Duke4.net Forums: Mapster32 crash - Duke4.net Forums

Jump to content

Page 1 of 1
  • You cannot start a new topic
  • You cannot reply to this topic

Mapster32 crash

User is offline   Kyanos 

#1

Latest build quits silently to desktop when switching from 3D mode (openGL) to 2D mode (8bit obviously.) It's probably tied up with switching the render mode because 3D mode software to 2D works fine.

This post has been edited by Drek: 23 June 2018 - 09:23 AM

0

User is offline   Hendricks266 

  • Weaponized Autism

  #2

Does the log file say anything, particularly with a debug build?
0

User is offline   Kyanos 

#3

Here you go tried it out just now, it crashed with a popup mentioning the exception and asked me to send it to you.

Quote

Caught exception 0xC0000005 at 0x000007FEFE755A83: Access violation.
0x000007FEFE755A83 : C:\Windows\system32\msvcrt.dll : mbscpy
0x00000000004A8850 : E:\Users\Desk\Downloads\eduke32_win64_debug_20180620-6927\ mapster32.debug.exe : /var/synthesis/eduke32/source/build/src/engine.cpp (12774) : in function (videoSet2dMode)
0x00000000004605EE : E:\Users\Desk\Downloads\eduke32_win64_debug_20180620-6927\ mapster32.debug.exe : /var/synthesis/eduke32/source/build/src/build.cpp (3365) : in function (overheadeditor)
0x00000000004736B9 : E:\Users\Desk\Downloads\eduke32_win64_debug_20180620-6927\ mapster32.debug.exe : /var/synthesis/eduke32/source/build/src/build.cpp (1488) : in function (editinput)
0x000000000042E1C9 : E:\Users\Desk\Downloads\eduke32_win64_debug_20180620-6927\ mapster32.debug.exe : /var/synthesis/eduke32/source/duke3d/src/astub.cpp (10925) : in function (ExtCheckKeys)
0x0000000000453A57 : E:\Users\Desk\Downloads\eduke32_win64_debug_20180620-6927\ mapster32.debug.exe : /var/synthesis/eduke32/source/build/src/build.cpp (11091) : in function (CallExtCheckKeys)
0x000000000047212B : E:\Users\Desk\Downloads\eduke32_win64_debug_20180620-6927\ mapster32.debug.exe : /var/synthesis/eduke32/source/build/src/build.cpp (846) : in function (app_main)
0x0000000000517686 : E:\Users\Desk\Downloads\eduke32_win64_debug_20180620-6927\ mapster32.debug.exe : /var/synthesis/eduke32/source/build/src/sdlayer.cpp (499) : in function (WinMain)
0x0000000000516FE9 : E:\Users\Desk\Downloads\eduke32_win64_debug_20180620-6927\ mapster32.debug.exe : /var/synthesis/eduke32/source/build/src/mingw_main.cpp (96) : in function (main)
0x00000000004013F7 : E:\Users\Desk\Downloads\eduke32_win64_debug_20180620-6927\ mapster32.debug.exe : /build/mingw-w64-crt/src/mingw-w64-v5.0.3/mingw-w64-crt/crt/crtexe.c (336) : in function (__tmainCRTStartup)
0x00000000004014FB : E:\Users\Desk\Downloads\eduke32_win64_debug_20180620-6927\ mapster32.debug.exe : /build/mingw-w64-crt/src/mingw-w64-v5.0.3/mingw-w64-crt/crt/crtexe.c (186) : in function (WinMainCRTStartup)
0x00000000774F652D : C:\Windows\system32\kernel32.dll : BaseThreadInitThunk
0x000000007772C541 : C:\Windows\SYSTEM32\ntdll.dll : RtlUserThreadStart
Sat Jun 23 13:45:38 2018
---------------

0

User is offline   Kyanos 

#4

I bisected revisions to come up with R6922 as the culprit, probably the glsuface stuff.
1

User is offline   Kyanos 

#5

This fixes it but it's sort of ugly and I'm sure you guys will do better, I'm interested in how you'll solve it.

adding these changes to engine.cpp, an extern int GLS_init to monitor with
~line 9905
        glsurface_initialize({xdim, ydim});
        GLS_init = 1;


~line 12767 re-initialize the glsurface if just coming back from 3D mode
#ifdef USE_OPENGL
        fxdim = (float) xres;
        fydim = (float) yres;
        if (GLS_init==0)
        {
            glsurface_initialize({xdim, ydim});
            GLS_init=1;
        }
#endif


And a toggle in Build.cpp
        if (keystatus[buildkeys[BK_MODE2D_3D]])
        {
            updatesector(pos.x,pos.y,&cursectnum);
            if (cursectnum >= 0)
            {   
                GLS_init = 0;
                keystatus[buildkeys[BK_MODE2D_3D]] = 2;
            }


This post has been edited by Drek: 23 June 2018 - 01:49 PM

0

User is offline   pogokeen 

  • EDuke32 Developer

#6

TerminX and I pushed some fixes for these issues. You should find things are more stable with r6932.

Sorry for the trouble, please let me know here if you run into any more of these issues!
1

User is offline   Paul B 

#7

View Postpogokeen, on 25 June 2018 - 09:28 AM, said:

TerminX and I pushed some fixes for these issues. You should find things are more stable with r6932.

Sorry for the trouble, please let me know here if you run into any more of these issues!

I think I'm getting the same problems here with Windows 10 1803. This is from Mapster32 r6969.

Caught exception 0xE06D7363 at 0x00007FFABCD1A388: (null)
0x00007FFABCD1A388 : C:\WINDOWS\System32\KERNELBASE.dll : RaiseException
0x00007FFABE7FAEEA : C:\WINDOWS\System32\msvcrt.dll : CxxThrowException


It's interesting that mapster.exe seems to load but not the debug version.

Please see my attached log file. Mapster.exe seems to run fine, this log file is generated from: mapster32.debug.exe

Attached File(s)



This post has been edited by Paul B: 31 August 2018 - 07:06 AM

0

User is offline   Paul B 

#8

View PostPaul B, on 31 August 2018 - 06:52 AM, said:

I think I'm getting the same problems here with Windows 10 1803. This is from Mapster32 r6969.

Caught exception 0xE06D7363 at 0x00007FFABCD1A388: (null)
0x00007FFABCD1A388 : C:\WINDOWS\System32\KERNELBASE.dll : RaiseException
0x00007FFABE7FAEEA : C:\WINDOWS\System32\msvcrt.dll : CxxThrowException


It's interesting that mapster.exe seems to load but not the debug version.

Please see my attached log file. Mapster.exe seems to run fine, this log file is generated from: mapster32.debug.exe



Still getting these strange errors with only the Debug version of Mapster 20181027-7136. Oddly enough this problem only seems to occur on my laptop and not my desktop Windows 10. Not sure if this is an isolated problem with my laptop or whether other people experience this. Windows 10 right click file context menu lets me choose whether to run Mapster using integrated Intel GPU or High performance Nividia GPU both end with the same result. I blame Windows 10 BS!

This post has been edited by Paul B: 27 October 2018 - 09:26 PM

0

User is offline   Forge 

  • Speaker of the Outhouse

#9

View PostPaul B, on 27 October 2018 - 09:23 PM, said:

Still getting these strange errors with only the Debug version of Mapster 20181027-7136. Oddly enough this problem only seems to occur on my laptop and not my desktop Windows 10. Not sure if this is an isolated problem with my laptop or whether other people experience this. Windows 10 right click file context menu lets me choose whether to run Mapster using integrated Intel GPU or High performance Nividia GPU both end with the same result. I blame Windows 10 BS!

most likely a registry or profile corruption on your computer.

Command prompt (as admin) and type regsvr32 KernelBase.dll then reboot

if you get the kernelbase.dll was loading but the entry point was not found message - then something in your registry is messed up.
0

User is offline   Paul B 

#10

View PostForge, on 27 October 2018 - 10:18 PM, said:

most likely a registry or profile corruption on your computer.

Command prompt (as admin) and type regsvr32 KernelBase.dll then reboot

if you get the kernelbase.dll was loading but the entry point was not found message - then something in your registry is messed up.



Yea, I don't know about your first recommendation. I don't just like randomly registering windows DLL's. If another kernelbase has taken priority as a result of some system environment path change then i'd have to locate the original path of the dll before running regsvr32. (not all dll's need to be registered or can be registered which doesn't always point to registry corruption or problems).

This is a very specific problem with the Debug mode of Mapster & Eduke32. If the DLL wasn't registered properly then Mapster itself probably wouldn't run among other things. I also don't think profile corruption is an issue. I was hoping someone else may have encountered a similar problem. It's not really a show stopper as I still can use Mapster just not with debugging mode. I'm pretty sure I can revert back to an older Eduke Debugger exe and it works.

So I just grabbed the Mapster Debugger r5948 and it works just fine. The newer debuggers crash and I noticed after the crash it leaves an Eduke32.exe running task in the background that never gets closed properly unless it is manually terminated from taskmgr. This glitch is a real bugger!

This post has been edited by Paul B: 28 October 2018 - 12:28 AM

0

User is offline   Forge 

  • Speaker of the Outhouse

#11

View PostPaul B, on 27 October 2018 - 11:27 PM, said:

I don't just like randomly registering windows DLL's.

No, you just randomly install programs and then complain when they don't work.

It works on the laptop, but not the desktop.
Is the laptop a 32 bit O/S & the desktop a 64 bit O/S?

that could indicate a code issue in the compiler

View PostPaul B, on 27 October 2018 - 11:27 PM, said:

The newer debuggers crash and I noticed after the crash it leaves an Eduke32.exe running task in the background that never gets closed properly unless it is manually terminated from taskmgr.

Unless this is something specific to the debugger mapster, I've never seen mapster load eduke32.exe all by itself. (even (t)est has to be launched by the user)
So are you running both programs at the same time, out of the same directory, all the time?

This post has been edited by Forge: 28 October 2018 - 07:03 AM

0

User is offline   Hendricks266 

  • Weaponized Autism

  #12

What happens if you click Continue when the backtrace DLL presents its options dialog?
0

User is offline   Paul B 

#13

View PostHendricks266, on 28 October 2018 - 08:31 AM, said:

What happens if you click Continue when the backtrace DLL presents its options dialog?


Well at first after I choose the start button from the Mapster debug launcher the entire Mapster Window goes white. If I use either the ignore or continue buttons during launch it ends with a crash which is reported in the event viewer application log:
The program mapster32.debug.exe version 1.9.9.9 stopped interacting with Windows and was closed. To see if more information about the problem is available, check the problem history in the Security and Maintenance control panel.
Process ID: 5564
Start Time: 01d46ef53d48a64f
Termination Time: 201
Application Path: C:\EDuke32\mapster32.debug.exe
Report Id: 88b09ba0-0985-47b7-8c25-678a7bffdb24
Faulting package full name:
Faulting package-relative application ID:

After checking the problem history in the Security and Maintenance control panel. There is nothing listed.

@ Forge - Running both 64Bit Windows 10 Pro 1803. As far as my windows installation its pretty static. I don't just install software for the heck of it. Nothing really changes in my environment. But I realize this isn't the case for most people.

Attached thumbnail(s)

  • Attached Image: MapsterDebug.png


This post has been edited by Paul B: 28 October 2018 - 11:43 AM

0

User is offline   Forge 

  • Speaker of the Outhouse

#14

View PostPaul B, on 28 October 2018 - 11:40 AM, said:

@ Forge - Running both 64Bit Windows 10 Pro 1803. As far as my windows installation its pretty static. I don't just install software for the heck of it. Nothing really changes in my environment. But I realize this isn't the case for most people.

calm down Beavis. You won't try to register the dll, (a microsoft dll on your microsoft o/s) so I'm giving you grief for it - Also makes it harder to figure out if it's a problem with the program, or the visual basic / .net on your comp.
I'm not a c++ guru, but it would help if one of the two culprits was eliminated from committing the crime
1

User is offline   Mark 

#15

Its "settle down Beavis"
0

User is offline   Paul B 

#16

View PostForge, on 28 October 2018 - 12:32 PM, said:

calm down Beavis. You won't try to register the dll, (a microsoft dll on your microsoft o/s) so I'm giving you grief for it - Also makes it harder to figure out if it's a problem with the program, or the visual basic / .net on your comp.
I'm not a c++ guru, but it would help if one of the two culprits was eliminated from committing the crime


Okay, I've gone through each commit and the culprit seems to start between 6918-6922 and anything later.
r6922 | pogokeen | 2018-06-13 12:15:26 -0700 (Wed, 13 Jun 2018) | 1 line

Add glsurface.cpp/h and tilepacker.cpp/h to Visual Studio Project
------------------------------------------------------------------------
r6921 | pogokeen | 2018-06-13 12:15:22 -0700 (Wed, 13 Jun 2018) | 2 lines

Don't allow GL modes if GL Major Version is less than 2.
Additionally, fix build issues for RENDERTYPE=WIN introduced in r6827.
------------------------------------------------------------------------
r6920 | pogokeen | 2018-06-13 12:15:19 -0700 (Wed, 13 Jun 2018) | 1 line

sdlayer.cpp: don't alter desktop composition using win_setvideomode() if we're creating an SDL window for rendering GL
------------------------------------------------------------------------
r6919 | pogokeen | 2018-06-13 12:15:16 -0700 (Wed, 13 Jun 2018) | 1 line

When OpenGL is available, prefer using the new glsurface to blit the classic renderer's 8-bit buffer to the screen.
------------------------------------------------------------------------
r6918 | pogokeen | 2018-06-13 12:15:10 -0700 (Wed, 13 Jun 2018) | 1 line

tilepacker.cpp: instead of maintaining the heap on each insertion, maxheap_buildHeap() on tilepacker_pack()
------------------------------------------------------------------------
See http://svn.eduke32.c...repname=eduke32 for more details.

While this fixes it for the Mapster Debug it doesn't fix it for the Eduke32 Debug. Yet they are the same crashes. Sorry about that Forge, I just get all worked up when I click Debug and all I see is a white screen of death! =P

Eduke32 Debug Breaks between r4469-4472
r4472 | hendricks266 | 2014-05-17 05:36:40 -0700 (Sat, 17 May 2014) | 1 line

Replace some constants with preprocessor macros.
------------------------------------------------------------------------
r4471 | hendricks266 | 2014-05-17 05:36:03 -0700 (Sat, 17 May 2014) | 1 line

Fix two Polymer crashes related to corrupted maps.
------------------------------------------------------------------------
r4470 | hendricks266 | 2014-05-17 05:35:46 -0700 (Sat, 17 May 2014) | 1 line

Fix ebacktrace on 64-bit. New DLLs all around.
------------------------------------------------------------------------
r4469 | hendricks266 | 2014-05-17 05:35:03 -0700 (Sat, 17 May 2014) | 1 line

Add support for an SDL2-based wm_ynbox based on SDL2's testmessage.
------------------------------------------------------------------------
See http://svn.eduke32.c...repname=eduke32 for more details.

The last known good working Eduke32 Debug for my laptop is version: r4468. and I have a feeling that r4470 | hendricks266 | 2014-05-17 05:35:46 -0700 (Sat, 17 May 2014) | 1 line

Fix ebacktrace on 64-bit. New DLLs all around. is what stopped it from working on my laptop.

This post has been edited by Paul B: 28 October 2018 - 03:39 PM

1

Share this topic:


Page 1 of 1
  • You cannot start a new topic
  • You cannot reply to this topic


All copyrights and trademarks not owned by Voidpoint, LLC are the sole property of their respective owners. Play Ion Fury! ;) © Voidpoint, LLC

Enter your sign in name and password


Sign in options