C / C++ / MFC

Using MSDN from .NET with MSVC6

18-Aug-02 10:19

A usual user account should never have assigned SeDebugPrivilege. This is one of the most powerful privileges and can easily be used to make yourself an Admin!

UGenn, what exactly do you want to do? Do you try to get this informations about a foreign process or about your own process? And what information do you need? "Privilege level" is a bit ambigious, because the access rights of a process are spread over a couple of places. Do you just want to know if the process is running under the LocalSystem account? Or if the user account the process is running under is a member of the Admins group? Anything about assigned and enabled privileges?

I am quite sure I can help you, but I need more input Roll eyes | :rolleyes:

--

Daniel Lohmann

http://www.losoft.de
(Hey, this page is worth looking! You can find some free and handy NT tools there Big Grin | :-D

)

Marcin Jakubowski18-Aug-02 4:52

Marcin Jakubowski

18-Aug-02 4:52

Hi,
Is it possible to use MSDN from Microsoft Visual Studio .NET with Microsoft Visual C++ 6.0? How to install it?

Re: Using MSDN from .NET with MSVC6

Michael P Butler18-Aug-02 8:11

Michael P Butler

18-Aug-02 8:11

No. You need MSDN October 2001 or earlier to work with VC6.

From Jan 2002, MSDN only integrates with VS.NET

Michael Smile | :)

Time flies like an arrow. Fruit flies like a banana

Debug Assertion in Debug DLL (not Release)

User 665818-Aug-02 3:45

User 6658

18-Aug-02 3:45

Hi!

I've got a very strange problem:
When unloading my DLL from my EXE using FreeLibrary(hDLL) I get a debug assertion:

File: dbgheap.c
Line: 1011

Expression: _CrtIsValidHeapPointer(pUserData)

What the hell does that mean? Neither my DLL nor my EXE do allocate memory using new. The only thing my DLL does is getting the contents of a database that is within the DLL as a resource. But the MSDN says that it's not necessary to free the resource on Win32 platforms.

In the Release version of my DLL nothing strange happens, everything works fine.

So why do I get this debug assertion? Confused | :confused:

thanks in advance
Greg

modified 12-Sep-18 21:01pm.

Re: Debug Assertion in Debug DLL (not Release)

User 665818-Aug-02 4:11

User 6658

18-Aug-02 4:11

Ok I found the mistake, but still have one question:

This is the code that allocates some memory:

GeoIP *gi = (GeoIP *)malloc(sizeof(GeoIP));

gi->cache = (unsigned char*)malloc(sizeof(char) * dat_size);
if (gi->cache != NULL) 
    gi->cache = (unsigned char*)rc_dat;
else 
    gi->cache = NULL;

This allocates memory for the contents of the db from the resource.

In the DLL_PROCESS_DETACH I call this code:

//	if (gi->cache != NULL)
//		free(gi->cache);

	if(gi != NULL)
		free(gi);

When I comment out the code above I don't get the assertion, but if not, the debug assertion occurs. It seems like gi->cache points to a non-null memory but free() still can't free the memory Confused | :confused:

Do I still need to free the gi->cache structure?

modified 12-Sep-18 21:01pm.

Re: Debug Assertion in Debug DLL (not Release)

C^hris L^osinger18-Aug-02 17:34

C^hris L^osinger

18-Aug-02 17:34

gi->cache isn't pointing at anything. in the first code sample, you allocated memory, then threw away the pointer when you set gi->cache to NULL or to rc_dat.

yes, you still need to free the memory, but you need to have the address of the memory before you can free it

-c

"To announce that there must be no criticism of the president, or that we are to stand by the president right or wrong, is not only unpatriotic and servile, but is morally treasonable to the American public."
-- Theodore Roosevelt

image effects!

Re: Debug Assertion in Debug DLL (not Release)

Mustafa Demirhan19-Aug-02 13:51

Mustafa Demirhan

19-Aug-02 13:51

I do not understand this part.

GeoIP *gi = (GeoIP *)malloc(sizeof(GeoIP));<br />
gi->cache = (unsigned char*)malloc(sizeof(char) * dat_size);<br />
if (gi->cache != NULL)     <br />
    gi->cache = (unsigned char*)rc_dat;<br />
else     <br />
    gi->cache = NULL;

You allocate some memory and then you loose the pointer because you modify the pointer gi->cache. gi->cache now points to a different location and you cannot delete it if it is not on the heap. Also, you will surely have a memory leak; because you lost the pointer for the newly allocated memory space.

Mustafa Demirhan
http://www.macroangel.com
Sonork ID 100.9935:zoltrix

<nobr>They say I'm lazy but it takes all my time

Problem with monitoring threads !!!

tongc18-Aug-02 3:41

18-Aug-02 3:41

Hi !!

I have 5 threads that takes the milisecond that take 10000, 15000, 20000, 25000, 30000 milisecond as parameter respectively. In the main thread i pass these parameter to the CreateThread function, then display how many threads are still running every second. It seems working seamlessly, but when i really notice, the first thread which I pass 5000 milisecond in only display 9 times(it should be 10 times), whereas the thread with 250000 as parameter is displayed 6 times. Where as each thread is 5 units away each the other.

this is my code

INT volatile runningThread = 5;	//number of running thread.

void Thread(DWORD time)	
{
	
	Sleep(time);
	
}

void main()
{

	INT i;
	HANDLE handle[5];
	DWORD threadID[5];

	DWORD time[5];
	time[0] = 10000;
	time[1] = 15000;
	time[2] = 20000;
	time[3] = 25000;
	time[4] = 30000;

	BOOL flag[5];
	for(i = 0; i < 5; i++)
		flag[i] = TRUE;
	
	for(i = 0; i < 5; i++)
	{
		handle [i] = CreateThread(0, 0, (LPTHREAD_START_ROUTINE)Thread 
									, (VOID *)time[i], NULL, &threadID[i]);
		
		if (handle[i] == 0)
			cout << "Create thread " << i << "failed " << GetLastError() << endl;
		
		cout << threadID[i] << endl;
	}
	
	while(WaitForMultipleObjects(5, handle, TRUE, 1000) == WAIT_TIMEOUT)
	{
		for (i = 0; i < 5 ; i++)
		{
			if(WaitForSingleObject(handle[i], 0) == WAIT_OBJECT_0 && flag[i] == TRUE)
			{
				runningThread--;
				flag[i] = FALSE;
			}
		}
		printf(" %d ", runningThread);
	}
	
	
	for(i = 0; i <5; i++)
	{
		if(!CloseHandle(handle[i]))
			cerr << "Failed to close process. Error number" << GetLastError() << endl;
		
	}

	cout << "\nAll threads have been terminated " << endl;

thanks in advance!!!

ADDED(after modification) : I've noticed one thing, that is if i debug it step by step (F11) then it works ok, but when running Ctrl+F5 it just doesnt seem run right!!!

valikac18-Aug-02 9:41

valikac

18-Aug-02 9:41

You code looks okay. However, I am not sure if your design will work. For example, in the first loop that creates when threads and assign them to the handles, time elapses from the first thread. Again, your second loop also takes time to iterate.

Kuphryn

Daniel Lohmann18-Aug-02 10:34

18-Aug-02 10:34

So you moved from processes to threads now? Big Grin | :-D

I am not sure if I really understood your question (what do you mean with "thread x display only y times?). Therefore it may be that I completly lost your point.

However, thread execution is not a deterministic thing. Depending on other ready-to-schedule thread it may be, that your 5000 msec thread is started after the 25000 msec thread. On a machine with more than one CPU they may be started at really the same time. Besides timing with Sleep, WaitForXXX and so on is not completly precise, standard timer resolution is about 10 msec on NT and about 50 msec on Win9x. This means your 5000 msec thread might sleep just 4990 msec.

Long story, short message: A threads execution time is not reliable and it is just bad design to relay on it.

--

Daniel Lohmann

http://www.losoft.de
(Hey, this page is worth looking! You can find some free and handy NT tools there Big Grin | :-D

)

tongc18-Aug-02 16:41

18-Aug-02 16:41

Hi Daniel! It's good hear from u!

Sorry, i was abit rush last night! But here is the requirment

<quote>
write a Win32 Console application that starts five threads, passing each a parameter which is the number of milliseconds that this thread is to simulate execution. The parameter must have a different value for each thread invocation, and should be in the range 20 to 40 seconds. You may predetermine these values, or determine them randomly. Your main thread should monitor these threads, outputting a digit every second indicating how many threads are still running. Your program should terminate after all five new threads have terminated.

I basically have done all of this but the result is as i explained above. Where as the first thread should run for 5 second it only run in 4, however other threads are running ok.

Do you notice any bugs or logical error??

Thanks in advance !!! Smile | :)

Daniel Lohmann19-Aug-02 1:12

19-Aug-02 1:12

tcuong wrote:
Do you notice any bugs or logical error??

No, the code looks fine. I assume it is a timing problem: As I wrote the timing precision on NT is about 10 msec. If your Sleep( 5000 ) actually returns after 4990 msec and maybe the WaitForMultipleObjects(..., 1000) returns after 1010 msec it may well happen you "loose" one active messsage of the thread.

To verify this you could modify the timings by small amounts. I'd also recommend to add printfs before and after the Sleep() statement and monitor the output sequence.

--

Daniel Lohmann

http://www.losoft.de
(Hey, this page is worth looking! You can find some free and handy NT tools there Big Grin | :-D

)

tongc20-Aug-02 5:19

20-Aug-02 5:19

Is there any other solutions? To conclude this thread, when dealing with thread, there is no way that the time allocated is run precisely when the thread executes , doesnt it???

Thanks !!

Daniel Lohmann20-Aug-02 8:32

20-Aug-02 8:32

Yep.
A thread may get or loose the CPU at any time, therefore precise execution times are not possible in general.

--

Daniel Lohmann

http://www.losoft.de
(Hey, this page is worth looking! You can find some free and handy NT tools there Big Grin | :-D

)

Mustafa Demirhan19-Aug-02 13:56

Mustafa Demirhan

19-Aug-02 13:56

As the others said, exact timings wont work!

Instead of monitoring each second, you can monitor twice each second. Then in the end; if you get an odd number (19 for the first thread for example) add 1 to it and divide it by 2. If you get an even number, then this means that the timings worked perfectly (which won't me the case for most of the time).

Mustafa Demirhan
http://www.macroangel.com
Sonork ID 100.9935:zoltrix

<nobr>They say I'm lazy but it takes all my time

Hot/mouseover

DanYELL18-Aug-02 3:33

DanYELL

18-Aug-02 3:33

Does anyone have any code where a text box
is highlightedin red when the mouse hovers
over it or when it has focus? There is a
demo project on codeguru, but it crashes
easily and that worries me.

Please, any response any one can give me
will be greatly appreciated.

Sincerely,
Danielle (an overworked graduate student)

excepting a flag????

Eric Cashwell18-Aug-02 2:39

Eric Cashwell

18-Aug-02 2:39

Please don't laugh to hard but how do you except a command line parameter such as a flag?
Ex: MyProj.exe /s

Thanks Eric

Re: excepting a flag????

tongc18-Aug-02 3:09

Easy question (hopefully)

18-Aug-02 3:09

In your main program put something like this:

void main(dataType flag)
{
  switch flag:
  case: s
       printf("hello your input param is s"
      break;
   case: 

   .............

   .............

}

whereas dataType is the dataType you want, such as integer or whatever flags you want. In the main program you can have switch case statement according to each input parameter Smile | :)

Hope this helps!!!

Re: excepting a flag????

Eric Cashwell18-Aug-02 3:15

Eric Cashwell

18-Aug-02 3:15

I knew that it was very simple. Thank you.

Eric

Re: excepting a flag????

Eric Cashwell18-Aug-02 4:25

Eric Cashwell

18-Aug-02 4:25

Very sorry for waisting your time Unsure | :~

I figured it out

void main(int argc, char* flag[])
{
// acrgc = number of string param
// flag[1] = first string param
// ...
// flag[n] = n string param
}

Thanks for your time though

Eric

Jh318-Aug-02 2:17

Jh3

18-Aug-02 2:17

how can i update the dialogs of my dialog application in my own class. Currently i've inherited from the *Dlg class, and even though the compiler returns no errors, i get a error for the command UpDateData(TRUE) while the program is running, saying Debug Assertion Failed.

Anyone know what the problem is?

Re: Easy question (hopefully)

Stephane Rodriguez.18-Aug-02 2:26

Stephane Rodriguez.

18-Aug-02 2:26

Probably something wrong in the DDX variable mapping declared in ::DoDataExchange(CDataExchange* pDX).
Be sure to use the Dialog classwizard to add or modify those members. Also check if member types are adequate (MSDN MFC Technical Note can help for that).

Let's swallow a small raisin.

Re: Easy question (hopefully)

tongc18-Aug-02 3:44