implement ienumerable with ienumerable<T>

Question

What is the need to implement IEnumerable (Non-generic) with IEnumerable (Generic) interface in a generic collection class?
A code example on msdn states (Link - http://msdn.microsoft.com/en-us/library/9eekhta0(v=vs.110).aspx )

public class App
{
// Excercise the Iterator and show that it's more 
// performant. 
public static void Main()
{
    TestStreamReaderEnumerable();
    TestReadingFile();
}

public static void TestStreamReaderEnumerable()
{
    // Check the memory before the iterator is used. 
    long memoryBefore = GC.GetTotalMemory(true);

    // Open a file with the StreamReaderEnumerable and check for a string. 
    var stringsFound =
        from line in new StreamReaderEnumerable(@"c:\\temp\\tempFile.txt")
        where line.Contains("string to search for")
        select line;


    Console.WriteLine("Found: " + stringsFound.Count());

    // Check the memory after the iterator and output it to the console. 
    long memoryAfter = GC.GetTotalMemory(false);
    Console.WriteLine("Memory Used With Iterator = \t"
        + string.Format(((memoryAfter - memoryBefore) / 1000).ToString(), "n") + "kb");
}

public static void TestReadingFile()
{
    long memoryBefore = GC.GetTotalMemory(true);
    StreamReader sr = File.OpenText("c:\\temp\\tempFile.txt");

    // Add the file contents to a generic list of strings.
    List<string> fileContents = new List<string>();
    while (!sr.EndOfStream) {
        fileContents.Add(sr.ReadLine());
    }

    // Check for the string. 
    var stringsFound = 
        from line in fileContents
        where line.Contains("string to search for")
        select line;

    sr.Close();
    Console.WriteLine("Found: " + stringsFound.Count());

    // Check the memory after when the iterator is not used, and output it to the console. 
    long memoryAfter = GC.GetTotalMemory(false);
    Console.WriteLine("Memory Used Without Iterator = \t" + 
        string.Format(((memoryAfter - memoryBefore) / 1000).ToString(), "n") + "kb");

}
}

// A custom class that implements IEnumerable(T). When you implement IEnumerable(T),  
// you must also implement IEnumerable and IEnumerator(T). 
public class StreamReaderEnumerable : IEnumerable<string>
{

private string _filePath;
public StreamReaderEnumerable(string filePath)
{
    _filePath = filePath;
}


// Must implement GetEnumerator, which returns a new StreamReaderEnumerator. 
public IEnumerator<string> GetEnumerator()
{
    return new StreamReaderEnumerator(_filePath);
}

// Must also implement IEnumerable.GetEnumerator, but implement as a private method. 
private IEnumerator GetEnumerator1()
{
    return this.GetEnumerator();
}
IEnumerator IEnumerable.GetEnumerator()
{
    return GetEnumerator1();
}

} 

// When you implement IEnumerable(T), you must also implement IEnumerator(T),  
// which will walk through the contents of the file one line at a time. 
// Implementing IEnumerator(T) requires that you implement IEnumerator and IDisposable. 
public class StreamReaderEnumerator : IEnumerator<string>
{
private StreamReader _sr;
public StreamReaderEnumerator(string filePath)
{
    _sr = new StreamReader(filePath);
}

private string _current;
// Implement the IEnumerator(T).Current publicly, but implement  
// IEnumerator.Current, which is also required, privately. 
public string Current
{

    get
    {
        if (_sr == null || _current == null)
        {
            throw new InvalidOperationException();
        }

        return _current;
    }
}

private object Current1
{

    get { return this.Current; }
}
object IEnumerator.Current
{
    get { return Current1; }
}

// Implement MoveNext and Reset, which are required by IEnumerator. 
public bool MoveNext()
{

    _current = _sr.ReadLine();
    if (_current == null)
        return false;
    return true;
}

public void Reset()
{
    _sr.DiscardBufferedData();
    _sr.BaseStream.Seek(0, SeekOrigin.Begin);
    _current = null;
}
// Implement IDisposable, which is also implemented by IEnumerator(T). 
private bool disposedValue = false;
public void Dispose()
{
    Dispose(true);
    GC.SuppressFinalize(this);
}
protected virtual void Dispose(bool disposing)
{
    if (!this.disposedValue)
    {
        if (disposing)
        {
            // Dispose of managed resources.
        }
        _current = null;
        _sr.Close();
        _sr.Dispose();
    }

    this.disposedValue = true;
}

 ~StreamReaderEnumerator()
{
    Dispose(false);
}
// This example displays output similar to the following: 
//Found: 2 
//Memory Used With Iterator =     33kb 
//Found: 2 
//Memory Used Without Iterator =  206kb
}

Answer 1

The need is simply because of how the interfaces are implemented. In .NET 1.1, there were no generics, so IEnumerable has no generic support and only exposes object . This introduces boxing (and is why the compiler also supports a pattern-based implementation for foreach , independent of IEnumerable ). In C# 2, we got IEnumerable<T> . It is useful to consider all typed iterations as also being untyped, hence:

IEnumerable<T> : IEnumerable

where IEnumerable<T> re-declares GetEnumerator as with the generic type. However, since these are now different methods with different signatures, it needs 2 different implementations.