Re: Bug reports for new RC/RR compilation
Quote:
Originally Posted by
Point Blank
All CTD's are derived from siege_settlement commands. We are working on solving this.
I get this on 156th turn with using 2tpy.
12:02:19.483 [game.script.exec] [trace] exec <if> at line 6033 in mods/Stainless_Steel_6/data/world/maps/campaign/imperial_campaign/campaign_script.txt
12:02:19.483 [game.script.exec] [trace] exec <siege_settlement> at line 6033 in mods/Stainless_Steel_6/data/world/maps/campaign/imperial_campaign/campaign_script.txt
12:02:19.530 [bink] [debug] [data/fmv/faction/minor_win.bik]Pause on
12:02:19.530 [bink] [debug] [data/fmv/faction/minor_win.bik]Pause on
12:02:21.890 [system.rpt] [error] Medieval 2: Total War encountered an unspecified error and will now exit.
I've modified the Sofia siege script to look like Scopia's and Adrianople's, will check out now.
Well, that was silly expecting a change without actually modifiyng the script.
Quote:
Originally Posted by
Point Blank
Yes but the script is saved with the game so you'll need a restart. I'm using TATW as a testbed for ways of fixing this issue.
Brb /wristing.
Re: Bug reports for new RC/RR compilation
Yes but the script is saved with the game so you'll need a restart. I'm using TATW as a testbed for ways of fixing this issue.
Re: Bug reports for new RC/RR compilation
Quote:
Originally Posted by
impspy
Is the daytime summer sky in Ireland supposed to look like this? (no it is not a night battle)
http://i528.photobucket.com/albums/d.../brokensky.jpg
I was turning the unused knights_templar faction into Ireland, and I got this sky sieging an English fort in Dublin province.
btw I fixed the problem (apparently midday_clear was not working) by using a newer version of weather_db found in skymod 2 alpha; all credit for the fix goes to horsearcher.
http://www.filefront.com/15141967/weather_db.xml/
Put weather_db.xml in ss6/data and click yes to overwrite, save game compatible.
Re: Bug reports for new RC/RR compilation
For some strange reason the events aren't occuring in my game. I've reinstalled a few times correctly and still nothing. The TO or Mongols don't emerge... or any other notifications for the events for that matter. Game runs, just no events. Anyone have a clue as to why its not working?
Re: Bug reports for new RC/RR compilation
steelmane - sounds like the campaign_script has failed.
have a look in the system log in the main M2tw/logs folder and check for campaign_script.txt -
see if there's any error reports
(in fact check to see if there's any error reports generally.
Re: Bug reports for new RC/RR compilation
Rozanov,
Here is the Syslog file.
Re: Bug reports for new RC/RR compilation
Steelmane - I couldn't spot anything - is your the standard RC/RR campaign_script or an amended one?
Re: Bug reports for new RC/RR compilation
Not sure if this is a bug but the battle AI seems to be very passive using this compilation - particularly with stuff like feudal knights (not general's bodyguard). I'm hitting the 'show me how' button at the start of every battle to enable the AI script and then leaving the advisor tab thing open (it doesn't close when I click show me how). Am I doing anything wrong or is this just an unfortunate thing that modded AI can't do much about? It's kinda sad when the enemy has a bunch of elite cavalry and they do them less good than some random infantry.
Re: Bug reports for new RC/RR compilation
Quote:
Originally Posted by
Point Blank
Yes but the script is saved with the game so you'll need a restart. I'm using TATW as a testbed for ways of fixing this issue.
So what is the suggested quick & dirty patch to prevent CTD's while we wait for a permanent fix ?
Shall we comment out all "siege_settlement" statements ? Is that enough ?
Re: Bug reports for new RC/RR compilation
Morfans - yes you can try that - comment out all the lines relating to the uprising emergence if you're not sure exactly what to comment out.
it means you'll need to start a new campaign.
exactly what causes the problem I don't know - I'm sure we never used to have this issue - so perhaps it's a feature added recently that's
messing with save games. (the siege_settlement part seems to fire OK whatever way you script it but causes a problem on saved games after it has fired.
If there was an easy fix it would have been posted by now (I hope!)
Re: Bug reports for new RC/RR compilation
Quote:
Originally Posted by
Rozanov
exactly what causes the problem I don't know - I'm sure we never used to have this issue - so perhaps it's a feature added recently that's
messing with save games. (the siege_settlement part seems to fire OK whatever way you script it but causes a problem on saved games after it has fired.
If there was an easy fix it would have been posted by now (I hope!)
Yes, siege_settlement itself is used successfully in other mods.
Could the pb be associated to the event counters used to activate/deactivate the siege command ? As if saved files erroneously stored the flags firing these event and on retrieving the saved data the game engine entered an infinite loop ...
Re: Bug reports for new RC/RR compilation
Quote:
Originally Posted by
Morfans
Could the pb be associated to the event counters used to activate/deactivate the siege command ? As if saved files erroneously stored the flags firing these event and on retrieving the saved data the game engine entered an infinite loop ...
I have just noticed that in the script the various event counters are generally tested to be either equal to zero or one - so if for some reasons, these numbers are stored as neither exactly 0 nor 1, the engine gets screwed up and must hang there forever ...
Has anybody ever tried to define these "event counter" tests unambiguously, eg having the set flag defined as 1, the reset flag defined as -1 and test for positive/negative values in the relevant "if" statements ?
Re: Bug reports for new RC/RR compilation
Morfans - um could be
(I really couldn't comment, I understand very little about scripting code.)
hopefully PB will see that and let us know if it's the problem.
Re: Bug reports for new RC/RR compilation
Quote:
Originally Posted by
Rozanov
Morfans - um could be
(I really couldn't comment, I understand very little about scripting code.)
hopefully PB will see that and let us know if it's the problem.
Looking into the code, it seems the counters that trigger siege events are generally not initialized.
For instance, at a given turn, a monitor is started that checks if Sofia is a rebel town and, if it is, generates certain events, among which the siege of the town. At the end, the sofia_rebel counter is set to 1. However, if Sofia is a rebel town, sofia_rebel is left undefined.
If the player saves his/her game at this turn and Sofia is slave, sofia_rebel is not stored anywhere.
Upon retrieving the data from the saved file, sofia_rebel is unknown and when it is tested on subsequent turns to see whether it's time to attack Sofia, the engine hangs.
That would imply the bug is only present if one of the settlements that are supposed to be sieged is a slave town at the start of the relevant monitor - something I cannot check, because I am restarting with SS RR after many months ...
Should that be the case, the fix would be simply to add initial counter assignments:
monitor_event FactionTurnStart FactionType Slave
and I_TurnNumber ...
set_event_counter sofia_rebel 0
set_event_counter bulgarians_rebel 0
if not I_SettlementOwner Sofia = Slave
spawn_army
etc
Re: Bug reports for new RC/RR compilation
I've got a problem with textures:
I can run custom battles, but when I do certain icons (namely the current orders such as firing, reloading, etc. and the fatigue/morale symbols) appear as white boxes. Additionally, the entire Campaign UI is messed up in a similar fashion - all white boxes and whatnot.
It shouldn't be my video card as vanilla M2TW worked fine, so I'm assuming that the mod changed something that is now causing this problem.
Just as a side note that the actual models of the units, the unit cards and the maps all look fine. It's just the interface.
Re: Bug reports for new RC/RR compilation
Quote:
Originally Posted by
Morfans
Looking into the code, it seems the counters that trigger siege events are generally not initialized.
For instance, at a given turn, a monitor is started that checks if Sofia is a rebel town and, if it is, generates certain events, among which the siege of the town. At the end, the sofia_rebel counter is set to 1. However, if Sofia is a rebel town, sofia_rebel is left undefined.
If the player saves his/her game at this turn and Sofia is slave, sofia_rebel is not stored anywhere.
Upon retrieving the data from the saved file, sofia_rebel is unknown and when it is tested on subsequent turns to see whether it's time to attack Sofia, the engine hangs.
That would imply the bug is only present if one of the settlements that are supposed to be sieged is a slave town at the start of the relevant monitor - something I cannot check, because I am restarting with SS RR after many months ...
Should that be the case, the fix would be simply to add initial counter assignments:
monitor_event FactionTurnStart FactionType Slave
and I_TurnNumber ...
set_event_counter sofia_rebel 0
set_event_counter bulgarians_rebel 0
if not I_SettlementOwner Sofia = Slave
spawn_army
etc
Rechecked: emergent factions (TO, Mongols) are scripted with all counters properly declared and initialized - only the "national uprisings" have undeclared siege_settlement counters.
Do we know when siege_settlement issues happen ? Only on Bulgarian/Wallace/4thCrusade events ?
Re: Bug reports for new RC/RR compilation
Morfans - definitely the Bulgarian uprising and venetian crusade, but also the TO emergence.
I don't think the Mongols are problem - but then I've only ever played up to there once!
all the event counters etc should be declared at the start of the script and set to 0 I would have thought.
Re: Bug reports for new RC/RR compilation
Quote:
Originally Posted by
Rozanov
Morfans - definitely the Bulgarian uprising and venetian crusade, but also the TO emergence.
I don't think the Mongols are problem - but then I've only ever played up to there once!
all the event counters etc should be declared at the start of the script and set to 0 I would have thought.
This is the list of counter declarations in the script:
Line 5783: declare_counter wallace_spawned
Line 6345: declare_counter attack_palanga
Line 6368: declare_counter attack_riga
Line 7292: declare_counter attack_urgench
Line 7316: declare_counter attack_khiva
Line 7339: declare_counter attack_konjikala
Line 7362: declare_counter timurids_konjikala
Line 7363: declare_counter timurids_yazd
Line 7364: declare_counter timurids_kerman
So actually no sofia/scopia/adrianople_rebel counters are declared.
The TO code seems ok under this respect, so there must be some other bug ...
Re: Bug reports for new RC/RR compilation
well, I don't know if it's the right thread for this but in EDU we see two mongol units with same stats (Disamounted Turhagut and Dismounted Turhagut).
I've already put another unit in that spot but would like to confirm. It's a duplicated unit, right ? This compilation reached the max units number ?
Re: Bug reports for new RC/RR compilation
C: looks like a bug - mistype.
you'll need to check other files to see which one is actually used in game.