MegaGlest Forum
Archives (read only) => Glest Advanced Engine => Bug reports => Topic started by: Psychedelic_hands on 11 January 2011, 14:21:50
-
Crash
Version: Advanced Engine 0.3.2
Time: Wed Jan 12 01:16:33 2011
Description: Access violation (Reading address: 0x00000000)
Call Stack:
Frame Code address
0x00B4F3D8 0x0121E1F1
0x00B4F3F0 0x0126D505
0x00B4F400 0x0126BAA5
0x00B4F434 0x01274A85
0x00B4F45C 0x0133DF82
0x00B4F4A0 0x013973B5
0x00B4F4CC 0x762CF8D2 GetMessageW+0x93
0x00B4F544 0x762CF794 GetWindowLongW+0x115
0x00B4F574 0x762D06F6 CallWindowProcW+0x75
0x00B4F594 0x762D069C CallWindowProcW+0x1b
0x00B4F5D0 0x6709834F wglSwapBuffers+0x2ed
0x00B4F5FC 0x762CF8D2 GetMessageW+0x93
0x00B4F674 0x762CF794 GetWindowLongW+0x115
0x00B4F6D8 0x762D0008 PeekMessageW+0x269
0x00B4F6E8 0x762C5A2C DispatchMessageA+0xf
0x00B4F774 0x01396C15
0x00B4FCEC 0x0127D857
0x00B4FD3C 0x0127DA2D
0x00B4FD48 0x75AD4911 BaseThreadInitThunk+0x12
0x00B4FD88 0x7724E4B6 RtlInitializeExceptionChain+0x63
0x00B4FDA0 0x7724E489 RtlInitializeExceptionChain+0x36
=======================
I wasn't doing anything out of ordinary.
Btw I've posted a few of these threads and no one has replied, is anyone fixing these? or am I reporting it wrong? Sorry if I seem demanding which is not my intention, I just want to make sure these bugs get fixed.
-
I've had a look at your posts with no replies. The call stacks don't show much and you aren't able to reproduce. So apart from the information that there is an access violation somewhere in the code we haven't got much to go by (unless there's something I'm not seeing). Do you have the symbols file (.pdb) next to your exe?
As a general rule make sure it has occured a few times so you can give us some indication of how to reproduce.
Thanks for spending the time to report. It's just hard to fix with the information given.
-
Sorry, I'll try to reproduce it before next time.
I do have a .pdb file? Am I supposed to do anything with it?
And thanks for replying :) And sorry I'm not giving enough info.
-
I've been looking into why the stack traces are so barren... someone messed up when they built the 0.3.2 windows release :look: While the linker was generating a pdb, the compiler wasn't generating info for it, and so none of our symbols were in it. This is why the Windows functions are correctly displayed, but not the 'interesting bits' that we want.
I'm going to upload replacement exe, pdb & map files for 0.3.2, and I'm just now trying to coerce CMake into doing what I want, so the correct compile & link flags will be 'baked in' from now, and I wont be able to screw it up again ;)
For this one, I tried to look up the code address in a map file generated from tag 0.3.2, but I'd messed with flags by hand when building 0.3.2, and the address didn't make sense.
However, looking at the windows functions further up the stack, this looks like it was probably in the GUI code somewhere.
@Psychedelic_hands:
If you remember when this was, can you try to locate the crash screen-shot (they a date & time stamped now).