Allegro.cc - Online Community

Allegro.cc Forums » The Depot » PixelBots Combat Prototype 1

This thread is locked; no one can reply to it. rss feed Print
 1   2   3 
PixelBots Combat Prototype 1
Edgar Reynaldo
Major Reynaldo
May 2007
avatar

It crashed again. The dump file is 110 MB and I can't upload that from here, and it would still take a while anywhere else. And DEP closed the program too.

This is as far as the console got :

Start
system.Init()

And this is what the problem details said :

Problem signature:
  Problem Event Name:	BEX
  Application Name:	PixelBots.exe
  Application Version:	0.0.0.0
  Application Timestamp:	51ee204f
  Fault Module Name:	StackHash_d34b
  Fault Module Version:	0.0.0.0
  Fault Module Timestamp:	00000000
  Exception Offset:	060d2ac0
  Exception Code:	c0000005
  Exception Data:	00000008
  OS Version:	6.0.6001.2.1.0.768.3
  Locale ID:	1033
  Additional Information 1:	d34b
  Additional Information 2:	20752c25712bf3f7bc825928e900516d
  Additional Information 3:	f29a
  Additional Information 4:	c436154ffc56d491095fa90afd9656e7

Read our privacy statement:
  http://go.microsoft.com/fwlink/?linkid=50163&clcid=0x0409
 

And it didn't even get as far as Crash Reporter to run, so there are no logs.

Edit-

http://answers.microsoft.com/en-us/windows/forum/windows_other-windows_programs/problem-event-name-bex-error-message/cf5baf73-0877-4070-abfb-a2c3a17a9e10

http://technet.microsoft.com/en-us/library/cc738483%28WS.10%29.aspx

Max Savenkov
Member #4,613
May 2004
avatar

Audric said:

Are you going to keep the top-view sight ? It may be an adequate view for a plane or car, but not so good to represent a combat robot (on wheels, tank treads, biped, quadruped etc).

Top-down view is the simplest one for graphics. It will definitely stay this way during prototyping, because I can create tiles & robots for it myself. For the game itself, I was thinking about JRPG-like view. Of course, the best option would be isometric view, but I simply can't create necessary art myself, and I don't have an artist who could help me for free (and I'm not going to sink money into this project until I'm pretty sure I can finish it).

It crashed again. The dump file is 110 MB and I can't upload that from here, and it would still take a while anywhere else. And DEP closed the program too.

So, it crashes in render initialization (as was expected, more or less). I have to ask (should have done it earlier!): do any OTHER Allegro projects that are using OpenGL run normally on your machine? I remembered a case now, with my own Netbook a while ago, which simply could not initialize OpenGL when Allegro asked for it, and crashed inside driver. Because Intel's drivers for Vista & Win7 are so bad I want to weep over them (namely, they do not seem to support OpenGL 1.2), even though XP and Linux drivers are OK.

If this is the case, then we have a problem here... I can't give you fullscreen DirectX version, because Allegro crashes, and OpenGL version won't work too. If so, I think I will do two things: first, I would spend some time trying to isolate that crash and check if it is my mistake, or an error in Allegro. Then, if this error can't be fixed quickly, I will proceed with my current plan: to make resizeable GUI (instead of pile of hacks and hardcode which was used in this prototype) which will work under any resolution without scaling. This may take some time, however, about a week at least.

EDIT: A-ha! It seems like I have found the culprit. Things stop crashing when I do not initialize TTF addon. I will try to recompile Allegro with a newer version of FreeType.

EDIT2: Recompiling did not help, but the other thing did. It seems I have to call al_init_ttf_addon BEFORE creating display, but only in fullscreen modes with non-desktop resolution. This isn't mentioned anywhere in docs, I believe, and I think it's a bug, but for now moving initialization code fixed fullscreen mode in DirectX. I have to run to work right now, but I'll have another version for you to try soon.

Audric
Member #907
January 2001

I've played around with a more symbolic representation, in side view. This would need the rotations to be disabled.
607841607842607843607844607845607846

Mark Oates
Member #1,146
March 2001
avatar

maybe add a little sense of depth

607847

could be cool :)

--
Visit CLUBCATT.com for cat shirts, cat mugs, puzzles, art and more <-- coupon code ALLEGRO4LIFE at checkout and get $3 off any order of 3 or more items!

AllegroFlareAllegroFlare DocsAllegroFlare GitHub

Edgar Reynaldo
Major Reynaldo
May 2007
avatar

Max Savenkov
Member #4,613
May 2004
avatar

Audric, Mark Cool! That's closer to what I was thinking about for final version. I was also planning to use 3 sprites for each robot: side view which could be flipped, front view and back view. This could prove to be problematic, though, because it may be hard to tell some bots from each other in front/back view... But using only side view would look a bit wrong, even with added depth, I think.

Edgar, I don't think there is a REAL BEX issue here. I run Windows 7 with the same protection mechanism turned on, and it does not throw a warning, and nobody else has reported such issue. Anyway, here's a scan result from VirusTotal - clear.

I think it's really a driver issue, and that's why you are seeing BEX error. Anyway, I have prepared a new test version for you to try, which is now again uses DirectX instead of OpenGL and should therefore work even on Intel cards with bad drivers. It also handles different resolutions slightly better, or so I hope.

Here it is: pixelbots_proto01_debug.zip

ALL: Now that Shadowrun Returns is out my productivity is at risk of dropping further. Hey, come on, I'm just researching modern tactic combat games!

Chris Katko
Member #1,881
January 2002
avatar

Ah, that seems to be a problem. UI and window size is really not adaptable. 1024x768 is a minimum :( Let's see if I can hack resolution selection and scaling into that code quickly...

FYI, if you want people to play your game, you shouldn't expect more than 1024x600 because that is a very common netbook resolution. My Asus Eee PC, for instance, causes me great distress when forms are cut off (most commonly in Linux.)

As always, congratulations on your progress so far, and good luck!

-----sig:
“Programs should be written for people to read, and only incidentally for machines to execute.” - Structure and Interpretation of Computer Programs
"Political Correctness is fascism disguised as manners" --George Carlin

Max Savenkov
Member #4,613
May 2004
avatar

Chris, yes, I'm aware of that, but I decided to cut some corners for prototype. May be too many :)

Edgar Reynaldo
Major Reynaldo
May 2007
avatar

This time your game opened the options window and I went to start a fullscreen game and it crashed. You didn't include debugging symbols so I can't debug it. And, Task Manager couldn't even kill it. I had to kill it with the parent process running it. Then I ran it in windowed screen and it worked, although I got my butt handed to me by a 200hp warrior I could only do 3 damage to.

Max Savenkov
Member #4,613
May 2004
avatar

Symbols are present in archive, "bin\PixelBots.pdb". If they does not work somehow, you can just post numeric addresses from stack of main thread here and I'll try to decode them with .map file.

According to my (admittedly incomplete) statistics, of 8 games recorded 4 were won, including one on Hard difficulty, so "butt handing" part is optional :) I have to admit that battle with boss currently lacks any elegance, though, because he does not have special AI and can't even use abilities.

Edgar Reynaldo
Major Reynaldo
May 2007
avatar

When I run it through WinDbg, it immediately crashes with a BEX error again. And if I run it outside the debugger in fullscreen it crashes. I don't really know how to use WinDbg, but here is the output :

CommandLine: C:\Games\PixelBots_pre_alpha-MaxSavenkov\bin\PixelBots.exe
Symbol search path is: *** Invalid ***
****************************************************************************
* Symbol loading may be unreliable without a symbol search path.           *
* Use .symfix to have the debugger choose a symbol path.                   *
* After setting your symbol path, use .reload to refresh symbol locations. *
****************************************************************************
Executable search path is: 
ModLoad: 00270000 002ee000   PixelBots.exe
ModLoad: 77c00000 77d28000   ntdll.dll
ModLoad: 76830000 7690c000   C:\Windows\system32\kernel32.dll
ModLoad: 69470000 6964c000   C:\Games\PixelBots_pre_alpha-MaxSavenkov\bin\allegro_monolith-5.1.dll
ModLoad: 76fc0000 7705d000   C:\Windows\system32\USER32.dll
ModLoad: 76690000 766db000   C:\Windows\system32\GDI32.dll
ModLoad: 76ef0000 76fb6000   C:\Windows\system32\ADVAPI32.dll
ModLoad: 76920000 769e2000   C:\Windows\system32\RPCRT4.dll
ModLoad: 770f0000 77c00000   C:\Windows\system32\SHELL32.dll
ModLoad: 76b40000 76bea000   C:\Windows\system32\msvcrt.dll
ModLoad: 77d80000 77dd8000   C:\Windows\system32\SHLWAPI.dll
ModLoad: 766e0000 76824000   C:\Windows\system32\ole32.dll
ModLoad: 77070000 770e3000   C:\Windows\system32\COMDLG32.dll
ModLoad: 72fc0000 73045000   C:\Windows\WinSxS\x86_microsoft.windows.common-controls_6595b64144ccf1df_5.82.6001.18523_none_886c608850a2f36f\COMCTL32.dll
ModLoad: 74d90000 74dc2000   C:\Windows\system32\WINMM.dll
ModLoad: 76d90000 76e1d000   C:\Windows\system32\OLEAUT32.dll
ModLoad: 74d50000 74d89000   C:\Windows\system32\OLEACC.dll
ModLoad: 763d0000 763d7000   C:\Windows\system32\PSAPI.DLL
ModLoad: 6b190000 6b25b000   C:\Windows\system32\OPENGL32.dll
ModLoad: 6d480000 6d4a3000   C:\Windows\system32\GLU32.dll
ModLoad: 6a400000 6a4e5000   C:\Windows\system32\DDRAW.dll
ModLoad: 747f0000 747f6000   C:\Windows\system32\DCIMAN32.dll
ModLoad: 76470000 765fa000   C:\Windows\system32\SETUPAPI.dll
ModLoad: 73050000 7305c000   C:\Windows\system32\dwmapi.dll
ModLoad: 74ac0000 74c6b000   C:\Windows\WinSxS\x86_microsoft.windows.gdiplus_6595b64144ccf1df_1.0.6001.18551_none_9e7a1850c9c1b3dc\gdiplus.dll
ModLoad: 71640000 716ff000   C:\Windows\system32\MSVCR100.dll
ModLoad: 76d20000 76d4d000   C:\Windows\system32\WS2_32.dll
ModLoad: 76910000 76916000   C:\Windows\system32\NSI.dll
ModLoad: 61b80000 61b98000   C:\Games\PixelBots_pre_alpha-MaxSavenkov\bin\zlib1.dll
ModLoad: 71410000 71479000   C:\Windows\system32\MSVCP100.dll
(ec8.e34): Break instruction exception - code 80000003 (first chance)
eax=00000000 ebx=00000000 ecx=0019f3ec edx=77c596f4 esi=fffffffe edi=00000000
eip=77c47b0e esp=0019f404 ebp=0019f434 iopl=0         nv up ei pl zr na pe nc
cs=001b  ss=0023  ds=0023  es=0023  fs=003b  gs=0000             efl=00000246
*** ERROR: Symbol file could not be found.  Defaulted to export symbols for ntdll.dll - 
ntdll!DbgBreakPoint:
77c47b0e cc              int     3

Max Savenkov
Member #4,613
May 2004
avatar

This output does not contain callstack beyond the last address, which is DbgBreakPoint in ntdll. There are two useful commands you can try in WinDbg:

!analyze -v

will dump a lot of info, some of which may be useful

And you can view callstack by simply typing

k

or by using View->Callstack (there are also windows for switching current thread there).

Edgar Reynaldo
Major Reynaldo
May 2007
avatar

Didn't get much of anything useful out of WinDbg. Couldn't load symbols even when I set it with .sympath c:\blah...

I don't know, but allegro seems to successfully set the video mode to 1280 x 800 fine, but it chokes on 1024x768 fullscreen for some reason. Used ex_display_options and things are screwed up and it gags on the window for a while and keeps spawning new windows... :P

Max Savenkov
Member #4,613
May 2004
avatar

Is 1024x768 present in list of supported modes at all? As to spawning of new windows, it seems to me that Allegro tries to find a workable resolution when requested one fails. That can lead to a prolonged start up and maybe even a crash.

Edgar Reynaldo
Major Reynaldo
May 2007
avatar

ex_monitorinfo says there are two 1024x768 modes supported, and my graphics driver can change the res to 1024x768 no problem.

I tried using A5.1.7 to set 1024x768 fullscreen and while it worked with OpenGL, it took a long time and showed a subsection of my desktop for a while before it cleared the screen. Using DirectX was much faster and cleaner every time, but in the end both set the resolution successfully.

What does your screen setup code look like? This is all the code I used to test fullscreen 1024x768.

#SelectExpand
1 2 3#include <cstdio> 4 5 6#include <allegro5/allegro.h> 7#include <allegro5/allegro_direct3d.h> 8 9int main(int argc , char** argv) { 10 11 12 if (!al_init()) {return 1;} 13// int DRIVER = ALLEGRO_OPENGL; 14 int DRIVER = ALLEGRO_DIRECT3D; 15 16 al_set_new_display_flags(ALLEGRO_FULLSCREEN | DRIVER); 17 ALLEGRO_DISPLAY* display = al_create_display(1024,768); 18 if (!display) {return 2;} 19 20 al_clear_to_color(al_map_rgb(255,255,255)); 21 al_flip_display(); 22 23 al_rest(7.0); 24 25 return 0; 26}

Max Savenkov
Member #4,613
May 2004
avatar

Here's main part of Reinit function that I use to close windowed mode and go to fullscreen:

#SelectExpand
1void AllegroRender5::Reinit( ERenderType type, bool fullscreen, int w, int h, bool vsync ) 2{ 3 al_set_new_display_option( ALLEGRO_VSYNC, vsync ? 1 : 2, ALLEGRO_SUGGEST ); 4 al_set_new_display_flags( ( m_fullScreen ? ALLEGRO_FULLSCREEN : ALLEGRO_WINDOWED ) | ( type == RENDERTYPE_DIRECT3D ? ALLEGRO_DIRECT3D : ALLEGRO_OPENGL ) ); 5 6 al_unregister_event_source( m_eventsQueue, al_get_display_event_source( m_display ) ); 7 al_destroy_display( m_display ); 8 m_display = al_create_display( w, h ); 9 if ( !m_display ) // Fallback 10 Reinit( type, false, 800, 600, true ); 11 else 12 al_register_event_source( m_eventsQueue, al_get_display_event_source( m_display ) ); 13 14 assert( m_display && "Failed to set new display mode" ); 15}

By the way, current code does not try to set 1024x768. I modified it to try to get your desktop resolution and set it as fullscreen. Maybe this is wrong. Here's code:

    ALLEGRO_MONITOR_INFO info;
    if( al_get_monitor_info( 0, &info ) )
    {
      render.Reinit( RENDERTYPE_DIRECT3D, true, info.x2-info.x1, info.y2-info.y1, true ); 
    }
    else // Keep current resolution (800x600 probably)
      render.ToggleFullscreen();

Edgar Reynaldo
Major Reynaldo
May 2007
avatar

Well, I checked al_get_monitor_info's return information and I get 0,0 to 1280,800 on adapter 0 on my laptop, so that shouldn't be the problem.

Do you have mingw setup? I can't debug with MSVS, but I can with gdb and g++, if you include debugging symbols.

I can see how your code would crash if you fail to set the successful resolution though, and that is because you would be unregistering the event source from the display twice and on a destroyed display the second time through. Unless al_unregister_event_source accounts for that somehow by checking if that source is actually registered. If you call Reinit for windowed mode you will also be destroying the display twice. Don't know why it fails in the first place though.

That's all I can see from what you've shown me, and I don't know what else would cause it.

Max Savenkov
Member #4,613
May 2004
avatar

Unfortunately, I don't have MinGW installed, or makefiles to use with it. I can upload a version with full debugging information, which could maybe help with WinDbg. Also, you can try to get callstack information after crash with ProcessExplorer (Process properties -> Threads -> Stack. It also can create mini and full memory dumps.

Edgar Reynaldo
Major Reynaldo
May 2007
avatar

Are you using OpenGL in that last exe you gave me? pixelbots.exe's top frame is inside atioglxx.dll. What is the development support like for OpenGL with MSVS? What versions are supported?

Call stack for pixelbots.exe at time of crash on main thread :

ntkrnlpa.exe!KeWaitForMultipleObjects+0xab7
ntkrnlpa.exe!KeWaitForSingleObject+0x492
ntkrnlpa.exe!PsGetCurrentThreadTeb+0x377
ntkrnlpa.exe!KeInsertQueueDpc+0x670
ntkrnlpa.exe!KeWaitForSingleObject+0x492
ntkrnlpa.exe!KeAreApcsDisabled+0x5b5
ntkrnlpa.exe!RtlGenerate8dot3Name+0x777
ntkrnlpa.exe!RtlGenerate8dot3Name+0x1944
ntkrnlpa.exe!RtlGenerate8dot3Name+0x107c
ntkrnlpa.exe!ZwQueryLicenseValue+0xbc6
ntkrnlpa.exe!ZwAlpcSendWaitReceivePort+0x11
ntkrnlpa.exe!PsAssignImpersonationToken+0x376
ntkrnlpa.exe!ExfReleasePushLockShared+0xd10
ntkrnlpa.exe!IoGetRequestorProcessId+0x2a0
ntkrnlpa.exe!KiCoprocessorError+0x173
ntkrnlpa.exe!ZwQueryLicenseValue+0xbc6
ntdll.dll!KiFastSystemCallRet
atioglxx.dll!DrvPresentBuffers+0xa2011

Kris Asick
Member #1,424
July 2001

I'm absolutely terrible at strategy games (save for anything made by Bullfrog before EA bought 'em out) so I probably won't be trying out yours, but it's awesome that my own stuff served to help inspire you towards your own designs! :)

The name seems odd though... You call them "Pixel" bots except they don't exactly look pixelated or such. The "Pixel" in "PixelShips" is referring to the fact that each ship is made up of component pixels that are simply tiny little blocks which fit together, thus when you destroy a ship, you need to pick up all of the pixels to successfully collect it. Even in the design stuff for PixelShips 2, the ships retain this pixelated look, just with higher resolution graphics, and there's even an emphasis on mixing and matching pixels to form new ships.

Unless you're referring to the fact that the maps just look like big pixel squares, but then that just makes the title even more misleading since it's the world that's pixelated, not the bots.

*shrugs* Ah well. It's your game you can name it whatever you want. ;)

--- Kris Asick (Gemini)
--- http://www.pixelships.com

Max Savenkov
Member #4,613
May 2004
avatar

Kris, it's a working title and it mostly reflects roots of game's idea, but not really related to actual gameplay. I think I will change it later.

Edgar, that's very strange! The latest debug version should not be using OpenGL at all, only DirectX. How it ended up in atioglxx.dll is a mystery to me, unless ATI routes DirectX calls through it intentionally (which is silly). Are you sure you're using the latest version I uploaded? I have chcecked active threads with ProcessExplorer for it, and I only see nvd3dum.dll, which is NVidia's D3D driver (when I run OpenGL version, I see nv3ogl.dll).

Edgar Reynaldo
Major Reynaldo
May 2007
avatar

Max Savenkov
Member #4,613
May 2004
avatar

Here it is. Actually, yes, I should begin numbering my versions, you're right, at least with build data/time.

Edgar Reynaldo
Major Reynaldo
May 2007
avatar

No, sorry, it is still crashing in the same place when I try to set full screen.

According to Process Explorer

17 atioglxx.dll!DrvPresentBuffers+0xa2011

I'm not having much luck with WinDbg either. Even when I set the symbol search path to your bin directory it says it still can't find debugging symbols. Your symbols don't work for some reason. Funny thing is when I run the program from WinDbg it crashes with a BEX error before the game's window even opens, instead of crashing when I try to run full screen.

I don't know but maybe there is some kind of binary incompatibility somewhere. That's the only reason I can think of why this would happen, because I have no problems setting 1280x800 on my laptop with A5.

Sorry I can't help more.

Here's !analyze -v if it helps. Check out the last control transfer part. It tells you the last two places where the instruction pointer was.

Edit- Nevermind, ln 77877b0e lists ntdll!DebugBreakPoint as the last symbol accessed.

#SelectExpand
10:000> !analyze -v
2****************************************************************************
3* * 4* Exception Analysis * 5* *
6****************************************************************************
7 8***** OS symbols are WRONG. Please fix symbols to do analysis. 9
10**********************************************************************
11***
12***
13*** Your debugger is not using the correct symbols
14***
15*** In order for this command to work properly, your symbol path
16*** must point to .pdb files that have full type information.
17***
18*** Certain .pdb files (such as the public OS symbols) do not
19*** contain the required information. Contact the group that
20*** provided you with these symbols if you need this command to
21*** work.
22***
23*** Type referenced: ntdll!_PEB
24***
25**********************************************************************
26*** ERROR: Symbol file could not be found. Defaulted to export symbols for C:\Windows\system32\kernel32.dll -
27**********************************************************************
28***
29***
30*** Your debugger is not using the correct symbols
31***
32*** In order for this command to work properly, your symbol path
33*** must point to .pdb files that have full type information.
34***
35*** Certain .pdb files (such as the public OS symbols) do not
36*** contain the required information. Contact the group that
37*** provided you with these symbols if you need this command to
38*** work.
39***
40*** Type referenced: nt!IMAGE_NT_HEADERS32
41***
42**********************************************************************
43*** ERROR: Symbol file could not be found. Defaulted to export symbols for C:\Windows\system32\ole32.dll - 44 45FAULTING_IP: 46ntdll!DbgBreakPoint+0 4777877b0e cc int 3 48 49EXCEPTION_RECORD: ffffffff -- (.exr 0xffffffffffffffff) 50ExceptionAddress: 77877b0e (ntdll!DbgBreakPoint) 51 ExceptionCode: 80000003 (Break instruction exception) 52 ExceptionFlags: 00000000 53NumberParameters: 3 54 Parameter[0]: 00000000 55 Parameter[1]: a20b0d78 56 Parameter[2]: 778896f4 57 58FAULTING_THREAD: 000016ac 59 60DEFAULT_BUCKET_ID: WRONG_SYMBOLS 61 62PROCESS_NAME: PixelBots.exe 63 64ADDITIONAL_DEBUG_TEXT: 65Use '!findthebuild' command to search for the target build information. 66If the build information is available, run '!findthebuild -s ; .reload' to set symbol path and load symbols. 67 68MODULE_NAME: ntdll 69 70FAULTING_MODULE: 77830000 ntdll 71 72DEBUG_FLR_IMAGE_TIMESTAMP: 4cb733dc 73 74ERROR_CODE: (NTSTATUS) 0x80000003 - {EXCEPTION} Breakpoint A breakpoint has been reached. 75 76EXCEPTION_CODE: (HRESULT) 0x80000003 (2147483651) - One or more arguments are invalid 77 78EXCEPTION_PARAMETER1: 00000000 79 80EXCEPTION_PARAMETER2: a20b0d78 81 82EXCEPTION_PARAMETER3: 778896f4 83 84MOD_LIST: <ANALYSIS/> 85 86PRIMARY_PROBLEM_CLASS: WRONG_SYMBOLS 87 88BUGCHECK_STR: APPLICATION_FAULT_WRONG_SYMBOLS 89 90LAST_CONTROL_TRANSFER: from 778a2fe6 to 77877b0e 91 92STACK_TEXT: 93WARNING: Stack unwind information not available. Following frames may be wrong. 940029f958 778a2fe6 7ffdf000 7ffdc000 778f51b8 ntdll!DbgBreakPoint 950029fa9c 7786094d 0029fb2c 77830000 71b8cd79 ntdll!RtlTimeToElapsedTimeFields+0x446f 960029fb08 7786dfc7 0029fb2c 77830000 00000000 ntdll!RtlSetHeapInformation+0x69 970029fb18 00000000 0029fb2c 77830000 00000000 ntdll!LdrInitializeThunk+0x10 98 99 100FOLLOWUP_IP: 101ntdll!DbgBreakPoint+0 10277877b0e cc int 3 103 104SYMBOL_STACK_INDEX: 0 105 106SYMBOL_NAME: ntdll!DbgBreakPoint+0 107 108FOLLOWUP_NAME: MachineOwner 109 110IMAGE_NAME: ntdll.dll 111 112STACK_COMMAND: ~0s ; kb 113 114BUCKET_ID: WRONG_SYMBOLS 115 116FAILURE_BUCKET_ID: WRONG_SYMBOLS_80000003_ntdll.dll!DbgBreakPoint 117 118Followup: MachineOwner 119---------

Max Savenkov
Member #4,613
May 2004
avatar

I'm out of ideas for now. Let's wait for the second prototype. It will have resizeable GUI, so anyone will be able to play in windowed mode in any resolution and also I will try to integrate better crash reporting tool (although so far I'm finding Google BreakPad somewhat strange under Windows).

 1   2   3 


Go to: