RevolutionDCM for BTS

[keldath]

hey glider,

can you tell me plz how can i disable one of the bug tabs/revolution tabs ?

i want to hide the unit renaming tab.

thanks !

 

[glider1]

@Kel
Yeah that's simple. Just comment out or delete this line in assets/config/init.xml:
<tab module="BugUnitNameOptionsTab"/>
You will find it at line 177 of that file.

Cheers.

 

[keldath]

great thanks buddy.

 

[Duuk]

There is one mod I've never been able to re-find that I saw when civ first came out.

It displayed in text the "hidden" modifiers that the AI used to determine relations. Stuff like "Your country is much smaller than mine" or "You have a tech lead".

If these could be re-added to the display, I'd be happy as a clam.

 

[Duuk]

...and one more unit I'd love added, but it looks like it would require harder work that I'd hoped.

Engineer. I know in one of the old civ games you could essentially transfer production from one city to another, and I'd love to see a unit to make life on new cities later in the game easier.

Engineer has...

Cost to build: 300
Moves: 1
Treated like a worker for combat
Adds Mission: Works Project (+100 production, a one-time boost)

Now, I understand this could be used to hurry wonders, but basically: Who cares? You're losing 67% production from one city to increase in another. The main use for this is hurrying lighthouses, granaries, and courthouses in new cities far away, and that's how the AI should use it.

Edit: After some hunting, I'm now 100% sure it would require a tweak to the dll to do this, since that is where I see how GreatWorks for culture work. Essentially, this is a copy of GreatWorks, but the bonus is production.

 

[phungus420]

Critical but in RevDCM 2.00 more information here:

http://forums.civfanatics.com/showpost.php?p=8081440&postcount=459

I realize that it's ocurring in Legends of Revoltution, and not a Pure RevDCM game. But the failed assert that is causing this in in CvSelectionGroupAI.cpp Also I've checked all art, and isolated the player and messed with things. It's not really possible this bug is caused by the modifications in Legends, it must be intrinsic to RevDCM, and I'm nearly certain the BetterAI 0.70f component specifically.

 

[glider1]

Thanks for the work trying to isolate it. I do not know where to start as not only is the problem found in a modded RevDCM but also in a debug dll compiled somewhere else and not a final release dll. It's going to difficult to pin it down but keep trying and I will keep an eye on developments.

@Duuk
The place to add these excellent ideas for these types of small mods is now in the RevDCM WoC forum:
http://forums.civfanatics.com/showthread.php?t=320428&page=3

The philosophy is to palm off as much of the smaller mods to WoC and integrate only the fundamental stuff into RevDCM itself.

Cheers.

 

[phungus420]

Guess I'll have to revert to 1.02 then...

Is RevDCM going to just accept instability now (and thus become uneusable for mods like Legends), or is this going to be fixed?

I've been using RevDCM as a base mod for a while now, as you know. The main reason I picked it is because it has always been stable (apart for that bug with the diplomacy text, which you fixed the next day... delete the autolog.ini). This is exactly what I thought would happen with WoC, critical bugs would be accepted and RevDCM would only be useable for RoM.

 

[glider1]

@Phungus
Stay cool. It will get sorted. There is no point in fixing something until the problem is pinpointed more precisely and more evidence accumulates. Yes please do revert to 1.02 if you think that will help. The only improvement was Better BTS AI 0.7.

Since it has not been reported on the better bts forum, you may well have a point that it is an interaction between revolutions and better bts. However there simply is not enough evidence accumulated by more players.

Hang in there.
Cheers.

 

[glider1]

@Phungus
I simply cannot download your mod due to an expensive ISP download limitation for myself. Therefore if you specify what "conditions" you think you have found this bug in, I will try to replicate those "conditions". Preferably would be that you do not run your mod as a debug build but instead as final release. Conditions are like:

what map size
what map type
how many starting civs
what options enabled
what revolutions options enabled
what bts options
etc

Cheers.

 

[phungus420]

Dang. Unfortunately I think it's rare to occur. There have been 30 some odd downloads of the recent 0.9.2 test build, and this is the first reported critical bug, and numerous reported completed games. I don't think it's really possible to replicate with similar conditions. I'll see what I can do about converting the save to base BtS, the problem is to do that I need to convert to a WB save, and I think doing so removes all wars and such going on, and other pertinent information that will likely stop this bug being triggered...

I do think it has to do with spys though, call it a hunch but things like the Log, and messing around with the spys seem to point to that.

 

[phungus420]

Yeah, have to do a WB save.

I have a solution though, it will take a while to get it to work.

I can change all the art references to default BtS references, remove the art, and you could access the save through that. It wol't look pretty, but that doesn't matter, and Will take a few hours to set that up.

Edit:
Yeah, I can't replicate it when I pair it down to RevDCM 2.00. Lots of things need to change, and it just doesn't trigger. I do have a functional converted save though, but no more CvSelectionGroup assert failure.

 

[glider1]

Ok I'm treating the absense of the reoccurence of this "bug" as good news. If it ever resurfaces, record down the game starting conditions and options so that it can be replicated in a new environment.

Don't forget that you can turn off options while the game is running. If you have autosave every turn, you can then turn the options off prior to the crash and see what happens. Also note that sometimes debug dll's indicated problems that can be ignored without fatality. A debug dll is like driving a rally car down a dirt road in a forest at 100mph. There are some trees you can ignore but others you simply cannot.

Cheers.

 

[keldath]

hi...

i tested many games with my dunemod that has revdcmwoc, ive had a ctd , but on load game it wont return, not sure its revdcm connected cause i had this with previous bersion of my overlord2 mod,

so basically, just wanted to say im very happy with this build of revdcm and woc.

i suggest you add unit statistics for next release

 

[cripp7]

I tried the 2.00 and when I pull up the options I get a blank screen and can't exit it out. I have to alt-tab out then end it. Of course I had to set the customassets =1.

 

[phungus420]

Still working on converting the save from gatlin gun so that it is a small download and trips the assert from pretty much a pure RevDCM mod. This is very hard. I have been making progress though. It seems I can wipe the changed units in the Legends mod, and still get it to trip the assert, but mucking with the leaders and civs makes it fails to keep it tripping. It seems a certainty now this is RevDCM based, as a straight RevDCM core along with straight RevDCM units still trips this failed assert, and I can't see how modified leaderheads and civs could effect CvSelectionGroup, anyway, will continue on this.

Edit: This is so friggin tedious....

 

[GatlingGun]

Still working on converting the save from gatlin gun so that it is a small download and trips the assert from pretty much a pure RevDCM mod. This is very hard. I have been making progress though. It seems I can wipe the changed units in the Legends mod, and still get it to trip the assert, but mucking with the leaders and civs makes it fails to keep it tripping. It seems a certainty now this is RevDCM based, as a straight RevDCM core along with straight RevDCM units still trips this failed assert, and I can't see how modified leaderheads and civs could effect CvSelectionGroup, anyway, will continue on this.

Edit: This is so friggin tedious....

Is the bug occuring in python or in one of the dlls? I have quite a bit of generic python expertise if that helps... although my knowledge of civ infrastructure is close to nil

 

[BobTheTerrible]

Well, I don't know if this was ever fixed, but since sometimes rebels spawn spies before they have any cities, if the spy successfully completes a mission and has no palace to return to, the game doesn't like this very much. I remember a spinoff mod of RevDCM fixing this but I don't remember if RevDCM fixed this itself.

 

[phungus420]

Is the bug occuring in python or in one of the dlls? I have quite a bit of generic python expertise if that helps... although my knowledge of civ infrastructure is close to nil

Fairly certain this is a dll AI logic issue. I can't see how it could be related to Python. Anyway, gotten close, stripped out everything that isn't RevDCM except the civs and leaderheads. Going to tweak this a little more, I may be able to go pure RevDCM WB save that shows this. However it is appearing like I will need to include a couple modded civs and leaders from legends, and as converting these has caused the bug not to trip. So far though I've been able to slowly wittle this save with a scapal and it's working, but every time I try to be a little more blunt in my aproach, the failed assert stops triggering . Anyway if I do end up needing to include a modded civ and leaderhead reference, I'll set it up so that the art defines calls for a default leaderhead in civ. So this will be small, no art will need be included. This has been a serious PITA tedious work, I hope once I get it to you you can use it to find and squash this bug.

Also gatling, there is a non critical assert failure that is caused by the game searching to see if the city has No_Religion. Obviously this is a bug, and while it doesn't crash, I'm sure it's doing no good. This one is easily reproduceable, just play a game of RevDCM with a debug dll, it'll will trip from turn 1, and pop up at least a dozen times per turn. Now this bug I think may be caused by Python. The inquisitions component specifically.

 

[glider1]

@Keldath

so basically, just wanted to say im very happy with this build of revdcm and woc.
i suggest you add unit statistics for next release

Thanks for the testing. Yeah WoC and RevDCM are really no big deal, with most of the "overhead" being WoC code that loads in during game initialization. Suspect that your load CTD's are possibly because there is some small difference between versions in your mod or a problem in the merge itself. If the WoC modules folder is empty (except for the two MLF files) then WoC does nothing. A possible debug way to distinguish between WoC and not is to load up your mod from the "my documents" folder, upon which WoC completely goes dead altogether, but at least the mod loads as per normal. There is a possible "issue" with the font's in version 2.00 of RevDCM that is currently being looked into.

@Phungus
Yeah debugging is tedious at times no question. Not all debug asserts are fatal. Ultimately, break pointing is your best tool.

@Bob the T

if the spy successfully completes a mission and has no palace to return to, the game doesn't like this very much. I remember a spinoff mod of RevDCM fixing this but I don't remember if RevDCM fixed this itself.

That is an interesting scenario. Whatever is "fixed" in Revolutions is now fixed in RevDCM because they are one and the same since RevDCM 1.00.

@GatlingGun

Is the bug occuring in python or in one of the dlls? I have quite a bit of generic python expertise if that helps... although my knowledge of civ infrastructure is close to nil

I'm probably somewhere in between. My python is not bad but not that good, but I do have some knowledge about the civ infrastructure which is not too bad. It will be handy to have someone around to ask the occasional question about Python.

Cheers.