automation problem. excel - C / C++ / MFC Discussion Boards

Thanks a lot! Works perfectly Smile | :)

modified 12-Sep-18 21:01pm.

Recently I got a bug with a program I wrote under W98 and ported to XP.
Actually, I just changed the allocation of my data arrays to GlobalAlloc - instead of malloc - and the program ran fine again.
I'm happy because I found the solution but now I wonder about all the other allocations I made using either the new operator or malloc and wonder which of them I should change to GloballAlloc and which of them are good. How many of them Frown | :(

So, is there an article over there explaining the strategy to follow ? What are the exact rules ? Differences between Wx and NT and so on.
Thanks,

Yarp

In all honestly, changing the allocation to GlobalAlloc didn't fix the bug, you just changed the execution enough to where the bug didn't bite you.

IMHO, you have a memory overrun somewhere. 9x and NT/W2K/XP execute programs very differently. Their memory footprint differ. Thus memory corruption problems might only show up in one flavor of the OS.

Change your GlobalAlloc back to malloc and try to find the real problem.

Tim Smith

I know what you're thinking punk, you're thinking did he spell check this document? Well, to tell you the truth I kinda forgot myself in all this excitement. But being this here's CodeProject, the most powerful forums in the world and would blow your head clean off, you've got to ask yourself one question, Do I feel lucky? Well do ya punk?

It's a problem between GetDIBIts and StretchDIBits, the DIB is probably selected in a DC when I try to overwrite the bits and that was the reason why the program fails. Locking the memory solved the problem the other way out is using some CriticalSection somewhere.

Thank for the info, I'm quite surprised but unfortunately I was trying to fix this bug since 1 month and couldn't find the way out.

Yarp

I saw a similar problem a while back on some code written for Win98 and rebuilt on XP.

I agree with Tim Smith. You didn't find a solution to the problem, you just found a way to hide the symptom of a nasty memory bug.

If you don't find it now you'll never be sure when it will come back to haunt you! Unsure | :~

If you continue to do the same things you always did,
don't be surprised if you get the same results you always got.

You didn't idenitify the bug associated with using the malloc function, but the _malloc_dbg might be useful in resolving the bug to a specific malloc call.

If the bug occurs when you read or write to the memory, you may be trying to read or write to unallocated memory. I know that GlobalAlloc has a 16 byte granularity. In other words if you allocate 1 byte of memory using GlobalAlloc, Win32 actually allocates 16 bytes. If you try to read or write to the 2nd or 3rd byte, etc then you probably won't get into trouble. I am not sure if this is true for malloc. Maybe something to look into.

I may be way off base, if so, then I am sure the CP gurus will correct me

It's a problem between GetDIBIts and StretchDIBits, the DIB is probably selected in a DC when I try to overwrite the bits and that was the reason why the program fails. Locking the memory solved the problem the other way out is using some CriticalSection somewhere.
Thank for the info, I will try _malloc_dbg.

Yarp

I had created "excel", that fromat, and print excel file. it is based on the article,
http://support.microsoft.com/directory/article.asp?ID=kb;en-us;Q178749 .
I followed every steps what it said on the article. and the execl program is working perfectly.

i also have another application that does printing job. it print any kind of files using shellexecuteex. anyway, i
1) opened classWizard
2) clicked on Automation tab
3) clicked on add class-> from a library
4) find execl9.olb , which is in c:\program\microsoft office\office
5) selected the classes that i need from excel9 lib.
6) added, the following codes in CMyprintApp::IniInstance() function, which loads and enables the COM services library;

<br />
     if(!AfxOleInit())<br />
  {<br />
    AfxMessageBox("Could not initialize COM dll");<br />
    return FALSE;<br />
  }<br />
<br />
  AfxEnableControlContainer();<br />
      .......<br />
<br />

7) in printExcel function, which is for MyPrint program, and onRun button function, which is for the execl program. i added the following code
[code]
_Application app; // app is an application object.
_Workbook book; // workbook object
_Worksheet sheet; // worksheet object
Workbooks books;
Worksheets sheets;

Range range; // used for Microsoft excel 97 component
LPDISPATCH lpDisp; //often reused variable.

//common OLE variants. Easy variants to use for calling arguments.
COleVariant covTrue((short)TRUE),covFalse((short)FALSE),covOptional((long)DISP_E_PARAMNOTFOUND, VT_ERROR);

//start microsoft excel, get application object, and attach to app object,
if(!app.CreateDispatch("Excel.Application"))
{
AfxMessageBox("Couldn't CreateDispatch() for Excel");
return;
}
[/code]

those steps are the steps that i did for the excel program. but the function is working fine for excel program but not for Myprint program.
I got the message "Couldn't CreateDispatch() for Excel " for myPrint program.

can anyone tell me why??

Ok, I'm developing a graphics application (think along the lines of InDesign) and have hit a problem with saving Grayscale TIFF images. When I import a grayscale image, it comes in with the colorspace flag set to ImageFlagsColorSpaceGRAY. When I use the function Save on this image, the image output is an ImagesFlagsColorSpaceRGB. The encoder used is image/tiff and it doesn't matter if I save it to a file or stream. Do I need to include any encoder params with this call (I currently pass NULL) or is this another problem with GDI+.

Thanks

Andy Brace
SW Development Engineer II
Qwest
A clear conscience is usually the sign of a bad memory.

Hey Folks,

Does anybody know of a way of determining which DLLs an EXE depends on pragmatically? Tools like Dependancy Walker work great, but I'm in the process of writing an About->File Info dialog, which lists all the DLLs/Versions. I could easily hardcode these values, but would much prefer to do it pragmatically, rather than updating the static list when the list changes.

Was leafing through a few development magazines recently, saw various references to CP, good to see Smile | :)

Thanks.
Mike.
doner@obtain.com

If you want functionality like Dependency Walker, you would have to read the PE header of the executable and locate the imports table. It lists the various modules the EXE depends on along with the functions from the modules that are needed. I can't say specifically where you can get that info but I've come across it on several occassions in the past.

There was an overview of the PE executable format in a recent MSDN magazine.
http://msdn.microsoft.com/msdnmag/issues/02/02/PE/PE.asp

Another approach is to examine the moudules that are currently in use. You can use the toolhelp32 functions for that. Basically you would create a snapshot with CreateToolhelp32Snapshot to get a snapshot of the modules in use by the process. Then call Module32First and successviely call Module32Next to get info about each module used by the process. Don't forget to call CloseHandle on the snapshot's handle when done.

Thanks for the info, great article to read... Managed to get it working!

Mike.

I've also been working in this area, and I found that the techniques used with the demo articals from MSDN (like PEDump, YAHU, etc.) don't cover calls made to a library that was loaded using LoadLibrary. Have you managed to get the export information from LoadLibrary calls? And if so could you give me a brief pointer as to how? (If it's even possible, I dunno...)

Dylan Kenneally
London, UK

Dylan,

I dont know of any way to determine the dependancies, as far as LoadLibrary goes. I opened up one of my EXE's who does a LoadLibrary() in a hex editor, and by the look of things, the DLL name was in the text area of the exe.

If you stumble across how to do this, let me know Smile | :)

Mike.
Ontario Canada.

Hi!

I have a win2000 service written in visual c++ 6.0 that has to access files from a shared folder from the local network neighborhood.

The shared folders are password protected, so I have to create a connection to that folder. I tried to use WNetAddConnection2.

But WNetAddConnection2 fails and returns the error code no. 1312.

I looked it up in the winapi headers, error 1312 is called ERROR_NO_SUCH_LOGON_SESSION, the explanation given is :

'A specified logon session does not exist. It may already have
been terminated.'.

The list of files/locations is taken from a database, so is dynamically changing over time.

Has anybody any idea how could I access files shared on other computers without logging in to the system?

Thanks, Alma

The solution i found is to use the CreateService function with user and password parameters set.
Ex:
--------------------------------------------------------------------------------

CString user = _T(".\\User");
CString pswd = _T("pswd");
LPSTR u = user.GetBuffer(user.GetLength()+1);
LPSTR p = pswd.GetBuffer(pswd.GetLength()+1);

SC_HANDLE hService = ::CreateService(
hSCM, m_szServiceName, m_szServiceName,
SERVICE_ALL_ACCESS, SERVICE_WIN32_OWN_PROCESS,
SERVICE_DEMAND_START, SERVICE_ERROR_NORMAL,
szFilePath, NULL, NULL, _T("RPCSS\0"), u, p);

--------------------------------------------------------------------------------

To create the connection:
--------------------------------------------------------------------------------

NETRESOURCE nr;

char* a = new char[MAX_PATH];
memset(a, 0, MAX_PATH);

char* b = new char[MAX_PATH];
memset(b, 0, MAX_PATH);

strcpy(a, (char*)"\\\\alma\\geza");
strcpy(b,(char*)"x:");

char pswd[20] = "pswd";
char account[20] = "user";

nr.lpRemoteName = a;
nr.lpLocalName = b;
nr.lpProvider = NULL;
nr.dwType = RESOURCETYPE_DISK;

DWORD flags = CONNECT_UPDATE_PROFILE;
DWORD res;

res = WNetAddConnection2(&nr, (char*)pswd, (char*)account, flags);

delete []a;
delete []b;

if (res == NO_ERROR)
return TRUE;
else
return FALSE;

I have the following:

typedef std::list<CMyObject*> myObjectList

The list is populated and is already in alphabetic order by name.

Each item can be either selected or unselected. I want to bubble the selected items to the top of the list without affecting their relative positions to one another. In other words, I may have the following (Selected items are indicated with the "[]" bracket pair):

A
[B]
C
D
[E]
[F]

I want to end up with this:

[B]
[E]
[F]
A
C
D

What's the best way to do this? Should I just remove each item one at a time (starting at the end of the list) and re-insert it at the top of the list? Is there some exotic list function that I can use?

------- signature starts

"...the staggering layers of obscenity in your statement make it a work of art on so many levels." - Jason Jystad, 10/26/2001

Please review the Legal Disclaimer in my bio.

------- signature ends

if you want exotic, you could use std::stable_sort with a custom predicate which says that A, C and D are equivalent and that B,E and F are equivalent (but less than A,C,D).

stable_sort preserves the (relative) order of equivalent items.

-c

Being just contaminates the void.
--Robyn Hitchcock

Image Processing - easy, like Sunday morning.

I came up with this, but I have't tried compiling it yet. What do you think?

CMyList::iterator iCurr;
CMyList::iterator iNext;

bool bMovedItem = true;

while (bMovedItem)
{
   iCurr      = myEntryList.begin();
   iNext      = iCurr + 1;
   bMovedItem = false;
   while (iCurr < myEntryList.end())
   {
      bool currSelected = (*iCurr)->mySelected;
      bool nextSelected = (*iNext)->mySelected

      if (!currSelected && currSelected != nextSelected)
      {
         myEntryList.reverse(iCurr, iNext);
         bMovedItem = true;
      }
      iCurr++;
   }
}

I'm certainly no STL expert (more accurately, I'm a rank beginner with a complete disdain for STL - grin).

------- signature starts

"...the staggering layers of obscenity in your statement make it a work of art on so many levels." - Jason Jystad, 10/26/2001

Please review the Legal Disclaimer in my bio.

------- signature ends

it might work Smile | :)

here's my top-of-the-head STL version:

bool operator < (const CListItem &a, const CListItem &b)
{
  // a is less than b only if a is selected and b isnt
  return (a.IsSelected && !b.IsSelected);
}

...

std::stable_sort(myList);

-c

Being just contaminates the void.
--Robyn Hitchcock

Image Processing - easy, like Sunday morning.

Personally?

I would dump the list to a text file, exec the COMMAND.COM sort program and then read back in the text file. Smile | :)

Tim Smith

I know what you're thinking punk, you're thinking did he spell check this document? Well, to tell you the truth I kinda forgot myself in all this excitement. But being this here's CodeProject, the most powerful forums in the world and would blow your head clean off, you've got to ask yourself one question, Do I feel lucky? Well do ya punk?

now that's exotic!

or... fire up an instance of Access, shoot the data into a special table, read it back with SQL, then pick the data out of the recordsets!

-c

Being just contaminates the void.
--Robyn Hitchcock

Image Processing - easy, like Sunday morning.

I would try that but we're on CE... Smile | :)

------- signature starts

"...the staggering layers of obscenity in your statement make it a work of art on so many levels." - Jason Jystad, 10/26/2001

Please review the Legal Disclaimer in my bio.

------- signature ends

One solution would be to use stable_partition. This will reorder the list based on a predicate which returns true or false, without changing the relative order of any of the items in the same grouping.

The following example shows a simple implementation:

#include <iostream>
#include <list>
#include <algorithm>
#include <functional>

void PrintChar( char ch )
{
std::cout << "'" << ch << "'" << std::endl;
}

struct Partition : public std::unary_function< char, bool >
{
bool operator()( char ch )
{
return ( ch == 'B' || ch == 'E' || ch == 'F' );
}
};

int main(int argc, char* argv[])
{
std::list<char> myList;

myList.push_back( 'A' );
myList.push_back( 'B' );
myList.push_back( 'C' );
myList.push_back( 'D' );
myList.push_back( 'E' );
myList.push_back( 'F' );

std::cout << "before:" << std::endl;
std::for_each( myList.begin(), myList.end(), PrintChar );

std::stable_partition( myList.begin(), myList.end(), Partition() );

std::cout << "after:" << std::endl;
std::for_each( myList.begin(), myList.end(), PrintChar );

return 0;
}

Best regards,
John

I have been taked with creating a TCP/IP file server app that can impliment drag and drop functionality to explorer.

My problm is that the app lists only the file names and attributes, that it is passed without a refrence to the original location. I need to be able to get the target of a drag and drop opperation onto the explorer window, so I can pass that target back to the calling app.

I've spent countless hours trying to figure out how to get the drop target on the drag opperation to explorer without any success.

Any help would be greatly appreciated.

Life is like a warm bath, the longer you're in the more wrinkled you get.