Why can I not declare a variable with the same name as a variable in an anonymous method

Question

I am creating a method to load default users into the identity database that comes with the MVC5 template. I have ended up with a compile error on the line (see full code below)

var user = CreateUser("email@email.email",true);

The exact error is Error 16 A local variable named 'user' cannot be declared in this scope because it would give a different meaning to 'user', which is already used in a 'child' scope to denote something else

I do not understand how this is happening, and what I find equally baffling is a soon as I comment out this line the next line cannot find variable user. If the compile refuses to let me use user due to a previous declaration, then how can it fail to find user after the duplicate declaration is removed.

I think this is related to closure/inline methods holding on to the name but am not sure. I also know I can fix this by just changing the variable name, I am more interested in the why.

Here is a simplified example of the method I am currently writing:

    internal static void CreateDefaultUsers(IdentityDb context)
{
    var defaultPassword = "admin";
    var userManager = new UserManager<User>(new UserStore<User>(context));
    var roleManager = new RoleManager<UserRole>(new RoleStore<UserRole>(context));

    Func<string, User> CreateUser = (string email) =>
    {
        User user = context.Users.SingleOrDefault(u => u.UserName == email);

        if (user == null)
        {
            user = new EziOrderUser { 
            FirstName = email.Substring(0, email.LastIndexOf("@")),
            Email = email, UserName = email, EmailConfirmed = true };

            userManager.Create(user, defaultPassword);

            userManager.AddToRole(user.Id, "User");
        }

        return user;
    };

    var user = CreateUser("email@email.email"); // <-- Error here
    userManager.AddToRole(user.Id, "Administrator"); // <-- Then here
}

(Have not checked that this sample compiles, it does illustrate the problem however)

Answer 1

Basically it's a feature made to make life easier for you. :-) If you could do this, you can get type clashes.

{
// scope 1

Func<string, User> CreateUser = (string email) =>
{
    // scope 2
    User user = context.Users.SingleOrDefault(u => u.UserName == email);

    // ...**1
    return user;
};

// **2
var user = CreateUser("email@email.email");
userManager.AddToRole(user.Id, "Administrator"); // <-- Then here

In your example, what is the type of User at **2 and **1? Say that CreateUser returns a string - that would infer that in scope 1 User can get both type string and type User. Since scope 2 is closed, you can argue that this is the correct one to use, but that will make it harder for developers to read the code. Enforcing that you can only use a name once per scope makes it easier.

Not all languages have this behavior, f.ex. C++ allows you to do these kinds of things. (which is probably the reason they did this in the first place :-)

By the ways, it's okay to re-use names, as long as they don't appear in the same scope. For example:

Func<string, User> CreateUser = (string email) =>
{
    User user = context.Users.SingleOrDefault(u => u.UserName == email);

    // ...
    return user;
};
{
    // New scope - user is not contained in a shared scope. This is OK.

    var user = CreateUser("email@email.email");
    userManager.AddToRole(user.Id, "Administrator"); 
}