C# to C++ Array?

Question

I have a library provided, which i know uses C++. I imported the DLL as such:
[DllImport("pst")] private static extern int pst_get_sensor(ref PSTSensor sensor); The PSTSensor is a struct, in a C++ example it was defined as such:

struct PSTSensor
{
        char    name[80];       /**< Device name */
        int     id;             /**< Device identifier (for other tracking interfaces */
        float   pose[16];       /**< Device pose estimate as row-major matrix */
        double  timestamp;      /**< Time the data was recorded */
};

The problem is, that besides an Int and a Double, it uses a Float, and more importantly an array. An array is something way different between C# and C++. When using this struct to call pst_get_sensor(ref sensor); the entire development environment crashes on me when running the code.

I currently do my struct in C# like this:

struct PSTSensor{
        public char[] name;
        public int id;
        public float[] pose;
        public double timestamp;
    }

My question is, how do i create an array that C++ understands, and properly returns? Is this even possible?

Thanks a lot in advance, Smiley

David Heffernan · Accepted Answer · 2013-11-22 13:31:59Z

6

Your C# struct needs to define the inline array lengths. And the name field is best declared as a string in the C# code. That makes it much easier for you to work with it.

[StructLayout(LayoutKind.Sequential, CharSet=CharSet.Ansi)]
struct PSTSensor
{
    [MarshalAs(UnmanagedType.ByValTStr, SizeConst = 80)]
    public string name;
    public int id;
    [MarshalAs(UnmanagedType.ByValArray, SizeConst = 16)]
    public float[] pose;
    public double timestamp;
}

The struct layout attribute is optional here, because it just re-states default values. But you may choose to keep it in order to be explicit about your intent.

edited Nov 22, 2013 at 13:31

answered Nov 22, 2013 at 13:21

David Heffernan

616k46 gold badges1.1k silver badges1.5k bronze badges

Sign up to request clarification or add additional context in comments.

3 Comments

Smileynator Over a year ago

Ok, let me try this out. In the mean time. I probably will need to read up on what marchalling is, and what unmanagedTypes are (i do not code in C++ yet, so that makes this so much harder to understand)

David Heffernan Over a year ago

This MSDN topic will help: msdn.microsoft.com/en-us/library/…

Smileynator Over a year ago

It works, awesome! i decided to keep the optional line for future reference. Thanks a million, 3 teachers could not get me my answer, and i saw marshaling in other context, but i did not get what it does or how to implement it. Thanks!

kunal · Accepted Answer · 2013-11-22 13:40:34Z

1

You can use fixed keyword with allowing unsafe code option in properties

unsafe struct PSTSensor
{
    public fixed char name[80];
    public int id;
    public fixed float pose[16];
    public double timestamp;
}

This will be the declaration and for using passing it to cpp function you can call like this

unsafe 
{
   PSTSensor stSensor = new PSTSensor();
   testfunc(stSensor); //function call cpp
}

answered Nov 22, 2013 at 13:40

kunal

9669 silver badges16 bronze badges

3 Comments

David Heffernan Over a year ago

But now your struct is much more onerous to use and you are compelled to use unsafe code. Yes you can do this, but you should not.

kunal Over a year ago

i agree for unsafe ,but how is it more onerous?

David Heffernan Over a year ago

Ask yourself how you read and write the fixed arrays? And then compare with the variant that does not use unsafe code. Finally, not that char is two bytes wide in C#. So you'd need to use fixed byte name[80] and how would you access that?

Collectives™ on Stack Overflow

C# to C++ Array?

2 Answers 2

3 Comments

3 Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

3 Comments

3 Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related