Managed C++/CLI Discussion Boards

It's called tough love. If you want the help, you need to follow the protocol.

"We make a living by what we get, we make a life by what we give." --Winston Churchill

Member 2021022 wrote:
I see Diplomacy is not one of your strong suites!!

Really? That's what you are going with? You are the guy that has misused the site and you are going to talk about diplomacy. I see Logic is not one of your strong suites.

Member 2021022 wrote:
I'm porting a WindowsControlLibrary solution from DotNet2003 to 2005

The first thing I'd do is stop using the old syntax...otherwise it's not really a port...

Mark

Mark Salsbery
Microsoft MVP - Visual C++
Java | [Coffee]

Is there porting guide you can direct me to that has examples of the newer syntax?

regards,
-dan

Member 2021022 wrote:
Is there porting guide

No. And if there was one it wouldn't be provided by Microsoft and it wouldn't be located on MSDN and you certainly could not find it using Google. Do you have to have someone wipe your arse for you? Just trying some more of that thar diplormacy thang.

Managed Extensions for C++ Syntax Upgrade Checklist[^]

Mark Salsbery
Microsoft MVP - Visual C++
Java | [Coffee]

This was extremely useful. Thanks
-dan

Hello,

I am using C++/CLI and I am trying to use the function SaveFileDialog but when I run the application in Debug mode when the function ShowDialog()is called I get the error:

"An unhandled exception of type 'System.Threading.ThreadStateException' occurred in System.Windows.Forms.dll

Additional information: Current thread must be set to single thread apartment (STA) mode before OLE calls can be made. Ensure that your Main function has STAThreadAttribute marked on it. This exception is only raised if a debugger is attached to the process."

And the program quit.

In Release mode the functions SaveFileDialog and OpenFileDialog looks like to work properly.

Can anybody help me wirh this error?

Thank you

Exception message:
Current thread must be set to single thread apartment (STA) mode before OLE calls can be made.

[STAThread]
int main()
{
    return 0;
}

Eslam Afifi

And if I have no main?

I have a dll with an exported function.

When you call this function a Windows Form is opened. In this Windows Form I am trying to add a button "Save File" and "Load File" but here is where I get the exception.

Thank you for you answer

I think you must either make the windows forms STA or run your dll code in a separate STA thread.

// C# 3.0 code
var thread = new Thread(() => { new SaveFileDialog().ShowDialog(); });
thread.SetApartmentState(ApartmentState.STA);

thread.Start();

Eslam Afifi

I think that should be the way, because the *.exe file that use my dll it is not a managed project, so it hasn´t the property STA.

Thank you,

You're welcome.

Eslam Afifi

hi. i have a class like this :

public ref class MyClass : public IDisposable
{
.
.
.
public: MyClass();
private: void Wait();
.
.
.
}

and the cpp file:

MyClass::MyClass()
{
    // Create a new thread to wait for events
    Thread^ t = gcnew Thread(gcnew ThreadStart(&MyClass::Wait));
    t->Name = "Event Thread";
    t->Start();
}

in compile i got this error :
"error C3350: 'System::Threading::ThreadStart' : a delegate constructor expects 2 argument(s)"

I've looked in msdn and saw a example that exactly used this way for threading but there was no error in compilation.
thank you for help.

sometimes 0 can be 1

erfi wrote:
I've looked in msdn and saw a example that exactly used this way for threading

So why didn't you post a link to the documentation you claim to have seen?

Try looking at the example code in this documentation[^]

erfi wrote:
I've looked in msdn and saw a example that exactly used this way

Exactly?

I have a feeling the Wait() function was static in the example.

Using a non-static function as you've shown won't work because a MyClass
object (instance) is required to make the call to Wait(). You'd have to pass
an instance to the delegate constructor.

Mark

Mark Salsbery
Microsoft MVP - Visual C++
Java | [Coffee]

Hi;

I'm finally digging into Reflection and thought, as an exercise, to create an instance of MessageBox and then to modify some of it's (private) properties, such as the font.

Using GetFields() to extract the FieldInfo, all I can get a few items out of it (static values which are, effectively, the return values). As for get Constructor: I get nothing. Creating an instance wants args (for the constructor) which I can't seem to get.

Testing on a very simple home-made method, I can read/write to public/protected/private fields, and the above listing methods (for fields, constructors) works fine.

Any comments and/or steering suggestions for my quest? (code snippets below)

Balboos

Reflection::Assembly ^a = Reflection::Assembly::GetAssembly(MessageBox::typeid);
// Gets all types/methods/whatever defined in the above assembly.
array<type^>^ typeArray = a->GetTypes(); // MessageBox is at index=1213

Type ^TypeToReflect = typeArray[1213];

Object ^n = TypeToReflect->InvokeMember("IDOK", BindingFlags::GetField | BindingFlags::Public | BindingFlags::NonPublic | BindingFlags::Instance | BindingFlags::Static, nullptr, TypeToReflect, nullptr );

// This returns a short list of items; mainly IDOK, IDCANCEL, etc.
array<system::reflection::fieldinfo xmlns:system="#unknown"> ^fields = TypeToReflect->GetFields(
        BindingFlags::Public | BindingFlags::NonPublic | BindingFlags::Instance | BindingFlags::Static);

// This returns an empty array
array<system::reflection::constructorinfo> ^Members = TypeToReflect->GetConstructors();

</system::reflection::constructorinfo></system::reflection::fieldinfo>

"The difference between genius and stupidity is that genius has its limits." - Albert Einstein

"How do you find out if you're unwanted if everyone you try to ask tells you to stop bothering them and just go away?" - Balboos HaGadol

"It's a sad state of affairs, indeed, when you start reading my tag lines for some sort of enlightenment. Sadder still, if that's where you need to find it." - Balboos HaGadol

MessageBox is a static class, there isn't any constructor, there is just one useful method and all it does is wrap the native MessageBox function. It probably is the worst class for getting acquainted with reflection!

Smile | :)

Luc Pattyn [Forum Guidelines] [My Articles]

- before you ask a question here, search CodeProject, then Google
- the quality and detail of your question reflects on the effectiveness of the help you are likely to get
- use the code block button (PRE tags) to preserve formatting when showing multi-line code snippets

modified on Tuesday, March 31, 2009 12:01 PM

Thanks - I can't stop spinning those wheels (an exhausting job, this running in place).

Ok - So it's a bad choice.

Still, why can't I get to any of it's internal info (particularly members), aside from the enum of return values, via reflection? I suppose when I start to understand what I'm doing, that will become clear. The link to the native version brings to mind MessageBoxA (of what seems so long ago).

I'll certainly choose another target that can be interesting. The static classes, like MessageBox, seemed to hold so much promise for tweaking fun.

Once again, thanks.

Balboos

I have a Windows service written in C#, that needs to load a C++ SDK. I have written CLI classes, that acts as a bridge. The problem is that my CLI code fails at the point where it makes call to the C++ SDK. I am not getting any error message also. It just hangs. If I try the same from Win Forms application, I did not see any problem. For me it looks like there is issue only when using Windows service. Is it possible to consume CLI\C++ dll's inside windows serive?

Please clarify me on this.. Any link related to the problem will be much helpful.

BicycleTheif wrote:
Is it possible to consume CLI\C++ dll's inside windows serive?

Yes.

Mark Salsbery
Microsoft MVP - Visual C++
Java | [Coffee]

Hi Everyone:
Recently while I've been trying to watch the TV, there are pop-up ads that cover up the bottom quarter of the screen that occur at regular intervals. This is especially annoying when the program contains subtitles. This occurs on every commercial TV network.

Has anyone thought about making a device we can attach to our cable boxes that can filter out the pop ups? Since there is one for the internet I thought there might be one for cable TV.

I wouldn't know where to begin.

Thanks,

Quecumber256

I'm working on a mixed-mode project, and I came across an impossible situation today. One piece of my code managed code was not working, and I narrowed it down to the fact that it's evaluating a FALSE value as TRUE!

Don't believe me? Take a look at the screen shot of my debugger.

Line 1458 contains the IF clause that is mis-evaluating. You can see the step-trace arrow has navigated into line 1459. The very explicit comparison "...bCamFollowsTrain == true" evaluates to true even though the debugger watch window very clearly shows that bCamFollowsTrain == false. So therefore false == true!

If I continue step-tracing into the CameraMove() call at 1460, that method also contains a check on the same variable and also evaluates to "true" even though the watch window continues to show that its value is "false".

All I can think here is memory corruption. Does anyone else have any ideas? Thanks.

modified on Saturday, March 28, 2009 6:21 PM

I found the problem...

First, the dialog where the user can set the value of bCamFollowsTrain had backwards logic. When the "yes" radio button was selected, bCamFollowsTrain got assigned a value of FALSE, and vice versa. That's my bad. Once I fixed that, my code functioned as expected.

So then the question became "Why is the Watch window showing me the wrong value?" I inserted a line to explicitly update the value of bCamFollowsTrain just before the IF statement. In step tracing, I noticed then that the value in the Watch window shows correctly, regardless of whether I assign true or false.

The user updates this value via a modeless dialog box. Since the dialog is modeless, the update is performed by delegate callback in the RadioButton_SelectedIndexChanged event. When I removed the explicit assignment code and went back to using this dialog, the Watch window began showing the wrong value again. In fact, I was able to duplicate the situation in reverse, making the debugger show bCamFollowsTrain == TRUE when I had set it to FALSE (and when step-tracing took the FALSE path).

Therefore, my conclusion is that it seems the Watch window has some issues with picking up on variables changed via delegate callbacks. As of right now, my code is working completely as expected while my debugger is still showing incorrect values.

*sigh*

Hi
I currently working on building a VC++.NET wrapper for a C++ DLL. Everything compiles fine, but for some reason any methods that require pointers to be passed do not appear to be public in the wrapper class. As an example:

The prototype for a function in the C++ DLL might be:

const char* error_msg( );

so I create a corresponding function in the VC++.NET wrapper class as:

String^ o10xxBase::flash_m::error_msg()
{
	return(gcnew System::String (m_flash->error_msg()));
}

This seems to work fine. Another example that requires input parameters... the prototype for the C++ function might be:

void WriteIoRegisterDW( unsigned int addr, unsigned int data );

so I create a corresponding function in the VC++.NET wrapper class as:

void o10xxBase::o10xxIo_m::WriteIoRegisterDW( unsigned int addr, unsigned int data )
{
	(m_o10xxIo->WriteIoRegisterDW( addr, data ));
}

Again this works fine it seems. However if the C++ prototype requires a pointer input things don't work so well. For example, prototype for the C++ function might be:

void read_fifo( short *p_Buffer, int count );

so I create a corresponding function in the VC++.NET wrapper class as:

void o10xxBase::adi_9446_a2d_m::read_fifo( short * p_Buffer, int count )
{
	(m_adi_9446_a2d->read_fifo( p_Buffer, count ));
}

This compiles with no errors, and yet the function does not appear in the wrapper class even though it is listed as public. Seems to be some issue I am overlooking with the movement of pointers from managed to unmanaged code.

Any thoughts most welcome.

Kind regards, Kurt