C / C++ / MFC

9-Jan-12 4:45

Hi all,

I have a problem when trying to resize main window of my application. I spent some time reserching possible solutions, but so for nothing works for me. I guess I make some basic mistake.
So - the code bellow is executed when user clicks on one of the buttons. I tried 3 different approaches to achieve it, everytime getting assertion error (looks to me like the window handle is invalid...)
My first approach:

C++

WINDOWPLACEMENT wpc;
CRect rc(0, 0, 0, 0);

    this->GetWindowRect(&rc);
    rc.right += 50;  // make it wider

    this->GetWindowPlacement (&wpc);
    wpc.rcNormalPosition = rc;
    this->SetWindowPlacement(&wpc);

Second approach:

C++

CRect rc(0, 0, 0, 0);

this->GetWindowRect(&rc);
rc.right += 50;

ScreenToClient( &rc );

    this->MoveWindow(rc);

and the third one:

C++

CRect rc(0, 0, 0, 0);

this->GetWindowRect(&rc);
rc.right += 50;

ScreenToClient( &rc );

this->SetWindowPos(this->m_hWnd, rc.left, rc.top, rc.Width(), rc.Height(), SWP_SHOWWINDOW);

The SetWindowPos() function doesn't return assertion error when I use this combination of flags: SWP_SHOWWINDOW|SWP_NOSIZE. However then it doesn't resize the window and it's useless for me.

I also checked than 'this' pointer is not corrupted by checking the
window text:

C++

char title[80];
		::GetWindowText (m_hWnd, title, sizeof(title));

It comes as I expect so I guess 'this' pointer itself is OK.
Any ideas?

Albert Holguin9-Jan-12 4:58

9-Jan-12 4:58

what assertion error are you getting? when you call this, that's the object instance of whatever object you're calling that within, what object type is that?

btw... if the error is telling you that the window handle is invalid, then you're probably not passing the argument correctly (it usually does a check as to whether that handle corresponds to any available windows)...and you're using SetWindowPos() wrong, look at the documentation again:
http://msdn.microsoft.com/en-us/library/a1yzfz6d%28v=VS.80%29.aspx[^]

blackbolek9-Jan-12 10:01

9-Jan-12 10:01

I'm going through the code line by line in debug mode and on the line with SetWindowPos() Visual Studio comes with 'Debug Assertion Failed!' window. It doesn't give me any specific information. When I click on 'Retry' then it takes me to 'wingdi.cpp' file, to BOOL CDC::Attach(HDC hDC) function with yellow arrow at the first line of it:

C++

ASSERT(m_hDC == NULL);      // only attach once, detach on destroy

So that's why I assumed it's something wrong with the window handle.

My this is 'CGraphDlg* const' type. CGraphDlg is my main dialog class:

C++

class CGraphDlg : public CDialog

I've changed the way I'm using SetWindowPos() to that:

this->SetWindowPos(&CWnd::wndBottom, rc.left, rc.top, rc.Width(), rc.Height(), SWP_SHOWWINDOW);

but I'm still getting the same problem.

I guess I can't use this here. I need some other pointer to the main window, but I'm not sure how to get it.
Or is this problem occurring because I'm using dialogs?

Albert Holguin9-Jan-12 10:18

9-Jan-12 10:18

That assert is telling you the device context is NULL, a separate issue. Where are you calling this from? From the OnPaint()? If so, that may be causing the issue (although I'm not sure it would or not). Try pulling the call to some other method that doesn't handle the device context.

blackbolek9-Jan-12 22:01

9-Jan-12 22:01

I'm calling it from push button event.
I moved the resizing code to the OnInitDialog() and the window was resized without any problems. I'm drawing a graph in this window. I do that in InitGraph() function which is called from InitDialog. So if I place the SetWindowPos() before call to InitGraph() then it works. If I place it after, then I have an error.

Looks to me that I do something wrong with DC in InitGraph():

C++

void CGraphDlg::InitGraph()
{
	
        CRect rectControl5;		// area of control being moved - divider line above controls
	CBitmap image;			// bitmap with logo

	// This functions sets up the bitmap in memory. Called from OnInitDialog.

	CDC * pDC = GetDC();
	
	if (!not_first_paint)
	{
		image.LoadBitmap(IDB_GRAPHBLANK);	// load in a blank BMP of the desired size
		dcGraphMem.CreateCompatibleDC(pDC);
		// Select the bitmap into the new DC in memory.
		dcGraphMem.SelectObject(&image);
	}

	// save window size
  	GetClientRect(GraphWindow);
	YSCALE = GraphWindow.Height() - 191;		
	graphx = GraphWindow.Width() - 212;		// current graph working area width
	
	// adjust size of the bitmap to the actual window size
	StretchBlt (dcGraphMem, 0, 0, GraphWindow.Width(), GraphWindow.Height()-25, dcGraphMem, 0, 0, 1600, 1044 , SRCCOPY);

	
	ReleaseDC(pDC) ;


	m_divline.GetParent()->GetClientRect(GraphWindow);
	m_divline.GetWindowRect(rectControl5);
	m_divline.MoveWindow(0, GraphWindow.Height()-BOTTOMCONTROLS-2, rectControl5.Width(), rectControl5.Height());

	image.DeleteObject();


	// now we know we went through graph initialization, 
	not_first_paint = true;


	return;
}

Albert Holguin10-Jan-12 3:27

10-Jan-12 3:27

How often do you call this? ...if it's called every time the graph is painted, you're trying to resize the window every single time and that's probably not what you want.

blackbolek9-Jan-12 23:30

9-Jan-12 23:30

I found what the problem is.
I tried to resize the window with DC still attached to it. That's why I had the assertion failure.

I added this code before resizing:

C++

if (dcGraphMem.m_hDC != NULL)
	dcGraphMem.DeleteDC();

and this after resizing:

C++

CDC * pDC = GetDC();

	if (dcGraphMem.m_hDC == NULL)
		dcGraphMem.CreateCompatibleDC(pDC);

and now it works OK.

Thank you all for your help. Thumbs Up | :thumbsup:

Richard MacCutchan10-Jan-12 2:31

10-Jan-12 2:31

You should not be holding a DC like this in the first place. You only need a DC at the time you need to display anything, which is in your OnPaint() function, it has no purpose when resizing.

Unrequited desire is character building. OriginalGriff
I'm sitting here giving you a standing ovation - Len Goodman

Albert Holguin10-Jan-12 3:29

10-Jan-12 3:29

If you have to do this... you may be resizing the window in a method where it's not appropriate to do so. See Richard's comment about the device context. You should only be holding the DC when you're actually doing any GDI drawing, otherwise, it should be released. That's sort of what's causing your problem in the first place, you're making a call to a method that requires the DC when you're already holding it up.

Hope you understand all this gibberish... good luck! Smile | :)

blackbolek10-Jan-12 20:52

10-Jan-12 20:52

Well I'm not holding the DC. What I do is open DC in the InitDialog for short time only to create compatibile memory DC, and then I draw my graph in memory DC. When necessary I open the DC again and copy the memory DC into DC.

C++

if (dcGraphMem.m_hDC == NULL)			// if memory device context not created yet
	{
		CDC * pDC = GetDC();
		dcGraphMem.CreateCompatibleDC(pDC);	// create one
		
		ReleaseDC(pDC) ;
	}

then I operate on dcGraphMem and when it's ready I do that:

C++

CDC * pDC = GetDC();

		if(pDC == NULL)
		{
			// Not able to get the screen
			return;
		}


		pDC->BitBlt(0, 0, GraphWindow.Width(), GraphWindow.Height()-BOTTOMCONTROLS-2, &dcGraphMem, 0, 0, SRCCOPY);

		ReleaseDC(pDC);

I thought that memory DC is not linked with window directly, however it looks like I need to delete it before I resize the window, and then create it again. Why is that?

Albert Holguin11-Jan-12 7:55

11-Jan-12 7:55

You don't need to delete it, you just can't hold it up, it's not hooked to the window directly, but your application won't be able to attach to the drawing context twice (it will eventually fail, as you've found out). You also should be drawing on your InitDialog() (this is for initializing controls really), that's not the proper place for that.

Any drawing should occur on OnPaint(), alternatively, no resizing should occur within OnPaint() (can't think of any legitimate reason to do so). Reason to keep all drawing in OnPaint()... that's the method the framework will call every time the framework determines your window needs to be redrawn, for whatever reason, anything from a resize to a window getting placed over your window then removed, etc, will trigger a call to OnPaint().

Richard MacCutchan9-Jan-12 5:50

9-Jan-12 5:50

blackbolek wrote:
The SetWindowPos() function doesn't return ...

The call to ScreenToClient() is changing the values in rc so that is not going to do what you want.

blackbolek wrote:
I also checked than 'this' pointer is not corrupted by checking the window text:

But you did not call it with the this pointer so the code is not the same as all your other calls to windows functions.

Unrequited desire is character building. OriginalGriff
I'm sitting here giving you a standing ovation - Len Goodman

Albert Holguin9-Jan-12 7:26

9-Jan-12 7:26

To add on to what Richard said... using this on a window object is not the same as using it's handle. Reason is, this refers to the object instance and not the handle to the window, which is different (they are linked together when a window is created, not when the object is created, but even then, they are not the same).

...Just in case in needs to be clarified....

blackbolek9-Jan-12 10:14

9-Jan-12 10:14

With ScreenToClient() commented off it gives me still the same 'Debug Assertion Failed' message.

Also I've modified the line which gets the window title to this:

C++

::GetWindowText (this->m_hWnd, title, sizeof(title));

And I still get proper window title.

Luc Pattyn9-Jan-12 11:07

Luc Pattyn

9-Jan-12 11:07

blackbolek wrote:
::GetWindowText (this->m_hWnd, title, sizeof(title));

and that is how I would attempt to get and set window position and size too, with an explicit hWnd, and not trough this->...

Smile | :)

Luc Pattyn [My Articles] Nil Volentibus Arduum

Richard MacCutchan9-Jan-12 21:19

Top level menu - Message on losing focus

9-Jan-12 21:19

You need to get your debugger working and check exactly where this assertion occurs and which variable it is complaining about. Also you should be calling these Windows functions via the class methods. If ::GetWindowText() works as shown above then it should also work with the following:

C++

GetWindowText(title, sizeof(title));

which is the correct way to code it.

Unrequited desire is character building. OriginalGriff
I'm sitting here giving you a standing ovation - Len Goodman

RavishB9-Jan-12 3:22

RavishB

9-Jan-12 3:22

Hi,

This is in reference to a MFC COM project I'm working on VC6.
The question is in regard to the top level menu.
The project contains a top level menu, say "Options"

I need to do certain “operations”, say toggle a BOOL value when the menu pane is open and when the menu pane is closed.

On selecting “Options”, WM_INITMENUPOPUP message is received and I can handle the functionality under this and make the BOOL variable TRUE.

The issue arises when the menu pane closes without selecting any menu option, I do not receive any windows message to handle and there is no confirmation that the menu pane is closed. Hence no location to toggle the BOOL variable.

I’ve checked on WM_UNINITMENUPOPUP message, but this message is not handled for VC6.

I’ve tried to retrieve the HMENU handle to the menu, but this is not possible as the menu is created in a different project altogether and hence I cannot use MENUBARINFO structure to find out if the menu is open or not.

Hence, the current challenge I face is whether or not the menu pane has been closed.

Is there any way to know if the menu pane is visible using the ids of the menu items or any other approach?

Thanks in advance for your thoughts

Regards,
Ravish

Re: Top level menu - Message on losing focus

Code-o-mat9-Jan-12 6:14

Code-o-mat

9-Jan-12 6:14

How about WM_EXITMENULOOP[^]?

> The problem with computers is that they do what you tell them to do and not what you want them to do. <
> If it doesn't matter, it's antimatter.<

Log event message

_Flaviu8-Jan-12 22:58

_Flaviu

8-Jan-12 22:58

I programatically write an information into system event log ... and goes well, except that I read this in event viewer :

C++

The description for Event ID 1 from source TestEventLog cannot be found. Either the component that raises this event is not installed on your local computer or the installation is corrupted. You can install or repair the component on the local computer.
If the event originated on another computer, the display information had to be saved with the event.
The following information was included with the event: 
My test message
the message resource is present but the message is not found in the string/message table

and my message is only 'My test message' ... why don't I see only my message ?

Re: Log event message

Chandrasekharan P8-Jan-12 23:38

Chandrasekharan P

8-Jan-12 23:38

This link should help you solve your problem. [^]

Every new day is another chance to change your life.

WinSNMP

msr_codeproject8-Jan-12 19:40

msr_codeproject

8-Jan-12 19:40

Hi all, I want to know WinSNMP supports which versions of SNMP. Please share some info.

Regards
msr

Re: WinSNMP

Richard MacCutchan8-Jan-12 22:55