Bug reports and technical issues

[Hippo8085]

Just as a wild guess without opening the game to see for myself: on line 459 of CvRFCEventHandler.py it should be con.iCivicEgalitarianism instead of con.iEgalitarianism

 

[Manuss]

Spoiler :

 

[Leoreth]

Alright, I've committed a fix.

 

[citis]

I know it is minor fix, but...

 

[Leoreth]

It's part of the base game. I don't know why it cannot be found.

 

[citis]

It's part of the base game. I don't know why it cannot be found.

Because it is part of a different mod. The RFC DoC doesn't have access in Gods Of Old mod files. You have to include it in your files.

 

[Leoreth]

Oh, you're right.

 

[adhiraj.bose]

I didn't raise this before. But the videos for some of the wonders *( basically all the wonder videos from the base BTS ) don't show. Just some graphics and the music. Its on version 1.11. I never raised it because it didn't affect the game as much. But what exactly is the matter with this ?

 

[Leoreth]

I already know about this. The SVN isn't affected as it relates to the packaging of the mod.

 

[Hamir]

Has the way the Japanese UP works changed / or is something wrong?

I seem to recall Japanese units needed fewer experience points per promotion, but it seem to be the same as other Civs at the moment (i.e. 2 / 5 / 10)

(Latest SVN, 600AD)

 

[citis]

Thanx Leoreth for the new aztec 3rd goal! However, does it count native, barbarian and indi units as "old world" units?

 

[Leoreth]

Oh, good point, it probably does.

 

[citis]

Traceback (most recent call last):
File "BugEventManager", line 402, in _handleDefaultEvent
File "CvRFCEventHandler", line 174, in onCityAcquired
File "Stability", line 56, in onCityAcquired
File "Stability", line 280, in checkStability
File "Stability", line 1168, in calculateStability
File "Stability", line 1571, in calculateEconomicGrowth
File "RFCUtils", line 1801, in linreg
ZeroDivisionError: float division
Traceback (most recent call last):
File "BugEventManager", line 402, in _handleDefaultEvent
File "CvRFCEventHandler", line 618, in onBeginGameTurn
File "RiseAndFall", line 1141, in checkTurn
File "RFCUtils", line 494, in cultureManager
RuntimeError: unidentifiable C++ exception
Traceback (most recent call last):
File "BugEventManager", line 402, in _handleDefaultEvent
File "CvRFCEventHandler", line 777, in onChangeWar
File "Stability", line 83, in onChangeWar
File "Stability", line 280, in checkStability
File "Stability", line 1168, in calculateStability
File "Stability", line 1571, in calculateEconomicGrowth
File "RFCUtils", line 1801, in linreg
ZeroDivisionError: float division

I propose to change the line 1800 as follows:
det=n*Sxx-Sx^2+1

Spoiler :

What kind of economic calculation is that? Kensyan economics?

The File "RFCUtils", line 494, in cultureManager is a known bug. Although not so frequent, a little annoying. Trying to find this function, I din't find it's definition, maybe there isn't such a function, can you use the plot.getCulture instead?

 

[Leoreth]

What kind of economic calculation is that? Kensyan economics?

It's not an economic calculation at all. I'm doing a linear regression over the commerce data points to smooth out all the spikes in the commerce graph.

 

[Leoreth]

Now I'm seeing that you're proposing to change linear algebra

I'm currently at a loss to see when such an error would occur. Can you give me a save where it happens?

Edit: as for the other bug, I'm aware of it, but CyCity::getCulture(int iPlayer) is clearly defined so that is not the problem.

 

[citis]

Now I'm seeing that you're proposing to change linear algebra

I'm currently at a loss to see when such an error would occur. Can you give me a save where it happens?

Edit: as for the other bug, I'm aware of it, but CyCity::getCulture(int iPlayer) is clearly defined so that is not the problem.

In linear algebra the gradient can be infinite (it is a vertical line), but have in mind that a vertical line ISN't a function (well, it is the only case, that a line isn't a function). I don't have problem with linear algebra, but the implementation is buggy . You should check for zero det and return infinite a and b .

It happens during autosave, waiting for the Dutch, so it is difficult to form or spot the exact autosave. I'll post one as soon as I find one.

CyCity refers to the m_pCity. I suppose that m_pCity is a CvCity pointer, so CyCity.getCulture calls CvCity.getCulture which is absent.

 

[Leoreth]

In linear algebra the gradient can be infinite (it is a vertical line), but have in mind that a vertical line ISN't a function (well, it is the only case, that a line isn't a function). I don't have problem with linear algebra, but the implementation is buggy . You should check for zero det and return infinite a and b .

I know that a vertical line isn't a function. But returning an infinite gradient is just as wrong and would only result in further errors.

The problem is that the input to the linear regression is a list of turn/commerce pairs one turn apart, so the linear regression should never run into situations where the determinant is 0. I'd rather address this problem than returning nonsensical values in case of nonsensical input.

It happens during autosave, waiting for the Dutch, so it is difficult to form or spot the exact autosave. I'll post one as soon as I find one.

Yep, that would help.

CyCity refers to the m_pCity. I suppose that m_pCity is a CvCity pointer, so CyCity.getCulture calls CvCity.getCulture which is absent.

It does, and CvCity::getCulture() does exist. The code wouldn't even compile if it didn't.

 

[Rorgloin]

The Byzantines still do not flip Roman cities in Egypt on the 3000 BC start.

 

[Leoreth]

I know, haven't be able to investigate yet.

 

[citis]

It seems there's a problem concerning the Dutch spawn...