Single Player bugs and crashes v35 plus (SVN) - After the 18th of August 2014

Fixed



You can try the latest svn i found something in CvPlot::getBestDefender. In a test turn the time for that function has gone from 1148 seconds to 343 seconds that helps a bit that turn went from 52 minutes to a bit under 40.

Thank you! Great job!

Finally Suppiluiluma capitulated after having only 5 cities left on the south of Arabian peninsula, where Saudi bomb Hausa nowadays, so i can peacefully build up my newly acquired middle east. On the bad side - Incan have built Three Gorges Dam ahead of me, no matter how hard i tried to beeline for hydroelectricity using all GPs i had and telling vassals what to research and having Encyclopedia too, and seems he will get first to Christo Redentor as well.

Next turn after capitulation took only 4 minutes to proceed. I will check how it goes further. (I have not updated SVN yet).
 
Thank you! Great job!

Finally Suppiluiluma capitulated after having only 5 cities left on the south of Arabian peninsula, where Saudi bomb Hausa nowadays, so i can peacefully build up my newly acquired middle east. On the bad side - Incan have built Three Gorges Dam ahead of me, no matter how hard i tried to beeline for hydroelectricity using all GPs i had and telling vassals what to research and having Encyclopedia too, and seems he will get first to Christo Redentor as well.

Next turn after capitulation took only 4 minutes to proceed. I will check how it goes further. (I have not updated SVN yet).

Two AI players in that game have over 2500 units and i saw one stack with at least ~1500 units. How long the turns take depends on the actions of such huge stacks and the units around them so i guess the turns in that game can take from ~4 minutes to a few hours. There is still room to speed it up but that takes time.
 
Two AI players in that game have over 2500 units and i saw one stack with at least ~1500 units. How long the turns take depends on the actions of such huge stacks and the units around them so i guess the turns in that game can take from ~4 minutes to a few hours. There is still room to speed it up but that takes time.

Aye, i saw those 2 stacks standing in front of each other. Cholan and Yu. in one - 400 of each - culverins, falconets, motars and other had 650+ great bombards + other units. I had 0.4 and 0.5 ratio to them at that time, now its 0.6 and 1.0. Seems that Yu stack got wasted - it was near Cholan city and was getting damage it seems.

Well, i am glad those huge games serve a purpose at making mod better :)

[edit] 10-15 turns passed - turns are fast now - well 3-4 minutes a turn or so. Since it takes 5-10 mins to make my own move it's very satisfying.
 
Two AI players in that game have over 2500 units and i saw one stack with at least ~1500 units. How long the turns take depends on the actions of such huge stacks and the units around them so i guess the turns in that game can take from ~4 minutes to a few hours. There is still room to speed it up but that takes time.

Alberts, Im also aware of existence of these huge stacks in-game. What do you think about my idea to increase :hammers: cost of all units by x10? Or maybe you have other solution to decrease amount of units in-game?
 
Alberts, Im also aware of existence of these huge stacks in-game. What do you think about my idea to increase :hammers: cost of all units by x10? Or maybe you have other solution to decrease amount of units in-game?

I have option increased building and unit costs turned on in this game.

Also, i think huge stacks come not from overproduction (well, partly from this source too), but from ability to trade siege units, so more advanced civs accumulate more of those from trading with civs that are falling behind, and with time this gap only increases.

What should be looked at - is HUGE amounts of surgeons, veterinarians, machinists etc.
 
Alberts, Im also aware of existence of these huge stacks in-game. What do you think about my idea to increase :hammers: cost of all units by x10? Or maybe you have other solution to decrease amount of units in-game?

You mean the production costs right?

Maybe they could be a bit higher but the same goes for the unit maintenance costs.


But the whole game enconomy is totally wrong because of some huge errors in the code. Until that is fixed it makes no sense to make such balance changes. I just need to find motivation and time to fix that mess.
 
Alberts, Im also aware of existence of these huge stacks in-game. What do you think about my idea to increase :hammers: cost of all units by x10? Or maybe you have other solution to decrease amount of units in-game?

I have option increased building and unit costs turned on in this game.

Also, i think huge stacks come not from overproduction (well, partly from this source too), but from ability to trade siege units, so more advanced civs accumulate more of those from trading with civs that are falling behind, and with time this gap only increases.

What should be looked at - is HUGE amounts of surgeons, veterinarians, machinists etc.
I would strongly suspect the huge amounts of healer units are due to the huge amounts of other units they are being trained to accommodate.

There's a number of big big big things that need to be done for the AI to bring their stacks back into balance. I don't think we can any longer continue to allow siege units to simply be assigned city attack AIs - they need their own, both to help guide their promo selections but also to help us more finely refine the amount each 'attack' stack really needs. The early game ones throw off some assumptions there because they not only evaluate to be the best siege weapon for the stack but also the best attack unit for the stack which leads them to be vastly overbuilt and when they end up leading the attack stack the whole stack assumes it won't be able to attack when the city defenders can't be further injured by the leader... oy vay so many problems there. The whole coding that handles the building of city attack stacks needs to be vastly restructured. Not a simple task.

And after that then we can look further at the healer units that still don't seem to be quite right in maintaining appropriate numbers and being in the places they should be when they should be.

And then transportation for various purposes...

And so on.

Increasing the production costs of units will only lead to nations taking longer to figure it's ok for them to use the attack stack they have and stall war efforts if severe enough.
 
You mean the production costs right?

Maybe they could be a bit higher but the same goes for the unit maintenance costs.


But the whole game enconomy is totally wrong because of some huge errors in the code. Until that is fixed it makes no sense to make such balance changes. I just need to find motivation and time to fix that mess.

@alberts2,
If I could afford to fly to Germany I would take you out to your favorite Gasthaus buy you the biggest/best meal and and as much beer as you wanted for the evening. :D Wife and I were stationed in Germany in 1974-76. Loved the Country, People, the Food and the beer/wine. Had a favorite Ratskeller(sp?) in Furth Im Wald that we would go play cards at while eating curry wurst, kasebrot, and much beer. We even got to see the 700th Anniversary re enactment of the Drakenfest (Slaying of the Dragon) in Furth. Good times were had by all! :D

JosEPh
 
Tactical Fusi is not working properly (behaves as a normal nuke restricted to just one cell). And I get this on explosion:

PY:Nuke detonated at 8, 7
20:59:56 TRACE: Error in nukeExplosion event handler <bound method NukeAfterEffects.onNukeExplosion of <NukeAfterEffects.NukeAfterEffects instance at 0x1E406B98>>
20:59:56 TRACE: global name 'iPlot' is not defined

I have put a fix for the iPlot problem on the SVN. I have not tested it as I don't have the facilities to do so at the moment.

Heads up for modders, I will be away from my modding computer from next Saturday for an unknown length of time, but at least a week or two. My internet access will be variable over that time also. Mum is having her other hip replaced and needs someone there for awhile, ie until I start getting in the way;).
 
Two AI players in that game have over 2500 units and i saw one stack with at least ~1500 units.
I counted more times ago.

I still think the best fix to increase the unit maintenance costs, and, if the cots per turn get too big, not demolish a building but demolish a unit first.

edit
since last SVN there is a missing image

edit2
German charset not correct (UTF8 failure?)
 
build 8538 crashed and this was the error in the logs:
[2139.210] info type 'BUILDING_BODY_EXCHANGE' not found, Current XML file is: modules\My_Mods\Traditions\Work_Ethic_CIV4BuildingInfos.xml
 
when tunnels are possible?

edit
I opened world builder and had a look what the factions do (or not)
AI factions have too much money and don't use it.
Sample: Dido, has 1 City size 17, lost all other, in main city there are 24 surgeons, more than 800k gold (not shown in game for trading)
 
I have put a fix for the iPlot problem on the SVN. I have not tested it as I don't have the facilities to do so at the moment.

Heads up for modders, I will be away from my modding computer from next Saturday for an unknown length of time, but at least a week or two. My internet access will be variable over that time also. Mum is having her other hip replaced and needs someone there for awhile, ie until I start getting in the way;).

Nope, the issue remains for Tactical Fusi at least. Now it says:

PY:Nuke detonated at 8, 7
20:05:51 TRACE: Error in nukeExplosion event handler <bound method NukeAfterEffects.onNukeExplosion of <NukeAfterEffects.NukeAfterEffects instance at 0x1E406B98>>
20:05:51 TRACE: global name 'iLattitude' is not defined


The result is that the Fusi behaves as a regular Nuke (w/o turning tile into water).
 
more ram usage?
after last SCN I notice that I need ~ 120 MB more ram after load sav than before.
(before it was ~ 1780 MB, after its ~ 1920
Means, that now I can do one turn less (2-3 tuns only) before new reload, than before (3-4).
 
Looks like I've got some kind of marsh generated on a peak tile and "floating" above it.

aenfGU2.jpg
 
For some reason the strategy for the BeastMaster has been set to the Great General. This is completely wrong. I will set it back or remove as the beastmaster is not a unit anyone can get. It is just a graphics thing with the tamed animals.

Nope, the issue remains for Tactical Fusi at least. Now it says:

PY:Nuke detonated at 8, 7
20:05:51 TRACE: Error in nukeExplosion event handler <bound method NukeAfterEffects.onNukeExplosion of <NukeAfterEffects.NukeAfterEffects instance at 0x1E406B98>>
20:05:51 TRACE: global name 'iLattitude' is not defined


The result is that the Fusi behaves as a regular Nuke (w/o turning tile into water).

edit is is the same bit of code as the Plot problem. It was added in to handle the fact that we have three types of coast and ocean, viz polar, temperate and tropical.

That is a totally different error:mischief:. I will investigate. As I said before the whole of the PythonErr.Log file as an attachment to the post will help.

Looks like I've got some kind of marsh generated on a peak tile and "floating" above it.

Peaks are added after, iirc. In that case it was a marsh with a resource which has the peak placed on it. Which map script are you using?
 
Back
Top Bottom