|
I've a problem I've never seen before. One application launches one or more child processes. Rarely, one of those child processes doesn't completely exit. It doesn't show up in task manager, but does show up in Resource Monitor, Process Explorer and by calling a CreateToolhelp32Snapshot/Process32First. When I try to open the process, it fails (and Process Explorer shows question marks for all handles, which would seem to me that they have been closed.)
Anyone have a clue?
PS. This is on Windows 7 64 and both the main application and child applications are 64-bit. Both are in C++ and use Qt. I just inherited this project so am not yet familiar with all the wiring, but I've never seen this behavior before.
|
|
|
|
|
Hey Joe,
Unfortunately... it means there is a bug in your code somewhere. I call them 'Zombie processes' and they are generally caused by a handle leak. You can create a process zombie by launching a child process duplicating and never closing its handle. The process does not show up in the task manager because the process has exited. The zombies are consuming system resources and next will be your brain. 
Best Wishes,
-David Delaune
|
|
|
|
|
I am trying to set the font for radio button using below code in my project but font does not set. Font and color does not set for text on radio button.
I have a dialog base application (MFC), at dialog there are some radio buttons
In dialog class, Oninitdialog() i use
CFont* pFont = GetDlgItem( IDC_RADIO1 )->GetFont();
LOGFONT LogFont = { 0 };
pFont->GetLogFont( &LogFont );
LogFont.lfItalic = TRUE;
LogFont.lfWidth = 9;
LogFont.lfHeight = 12;
LogFont.lfWeight = FW_BOLD;
memcpy(LogFont.lfFaceName,"Verdana",7);
m_StaticFont.CreateFontIndirect( &LogFont );
GetDlgItem( IDC_RADIO1 )->SetFont( &m_StaticFont );
in header file
CFont m_StaticFont;
To set the color for text on radio button, i use OnCtlColor (WM_CTLCOLOR)
HBRUSH CtestxmlDlg::OnCtlColor(CDC* pDC, CWnd* pWnd, UINT nCtlColor)
{
///
if(pWnd->GetDlgCtrlID() == IDC_RADIO1)
pDC->SetTextColor(RGB(255,0,0));
////
}
But when i use above code in a test program it works. Please suggest what's wrong.
I have debug my project and found OnCtlColor calls and it set the text color also.
Please help.
|
|
|
|
|
How is this test program different from the dialog based program that you mentioned first?
|
|
|
|
|
Both are dialog based applications. Test program contains only one radio button.
In main program there are lot of controls and work against them
|
|
|
|
|
As you can see yourself, the code works.
So that means what you've done is correct.
That must mean, there is something else going wrong in the first program.
Maybe a wrong ID has been used.
You could debug the code to get the answer.
|
|
|
|
|
The font is set but text color is not set in first program.
This is the problem with radio buttons. If i add a new radio button then also text color is not set.
But for static control the text color is set.
It means something needs to be supported for radio button
|
|
|
|
|
I found the reason why.I am using XP theme. Below url also says that
http://www.go4expert.com/forums/showthread.php?t=16457&page=2
If i remove xp theme then color is set for radio text.
Is there any solution for that problem.It's require for me to use XP theme.
XP theme
="1.0"="UTF-8"="yes"
<assembly
xmlns="urn:schemas-microsoft-com:asm.v1"
manifestVersion="1.0">
<assemblyIdentity
processorArchitecture="x86"
version="5.1.0.0"
type="win32"
name="appname.exe"/>
<description>appname</description>
<dependency>
<dependentAssembly>
<assemblyIdentity
type="win32"
name="Microsoft.Windows.Common-Controls"
version="6.0.0.0"
publicKeyToken="6595b64144ccf1df"
language="*"
processorArchitecture="x86"/>
</dependentAssembly>
</dependency>
</assembly>
|
|
|
|
|
Hi all,
I just want to know one thing that when we insert items in List Control its memory increases by 4k. I wanted to know that this memory increase also depends on number of columns we insert in each row.
Thanks in advance
|
|
|
|
|
Did you give it a try? I mean just include another column and see for yourself
I know I am coward since the day I know that fortune favors the brave
|
|
|
|
|
Hi all
I have the following code
CString s;
s = "CSECT";
do {
memset(&buffer[0],0x20,121);
pFile->ReadString(&buffer[0],123);
} while(s.Find((LPCTSTR) &buffer[0]) == -1);
When buffer has the following the return code remains -1
000000 00000 000BC 10 TESTA CSECT
|
|
|
|
|
As I read the docco[^], you've got your find inside-out. Your code is looking for the contents of the buffer inside "CSECT". No great surprise that it's not found.
Software rusts. Simon Stephenson, ca 1994.
|
|
|
|
|
|
Several things could be going wrong here.
First - As Peter said, you've got it inside out.
Second - Even though you're assigning a non-unicode string to a CString object, it would be converted to a unicode string if you're doing a unicode build.
Third - Looks like the buffer contains binary data. If so, it would not work with a string object since it would look at a 0 as the null terminating character to end the string.
So, I recommend you use the memcmp function to find the string - memcmp(buffer, (LPVOID)(LPCTSTR)s, 121);
|
|
|
|
|
Hello,
is there a convenient way to pass a pointer to a data member so that it is recognized as a template argument?
The following does not compile under VC++ 2008:
template<class Parent, typename Type, Type Parent::*Member>
void AddMember(Type Parent::*Member)
{
return true;
}
...
AddMember(&TestObject::i);
Alex
|
|
|
|
|
Not completely sure... but did you try just this:
void AddMember(Type *Member)
If you're calling a method for it, no need to specify Parent::*Member, where the pointer comes from is up to the caller.
|
|
|
|
|
Try this -
template<class Parent, typename Type>
void AddMember(Type Parent::*Member)
{
return true;
}
|
|
|
|
|
Thank you for your answers.
Please look at this example:
struct TestObject
{
int i;
};
template<int TestObject::*Member>
void test() {}
...
test<&TestObject::i>();
This way, I can call the function template so that the member pointer is passed as a (non-type) template argument. However, in my case, I would like that parent and type are template parameters, too:
struct TestObject
{
int i;
};
template<class Parent, typename Type, Type Parent::*Member>
void test() {}
...
test<Testobject, int, &TestObject::i>();
However, this is not very convenient to call test() as you have to specify all template arguments manually.
If you pass &TestObj::i as a argument to the function, Parent and Type are deduced - but not the Member...
Alex
|
|
|
|
|
I understand that you want the parent and type to be template parameters.
So let me repeat my earlier code snippet -
template<class Parent, typename Type>
void AddMember(Type Parent::*Member)
{
return true;
}
Here Parent and Type are template parameters and the calling is also convenient - AddMember(&TestObject::i);
Here Parent will be deduced to TestObject and Type will be deduced to int.
And Member will point to i.
I think this is as convenient as it can get.
|
|
|
|
|
I'm having trouble either launching the 2 correctly, or getting them to run together, not sure which one.
I want to run a fake progress bar update, and run a program in cmd.exe together at the same time.
This current code runs the CreateProcess, and before I wait for the Process, I create a thread to run the progress program. The process runs, and then the progress program runs when the process is complete. I'm trying to run them simultaneously. I'm not waiting for the progress bar to finish, because it's decorative.
if (CreateProcess( lpFile_Create, szParameters_Create, NULL, NULL, FALSE, CREATE_NO_WINDOW, 0, szWindowsTempDir, &si, &pi) )
{
PIIS7_INSTALL_PROGRESSDATA pDataArray;
pDataArray = (PIIS7_INSTALL_PROGRESSDATA) HeapAlloc(GetProcessHeap(), HEAP_ZERO_MEMORY,
sizeof(IIS7_INSTALL_PROGRESSDATA));
if( pDataArray == NULL )
{
ExitProcess(2);
}
else {
pDataArray->hStatusMessage = pzStatusMessage;
pDataArray->hProgressBar = pzProgressBar;
hProgressBar = CreateThread( NULL, 0, CA_IIS7_ProgressBar_Install, pDataArray, 0, dwProgressBarID );
}
WaitForSingleObject(pi.hProcess, INFINITE);
GetExitCodeProcess(pi.hProcess, (unsigned long *)&exitstatus);
CloseHandle(hProgressBar);
CloseHandle(pi.hProcess);
CloseHandle(pi.hThread);
if(pDataArray != NULL)
{
HeapFree(GetProcessHeap(), 0, pDataArray);
pDataArray = NULL;
}
|
|
|
|
|
Did you state what the problem is? My apologies if I missed it.
The difficult we do right away...
...the impossible takes slightly longer.
|
|
|
|
|
I can get the CreateProcess to go, or the CreateThread, but not both.
|
|
|
|
|
So when the process starts, and it gets to the CreateThread call, did you call GetLastError to see why CreateThread fails?
The difficult we do right away...
...the impossible takes slightly longer.
|
|
|
|
|
Could it be that since you're blocking the thread in the WaitForSingleObject function, that the program is unable to process the window messages necessary to update the progress bar?
And then when the main thread is released, all the messages are processed?
The difficult we do right away...
...the impossible takes slightly longer.
|
|
|
|
|
I think your right. But I'm not sure in how to proceed, I was thinking WaitForMultiple, but I thought it was only for threads with the same handle.
Would it be possible to offer a hint?
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
