🚀 go-pugleaf

I have a problem finding how to get around the limitation imposed by the
absence of destructors in Java.

The destructor in C++ was very useful to control not just deallocation
(which is done automatically by the garbage collector in Java) but alse
to restore a previous state when the state of something changes
temporarily within a limited scope, the scope of a method for instance.

Let me exemplify this :

Suppose that I want to change the cursor of a frame during a certain
operation and restore it to its previous state when I leave the method.

What I will do in Java is :

method() {
  Cursor oldCursor = getCursor();
  setCursor(new Cursor(Cursor.WAIT_CURSOR));

  // treatment ...

  // everytime I come across a return in my algorithm, just before the
return statement :
  setCursor(oldCursor);

  //...
}

Since this is rather boring, in C++ what I used to do in such a case was
to create an object, say "CursorContext", with an implementation
resembling this :

class CursorContext
{
private:
  Cursor oldCursor;
  Frame f;
public:
  CursorContext(Frame f, Cursor newCursor)
  {
    oldCursor = f.getCursor();
    f.setCursor(newCursor);
    this.f = f;
  }

  ~CursorContext()
  {
     f.setCursor(oldCursor);
  }

};


Then code in my method becomes (assuming I have created a global object
called waitCursor for example) :


MyFrame :: method(){
  CursorContext cursorContext(this, waitCursor);

  // treatment ...

  // when I come across a return statement, nothing to do, since
destructor of local object is called, restoring old cursor
}

Is there any way do to something approaching in Java ?

Thanks for any tips.
Bruno

--
P.S. Please remove nospam. prefix to reply to sender. Thanks
Veillez � supprimer le pr�fixe nospam. pour me r�pondre. Merci