Single Player bugs and crashes v38 plus (SVN) - After the 20th of February 2018

[george585]

I've tried hacks to the mod name before and they didn't work to do anything but crash the game on load. Toffer said he had the same problem with that as well so I didn't try TOO hard to sort it out further when that's what happened with my efforts.

I think you can fix it easily by making a directory junction to the mod folder:

cmd.exe
cd <Path to Beyond the Sword\Mods>
mklink /j <Different name of the mod (C2C in my case)> Caveman2Cosmos

This will create a new "folder" (junction) to the existing folder (Caveman2Cosmos mod folder). To the game engione it will appear transparently as a differently-named mod. Hope it helps.

More on mklink: https://docs.microsoft.com/en-us/windows-server/administration/windows-commands/mklink

 

[Yudishtira]

(10528) The disappearing animals are mostly fixed, but I still regularly get the sitch that you have to end the hunter's move before the animals appear. He doesn't have to be out of moves, space to skip turn works just as well, but I'm too lazy to skip turn every tile, and would appreciate not having to.

 

[Toffer90]

(10528) The disappearing animals are mostly fixed, but I still regularly get the sitch that you have to end the hunter's move before the animals appear. He doesn't have to be out of moves, space to skip turn works just as well, but I'm too lazy to skip turn every tile, and would appreciate not having to.

I'm pretty sure you are imagining this, if you can find an example where doing a recalc make an invisible animal visible, then you have found a bug in the hide and seek system, as the animal should then have been visible before doing the recalc.
I experience that animals become visible the moment I move a unit that can see them close enough to see them.
Remember that all units have stronger ability to see invisible units when the invisible unit is on the same tile as the observing unit. meaning that your hunter may not see the animal that is one tile to the left, but the moment you move the hunter to the left the animal becomes visible, the hunter has often used up all its movement points the moment the animal becomes visible because this often happens in forests/jungle/bamboo where animals have high enough camouflage to not be seen from an adjacent tile.

 

[Niveras]

Guaranteed crash at end of turn here. This game was started in SVN 10524, but I updated to 10537 in an attempt to salvage it (as 10526 indicated a divide-by-0 fix that I thought might save it) . Due to the update, it will ask to recalc when you load. You can do so or not as you prefer, result seems the same. Also PPIO 0.6.1.1 (for SVN 10537).

Incidentally, another oddity with the barb civs is that some of them have multiple cities that are highly disconnected. Korea, for example, has their capital Seoul in about the northeast, but also has Pyongyang in the far southwest. Byzantium was a very recently founded civ and started with three cities: I haven't explored their area yet, but given visible claimed territory it appears they might have two in the far southeast and one on the far north west coast. I haven't noticed if other civs have similar oddities (although Arabia has been a minor civ for a very long time). I have seen a python (maybe?) error come up occasionally in the notification text under the research bar. I believe it made some reference to barbarians and/or barb civs, but unfortunately I did not think to write it down and report it directly. I mention it only because it may also be related to, for example, Arabia's lack of progression into a normal civ.

I zipped the minidump as the forum indicated it did "not have an allowable extension."

 

[Yudishtira]

I'm pretty sure you are imagining this

So much for keeping an open mind .

if you can find an example where doing a recalc make an invisible animal visible, then you have found a bug in the hide and seek system, as the animal should then have been visible before doing the recalc.

When did I mention a recalc?

I experience that animals become visible the moment I move a unit that can see them close enough to see them.
Remember that all units have stronger ability to see invisible units when the invisible unit is on the same tile as the observing unit. meaning that your hunter may not see the animal that is one tile to the left, but the moment you move the hunter to the left the animal becomes visible, the hunter has often used up all its movement points the moment the animal becomes visible because this often happens in forests/jungle/bamboo where animals have high enough camouflage to not be seen from an adjacent tile.

All very true, but as I think I said, this happens without moving. You move to a tile, the animals can't be seen. You hit space, the animals show up.

Tbrd fixed a similar problem (animals were disappearing during the hunter's move, when he got adjacent to them) just recently, so let's hear what he has to say.

 

[glaucus.b]

Revision 10537 with latest PPIO modmod: I have built Worldview-Cannibalism but I cannot butcher captives. Do I need another building or something else to butcher captives?

 

[Toffer90]

When did I mention a recalc?

You didn't, I'm saying that a recalc will make sure that animals that you should be able to see are actually seen.
If there is a bug that makes an animal you should see invisible then a recalc will make it visible.

All very true, but as I think I said, this happens without moving. You move to a tile, the animals can't be seen. You hit space, the animals show up.

Probably because the animal comes closer to your hunter, or moves into a plot where it has less invisibility boost from feature/terrain, during the end turn sequence.'

Tbrd fixed a similar problem (animals were disappearing during the hunter's move, when he got adjacent to them) just recently, so let's hear what he has to say.

I remember there was a bug like that not too long ago, and back then I often hit recalc to check if there was any animals that I should see around my hunters.
From what I've experienced lately, the bug is gone now.

 

[MattCA]

Revision 10537 with latest PPIO modmod: I have built Worldview-Cannibalism but I cannot butcher captives. Do I need another building or something else to butcher captives?

that and sacrifice has been broke for a while I believe. I noticed its not on the dll but I thought the point of the outcome missions is so you don't need the c++. is that right?

Here's what u guys were talking about the other day

 

[T-6000]

I've been having a few crash bugs in my current game (mid-ancient-era or so, so well before I'd usually get the barrage of crashes that presumably come from running out of memory later in the game). The crashes in this game have generally been repeatable and the current one is a consistent end-of-turn crash. They've largely seemed to be correlated with Denmark (AI), who I've had a few squabbles with over territory. The crash can be reproduced by loading the saved game and ending the turn.

More information that may or may not be relevant:

• The game was started on SVN revision 10504 and I've been playing on that version. Updating to the current revision doesn't avert the crash.
• I've made one change to the XML, setting the Arsonist unit's combat strength down to 5 (as I've previously found them to be unreasonably disruptive to the gameplay when they come along).
  
• Going back a turn and replaying leads to a crash at the end of this same turn. This still happens if I end the war, though there could still be combat happening since Denmark has recently started using some of the stealth/criminal units.
• The previous crashes related to Denmark cheekily founding a city on the grassland peninsula south of my capital during enforced cease-fire turns; when I went to war to clear it out, the game would reliably crash during my turn when I attacked the city and/or their nearby units. I avoided this by going back to a previous saved game (~12 turns back) and preventing the city from being founded in the first place. In that alternate timeline, the crashes did start at roughly the same turn (though if I recall correctly there were no end-of-turn crashes in that timeline).
  
• Attacking units or cities in the current war has not resulted in any immediate crashes.
  
• Using the worldbuilder to delete various Danish cities also causes semi-consistent crashes (though I don't know if that's just how the worldbuilder is since I normally never use it). Reproduced fairly reliably by erasing the smaller Danish cities; it tends to happen after deleting 1-3 of them though the specific city that causes the crash varies. Notably I can use the worldbuilder to wipe France and the Ottomans off the map entirely without any issues.

 

[Yudishtira]

You didn't, I'm saying that a recalc will make sure that animals that you should be able to see are actually seen.
If there is a bug that makes an animal you should see invisible then a recalc will make it visible.
Probably because the animal comes closer to your hunter, or moves into a plot where it has less invisibility boost from feature/terrain, during the end turn sequence.'
I remember there was a bug like that not too long ago, and back then I often hit recalc to check if there was any animals that I should see around my hunters.
From what I've experienced lately, the bug is gone now.

I don't get what you're problem is. How can the animal have moved between me moving my unit and hitting the space key?

The only recalc I know is ctrl-shift-T, and it takes over 10 minutes. So great advice that I should do this every time my hunter (well every unit really) moves one tile.

 

[Thunderbrd]

I second that. In fact I think I reported that too last year, if I recall correctly. I was supposed to get 2 free techs in one turn (I think one was due to completing a Wonder, the other due to researching a tech), but got only one in then end.

I experienced the problem before anyone had reported it but it's not important enough to deal with yet. It's on the nearly infinite list of low priority fixes since it at least works if you get one and if you know it's an issue, it means try not to time getting two in one round. Easily avoided in play. It's something I will look into when there aren't a lot more critical things to attend to.

I think you can fix it easily by making a directory junction to the mod folder:

cmd.exe
cd <Path to Beyond the Sword\Mods>
mklink /j <Different name of the mod (C2C in my case)> Caveman2Cosmos

This will create a new "folder" (junction) to the existing folder (Caveman2Cosmos mod folder). To the game engione it will appear transparently as a differently-named mod. Hope it helps.

More on mklink: https://docs.microsoft.com/en-us/windows-server/administration/windows-commands/mklink

Yeah, I'm not bothering with all that just because people can't follow the instructions we clearly gave. Sorry.

(10528) The disappearing animals are mostly fixed, but I still regularly get the sitch that you have to end the hunter's move before the animals appear. He doesn't have to be out of moves, space to skip turn works just as well, but I'm too lazy to skip turn every tile, and would appreciate not having to.

I get this at times too. I have tried to figure it out but haven't the slightest clue why this is happening. I think it has something to do with when the graphics engine refreshes.

Guaranteed crash at end of turn here. This game was started in SVN 10524, but I updated to 10537 in an attempt to salvage it (as 10526 indicated a divide-by-0 fix that I thought might save it) . Due to the update, it will ask to recalc when you load. You can do so or not as you prefer, result seems the same. Also PPIO 0.6.1.1 (for SVN 10537).

Incidentally, another oddity with the barb civs is that some of them have multiple cities that are highly disconnected. Korea, for example, has their capital Seoul in about the northeast, but also has Pyongyang in the far southwest. Byzantium was a very recently founded civ and started with three cities: I haven't explored their area yet, but given visible claimed territory it appears they might have two in the far southeast and one on the far north west coast. I haven't noticed if other civs have similar oddities (although Arabia has been a minor civ for a very long time). I have seen a python (maybe?) error come up occasionally in the notification text under the research bar. I believe it made some reference to barbarians and/or barb civs, but unfortunately I did not think to write it down and report it directly. I mention it only because it may also be related to, for example, Arabia's lack of progression into a normal civ.

I zipped the minidump as the forum indicated it did "not have an allowable extension."

I'm stuck working on some things at the moment and will take a look at crashes later - probably next weekend. The barbarian thing is known and came on when we changed things so that the barb nations pick up more than one city when they initialize. It's a python thing the way this is setup.

I've been having a few crash bugs in my current game (mid-ancient-era or so, so well before I'd usually get the barrage of crashes that presumably come from running out of memory later in the game). The crashes in this game have generally been repeatable and the current one is a consistent end-of-turn crash. They've largely seemed to be correlated with Denmark (AI), who I've had a few squabbles with over territory. The crash can be reproduced by loading the saved game and ending the turn.

More information that may or may not be relevant:

• The game was started on SVN revision 10504 and I've been playing on that version. Updating to the current revision doesn't avert the crash.
• I've made one change to the XML, setting the Arsonist unit's combat strength down to 5 (as I've previously found them to be unreasonably disruptive to the gameplay when they come along).
  
• Going back a turn and replaying leads to a crash at the end of this same turn. This still happens if I end the war, though there could still be combat happening since Denmark has recently started using some of the stealth/criminal units.
• The previous crashes related to Denmark cheekily founding a city on the grassland peninsula south of my capital during enforced cease-fire turns; when I went to war to clear it out, the game would reliably crash during my turn when I attacked the city and/or their nearby units. I avoided this by going back to a previous saved game (~12 turns back) and preventing the city from being founded in the first place. In that alternate timeline, the crashes did start at roughly the same turn (though if I recall correctly there were no end-of-turn crashes in that timeline).
  
• Attacking units or cities in the current war has not resulted in any immediate crashes.
  
• Using the worldbuilder to delete various Danish cities also causes semi-consistent crashes (though I don't know if that's just how the worldbuilder is since I normally never use it). Reproduced fairly reliably by erasing the smaller Danish cities; it tends to happen after deleting 1-3 of them though the specific city that causes the crash varies. Notably I can use the worldbuilder to wipe France and the Ottomans off the map entirely without any issues.

See note above about crashes at the moment. In a test I also encountered a graphic one I couldn't get past which I think I've seen before. I believe we have a bad unit under some cultural define. Not saying this is that... just saying art crashes like these are outside of my ability to do anything about and it's one of the few reasons we're getting crashes right now. It'll take someone with a lot of time on their hands to go into worldbuilder with representative players of each of the cultural artstyles (I'm almost certain the Neanderthal is one with a problematic unit) to go in and place every unit in the game until they find the one that crashes when it's seen on the screen.

 

[Johnny Canuck]

If I turn off Alt-Timelines, I get a number of error messages when loading the game, all of which refer to the Pepper2000 module. I am able to click past the error messages and get the game to load.

 

[Yudishtira]

Khmer city name still (10528) coming up "I?varapura".

 

[Yudishtira]

I get this at times too. I have tried to figure it out but haven't the slightest clue why this is happening. I think it has something to do with when the graphics engine refreshes.

Graphics engine sounds about right, except that the fog of war is refreshed every tile (if you move one tile at a time, and without having to end your move), and any "old-school" units that become visible from that refresh, become visible (and that includes old-school invisible units like Rogues and Assault Ships).

When I say the animal is not visible, I mean it is not graphically represented on the map. In some (and perhaps all) cases, it does show up on mouseover of the tile, and on the Military Advisor.

 

[raxo2222]

If I turn off Alt-Timelines, I get a number of error messages when loading the game, all of which refer to the Pepper2000 module. I am able to click past the error messages and get the game to load.

Pepper module references stuff in alttimelines meaning alt timelines cant be disabled.

 

[Yudishtira]

Pepper module references stuff in alttimelines meaning alt timelines cant be disabled.

1. Do you know for certain that turning off alt-timelines results in fatal errors? So far the poster has not encountered any.

2. Unless y'all have had The Discussion and reached A Conclusion, alt-timelines are optional (DH and possibly others have worked hard to make it so), meaning this is a bug in Pepper's module.

3a. How late in the game does content from Pepper's module start getting encountered?

3b. How easy is it to turn off said module?

 

[raxo2222]

1. Do you know for certain that turning off alt-timelines results in fatal errors? So far the poster has not encountered any.

2. Unless y'all have had The Discussion and reached A Conclusion, alt-timelines are optional (DH and possibly others have worked hard to make it so), meaning this is a bug in Pepper's module.

3a. How late in the game does content from Pepper's module start getting encountered?

3b. How easy is it to turn off said module?

1. Even if there are no fatal errors still something could be broken.

2. I guess DH didn't knew, that Pepper module reference stuff in alt timelines.

3. Some stuff start coming at Renaissance/Industrial eras.
This module does changes to Earth stuff, so you can't really disable it.

 

[raxo2222]

@Thunderbrd I was testing new national culture scaling (complex leaders), so to speed up process I bumped up palace culture production.
It seems like culture "velocity" calculation overflows somewhere.

You can't generate more than 100 000 culture per turn.

I added file with modified Palace NW so you can test it out by doubling culture output.
As for modifiers I had to counter initial reduction in culture.

 

[MattCA]

passed 3 turns with T-6000 save. no crash. checked my dll, its at svn 10523. I might update and try again.
did 3 turns with Nivaras and got a non-fatal msg from python - variable referenced too early.
I only spent a second to fix and I don't mean to waste any1s time with this.

Spoiler BarbarianCiv.py :

// current
line 428:        if -1 in aList:
line 429:            iStd = -1

line 460:        else: aList.append(iStd)



// mine (Matt's)
line 428:        iStd = -1
line 429:        if -1 in aList:

line 460:        else: aList.append(iStd)

file is PPIO v0.6.1.2

 

[raxo2222]

passed 3 turns with T-6000 save. no crash. checked my dll, its at svn 10523. I might update and try again.
did 3 turns with Nivaras and got a non-fatal msg from python - variable referenced too early.
I only spent a second to fix and I don't mean to waste any1s time with this.

Spoiler BarbarianCiv.py :

// current
line 428:        if -1 in aList:
line 429:            iStd = -1

line 460:        else: aList.append(iStd)



// mine (Matt's)
line 428:        iStd = -1
line 429:        if -1 in aList:

line 460:        else: aList.append(iStd)

file is PPIO v0.6.1.2

could be bug in PPIO @Toffer90