I have a CButton as part of a CDialog screen. Everything works fine, but after I reinitialize a hardware device, when I click anywhere on the screen (even outside the dialog) the event for the CButton in question fires. I've run through the code in debugger mode, and the app literally thinks the button in question is being pressed. Thing is, its NOT being pressed. I've experimented with clearing the message pump, but that didn't seem to help. If you need code, let me know and I'll see what I can drum up without giving away how our app works. I get the feeling this is an edge case that others have run into and know how to solve without code.
Thanks in advance!
Related
Fig1:
Fig2:
As in Fig1, after I clicked radio button, the radio button seems not get correctly painted, seems 50% transparent.
As in Fig2, after I move my mouse to hover the radio button again, it got correctly painted.
I have no idea about what is going on behind.
What are the possible reasons?
By the way, after I pressed a shortcut key PrtSc to get a screenshot(a freeware: Greeshot), this problem disappeared.
One hint I just found is: the dialog holding the radios were in modeless mode. The problem disappeared after I showed it as a modal dialog.
Actually, to achieve below goals, I am implementing the message pump for this dialog. Perhaps I am doing something wrong in the message pump logic. Continue to check the pump.
My Goal: user can interact with other windows while showing this dialog, one exception is: user should not be able to interact with the parent dialog of this dialog in question, so that database transaction may be issued prematurally.
I used to have a similar problem within wxWidgets. Turns out it's a long lasting bug that really never had been fixed. I now use a workaround with a simple wrapper function that 'enables' the element (places focus on it) and then 'unfocuses' it. Not sure in your current setup.
As a result, it seems I did something wrong in the message pump for the dialog in question.
I am implementing the message pump for this dialog.
To avoid user interact with the parent dialog of the dialog in question, I was ignoring message for the parent dialog and to children of parent dialog.
After I changed the above logic to below, the problem got solved.
EnableWindow(hParent, FALSE);
EnableWindow(hParent, TRUE);
So, I think I was doing something wrong in ignoring message, not sure exactly where.
So I have a MFC application which flick when we do some action.
So I figured I would set SetRedraw(false) and set it to true at the end of the function.
The application doesn't refresh anymore but if I click on it while SetRedraw(false), my cursor is not catched by my application, it goes throught it and set focus on the application below.
Anyone has some kind of idea how I could fix that.
I ended up using CWnd::LockWindowUpdate instead after some research.
It freezes the update but doesn't act if the window was transparent.
I'm trying to use SendMessage to post mouse clicks to a background window (Chrome), which works fine, but brings the window to front after every click. Is there any way to avoid that?
Before anyone says this is a duplicate question, please make sure that the other topic actually mentions not activating the target window, because I couldn't find any.
Update: aha, hiding the window does the trick, almost. It receives simulated mouse/keyboard events as intended, and doesn't show up on screen. However, I can just barely use my own mouse to navigate around the computer, and keyboard input is completely disrupted.
So my question is, how does sending messages to a window affect other applications? Since I'm not actually simulating mouse/keyboard events, shouldn't the other windows be completely oblivious to this?
Is it possibly related to the window calling SetCapture when it receives WM_LBUTTONDOWN? And how would I avoid that, other than hooking the API call (which would be very, very ugly for such a small task)?
The default handling provided by the system (via DefWindowProc) causes windows to come to the front (when clicked on) as a response to the WM_MOUSEACTIVATE message, not WM_LBUTTONDOWN.
The fact that Chrome comes to the front in response to WM_LBUTTONDOWN suggests that it's something Chrome is specifically doing, rather than default system behaviour that you might be able to prevent in some way.
The source code to Chrome is available; I suggest you have a look at it and see if it is indeed something Chrome is doing itself. If so, the only practical way you would be able to prevent it (short of compiling your own version of Chrome) is to inject code into Chrome's process and sub-class its main window procedure.
(Maybe this is a beginner question for people experienced with Windows message queue handling and dialg boxes. Unfortunately this is not my area of expertise, so please be gracious to me.)
I have a quite simple C++ program in terms of user interface.
It should not be a real console program of several reasons, but it runs unattended.
It is basically a MFC stub without showing a window at all. But sometimes it shows message boxes like :
MessageBox("Question,"XX",MB_YESNO) or so.
The problem is, that, when two questions are posed after each other, sometimes Windows seems to save the mouse click or keyboard click or the user wanted to klick only once, but the hardware send two clicks. So the user has not real possibility to answer the second question, but yes or no is answered from the "ghost" click before.
(There is a word for it, but I don't know it in English. I hope, you got my point though.)
In the command line there is a fflush() for such things. How to handle it here?
I would even use a customized Messagebox, if I find somewhere ready code for it
(and I have not to write it :-)
But I thought, there might be an easy-to-use snippet to delete the message queue of the app before the next messagebox is shown. But all I know about Windows messages is that they exist ;-(
Can anybody help me?
Philm,
Message box is a modal window, hence it has own message loop. When message box is dismissed, its message does not exist, no mouse messages.
Mouse click messages are always posted to a window under the cursor, unless GetCapture is called.
From what you claim, you do not show any window, so no mouse messages posted in the que?
The only way to resolve your problem would be to debug your application or the test project that you write to duplicate this problem. Can you write it and post somewhere for downloading?
I have this XP app (win32 C++) that I am just now testing under Windows7 (vista actually, but it does the same thing in windows 7).
I'm surprised that virtually the only issue I'm encountering is the following:
None of the scroll bars in a complex modelless dialog are functioning correctly. The main problem is the scroll thumb is not responding - just stays locked in position if you try and move it. Have had no issues going all the way back to win98, win2000, and winxp. Only in windows 7/Vista just now
But there is no commonality in the scrollbars in this dialog to explain it: One is in a plain richtext control created through a resource file. Another is in a richtext created through CreateWindow. And yet a third scrollbar is in a custom window class. None of them are working correctly (although you can make them scroll by right clicking and selected "Scroll Here".)
So I'm presuming maybe most encountered this a few years ago when porting to Window7/Vista for the first time, but I'm not finding anything in google now.
For modeless dialogs, you have to run IsDialogMessage in the main application GetMessage Loop, so messages for modeless dialogs are not subject to TranslateMessage and DispatchMessage. So I was doing that previously. However, Vista/Win7 doesn't like WM_MOUSEMOVE, and WM_LBUTTONDOWN and WM_LBUTTONUP to be bypassed like that for the dialog (i.e. they need to stay in the main App message loop). At least this was the problem in my case. I check for those message types now in the main message loop and that solved my problem. Can't explain it necessarily. Also couldn't explain why no one's encountered this previously (could be some idiosyncracy of my set up I guess). THanks for those who looked into this.