I'd suggest that you do a few more tests and report what you find:

1. Move the old mptrack.ini file out of the current directory, and restart so that MPT makes a new INI file. You'll hate to hafta reset all the paths and settings, but it'll help you narrow down the source of the problem.

2. Is there a certain point where it crashes (e.g., clicking on a button, at a given time after starting up, only during playback, etc.)? Keep making it crash to see if you can replicate the behavior.

3. Often i've noticed un-asserted crashes tend to be memory run-ins or access problems.

4. I've also noticed some applications within and outside of Windows have a problem with DEEPLY nested pathnames. Try pulling your MPT folder out to the desktop and running it from there. Also try this with your plugin folders.

5. I believe there is a debug or logging command (Jojo would know this), so research that and see if that helps.

Tell us what you find here...

3. The access "problems" you are talking about are called access violations and happen when accessing memory that has been freed before or trying to access a null pointer or f.e. when accessing a vector element that is out of range.

4. Only happens with path names close to 250 characters.

5. There isn't, but I can provide additional files for debugging if you have Visual Studio installed (preferrably VS2008, I guess any flavour (Trial/Pro/...) should do).

I am currently testing in various scenarios, but is still hard to get results.

@Saga_Musix

5. I have VS2010 Pro Trial, but not installed. I can install for more analysys on this bug.

Please tell me more what else should I do.

I'll send you a PM with additional information on how to debug using some additional files.

I guess that you sent me a debug build of OpenMPT, but un near six hours of testing I cannot make it crash, just one time at this time and JIT debugger don't have recognized the error (only the "Error found in program" [send bug report to MS or whatever I don't know how it is in english :P ]. This last error I think that is related with VST's but I am still unsure.

Maybe it was related to a recent fix then. Have you, by any chance, used the Cleanup dialog? I know that there was a crash in that dialog in 1.19.01.03.

No, I have not use it.

Instead of crashin it only hangs (for the 1.19.01.04). The window disappear but the process remains on menory, perhaps to be a bug in a plugin bundled with Acoustica Mixcraft. http://www.acoustica.com/mixcraft/download.htm

.\Acoustica Mixcraft 5\VST\Acoustica\Acoustica Instruments.dll
.\Acoustica Mixcraft 5\VST\Acoustica\Acoustica Expanded Instruments.dll

Has the bug occurred in previous versions? Yes, but not in the same way.

You might have to click the window if it seems to be hanging. You should get a crash dialog then.

It hangs, but you need to force task kill.

Here is a result of JIT:
Unhandled exception at 0x000c6c93 in mptrack.exe: 0xC0000005: Access violation reading location 0x000c6c93.

The addresses are unluckily not that helpful. Didn't you manage to create a debugging instance as described in my PM? You should get the chance to do so when you get the crash prompt.

The bug says that were on CControlBar, but I cannot remember the section of barcore.cpp

If I am not wrong were occured on this line: LRESULT CControlBar::WindowProc(UINT nMsg, WPARAM wParam, LPARAM lParam)
and I can remember something here but in line 507

That function would be part of the MFC library, but I can't even find it being used anywhere...

I have an active current bug (in debugger) What data do you need?

I think that you get this useful http://www.mediafire.com/file/fccxwy4oqogl14n/dbg-mptrack.7z

EDIT: Updated link.

Those screenshots appear to be a bit more useful. However you need revision 859 of the code so the debugger will give reliable results (one of the screenshots say that sndmix.cpp doesn't match, so it might give wrong results). With revision 859, you should be able to see the exact line where it crashed and the current state of all variables.

It's the 859 downloaded as tarball on SVN repository. I don't why appear that warning if I have the r859

Let me some time to provide feedback; I'm currently not working on songs that crash OpenMPT. Also, how useful is dump openmpt memory(?) when debugging from VS? I guess it's useful but I don't know if you need other files for debugging. Another issue I have, is a hang of OpenMPT, but not a crash. Please let me know what you need for proper debugging.

Still pending, I'll go into v1.19.01 and debug (try to) if I am able.

I do a lot of test regarding this issue on XP but I cannot reproduce, now I have updated my main system to Windows 7 and tested too. Still without successful results, so I have decided to close this.