|
ivanris wrote: what are the arguments for this function, if any?
That all depends on what CWnd object it is created in the context of. For example:
BOOL CComboBox::Create( DWORD dwStyle, const RECT& rect, CWnd* pParentWnd, UINT nID );
BOOL CButton::Create( LPCTSTR lpszCaption, DWORD dwStyle, const RECT& rect, CWnd* pParentWnd, UINT nID );
BOOL CSplitterWnd::Create( CWnd* pParentWnd, int nMaxRows,
int nMaxCols, SIZE sizeMin,
CCreateContext* pContext, DWORD dwStyle = WS_CHILD | WS_VISIBLE |WS_HSCROLL | WS_VSCROLL | SPLS_DYNAMIC_SPLIT,
UINT nID = AFX_IDW_PANE_FIRST );
"Money talks. When my money starts to talk, I get a bill to shut it up." - Frank
"Judge not by the eye but by the heart." - Native American Proverb
|
|
|
|
|
Pattern brush on Win98 can be created from a bitmap larger than 8x8 pixels.
The brush can be used by simple draw function, such as FillRect().
But, Region functions on Win98 use only first 8x8 pixels of the brush to paint, such as FillRgn(), other part of the brush is ignored.
There is no problem on XP, which uses whole brush to paint.
Many my customers still use Win98, so I want to create a FillRgn() function for Win98, which can use whole brush to paint any region as XP does.
Do you have any idea to help me to create the function?
|
|
|
|
|
the standard brute-force flood-fill algorithm is:
floodfill(x, y)
{
if (pixel(x,y)==searchColor)
{
paint(x,y,fillColor)
floodfill(x+1, y+1)
floodfill(x+1, y-1)
floodfill(x-1, y+1)
floodfill(x-1, y-1)
}
}
(even better, use a stack to eliminate the recursion - just push the 4 points onto the stack instead of doing the recursion, then loop and pop)
a patterned fill would replace paint(...) with a function to copy a pixel from the brush to the output pixelout = pixelBrush(x % brushW, y % brushH).
the fact that your brush could contain 'targetColor' complicates things, however - the standard algorithm would loop forever if you painted the output pixel the same color as the search color. one way to handle this would be to make a copy of your source image and do the standard floodfill on it, in memory, while also doing the patterned fill on the original image:
if (pixel(x,y)==searchColor)
{
paint(x,y,fillColor)
patternpaint_output_image(x,y)
...
}
if memory is an issue, you could use a monochrome/grayscale 'mask' image for the copy - set all searchColor pixels set to 0 and all others to 1. then do the standard floodfill on that image, while also putting brush pixels on your output image.
and, for speed, you'll need to do this on the actual pixels (with a DIB), and not using DC functions.
|
|
|
|
|
Hi Guys
Can anyone explain when it will quit from following loop
while(*d++=*s++); please explain whats it doing
|
|
|
|
|
The ++ is an increment, and is being used as a post-increment.
*d++ = *s++
is semantically equivalent to
*d = *s;
d = d + 1;
s = s + 1;
It is performing a copy.
It will stop copying when it has copied a zero value for the basic type, since in C/C++ you can use most expressions as values. If what s and d point to is a non-simple type (such as a class), then it may also invoke a conversion cast operator.
The basics of this is explained in the C Programming reference by Kernighan & Ritchie, and in many C++ beginner's books.
Steve S
Developer for hire
|
|
|
|
|
NitinPatil wrote: please explain whats it doing
It is creating a potential buffer overflow vulnerability. 
led mike
|
|
|
|
|
From the looks of it, it is doing a C-style string copy operation. I imagine your code looks something like the following:
void f(char* const s, char* d)
{
while (*d++ = *s++);
}
which is pretty much the same as writing strcpy(d, s);, and suffers from the same memory issues (say, if s does not have a terminating null character, or d is smaller than s).
If you decide to become a software engineer, you are signing up to have a 1/2" piece of silicon tell you exactly how stupid you really are for 8 hours a day, 5 days a week
Zac
|
|
|
|
|
To put it simply, the loop will end when the expression "(*d++=*s++)" becomes zero. It will be zero when *d++ becomes zero. The *d++ will become zero when a zero is written to its address (which means that a zero will be copied from the *s++ source).
This code copies data from the address pointed to by "s" into the address pointed to by "d". After using both addresses "s" an "d" they will be incremented, thus pointing to the next element to copy.
The loop will stop after copying a null byte. So, most probably, it is an algorithm to copy null-terminated strings.
|
|
|
|
|
how i can send text from my program to testbox in external program
|
|
|
|
|
Have you looked at SetWindowText()?
"Money talks. When my money starts to talk, I get a bill to shut it up." - Frank
"Judge not by the eye but by the heart." - Native American Proverb
|
|
|
|
|
Do you mean like keystrokes ( keyboard emulator ) ?
Do you want to send any data ( text or whatever kind of data ) to another application / program that running ?
AFAIK, there is a library on .NET to do it, called SendKeys.
But if you do not use any .NET then you may have to search on WINAPI documentation.
|
|
|
|
|
Why don't you store your text into an ASCII-file and read this file in the external program?
Good luck, Gerard
|
|
|
|
|
He can send his text direct to a program
|
|
|
|
|
|
yes it solved but some program didnt recive the text so i ahve its handle
|
|
|
|
|
As a security measure, we're trying to stop an application of our own creation having access to explicitly linked DLLs unless those DLLs are "licensed" for the particular installation of the application on a single PC. Each DLL typically exports the same basic functionality.
Thought occurs that a sneaky user could, knowing the name of a licensed DLL, rename another DLL to that name and get around the licence issue.
From looking at the binary of a DLL, the compiled name of the DLL is embedded therein, just before the list of exported functions. Is there a method for accessing this name programmatically, to check that it matches the actual physical filename (otherwise, the application will refuse to load the DLL)?
|
|
|
|
|
I have a problem related to a window that in some cases owns a context/popup menu.
Here is the code for the creation and tracking of the popup menu:
LRESULT
MyWindowClass::OnCreateAndTrackPopupMenu(UINT uMsg, WPARAM wParam, LPARAM lParam, BOOL& bHandled)
{
HMENU aMenu = ::LoadMenu(_Module.GetResourceInstance(),(LPCTSTR)IDR_CONTEXT_MENU);
if(aMenu != NULL)
{
HMENU hSubMenu = ::GetSubMenu(aMenu,0);
if(hSubMenu != NULL)
{
SetForegroundWindow(m_hWnd);
TrackPopupMenuEx(hSubMenu,TPM_LEFTALIGN|TPM_TOPALIGN,pt.x,pt.y,m_hWnd,NULL);
::PostMessage(m_hWnd, WM_NULL, 0, 0);
}
::DestroyMenu(aMenu);
}
return 0;
}
My application response to a window message that should deal with the destruction of the window above:
LRESULT
MyMainClass::OnDestroyWindow(UINT uMsg, WPARAM wParam, LPARAM lParam, BOOL& bHandled)
{
::PostMessage(MyWindowClass->m_hWnd,WM_CANCELMODE,0,0);
MyWindowClass->DestroyWindow();
delete MyWindowClass;
MyWindowClass= NULL;
return 0;
}
The problem: If the popup menu is present when the destruction (DestroyWindow()) takes place, the application will crash ("ERROR - Object deleted before window was destroyed"), because the popup menu still is active (still waiting for TrackPopupMenuEx() to return because WM_CANCELMODE never reach the window before it is destroyed).
So the question is, how can I make sure that the OnCreateAndTrackPopupMenu-function return before DestroyWindow() is executed?
john
|
|
|
|
|
you can do something like :
You call DestroyWindow ( to delete the Window object ), this will eventually call PostNcDestroy, in which you will do a delete this, to delete the C++ object,and before that, send a message ( SendMessage ) to the caller of the window to set its pointer to NULL.
For example :
void MyAutoDeleteDialog::PostNcDestroy()
{
BaseDialog::PostNcDestroy();
AfxGetMainWnd()->SendMessage( WM_AUTO_DELETE_DIALOG_CLOSED, 0, 0 );
delete this;
}
|
|
|
|
|
Thanx for the quick response. I understand your solution and what it does, but it doesn't seem like it solves my problem. I need to cancel the dialog before the object is deleted by using WM_CANCELMODE. How does this fit into the solution?
john
|
|
|
|
|
|
This doesn't do it for some reason. Maybe WM_CANCELMODE IS received synchronically, but TrackPopupMenu() never returns anyhow, before it is too late and the object is deleted - crash  .
john
|
|
|
|
|
Hi Everyone, i am still having troule with this over a wan... can anyone shed any light on it for me ?? i am getting kinda desperate now !!
CFtpConnection Thread......
Thanks in advance
Lee
|
|
|
|
|
try to find sample programs in VC disks (MSDN).
FTP is very easy.
|
|
|
|
|
i have done... and the way i have written it works on the local network, but i have a router set up with 3-5 pcs on it and am using port forwarding, if i leave the ftp port at 21 it works fine, if i change the port to anything else i can connect and create folders etc, but cannot do file searches or transfer files... i need to have it on different port as i have upto 5 pcs to forward to...
please help....
|
|
|
|
|
Open your MSDN -> choose Index tab -> Type: FTP Sessions -> Enter
Sincerely
Thangnvhl
|
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
