Re: Get/Kill focus in MFC - C / C++ / MFC Discussion Boards

Yes, you are right. It is about repainting the dialog border.
I tried SetWindowPos with SWP_NOACTIVATE|SWP_SHOWWINDOW and the first parameter as &wndTopMost and &wndTop. With &wndTop the window is invisible, with &wndTopMost it is visible. On both occasions the dialog window loses focus and the frame is repainted.
It happens no matter what the first parameter of the SetWindowPos you passed.

geoyar

How does your code look like now?

> The problem with computers is that they do what you tell them to do and not what you want them to do. <
> "It doesn't work, fix it" does not qualify as a bug report. <
> Amazing what new features none of the programmers working on the project ever heard of you can learn about when reading what the marketing guys wrote about it. <

Sorry, I have deleted all SetWindowPos from my code.
I remember that in my other app I used a modeless dialog box with CStatic in it over the modal dialog box. I have no problem with focus, placement and visibilty of this box.

So I created a modeless dialog box CTipDlg m_tipDlg and added to OnRButtonClick:

OnRButtonDown(UINT nFlags, CPoint point)
{
m_tipDlg.Create(IDD_TIPDLG, this);
m_tipDlg.SetLayeredWindowAttributes(RGB(10, 10, 10), 0, LWA_COLORKEY);
m_tipDlg.ShowWindow(SW_SHOW);

.....................................................................
}

Now I see two dialogs, one modal, one modeless(layered, transparent and not topmost), with two active frames. The focus stays on control I have clicked.
It is the test of concept for now. It might not work for me at the end. But focus behaves as I wanted.

Of course it is overkill, because all I need is a visible window.
Of course I will look in CDialog MFC code.

Any help will greatly appreciated.

geoyar

I'd like to help, but i don't see how...

> The problem with computers is that they do what you tell them to do and not what you want them to do. <
> "It doesn't work, fix it" does not qualify as a bug report. <
> Amazing what new features none of the programmers working on the project ever heard of you can learn about when reading what the marketing guys wrote about it. <

I coded a modeless dialog in and it works like the topmost window: the main dlg frame is changing. So I am again at square one.

geoyar

Finally, I did it.
It is about processing WM_NCACTIVATE in the main dialog.
Just like

BOOL CSliderGDIDlg::OnNcActivate(BOOL bActive)
{
if (bActive)
return CDialog::OnNcActivate(bActive);
return FALSE;
}

Of course the condition might and should be more sofisticated.

geoyar

All right, nice that it works, i don't see why this is better than the other solution but i don't see the whole picture so...congratulations. Thumbs Up | :thumbsup:

> The problem with computers is that they do what you tell them to do and not what you want them to do. <
> "It doesn't work, fix it" does not qualify as a bug report. <
> Amazing what new features none of the programmers working on the project ever heard of you can learn about when reading what the marketing guys wrote about it. <

After a bit of trying and debugging this is what I got:

1. You are right: SetWindowPos with SWP_NOACTIVATE does not send WM_NCACTIVATE to other windows. So

SetWindowPos(&wndTopMost, x, y, cx, cy, SWP_SHOWWINDOW|SWP_NOACTIVATE);

does not change the main dialog/window frame. But control that have received mouse right click and created popup topmost window does lose focus. I had to restore focus manually, using SetFocus().

2. If the position and size of the window are not changing SetWindowPos will not redraw window if you do not call Invalidate() or InvalidateRect/Region() before call to SetWindowPos.

3. The window must be created without WS_VISIBLE. Other way even with WS_EX_NOACTIVALTE main dialog frame changes.

This is a code:

BOOL CTipWnd::CreateTipWnd()
{
BOOL bRes = FALSE;
bRes = CreateEx(WS_EX_LAYERED|WS_EX_TRANSPARENT|WS_EX_TOPMOST|WS_EX_NOACTIVATE,
AfxRegisterWndClass(CS_HREDRAW|CS_VREDRAW|CS_SAVEBITS),
NULL,
WS_POPUP,
0, 0, 0, 0,
NULL,
NULL,
0); // Not visible
if (bRes)
{
CRect tipWndRect = SetTipWndRects();
bRes = SetLayeredWindowAttributes(bkgndCol.ToCOLORREF(), 0, LWA_COLORKEY);
SetWindowPos(&wndBottom, tipWndRect.left, tipWndRect.top, tipWndRect.Width(),
tipWndRect.Height(), SWP_NOACTIVATE); // Still not visible
}

return bRes;
}

void CTipWnd::UpdateTipWnd(..., bool bShow)
{
// Changing window layout
...........................
// Done. Now redraw and show it
if (bShow)
{
Invalidate(FALSE);
CRect tipR;
GetWindowRect(&tipR);
SetWindowPos(&wndTopMost, tipR.left,tipR.top, tipR.Width(), tipR.Height(),
SWP_SHOWWINDOW|SWP_NOACTIVATE);
}
}

And in calling control

void CSliderGdiCtrl::OnRButtonDown(UINT nFlags, CPoint point)
{
m_tipWndPtr = new CTipWnd;
m_tipWndPtr->SetOwner(this);
..............
BOOL bRes = m_tipWndPtr->CreateTipWnd();
if (!bRes)
{
delete m_tipWndPtr;
m_tipWndPtr = NULL;
}
else
{
UpdateTipWnd();
SetFocus();
}
}
..................
}

Thank you so much for your help.

geoyar

Yourwelcome.

> The problem with computers is that they do what you tell them to do and not what you want them to do. <
> "It doesn't work, fix it" does not qualify as a bug report. <
> Amazing what new features none of the programmers working on the project ever heard of you can learn about when reading what the marketing guys wrote about it. <

I have MFC-based SDI GUI application.

Also I have a thread which contains custom messages queue. Each of these messages should be posted to GUI, most of them adds new items to ListView.

If I'm just doing PostMessage( WM_MY_CUSTOM_MESSAGE ) to main GUI thread I have 2 undesirable effects:
1. Main thread queue is overflowed and I have message-leaking;
2. GUI becomes unresponsible - windows aren't repainted, I can't push toolbar and dialog buttons and so on;

Question:
How can I organize posting messages to GUI from the secondary thread and avoid 2 above effects (especially the second one)?

In your app where the message is handled, you can try to SetRedraw(FALSE) when processing the first message and then SetRedraw(TRUE) when all messages are received. You'll likely need to add a complete indicator with the message to accomplish this. Smile | :)

Chris Meech
I am Canadian. [heard in a local bar]

In theory there is no difference between theory and practice. In practice there is. [Yogi Berra]

Is there a tool available to process source code to detect mismatches between the sprintf format string and the arguments that follow? For example specifying a %s and neglecting to provide a c string to satisfy it, etc.

I use hundreds of sprintf's to format log messages and am looking for a way to insure correctness without eyeballing each and everyone of them.

When this happens during runtime in my code the program thread seems to hang and when it's the main server thread the other secondary threads cease to get serviced and as they are designed to timeout, they do, and a programmatic exit occurs.

If you're using C++ don't fart about with sprintf, use std::stringstream - you'll avoid problems of lobbing the wrong thing at a particular format specifier.

If you're stuck using C then there's not a lot I can suggest although PC-Lint is pretty good at finding errors in the printf family of functions.

Cheers,

Ash

Alan Kurlansky wrote:
detect mismatches between the sprintf format string and the arguments that follow?

Yes, there are a couple of tools available. Have a look at the following:

Splint checks for vulnerabilities and coding mistakes
PC-Lint the "classic" code analysis tool
Dr Dobbs had a full article about code analysis
also see Wikipedia: tools for static code_analysis for C and C++

Happy debugging! Smile | :)

M

PS: This post is for developers who prefer format() over STL string streams.
PPS: And if you are one of them, you could have a look at Boost Format library. Smile | :)

Chat in Europe Java | [Coffee]

Now with 24% more Twitter

I have a C++ dll, lib and the corresponding header file. I would like to create a dummy wrapper dll that simply returns True from some of the functions in the dll, so that when testing, I can replace the dll with the test one and have the application still work without recompilation.
Some function however (error handling) I would like to keep. My approach is as follows.

1. Copy the .h file in my local project folder and remove the extern keyword for the functions I would like to override, than add implementation for those functions.
2. All functions are declared __declspec( dllexport ) so that they are included in the dll.
3. I include the original lib file as an Additional dependency in the project, so I can link to the function definitions for the functions I'm not overriding. The project is set to DLL (I'm using VS)

Everything compiles fine, but when I use depends to look at the dll, I only see the functions I've implemented, not the ones from the lib.

Any ideas?

So you expect the dll to automagically implement the functions you didn't write? If you find a compiler to do that let me know (private email please, I don't want everyone else to know).

Phil

The opinions expressed in this post are not necessarily those of the author, especially if you find them impolite, inaccurate or inflammatory.

You really have a fantastic signature! Great!

I expect to link to the static lib file and export those implementations. Or are those just stub functions that load the dll code?

Ah ha, just seen one point of confusion you've got...

When you produce a DLL you can produce something called an import library. This is nothing like an ordinary static library (although you can add object files to it if you want) but essentially just a lump of data describing the functions exported from the DLL - essentially just the function's names and position in the exported function table. They don't contain any code.

So if you link with a DLLs import library then all that means is that the corresponding DLL will be automatically loaded when the program starts and the loader can fixup the addresses of the imported functions.

Cheers,

Ash

Yeah I realized that after a while Smile | :)

I wouldn't try using library interpositioning for this sort of thing. While you can implement a DLL that calls selectively forwards functions to another DLL it ends up being a bit of a headache [1]. The approach I tend to use is...

- wrap the DLL in a class
- extract an interface class out of the class you've implemented (all the functions are pure virtual)
- implement the interface again, passing the bits you want through to the DLL and reimplementing the bits you don't want to
- select at runtime or compile time (depending on the type of testing you want to do) which implementation of the interface you want to use

This method is something to consider with any third party library so when it doesn't work properly you've got a central point to either replace it or patch its functionality.

Cheers,

Ash

[1] If you really want to do use library interpositioning then...

- reimplement the entire interface of the wrapped DLL, starting off by just returning succeeded, true, whatever is appropriate
- reimplement the functions you want to pass through by (a) manually loading the wrapped DLL and (b) calling the wrapped function
- reimplement the bits you don't want to pass through however you want

This saves the balls ache of link order of the wrapping and wrapped libraries.

So I can't simply use the external function definitions to expose those functions in the dll and statically link the lib file to the dll for the actual implementation??

Wrapping the dll in a class sounds like an overkill, I wanted something quick and dirty. If I try to use the library interpositioning you described at the bottom of your post, won't the functions I import from the wrapped dll clash with the functions I defined in the wrapper, being the same name and signature??

Not sure what you're talking about in the first paragraph (sorry!). When you define a function you're actually writing the code for it. At least in C++ terms, not sure what the C standard says about definitions.

Wrapping the DLL in a class may or may not be overkill. It's guarenteed to work and gives you a nice boundary to unit test around by adding something that's easily substitutable. If the DLL has a couple of hundred functions you're trying to use then it might be a headache (but then the DLL itself is likely to already be that) but for 20 or so functions it's less than an hours work.

As to the interpositioning I described... Why would there be any clash of names? You're not linking with the wrapped DLL so there are no names from it to clash with.

Cheers,

Ash

So doing dll import would not name clash with my local functions ?

Nope, I didn't say that. I said that if you don't link with the DLL then there'll be no names to clash with.