Data Conversion Problem

[azapf1972]

Pete,

first of all, thanks for that outstanding piece of software called FSUIPC... I've been playing with it for a few days now and tried to use it with C# (I know, I know, you don't use C#...)

The C# Code contained in the SDK works fine - it just left me with one problem: I was unable to retrieve and values that you defined as FLOAT (i.e. AI Traffic Block LAT, LON, ALT) with the FSUIPC_Get methods provided.

I have been pushing if forwards and backwards and finally wrote myself another method overload for FSUIPC_Get to handle floats.

		public bool FSUIPC_Get(ref int Token, ref float Result) 
		{
			int Size = 4;    // 2 bytes in an int
			float[] myFloat = new float[1];

			if ((Token < 0) || (Token > IPC_BUFFER_SIZE - (4 + Size)) ) 
			{ //Token out of range
				Result = 0;
				return false;
			}

			IntPtr heapbuf = Marshal.AllocHGlobal(Size);
			Marshal.Copy(IPC, Token + 4, heapbuf, 4);
			Marshal.Copy(heapbuf, myFloat, 0, 1 );

			Result = myFloat[0];

			Marshal.FreeHGlobal(heapbuf);
			if (IPCdr[Token] ) 
			{
				IPCdr[Token] = false;
				return true;
			} 
			else 
			{    // if (data ready flag not set, function returns false and value found
				return false;
			}
		}

Now, most of the code is a simple copy from one of the existing ones - I just adopted the following section

IntPtr heapbuf = Marshal.AllocHGlobal(Size);
Marshal.Copy(IPC, Token + 4, heapbuf, 4);
Marshal.Copy(heapbuf, myFloat, 0, 1 );

Result = myFloat[0];

and of course had to define

float[] myFloat = new float[1];

.

Unfortunately, the Marshal.Copy does only provide an overload for a float[] Array - not a single float, hence that funny one-element array.

To make a long story short: it works - I can now define a float variable in the code, pass it to FSUIPC_Get and get the proper results from the AI Traffic tables.

What I want to know: is there any other (simpler) way and how have others using C# overcome my initial problem of not being able to read a float properly...

Any thoughts are welcome - I want to learn...

Andreas

[Pete Dowson]

I was unable to retrieve and values that you defined as FLOAT (i.e. AI Traffic Block LAT, LON, ALT) with the FSUIPC_Get methods provided.

That's odd. Doesn't C# support the standard Intel 32-bit floating point format?

What I want to know: is there any other (simpler) way and how have others using C# overcome my initial problem of not being able to read a float properly...

Well, I do hope there will be someone who knows C# enough. Seems odd the way Microsoft develop these languages, making them more and more restrictive. I was surprised at many of the deficiencies and quirks of VB.

Regards,

Pete

[Rhysa]

I had this problem too trying to read an Unsigned Int32 in VB.Net. So I too added another Get method that looked like this:

    Overloads Function FSUIPC_Get(ByRef Token As Integer, ByRef Result As UInt32) As Boolean
        Dim Size As Integer = 8    ' 8 bytes in a Double
        Dim InBuf(8) As Byte
        Dim i As Integer
        If (Token < 0) Or (Token > IPC_BUFFER_SIZE - (4 + Size)) Then 'Token out of range
            Result = Convert.ToUInt32(0)
            FSUIPC_Get = False
            Exit Function
        End If
        Result = BitConverter.ToUInt32(IPC, Token + 4)
        If IPCdr(Token) Then
            IPCdr(Token) = False
            FSUIPC_Get = True
        Else    ' If data ready flag not set, function returns FALSE and value found
            FSUIPC_Get = False
        End If
    End Function

Now the only lines that I had to chage was

Overloads Function FSUIPC_Get(ByRef Token As Integer, ByRef Result As UInt32) As Boolean

Result = BitConverter.ToUInt32(IPC, Token + 4)

Now C# and Vb.net are pretty similar and the only Get method in my code that uses Marshal.Copy is the string handler.

It seems the VB.net and C# sdks are very different.

[azapf1972]

That's odd. Doesn't C# support the standard Intel 32-bit floating point format?

Pete,

C# supports both 32bit and 64bit floating point variables - no problem there. What happened (at least to me - and that might just be because i did not find a "proper" way doing it...) is that when calling FSUIPC_Get, one has to pass a variable reference which will then take the result of the _Get call.

In the C#-SDK, there are four overloads declared for FSUIPC_Get, each differing in the datatype of the receiving variable. And although 32 bit are 32 bit (memory wise), the compiler does not allow to pass a float variable when the overload is declared to receive an int parameter - the compiler throws an error.

Also, loading the 4 byte into an integer and then casting to a float did not work either because the value was already interpreted before casted...

I will continue to look into this...maybe there is a different solution, maybe the SDK could use another _Get overload (but I am not sure mine is the best way of doing it.... I am rather sure it is currently not...)

Andreas