Probably something simple, but I'm SDK stupid

antaine · Apr 1, 2010

So I'm trying to merge the RFC Marathon elements into my own mod of RFC. I used Winmerge to merge in several thousand lines of code to both the python and the c++ files.

But when I try to recompile the DLL I get the following list of errors. I wouldn't be surprised if there's just one dumb thing I forgot to do, but I don't understand what the error log is trying to tell me. Can anyone help me translate from C++ to idiot?

(ps, this was my second attempt...it takes me 3-4 hours for it to compile to this point, so I'm hoping insight will help me get this fixed today instead of weeks of compiling or recompiling via trial and error)

halp... :confused:

General Tso · Apr 1, 2010

You can't merge two source files together and expect them to work. The people who wrote the code for each file didn't design it to work with the other file.

antaine · Apr 1, 2010

but the mod I'm looking to merge was designed to work with RFC. I just successfully added a civ to RFC's mod, and this other mod adds a different year-counting feature to the original RFC to enable marathon and epic speeds...so they were both designed to work with RFC.

They shouldn't really conflict with each other as the marathon mod changes the turn mechanism while my own mod changes tile properties for one civ and changes a settler map (the rest of my alterations are in python).

I was just able to successfully merge the marathon mod files and the original RFC files (I ran it to make sure they work), re-make all my alterations, and am trying a clean recompile from scratch.

embryodead · Apr 1, 2010

The SDK part RFC Marathon should merge easily with your mod, don't think there could be any overlapping. There messages are rather vague, looks like all the functions from CvEventReporter.cpp are missing but I don't really know what it means.

Let me get this straight: you merged RFC Marathon into RFC (kind of redundant but OK) and it works?
Then you re-add your stuff in CvRhyes.cpp and it doesn't work? What other files did you change apart from CvRhyes?

antaine · Apr 1, 2010

Well, the first time, I merged v2 and v3 (and made sure to include the extra files in that other package, marathon-3x-files) into my RFC Irish. When I tried to recompile the DLL I got the error messages printed here.

I figured there might be a problem due to files created when I compiled my mod etc, so I then merged the v2 and then v3 into RFC (but did not recompile the DLL, I was testing everything else at that point). Then I re-made all my changes into this re-modded mod of a mod and tried recompiling. 3 hrs later, I got the same set of error messages.

I'm now finishing a recompile in which I weeded out the CAR mod, and just included your stuff. I seem to be having better luck with that, it seems that if I reject all lines associated with the CAR mod (that's the stuff marked "Sanguo," no?) I don't get the error messages.

The problem right now is that while *almost* every line associated with Sanguo is signed, there are a few stray lines that are not (I just ran into a bunch in CvTeamAI.cpp), so I "build" and then get errors which make enough sense to me to see what lines are causing the problems, change those lines back and try again.

The good thing is that at this point, I don't have to do the compile from scratch. I can get the errors, change the lines, hit "build" and pick up from where I left off (no more 3 hour waits...yay!).

I should have the rest of the lines weeded out within the hour and a DLL compiled that I can try and see if it runs or crashes.

antaine · Apr 1, 2010

alright, eliminating CAR seemed to work, but I'm getting a python error now,

File "RiseAndFall", line 330, in getExileData

EOFError

the line in question reads:

scriptDict = pickle.loads( gc.getGame().getScriptData() )

the lines around it read (it's the second line, below):

def getExileData( self, i ):
scriptDict = pickle.loads( gc.getGame().getScriptData() )
return scriptDict['lExileData']

def setExileData( self, i, iNewValue ):
scriptDict = pickle.loads( gc.getGame().getScriptData() )
scriptDict['lExileData'] = iNewValue
gc.getGame().setScriptData( pickle.dumps(scriptDict) )

I'm not sure what an EOFError is, how do I fix it?

embryodead · Apr 2, 2010

You have a bad luck for vague, uninformative errors. EOFError means End Of File Error that pickle gets, but that in my experience doesn't have anything to do with what it says. I had this error when I was caching the years/turns in a dictionary, even though I didn't use pickle at all.

Only thing that comes to mind is that getExileData should not be called at the beginning of the game. Maybee the human player was killed/collapsed pre-maturely. Make sure you converted all the years/turns in Consts.py, and that they match the years from CvRhyes, and that at least all the functions using tBirth have been converted to getTurnForYear...

antaine · Apr 2, 2010

if this helps, this is the turn by turn python log

Spoiler :

antaine · Apr 2, 2010

in CvRhyes, you have the following:

int startingTurn[27] = {0, 0, 0, 0, 50, 84, 86, 90, 97, 121, 145, 177, 183, 187, 193, 196, 203, 205, 207, 213, 220, 234, 236, 240, 241, 249, 346}; //332 for 1733

// edead: replaces the above to allow different game speeds
int startingTurnYear[27] = {-3000, -3000, -3000, -3000, -1600, -850, -820, -753, -660, -300, 60, 551, 620, 655, 720, 750, 820, 840, 860, 920, 989, 1130, 1150, 1190, 1200, 1280, 1775};

should the first int starting turn etc be commented out? I've compiled it both ways, and it makes no difference to the eoferror, but it struck me as odd.

In any event, it seems that I had forgotten to merge the lines into culturelevelinfos.xml...once I fixed that, that seems to have fixed the problem.

Shame that I can't seem to get the CAR mod portions to work, as the game runs somewhat slower through CrossOver than I imagine it would normally, but I could be wrong about that...maybe someday I'll get that stuff in there...

embryodead · Apr 2, 2010

Yeah, from the log you posted it's clear it was because of culture levels (cultural victory on 2nd turn).

The first array in CvRhyes is unused and not commented out because I'm lazy

The speed loss in RFCMarathon is negligible - through tests I wasn't able to tell if it is slower at all. The CAR mod was added just because it was available and any speed gain is good.

antaine · Apr 2, 2010

Thanks...I won't bother with CAR then...too much to do in real life than debug for negligible speed increases

I got it running, for anyone following this thread...the most updated version appears in the first post of this thread

http://forums.civfanatics.com/showthread.php?t=359542

the marathon/epic version is v2...still a few cosmetic, "housecleaning" things to do with it, but it's basically done unless a major bug/imbalance presents itself that wasn't present in the regular version.

Regarding speed, however, does enabling debug in the ini slow things down appreciably?

embryodead · Apr 2, 2010

Yes, both logging enabled in the INI and using debug DLL which you did, decrease the game speed greatly.

Probably something simple, but I'm SDK stupid

antaine

King

General Tso

Panzer General

antaine

King

embryodead

Caliph

antaine

King

antaine

King

embryodead

Caliph

antaine

King

antaine

King

embryodead

Caliph

antaine

King

embryodead

Caliph

Similar threads