Notes |
|
|
|
|
|
Autosave consists of three files. |
|
|
|
|
|
|
In my first attempt it is not reproductible. Did you use any special strategy? |
|
|
|
No special strategy, i will upload logs later tonight, not at pc right now. |
|
|
|
it seems to happen only aggains stronghold, i captured a necroplis without problem |
|
|
|
maybe it's escape tunnnel support bug? |
|
|
|
i don't know, does the ai builds anything? or what do you mean? |
|
|
(0001915)
|
Ivan
|
2011-08-05 23:12
|
|
Probably one of multithreading bugs we have from time to time. It looks that on Win it is possible to delete locked mutex but on Linux this results in assertion failure.
I'll try to reproduce this myself.
lssjbroll, can you create core dump or get stacktrace? This may make things more simple. |
|
|
|
i will as soon as i find out how to compile vcmi with debug symbols. any pointers? |
|
|
|
is that what you want? or do you need something else? |
|
|
(0001921)
|
Ivan
|
2011-08-06 18:45
|
|
That's what I need. Don't see how to fix this though.
It looks that there is still an active stack in the battle. Maybe last creature got morale?
Crash happens in CBattleInterface::givenCommand destructor.
The only other place in code where givenCommand used is CPlayerInterface::activeStack - called when a stack got turn |
|
|
|
only if morale is happening every time? maybe the game expects another stack to heve turn, but that stack doesn't exist, but again i'm no dev so this kinda goes over me.
if you need anything else to try, drop a line :) |
|
|
(0001932)
|
Tow
|
2011-08-25 15:50
|
|
Fixed in r2307.
Nasty, little bug. First SetActiveStack packet was duplicated. |
|