I did some mock-games to test kurio OOS. Here are my findings:
They look similar the logs I obtained... Random changes that do not seem related to the presence of the Kurios in another civilizations. For any future cases, could you please set the SynchLog variable in CivilizationIV.ini to 1? It generates the generation of an additional log file, MPLog, which includes additional information and should be sent along with OOSLog. In a few of my recent testing games for the Kuriotate OOS the MPLog files were identical except for ASYNC random values, though
Also, with regard to Game 2, bear in mind that there are also OOS issues when neither the Kuriotates nor the Infernals are present in the game. They are very rare, but I have seen them happening.
Tholal: In the link you posted I saw how to run multiple instances of the game on a single computer. Knowing that this is possible, I have modified the mechanism by which the MPLog file is generated to force it to have different names depending on the player active in the game. This means that it is now possible to debug OOS issues and obtain logs for it with a single computer.
The process goes as follow. To be able to launch multiple instances of the mod, you need to add "multiple" to the parameters of Civ4BeyondSword when launching it. For example, I'm using:
Code:
"C:\Program Files (x86)\2K Games\Firaxis Games\Sid Meier's Civilization 4 Complete\Beyond the Sword\Civ4BeyondSword.exe" multiple mod=\FFH2Dev
To generate synchronization logs with different names, you need the modified CvRandom file attached to this post. You will also need to connect with different player names to the multiplayer game, since the file will be called SynchLog - [USERNAME].txt.
This approach also allows to do something that is quite helpful in finding OOS errors; having both synchronization logs open at the same time with an application such as WinMerge. You can press F5 every few turns on WinMerge to check if something is already wrong, before the game catches the desynch and throws an OOS error. The only problem I have noticed with this approach is that sometimes you will get errors when creating autosaves (I wouldn't trust those too much when using this method).
I have also created a testing guide which gives instructions on how to enable logs and crash dump generation and how to provide useful information along with bug reports, intended for users. The tutorial speaks about ExtraModMod specific stuff such as beta builds and links to my tracker, but it should be trivial to adapt it to More Naval AI if you want to use it. You can check it here:
https://bitbucket.org/Terkhen/extramodmod/wiki/Testing (feedback and suggestions to improve it are welcome). This guide is already updated for my SynchLog change.
I'm attaching three additional sets of logs for the Kuriotate issue, this time with synchronization logs too. The synchronization logs show no differences besides ASYNC random values, and therefore do not give us any new information. In one of the games I forgot to add AIs and I got no OOS errors at all (that game was played until turn 80 or so). I wonder how the presence of a human Kuriotate player can affect AIs that have not even met that player, though...
EDIT: I have made similar modifications to how the BBAI logs are generated. They now include the player name in the text file's name (to allow to differenciate them more easily when getting logs from multiple players and to allow to gather them when using multiple instances on the same computer). The log file no longer includes the timestamp, as that made WinMerge comparison between logs of different players very complicated. In case you are interested, I'm attaching the modified file to this post.