvPlayerAI::AI_conquerCity

/************************************************************************************************/
/* UNOFFICIAL_PATCH                       06/14/09                       Maniac & jdog5000      */
/*                                                                                              */
/*                                                                                              */
/************************************************************************************************/
/* original bts code
		CvEventReporter::getInstance().cityAcquiredAndKept(GC.getGameINLINE().getActivePlayer(), pCity);
*/
		CvEventReporter::getInstance().cityAcquiredAndKept(getID(), pCity);
/************************************************************************************************/
/* UNOFFICIAL_PATCH                        END                                                  */
/************************************************************************************************/

 

[Bill Bisco]

I have a very strong suspicion that its the AI in Multiplayer games that's causing OOSes

 

[thealien_83]

If I can get ahold of any friends to experiment with, I shall endeavor to do so. But getting them together to play is like wrangling cats into a bag even at the best of times, much less now, so we'll see how that pans out.

 

[idiot2]

I have a very strong suspicion that its the AI in Multiplayer games that's causing OOSes

Come to think of it, in our test game we eliminated the AIs early...

 

[Tamman]

Just ran 2 games global highlands 2 humans 5 ai's no optional settings

Version 1
OOS turn 174

Version 2
400 turns before stopping no OOS's

Will try to get to version 3 & 4 soon

 

[xienwolf]

BTW for those of you with a nice computer that can handle it, you can test for OOS issues by yourself.

Change your FfH shortcut from:

"C:\Program Files\Firaxis Games\Sid Meier's Civilization 4\Beyond the Sword\Civ4BeyondSword.exe" mod=\Fall from Heaven 2

to

"C:\Program Files\Firaxis Games\Sid Meier's Civilization 4\Beyond the Sword\Civ4BeyondSword.exe" multiple mod=\Fall from Heaven 2


Then you can do a LAN game against yourself, If Kael is interested in enabling AIAutoPlay in multiplayer for these tests, you can see how to do so (and make it automatically stop when an OOS happens) in Fall Further, Note that AutoPlay isn't as good as real playm since the players may notice that an OOS occurs right after seeing something happen, and in autoplay you see little. But with an improved OOS logger the Autoplays become more useful.

 

[Goodgimp]

I'll download and help test this sometime next week. In the meantime, I have some speculation.

My friends and I have recently returned to FFH2 and have been playing it over the LAN. We've noticed consistent behavior from several games that we've run:

1. OOS is basically absent for the first 100 or so turns in the game (Normal Speed)
2. OOS issues come in a deluge during midgame
3. Towards late game, OOS issues drop dramatically or disappear entirely.

The strange thing is, though, is we just played two games back to back. One game had OOS issues terribly bad, and we even ended up having to abandon the game due to a repeatable crash that happened around turn 250.

So we started up a new game. The weird thing was, we played that game until turn 300+, for a few hours straight, and have not had a single OOS the entire time. Not one. This is highly unusual, we will always see multiple OOS issues by now.

I'm wondering if the desyncs may be caused by specific AI/Civs in the game? Game options have remained consistent between every run, the only factor changing is the Civs that are in the game. I do know that the Balseraphs have been in each of our games, but this desync-less one they got wiped out in the first 5 turns or so. Could be coincidence, but I'm really wondering if it has something to do with a particular Civ...

 

[dieffenbachj]

DLL4: No out of sync errors by turn 300, when the map ended. Huge, quick, all AI except for 1 other player.

On a side note, I hope that these DLL changes were intended to break gameplay in order to find the problem--in the test game, the AI never built anything but warriors. Ever. Not even palisades.

And with their cities swarming with barbarians, they never attacked the barbarians--only waited for the barbarians to attack their cities. But since the barbarians were outside the cities, the AI couldn't (or didn't want to) even improve the terrain around their cities.

So, in essence, it was a match between me, my friend, and a smattering of civilizations with one un-improved city.

So there were no out of sync errors... but, uh, if the sync errors were AI-related caused by anything besides warriors and the Fishing technology? Of _COURSE_ there were no out of sync errors. That's all there was on the map!

..oh, and the barbarians were ruthless. A touch TOO prevalent. But that's probably because the AI wasn't bothering to clean them up.

 

[Emptiness]

in the test game, the AI never built anything but warriors. Ever. Not even palisades.

What settings did you use for your game?

 

[dieffenbachj]

What settings did you use for your game?

Well, it was Huge... everything else was standard in terms of map options (Temperate, Erebus, Medium Sea Level, Ancient Era) and time was on Quick

For gameplay options, we had Permanent Alliances, City Flipping After Conquest, Blessings of Amathon, and Living World.

EDIT: And All Unique Features!

 

[Emptiness]

Strange. The current AI tends to behave like that when it's in war (does not build buildings or settlers), so I thought maybe you were using the "Always War" option. Did you or your friend declare war on the AIs immediately upon contact?

 

[Sephi]

The test dll he used disabled about everything

 

[Mr.Black]

Hello, I have a big Problem with OOS in Lan-Games

It's simply not playable!

What happend:
I tried to start a game with a friend. I have Version 0.41 and Patch J.
The Download File of the Game, and the Patch were the same, he got the files directly from my PC.
When we tried to start a Lan-Game, our computers were not able to see each other. But it was not a problem of the Network, the PC's had no problems to see and Ping each other out of the Mod.

So we made a direct connection and started a game. When the game started, the OOS massage came, so we were not able to make only one round together.

Has anybody an Idea, what we could do?

 

[Goodgimp]

Mr. Black, it really sounds like something is not installed right, especially after saying that you couldn't see each when trying to join via the LAN.

It's a pain, but if you can't figure it out you may want to try uninstalling and reinstalling both the game and the mod to make sure there aren't any modified files laying around.

 

[Emptiness]

Actually, I was under the impression players often cannot see each other's games (which is why Hamachi is commonly used) and that an OOS on the first turn of MP games was very common. The solution to this OOS, unless I'm mistaken, is to reload from the autosave. If I'm correct, then reinstalling won't correct either of those problems.

 

[dieffenbachj]

Ran DLL 2 last night. We got 100-150 turns in before the game ended (barbarians, oh god, the barbarians...) with no OOS. Farther than we've ever gotten with regular patch j without OOS, so... improvement, I suppose. Might play again tonight, might not.

 

[Goodgimp]

Sorry to bring this up again, but I played another Multiplayer LAN game (2 human, 10 ai) making sure specifically to exclude the Balseraphs but leaving all other slots as Random Good / Random Neutral. Once again, not a single desync the entire way through the game.

One that game was completed, I started a new game and included the Balseraphs (Keelyn). Game desynced about 7 times by turn 300.

EDIT: Can anyone else reproduce this with excluding the Balseraphs? I've yet to generate an OOS as long as they're not in the game in multiple attempts now.

 

[EviLive]

I dont really think it is just one civ, myself and several friends had been playing this mod (thank you so much for making it) at least every other night, and depending on the players there it was everywhere from 2 to 6 people and then some amount of AI, so about 6 to 12 usually. When someone was not on for the night we started a new game so at any given time we had maybe 5 or 6 games going at a time.

Each of these games was under patch J as we tended to discontinue any games frome earlier patches, however starting with the major AI patch and continuing on til patch J every one of these games desynched. Usually to the point that trying to reload only caused another desynch 2 turns later. Sometimes we were able to go back some turns and get it to prolong the desynch another 0 turns or so, and sometimes we went back and it desynched earlier.

Because we really do love the mod we decided to try and work with it, over a few weeks time we ran games with all different settings and races trying to pinpoint something, without the abbility to open worldbuilder in MP games or to convert MP to SP it does get hard.

All we came up with is that we do not think it is any one civ, nor any one unit, we think it is an event, an event as in the making of a particular unit, or the use of a certain spell or ability, maybe a specific building, because in all of our games we had different civs and different settings but they all went down and this usually happened after turn 90 (we had one game that oos at 90 the rest well after) but usually before 200.

I had posted some of those games and issues before, though at the time few others were seeing these issues, though I realize part of this was not as many people playing MP as SP. We eventually just took a break from civ for a bit, but seeing this thread now I think we will start looking in to it again in hopes of some fixes for it.

 

[Renaud]

Is it true that patch 'd', which I believe is the patch level of the base mod file, is stable for MP? Or do we have any recommendations for a later patch level that is really stable in MP. It would be nice to have all the updates, especially the graphical updates from later patches, but the priority has to be a stable game.

I never have crashes in single player, but have only played MP once and it was stable with no crashes, but we only played to turn 220 or so before folks left the game and we didn't save it (I believe we were playing patch g).

 

[Sephi]

if you tried all the dlls Kael posted you might also want to test my modmod. I merged my fix which seems to work in Wildmana into FFH patch j source.

http://forums.civfanatics.com/showthread.php?t=346039