Fall Further - The Original Thread

[Teg_Navanis]

Reproduce this...
(...)

That's the same problem we discussed here.

To give the short answer, some of your units' statistics are not properly initialised.

The bug is not in the code that's listed in the traceback, but something must have gone wrong earlier, when the unit that caused the popup was created. Did you start the game in which this happened with an earlier version of FallFurther? Have you manually changed any of the settings or files? If not, I'll have a look at the settings/files in the mod tomorrow to see if there is perhaps some way the data is not properly initialised (or deleted afterwards).

 

[Ambreville]

That's the same problem we discussed here.

To give the short answer, some of your units' statistics are not properly initialised.

The bug is not in the code that's listed in the traceback, but something must have gone wrong earlier, when the unit that caused the popup was created. Did you start the game in which this happened with an earlier version of FallFurther?

Nope.

Have you manually changed any of the settings or files?

Nope.

If not, I'll have a look at the settings/files in the mod tomorrow to see if there is perhaps some way the data is not properly initialised (or deleted afterwards).

Good luck.

 

[Daijin]

I was wondering how we were doing with the Multiplayer side of the field. Too bad it's apparently "not good"


To help isolate when/how the OOS is happening, you need to get a feel for what kind of actions cause it. Disabling all events is a classic quick check. If your first few turns tend to always be OOS free, then ideally it is a feature available at later Techs, so you start to track what Tech level it happens at to narrow things down. Maybe play some Human only games and really Specialize on the Tech Trees (I'll do Archery line, you do Melee Line) to see if you can figure things out like "If we avoid the priest line, no OOS issues."

If the OOS happen right from the start of the game, then try playing with which Civilizations you are using, see if it is specific ones.


Biggest help would probably be to figure out if the issue happens just with FF, or also with FfH. Knowing that cuts down on where we have to look considerably.

After some deep thought, and then testing to see if it was the culprit. It appears to be the "Flavour Start" option that does the OOS issue.
No longer having any OOS with that option off.

 

[xienwolf]

Interesting. Pity you didn't post that 2 hours earlier when I finally got my wife to sit down for a game of FF Hit an OOS on the end of the first turn, wound up doing it HotSeat style.

 

[Jean Elcard]

After some deep thought, and then testing to see if it was the culprit. It appears to be the "Flavour Start" option that does the OOS issue. No longer having any OOS with that option off.

Interesting and very strange at the same time. I really can't think of a reason, how Flavour Start could cause an OOS. All the stuff it does happens before the actual game starts. After this, its code is practically non-existent, because it's not called again until you start another game.

At least I know it was working without OOS in earlier versions. In conjunction with Fall from Heaven, though. I played a game myself together with some friends two weeks ago. To find out, what's causing the newly arisen OOS errors I would have to take out junks of the new code since then one at a time and see, if the OOS disappear. But to test for OOS I would also need two computers and this is the biggest problem at hand here.

I would be interested in knowing, if anybody else can confirm, that the Flavour Start option alone causes the OOS. Furthermore, if somebody has the time to try it and knows how to do it, I would like to know, if the OOS vanish, if you switch off (set it to False) one off the following two options (or both of them) inside the file "CvFlavourInterface.py":

ASSIGN_FLAVOURFUL_STARTING_LOCATIONS
ASSIGN_FLAVOURFUL_UNIQUE_IMPROVEMENT_LOCATIONS

It's important that you apply the same changes on both computers! Otherwise OOS are guaranteed.

 

[Elm]

Using the Sun and Shadow map posted earlier in this thread -- those python errors have started coming up every turn, very early in the game. Very annoying. Guess I will turn of python error reporting.

 

[Vehem]

Using the Sun and Shadow map posted earlier in this thread -- those python errors have started coming up every turn, very early in the game. Very annoying. Guess I will turn of python error reporting.

Does that map include pre-placed units (other than the normal Settler/Warrior/Scouts)?

 

[Demus]

yep, an entire minotaur island for example, cities pre-placed aswell

 

[Rex rgis of Ter]

This can be fixed. I'll post another version with only setttlers and warriors, and redo the western islands.

 

[Teg_Navanis]

Concerning the python exception Ambreville reported: uncommenting lines 51 and 171 of UnitStatisticsTools.py will stop it from happening.

 

[Vehem]

This can be fixed. I'll post another version with only setttlers and warriors, and redo the western islands.

Thanks - but won't be needed with any luck because...

Concerning the python exception Ambreville reported: uncommenting lines 51 and 171 of UnitStatisticsTools.py will stop it from happening.

...added this - will be included in the next release (either 042 or possibly a Patch E). Thanks Teg.

 

[Ambreville]

After trying half a dozen new games, I keep getting repeat CtD's like the ones I described yesterday or the day before when running Kahd. These don't happen with other races --> ergo: there must be something wrong with the Kahdi race (even if your fancy computers can't reproduce them from my saves).

 

[Jean Elcard]

After some deep thought, and then testing to see if it was the culprit. It appears to be the "Flavour Start" option that does the OOS issue.
No longer having any OOS with that option off.

Eureka! I found the problem! The OOS issue caused by FlavourMod in multi player games will be fixed with the next patch to it.
It had something to do with civ4s message system. I will have to disable the in-game message about the quality of the starting plot your civ got for MP games. That's all. No big loss.

 

[Ambreville]

I found a CtD that I could tie to one of my actions. I played Deirdra (the explorers?), with Runes of Kilmorph. I attacked a Jonas Endain worker goblin with a Stonewarden, and that caused a CtD...

Man, this mod is really unstable.

 

[xienwolf]

Sweet, got a save for us to play with on this one?

 

[Valkrionn]

I found a CtD that I could tie to one of my actions. I played Deirdra (the explorers?), with Runes of Kilmorph. I attacked a Jonas Endain worker goblin with a Stonewarden, and that caused a CtD...

Man, this mod is really unstable.

To be honest, I've never actually had a CTD from Fall Further, it's always played fine for me. I'm really kind of amazed you find so many.

 

[Vehem]

Eureka! I found the problem! The OOS issue caused by FlavourMod in multi player games will be fixed with the next patch to it.
It had something to do with civ4s message system. I will have to disable the in-game message about the quality of the starting plot your civ got for MP games. That's all. No big loss.

Excellent - we'll incorporate that for 042 then Thanks Jean.

(even if your fancy computers can't reproduce them from my saves).

To be honest, I've never actually had a CTD from Fall Further, it's always played fine for me. I'm really kind of amazed you find so many.

Aye - the crashes that I've encountered throughout development generally have an identifiable cause and I can trace that to fix it. There have been some problems and on the whole, the community here has been great in reporting them and helping us to sort them out.

I don't think the "fanciness" of the computers is a factor however. I do tend to run mine toward the top end of graphics etc, but I know that Xienwolf runs his at the lowest resolution so between us we should be seeing the game from most angles. A lot of the players will have completely different setups again and have few problems.

EDIT: One thing has just caught my attention though - I took a look at your earlier post regarding your system specs Ambreville. Everything listed seems to be fine for Civ - plenty of memory, decent processor etc - but I can't see your graphics card listed. I did just realize that a non-shader card may have some trouble as I believe some of the graphics we've borrowed might not have non-shader versions. Sometimes you'll get away with that, sometimes it may cause a problem (depends on the level to which the graphics rely on the shaders).

If that is the case, there's a fairly straightforward way to help identify any such units (if a little time-consuming) and we can probably rework some graphics to solve that problem. This would actually explain why the error is non-reproducible.

In anycase - can you post the spec for your graphics card and I'll investigate whether that may be the cause.

 

[Ambreville]

One thing has just caught my attention though - I took a look at your earlier post regarding your system specs Ambreville. Everything listed seems to be fine for Civ - plenty of memory, decent processor etc - but I can't see your graphics card listed. I did just realize that a non-shader card may have some trouble as I believe some of the graphics we've borrowed might not have non-shader versions. Sometimes you'll get away with that, sometimes it may cause a problem (depends on the level to which the graphics rely on the shaders).

If that is the case, there's a fairly straightforward way to help identify any such units (if a little time-consuming) and we can probably rework some graphics to solve that problem. This would actually explain why the error is non-reproducible.

In anycase - can you post the spec for your graphics card and I'll investigate whether that may be the cause.

I have an NVidia graphics card, but I can't figure out a way to get the specs listed to save my life...

---------------------------------

EDIT

After all, here's what I found:

Spoiler :

System Information report written at: 08/06/2008 07:31:19 AM
[Display]

Item Value
Name NVIDIA GeForce FX 5200
PNP Device ID PCI\VEN_10DE&DEV_0322&SUBSYS_00000000&REV_A1\4&26E5833C&0&0008
Adapter Type GeForce FX 5200, NVIDIA compatible
Adapter Description NVIDIA GeForce FX 5200
Adapter RAM 256.00 MB (268,435,456 bytes)
Installed Drivers nv4_disp.dll
Driver Version 6.14.10.9371
INF File oem25.inf (nv4_NV3x section)
Color Planes 1
Color Table Entries 4294967296
Resolution 1024 x 768 x 60 hertz
Bits/Pixel 32

 

[Demus]

for starters, rightclick your desktop, select properties, then go to the "settings" tab (or something like that, translated from dutch). This works for XP anyway, showing your version of your graphics card.

 

[Ambreville]

Thanks. I finally figured it out (see edit).