Help please with pointers to callbacks in structs

[Brian]

I am very new to C# programming and have run into a problem.

I apologize for the repost of this article. For some reason, my news
reader attached it to an existing thread.

First off, I have an SDK that I have written for C/C++ and would like to
port it to C# if at all possible.

Basically, I've got a structure that I need to pass to the C-style DLL
(i.e. no mangled names - everything is __stdcall). This structure needs
to contain pointers to callback functions that the DLL can use to notify
the C# application when certain events occur. In addition, I need to use
SetEnvironmentV ariable and pass a string containing a pointer to a
character array for special processing by the DLL. I've included some
code that shows basically everything I'm looking to do.

To save time, I understand how to get access to exported functions within
DLLs.
// this structure will be filled by the DLL and will be passed
// to the callback for function2, (i.e. FnPtr2's typedef)

typedef struct _toPassAsArgume ntStruct
{
int a;
int b;
double c;
} TO_PASS_AS_ARGU MENT;
// these are the typedef's for two callback functions that will be
// stored in the structure, DEMOSTRUCT

typedef BOOL (__stdcall *FnPtr1)(DWORD i, char *path, DWORD pathLength);
typedef BOOL (__stdcall *FnPtr2)(TO_PAS S_AS_ARGUMENT *structure);
typedef struct _demoStruct
{
FnPtr1 function1;
FnPtr2 function2;
} DEMOSTRUCT;
char global_text[5] = {"test"};
// MyFunction1 and MyFunction2 are called by the DLL

BOOL MyFunction1(DWO RD i, char *path, DWORD pathLength)
{
strcpy(path, "unknown");
return TRUE;
}

BOOL MyFunction2(TO_ PASS_AS_ARGUMEN T *structure)
{
structure->a = 5;
return TRUE;
}
// simple_function is called by the non-DLL (i.e. the application)

void simple_function ()
{
char text[128] = {0};

DEMOSTRUCT ds = {0};

// Are these four statements possible in C#?

ds.function1 = MyFunction1;
ds.function2 = MyFunction2;

// the contents of 'text' will be passed to the DLL.
sprintf(text, "%x", &global_text );

SetEnvironmentV ariable("test", text);
}
--

TIA,
Brian

[Brian]

> You could use sprintf (c/c++ runtime) but there is a difference in calling

convention, sprintf uses __cdecl. You can call it from c# but you can't
have variable arguments therefore all arguments must be declared eg :
Oh, yeah.... I forgot about that little gotcha.
[DllImport("msvc rt.dll", CharSet=CharSet .Ansi,
CallingConventi on=CallingConve ntion.Cdecl)]
public static extern int sprintf(IntPtr pDest, String format, int i, String
s); And I don't think you need it, you can use .Net formatting capabilities :
string s = Format( "{0} {1}", 10, "test" ); // eg
lstrcpyA( pData, s );
That's what I'll recommend should anyone ask about it. Obviously, sprintf
would be a pain to use.

Out of morbid curiosity, any particular reason why C# is so 'protective' It's not just c#, all net languages compile to ilasm where most of the
restrictions come from.
The idea is to produce more secure and stable applications. of its data? In the land of C (sort of a contradiction there,
phonetically speaking), it's possible to do just about anything to
anybody. Altough I like c it leads to hard to find bugs.
But entomology is fun! :)

C# seems to want keep a tight reign of its data. I'm assuming
its for security reasons, but it sure makes it rough on the 'die-hard' C
programmers.

You should just forget about pointers when you use c# and learn the class
library which has a lot of functionality.

I shall forget. Now what were we talking about??? :) Seriously though,
I've got a couple of books on C#, so hopefully those plus the very helpful
suggestions I've gotten from you will be enough to see me through.

Thanks again,
Brian

[Brian]

I had a problem with the lstrcpyA function. I watched the memory on the
destination string, and only the first character of the src string was
copied due to its unicode nature. I wrote the following function to take
care of the issue. Is there a better way. The destination string is a
non-unicode string.

public int CopyString(stri ng src, IntPtr dest, UInt32 len)
{
byte [] copiedSrc;
if (len < src.Length)
{
return 0;
}

copiedSrc = new byte [src.Length];

for (int i=0; i<src.Length; i++)
{
byte b = (byte)src[i];
copiedSrc[i] = b;
}

lstrcpyA(dest, copiedSrc);

return src.Length;
}
<big snip>

Thanks,
Brian

[BMermuys]

Hi,

<Brian> wrote in message news:10******** *****@corp.supe rnews.com...

I had a problem with the lstrcpyA function. I watched the memory on the
destination string, and only the first character of the src string was
copied due to its unicode nature.
[DllImport(@"ker nel32.dll", CharSet=CharSet .Ansi)]
public extern static void lstrcpyA(IntPtr p, string s);

Because of the CharSet.Ansi parameter the framework should first re-code the
unicode string (s) into an ansi one before copy. It works for me, don't
know why you see this behaviour.
I wrote the following function to take
care of the issue. Is there a better way. The destination string is a
non-unicode string.
There's a better way (managed) to convert an unicode string into ansi bytes
and then you don't need lstrcpyA anymore but you can use Marshal.Copy :

public int CopyString( string src, IntPtr dest, UInt32 len )
{
// truncate if src too long
if ( src.Length > (len-1) ) src = src.Substring(0 , len-1);

// Get ansi bytes including null terminator
byte [] temp = System.Text.Enc oding.Default.G etBytes( src + '\0' );

Marshal.Copy( temp, 0, dest, temp.Length );
return src.Length;
}

hth,
Greetings

public int CopyString(stri ng src, IntPtr dest, UInt32 len)
{
byte [] copiedSrc;
if (len < src.Length)
{
return 0;
}

copiedSrc = new byte [src.Length];

for (int i=0; i<src.Length; i++)
{
byte b = (byte)src[i];
copiedSrc[i] = b;
}

lstrcpyA(dest, copiedSrc);

return src.Length;
}
<big snip>

Thanks,
Brian