Reinforcement CTD Fix (Please Test)

[Kull]

We are pretty sure that the Reinforcement CTD has been solved. The "final fix" has yet to be achieved, but we are 99% certain that it's related to the awarding of Post-battle traits. The attached file contains a revised export_descr_character_traits files. After unzipping, please drop it into the Data/ subdirectory of your v.74 EB game folder.

1) The good news is that you can use this file immediately, even in the middle of an ongoing campaign. In fact, if you had a prior campaign that could not be completed due the Reinforcement CTD, you can now resurrect it.

2) The bad news is we took a "brute force" approach and have deactivated all post-battle traits. Ultimately we'd like to narrow the problem down and reinstate as many of these as possible, but given your frustration with this bug, it was felt that you'd accept the current fix as the lesser of two evils.

What we would like from you: Confirmation. If you have save games that always result in a Reinforcement CTD, please try them again after installing the new traits file - and report back here with the results.

Thank you for your patience and support.

ftp://europabarbarorum.org/ReinforcementCtdTraits.rar

NOTE: The zip contains a second file called "BACKUP". As you might suspect, it's the original export_descr_character_traits file, which can be used to restore the v.74 installation should you so desire. Please extend your thanks to therother who created the new Traits file and who (along with MANY others) has assisted in tracking down this most difficult of bugs.

[Ecthelion]

I've uninstalled RTW already so unfortunately I can't test this. But when I was fooling around with this bug, I took a similar approach and it avoided the CTD. The only downside being that the general gets no attributes as a result of battles.

[Turkeys!]

i tryed it and it works perfectly no ctds at all

[bovi]

Seems to work, although I miss the triggers... I've got a relatively simple battle saved which consistently CTDs with all triggers enabled, so it doesn't take all that long to play it through. I'm currently uncommenting chunks of the triggers from the attached traits file and replaying the battle to see if I can find the guilty one (or at least its general vicinity) for my particular CTD, will post back when/if I'm successful.

[bovi]

Apparently one or more of the hate_n_fear triggers are part of the trouble for that battle (Pyrrhos gaining control over Macedon's Pella, with small amounts of defenders but a few reinforcements). When enabling chunks of triggers originally enabled in 0.74 from top to bottom in the attached traits file, the battle starts to crash once these traits are enabled. However, having all the triggers of 0.74 enabled except the hate_n_fear triggers also causes a crash, so there seems to be no single guilty trait. I'm not sure this information should be in the current thread as it may make it cluttered, feel free to point me to another place more suitable for further results of my investigation.

[Leão magno]

seens solved to me!!!

[bovi]

I concede defeat at seeking out the trigger(s) in question . Whether it CTDs varies even though the outcome of the fight is nearly identical, and any number of triggers seem to cause the crash. Without debugging tools it seems impossible to find the errors . I only hope that you will succeed where I have failed, so the 0.8 will be stable and with the most possible content!

[LorDBulA]

Could you provide list of triggers you founed that coused CTD?

[bovi]

.

I thought I had found one within the hate_n_fear set, as I played the given battle without crash until I enabled those. However, playing the battle again with the same traits file didn't cause a crash every time, even though the battle had pretty much the same result, varying between clear and heroic victory. It crashed often though. I also replayed the battle a couple of times with the "last known good" traits file, which unfortunately revealed itself to CTD too! I *think* the frequency was less without that group of triggers enabled.

The random occurence of the CTD makes it hard to pin down specific triggers, and I'm pretty sure that's why it hasn't been fixed yet by you obviously competent makers of EB. I have a suspicion it is because some trigger reacts twice on the same battle. Most triggers have a chance to yield an effect, and I've read somewhere else on this forum that there's a bug causing post battle triggers to run once per enemy army involved. The more triggers that are enabled, the greater chance that one of them gets its effect more than once. Given that there's no log that I'm aware of, of which triggers are being run and effected, I'd given up on testing them...

But I thought of another possible test just now. I'll reduce the chance of every post-battle trigger to 0 and see if it consistently does or does not CTD :hmmm:. It could possibly strengthen my theory of double effect causing the crash, or at the very least kill it. Then, if it runs fine I'll set the chance to 100 for groups of triggers, like I enabled them previously, and see where it screws up.

[BozosLiveHere]

.
But I thought of another possible test just now. I'll reduce the chance of every post-battle trigger to 0 and see if it consistently does or does not CTD :hmmm:. It could possibly strengthen my theory of double effect causing the crash, or at the very least kill it. Then, if it runs fine I'll set the chance to 100 for groups of triggers, like I enabled them previously, and see where it screws up.

I don't think you can set Chance to 0, the inferior limit is 1. But it is a nice idea actually, thanks for the help.

About the double effect, it certainly can't happen in .74: RTW 1.2 had a bug where if you manually played battles, the PostBattle triggers would be tested twice. To fix it, we created the Bug and Fixed traits, that were placed at the end of the trigger section. As the engine did it's first pass through the triggers, the Bug trait was awarded and then, because all PostBattle triggers had the and not Trait Bug = 1, the traits wouldn't be awarded a second time.

As TA said, what's really puzzling me in our latest build is the fact that autoresolve always avoids the CTD. As you know, there aren't any traits that are awarded only if you manually play your battles.

[Teleklos Archelaou]

The thing that BozosLiveHere has been most puzzled wth lately is the fact that the CTD never happens with an autoresolve though, and why the triggers all seem to work when you do autoresolve, but not when you fight it out. There doesn't seem to be anything different though - it's very irritating to say the least. There has to be a process we are unaware of that happens in one and not the other. Could it be something to do with the way the results are presented? Any variance there? Anything else?

[bovi]

Are you sure that the execution order of the triggers is guaranteed to follow the order in the traits file? Then my test doesn't give any useful information anyway, so I'm cancelling it.

If the order is guaranteed, it appears to me to be another bug: Triggers may affect too few battles' results, among them only one defending battle each turn. I'm not certain how the antitraits work, so I may be wrong about this though...

My case: The "Bug" trait is always awarded at the end of any post battle execution. It is reset by the "Fix" trait when

• Post battle, after all triggers have failed their conditions. Immediately reversed by the Bug trait in the next trigger, so the Bug trait persists into the next trigger run *and* the next fight (unless not, discussed further below).
• "CharacterTurnStart" - I suppose this means the character is selected in the campaign map?
• "CharacterTurnEnd" - I suppose this means the character is deselected in the campaign map?

If my assumption above is correct, the player will have to deselect the army before continuing to the next fight in order to have triggers work in the next fight. Consecutive defending fights will give no such opportunity.

A correct fix trigger would be to reset the Bug trait *before* a battle commences. I'm not sure if this is possible, but I've seen a PreBattleWithdrawal designator in the file for WhenToTest, so perhaps there is one for pre-autoresolve and pre-battlefield as well?

I see the post-battle Fix trigger affects the Fix trait by 2. This strikes me as odd, 1 should be enough... Does this mean the Bug award is reduced by 1, to 0, and then Fix is set to 1? This would cause triggers to run again the third time around, wouldn't it, as the Bug trigger only reduces the Fix trait to 0 rather than increase the Bug trait to 1? If traits awarded always kill existing antitraits and then have full effect, the 2 has no real effect and should rather have the value 1 like in the characterturn fix-triggers; 1 Bug -> 2 Fix -> 1 Bug. But if the effect is 1 Bug -> 1 Fix -> 0 Bug, then it should in theory work correctly if there are 1 or 2 enemy armies, but no more than that.

Of course, my test battle only involves 2 enemy armies, so there must be more to it than this potential discovery. And the crashes are a lot more important to fix than ensuring the triggers are not being bypassed in a few battles...

I have no idea why autoresolve should work better. There are a huge lot of traits I've never seen on characters in my games, as well as many hidden ones, and I take your word that every trait can be awarded either way. The processing must be different based on type of resolve though, as there is the bug of multiple passes through the triggers when playing on the battlefield and not using autoresolve. Something affecting the post battle traits. I just wish CA would include debug log functionality in their moddable content... This would be infinitely easier to grapple then.

[bovi]

I practically never use autoresolve either, BTW, it's too unpredictable.

[bovi]

Okay, now I'm pretty sure I've found one... I've upped the chance of every post battle trigger to 100%, and it behaves a lot more consistently now. The EB036_Large_Battle triggers seem to always crash in my battle. As the only trait affected is the battlepoints and there are plenty other triggers that affect it, I suppose the crash comes from some of the conditions. Both NumEnemiesInBattle and UnitSize is used for the first time in this set of triggers, and I've enabled them in a linear fashion... As the NumEnemiesInBattle variable isn't used in any of the vanilla triggers, a potential CTD bug while calculating this number from multiple armies may have slipped by CA's attention. I don't understand why the chance of effect should make any difference though, unless this is used before the actual conditions .

http://rapidshare.com/files/1001883/ebtest.rar has the relevant files (savegame, last working traits file and the crashing traits file) if you all would like to help out testing... It's very repetitive and frankly boring work so I haven't been making much progress each time . There's also a small java class I've made to ease the commenting/uncommenting of the traits file. Use it if you know how, I won't be documenting or explaining it if you don't. Note that the (un)commentBattleTriggersWithPattern functions appear to have a bug, the game crashed upon loading after performing my last edit using them.

[Foot]

Great work, bovi. I'll make sure our guys hear about this stuff.

Foot

[bovi]

Good news! Try http://rapidshare.com/files/1043528/...ter_traits.rar out. As I disabled the EB036-triggers (and ironed out a few bugs in my java utility, which left a few non-code lines uncommented), every other trigger seems to work, at least in my particular battle. Note that the chance is set to 100% for every post-battle trigger, so it's not the real deal, but it's a lot easier to find out whether it crashes or not with this one. If your game still crashes, feel free to post the savegame here, and I may take a look at it when I get the chance.

What's weird is, both the conditions are used in other triggers, which seem to work fine .

[bovi]

Must have done something wrong that time so it worked. Tried it again, and it's crashed three times straight

[Teleklos Archelaou]

Even if you can't seriously do any betatesting bovi, I think it would be helpful for us if were brought onto the beta test team so that any monkeying around with the code or problems you do it would be with the newest build.

[BozosLiveHere]

Ok, thanks to bovi's brilliant find, I now present you a version of export_descr_character_traits that hopefully isn't plagued by the Reinforcement CTD. The good part is that it's fully functional, so the worse that can happen if you use this new version is that you don't see any improvement. Here it is:

http://rapidshare.com/files/1047634/...ts.rar.html%22

[Shigawire]

Thank you for your excellent work Citizen Bovi!

All hail BozosLiveHere, LordBula, Bovi and Kull!