If I did not get this terribly wrong, this behaviour is strange for me. Rather than explaining, I'll post a sample code below and please tell me why does I get output x and not y.
private void button1_Click(object sender, EventArgs e)
{
List<int> l = new List<int>() { 1, 2, 3 };
Fuss(l);
MessageBox.Show(l.Count.ToString()); // output is 5
}
private void Fuss(List<int> l)
{
l.Add(4);
l.Add(5);
}
Output should, I assume would be 3. But I get the output as 5. I understand the output can be 5 if I do this:
private void button1_Click(object sender, EventArgs e)
{
List<int> l = new List<int>() { 1, 2, 3 };
Fuss(ref l);
MessageBox.Show(l.开发者_如何转开发Count.ToString()); // output is 5
}
private void Fuss(ref List<int> l)
{
l.Add(4);
l.Add(5);
}
It does not act like its passed by ref.
void ChangeMe(List<int> list) {
list = new List<int>();
list.Add(10);
}
void ChangeMeReally(ref List<int> list) {
list = new List<int>();
list.Add(10);
}
Try it. Do you notice the difference?
You can only change the contents of list (or any reference type) if you pass it without a ref (because as others have said, you are passing a reference to the object on the heap and thus change the same "memory").
However you cannot change "list", "list" is a variable that points to an object of type List. You can only change "list" if you pass it by reference (to make it point somewhere else). You get a copy of the reference, which if changed, can only be observed inside your method.
Parameters are passed by value in C# unless they are marked with the ref
or out
modifiers. For reference types, this means that the reference is passed by value. Therefore, in Fuss
, l
is referring to the same instance of List<int>
as its caller. Therefore, any modifications to this instance of List<int>
will be seen by the caller.
Now, if you mark the parameter l
with ref
or out
, then the parameter is passed by reference. What this means is that in Fuss
, l
is an alias for storage location used as a parameter to invoke the method. To be clear:
public void Fuss(ref List<int> l)
called by
List<int> list = new List<int> { 1, 2, 3 };
Fuss(list);
Now, in Fuss
, l
is an alias for list
. In particular, if you assign a new instance of List<int>
to l
, the caller will see that new instance assigned to the variable list
as well. In particular, if you say
public void Fuss(ref List<int> l) {
l = new List<int> { 1 };
}
then the caller will now see a list with one element. But if you say
public void Fuss(List<int> l) {
l = new List<int> { 1 };
}
and call by
List<int> list = new List<int> { 1, 2, 3 };
Fuss(list);
then the caller will still see list
as having three elements.
Clear?
ByRef and ByVal only apply to value types, not to reference types, which are always passed as though they were "byref".
If you need to modify a list discreetly, use the ".ToList()" function, and you'll get a clone of your list.
Keep in mind that if your list contains reference types, your "new" list contains pointers to the same objects that your original list did.
Lists are already reference types, so when you pass them to a method, you are passing a reference. Any Add
calls will affect the list in the caller.
Passing a List<T>
by ref
behaves essentially like passing a double-pointer to that list. Here's an illustration:
using System;
using System.Collections.Generic;
public class Test
{
public static void Main()
{
List<int> l = new List<int>() { 1, 2, 3 };
Fuss(l);
Console.WriteLine(l.Count); // Count will now be 5.
FussNonRef(l);
Console.WriteLine(l.Count); // Count will still be 5 because we
// overwrote the copy of our reference
// in FussNonRef.
FussRef(ref l);
Console.WriteLine(l.Count); // Count will be 2 because we changed
// our reference in FussRef.
}
private static void Fuss(List<int> l)
{
l.Add(4);
l.Add(5);
}
private static void FussNonRef(List<int> l)
{
l = new List<int>();
l.Add(6);
l.Add(7);
}
private static void FussRef(ref List<int> l)
{
l = new List<int>();
l.Add(8);
l.Add(9);
}
}
A variable, parameter, or field of type "List", or any other reference type, doesn't actually hold a list (or object of any other class). Instead, it will hold something like "Object ID #29115" (not such an actual string, of course, but a combination of bits which means essentially that). Elsewhere, the system will have an indexed collection of objects called the heap; if some variable of type List holds "Object ID #29115", then object #29115 in the heap will be an instance of List or some type derived therefrom.
If MyFoo is a variable of type List, a statement like 'MyFoo.Add("George")' won't actually change MyFoo; instead, it means "Examine the object ID stored in MyFoo, and invoke the "Add" method of the object stored therein. If MyFoo held "Object ID #19533" before the statement executed, it will continue to do so afterward, but Object ID #19533 will have had its Add method invoked (probably altering that object). Conversely, a statement like "MyFoo = MyBar" will make MyFoo hold the same object-id as MyBar, but won't actually do anything to the objects in question. If MyBar held "Object ID #59212" before the statement, then after the statement, MyFoo will also hold "ObjectId #59212". Nothing will have happened to object ID #19533, nor object ID#59212.
The difference between ref and non-ref for reference types like List is not whether you pass a reference (that happens always), but whether that reference can be changed. Try the following
private void Fuss(ref List<int> l)
{
l = new List<int> { 4, 5 };
}
and you'll see the count is 2, because the function not only manipulated the original list but the reference itself.
Only primitive types like int, double etc. are passed by value.
Complex types (like list) are passed via reference (which is a passing pointer by value, to be exact).
lets explain it Easy :)
- "Fuss" in your first chunk of code knows that he should accept a refrence from a list type
- "Fuss" in your second chunk of code knows that he should accept nothing! but just a refrence to "sth else" which accidently this "sth else" is the same thing which was in your first chunk of code
精彩评论