In our application, we have specialized modules in DLLs, which get control over a window inside our main application window. We pass the window as HWND, not as a CWnd*. Inside the DLL, I use CWnd::FromHandle to get a CWnd*, which I use as a parent to create a CHtmlView inside that window.
As soon as this view is about to be shown, a debug assertion occurs in wincore.cpp:
if(pMap)
{
ASSERT( (p = pMap->LookupPermanent(m_hWnd)) != NULL ||
(p = pMap->LookupTemporary(m_hWnd)) != NULL);
}
ASSERT((CWnd*)p == this); // must be us
// Note: if either of the above asserts fire and you are
// writing a multithreaded application, it is likely that
// you have passed a C++ object from one thread to another
// and have used that object in a way that was not intended.
// (only simple inline wrapper functions should be used)
//
// In general, CWnd objects should be passed by HWND from
// one thread to another. The receiving thread can wrap
// the HWND with a CWnd object by using CWnd::FromHandle.
//
// It is dangerous to pass C++ objects from one thread to
// another, unless the objects are designed to be used in
// such a manner.
Here the line with LookupTemporary fails, altough I used it exactly as described in the comment. In release mode, or if I just place an empty AssertValid-method in my view class, everything works fine.
My question: how can this happen, and is it something I should care about?
Related
A general question that could help answer another question I asked before. I believe that COM objects only return a pointer to the created object. So when I try to initialize a COM object twice, do I make two COM objects in my app, or do I make two different pointers to the same object? or perhaps there are two copies of the object that the pointers point to their respective objects?
// Pointer to WebViewController
static wil::com_ptr<ICoreWebView2Controller> webviewController;
// Pointer to WebView window
static wil::com_ptr<ICoreWebView2> webviewWindow;
// Locate the browser and set up the environment for WebView
CreateCoreWebView2EnvironmentWithOptions(nullptr, nullptr, nullptr,
Callback<ICoreWebView2CreateCoreWebView2EnvironmentCompletedHandler>(
[hWnd, url](HRESULT result, ICoreWebView2Environment* env) -> HRESULT {
// Create a CoreWebView2Controller and get the associated CoreWebView2 whose parent is the main window hWnd
env->CreateCoreWebView2Controller(hWnd, Callback<ICoreWebView2CreateCoreWebView2ControllerCompletedHandler>(
[hWnd, url](HRESULT result, ICoreWebView2Controller* controller) -> HRESULT {
_com_error err4(result);
MessageBoxW(hWnd, L"Controller setup", 0, MB_OK);
MessageBoxW(hWnd, err4.ErrorMessage(), 0, MB_OK);
if (controller != nullptr) {
webviewController = controller;
webviewController->get_CoreWebView2(&webviewWindow);
}
// Add a few settings for the webview
// The demo step is redundant since the values are the default settings
ICoreWebView2Settings* Settings;
webviewWindow->get_Settings(&Settings);
Settings->put_IsScriptEnabled(TRUE);
Settings->put_AreDefaultScriptDialogsEnabled(TRUE);
Settings->put_IsWebMessageEnabled(TRUE);
// Resize WebView to fit the bounds of the parent window
RECT bounds;
GetClientRect(hWnd, &bounds);
webviewController->put_Bounds(bounds);
// Schedule an async task to navigate to Bing
webviewWindow->Navigate(url);
if (GetWindow(hWnd, GW_CHILD) != NULL) {
MessageBoxW(hWnd, L"not a NULL child window", 0, MB_OK);
};
// Step 4 - Navigation events
// Step 5 - Scripting
// Step 6 - Communication between host and web content
return S_OK;
}).Get());
return S_OK;
}).Get());
I called
CoInitialize(NULL, COINIT_APARTMENTTHREADED | COINIT_DISABLE_OLE1DDE); earlier in the code only once. My question is what happens when I run this block twice?
My question is what happens when I run this block twice?
In this specific case, executing the code twice will produce two distinct WebView2 instances. Both instances are then assigned to the same static variable, causing one instance to be irrevocably leaked.
Being an XY Problem this information isn't terribly useful. The real issue you are trying to solve is knowing when the WebView2 control is fully constructed. That question has been answered elsewhere.
So when I try to initialize a COM object twice, do I make two COM objects in my app, or do I make two different pointers to the same object? or perhaps there are two copies of the object that the pointers point to their respective objects?
That really depends on the implementation of the COM object. If it is implemented as a singleton, then instantiating it multiple times will return a pointer to the existing instance (creating a new instance if it doesn't exist yet). But, if the COM object is not implemented as a singleton, then each instantiation will return a pointer to a new instance.
I have followed this link to get the window handle of a ActiveX control
Sample Code from microsoft's site
// The following code should return the actual parent window of the ActiveX control.
HWND CMyOleControl::GetActualParent()
{
HWND hwndParent = 0;
// Get the window associated with the in-place site object,
// which is connected to this ActiveX control.
if (m_pInPlaceSite != NULL)
m_pInPlaceSite->GetWindow(&hwndParent);
return hwndParent; // Return the in-place site window handle.
}
But in my case I keep finding that "m_pInPlaceSite" is always NULL. I'm trying to run this code in my controls FinalConstruct. Is there something else I need to implement for the m_pInPlaceSite to be given a value? Or do I need to Query to get the value.
Thanks
FinalConstruct is way too early. In FinalConstruct your class is just being created and is not yet initialized. There is no "in place" site, there is no yet site at all.
Your control will be called by its owner, it will be given its site, then activated - only then you will possibly have m_pInPlaceSite available.
So I have this MFC dialog program I am working with. The dialogs are written but now I am having difficulty passing data around from dialog to dialog. I have the following structure _dlgDataHandler set up in a class derived from CWinApp and have have created a "new" statement for a pointer to this type.
//.......SRK.h file
class CSRK_App : public CWinApp
{
public:
CFSB_App();
// added the following data structure for data passing withing the program
typedef struct _dlgDataHandler {
char RepetitionRadio[24];
// another member
// yet another member and so on as necessary
} *dlgDataHandlerPtr;
// extern dlgDataHandlerPtr dlgDataHandler;
// Overrides
// ClassWizard generated virtual function overrides
//{{AFX_VIRTUAL(CSRK_App)
public:
virtual BOOL InitInstance();
//}}AFX_VIRTUAL
// Implementation
//{{AFX_MSG(CSRK_App)
// NOTE - the ClassWizard will add and remove member functions here.
// DO NOT EDIT what you see in these blocks of generated code !
//}}AFX_MSG
DECLARE_MESSAGE_MAP()
};
//....... SRK.cpp A pointer to a new dataHandler created in this block about 2/3 the way down
// CSRK_App initialization
BOOL CSRK_App::InitInstance()
{
AfxEnableControlContainer();
// Standard initialization
// If you are not using these features and wish to reduce the size
// of your final executable, you should remove from the following
// the specific initialization routines you do not need.
//SetRegistryKey(_T("Local AppWizard-Generated Aplications"));
#ifdef _AFXDLL
Enable3dControls(); // Call this when using MFC in a shared DLL
#else
Enable3dControlsStatic(); // Call this when linking to MFC statically
#endif
//CSRK_Dlg dlg;
CDialogMain dlg("SRK - Beta"); // added 12/27 **
m_pMainWnd = &dlg;
//const char* m_pszHelpFilePath = NULL;
//free((void*)m_pszHelpFilePath);
//m_pszHelpFilePath=_tcsdup(_T("c:\SRKHelp.rtf"));
// the following line added to allocate memory for the structure
dlgDataHandlerPtr dlgDataHandler = new _dlgDataHandler;
dlg.SetWizardMode(); // added 12/27 **
int nResponse = dlg.DoModal();
if (nResponse == IDOK)
{
// TODO: Place code here to handle when the dialog is
// dismissed with OK
}
else if (nResponse == IDCANCEL)
{
// TODO: Place code here to handle when the dialog is
// dismissed with Cancel
}
// Since the dialog has been closed, return FALSE so that we exit the
// application, rather than start the application's message pump.
return FALSE;
}
In the dialog .cpp files, there are five, I need to be able to get data from the AFX variables "m_" and load them into this dataHandler Structure (or another one like it) so that they can be used in other dialogs and parts of the program, specifically my actual code when all the dialog data collection is done. Someone said to use AfxGetApp() so that I could have a handle on the current instance but I do not understand what they are talking about. And yes I have read about it in many forums, I just don't get it. I further realize this is probably not the best way to do it. I am trying to learn MFC/OOP with what time I have available, but for now, I am just trying to get a handle on the basic process as I can tune it later once I understand how to collect and pass simple data around.
I further don't understand how calling AfxGetApp() will help me get a handle on the members of CSRK_App. It inherited CWinApps public members but AfxGetapp() can't see what CSRK_App has... can it?
First, to explain the AfxGetApp advice you have received. There is some extra hand-waving using 'new' and a pointer, but this is basically using a global variable for the structure that holds your data. This is not the best way to do what you are trying to do. There are many pitfalls.
AfxGetApp() is an MFC call that returns a pointer to your main App class derived from CWinApp.
If you want to use that returned pointer, you need to cast it as a CSRK_App* pointer with:
CSRK_App* pApp = static_cast <CSRK_App*> ( AfxGetApp());
Then you can use pApp->dlgDataHandlerPtr->... to access the variables you need.
Now, for the pitfalls. Someone else may chime in with a reason why the 'new' and the pointer are helpful, but I do not see any advantage to this approach as compared to just having a local variable dlgDataHandler inside your CSRK_App class. That would simplify the code.
The next issue is that all your data is public in a struct. Any dialog class that can call AfxGetApp can read or write any data in that struct. You have no way to control access.
Also, all of your dialog classes must now include SRK_App.h so they know the structure, and have access to all other variables in that App class.
A cleaner, object-oriented approach would be to declare the struct (class) for the data in a separate .h file that could be included in the dialog classes. Then, you would pass a pointer/reference to this data into the constructor of the dialog classes. The dialog class would have no need to know anything about the App class.
For an even higher level of segregation, the dialog classes can be written so they only get a copy of the dlgDataHandler class passed in before calling .DoModal(), and then after the DoModal call returns with IDOK, the App class can have control over which data from the dialog gets updated into the dlgDataHandler class. The advantage of this approach is that it insures that no matter how the dialog class is programed, the user can always "Cancel" the dialog without modifying any data.
It is necessary to switch off the "ClearType" property of the default font for all dialog controls. It is possible to do that for one control by setting
logfont.lfQuality = ANTIALIASED_QUALITY
There are a lot of suggestion how to do the same for modal dialogs (http://neelaakash.wordpress.com/2007/12/31/change-default-dialog-font-of-cdialog/ and others), but that should be done for non-modal dialogs (are instantiated with new and Create(...) methods). I've tried to do that myself:
Override 'Create' method, and modify dialog template:
BOOL CActivationChildDialogLicenseInfo::Create(UINT nIDTemplate,
CWnd* pParentWnd)
{
CDialogTemplate dlt;
int nResult;
// load dialog template
if (!dlt.Load(MAKEINTRESOURCE(nIDTemplate))) return -1;
// set your own font, for example “Arial”, 10 pts.
dlt.SetFont(L"Arial", 12);
// get pointer to the modified dialog template
LPSTR pdata = (LPSTR)GlobalLock(dlt.m_hTemplate);
// let MFC know that you are using your own template
m_lpszTemplateName = NULL;
InitModalIndirect(pdata);
// display dialog box
nResult = CActivationChildDialog::Create(nIDTemplate, pParentWnd);
// unlock memory object
GlobalUnlock(dlt.m_hTemplate);
return nResult ;
}
Seems like this method do nothing (it is called, I've checked that with putting break-point inside).
I've tried to call
nResult = CActivationChildDialog::Create(NULL, pParentWnd);
...but got a lot of ASSERTs.
I've also tried to override the 'OnSetFont' method:
void CActivationChildDialogLicenseInfo::OnSetFont(CFont *pFont)
{
CActivationChildDialog::OnSetFont(pFont);
LOGFONT logfont;
pFont->GetLogFont(&logfont);
LOGFONT logfont2=logfont;
pFont->DeleteObject();
logfont2.lfItalic = true;
logfont2.lfQuality = ANTIALIASED_QUALITY;
pFont->CreateFontIndirect(&logfont2);
}
That causes ASSERT during run-time and resulted in a VERY big font being used (lost default font settings, doesn't accept new specified settings)... I don't know why.
Please advise, how can I change a default dialog font that will be "inherited" by all dialog controls?
Thank you very much.
First off: the simple, reliable way to do this is to create the dialog and then send WM_SETFONT (or call SetFont()) to the dialog and each control in it. I'll show you how to do that below, but first, here's why the two strategies you've already tried didn't (and can't) work:
Modifying the dialog template
First off, you should be calling CDialog::CreateIndirect() if you wish to use a dialog template that you've already loaded.
But don't bother. The dialog template contains only the face name and point size - it does not allow you to specify other LOGFONT values such as lfQuality. If it did, you could simply specify that in your resource definition and avoid writing any runtime code at all!
Intercepting WM_SETFONT
In theory, you could make this work. But it's not practical. Your code has several problems: first off, you'd have to intercept this message for every child control in order for it to do anything useful: the dialog itself probably doesn't render any text. But worse, you're passing the original font to the base class (which hands it to the default window procedure, which stores it internally for later use) and then immediately destroying it - this means the dialog (and everything else using that font, including all of the child controls) will be trying to draw text using a bogus font, and reverting to the default font as a result. Finally, you're creating a new font attached to a temporary object (pFont) created and destroyed by MFC - internally, the CFont object you're working with will be detached from the font handle and destroyed, leaking a handle to a font object that nothing uses.
Leaky abstractions: a note on HFONT and CFont
HFONT is a handle type that Windows uses to represent a font object. Like most of GDI, there are specific functions for creating fonts, and the general-purpose DeleteObject() function for destroying them.
CFont is a light-weight wrapper class for HFONTs. A CFont instance can be attached and detached from an existing HFONT, or used to create a new one. If a CFont instance is still attached to a HFONT when its deconstructor executes, it will call DeleteObject() to destroy the underlying object. Internally, MFC utilizes temporary CFont instances that are attached and detached from HFONTs when calling various message handlers (such as OnSetFont). It's worth remembering that internally, Windows knows nothing about CFont, and a single HFONT may belong to 0 or more CFont instances at any given point in time.
A note on fonts and WM_SETFONT
When you create a new font - whether or not it is wrapped in a CFont object - you are the owner of that font, and it is your responsibility to destroy it once you are finished using it. Passing it to WM_SETFONT (CWnd::SetFont()) doesn't change ownership! This is actually quite useful, as it allows you to pass the same font to multiple windows without worrying about which one will destroy it - you're still the owner, and so you can (and must) destroy it yourself (once there are no windows still using it).
Finally - how to quickly create and set a font on a dialog and all its children
So you should now have enough background to understand the necessary steps:
Create the dialog
Create the desired font
Pass the font to the dialog and its children (by either sending WM_SETFONT messages, or by calling CWnd::SetFont... which itself send a WM_SETFONT message).
When the dialog is destroyed, also destroy your font.
Example
// define this as a class member - class destructor then handles step four!
CFont m_nonCleartypeFont;
BOOL CActivationChildDialogLicenseInfo::Create(UINT nIDTemplate,
CWnd* pParentWnd)
{
// step one: create dialog normally
BOOL nResult = CActivationChildDialog::Create(nIDTemplate, pParentWnd);
// step two: create custom font
// relying on destructor to destroy font once we're done with it
// so be careful to only create it once!
if ( NULL == m_nonCleartypeFont.m_hObject )
{
CFont* pOriginalFont = GetFont(); // use template font as... template!
// pull information from original font
LOGFONT logfont;
pOriginalFont->GetLogFont(&logfont);
// make font adjustments:
// specify italics
logfont.lfItalic = true;
// and non-cleartype antialiasing
logfont.lfQuality = ANTIALIASED_QUALITY;
// create our font based on adjusted information
m_nonCleartypeFont.CreateFontIndirect(&logfont);
}
// step three: set our custom font on the dialog and all children
SetFont(&m_nonCleartypeFont, FALSE);
// Send message to quickly set this font for all children.
// See documentation for SendMessageToDescendants()
// - this is actually the example given!
SendMessageToDescendants(WM_SETFONT,
(WPARAM)m_nonCleartypeFont.m_hObject,
MAKELONG(FALSE, 0),
FALSE);
return nResult;
}
After switching from VS2005 to VS2008 SP1, I found an issue that I can't explain.
A program works fine under VS2005 in both release and debug mode. Under VS2008, when entering the debugger an assert is raised.
If I let the program run (in debug or release mode), no assertion at all.
I spent almost two days on this and I don't understand what I do wrong.
Description of the program:
I have a MFC dialog based program that creates a user thread (CWinThread) that creates the main dialog of the application.
A worker thread loops infinitely and posts each second a message to the dialog. The message is processed in the gui thread.
Some parts of my code:
The InitInstance of the gui thread:
BOOL CGraphicalThread::InitInstance()
{
CGUIThreadDlg* pDlg = new CGUIThreadDlg();
pDlg->Create(CGUIThreadDlg::IDD);
m_pMainWnd = pDlg;
AfxGetApp()->m_pMainWnd = pDlg;
return TRUE;
}
The worker thread:
UINT ThreadProc(LPVOID pVoid)
{
do
{
AfxGetApp()->m_pMainWnd->PostMessage(WM_APP+1, (WPARAM)new CString("Hello"), NULL);
Sleep(1000);
}
while(!bStopThread);
return 0;
}
The dialog message handler is like this:
LRESULT CGUIThreadDlg::OnMsg(WPARAM wp, LPARAM lp)
{
CListBox* pList = (CListBox*)GetDlgItem(IDC_LIST1);
CString* ps = (CString*)wp;
pList->InsertString(-1, *ps);
delete ps;
return 1L;
}
This works perfectly fine with VS2005.
But with VS2008, but as soon as a put a breakpoint and enter the debugging mode, I have an assertion raised ???
wincore.cpp line 906
CObject* p=NULL;
if(pMap)
{
ASSERT( (p = pMap->LookupPermanent(m_hWnd)) != NULL ||
(p = pMap->LookupTemporary(m_hWnd)) != NULL);
}
ASSERT((CWnd*)p == this); // must be us
// Note: if either of the above asserts fire and you are
// writing a multithreaded application, it is likely that
// you have passed a C++ object from one thread to another
// and have used that object in a way that was not intended.
// (only simple inline wrapper functions should be used)
//
// In general, CWnd objects should be passed by HWND from
// one thread to another. The receiving thread can wrap
// the HWND with a CWnd object by using CWnd::FromHandle.
//
// It is dangerous to pass C++ objects from one thread to
// another, unless the objects are designed to be used in
// such a manner.
If I remove the GUI thread and create the dialog into the CWinApp thread, the problem doesn't occur anymore.
Does anybody have any idea?
Am I doing something wrong?
Thank you
// Note: if either of the above asserts fire and you are
// writing a multithreaded application, it is likely that
// you have passed a C++ object from one thread to another
// and have used that object in a way that was not intended.
// (only simple inline wrapper functions should be used)
//
// In general, CWnd objects should be passed by HWND from
// one thread to another. The receiving thread can wrap
// the HWND with a CWnd object by using CWnd::FromHandle.
//
// It is dangerous to pass C++ objects from one thread to
// another, unless the objects are designed to be used in
// such a manner.
#Ismael: I had already tried that the assert is still fired. The only way I found to remove the assert is to create the dialog into the CWinApp thread.
But this doesn't explain what happens since there's still the worker thread that post to the dialog every second.
Anyway , thanks.
#daanish.rumani: I've checked the wincore.cpp and the CWnd::AssertValid() is exactly the same (but there's of lot of differences in the rest of the files).
I would accept that a piece of code works with VS2005 and not VS2008, but
I can't see what I do wrong.
If I do something wrong, what is the correct way to proceed?
Why the assert is only fired when a breakpoint is hit and I step over the Sleep call?
I can run the program fine, even when its compiled in debug mode, as long as I don't enter the debugger.
Could it be a bug in the debugger?