Thread: Unhandled exception

When I debug my app there is no problem not even a warning. When I release the app again no warnings no issues while running the program.

When I run the app manually (double click on the actual icon)

Code:

An unhandled win32 exception occurred in ProgX.exe

It offers me to debug the app again. The app stops in WM_CREATE message. I suppose there might be lack of memory for the program but why does this happen only when I launch it manually?

Has this ever happened to you. Do you have an Idea what to do next?

I will certainly provide more information according to what you need to know about the app.

Thank you in advance for any help

Does the same also happen when you double-click on executable when it's compiled as a debug build? Try commenting out stuff, or using a bunch of calls to MessageBox() to triangulate on what exact line it's happening on.
It has never happened to me before, but I'd guess the difference lies in the environment the executable is run in.

Post the code that you have in your WM_CREATE case.

The debug built works fine if run manually. No prob with that one either. Only with the release.

Code:

LRESULT CALLBACK WndProc(HWND hWnd, UINT message, WPARAM wParam, LPARAM lParam)
{
        switch(message)
        {
		case WM_CREATE:
			ProgStart(hWnd);
			_pGame->Clean(hWnd);
			break;
        }
}

ProgStart(hWnd); Contains the following code void Sudoku::Start(HWND hWnd) where does the issue show up.

If it helps the actual point where the prog stops is in the highlighted line.

Code:

void Sudoku::Start(HWND hWnd)
{
	// All rects needed to be disabled after start
	for(int i = 0; i <= m_iRowCount; i++)
	{
		for(int j = 0; j <= m_iColCount; j++)
		{
			m_bOwner[i][j] = true;
			m_bActive[i][j] = false;
			m_bSolver[i][j] = false;
			bHighLightOwner = false;
			bHighLightSolver = false;
			bHighLightBase = false;
			m_iValue[i][j] = -1;
			m_iCorrect[i][j] = -1;
			for(int k = 0; k <= 8; k++)
			{
				m_iPossible[i][j][k] = k + 1; // At this point an unhandeled exception always occurs
			}
			iProgressBar++;
			m_bActive[i][j] = false;
		}
	}
}

All needed variables are declared so I dont believe there is a problem with invalid pointers.

For instance, if m_iRowCount or m_iColCount are 9 -- which would make sense, since that's how many rows and columns there are -- then i and j go over the limit.

Originally Posted by tabstop

For instance, if m_iRowCount or m_iColCount are 9 -- which would make sense, since that's how many rows and columns there are -- then i and j go over the limit.

Originally Posted by Elysia

The problem is that you do a buffer overrun (or buffer underrun) somewhere.

There might be a possibility of buffer overrun but why does it work properly in debug build and even while running the release built from the compiller?

Gordon, just try replacing

Code:

	for(int i = 0; i <= m_iRowCount; i++)
	{
		for(int j = 0; j <= m_iColCount; j++)
		{

with

Code:

	for(int i = 0; i < m_iRowCount; i++)
	{
		for(int j = 0; j < m_iColCount; j++)
		{

and see if the buffer still overflows

Well turns out I initialized my m_iRowCount and m_iColCount later then I first needed them so there I go. Sorry for bothering with that stupid problem.

Anyway thanks a lot everyone for help

Originally Posted by Elysia

Uninitialized variables... Visual Studio detects them and triggers a runtime error if you use them without initializing them first.
I'm surprised that you missed that, in case you are using VS.

I am surprised myself that VS let me do that... At least I can see that the starting function is not that useful in this case.