Crashes at end of turn around game turn 80 or later?

[Geoffrey of Villehardouin]

I started 4 campaigns recently in the Ultimate version of DIK. In all cases I had crashes at the end of the AI turn around game turn 80 or so. I could not find anything wrong with the characters and armies when I lifted the fog of war. The crashes seem to happen just as the new turn is about to start after most/all AI movements have been done. I have experienced them both in normal and multiplayer campaigns. Has anyone else experienced end of turn crashes?

[sullivanclan1]

I have to! This is a common error this mod. Idk how to solve it either. Have you taken a look at the system log?

[alhoon]

I don't know this mod but perhaps an event that fires up causes the crush?

[Geoffrey of Villehardouin]

Thank you both for replying. A common error! - that is unfortunate!

The error log complains about some ancillaries. However, there are thousands of lines after that and before the log ends with the "encountered an unexpected error" message. I am attaching the error log and the campaign script (zip file). The last error is about an entry in the campaign script in line 1921 about an heir ancillary. The campaign script says that the ancillary relates to faction heir turk and the error is that the faction "turk" does not exist, which of course does and is still alive, although maybe it should be turks not turk?

One remote possibility is that some kind of temp file or similar is being generated and it may have reached a size limit but I cannot put my finger to it. I remember having once some type of a huge temp file being generated in another mod during development, which got so big that the campaign would unexpectedly crash mid-game. A user had reported it to Gigantus and he had resolved it.

[Gigantus]

The log ends (nearly half a second, an eternity, after the last entry) when processing trait triggers, which leads to the assumption that something is messed up there - maybe something like the BecomesFactionLeader bug.
That's apart from numerous errors in that file, none of which should be fatal, the triggers simply won't execute.

My suggestion: make a backup of the EDCT file, then completely remove the trigger section and see if the round progresses. Don't forget to enable the default EDCT again but keep the trigger less version handy for next time.

Edit: the error appears to be in the 'savoy' faction's turn.

[Geoffrey of Villehardouin]

The log ends (nearly half a second, an eternity, after the last entry) when processing trait triggers, which leads to the assumption that something is messed up there - maybe something like the BecomesFactionLeader bug.
That's apart from numerous errors in that file, none of which should be fatal, the triggers simply won't execute.

My suggestion: make a backup of the EDCT file, then completely remove the trigger section and see if the round progresses. Don't forget to enable the default EDCT again but keep the trigger less version handy for next time.

Edit: the error appears to be in the 'savoy' faction's turn.

Wow, instant response! I removed the BecomesFactionLeader section from the EDCT as you suggested but the crash persists. I suspect it may have something to do with the Turkish leader, because of the spelling error. In this campaign, the crash happened during the Savoy game turn 80 but subsequently, after moving a character that was centred on just before the crash to another location and re-launching the campaign, it is happening at the very end of the Rebels turn. In other campaigns it happened at game turn 84, 86 and I think at game turn 68 in a hotseat test campaign. At that point (i.e.68-86 game turns into the campaign), a young heir might be replacing a dead king, so it may have something to do with the BecomesFactionLeader trigger. However, the mod does not crash at the Turks' turn.

[Gigantus]

I was actually suggesting to completely remove the trigger section as it may not necessarily be the leader bug, but rather a different event\trigger.

That leader bug only causes a crash when the FL and FH get killed in the same battle and they are the last adult males in the family tree. It's a 'last stand' situation.

[Geoffrey of Villehardouin]

I was actually suggesting to completely remove the trigger section as it may not necessarily be the leader bug, but rather a different event\trigger.

That leader bug only causes a crash when the FL and FH get killed in the same battle and they are the last adult males in the family tree. It's a 'last stand' situation.

Sorry, Gigantus, I am not as much of an expert as you - to say the least! I edited out that section from the EDCT. I thought the campaign script is not savegame compatible. Then removing the trigger from it will not make a difference for this campaign. Is that not the case?

Maybe if you could say what to remove from which file, it will be clearer to me.

[Gigantus]

To get around possible crashes caused by EDCT triggers (like the BecomesFactionLeader) simply remove all triggers from the file (after making a backup). The change is save game compatible as the traits themselves remain in the EDCT.

Once you gotten past the crash simply re-instate the original EDCt again.

[Geoffrey of Villehardouin]

I corrected the trigger about faction heir turk into turks in the campaign script and I no longer get that error at game turn 80 or so. However, there are a bunch of other errors now around game turn 80-84 or thereabouts that have to do with character ancillary triggers. The error log complains that the ancillary description does not exist. It exists in the EDCT but perhaps it should also exist in another file, perhaps a text file, and it is missing there. Editing the EDCT in the way that you suggested does not resolve the mid-game crash. I will keep troubleshooting but I would probably have more of a chance of success if I was more used to and more active at modding everything. As for Broken Crescent, many Broken Crescent scripts have been removed from the latest version 2.4.2 but not the buildings, event announcements, armies on a holiday or other things associated with them. Without a script things are not going to happen. That's a different issue.

[Gigantus]

The ancillary description line exists in EDA, which is then linked to it's entry in export_ancillaries. A missing description, while sloppy, is not fatal afaik. Simply adding a requisite dummy entry will remove that error message. Nothing worse then sloppy coding when it comes to bug hunting...

I understand there is the possibility to trigger ancillaries in the EDCT, if it complains there then I think it would complain about non-existence of the ancillary, not it's description.

Leftovers: happens when modders not familiar with the script work on it, eg do not trace if there are 'dependents' when removing stuff.