ATL / WTL / STL

how do i copy a CImage DIB to clipboard

30-Oct-12 6:04

You have not defined the types in your map template; the above code does not even compile.

One of these days I'm going to think of a really clever signature.

Member 951640015-Oct-12 11:15

Member 9516400

15-Oct-12 11:15

im trying to copy a cimage dib to the clipboard. the second
memcpy fails with a read access violation. can anyone help?

C++

CImage tmpImage = pDoc->m_imageArray[0];
     
    int w = tmpImage.GetWidth();
    int h = tmpImage.GetHeight();
    int Bpp = tmpImage.GetBPP();
     
    BITMAPINFOHEADER bmInfohdr;
    bmInfohdr.biSize = sizeof(BITMAPINFOHEADER);
    bmInfohdr.biWidth = w;
    bmInfohdr.biHeight = -h;
    bmInfohdr.biPlanes = 1;
    bmInfohdr.biBitCount = Bpp;
    bmInfohdr.biCompression = BI_RGB;
    bmInfohdr.biSizeImage = w*h*Bpp;
    bmInfohdr.biXPelsPerMeter = 0;
    bmInfohdr.biYPelsPerMeter = 0;
    bmInfohdr.biClrUsed = 0;
    bmInfohdr.biClrImportant = 0;
     
    BITMAPINFO bmInfo;
    bmInfo.bmiHeader = bmInfohdr;
    bmInfo.bmiColors[0].rgbBlue=255;
     
     
    void* pBits = tmpImage.GetBits();
    HANDLE hData = ::GlobalAlloc (GMEM_MOVEABLE, sizeof(BITMAPINFO) + w * h * 3);
    LPVOID pData = (LPVOID) ::GlobalLock (hData);
    LPBYTE p_imagebits;
    p_imagebits  = (LPBYTE)pData + sizeof(BITMAPINFO);
    
 
    memcpy(pData,&bmInfo,sizeof(BITMAPINFO));
     
           
    DWORD dwBytes = ((DWORD) w * Bpp) / 32;
    
    if(((DWORD) w * Bpp) % 32) {
       dwBytes++;
    }
    dwBytes *= 4;
   
    unsigned long m_dwSizeImage = dwBytes * h; // no compression
     
     
    memcpy (p_imagebits, pBits, m_dwSizeImage);
    
 
    ::GlobalUnlock (hData);
     
    COleDataSource* pods = new COleDataSource;
    pods->CacheGlobalData (CF_DIB, hData);
    pods->SetClipboard ();

Re: how do i copy a CImage DIB to clipboard

chaau15-Oct-12 11:41

chaau

15-Oct-12 11:41

It fails because the memory allocated by GlobalAlloc is not enough. See, how you use sizeof(BITMAPINFO) + w * h * 3 value to allocate memory, and then you are trying to calculate the m_dwSizeImage value, which is much greater then requested.

You have got two options:
Either move this chunk of code after you have calculated the image size, so that your code becomes:

C++

DWORD dwBytes = ((DWORD) w * Bpp) / 32;

if(((DWORD) w * Bpp) % 32) {
   dwBytes++;
}
dwBytes *= 4;

unsigned long m_dwSizeImage = dwBytes * h; // no compression

void* pBits = tmpImage.GetBits();
HANDLE hData = ::GlobalAlloc (GMEM_MOVEABLE, sizeof(BITMAPINFO) + m_dwSizeImage);
LPVOID pData = (LPVOID) ::GlobalLock (hData);
LPBYTE p_imagebits;
p_imagebits  = (LPBYTE)pData + sizeof(BITMAPINFO);


memcpy(pData,&bmInfo,sizeof(BITMAPINFO));
memcpy (p_imagebits, pBits, m_dwSizeImage);

Or, just use a solution[^] from Codeguru

BTW, it looks like you have calculated the m_dwImageSize incorrectly anyway

Why does this fail - CFile / SeekToEnd?

charlieg10-Oct-12 6:43

10-Oct-12 6:43

Code snippet below. The SeekToBegin fails because the CFile object has an invalid file handle. But the open returns true, so the code proceeds to die an ugly death. The code is called within a tight loop passing through a couple of hundred files. It feels like I'm overrunning the hard drive or file system, but I have no idea how I might do that... any thoughts are greatly appreciated.

C++

// in a loop of many files....
// do some compression.
// now open a temp file to dump raw data into:

FilePath = (CString)directoryBuf;
FilePath += (CString)"\\uncompressed.unc";
CFileException fe;

if(UncompFile.Open(FilePath, CFile::shareExclusive | CFile::modeCreate | CFile::modeReadWrite), &fe)
{
  UncompFile.SeekToBegin();
  UncompFile.Write(outBuf, c_stream.total_out);
  UncompFile.SeekToBegin();
  UncompFile.Write(outBuf, c_stream.total_out);
  UncompFile.Close();

  cmpCRC = helper.ComputeFileCRC( FilePath );

  UncompFile.Remove((LPCTSTR)FilePath);

  if(uncCRC != cmpCRC)
  {
    ErrorString = (CString)infoStruct->sourceFileName  + "\n failed CRC test.";
    return ErrorString;
  }

Charlie Gilley
<italic>You're going to tell me what I want to know, or I'm going to beat you to death in your own house.

"Where liberty dwells, there is my country." B. Franklin, 1783
“They who can give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety.” BF, 1759

modified 10-Oct-12 13:11pm.

charlieg10-Oct-12 9:31

10-Oct-12 9:31

Okay, I made one change. Just prior to the .Close, I now invoke .Flush().

The main difference between my environment and the old environment is that my laptop is much faster with a solid state disk. Is it possible that there is a race condition inside of CFile? Seems like I'm grasping at straws.

chaau10-Oct-12 11:23

chaau

10-Oct-12 11:23

Why is this code duplicated?

C++

UncompFile.SeekToBegin();
UncompFile.Write(outBuf, c_stream.total_out);

FYI, Remove() is a static method. It will be correct to call it like this:

C++

CFile::Remove((LPCTSTR)FilePath)

The method Remove() can throw an exception
Example here[^]

charlieg11-Oct-12 2:20

11-Oct-12 2:20

Duplicated code: I don't know. I didn't write this code Smile | :)

. Not ducking the question, and it did occur to me that if I was creating a file, the initial position would be zero. In testing, I commented out that line only to have the same exception thrown on the write method - bad file handle.

Static Remove and exceptions: noted.

Richard MacCutchan10-Oct-12 21:45

10-Oct-12 21:45

C++

FilePath = (CString)directoryBuf;
FilePath += (CString)"\\uncompressed.unc";

What is the purpose of the casts in the above statements?

In the rest of your code you are not checking the contents of fe after the open so you cannot be sure you have a valid handle at that point. It is always advisable to check the return status of every function call dealing with files, rather than assuming they all succeed.

One of these days I'm going to think of a really clever signature.

charlieg11-Oct-12 2:30

11-Oct-12 2:30

Inherited code, probably done for no particular reason. They're gone now.
Function returns: yup, not going to argue the point.

I'm going to condense this code down to a very tight little application and hammer a hard drive. Adding the flush seemed to stabilize things, but I'm suspicious of memory corruption.

Stay tuned for the next episode of "As the Code turns..." will Charlie find happiness?

charlieg11-Oct-12 4:53

11-Oct-12 4:53

Richard - the fe content only becomes important if the open call fails. Note the weirdness, it's returning true yet with an invalid handle. In the debugger, I can see that the content of the file exception structure is "no error".

Richard MacCutchan11-Oct-12 11:19

11-Oct-12 11:19

charlieg wrote:
it's returning true yet with an invalid handle.

I find that very hard to believe; I think you are perhaps misreading something.

One of these days I'm going to think of a really clever signature.

charlieg11-Oct-12 11:34

11-Oct-12 11:34

if(UncompFile.Open(FilePath, ...)
{
UncompFile.SeekToBegin(); // <--- asserts here. bad file handle...

I don't know how I could misread that. I'm all ears Smile | :)

or eyes....

Richard MacCutchan11-Oct-12 12:01

11-Oct-12 12:01

charlieg wrote:
I don't know how I could misread that.

I suspect there is a lot more going on in that code than meets the eye. You cannot open a file successfully and then immediately find that the handle is bad; it implies a serious bug in the Win32 SDK which would affect just about every system on the planet. The reality is that there is a bug in your code that has, as yet, gone undetected, but without a lot more information we cannot begin to guess where it is.

One of these days I'm going to think of a really clever signature.

charlieg12-Oct-12 5:54

12-Oct-12 5:54

I completely agree. Reaching for a bug in Win32 is silly....

and apologies to all for putting this question in the wrong friggin forum.

Stephen Hewitt11-Oct-12 12:58

Stephen Hewitt

11-Oct-12 12:58

Lose all the casting. The cast to CString is implicit; it's not needed and is ugly. The cast from CString to LPCTSTR is in the same boat. As to your problem, I didn't look, I was distracted by all the unsavoury casting.

Steve

charlieg18-Jan-13 1:47

Keep program open and running

18-Jan-13 1:47

This is just too funny. sigh, in a sad sort of funny....

So, I have managed to muddle along as there simply is no time to re-write this app. Yet. I can get it to run in the debugger under Windows 7, so that's what I've been doing.

Fast forward three months.

New laptop, new Windows 7 install (should be basically the same, but who really knows?), new VM, etc.

Guess what's running now? Mad | :mad:

Thumps along just fine. I know the bug is still there, I'm not that dellusional.

Lucidation8-Oct-12 11:23

Lucidation

8-Oct-12 11:23

First off, I'm still new to coding in C++ and so I hope this is the right forum for my question...

I'm coding a very small utility for a niche need we've got where I work. Right now I've got things coded so that it uses CreateProcess to fire off another program we run. What I need to do is then have this utility stay open and running (the program that it kicks off can terminate the original program once it's done doing its thing - and I can't just wait for a return code because the program that gets kicked off may kick off others, and one of those might be the final item to run).

Right now it launches our program and then, since it's done doing all the code, it finishes up and quits. Is there any way to tell the thing to stay open and running? Kind of like a service (but not a service because I don't want to install anything).

Richard MacCutchan8-Oct-12 22:14

8-Oct-12 22:14

Just add a loop with a timer, and whatever other functions you need. The sequence would then be something like:

while TRUE
    if applicationFlag is TRUE
    then
        start external application
        set applicationFlag FALSE
    end if
    sleep for some seconds
end while

You will need to add some other decision code to set the applicationFlag to TRUE at some point.

One of these days I'm going to think of a really clever signature.

Lucidation9-Oct-12 11:14

Lucidation

9-Oct-12 11:14

What if the program remains open for an hour or three (I hope they close it sooner, but I have to think of the "what ifs")? Is the Sleep command going to be the right way to go, or am I going to create a problem when using it over an extended period of time?

Richard MacCutchan9-Oct-12 11:19

9-Oct-12 11:19

The other alternative is to use a timer: Timers Tutorial[^] is a great article which shows how to implement a few.

One of these days I'm going to think of a really clever signature.