C / C++ / MFC

19-Apr-09 1:47

maxMESA wrote:
My question is how can i load this dll in my activex. I am using LoadLibrary(x.dll) and it searches for dll in windows/system32 folder.

Construct the full path to the DLL and pass that to LoadLibrary.

You can get the ActiveX's path using GetModuleFileName[^], either by storing the ActiveX DLL's HINSTANCE in it's DllMain or (and I think this is better, as there's no global varible involved) by using GetModuleHandle[^] or GetModuleHandleEx[^] to get the ActiveX DLL's HMODULE. In fact, if you put the function below in your ActiveX DLL (it has to be in that DLL), you'll get the ActiveX DLL's path - the function asks for the module containing itself.

std::string GetThisModulesFileName()
{
   HMODULE hMod;
   char path[MAX_PATH];
   GetModuleHandleExA(GET_MODULE_HANDLE_EX_FLAG_FROM_ADDRESS, reinterpret_cast<lpcstr>(&GetThisModulesFileName), &hMod);
   GetModuleFileNameA(hMod, path, MAX_PATH);
   return std::string(path);
}</lpcstr>

You might want to add some error checking for production code Sniff | :^)

Once you've got the ActiveX DLL's path, it's simple to extract the directory with PathRemoveFileSpec[^] and then add on the DLL's name with PathCombine[^].

Java, Basic, who cares - it's all a bunch of tree-hugging hippy cr*p

Re: activex dll problem

maxMESA19-Apr-09 7:14

maxMESA

19-Apr-09 7:14

ok i solved. i loaded dll to sys32 folder with an windows inf file as told here

http://msdn.microsoft.com/en-us/library/aa751974(VS.85).aspx[^]

point reference VS value

sharion18-Apr-09 21:18

sharion

18-Apr-09 21:18

We know that when point variable as parameter of function,operations done in the function to the parameter actually willl effect the variable that is passed into the function.That is called pass by reference.

Based on above viewpoint,I have an confusion. For example,

void GetMemory(char* p)
	{
		p =  new char[100];
	}
int main()
{
   char* str = NULL;
   GetMemory(str);
   /* */
}

After GetMemory(str) is complete, str is still NULL and no point variable is related to memory allocated.

About the result, one explanation is p in function GetMemory is only a copy of variable str. This explanation may be right but it seems that it contradicts pass by reference. Now that passing by reference, where is the copy come from. And operations done to variable p is equal to operations to str,why str is NULL finally.

Appreciate your kind very much and hope anyone give me a clear explanation. Thanks in advance.

Stuart Dootson18-Apr-09 22:53

18-Apr-09 22:53

sharion wrote:
About the result, one explanation is p in function GetMemory is only a copy of variable str. This explanation may be right but it seems that it contradicts pass by reference. Now that passing by reference, where is the copy come from. And operations done to variable p is equal to operations to str,why str is NULL finally.

There is no contradiction. When you pass a pointer, the thing that is pointed at is passed by reference. So, in your case, the character(s) pointed at by str are passed by reference. str, however, is passed by value.

To pass str by reference, you need to add another level of indirection, either by passing a pointer to str or passing a reference to str, like this:

Pass a pointer to str:

void GetMemory(char** p)
	{
		*p =  new char[100];
	}
int main()
{
   char* str = NULL;
   GetMemory(&str);
   /* */
}

Pass a reference to str:

void GetMemory(char*& p)
	{
		p =  new char[100];
	}
int main()
{
   char* str = NULL;
   GetMemory(str);
   /* */
}

Java, Basic, who cares - it's all a bunch of tree-hugging hippy cr*p

sharion19-Apr-09 1:35

sharion

19-Apr-09 1:35

Thanks for your respond.

Stuart Dootson wrote:
When you pass a pointer, the thing that is pointed at is passed by reference.

Still feel some confusion.

void GetMemory(char* p)
{
    p =  new char[100];
    *(p + 10) = 'a'; /* following comment is true or not?*/
                     /* operate to the memory block 
                        that you called passed by reference
                     */
}
int main()
{
   char* str = NULL; 
   GetMemory(str);
   /* */}

We believe str is still NULL, where I have mistaken?

Luc Pattyn19-Apr-09 1:49

19-Apr-09 1:49

this is what occurs in sequence:

char* str = NULL;   // str is a char pointer, set to NULL
GetMemory(str);
     void GetMemory(char* p)  // p is a char pointer set to NULL
     p =  new char[100];      // the old p is overwritten by a new pointer, pointing to a 100 byte buffer
     *(p + 10) = 'a';         // the new p is used to write some into that buffer
     return;
... // back from GetMemory    // str has never been changed

and this is how the callee would be able to modify the caller's pointer

void GetMemory(char** p)    // accept pointer to pointer to char
{
    p =  new char[100];
}
int main()
{
   char* str = NULL; 
   GetMemory(&str);   // pass the address of str (=pointer to pointer)
   // str will have changed now

Luc Pattyn [Forum Guidelines] [My Articles]

Avoiding unwanted divs (as in "articles needing approval") with the help of this FireFox add-in

Stuart Dootson19-Apr-09 1:52

19-Apr-09 1:52

sharion wrote:
*(p + 10) = 'a'; /* following comment is true or not?*/
/* operate to the memory block
that you called passed by reference
*/

Not true.

You passed NULL to GetMemory. That NULL was copied into p (that's what pass by value does). You then changed p's value to the result of the memory allocation and altered that memory block.

So, in effect, you passed a reference to the memory starting at NULL, then threw away that reference in GetMemory by doing

p =  new char[100];

sharion wrote:
We believe str is still NULL, where I have mistaken?

No mistake, str will remain NULL with your code. Did you look at the code samples I provided you?

Java, Basic, who cares - it's all a bunch of tree-hugging hippy cr*p

sharion19-Apr-09 3:21

sharion

19-Apr-09 3:21

Thanks for guy's repond.

Of course your solution to keeping str changed is complete right. I feel awkward about your explanation. Based on experience previous, I take it for granted that whatever you have done to p have the same effect to str. That is passing by reference what I previously thought. Now your explanation is opposite to my thought and I find it hard to drive me from the viewpoint out-of-date.That is my puzzle.

Thanks a lot.

Stuart Dootson19-Apr-09 3:52

19-Apr-09 3:52

sharion wrote:
Based on experience previous, I take it for granted that whatever you have done to p have the same effect to str. That is passing by reference what I previously thought

What you are actually doing in your code is passing the pointer str by value. At the same time, you are passing by reference is what str points to.

C always passes parameters by value. So, when you want to pass a parameter by reference, you need to get a pointer to the value and pass the pointer by value to the function. So, this function takes x by value and returns a value calculated from x.

int Double(int x)
{
   return x + x;
}

while this function does the same, but passing x by pointer. To show that the actual value you pass to the function is not the one you're passing by pointer, I've changed the parameter to pX, meaning 'pointer to X'.

int Double(int *pX)
{
   return (*pX) + (*pX);
}

Now, let's move on to some string functions. Let's calculate the length of a C string:

int StringLength(char* s)
{
   int length = 0;
   while (*(s + length) != 0)
      ++length;
}

Now, what is s? It's a string, right? But in C, a string is just a pointer to the first character of the string! So, we're passing a pointer to the sequence of characters that make up the string. And we're passing that pointer by value. What we're passing by pointer is the characters that make up the string.

This page[^] and the ones it links to (especially the one about modifying function arguments[^] might help explain it better than I can - I've never been a good teacher Smile | :)

Java, Basic, who cares - it's all a bunch of tree-hugging hippy cr*p

Luc Pattyn19-Apr-09 4:05

19-Apr-09 4:05

I fully agree, except maybe with "I've never been a good teacher".
"C always passes parameters by value" basically says it all.

Of course the problem is other languages (try and) hide the notion of pointers with keywords such as ByRef (VB) or ref (C#).

Smile | :)

Luc Pattyn [Forum Guidelines] [My Articles]

Avoiding unwanted divs (as in "articles needing approval") with the help of this FireFox add-in

Stuart Dootson19-Apr-09 4:24

19-Apr-09 4:24

Try something like Ada[^] Smile | :)

To paraphrase - we'll pass some things by value, some by reference...and we're not telling which Smile | :)

It does work well, however, as Ada also has the concept of in, out and in out parameters, so you use that rather than having to interpret how to use 'by value' or 'by reference' parameters to get what you want.

Java, Basic, who cares - it's all a bunch of tree-hugging hippy cr*p

Luc Pattyn19-Apr-09 4:37

19-Apr-09 4:37

I've read the Ada specs when they first emerged in the eighties, but never actually used the language.
Is it still in active use?
is there an ADA#?
should we have a forum on it? Big Grin | :-D

Luc Pattyn [Forum Guidelines] [My Articles]

Avoiding unwanted divs (as in "articles needing approval") with the help of this FireFox add-in

Stuart Dootson19-Apr-09 4:56

19-Apr-09 4:56

Luc Pattyn wrote:
Is it still in active use?

Yes, especially in safety-critical applications - the company I work for uses it. It's also becoming more actively used in some security areas, especially when used in conjunction with SPARK[^].

Luc Pattyn wrote:
is there an ADA#?

Yes[^], although it's not really active.

Luc Pattyn wrote:
should we have a forum on it?

Java, Basic, who cares - it's all a bunch of tree-hugging hippy cr*p

Luc Pattyn19-Apr-09 5:00