Hyborem crash bug reports/search for cause of bug

[smjjames]

ok, I'm getting a crash here, however I'm not sure if this is because the patch isn't working for this save (it's from an old save, not a new game) or it's from something else. It crashes on end turn. I had already done a victory as the Lucichurip (sp?), so maybe that has something to do with it, or maybe it doesn't like spawning on tundra, I have no idea.

I'll try a new game and place a settler on tundra to see whether it happens there as well. It's possible the patch doesn't work for old saves.

 

[Kael]

cool. So, what did you do with the code that stopped it from crashing and still allowed the flames to appear? I'm just really curious here since this is such a strange crash bug for me. Which now seems to be 100% fixed for me.

This all started with the desire to be able to disable Hell Terrain. In 0.30 I had one python function called "doMapTurn" that would go through each plot on the map each turn and do various things.

When I added the ability to disable Hell Terrain I broke that function into 2 parts. The part that handled Hell Terrain was called doHellTerrain and if you had Hell Terrain disabled it was completly skipped. The rest of the doMapTurn functions were moved into the SDK for performance reasons. The end result was that if you disabled hell terrain you didnt need to run a python script that touched every plot every turn, which was nice.

Meanwhile in the SDK I had to add functions to makeup for the non-Hell Terrain things. Two of those were the spread of flames (so that they spread to surrounding tiles if their is flammable material there) and the random expiration of flames (so they eventually burn themselves out). Calbrenar isolated it down last night to the expiration process. If flames never expire we never got a crash.

So we tried out a bunch more tests and figured out that expiration in the forests and typical situations never seemed to cause crashes, it was only expiration of fires on burning sands. I still don't know why its a problem (especially since it only effects Vista). If I had to guess I would say that the doHellTerrain python script is adding flames to the tile and the SDK is randomly removing flames from the tile and those two process are conflicting with each other (specifically one is in the process of generating the flames and animation when the other orders it down).

The change in this latest test patch was to never expire flames on burning sands tiles.

It was a hard problem to diagnose mostly because I couldn't recreate it. Calbrenar was up to after 2am with me last night running tests, reporting results and troubleshooting on his end while I went through code on mine. Without that this issues would have taken a huge amount of back and forth through the forum to get resolved. But now I'll be able to post patch "e" tonight with this fix included.

 

[Kael]

ok, I'm getting a crash here, however I'm not sure if this is because the patch isn't working for this save (it's from an old save, not a new game) or it's from something else. It crashes on end turn. I had already done a victory as the Lucichurip (sp?), so maybe that has something to do with it, or maybe it doesn't like spawning on tundra, I have no idea.

I'll try a new game and place a settler on tundra to see whether it happens there as well. It's possible the patch doesn't work for old saves.

Thats definitly a different issue, it crashes for me too. I'll check it out.

 

[smjjames]

I just tried a new game and it does not crash, I think there is either a conflict between old saves where it still wants to not wait at end of turn vs the new patch where it looks like you added the wait at end of turn fix.

So I think it's that it does not work well with old saves, but with new games it's just fine and I just checked it out.

 

[Kael]

I just tried a new game and it does not crash, I think there is either a conflict between old saves where it still wants to not wait at end of turn vs the new patch where it looks like you added the wait at end of turn fix.

So I think it's that it does not work well with old saves, but with new games it's just fine and I just checked it out.

Maybe, thats actually a patch "e" DLL so it may be a little off. I'll have a full patch "e" up in a little bit.

You save game is weird, I assume you were playing as the Luchuirp and you scraped all your units to make it easy to conquer your former empire with Hyborem? Sorry, I mean "sacrificed to the coming fiery lord". But for some reason it triggered the religious victory condition (according to the map there is no way the Luchurip should qualify for that). Your crash appears to be a problem handlign the bogus victory condition.

 

[smjjames]

Maybe, thats actually a patch "e" DLL so it may be a little off. I'll have a full patch "e" up in a little bit.

You save game is weird, I assume you were playing as the Luchuirp and you scraped all your units to make it easy to conquer your former empire with Hyborem? Sorry, I mean "sacrificed to the coming fiery lord". But for some reason it triggered the religious victory condition (according to the map there is no way the Luchurip should qualify for that). Your crash appears to be a problem handlign the bogus victory condition.

Yea thats what I was planning on doing, removing all units so that I could remove them from the game easily lol. I had already triggered the religious victory condition a while ago on turn 197 or so. Although there seems to be something wierd with that victory condition and I'm not sure if it is triggering right because I had started a turn with 89% religious spread and the next turn when I had spread OO to a few more cities it triggered.

Pretty wierd that it was triggering the same win again. The azer was a random event though, but it didn't matter since I wanted to conquer my old empire quickly anyways.

 

[Blackmantle]

I have to say again that my rig is not running vista but win2k. But very likely the fix will work for 2k as well i guess. So no need to make much more fuss.
(also had a smaller ctd issue with regular forest fires, but extinguishing them fixed it)

Thanks for the greath support in spite of the problems of pinning the issue down.

Also thanks to Calbrenar for your effort of finding the cause of the bug.

 

[Yaotzin]

I have to say again that my rig is not running vista but win2k. But very likely the fix will work for 2k as well i guess. So no need to make much more fuss.
(also had a smaller ctd issue with regular forest fires, but extinguishing them fixed it)

Thanks for the greath support in spite of the problems of pinning the issue down.

Yeah, it's not Vista only, I've never run anything but XP and I was getting it. Pretty strange. Fixed anyway so all is well

 

[smjjames]

I have to say again that my rig is not running vista but win2k. But very likely the fix will work for 2k as well i guess. So no need to make much more fuss.
(also had a smaller ctd issue with regular forest fires, but extinguishing them fixed it)

Thanks for the greath support in spite of the problems of pinning the issue down.

Also thanks to Calbrenar for your effort of finding the cause of the bug.

and me for bieng so persistent about it

also Thanks to Kael for fixing this issue