Single Player bugs and crashes - After the 13th July 2013

[demmalition1]

Do you have Viewports "on?"

Errr..., no. Should I? How do I do so? It's a "large" sized map, but would it help? Switch to a smaller map size?

Also, my previous save loads, but not the recent one I linked. Should I submit it as well for testing purposes?

 

[demmalition1]

Here's the previous save of the 2nd game before the crash (the 2nd one, not the 1st one with the old .dll file). The .dll in this is updated, and this save works, but not the one I made in comment #359.

 

[Snofru1]

The graphics of the Hero "Lus Alves de Lima e Silva" is still hovering over the ground and facing towards the sky (it is a mounted unit and I think even in the Galactig Era flying horse aren´t an option...

 

[Zengdar]

Is this V31, or V31 with the patch DLL? If it's V31 (original version) download the patch DLL from the patches thread and it will almost certainly resolve your problem.

I downloaded the mod it like 5 days ago. I might consider going with SVN I guess. Anyway, thanks for the tip, gonna try it after work.

 

[Koshling]

Rev 5890
I changed Nebuchadnezzar's Revolutionary to Megalomaniac and CTD occurred. Was it really the reason of this CTD? attached save and minidump.

And the flag of Chippewa is just white? (Screenshot)

No - the minidump clearly points to a new piece of code (added yesterday) that is related to city founding. I can't see immediately what's wrong woth it, but your save crashes for me too, so I'll have no trouble figuring it out when I have some C2C debug time later today...

Hey, I've given up on trying to resurrect my earlier game and started a new one after installing the DLL patch. Works fine, but now my new game won't load... crashes quite often (every 5-10 turns) and it's quite frankly annoying. Now it's back to the "initializing" bug again while loading with corrupt saves despite the patch.

Should I do a fresh install of BTS (Steam ed.)? A fresh DL of the mod and patch? I'm running out of ideas here.

Your minidump is 0 bytes, which is USUALLY a symptom of running out of memory for some reason. I'll see if I can reproduce it later today from your save, but I suspect I won't be able to given the above. Can you try with viewports on and see if that makes any difference?

 

[Koshling]

Here's the previous save of the 2nd game before the crash (the 2nd one, not the 1st one with the old .dll file). The .dll in this is updated, and this save works, but not the one I made in comment #359.

This save loads ok for you? It doesn't for me, so if you can confirm I haven't misunderstood, and it does for you, then I'll spend time analyzing the details...

Also I have another point of confusion - the date on this save is 7090BC but on the other one it's 8725BC, yet you state that this one is from BEFORE the crash (i.e. - earlier), which makes the dates the wrong way around. Are they the same game?

Edit - and both saves were SAVED by the patched version right?

Edit 2 - actually, I can check from the save header exactly what it was saved by, and the answer is 5709, which is the v31 original. You must not have overwritten the DLL with the patch correctly. The file CvGameCoreDLL.dll in the 'Assets' folder in your mod folder ('C2C Release' in your case) should be the one from the patches thread, dated July 26th, size 8508K. Did you possible copy it to the wrong folder (like say the mod folder itself rather than Assets?)

Anyway, based on the evidence that:

1) The save header says it was saved by rev 5709 (v31 original); and
2) The save corruption exactly matches the known one that can occur with v31 original

I would say that the patch is not applied.

Edit 3 - to tell for certain what is running hover over the flag in-game. The second bullet point gives the revision number of the DLL. It will say something like:
Caveman2Cosmos build XXXX (game saved from build YYYY)
where XXXX is the revision you are running now, and YYYY is the revision that was running when the save you loaded was created

 

[fox456]

Could someone please tell me what file(s) could be causing the following problem? When I reach the Civilization Selection Screen to set up a game, I get a CTD immediately when clicking any civilization to play.

I have been able to play this version (v31) already, but now it is not working. An older C2C version I still have (v27) allows me to choose a civ. Also, I reinstalled v31 and still have the problem.

 

[Zengdar]

Is this V31, or V31 with the patch DLL? If it's V31 (original version) download the patch DLL from the patches thread and it will almost certainly resolve your problem.

Ok, I just put the patch DLL in Beyond the Sword\Mods\C2C Release\Assets, my game is still no loading, just as before.

Here are again the minidump, the save you have it already.

Thx for help

 

[Talin2009]

Repeatable CTD (Immense map, viewports on, Reneissance era) at end of turn

just updated SVN (before that last update was at 28-07-13). CTD is still there.

"Your submission could not be processed because a security token was missing." when i try to upload zip file again!!! grrrrrrrrrrrrr. No idea what causing this.

Uploaded savegame here: http://zalil.ru/34664533 (download begins automatically in 10 seconds)

p.s. I also noticed small glitch - when i update SVN and load savegame from previous versions, any buildings that have been added ingame and should become obsolete at my ingame date become buildable again (you can see in this game i can build Rock Art building for example).

 

[nanovirus]

Ah, if you're running out of memory that could explain crashes and hangs. Because BTS is a 32-bit program it cannot use large quantities of physical memory, so if you have 4G you're at the most that can be used anyway. Large maps generally exceed this unless you turn viewports on. Try with viewports on (BUG settings) - they are slightly annoying to use at times, but they should allow you to continue...

Hi

Thanks for the tip

Tried running viewpoint but same result. When clicking end turn the game starts acting abnormal. The memory usage is 1 GB at start of end turn then i can see it going to 2.7 GB and eventually crashing. Takes about 2 minute.

When the game reaches 2.7 Gb of memory my system is at 100%.
Still it strange the game acts this way.

 

[Koshling]

Ok, I just put the patch DLL in Beyond the Sword\Mods\C2C Release\Assets, my game is still no loading, just as before.

Here are again the minidump, the save you have it already.

Thx for help

That's right - it's the SAVE file that is broken due to the version that SAVED it being the unpatched version - the bug was in the save routine. You'll have to play forward from an earlier save I'm afraid. Can you also verify what revision (build number) your installation claims to be now when you hover over the flag please. I'm not sure off hand what rev the latest patch was, but it would be useful to know for future reference.

 

[Koshling]

Hi

Thanks for the tip

Tried running viewpoint but same result. When clicking end turn the game starts acting abnormal. The memory usage is 1 GB at start of end turn then i can see it going to 2.7 GB and eventually crashing. Takes about 2 minute.

When the game reaches 2.7 Gb of memory my system is at 100%.
Still it strange the game acts this way.

Seems like you have a definite bug, which makes it odd that I cannot reproduce it. Could you turn AI logging on (BBAI logs to level 3 in the BUG logging screen, and logging enabled in your INI file) and after the next crash take a look at BBAI.log (in the logs folder of your BTS folder inside 'My Games'. It might be that the end of that log indicates some repeated attempt to do something that could explain things.

Failing that (if you're up for it) I could build you a special version of the DLL that looks for memory usage going over some threshold (say 2G) and force-crashes the game at the next attempted allocation - that would force a useful minidump that I might able to diagnose things from...let me know if you'd be prepared to try this.

 

[Koshling]

Repeatable CTD (Immense map, viewports on, Reneissance era) at end of turn

just updated SVN (before that last update was at 28-07-13). CTD is still there.

"Your submission could not be processed because a security token was missing." when i try to upload zip file again!!! grrrrrrrrrrrrr. No idea what causing this.

Uploaded savegame here: http://zalil.ru/34664533 (download begins automatically in 10 seconds)

p.s. I also noticed small glitch - when i update SVN and load savegame from previous versions, any buildings that have been added ingame and should become obsolete at my ingame date become buildable again (you can see in this game i can build Rock Art building for example).

The minidump lead me directly to this - a fix will be on the SVN later today. Meanwhile, a workaround is to do a recalc, which clears the underlying condition that leads to it crashing.

 

[dekk]

Hi.

first off: many thanks for this awesome mod

sadly, I recently ran into a repeatable CTD. It happens when other civs are doing their turn. I attached a 7z file including minidump and savegame.

hopefully you guys are able to fix the bug so I can continue this match.

cheers
dekk

 

[strategyonly]

Hi.

first off: many thanks for this awesome mod

sadly, I recently ran into a repeatable CTD. It happens when other civs are doing their turn. I attached a 7z file including minidump and savegame.

hopefully you guys are able to fix the bug so I can continue this match.

cheers
dekk

Have you looked in the patch thread, and installed it?? If not pls do so, then let us know. thx. . . SO

http://forums.civfanatics.com/showthread.php?t=434712

btw Welcome to CFC and trying this Great Mod, C2C

 

[nanovirus]

Seems like you have a definite bug, which makes it odd that I cannot reproduce it. Could you turn AI logging on (BBAI logs to level 3 in the BUG logging screen, and logging enabled in your INI file) and after the next crash take a look at BBAI.log (in the logs folder of your BTS folder inside 'My Games'. It might be that the end of that log indicates some repeated attempt to do something that could explain things.

Failing that (if you're up for it) I could build you a special version of the DLL that looks for memory usage going over some threshold (say 2G) and force-crashes the game at the next attempted allocation - that would force a useful minidump that I might able to diagnose things from...let me know if you'd be prepared to try this.

The minidump lead me directly to this - a fix will be on the SVN later today. Meanwhile, a workaround is to do a recalc, which clears the underlying condition that leads to it crashing.

okay

I don't notice anything strange in the log but yust te be shure i'm uploading them.
There is a RandomLogger log that is 1.2 GB that i'm not including don't know if that file size is normal or not. Can't even open it whit notepad++.

 

[Koshling]

okay

I don't notice anything strange in the log but yust te be shure i'm uploading them.
There is a RandomLogger log that is 1.2 GB that i'm not including don't know if that file size is normal or not. Can't even open it whit notepad++.

Nothing unusual in the log as you say. I'll build you a test DLL that will deliberately crash (producing a hopefully useful minidump) when memory allocations start ramping up dramatically. Will be posted here later today...

 

[demmalition1]

Whoops, wrong file!

To clarify:

---Game 2, post-patch---

-(attached) 9925 BC @ 8/5/2013, 10:36pm file is loading. Before game 2 crash with the dll patch.

-(comment # 359) 8725 BC @ 8/5/2013, 10:44pm file isn't loading. Game 2 post-patch.

---Game 1, pre-patch---

-(comment #362) 7090 BC @ 8/4/2013, 3:12am isn't working. It's from game 1 pre-patch and I accidentally uploaded it as I had filter:title instead of filter:date.

-(comment #328) 7050 BC @ 8/4/2013, 3:17 file won't work, since game 1 pre-patch.

EDIT:
I checked the dll file, it's up to date. Just the wrong file was uploaded in the previous comment it seems. I'll try a game 3 tonight after work with viewports enabled to try and see if anything is wrong. Thanks for all the help!

EDIT 2:
Should I just go straight to SVN beta?

 

[Koshling]

Nothing unusual in the log as you say. I'll build you a test DLL that will deliberately crash (producing a hopefully useful minidump) when memory allocations start ramping up dramatically. Will be posted here later today...

@Nanovirus - attached is a test DLL that will deliberately force a (semi-controlled) crash when memory allocation passes 2G. This should produce a minidump that has useful context that should help with the diagnosis. Run the turn (should crash) and post the minidump that results please, and we'll take it from there...

NOBODY EXCEPT NANOVIRUS SHOULD USE THIS DLL!

BTW - for me your game takes about 1.7G on load, but drops back to about 1.6G after I autoplay the next 10 turns.

 

[Koshling]

Whoops, wrong file!

To clarify:

---Game 2, post-patch---

-(attached) 9925 BC @ 8/5/2013, 10:36pm file is loading. Before game 2 crash with the dll patch.

-(comment # 359) 8725 BC @ 8/5/2013, 10:44pm file isn't loading. Game 2 post-patch.

---Game 1, pre-patch---

-(comment #362) 7090 BC @ 8/4/2013, 3:12am isn't working. It's from game 1 pre-patch and I accidentally uploaded it as I had filter:title instead of filter:date.

-(comment #328) 7050 BC @ 8/4/2013, 3:17 file won't work, since game 1 pre-patch.

EDIT:
I checked the dll file, it's up to date. Just the wrong file was uploaded in the previous comment it seems. I'll try a game 3 tonight after work with viewports enabled to try and see if anything is wrong. Thanks for all the help!

EDIT 2:
Should I just go straight to SVN beta?

Ok, I've checked the save headers for all three saves now, and I hate to sound like a broken record, but are you SURE the DLL is correctly installed? Please check what the text says in the flag hover to verify the exact build number.

All 3 saves are claiming they were saved from build 5709, which is the V31 original.

The image attached is what I get when I load your latest save in my latest build (which is later than the patch, so that number should be different to what you'll see with the patch). The 5709 it indicates as the 'saved by' number is the build number of the DLL that was used to save the file. The other is the build number of the version currently running (if you have the patch correctly installed this should NOT be 5709).