Public beta patch 1.14 rc21 (July 8, 2009)

[Gray_Lensman]

deleted

[Gray_Lensman]

deleted

[Gray_Lensman]

deleted

[Clovis]

Similar thoughts here, but there's also the possibility of linking the replacement process to the successful completion of the "1861 Threaten Richmond" event. Almost as a precursor however, in order to satisfy the "1861 Threaten Richmond" event, he (McDowell), would have had to already won at Manassas. I just can't picture McDowell satisfying the "1861 Threaten Richmond" event without winning at Manassas first, unless (gamewise) it was against a totally incompetent AI.

I was able to piece together your earlier contribution allowing McClellan to be placed wherever McDowell is located. Nifty and again much appreciated. Thanks. As a returned favor, you might want to check you SVF code to ensure the removal of the artillery element that also accompanies the Northeastern Virginia Army HQ unit. Testing revealed this element remained with McDowell after the HQ element was removed. This might possibly lead to the "rebuilding" of the HQ unit since only one element was removed.

After using the database/compile method, this is what I finally came up with that works: (Note that I added a few more conditionals to ensure the presence of all the necessary units prior to the event action.) (Also, note that for some reason I had to remove McDowell to remove his HQ, then reinstate him) ...Not exactly sure why yet.



For now, I'm not posting this reworked update until it's decided what the linkage should be and condideration of this being AI only. I don't particularly like the McClellan event firing and then McClellan staying in Cincinnati... Seems rather incomplete without additional rework.

Another item for those reading this...
Lyon/Price event timing overlap error for the 1861 July Scenarios has been fixed.

Thanks . I will investigate.

[enf91]

For the save game with the weird time requirement for blockade ships entering port, do you want the .trn or the .ord?

Also, you never answered my question about how a general can cause a strength penalty if his ratings cause only bonuses.

[hattrick]

I'm unable to reproduce the crash. This could be for several reasons.

1.) An incorrect installation... Did you install th v1.13b comprehensive patch prior to this public beta patch.

2.) Possibly already fixed in the latest update work that I've done.

3.) Different settings or not enough save files. (might need to go back another turn).

Hi Gray,

1.) Yes, its a correct installation.

2.) Yes, its very possible and most likely its already fixed. I do have a question though, can you tell from the logs if when the AI makes its decisions? Is it possible that when it changes those decisions from rerunning turns that it could effect turns from crashing, when previously they did? I'm just curious, always thought about this.

3.) Regarding the settings, I removed that installation and reinstalled with another patch, I just saved the saves. If you think they are important enough, I could try to remember.

Here are some earlier turns, just in case.

[Spharv2]

Getting a hard crash in Aug '63 (I think) of this game, can't tell for sure what the issue is, but I've tried three times to restart, and it crashes when processing each time.

[enf91]

Something really, really weird. As soon as I reached Feb 1862, the thing with ships needing 15 years to enter port disappeared. However, weather somehow affects their ability to enter port.

Also, when an error pops up, it is followed by something that says "the confirmation" and "yes" "no" "ignore" options because I activated the option that sends errors to you guys. Is it "yes" that does that?

[Gray_Lensman]

deleted

[Gray_Lensman]

deleted

[richfed]

This will be a code problem for Pocus to look into if it wasn't there in earlier versions. Thanks for pointing it out. To attempt to fix it in the database would require every single scenario's database to be altered for every single water region and then accompanied by every scenario being recompiled with no guarantee that it would fix the problem (and remain fixed if Pocus reworked the code after the fact.) The game code is supposed to compensate for this internally.

This seems to have happened once before --- I really don't recall the exact cirmstances, but maybe around the time the ROE buttons were instituted. Pocus did fix it at that time ...

[Gray_Lensman]

I received the exact same error last night while testing my mod. 1.13e-RC5.

Please post a saved game when you get this message, otherwise Pocus cannot reproduce the problem and he usually moves on to other items more immediate. To be honest, he's working on several things at once including 2 new games and without something to streamline the process, these posts are not very helpful.

[Gray_Lensman]

Pocus:

I think we can disregard this particular bug report unless/until it's confirmed via a saved game with unaltered files.

I don't think Pocus would want to debug my saved game. I have many many modifications to the units, models, regions and scenario with my mod. It would take some time to setup the environment. If Pocus wishes to debug this I'll be happy to put together the environment. I just wanted to confirm Clovis's post of the issue.

10:13:48 AM [Critical ] TFactions.GetFactionByUID The faction of UID -1 was not found
10:13:48 AM (Reporting) Host MakeStep
10:13:48 AM [Error ] THost.MakeStep Exception caught: Violation d'accès à l'adresse 005C87DF dans le module 'AACW.exe'. Lecture de l'adresse 00000000 DebugStep: Replay Creation
10:13:48 AM (Reporting) Host MakeStep
10:13:48 AM [Error ] THost.MakeStep Exception caught: Violation d'accès à l'adresse 005C87DF dans le module 'AACW.exe'. Lecture de l'adresse 00000000 DebugStep: Replay Creation
10:13:48 AM (Reporting) Host MakeStep
10:13:48 AM [Error ] THost.MakeStep Exception caught: Violation d'accès à l'adresse 005C87DF dans le module 'AACW.exe'. Lecture de l'adresse 00000000 DebugStep: Replay Creation
10:13:48 AM [Error ] TAIEngine.ApplyAgenda Exception caught: Violation d'accès à l'adresse 0066364A dans le module 'AACW.exe'. Lecture de l'adresse 00000008 LogStep: Summary1003416 43. Confederate Detachment
10:13:48 AM [Error ] THost.LoadHostFile One or more error while generating AI files. Check the logs

edit> From the AIlog
10:13:48 AM (Reporting)
10:13:48 AM (Reporting) List of Agendas for 1003416 43. Confederate Detachment, 1102 Ft Monroe, VA, Pow:71, Count: 4(l 0/L 4/S 0)
10:13:48 AM (Reporting) Agenda 1 : Agenda: Intercept (Master), Value: 13200, Region: 1102 Ft Monroe, VA
10:13:48 AM (Reporting) Choosen Agenda: Agenda: Intercept (Master), Value: 13200, Region: 1102 Ft Monroe, VA

Shovelhead:

Are we talking about bugs found with the new v1.13e Public Beta patch and the files exclusively associated with it or are we talking about bugs found with a different modded files that just happen to be using the v1.13e executable? If it is indeed the latter topic, modded bugs should be discussed in the AACW Mod forums to keep from confusing the issue for Pocus to work with, otherwise we'll never get this thing sorted out for him to work with.

The whole point of a Public Beta patch is to post the new game version for gamers to voluntarily test for flaws and report on those flaws prior to it's being released as an "official" game update patch, not to provide a new test bed for Mods. (At least not until the patch itself is stable and released, at that point it is fair game to Mod and dissect to everyone's content).

edit> I should have posted something about bug reports and posts/comments added to this thread as having to be pertinent to this specific unaltered game patch file testing, so this is more my fault than anyone's. But for the future Shovelhead, when we start threads regarding a new Public Beta patch, we would like for all the voluntary posts contained therein to be related to testing that new specific beta patch with the unaltered files. That way Pocus does not have to sort thru all sorts of extraneous stuff that he cannot possible reproduce to fix the public beta patch itself. Sorry for the lack of explanation in the first post of the thread.

Note> The bolded text at the top of this post is only for Pocus informational purposes and is not meant to mean anything derogatory regarding the post itself.

[Gray_Lensman]

deleted

[Clovis]

Thanks for the feedback and the saved game files. I was finally able to confirm this with unaltered (RC5d) Public beta patch files.

edit> This is indeed confirmed with the latest v1.13e AACW.exe and again with unaltered files. Darn. We'll have to wait for Pocus to squash it or at least store it away somewhere with the rest of his nuts.

Pocus:

I posted the confirmation of this bug along with saved game and log files at Post #4 at this link: http://www.ageod-forum.com/showthread.php?t=14004

So indeed, without my post, now moved to the modding forum this bug posted for someone playing the unaltered version, would have been unoticed by you and the next patch released with this bug. Great result.

[enf91]

Not really a problem, just CSA AI quirks mostly from the April 1861 campaign.

1. The CSA has almost never defended Manassas in all the times I have played it. Maybe once back a few versions ago, but that was it. Instead, it keeps its Army of the Potomac by Richmond and sends some other force to attack Fort Monroe. Sometimes it uses Joe Johnston to attack Pittsburgh, where (under the system from the last version under which loyalty influenced military control) the entire force would be destroyed.

2. No matter what the overall situation is, the CSA regularly attacks/besieges Fort Pickens in force. I could have forces overrunning central and western Tennessee while this stack with some 300-400 strength points sits in western Florida. I suppose it could be there to defend against possible amphibious assaults that other players might launch, but the stack sits there for months on end doing absolutely nothing.

3. I saw once a CSA stack (1/16/1862 or 2/1) that had about 9 generals and 1 division. No other combat troops were in the stack.

[Gray_Lensman]

deleted

[Clovis]

Not really a problem, just CSA AI quirks mostly from the April 1861 campaign.

1. The CSA has almost never defended Manassas in all the times I have played it. Maybe once back a few versions ago, but that was it. Instead, it keeps its Army of the Potomac by Richmond and sends some other force to attack Fort Monroe. Sometimes it uses Joe Johnston to attack Pittsburgh, where (under the system from the last version under which loyalty influenced military control) the entire force would be destroyed.

2. No matter what the overall situation is, the CSA regularly attacks/besieges Fort Pickens in force. I could have forces overrunning central and western Tennessee while this stack with some 300-400 strength points sits in western Florida. I suppose it could be there to defend against possible amphibious assaults that other players might launch, but the stack sits there for months on end doing absolutely nothing.

3. I saw once a CSA stack (1/16/1862 or 2/1) that had about 9 generals and 1 division. No other combat troops were in the stack.

enf91,

The AI question is a complex one; as is, AACW AI is almost brilliant. What it lacks is the capacity to adapt his aggressivity and choose its objectives. But many variables are presiding the AI objective choices and the AI engine is made to choose different solutions in face of a situation.

There are some commands to module this. I've used them in the mod whose mention is prohibited here and some could think it would suffice to put the same events in the official. It will not because one of the main AI parameters is the balance of force. What i've done is so tied with those I made in the balance of forces porting them to another situation would give more curious results than interesting ones.

There are many other points. AI is fond of militias. If you're altering in any way militias without coping with AI reaction you will weaken the AI beyond any measure. Other example: heavy artillery. Currently, AI isn't warned adding a slow artillery unit to one stack will deprive it of any fast move...

Improving AI is possible and I don't doubt Pocus will improve it yet more. In any case, considering the effort needed to adapt my own work to the official game and the way the so called coordinator is acting with modder, the reply is for now no. If someone else wants to adapt my own work to the official version , I will eventually provide him some help. The task will be a long one. I estimate about 200 hundred hours to get a satisfying result for a modder. From what I knwow, Pocus needs a lot of time too to improve AI and the log files are jsust a testimony how hard he has worked on this.

[Rafiki]

[color="Blue"]clovis, this needs to stop. You seem to be having some badly hurt feelings and are lashing out, in particular at Gray Lensman, and it's becoming *really* counter-productive. I'm loosing sight of your actual grievances in all the bru-haha and grandstanding [/color]

[GraniteStater]

@Gray Lensman:

My apologies for not sending you any files thus far. From what I've read above, it seems to reflect my most recent experiences with (e).

To recap:

* Got to Oct 61 (US) and got McDowell torched in Richmond. Units trapped with no retreat. When this turn is loaded and told to process the next turn, a Critical Error message is displayed. Clicking OK (i. e., performing only user action possible) results in a bit of further processing and then a CTD after reaching the Supply Segment display. -- (e) version of about a week ago.

* Spurious Union units in/at Ft. Sumter in May or June 61 well after its fall. Same (e).

" CSA commanders from Outer Space in upstate NY; no real units with them, AFAIK - (d).

On the whole, the beta is a good direction. Having the AI not waste its efforts as the CSA in pointless pinpricks (a serious thrust is s. t. else) is a Good Idea; as the player, I can adapt and commision an anti-Raiding effort and have done so, but not having two regiments of Sioux warriors take a harbor in Illinois is a Good Thing. By the time late '62 is rolling around, I actually start to welcome these fruitless endeavors, because by that time, the Union has enough men to wheel around and crush invasions of almost any size.

I'll try to take the time, bite the bullet, and send you some files.

BTW, I'm running on a Vista laptop (3 GB RAM - plenty of muscle) and 1.13b is extremely stable.

[Gray_Lensman]

deleted

[ShovelHead]

For the foreseeable future is version 1.13b the official latest version of AACW?

[oldspec4]

Stress levels elevated. Time for a break. This particular Public Beta patch is no longer supported by me. Changes will be forthcoming in a later data patch (possibly middle June). If the developers are still busy then, we'll postpone it until such time as RoP and VgN have been released, which could be even later in the year.

Near term, I'll be issuing some QuickFixes for common errors contained in the v1.13b "official" patch.

Sorry for any inconvenience.

Thanks for the continued QuickFix support

[Gray_Lensman]

deleted

[tc237]

Thanks for all the hard work Gray, it is very much appreciated.

[enf91]

For some perplexing reason, during the April 1861 w/KY campaign, the mousewheel stops working in Late May 1861 and starts working again a few turns later.

edit: And it worked again right after I wrote this (alt-TAB). Weird.

[Comtedemeighan]

Its probably just your computer and not the game

[Gray_Lensman]

deleted

[enf91]

No, that's the thing. The mousewheel works perfectly in every situation. It's just with this game on that turn.

[Gray_Lensman]

deleted