MegaGlest Forum

Archives (read only) => Glest Advanced Engine => Bug reports => Topic started by: Mr War on 12 January 2011, 21:07:15

Title: Crash report - GAE.0.3.2 - Access violation
Post by: Mr War on 12 January 2011, 21:07:15

Whilst playing Japanese faction Mod

Crash
Version: Advanced Engine 0.3.2
Time: Tue Jan 11 19:35:45 2011
Description: Access violation (Writing address: 0x00000000)

Call Stack:
  Frame       Code address
  0x00000000  0x03EE6678 devProcessAttach+0x12a98

(http://www.freeimagehosting.net/uploads/4f23f5b354.jpg) (http://www.freeimagehosting.net/)

Title: Re: Crash report - GAE.0.3.2 - Access violation
Post by: hailstone on 13 January 2011, 00:11:26

I played Japanese faction 0.75a for about an hour on latest dev build and had no crash. Can you reproduce this? Do you have 0.3.2 that silnarm reuploaded to fix the missing debug symbols?

Title: Re: Crash report - GAE.0.3.2 - Access violation
Post by: Yggdrasil on 13 January 2011, 12:06:35

Yeah, please get the new binary pack:
https://forum.megaglest.org/index.php?topic=6283.msg65471#msg65471

and try to get it crashing again. That way we get a better stack trace and can fix this problem quicker.

Title: Same crash with improved info
Post by: Mr War on 14 January 2011, 23:25:07

Crash
Version: Advanced Engine 0.3.2
Time: Sat Jan 15 00:22:06 2011
Description: Access violation (Writing address: 0x00000000)

Call Stack:
  Frame       Code address
  0x017AF7B8  0x03E86678 devProcessAttach+0x12a98
  0x017AFF34  0x03F92C0B devProcessAttach+0x11f02b
  0x017AFF84  0x00AE8A0D
  0x017AFFD0  0x7743D0E9 BaseThreadInitThunk+0x12
  0x017AFFD4  0x7782198E RtlInitializeExceptionChain+0x36
  0x017AFFD8  0x00C61DEA PHYSFS_writeUBE64+0x23a9a
  0x017AFFDC  0x7FFDA000

=======================

Title: Re: Crash report - GAE.0.3.2 - Access violation
Post by: hailstone on 15 January 2011, 00:36:35

Looking at PHYSFS_write(file, pbase(), 1, size); in OFileStreambuf::overflow isn't the third parameter the size?

Title: Re: Crash report - GAE.0.3.2 - Access violation
Post by: silnarm on 15 January 2011, 01:32:27

Quote from: hailstone on 15 January 2011, 00:36:35

Looking at PHYSFS_write(file, pbase(), 1, size); in OFileStreambuf::overflow isn't the third parameter the size?

Yes, but that is fine. In this case the objects of interest are bytes, and we want to read 'size' of them.
Ultimately it shouldn't really matter, the operation should write size * count bytes, so even if they are the 'wrong way around', it should still work.

This is a rather odd one, the 'ProcessAttach' would suggest it is a DLL problem, but if this is happening in game, then they all would have been loaded long before . . .

Title: Re: Crash report - GAE.0.3.2 - Access violation
Post by: will on 15 January 2011, 12:26:36

Should there be any reading or writing whilst game is playing?

Do the entry points get executed each time a thread is created, or?

(I guess this is windows, so I have only vague memory on that point)

Title: Re: Crash report - GAE.0.3.2 - Access violation
Post by: Omega on 15 January 2011, 23:08:05

Quote from: will on 15 January 2011, 12:26:36

Should there be any reading or writing whilst game is playing?

I would assume there is, since the AI logs, etc are written.

Title: Re: Crash report - GAE.0.3.2 - Access violation
Post by: silnarm on 20 January 2011, 12:19:02

Found this little gem while perusing the Japanese faction topic...

Quote from: Mr War on 12 January 2011, 22:59:24

Windows. I got the same crash again - seems to be when Samurai appears on screen