AND2 and SVN Bug Reports - A New Dawn 2 ONLY

[Fozman]

Ran some tests on the game having this issue at a few different save points. Quickly played through about 10 turns to let the Asset changes settle.

Tested with 3 values:

• 4 * default (10,000)
• 10 * default (25,000)
• 0.25 * default (625)

At all 3 values, the "one-and-out" phenomenon occurs. However, at the higher values, Corporation movement is significantly reduced. So, the one turn in/out only occurs once or twice every 5-10 turns instead of constantly. This has the drawback of making Corporations hardly spread at all. With the lower value, the Corporation spam is relentless -- 8 to 10 cities every single turn. (2 Guilds in a 30 city empire.)

I looked around for other mentions of this code, and found a forum post from 2011. The code is slightly different as this was back in the RAND 1.76 days, but I agree with the analysis in the post:

As far as I can tell, this grabs the influence of an individual corporation presence in a city (from elsewhere in the file), compares it to the average, and if that particular corporation is lacking it's chucked into a random number generator. Rolling badly on the generator purges the corp from the city. Because it's compared to the average, generally about half the corporations are at risk of being purged at any given moment.

If this was still something located in the Python, I'd adjust the math myself. However, I'm not going to invest the time in setting up tools to edit the .dll. The two things I would try would be:

• Adjusting the way iRand interacts with iDiff via arithmetic. (i.e. reduce the "at-risk" purges from ~50% to the lowest quartile or even the bottom 10%)
• Adding another check with an if-statement to disallow a purge if the company has been in town for less than "x" turns. I realize this would possibly require a new variable and all the work associated with setting that up, but this would probably be the most effective solution.

I imagine that the Corporation spread/decay works similar to Religion, so perhaps there would be some insight in that code. Regardless, if this continues to happen in my games, I'm changing Corporations to manual and that will be the end of it.

 

[Qwarter]

Sorry for asking stupid questions, but how to install the snapshot from sourceforge?

 

[45°38'N-13°47'E]

Ran some tests on the game having this issue at a few different save points. Quickly played through about 10 turns to let the Asset changes settle.

Tested with 3 values:

• 4 * default (10,000)
• 10 * default (25,000)
• 0.25 * default (625)

At all 3 values, the "one-and-out" phenomenon occurs. However, at the higher values, Corporation movement is significantly reduced. So, the one turn in/out only occurs once or twice every 5-10 turns instead of constantly. This has the drawback of making Corporations hardly spread at all. With the lower value, the Corporation spam is relentless -- 8 to 10 cities every single turn. (2 Guilds in a 30 city empire.)

I looked around for other mentions of this code, and found a forum post from 2011. The code is slightly different as this was back in the RAND 1.76 days, but I agree with the analysis in the post:


If this was still something located in the Python, I'd adjust the math myself. However, I'm not going to invest the time in setting up tools to edit the .dll. The two things I would try would be:

• Adjusting the way iRand interacts with iDiff via arithmetic. (i.e. reduce the "at-risk" purges from ~50% to the lowest quartile or even the bottom 10%)
• Adding another check with an if-statement to disallow a purge if the company has been in town for less than "x" turns. I realize this would possibly require a new variable and all the work associated with setting that up, but this would probably be the most effective solution.

I imagine that the Corporation spread/decay works similar to Religion, so perhaps there would be some insight in that code. Regardless, if this continues to happen in my games, I'm changing Corporations to manual and that will be the end of it.

Before embarking on a mission to rewrite that part of the code, I would like to understand why this is happening in your game while it's never happened in any of my game since years (nor I had other reports on this problem). Did you change any (even apparently not related) xml file in any way? Which gamespeed/mapsize are you using? How many of your cities are experiencing this behaviour and how many cities do you have? Can you post a savegame? Thanks.

 

[45°38'N-13°47'E]

Sorry for asking stupid questions, but how to install the snapshot from sourceforge?

It's not a stupid question, you can read instructions here halfway through the page: http://anewdawn.sourceforge.net/pages/install/#developper-installation-svnenglish

 

[Ag0range]

I am sorry, has any method been found to be able to update from 1056 to 1058? Just wishing to taste all new changes without building up all the SVN

 

[Fozman]

Before embarking on a mission to rewrite that part of the code, I would like to understand why this is happening in your game while it's never happened in any of my game since years (nor I had other reports on this problem). Did you change any (even apparently not related) xml file in any way? Which gamespeed/mapsize are you using? How many of your cities are experiencing this behaviour and how many cities do you have? Can you post a savegame? Thanks.

Agreed, no sense going on a crusade if it was a fluke. Just presenting the ideas for documentation.

I did not change any files before or during the progress of this game, it was played on v1056. Since I've been trying to update, I've been getting the asset change warning when opening these saves. My only thoughts on why it would be occurring is due to transoceanic trade right around the time of Astronomy & Navigation (not all the AI would have it at this point). However, I had a Pangaea game that exhibited similar, but less frequent behavior (which is why it didn't bother me as much).

The savegame attached is at least 25 turns after Navigation. Gamespeed & mapsize are "default" (Marathon & Huge). If you look through the Event Log, the immediate flip-flop happens to about 1-4 cities per turn (out of 34, I think). It happens to different cities each time, but most are in the northeast quadrant where my core is.

 

[45°38'N-13°47'E]

I am sorry, has any method been found to be able to update from 1056 to 1058? Just wishing to taste all new changes without building up all the SVN

Not yet, I've uploaded rev1059 and the problem persists; I'm trying to upload a full installer for rev1059. You'll have to download 1GB again but for the moment I have no other solution.

 

[<Nexus>]

With the offline patcher I have downloaded a 16MB patch file (1057 to 1059) but could not test it yet (I'm at work). 16MB looks a little light for dll changes, right?

 

[45°38'N-13°47'E]

With the offline patcher I have downloaded a 16MB patch file (1057 to 1059) but could not test it yet (I'm at work). 16MB looks a little light for dll changes, right?

No, it should be ok.

 

[<Nexus>]

No, it should be ok.

...and it worked indeed.
However ML was so bugged I didin't know if it was because a failed patching or the programing. Now I read the bug reports and so it was the later.

 

[nionios]

I've experienced CTDs with 1059 in saved game and now experiencing constant CTD with 1061 at the very early turns of a new game.
Is Movement Limit the culprit?

 

[45°38'N-13°47'E]

I've experienced CTDs with 1059 in saved game and now experiencing constant CTD with 1061 at the very early turns of a new game.
Is Movement Limit the culprit?

Almost certainly. Strange, I think everything should be solved in rev1061. As a workaround, try switching off Terrain Damage if you're using it or switch it on if you're not using it. Can you also upload a savegame? Thanks

Edit: I ran some more test with and without Terrain Damage (and switching between them too), and I had no CTD.

 

[rtt4a]

No CTDs for me either.. but closing game to desktop has caused error messages few times. I think its related to AND checker somehow because I get no errors when lauching game via checker.

 

[IPEX-731BA5DD06]

SVN 1054 - Demand to ceding leadership on single leader civ.

Small bug here, when you cede leadership, on a single leader civ, Mayans, Native Americans, Sumerian's, there is no other leader available to take over. So you get no change, no running by AI nothing, just -2 "you refused to grant our request" if you at 1st don't accept change in leadership.

Solution - Disable this event for single leader civs, or enable the AI to be the leader inserted.

*Edit*, in addition, I can't build a palisade *edit* and walls. Every time I try to finish it form 1 turn, it locks the game up, either its been accidentally removed as a build for the Mayan's, or its not build-able at all?? Seems to have fixed its self, but 4th time playing save?? Maybe a tech I researched?? Ironworking

 

[nionios]

Almost certainly. Strange, I think everything should be solved in rev1061. As a workaround, try switching off Terrain Damage if you're using it or switch it on if you're not using it. Can you also upload a savegame? Thanks

Edit: I ran some more test with and without Terrain Damage (and switching between them too), and I had no CTD.

CTD with Terrain Damage either on or off.
I'm going to try full installer of rev1061.

 

[jsk119]

The workers under my capital Babylon don't have access to watermill on some tiles, even though they should cause it's a river. Also one of my workers is working on a watermill and my other workers cannot help him the watermill icon isn't there.

 

[45°38'N-13°47'E]

The workers under my capital Babylon don't have access to watermill on some tiles, even though they should cause it's a river. Also one of my workers is working on a watermill and my other workers cannot help him the watermill icon isn't there.

I know it happens sometimes, but I still don't know why. Sometimes it happens with other improvements too.

 

[Vokarya]

I know it happens sometimes, but I still don't know why. Sometimes it happens with other improvements too.

Might it be something with the <bRequiresRiverSide> tag? I thought it had something to do with you not being able to build Watermills on both sides of the same river. It's a different tag from <bRiverSideMakesValid>. Both are turned on for Watermill but not turned on for anything else.

 

[AntiochFarm]

I have a problem with the Launcher (and2_checker.exe) itself. I am using Civilization IV Complete, the rand_full_installer_2017-04_rev1055.exe which I downloaded from the web site, and launcher version 1.23.

Two issues:

1. I used the default installation folder of My Games/Beyond the Sword/Mods as recommended. Since then when I start up the shortcut, I get this error:

The mod installation was detected in double. This can cause unexpected
problems. Please check for the mod folder in these two locations and delete the
wrong one:
- My Documents/My Games/Beyond the Sword/Mods
- Game folder installation/Beyond the Sword/Mods

I think the Launcher is checking the same folder twice using different pathing procedures. The mod is not installed anywhere else. I uninstalled and then reinstalled in a different folder, which eliminated the problem.

2. I installed version 1055. The program alerts me that there is an update available. The program goes through the motions to update to 1061. Then it returns back to the and2_checker program and it is still 1055. There are no messages that the upgrade failed and, for that matter, I can see files being created in the folder temporarily, but it never changes from version 1055. The Launcher keeps encouraging me to update every time.

 

[nionios]

CTD again in new game with full installer 1061.