[Crash] Any way to know what caused this fatal error?

[_Guhun_]

So, I was playing a map for a few hours. There were only 2 players, me and another. I went to get some water, when I got back, the game had crashed.

The other played said he did nothing, not even move a unit. Any way to know what caused this error? I have no idea where to start.

 

[Dr Super Good]

The crash message is too ambiguous without the build symbols and source code of WC3 to even guess what might have been the cause. The crash was caused by an access violations and since none of the registers point towards 0 it most likely the result of accessing a deleted object, or caused by a buffer overrun.

Some buffer overrun errors can be caused by triggers leaking too many objects. Eventually leaks can cause WC3 to run out of some finite resource and crash.

 

[_Guhun_]

The crash message is too ambiguous without the build symbols and source code of WC3 to even guess what might have been the cause. The crash was caused by an access violations and since none of the registers point towards 0 it most likely the result of accessing a deleted object, or caused by a buffer overrun.

Some buffer overrun errors can be caused by triggers leaking too many objects. Eventually leaks can cause WC3 to run out of some finite resource and crash.

Hmmmm, I will check for any object leaks. Do you know if trying to load data from hashtable adresses that have not been writen to can cause this error? It might have been that, as well.

 

[TempMail]

Actually wrong. Crashlogs used to be useful until Blizz fucked them up with patches. There COULD be useful info, but since 1.28 - its literally pure trash.

 

[Dr Super Good]

Do you know if trying to load data from hashtable adresses that have not been writen to can cause this error?

They used to cause such an error. But they might have patched that.

Actually wrong. Crashlogs used to be useful until Blizz fucked them up with patches. There COULD be useful info, but since 1.28 - its literally pure trash.

99% of the time there were not helpful then, and chances are the same applies now. Without the source code and build symbols most of the stuff shown is meaningless.

There used to be a select few, highly specific, object type related crashes that could be spotted from the old crash log format that cannot be spotted from the new crash log format. All other crashes were as abstract then as they are now.

 

[TempMail]

between 99% and 100% faulty logs I prefer 1st kind, thanks