Specific Bug Reports

Wodan · Feb 8, 2007

Assert Failed

File: .\CvCity.cpp
Line: 472
Expression: (0 < GC.getNumProjectInfos())
Message: GC.getNumProjectInfos() is not greater than zero but an array is being allocated in CvCity::reset

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

this one came from trying Omens

Wow, a 2fer....

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

Assert Failed

File: .\CvGlobals.cpp
Line: 1236
Expression: eUnitNum > -1
Message:

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

Iustus · Feb 8, 2007

Mexico said:
as reported earlier in http://forums.civfanatics.com/showpost.php?p=5060449&postcount=142 and checked in rev.350, this bug is still present
when AI want move unti to city for upgrade, but path is blocked, game crashed

Woops. Thank you for catching this one. I looked into this, but I saw this code:

Code:

if (bCanPathToClosestCity && iClosestCityPathTurns < iUpgradeCityPathTurns)
{
	pThisTurnPlot = getPathEndTurnPlot();

So, I thought Blake already checked in a fix.

Of course, thats further down in the code.

Fixed now.

-Iustus

Cyrus_the_Great · Feb 8, 2007

Iustus, sorry but I don't have a save anymore for my Barbarians assert. It seems like I can't duplicate the error anyway. I will wait for the next build and go back and try it again. I'll be sure to post a save game.

Thanks to you and Blake, for making this version of Civ so much more appealing, with all your efforts.

Iustus · Feb 8, 2007

Regarding all scenarios...

I am not sure that it makes sense to 'remove' all the asserts that you get on Warlords scenarios.

The no projects in the list one is pretty easy to remove, but that does not deal with some of the others.

Some default python code is asking for units of type "GUN" and type "ARMOR" which does not exist in some of the scenarios. In this case, the only way to 'fix' it would be to either change the python code (out of scope of this project), or to never report when someone tries to get a type that does not exist (which I think would be bad).

In every case I have found so far, just ignoring the assert seems to work.

I would still like to see people report asserts they recieve in scenarios though. It is just that most of them are harmless.

The issue is that Firaxis put all these asserts in there, and I am reluctant to remove the Firaxis asserts just because Firaxis scenarios trigger them.

I will run through all the scenarios to see if I find any real problems.

Anyone have any thoughts on this?

-Iustus

Wodan · Feb 8, 2007

Iustus said:
Regarding all scenarios...

In every case I have found so far, just ignoring the assert seems to work.

Anyone have any thoughts on this?

Well, the real question is...

If we have a no-Assert build of BetterAI (like, the real/final v.1.0), what are the ramifications? Is it "ignore all" by default? Is Joe Newbie going to crash, or is he going to be fine?

If it ignores and everything's fine, then there you have your answer, methinks.

Wodan

shadyforce · Feb 8, 2007

This looks like the one you were referring to.

Code:

Assert Failed

File:  .\CvCity.cpp
Line:  472
Expression:  (0 < GC.getNumProjectInfos())
Message:  GC.getNumProjectInfos() is not greater than zero but an 
array is being allocated in CvCity::reset

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

I'm half way through the scenario and after ignoring the assert, there have been no problems.

mrgenie · Feb 8, 2007

Hi Iustus, Blake anyone, else playing betterAI..did anyone play with RAGING BARBARIANS options enabled? i get a CTD either on map creation or first couple of turns..can anyone second this?

shadyforce · Feb 8, 2007

Code:

Assert Failed

File:  .\CvGlobals.cpp
Line:  1236
Expression:  eUnitNum > -1
Message:  

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

Another assert, this time towards the end of the setup after starting a new Rome Scenario.

Iustus · Feb 8, 2007

Wodan said:
Well, the real question is...

If we have a no-Assert build of BetterAI (like, the real/final v.1.0), what are the ramifications? Is it "ignore all" by default? Is Joe Newbie going to crash, or is he going to be fine?

If it ignores and everything's fine, then there you have your answer, methinks.

Wodan

Yes, a non-assert build is basically ignore all by default. (It actually never even does the check).

My concern is that people will fail to report asserts in scenarios, because they get that spew of them at the start, and some during.

I would like to see all asserts continue to be reported, even though some are Firaxis bugs in the scenarios (or in their python code) which I am not going to fix.

-Iustus

KHSOLO · Feb 8, 2007

Ok here is trhe assert error:

Here is the saved game:

http://forums.civfanatics.com/uploads/66541/GUILLERMO_SCHMIDT_AD-2363.zip

Just start the save and click for next turn.

Thanxs for the hard work amigo, your work is really invaluable and very close to the communitys needs, congrats and S!!

bolt144 · Feb 8, 2007

Following the directions from Iustus, I was able to capture logs from an OOS error. I ran WinDiff and a screen shot is below of where the first non-Asynch difference was. I will attach the logs too.

Iustus · Feb 9, 2007

bolt144 said:
Following the directions from Iustus, I was able to capture logs from an OOS error. I ran WinDiff and a screen shot is below of where the first non-Asynch difference was. I will attach the logs too.

Fantastic, thank you. I believe I have found the source of the problem. The next build (sorry held it a day for this), should have the fix.

-Iustus

bolt144 · Feb 9, 2007

Iustus said:
Fantastic, thank you. I believe I have found the source of the problem. The next build (sorry held it a day for this), should have the fix.

-Iustus

Excellent. Looking forward to trying it out. Thanks!

Iustus · Feb 9, 2007

New build posted 2/09

These are known issues with this build (I should have put this in the release notes as well):

- Some scenarios will trigger harmless asserts (UNITCOMBAT_GUN, UNITCOMBAT_ARMOR, etc). These are caused by python code asking for things that do not exist. Please continue to report asserts, unless you are certain that your specific one was already reported. When you get one of these, you should be able to click ignore always and continue to play.

- Ironclads trigger two asserts when used. This is caused by an XML bug which Firaxis needs to fix (the sound effect seems to be missing or have a typo). You can click ignore always.

Please continue to report all asserts you see, but remember that many of them are harmless. Often the intent of the XML based asserts is to indicate to the Scenario developer that they have an error they need to fix. I only wish that the Firaxis developers had run their own assert build more often to fix these issues. The game should run fine though.

You can always download the non-assert version of BetterAI to compare results, but clicking "Ignore" should have the same effect as just running the non-assert version.

-Iustus

gnome · Feb 9, 2007

2/9 release

Assert Failed

File: .\CvPlayerAI.cpp
Line: 4018
Expression: ePlayer != getID()
Message: shouldn't call this function on ourselves

----------------------------------------------------------
(after ignoring)

Assert Failed

File: .\CvPlayerAI.cpp
Line: 4032
Expression: ePlayer != getID()
Message: shouldn't call this function on ourselves

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

Occurred when reloading save. Also occurred on previous build.

Mods (combined)
Better AI Handicaps
Look & Feel
Bad Ronald's Flags

Iustus · Feb 10, 2007

gnome said:
2/9 release

Assert Failed

File: .\CvPlayerAI.cpp
Line: 4018
Expression: ePlayer != getID()
Message: shouldn't call this function on ourselves

----------------------------------------------------------
(after ignoring)

Assert Failed

File: .\CvPlayerAI.cpp
Line: 4032
Expression: ePlayer != getID()
Message: shouldn't call this function on ourselves

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

Occurred when reloading save. Also occurred on previous build.

Mods (combined)
Better AI Handicaps
Look & Feel
Bad Ronald's Flags

This looks like a bug from one of those python mods. Someone is calling AI_getAttitude on themselves. I am guessing the "Look & Feel" mod, but I am not familiar with it. You want to look for a loop over every player calling AI_getAttitude. It should not call this on itself, it should skip the active player ID.

Edit: I found the trouble lines, in autologEventManager.py, as reported here in the Look & Feel thread.

-Iustus

gnome · Feb 10, 2007

Thanks, you rock!

herwin · Feb 10, 2007

Assert Failed

File: .\CvCityAI.cpp
Line: 5409
Expression: false
Message: infinite loop(!!!!!)

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

Unclear why...

WarISgood · Feb 10, 2007

Hello, I like all the changes being made in this *mod

I have a saved game that crashes me to desktop as soon as I end my turn. It does this repeatedly. I have no clue what could be causing it or I would say. It's a great game though that I'd love to continue.

Edit: Supposedly my saved game it too large at 828 kb. All I can add is that there are no other mods running, this is on vanilla civ with latest patch. I have a permanent alliance just recently started with Mao and there are a couple of wars going on. Rifling is about to be discovered on the NEXT (rifling has 1 turn to go before the crash) turn.

Thanks cf nz,

Here is the save
http://forums.civfanatics.com/uploads/20110/Latest.Civ4SavedGame

Edit: I am sorry if I bothered anyone with what turned out to be a simple fix. I needed to get rid of everything in my custom assets folder. Good day B-)

cf_nz · Feb 10, 2007

WarISgood said:
Supposedly my saved game it too large at 828 kb.

Use the Easy Upload system. See image for link locations.

Specific Bug Reports

Deity

King

Warlord

King

Deity

Chieftain

King

Chieftain

King

Warlord

Chieftain

Attachments

King

Chieftain

King

Warlord

Attachments

King

Warlord

Chieftain

Chieftain

Prince

Attachments

Similar threads