views:

672

answers:

1

Following up on my previous question concerning the Windows 7 taskbar, I would like to diagnose why Windows isn't acknowledging that my application is independent of javaw.exe. I presently have the following JNA code to obtain the AppUserModelID:

public class AppIdTest {

    public static void main(String[] args) {
        NativeLibrary lib;
        try {
            lib = NativeLibrary.getInstance("shell32");
        } catch (Error e) {
            System.err.println("Could not load Shell32 library.");
            return;
        }
        Object[] functionArgs = new Object[1];
        String functionName = null;
        Function function;
        try {
            functionArgs[0] = new String("Vendor.MyJavaApplication")
                    .getBytes("UTF-16");
            functionName = "GetCurrentProcessExplicitAppUserModelID";
            function = lib.getFunction(functionName);
            // Output the current AppId
            System.out.println("1: " + function.getString(0));
            functionName = "SetCurrentProcessExplicitAppUserModelID";
            function = lib.getFunction(functionName);
            // Set the new AppId
            int ret = function.invokeInt(functionArgs);
            if (ret != 0) {
                Logger.out.error(function.getName() + " returned error code "
                        + ret + ".");
            }
            functionName = "GetCurrentProcessExplicitAppUserModelID";
            function = lib.getFunction(functionName);
            // Output the current AppId
            System.out.println("2: " + function.getString(0));
            // Output the current AppID, converted from UTF-16
            System.out.println("3: "
                    + new String(function.getByteArray(0, 255), "UTF-16"));
        } catch (UnsupportedEncodingException e) {
            System.err.println("System does not support UTF-16 encoding.");
        } catch (UnsatisfiedLinkError e) {
            System.err.println(functionName + " was not found in "
                    + lib.getFile().getName() + ".");
        }
    }
}

The output of the application is seemingly gibberish:

1: ‹ÿU‹ìƒìL¡¬Ÿv3ʼnEüSV‹uƒ&
2: ‹ÿU‹ìƒìL¡¬Ÿv3ʼnEüSV‹uƒ&
3: ????????????????P???????????

Being aware of the fact that the output may be UTF-16, in (3) I attempted to convert a byte array from UTF-16. In all honesty I don't know if my approach here is right as (a) I don't know the size of a PWSTR and (b) I don't know if GetCurrentProcessExplicitAppUserModelID is indeed returning a byte array or string.

I'm aware that JSmooth will run the GUI process in a wrapper which simulates this effect. Launch4j claims to do the same, but doesn't appear to work. I am looking to have the AppUserModelID set regardless of the Java wrapper.

What is going wrong here?

+3  A: 

Hello Paul!

I didn't see your question before otherwise I would have given a try even without a bounty.

Here is what I came up with. Please note, as stated in the code itself, I didn't implement proper memory clean up with the CoTaskMemFree function (from Ole32.dll). So I suggest you take only the implementation for SetCurrentProcessExplicitAppUserModelID()

package com.stackoverflow.AppIdTest;

import com.sun.jna.Native;
import com.sun.jna.NativeLong;
import com.sun.jna.Pointer;
import com.sun.jna.WString;
import com.sun.jna.ptr.PointerByReference;

public class AppIdTest
{

  public static void main(String[] args) throws Exception
  {
    setCurrentProcessExplicitAppUserModelID(AppIdTest.class.getName());

    System.out.println(getCurrentProcessExplicitAppUserModelID());
  }

  // DO NOT DO THIS, IT'S JUST FOR TESTING PURPOSE AS I'M NOT FREEING THE MEMORY
  // AS REQUESTED BY THE DOCUMENTATION:
  //
  // http://msdn.microsoft.com/en-us/library/dd378419%28VS.85%29.aspx
  //
  // "The caller is responsible for freeing this string with CoTaskMemFree when
  // it is no longer needed"
  public static String getCurrentProcessExplicitAppUserModelID()
  {
    final PointerByReference r = new PointerByReference();

    if (GetCurrentProcessExplicitAppUserModelID(r).longValue() == 0)
    {
      final Pointer p = r.getValue();


      return p.getString(0, true); // here we leak native memory by lazyness
    }      
    return "N/A";
  }

  public static void setCurrentProcessExplicitAppUserModelID(final String appID)
  {
    if (SetCurrentProcessExplicitAppUserModelID(new WString(appID)).longValue() != 0)
      throw new RuntimeException("unable to set current process explicit AppUserModelID to: " + appID);
  }

  private static native NativeLong GetCurrentProcessExplicitAppUserModelID(PointerByReference appID);
  private static native NativeLong SetCurrentProcessExplicitAppUserModelID(WString appID);


  static
  {
    Native.register("shell32");
  }
}

Does it work for you?

At least here it correctly prints back:

com.stackoverflow.AppIdTest.AppIdTest

Gregory Pakosz
Brilliant! It works like a charm! I love it how without any further modifications my application name and icon appear for the pinned application. Thank you very much!
Paul Lammertsma
I'm glad I helped. I never used JNA before that's why I was curious about your question. However, I didn't have access to a Windows 7 computer until today
Gregory Pakosz
and again, don't read the value back in your real application, otherwise you need to enhance the implementation to correctly clean up the memory after having converted the appID back to a Java String
Gregory Pakosz
Does the memory linger until the application closes? If so, then it's not something I really need to worry about as the function will only be called once per run. Nevertheless, the current implementation only calls `setCurrentProcessExplicitAppUserModelID()`.
Paul Lammertsma
The worry is with `getCurrentProcessExplicitAppUserModelID()` and yes I would say when the application closes it's reclaimed anyway. The point of implementing `getCurrentProcessExplicitAppUserModelID()` was getting feedback from Windows 7 before answering you and also being curious about jna and passing a `WString` by reference
Gregory Pakosz
It's elegant and simple; I'll bear in mind that the memory needs clearing when using functions that take pointers as arguments.
Paul Lammertsma
not necessarily the case when using functions that take pointers as arguments -- but for this specific one it is, as stated in the function contract
Gregory Pakosz