Can I use a for loop to add anonymous ActionListener objects?

Question

I have a setListener() method that has the following inside:
for(int k = 0; k < buttons.length; k++)
    buttons[k].addActionListener(new ActionListener()
        public void actionPerformed(ActionEvent e)
            g2.setColor(colors[k]);
}I have a JButton array and a Color array and I was hoping I could add them quickly in one shot rather than manually adding an anonymous ActionListener 9 times (I have 9 components). It tells me I need to make k final for an inner class and when I tested it by removing the for loop and keeping k as a final integer, it works. Is there a medium such that I can achieve what I want more or less while respecting Java's syntax?
Any input would be greatly appreciated!
Thanks in advance!

Answer 1

s3a wrote:
The local variable exists on the stack for as long as the method is executing. Once the method ends, that variable no longer exists. The newly created object, however, can continue to exist long after the method has ended. So when it's referring to variable X, it cannot refer to the same X as the one in the method, because that variable may no longer exist.Sorry for picking on little details but I am still not fully satisfied.
Earlier I questioned if the local variable changed, but now let's say that that variable no longer exists at all, why does that even matter if the inner class copied the value and is keeping it for itself? What do you mean? The variable that existed in the method ceased to exist when the method ended. The inner class, however, can continue to live on, and can continue to refer to what we see as the same variable. However, it obviously can't be the same variable, since, in the world of the inner class, it still exists, while in the world of the method, it does not.
This is completely independent of whether the variable changes or not. Regardless or whether the variable changes we still need two copies--one that goes away when the method's stack frame is popped, and one that lives on as long as the inner object does.
Then, because there are two copies, the designers decided that the variable has to be final, so that they wouldn't have to mess with the complexity of keeping two variables in sync.
That explanation leads me to believe that there is no copy and that there is some kind of "permanent umbilical cord" between the inner class and the local variable.Wrong. There has to be a copy. How else could the inner class keep referring to it after the method ends?
Also, does marking it as final force it to be a permanent constant "variable" even if it's not a field "variable"? Or, similarly, does making a "variable" final make Java pretend that it is a field "variable"?Making a variable final does exactly one thing: It means the value can't change after it's been initialized. This is true of both fields and locals.
As for the "pointless" byte-counting, I really don't see how it's confusing unless you're an absolute beginner. If I see somebody using a byte, I assume they have a real reason to do so, not pointless byte-hoarding. Then when I see that it's just a loop counter, I wonder WTF they were thinking, and have to spend some time determining if there's a valid reason, or if they're just writing bad code. Using a byte for a loop counter is bad code.
I could then ask, why are people not using long instead of int and saying that that's using int is too meticulous?Part of it is probably historical. Part of it is because at 4 bytes, and int is exactly on word on the vast majority of hardware that Java has targeted since it came out. 8-byte words are becoming more common (64-bit hardware) but still in the minority.