This is clearly not appears like it wouldn't be a best practice. Can someone explain why it would not be a best practice or how this works? Any books or articles providing an explanation would be appreciated.
//The constructor
public Page_Index() {
//create a local value
string currentValue = "This is the FIRST value";
//use the local variable in a delegate that fires later
this.Load += delegate(object sender, EventArgs e) {
Response.Write(currentValue);
};
//change it again
currentValue = "This is the MODIFIED value";
}
The value that is output is the second value "Modified". What part of the compiler magic is making this work? Is this as simple as keeping track of the value on the heap and retrieving it again later?
[Edit]: Given some of the comments, changing the original sentence some...
currentValue is no longer a local variable: it is a captured variable. This compiles to something like:
Jon Skeet has a really good write up of this in C# in Depth, and a separate (not as detailed) discussion here.
Note that the variable currentValue is now on the heap, not the stack - this has lots of implications, not least that it can now be used by various callers.
This is different to java: in java the value of a variable is captured. In C#, the variable itself is captured.
You need to capture the value of the variable within the closure/delegate, else it can be modified, like you saw.
Assign currentValue to a variable local (inside) to the delegate.
That is the point; it really isn't a local variable any more - at least, not in terms of how we normally think of them (on the stack etc). It looks like one, but it isn't.
And for info, re "not good practice" - anonymous methods and captured variables are actually an incredibly powerful tool, especially when working with events. Feel free to use them, but if you are going down this route, I would recommend picking up Jon's book to make sure you understand what is actually happening.