Bug Reports and Technical Issues

[jorissimo]

Constant MAFs

Hi all,
I'm getting constant MAFs on DoC 1.15 as well as on the most recent Git version starting from about the 700s AD, or from the moment I start to play with naval units. This was with India. The same happens with Indonesia in the 600AD start, right from the start, making me suspect that naval units are the culprit. The 600AD world is, after all, a lot emptier than the 3000AD one would be around 600AD and so it can't just be the sheer amount of civs and units on the map that is causing it. My computer is 64 bits, has 8 gigs of ram, an i5 processor running at 2.6 Ghz, and civ 4 and mods such as RFC and RAND normally run like a charm on high graphics in all ages, early game, late game, you name it. I have been reading a lot about MAFs and what is causing them. I tried the 4 GB patch, tried tinkering with unit paging, put my laptop on airplane mode and turn off all other big programs and processes including the virus scanner when I play, put all graphics on low, turned off all animations etc., but the MAFs show no mercy. I even tried deleting the naval units, but to no avail. It seems like whatever I try, the MAFs come knocking, occurring practically every turn starting from the aforementioned moments. I really don't get it anymore. With a 64 bits, 8 GB RAM system and the 4 GB patch, what could be the problem? An interesting observation, however, in which the solution might lie, is the fact that, when I open task manager the moment the game freezes or shuts down, memory usage is always just below the 3 GB mark, and the line showing the performance's development through time shows a long, flat line right below the 3 GB mark. This suggests that the game is unable to utilize more than 3 GB even with the 4 GB patch, which should allow usage of up to 4 GB, right? Does anyone have any ideas or suggestions?

 

[Nyayr]

I assume some sort of code pushed out their ships from somewhere else to one of their cities. Usually coastal cities are preferred for that but I guess that's what happens when there isn't any.

Nah they never ever had a coast line... I prevented them from that by building the only coast city in their domain a few turns after they spawned. I decided to be a greedy Rome! Who wants all domains they can own historically. ^^ Even vassalized the Gaul, Germania and Spain.

Unless maybe they got an event that gave 4 galleons?

 

[Leoreth]

That's what I assume happened? You cannot build Galleons without water access.

 

[1SDANi]

That's what I assume happened? You cannot build Galleons without water access.

Could any Colonization related things given France Galleons?

 

[Leoreth]

Trading Company gives one Galleon, otherwise I cannot directly think of anything.

 

[1SDANi]

Trading Company gives one Galleon, otherwise I cannot directly think of anything.

What about Give Colonists in RhyesAndFall.py?

 

[Leoreth]

Oh, right, that's a good point. I'm going to check that.

 

[Nyayr]

What does count as a colony for Rome? Can't build the Trading company, even with 1 city in south America.

 

[Leoreth]

That should count, I recently changed something there though, if you could upload that game it would be helpful in seeing if it's a bug.

 

[jorissimo]

Fixed the MAF problem. Set bcdedit /set increaseuserva 4069 and downloaded an app called CFF explorer, which is basically a more complex version of the 4 GB patch that allows LAA flagging of dlls as well. So aside from the .exe which had already been LAA flagged, I flagged all of the dlls in the my games BtS folder. When I loaded up an existing Indonesia save, it still crashed. Loading up a new game meant no more crashes however. The only thing is that I'm not sure what actually fixed the problem, the whole bcdedit and the flagging, or just starting a new game. Or a combination of the two, with the made changes only coming into effect with the new game.

 

[Leoreth]

I think what actually solved your problem is that the game assets are now properly cached, since none of what you did should have any effect on the available memory beyond the 4GB you've already enabled for yourself.

 

[jorissimo]

Hmm makes sense actually. I re-installed the whole thing after all, so the cache was empty. So are you saying that it takes a couple of re-starts for all of the game assets to be cached?

 

[Leoreth]

Honestly, I have no clue what actually happens under the hood. It's just that I've observed that the mod (and other resource intensive mods) shows the behaviour where the initial starts are slow or even produce MAFs. It's similar to the loading time before the mod is even started, which often is really long until the mod has been started a couple of times and then becomes more bearable. I can only assume some kind of caching or efficient memory usage starts to kick in after a while.

 

[jorissimo]

That would explain it. The first couple of times I loaded the mod it took ages. Currently it loads quite fast.

 

[Nyayr]

That should count, I recently changed something there though, if you could upload that game it would be helpful in seeing if it's a bug.

Here it is, They don't have any colonial expenses, so don't add as a trade company options. Also changing to State Party don't affect them either, in a negative.

 

[jorissimo]

After playing a bit more, the caching explanation makes more and more sense. I've noticed that especially just after loading up a save the game is very sensitive to crashes. Even a brusque camera movement or a deviation from the automatic unit cycling can cause the game to crash. Right now, whenever I load up a save, I just let it sit there for a few minutes before I start playing. Then I say a little prayer and try not to touch anything, not select anything, not move the camera and just follow the automatic unit cycling until I feel like the game has become stable.

 

[Pedersen]

Tried a game as Russia. Was going well, but at current turn 460, I keep getting a CTD (crash to desktop) every time I end turn.

Using current GIT-version (or so I believe, updated yesterday). Have enabled pythons and such.

Not sure what the problem is, but thought I would share.

 

[Enyavar]

Behold the powerful Capital-Ruin of my Holy Rome shortly after a flip from Vikings.

The savegame is just before the flip occurs, turn 204. End the turn, agree to the flip, then end the follow-up turn and agree to play as Holy Rome: Vienna is both the capital in city-screen AND a ruin on world-screen.
Gameplay isn't hindered at all.

I have seen pretty much the same bug with Mongolia waaaay back in 1.15, but only once. Back then, I did a China->Mongolia flip, and Mongolia assumed the capital of Samarkand, which was displayed as a ruin.
For that reason I am inclined to believe that this is something of a graphical error and not at all related to the raze-city mechanic.

Note that in both cases, the AI-preferred capital spot was blocked: As Chinese, I had pre-settled the better spot 1N of Karakorum; as Vikings, I had pre-settled the better spot of Cologne instead of Frankfurt. Which is the reason that the AI spawns the capital in the second-preferred spot, but displays the already existing city there as a ruin. (at least, that's my guess)

NEVERMIND (Vienna fixed itself right when it first grew the population. Back then with Karakorum, it remained a ruin throughout my game. So this is a very minor and temporary bug.)

 

[1SDANi]

Behold the powerful Capital-Ruin of my Holy Rome shortly after a flip from Vikings.

The savegame is just before the flip occurs, turn 204. End the turn, agree to the flip, then end the follow-up turn and agree to play as Holy Rome: Vienna is both the capital in city-screen AND a ruin on world-screen.
Gameplay isn't hindered at all.

I have seen pretty much the same bug with Mongolia waaaay back in 1.15, but only once. Back then, I did a China->Mongolia flip, and Mongolia assumed the capital of Samarkand, which was displayed as a ruin.
For that reason I am inclined to believe that this is something of a graphical error and not at all related to the raze-city mechanic.

Note that in both cases, the AI-preferred capital spot was blocked: As Chinese, I had pre-settled the better spot 1N of Karakorum; as Vikings, I had pre-settled the better spot of Cologne instead of Frankfurt. Which is the reason that the AI spawns the capital in the second-preferred spot, but displays the already existing city there as a ruin. (at least, that's my guess)

NEVERMIND (Vienna fixed itself right when it first grew the population. Back then with Karakorum, it remained a ruin throughout my game. So this is a very minor and temporary bug.)

Same thing happens every time with Iran.

 

[Leoreth]

Here it is, They don't have any colonial expenses, so don't add as a trade company options. Also changing to State Party don't affect them either, in a negative.

Oh, I assume you started that game and updated while it was going? I introduced a new system to handle different continents (for performance) which is initialised at game start, so it doesn't work in ongoing games. Pleasantly surprised actually that the save was otherwise compatible, looks like my efforts to design for compatibility make a difference.