Managed C++/CLI

Re: Managed/Unmanaged: pinning ptr madness

27-Feb-03 1:56

Thank you for your answer.

It is obvious that passing a cast (non-pin) pointer to a native function is dangerous for heap objects, as the GC may compact the heap, move objects etc. But what about __value object? They are stored on stack, aren't they? Will stack object be move during collection, too?

How heavy would you say is pinning a pointer? Is it a real performance issue if I keep the pointer pinned only for some native calculations that won't last more then a view microseconds?

- Andre

VizOne28-Feb-03 0:31

Re: Managed/Unmanaged: pinning ptr madness

28-Feb-03 0:31

Me again!

Funnily, during a test I found out, that pinning the pointer first and then passing it to the native function works slightly faster...

Of course reinterpret_cast is an evil monster, but I have to use it here, haven't I?

Let's say, I have those structs:

<code>
[StructLayout(LayoutKind::Sequential, Pack=4)]
public __value struct ManagedStruct
{
float AddComponents();
float a, b, c;
};

#pragma pack(push, 4)
struct NativeStruct
{
float a, b, c;
}
#pragma pack(pop)
</code>

I guess these structs are guaranteed to be binarilly identical.

Now I have a function:
<code>
// the native version
#pragma unmanaged
float AddComponents(NativeStruct * p) { return p->a + p->b + p->c; }
#pragma

</code>

If I want to call this from within managed code via IJW, I'd do this:

<code>

float ManagedStruct::AddComponents()
{
ManagedStruct __pin * pinThis = this;
return ::AddComponents(reinterpret_cast<NativeStruct *>(pThis));

</code>

The reinterpret_cast is safe and ok, or not?

Is there another way of calling the native function?
- Andre

Jeff J28-Feb-03 22:25

Jeff J

28-Feb-03 22:25

Hi VizOne,

Sorry for the late reply.

VizOne wrote:
But what about __value object? They are stored on stack, aren't they? Will stack object be move during collection, too?

Although __value objects are typically stored on the stack, they can also be put on the heap. Stack objects are the same in .Net as they are in regular C/C++, so they won't move.

How heavy would you say is pinning a pointer? Is it a real performance issue if I keep the pointer pinned only for some native calculations that won't last more then a view microseconds?

I did some preliminary tests too, and I could barely measure any performance differences. I forced GC collections during them. The GC does not collect very often under normal circumstances, so I don't see pinning as costly either. Your tests jibe with mine.

Of course reinterpret_cast is an evil monster, but I have to use it here, haven't I?

I always do, and MC++ code that ships with .Net uses it too. I really think MS is just playing it "safe". As I recall, reinterpret_cast was just discouraged. The software world would grind to a halt if such casts were never used Wink | ;)

The reinterpret_cast is safe and ok, or not?

I usually call native functions from MC++ much like you do. As long as managed objects are pinned and the cast is appropriate, safety should not be of concern.

Cheers

Implementing IDisposable

VizOne24-Feb-03 9:49

24-Feb-03 9:49

Hi there!

I am currently currently working .net classes that implement IDisposable
following the msdn-guidelines, but I have a problem I hope you can help me
with:

Let's say these are my classes:

<br />
public __gc class B : public IDisposable<br />
{<br />
public:<br />
  // Destructor calls Dispose(bool)<br />
  ~B() { Dispose(false); }<br />
  // Implements interface<br />
  virtual void Dispose() { Dispose(true); GC::SuppressFinalize(this); }<br />
  // this is the workhorse, it handles manual and finalizer<br />
  // disposing differently as recommended by MS<br />
  virtual void Dispose(bool arg) { /* whatever */ }<br />
};<br />
<br />
<br />
public __gc class D : public B<br />
{<br />
public:<br />
  // overrides Dispose(bool), but ultimatly calls B::Dispose(bool)<br />
  virtual void Dispose(bool arg) { /* whatever */  B::Dispose(arg); }<br />
};<br />

Looks nice, but does not work Frown | :(

I can't do something like that:

<br />
D* d = new D();<br />
d->Dispose();<br />

as the hide-by-name convention does not allow that. In native C++ I could
use a using declaration:

<br />
public __gc class D : public B<br />
{<br />
  using B::Dispose;<br />
/*...*/<br />
};<br />

and everything was fine. However, this is not allowed in .net. Nevertheless
this implementation of IDispose is used throughout the framework. How did MS
do this? Did I miss something? Is it only possible in C#?

Thanks in advance.

- Andre (VizOne) Loker

Paul Selormey25-Feb-03 0:23

25-Feb-03 0:23

Try this template:

public __gc class Hello : public System::IDisposable 
{
public:
	Hello(void)
	{
	}

	virtual ~Hello(void)
	{
		Dispose(false);
	}

	////////////////////////////////////////////////////////////////////////////
	// The System::IDisposable interface implementations

	void Dispose()
	{
		Dispose(true);
	}

protected:
	void Dispose(bool disposing)
	{
	    // Normally clean up here....  

		if (disposing)
		{
			System::GC::SuppressFinalize(this);
		}
	}

};

Best regards,
Paul.

Jesus Christ is LOVE! Please tell somebody.

VizOne25-Feb-03 0:30

25-Feb-03 0:30

This was what I have done. However, I can't call Dipose() on an object of the derived class because it is hidden-by-name. I guess this template is only implemantable in C#, so I simply renamed Dispose(bool) to InternalDispose(bool), which works fine for me.

Nevertheless thanks for your answer.

- Andre

Paul Selormey25-Feb-03 0:49

25-Feb-03 0:49

The Dispose(Boolean) has nothing to do with .NET. The interface IDisposable only defines Dispose(), so you are doing anything wrong here.
The Dispose(Boolean) is defined as either private or protected, so external classes are not supposed to access it.

See the doc for IDisposable, there is an MC++ example given there to confirm your approach.

Best regards,
Paul.

Jesus Christ is LOVE! Please tell somebody.

VizOne25-Feb-03 0:58

25-Feb-03 0:58

Yes, I *do* know. And I don't want to access Dispose(bool), but Dispose(). I give you an example of what I mean:

<br />
public __gc class B : public IDisposable<br />
{<br />
public:<br />
    virtual ~B() <br />
    { <br />
        Dispose(false); <br />
    }<br />
<br />
    virtual void Dispose() <br />
    { <br />
        Dispose(true); <br />
    }<br />
<br />
protected:<br />
    virtual void Dispose(bool arg)<br />
    {<br />
        /* Clean up */ <br />
    }<br />
};<br />
<br />
public __gc class D : public B<br />
{<br />
protected:<br />
    virtual void Dispose(bool arg)<br />
    {<br />
        /* Clean up */<br />
        // call base classes implementation<br />
        B::Dispose(arg);<br />
    }<br />
};<br />
<br />
int _tmain(void)<br />
{<br />
    D * d = new D();<br />
    d->Dispose();<br />
    return 0;<br />
}<br />

This is implemented as suggested in MSDN, isn't it? However, it does not work.

d->Dispose() generates a C2660 saying that Dispose does not accept 0 parameters. This error is obviously caused by the fact that Dispose() is hidden-by-name.

I could use d->B::Dispose() but this looks ugly to me, so I renamed Dispose(bool) to InternalDispose(bool).

Hope I could explain what I meant.

- Andre

Paul Selormey25-Feb-03 1:48

25-Feb-03 1:48

VizOne wrote:
Hope I could explain what I meant.

Yeah, but you are missing the point main here.
VizOne wrote:

I could use d->B::Dispose() but this looks ugly to me, so I renamed Dispose(bool) to InternalDispose(bool).

Add some output to the D::Dispose or D::InternalDispose to see if it is being called!

A better picture is cast the D pointer to IDisposable, it will compile, but similar to the above you end up calling B::Dispose, it is similar to your d->B::Dispose() syntax.

int _tmain(void)
{
	D * d = new D();
	__try_cast<IDisposable*>(d)->Dispose();
	return 0;
}

The best way is to implement the same interface in D, since you will not get the expected result by renaming.

Also, you trying to convert the IDisposable::Dispose() to virtual, it will compile but will not act as virtual method.

Best regards,
Paul.

Jesus Christ is LOVE! Please tell somebody.

VizOne25-Feb-03 2:45

25-Feb-03 2:45

Paul Selormey wrote:
Also, you trying to convert the IDisposable::Dispose() to virtual, it will compile but will not act as virtual method

Well I am sure it *does* act as a virtual method. Excerpt from MSDN doc on managed c++:

"The following features are supported for __gc interfaces:

All methods of an interface are implicitly pure virtual. Neither the keyword virtual nor the suffix =0 is required on an interface method declaration, although both are allowed"

Here is a small test in which I have included everything I mentioned so far. It works perfectly for me:

<br />
public __gc class B : public IDisposable<br />
{<br />
public:<br />
    virtual ~B() <br />
    { <br />
        Console::WriteLine("B::~B()");<br />
        InternalDispose(false); <br />
    }<br />
<br />
    virtual  void Dispose() <br />
    { <br />
        Console::WriteLine("B::Dispose()");<br />
        InternalDispose(true); <br />
    }<br />
<br />
protected:<br />
    // renaming this function to InternalDispose solved my problem<br />
    virtual void InternalDispose(bool arg)<br />
    {<br />
        Console::WriteLine("B::InternalDispose(bool)");<br />
        /* Clean up */ <br />
    }<br />
};<br />
<br />
public __gc class D : public B<br />
{<br />
public:<br />
    // just to show that the function *is* virtual<br />
    void Dispose() <br />
    { <br />
        Console::WriteLine("D::Dispose()"); <br />
        B::Dispose();<br />
    }<br />
protected:<br />
<br />
    void InternalDispose(bool arg)<br />
    {<br />
        Console::WriteLine("D::InternalDispose(bool)");<br />
        /* Clean up */<br />
        // call base classes implementation<br />
        B::InternalDispose(arg);<br />
    }<br />
};<br />
<br />
int _tmain(void)<br />
{<br />
  B * b = new D();<br />
  b->Dispose(); // as virtual as can be *g*<br />
}<br />

The output is (as expected)
D::Dispose()
B::Dispose()
D::InternalDispose(bool)
B::InternalDispose(bool)

Of course, in my final version I will mark B::Dispose() as __sealed and throw away D::Dispose()

- Andre

Paul Selormey25-Feb-03 2:54

25-Feb-03 2:54

Nothing new, just re-implementing IDisposable to create the impression of a virtual method (in which case you do not have to use the "Internal" attachment) - have fun.

Best regards,
Paul.

Jesus Christ is LOVE! Please tell somebody.

Managed type???

Orkun GEDiK21-Feb-03 12:49

21-Feb-03 12:49

Hi,

I'm writing a wrapper class. So i have a DLL that contains too many C functions. Anyway, one function has void pointer for out parameter (2nd parameter) like following;

void Test(void* pIn, void* pOut);

Function is trying to fill pOut parameter with some info. The problem has started here. I want to convert void* value String or StringBuilder class. I can do it under unmanaged C/C++ such as following code;

char* cpTest = new char[100];
Test(.... , static_cast<void*>(cpTest));

Now, i want to do it under managed C++. Eventually, which way should i follow?

Thank for your response...

Orkun GEDiK

Paul Selormey23-Feb-03 1:46

23-Feb-03 1:46

You can pass the character array. Use the native structure System::IntPtr for the Test() function parameter.

You can then write

Test(..., System::IntPtr((void*)cpTest);

You can use it for both the in and out.

Best regards,
Paul.

Jesus Christ is LOVE! Please tell somebody.

Orkun GEDiK23-Feb-03 3:55

23-Feb-03 3:55

Thank you for your response. System has generated following error;

VoidPtr.cpp(19) : error C2440: 'type cast' : cannot convert from '__wchar_t __gc[]' to 'void *'
Can only convert a __gc array to or from Object * or Array *
VoidPtr.cpp(19) : error C2440: 'type cast' : cannot convert from '__wchar_t __gc[]' to 'void *'
Can only convert a __gc array to or from Object * or Array *

The source code at below;

#include "stdafx.h"

#using <mscorlib.dll>
#include <tchar.h>

using namespace System;
using namespace System::Runtime::InteropServices;

#pragma unmanaged
// Assume this block inside of DLL
void Test(void* pIn, void* pOut)
{
char* pTest = static_cast<char*>(pIn);
pOut = static_cast<void*>("Internal\0");
}

#pragma managed
void WrapperTest(String* strIn, String* strOut)
{
Test(System::IntPtr((void*)strIn->ToCharArray()), System::IntPtr((void*)strOut->ToCharArray()));
}

int _tmain(void)
{
String* strTestA = new String("Hello");
String* strTestB = new String("");
WrapperTest(strTestA, strTestB);
return 0;
}

Thank you...

Orkun GEDiK

Paul Selormey23-Feb-03 23:49

23-Feb-03 23:49

Now, I see the picture clearer. You do not need the WrapperTest function and the System::IntPtr I mentioned earlier.

Get the char* and pass it to the Test method:

const wchar_t __pin* psz = PtrToStringChars( pString )

The following is defined in vcclr.h header file included in VC.NET. This the same function used by the CString to convert the System::String to CString.

PtrToStringChars(String* pString)

Best regards,
Paul.

Jesus Christ is LOVE! Please tell somebody.

Orkun GEDiK24-Feb-03 10:48

24-Feb-03 10:48

Very good and useful information for me. Thank you. I've changed my code to following form;

<br />
#include "stdafx.h"<br />
<br />
#using <mscorlib.dll><br />
#include <tchar.h><br />
#include <vcclr.h><br />
<br />
using namespace System;<br />
using namespace System::Runtime::InteropServices; <br />
<br />
#pragma unmanaged<br />
// Assume this block inside of DLL<br />
void Test(void* pIn, void* pOut)<br />
{<br />
	char* pTest = static_cast<char*>(pIn);<br />
	pOut = ("Internal\0");<br />
}<br />
<br />
#pragma managed<br />
int _tmain(void)<br />
{<br />
	String* strTestA = new String("Hello");<br />
	String* strTestB = new String("");<br />
	wchar_t __pin* pTestA = PtrToStringChars(strTestA);<br />
	wchar_t __pin* pTestB = PtrToStringChars(strTestB);<br />
	Test(static_cast<void*>(pTestA), static_cast<void*>(pTestB));<br />
    return 0;<br />
}<br />

Now, i can pass the pointer into DLL function, but i can't get any data from DLL (2nd parameter of Test function. So, pOut should be filled some data at inside of DLL). Thank you for your help...

Best Regards...

Orkun GEDiK

Paul Selormey25-Feb-03 0:30

25-Feb-03 0:30

But pTestB is not a buffer to receive. Simply define a buffer and pass it in the out parameter. There is no need for the strTestB.

char szTestB[BufferSize] = {0};

should be sufficient.
In the Test() function, you can do something simply like this:

lstrcpy(pOut, "Internal");

Best regards,
Paul.

Jesus Christ is LOVE! Please tell somebody.

Orkun GEDiK25-Feb-03 10:59

25-Feb-03 10:59

Ok. That means, i can't use managed type for pOut parameter. Ok, now i'm happy. Thank you for your response. Good answer. Thank you.

Best Regards,
Orkun.

Orkun GEDiK

AppDomain question...

Orkun GEDiK21-Feb-03 11:28

Problems working witk wav data files

21-Feb-03 11:28

Hello,

I've question about AppDomain. Can i set or get any data between the processes (such proxy) that executed by "ExecuteAssembly" function. If there's a way, please send the code example.

Thank you.

Orkun GEDiK

matlab21-Feb-03 4:03

matlab

21-Feb-03 4:03

Hi!I have a problem working with Wav data files.I read a sample (short int) i do a cast into a double int to work with it. when i multiply or sum or divide for a constant and the results isn't entire the sound become saturated.The code seems like that:

[......]
fin.read((char*)&sample,sizeof(short));
double samdouble=sample;
samdouble=sample*0.5; // for example
sample=short(samdouble);
fout.write((char*)&sample,sizeof(short));
[.....]

I hope somebody can help me because i'm stopped in my project for that problem and i'm beginning to lose the expectancy

Unverifiable assembly 'GIS.View' failed policy check

Paul Selormey16-Feb-03 22:42

Re: Unverifiable assembly 'GIS.View' failed policy check

16-Feb-03 22:42

Hello All,
I compiled an MC++ codes and it includes some unmanaged classes.
The codes works alright in a local folder, but when I moved it to a network
folder (local network folder) I get the following IO exception:

-----------------------------------------------------------------------
An unhandled exception of type 'System.IO.FileLoadException' occurred in
WindowsApplication2.exe

Additional information: Unverifiable assembly 'GIS.View' failed policy check.
-----------------------------------------------------------------------

Is there no way to get MC++ assemblies to execute safely from a network folder?

Best regards,
Paul.

Jesus Christ is LOVE! Please tell somebody.

ThatsJustMe17-Feb-03 22:22

ThatsJustMe

17-Feb-03 22:22

I think that this has nothing to do specially with C++ and that
it happens also when trying to start a C#-programm from a network
share.

I think you can reconfigure your .NET-framwork with the utilities
which have been added to the star menu under Accessories, Management
(My XP is german, so these are guesses)

There you have to assign "FullTrust" to the network-share you want
to use.

Unfortunately I have not fully understood this stuff myself and find
it rather annoying.

Re: Unverifiable assembly 'GIS.View' failed policy check

Paul Selormey20-Feb-03 2:20

Need help with casting __gc-pointer to void* and then back again

20-Feb-03 2:20

Thanks for the information. I also got related article on the news group.

If the code isn't verifiable (if it contains unsafe code), it can cause 
security issues. If the assembly is loaded without sufficient permissions 
(such as when it is loaded from an untrusted location), then the code will 
not be allowed to execute.

To solve this problem, you can adjust the security on the destination 
machine. This can be done by caspol utility. 

For more information, please check following article:

http://msdn.microsoft.com/library/en-us/dnforms/html/winforms11122002.asp

For additional information about creating verifiable code in MC++, you may 
want to install .NET Framework SDK 1.1 Beta from:

http://msdn.microsoft.com/visualc/productinfo/visualc03/default.asp

The article titled "Producing Verifiable Components with Managed Extensions 
for C++" in the SDK maybe helpful.

Best regards,
Paul.

Jesus Christ is LOVE! Please tell somebody.

Anonymous16-Feb-03 12:52

Anonymous

16-Feb-03 12:52

Hi!

I want to code a wrapper for an existing c++-class.

The class has a function to define a callback and
when I set the callback I can give it some data
for me on the way: A void*. I can use it to
store stuff I need when I'm inside the callback.

This should be a common c++-concept and I think
the best thing for the managed proxy-class I'm
writing is a delegate, so that the C#-users who
call the function can give it a delegate which
is then called by the callback-proxy.

So I store the two delegates inside a class:

<br />
	__delegate void CallBackNewFile(String* fname,int size);<br />
	__delegate void CallBackBytes(int size);<br />
<br />
	__gc class ExtractInfo<br />
	{<br />
	public:<br />
		CallBackNewFile* newfile;<br />
		CallBackBytes* bytes;<br />
	};<br />

Now I want to give the callback-function
a void*:

<br />
		static bool Extract(String* cabfile,String* DestDir,bool DontMakeFolders,<br />
			CallBackNewFile* newfile,CallBackBytes* bytes)<br />
		{<br />
			CCabExtractor cab;<br />
			if (!cab.Create())<br />
				return false;<br />
			ExtractInfo __pin * inf=new ExtractInfo();<br />
			inf->newfile=newfile;<br />
			inf->bytes=bytes;<br />
			cab.SetCallback(&mycallback,inf);<br />
<br />

Please notice the "ExtractInfo __pin * inf=new ExtractInfo();". Is
that the right way of pinning?

Converting the gc-pointer into a void* seems to work!

BTW:
The SetCallback function has this declaration:

<br />
  void SetCallback (CALLBACKFUNC pCallback, void * pv)<br />

And now comes the problem:

My callbackfunction wants to do something with the passend void* pointer:

<br />
static bool __stdcall mycallback (void * pv, const char * lpszPath,  DWORD cb)<br />
		{<br />
			ExtractInfo* i=(ExtractInfo*)(pv);<br />
			if (lpszPath==NULL)<br />
			{<br />
				if (i->bytes)<br />
					i->bytes(cb);<br />
			} else {<br />
				if (i->newfile)<br />
					i->newfile(lpszPath,cb);<br />
			}<br />
			return true;<br />
		}<br />

The problem is the line "ExtractInfo* i=(ExtractInfo*)(pv);

I can't get this line right so that it compiles!

I tried several variants. I think it won't let me get
a __gc class out of a void* pointer.

BUT: That class HAS to be a __gc class, otherwise
the compiler complained about declaring delegates inside
it.

Can someone help me?

How do I transport a managed pointer/reference THROUGH
unmanaged code?
How can I cast a void* into a reference to a managed
object?

Many thanks in advance!!!

Re: Need help with casting __gc-pointer to void* and then back again

Paul Selormey16-Feb-03 19:26