C / C++ / MFC

18-Jul-07 22:05

I created a Dialog-based application and overriden CWinApp::OnIdle() but the focus is not coming inside the function. How to make the focus when the application don't have any message in message queue. I want to do this because, one acknowledgement message from H/W via USB port is missing between end of OnInitDialog() and display of dialog.
I want to check the message queue in OnIdle().
Any help would be greatly appreciable!!
Best Regards,
Suman

Mark Salsbery19-Jul-07 5:21

19-Jul-07 5:21

See here[^]

Mark Salsbery
Microsoft MVP - Visual C++

"Great job team! Head back to base for debriefing and cocktails."

rp_suman19-Jul-07 14:40

19-Jul-07 14:40

Hi Mark,
OnKickIdle() is running without problem.
Thanks for your great help!!
Best Regards,
Suman

Mark Salsbery19-Jul-07 14:56

Re: OnIdle() not getting focus [modified]

19-Jul-07 14:56

You're welcome!

Just like OnIdle(), you can also return non-zero from the message handler to get more idle
processing time.

Cheers,
Mark

Mark Salsbery
Microsoft MVP - Visual C++

"Remember the five Ds of dogeball: Dodge, Dip, Duck, Dive, and Dodge."
"If you can dodge a wrench, you can dodge a ball."

rp_suman19-Jul-07 17:46

19-Jul-07 17:46

If I try to get messages range upto 0x300, when closing the dialog, there is an error message saying "Debug Assertion Failed!" and the file name is: wincore.cpp in line 4563. Is there any range limit to get the messages.
My code is as follows:

LRESULT CIdleTestDlg::OnKickIdle(WPARAM wp, LPARAM lp)<br />
{<br />
	MSG msg;<br />
	if(GetMessage(&msg,NULL,0,0x300)){		<br />
			DispatchMessage(&msg);<br />
	}<br />
	return 0L;<br />
}

Best Regards,
Suman

-- modified at 0:13 Friday 20th July, 2007

Mark Salsbery19-Jul-07 20:14

Re: OnIdle() not getting focus [modified]

19-Jul-07 20:14

You should get messages the normal MFC way, through the message map.
WM_KICKIDLE is a message sent by the modal message loop in MFC.

This is from the docs:

The following are the ranges of message numbers. 

Range                  Meaning 
0 through WM_USER–1    Messages reserved for use by the system. 
WM_USER through 0x7FFF Integer messages for use by private window classes. 
WM_APP through 0xBFFF  Messages available for use by applications. 
0xC000 through 0xFFFF  String messages for use by applications. 
Greater than 0xFFFF    Reserved by the system.

What message are you trying to get?

Mark

Mark Salsbery
Microsoft MVP - Visual C++

"Remember the five Ds of dogeball: Dodge, Dip, Duck, Dive, and Dodge."
"If you can dodge a wrench, you can dodge a ball."

rp_suman19-Jul-07 20:57

19-Jul-07 20:57

I am trying to get MM_MIM_LONGDATA message. The value is 0x3c4.
But If I try to get upto range 0x300 or more the error is happening.
So I had doubt, there is some limit in message range. Thanks for info abt message limit.
I am trying to check message queue in OnKickIdle() because I have doubt an acknowledgement message from H/W via USB port is missing between end of OnInitDialog() and display of dialog if I use the normal MFC message map. I am trying to get USB MIDI port messages without missing in anyway.
Best Regards,
Suman

-- modified at 3:14 Friday 20th July, 2007

Mark Salsbery20-Jul-07 5:05

20-Jul-07 5:05

This shouldn't have anything to do with idle processing.

It doesn't make sense that you are missing a message between OnInitDialog and the displaying of
the dialog since you would have needed a valid HWND to get the message in the first place.

What window handle are you passing when you call midiInOpen()?
When are you calling midiInOpen()?

You may get better performance if you use a callback function instead of window messages.

If you choose to use window messages, you should be able to get them just like any other
window message - for example:

// Add to the dialog's .h file:
  
afx_msg LRESULT OnMIMLongData(WPARAM wp, LPARAM lp);
 
 
// Add to the dialog's .cpp file:
 
// Add to the dialog's message map
ON_MESSAGE(MM_MIM_LONGDATA, &CMyDlg::OnMIMLongData)
 
LRESULT CMyDlg::OnMIMLongData(WPARAM wParam, LPARAM lParam)
{
   HMIDIIN MIDIDeviceHandle = (HMIDIIN)wParam;
   MIDIHDR *pMIDIHdr = (MIDIHDR *)lParam;
 ...
   return 0;
}

Mark

Mark Salsbery
Microsoft MVP - Visual C++

"Remember the five Ds of dogeball: Dodge, Dip, Duck, Dive, and Dodge."
"If you can dodge a wrench, you can dodge a ball."

rp_suman22-Jul-07 19:09

22-Jul-07 19:09

Mark Salsbery wrote:
You may get better performance if you use a callback function instead of window messages.

Yes, If I use callback function, there is no problem.

Mark Salsbery wrote:
What window handle are you passing when you call midiInOpen()?
When are you calling midiInOpen()?

I am calling both midiOutOpen() and midiInOpen in OnInitDialog() function.
The window handle passed is the pointer to application's main window handle as in following code:

HMIDIIN hdlMidiIn = NULL;<br />
MMRESULT mReturnCode=MMSYSERR_NOERROR;  <br />
CWnd* pMainDialog = (CWnd*)AfxGetApp()->m_pMainWnd;<br />
HWND hWnd = pMainDialog->m_hWnd;<br />
mReturnCode = midiInOpen(&hdlMidiIn,uhInID,(DWORD)hWnd,(DWORD)NULL,CALLBACK_WINDOW);<br />

Where, uhInID is MIDI port number.
Now I changed the midiInOpen as follows:
mReturnCode = midiInOpen(&hdlMidiIn,uhInID,(DWORD_PTR)MidiInProc,(DWORD)NULL,CALLBACK_FUNCTION);
Where MidiInProc() is the callback function.
Now I am calling the MIDI message processing function directly by creating object of Dialog class inside the callback. But I want to just send the message, so that callback load will be reduced.
Best Regards,
Suman

Mark Salsbery23-Jul-07 5:08

23-Jul-07 5:08

The message will go to the window handle you pass to midiInOpen().

If you pass the main window handle, you won't receive a message in a dialog, unless the dialog is
your main window. Make sure you pass the handle to the window you want to get messages in Smile | :)

Mark

Mark Salsbery
Microsoft MVP - Visual C++
Java | [Coffee]

rp_suman26-Jul-07 18:44

26-Jul-07 18:44

Hi Mark,
I changed the midiInOpen() to use CALLBACK_FUNCTION as it works fine.
Callback is working well, if I use SendMessage() or directly call the function for MIDI message processing from callback. The PostMessage() is not working well with callback. So, I have implemented SendMessage(). Is there any issues using SendMessage() in callback? Thank you for your great help!!
Best Regards,
Suman Smile | :)

rp_suman19-Jul-07 22:46

19-Jul-07 22:46

Hi Mark,
If I dispatch the message from OnKickIdle() I am getting the missing message most of the times. The code is:

LRESULT CMIDIDlg::OnKickIdle(WPARAM wp, LPARAM lp)<br />
{<br />
	MSG msg;<br />
	if(GetMessage(&msg,NULL,0,MM_MIM_LONGDATA)){<br />
		if(msg.message != WM_KICKIDLE)<br />
			DispatchMessage(&msg);<br />
	}<br />
	return 0L;<br />
}<br />

I am going to try some other methods like Win 32 project instead of MFC etc.
Thanks for the help!!
Best Regards,
Suman
P.S. As Saturday & Sunday are holiday I will come again on monday.
Mark, Dont foget to relax & rest while hard work.

Mark Salsbery20-Jul-07 5:08

20-Jul-07 5:08

You shouldn't be doing a GetMessage() call here because you've halted message processing for the
window until you get a message. That's not good Smile | :)

See my other reply above.

rp_suman wrote:
Mark, Dont foget to relax & rest while hard work.

I work at home - I always get to relax Smile | :)

Have a nice weekend!
Mark

Mark Salsbery
Microsoft MVP - Visual C++

"Remember the five Ds of dogeball: Dodge, Dip, Duck, Dive, and Dodge."
"If you can dodge a wrench, you can dodge a ball."

rp_suman22-Jul-07 18:48

22-Jul-07 18:48

Mark Salsbery wrote:
You shouldn't be doing a GetMessage() call here because you've halted message processing for the
window until you get a message. That's not good

See my other reply above.

Yes, you are right. I think, If Message comes inside OnKickIdle() means it should also comes if there is no OnKickIdle().
Best Regards,
Suman

Checking a valid pointer

Electronic7518-Jul-07 21:37

Electronic75

18-Jul-07 21:37

Hello, I have a vector(CMyData) of an MFC class CMyData in CMyDocumentClass. Because program have to display different sets of Data I usually pass a pointer from a member of this data vector that should be displayed to different view classes. In CMyDataListView, CMyDataPlotView classes before doing anything I first check the validity of pointer with

<br />
if(!AfxIsValidAddress(m_pData, sizeof(CMyData), TRUE))<br />
   return;<br />

this line usually is passed without problem indication I can read and write to memory location pointed by m_pData.
but Next line fails!

<br />
if(!m_pData->IsKindOf(RUNTIME_CLASS(CMyData)))<br />
  return;<br />

It says Access violation writing location 0x000000000
checking Call Stack shows that AfxIsValidAddress(...) has been actually executed without problem.

Can someone explain what is going on please? How should I check that a memory access failure won't happen.
By the way I have to add that in CMyData I have multiple vectors but please note that all this problems occure before I try to access such variables. Actually in the process of checking validity of pointer.

Thanks Smile | :)

James R. Twine19-Jul-07 1:01

James R. Twine

19-Jul-07 1:01

If the CMyData data object contains pointers itself, it is perfectally reasonable for AfxIsValidAddress(...) to say that a pointer to the object itself is valid, but it cannot evaluate the object's internals to see if its contained pointers are valid. For example, the object itself may have stepped on its runtime-type information.

Peace!

-=- James
Please rate this message - let me know if I helped or not!<HR>If you think it costs a lot to do it right, just wait until you find out how much it costs to do it wrong!
Avoid driving a vehicle taller than you and remember that Professional Driver on Closed Course does not mean your Dumb Ass on a Public Road!
See DeleteFXPFiles

David Crow19-Jul-07 2:53

David Crow

19-Jul-07 2:53

Electronic75 wrote:
if(!m_pData->IsKindOf(RUNTIME_CLASS(CMyData)))

Set a breakpoint on this statement. Is m_pData a CMyData object? Does it point to the same address as it did when first initialized?

"A good athlete is the result of a good and worthy opponent." - David Crow

"To have a respect for ourselves guides our morals; to have deference for others governs our manners." - Laurence Sterne

Mike Dimmick19-Jul-07 3:04

Mike Dimmick

19-Jul-07 3:04

AfxIsValidAddress in VS6 calls IsBadReadPtr or IsBadWritePtr depending on whether bReadWrite is FALSE or TRUE respectively. However, it's been discovered in the last few years that these APIs are seriously flawed. See IsBadXxxPtr should really be called CrashProgramRandomly[^] for details.

Visual Studio .NET 2003 still calls IsBadXxxPtr in Debug builds, but in Release builds only checks for NULL. In Visual Studio 2005, the IsBadXxxPtr checks are completely removed, the function only checks for NULL.

It does look like you've called through a bad pointer - this could well mean that the virtual function pointer inside CMyData is NULL, which it will be if you used memset, for example.

Stability. What an interesting concept. -- Chris Maunder