Multiplayer Woes

[BlueGenie]

Going to keep trying until someone of us MP players give up.
We restarted again, this time from having freshly started CivIV/C2C games.
Again same settings/files.
At end turn we get OOS. This time Player 1 has the different value.
All logs we could find are attached.

Cheers

 

[AIAndy]

Going to keep trying until someone of us MP players give up.
We restarted again, this time from having freshly started CivIV/C2C games.
Again same settings/files.
At end turn we get OOS. This time Player 1 has the different value.
All logs we could find are attached.

Cheers

Thank you for contributing to this.
It is still quite weird with all random logs showing the same but the Rand being different at the time when the log is written.
Hmm, who was the first one to press end turn?

 

[BlueGenie]

This time Player2, then Player0, lastly Player1.

Cheers

 

[AIAndy]

This time Player2, then Player0, lastly Player1.

Cheers

It seems like my deactivation of the MinimizeAITurnSlices option was not successful (I checked with the wrong function, will upload the fix shortly). Did you have that all deactivated?

 

[BlueGenie]

We had that inactive, yes.

Cheers

 

[AIAndy]

I see two events for city building towards the end of the Python debug log. Which of the three players actually built a city in turn 1?

EDIT: Btw, I have now also tried using three Civ4 instances on my computer with the savegame you posted earlier and I could play on without OOS. Of course I can't really test significant delays between the speeds of the computers nor delays in the message passing and I have to tab between choosing end turn so there is no "Two people choose end turn shortly after each other".

 

[BlueGenie]

I know Player 0 did, not sure about the others. I can ask them tomorrow. *ponders* Actually I might know. Player 2 increased in points during turn 1 so I think he is the other one.

Cheers

 

[AIAndy]

I know Player 0 did, not sure about the others. I can ask them tomorrow. *ponders* Actually I might know. Player 2 increased in points during turn 1 so I think he is the other one.

Cheers

Then maybe I should check what happens if there is no city.

EDIT: Tried it with the savegame and 3 instances of C2C. Not building a city in the first turn for player 1 did not cause an OOS.

 

[Veilkost]

As much fun C2C is in SP, it seems like that in its current state it is unplayable in multiplayer. We have tried it several times now and the result is always the same: far too frequent OOSes. The best we had was like 400 turns w/o problems, then OOS every 20 turns, then 10, 5, 2..

Is there any advice on what settings to chose to not get those OOSes? Did anybody actually get the mod running successfully in MP? Also, any hints on what to do if a game does go OOS? I remember that RoM had OOSes too, but once the game was rehosted, everything went fine for hours usually. Tricks, please!

 

[ls612]

As much fun C2C is in SP, it seems like that in its current state it is unplayable in multiplayer. We have tried it several times now and the result is always the same: far too frequent OOSes. The best we had was like 400 turns w/o problems, then OOS every 20 turns, then 10, 5, 2..

Is there any advice on what settings to chose to not get those OOSes? Did anybody actually get the mod running successfully in MP? Also, any hints on what to do if a game does go OOS? I remember that RoM had OOSes too, but once the game was rehosted, everything went fine for hours usually. Tricks, please!

Post your OOS logs and your random number logs, and AIAndy can take a look at them. Also, are you using V25 or the SVN?

 

[Veilkost]

I am running v25 of C2C, no SVN version.

Can you tell me which files exactly you need? I have a lot of .txt files in my BtS folder, called OOSLog. Those?

 

[ls612]

I am running v25 of C2C, no SVN version.

Can you tell me which files exactly you need? I have a lot of .txt files in my BtS folder, called OOSLog. Those?

Yes, and anything named RandLog. Zip them up so that they will fit onto the forums.

 

[BlueGenie]

New try with the last fix you gave us AIAndy, this time with 2 players only, both with decent computers.
Managed initial start and a whooping 30 turns before a first OOS.
Still not sure exactly what logs you want/need to check so we're just handing all the logs we could find over to you.

Cheers

 

[AIAndy]

New try with the last fix you gave us AIAndy, this time with 2 players only, both with decent computers.
Managed initial start and a whooping 30 turns before a first OOS.
Still not sure exactly what logs you want/need to check so we're just handing all the logs we could find over to you.

Cheers

The likely culprit is some canTrigger event Python code calculating different for the two players. Did any of you get an event in that turn?

 

[BlueGenie]

Nope. AI might have, we had 2 in that game I think.

Cheers

 

[AIAndy]

Another thing I don't get though is why your BBAI logs are so different.
There might be some issue with AIs.

 

[BlueGenie]

Well, I don't know why, that's what I'm hoping you'll figure out though if there's something with the AI I could imagine an AI getting an event for one player but not the other would cause OOS.
If it's just AI doing things differently then Koshling might be able to figure this one out?

Cheers

 

[AIAndy]

Well, I don't know why, that's what I'm hoping you'll figure out though if there's something with the AI I could imagine an AI getting an event for one player but not the other would cause OOS.
If it's just AI doing things differently then Koshling might be able to figure this one out?

Cheers

Looking at that I found an OOS bug with the tech chooser popup that will cause an OOS quite often when only some human players get that popup or in a different sequence.
It is fixed in the latest SVN and was probably the culprit of your second turn OOS problems (if some players relied on selecting a tech in the first turn right away and some relied on the popup).

 

[BlueGenie]

Right. Sounds good.
Now, do we need to use the SVN version to keep testing for OOS (and inadvertently probably also inviting other problems) or can you give me a zip/rar-file to download with the needed fixes in?

Which reminds me, can we use the current DLL with the "older" version we are using or is it dependent on other things that are not in the plain 25 (or rather a couple of revisions before v25 came out, rev 3247 vs v25's rev 3259)?

Cheers

 

[AIAndy]

Right. Sounds good.
Now, do we need to use the SVN version to keep testing for OOS (and inadvertently probably also inviting other problems) or can you give me a zip/rar-file to download with the needed fixes in?

Which reminds me, can we use the current DLL with the "older" version we are using or is it dependent on other things that are not in the plain 25 (or rather a couple of revisions before v25 came out, rev 3247 vs v25's rev 3259)?

Cheers

I am not sure if there are any changes that would make it not work with the old assets but then I'd say try the SVN so any bug fixes in the meantime are also covered by some MP testing.
Mind though that I am away from my modding computer for the next weeks (holiday) so I won't be able to do any fixes in that time although I will be on the forum now and then.