What's new in C# 8.0

C# 8.0 adds the following features and enhancements to the C# language:

The remainder of this article briefly describes these features. Where in-depth articles are available, links to those tutorials and overviews are provided. You can explore these features in your environment using the dotnet try global tool:

  1. Install the dotnet-try global tool.
  2. Clone the dotnet/try-samples repository.
  3. Set the current directory to the csharp8 subdirectory for the try-samples repository.
  4. Run dotnet try.

Readonly members

You can apply the readonly modifier to any member of a struct. It indicates that the member does not modify state. It's more granular than applying the readonly modifier to a struct declaration. Consider the following mutable struct:

public struct Point
{
    public double X { get; set; }
    public double Y { get; set; }
    public double Distance => Math.Sqrt(X * X + Y * Y);

    public override string ToString() =>
        $"({X}, {Y}) is {Distance} from the origin";
}

Like most structs, the ToString() method does not modify state. You could indicate that by adding the readonly modifier to the declaration of ToString():

public readonly override string ToString() =>
    $"({X}, {Y}) is {Distance} from the origin";

The preceding change generates a compiler warning, because ToString accesses the Distance property, which is not marked readonly:

warning CS8656: Call to non-readonly member 'Point.Distance.get' from a 'readonly' member results in an implicit copy of 'this'

The compiler warns you when it needs to create a defensive copy. The Distance property does not change state, so you can fix this warning by adding the readonly modifier to the declaration:

public readonly double Distance => Math.Sqrt(X * X + Y * Y);

Notice that the readonly modifier is necessary on a read only property. The compiler doesn't assume get accessors do not modify state; you must declare readonly explicitly. The compiler does enforce the rule that readonly members do not modify state. The following method will not compile unless you remove the readonly modifier:

public readonly void Translate(int xOffset, int yOffset)
{
    X += xOffset;
    Y += yOffset;
}

This feature lets you specify your design intent so the compiler can enforce it, and make optimizations based on that intent.

Default interface methods

You can now add members to interfaces and provide an implementation for those members. This language feature enables API authors to add methods to an interface in later versions without breaking source or binary compatibility with existing implementations of that interface. Existing implementations inherit the default implementation. This feature also enables C# to interoperate with APIs that target Android or Swift, which support similar features. Default interface methods also enable scenarios similar to a "traits" language feature.

Default interface methods affects many scenarios and language elements. Our first tutorial covers updating an interface with default implementations. Other tutorials and reference updates are coming in time for general release.

More patterns in more places

Pattern matching gives tools to provide shape-dependent functionality across related but different kinds of data. C# 7.0 introduced syntax for type patterns and constant patterns by using the is expression and the switch statement. These features represented the first tentative steps toward supporting programming paradigms where data and functionality live apart. As the industry moves toward more microservices and other cloud-based architectures, other language tools are needed.

C# 8.0 expands this vocabulary so you can use more pattern expressions in more places in your code. Consider these features when your data and functionality are separate. Consider pattern matching when your algorithms depend on a fact other than the runtime type of an object. These techniques provide another way to express designs.

In addition to new patterns in new places, C# 8.0 adds recursive patterns. The result of any pattern expression is an expression. A recursive pattern is simply a pattern expression applied to the output of another pattern expression.

switch expressions

Often, a switch statement produces a value in each of its case blocks. Switch expressions enable you to use more concise expression syntax. There are fewer repetitive case and break keywords, and fewer curly braces. As an example, consider the following enum that lists the colors of the rainbow:

public enum Rainbow
{
    Red,
    Orange,
    Yellow,
    Green,
    Blue,
    Indigo,
    Violet
}

If your application defined an RGBColor type that is constructed from the R, G and B components, you could convert a Rainbow value to its RGB values using the following method containing a switch expression:

public static RGBColor FromRainbow(Rainbow colorBand) =>
    colorBand switch
    {
        Rainbow.Red    => new RGBColor(0xFF, 0x00, 0x00),
        Rainbow.Orange => new RGBColor(0xFF, 0x7F, 0x00),
        Rainbow.Yellow => new RGBColor(0xFF, 0xFF, 0x00),
        Rainbow.Green  => new RGBColor(0x00, 0xFF, 0x00),
        Rainbow.Blue   => new RGBColor(0x00, 0x00, 0xFF),
        Rainbow.Indigo => new RGBColor(0x4B, 0x00, 0x82),
        Rainbow.Violet => new RGBColor(0x94, 0x00, 0xD3),
        _              => throw new ArgumentException(message: "invalid enum value", paramName: nameof(colorBand)),
    };

There are several syntax improvements here:

  • The variable comes before the switch keyword. The different order makes it visually easy to distinguish the switch expression from the switch statement.
  • The case and : elements are replaced with =>. It's more concise and intuitive.
  • The default case is replaced with a _ discard.
  • The bodies are expressions, not statements.

Contrast that with the equivalent code using the classic switch statement:

public static RGBColor FromRainbowClassic(Rainbow colorBand)
{
    switch (colorBand)
    {
        case Rainbow.Red:
            return new RGBColor(0xFF, 0x00, 0x00);
        case Rainbow.Orange:
            return new RGBColor(0xFF, 0x7F, 0x00);
        case Rainbow.Yellow:
            return new RGBColor(0xFF, 0xFF, 0x00);
        case Rainbow.Green:
            return new RGBColor(0x00, 0xFF, 0x00);
        case Rainbow.Blue:
            return new RGBColor(0x00, 0x00, 0xFF);
        case Rainbow.Indigo:
            return new RGBColor(0x4B, 0x00, 0x82);
        case Rainbow.Violet:
            return new RGBColor(0x94, 0x00, 0xD3);
        default:
            throw new ArgumentException(message: "invalid enum value", paramName: nameof(colorBand));
    };
}

Property patterns

The property pattern enables you to match on properties of the object examined. Consider an eCommerce site that must compute sales tax based on the buyer's address. That computation is not a core responsibility of an Address class. It will change over time, likely more often than address format changes. The amount of sales tax depends on the State property of the address. The following method uses the property pattern to compute the sales tax from the address and the price:

public static decimal ComputeSalesTax(Address location, decimal salePrice) =>
    location switch
    {
        { State: "WA" } => salePrice * 0.06M,
        { State: "MN" } => salePrice * 0.75M,
        { State: "MI" } => salePrice * 0.05M,
        // other cases removed for brevity...
        _ => 0M
    };

Pattern matching creates a concise syntax for expressing this algorithm.

Tuple patterns

Some algorithms depend on multiple inputs. Tuple patterns allow you to switch based on multiple values expressed as a tuple. The following code shows a switch expression for the game rock, paper, scissors:

public static string RockPaperScissors(string first, string second)
    => (first, second) switch
    {
        ("rock", "paper") => "rock is covered by paper. Paper wins.",
        ("rock", "scissors") => "rock breaks scissors. Rock wins.",
        ("paper", "rock") => "paper covers rock. Paper wins.",
        ("paper", "scissors") => "paper is cut by scissors. Scissors wins.",
        ("scissors", "rock") => "scissors is broken by rock. Rock wins.",
        ("scissors", "paper") => "scissors cuts paper. Scissors wins.",
        (_, _) => "tie"
    };

The messages indicate the winner. The discard case represents the three combinations for ties, or other text inputs.

Positional patterns

Some types include a Deconstruct method that deconstructs its properties into discrete variables. When a Deconstruct method is accessible, you can use positional patterns to inspect properties of the object and use those properties for a pattern. Consider the following Point class that includes a Deconstruct method to create discrete variables for X and Y:

public class Point
{
    public int X { get; }
    public int Y { get; }

    public Point(int x, int y) => (X, Y) = (x, y);

    public void Deconstruct(out int x, out int y) =>
        (x, y) = (X, Y);
}

Additionally, consider the following enum that represents various positions of a quadrant:

public enum Quadrant
{
    Unknown,
    Origin,
    One,
    Two,
    Three,
    Four,
    OnBorder
}

The following method uses the positional pattern to extract the values of x and y. Then, it uses a when clause to determine the Quadrant of the point:

static Quadrant GetQuadrant(Point point) => point switch
{
    (0, 0) => Quadrant.Origin,
    var (x, y) when x > 0 && y > 0 => Quadrant.One,
    var (x, y) when x < 0 && y > 0 => Quadrant.Two,
    var (x, y) when x < 0 && y < 0 => Quadrant.Three,
    var (x, y) when x > 0 && y < 0 => Quadrant.Four,
    var (_, _) => Quadrant.OnBorder,
    _ => Quadrant.Unknown
};

The discard pattern in the preceding switch matches when either x or y is 0, but not both. A switch expression must either produce a value or throw an exception. If none of the cases match, the switch expression throws an exception. The compiler generates a warning for you if you do not cover all possible cases in your switch expression.

You can explore pattern matching techniques in this advanced tutorial on pattern matching.

using declarations

A using declaration is a variable declaration preceded by the using keyword. It tells the compiler that the variable being declared should be disposed at the end of the enclosing scope. For example, consider the following code that writes a text file:

static void WriteLinesToFile(IEnumerable<string> lines)
{
    using var file = new System.IO.StreamWriter("WriteLines2.txt");
    foreach (string line in lines)
    {
        if (!line.Contains("Second"))
        {
            file.WriteLine(line);
        }
    }
// file is disposed here
}

In the preceding example, the file is disposed when the closing brace for the method is reached. That's the end of the scope in which file is declared. The preceding code is equivalent to the following code that uses the classic using statement:

static void WriteLinesToFile(IEnumerable<string> lines)
{
    using (var file = new System.IO.StreamWriter("WriteLines2.txt"))
    {
        foreach (string line in lines)
        {
            if (!line.Contains("Second"))
            {
                file.WriteLine(line);
            }
        }
    } // file is disposed here
}

In the preceding example, the file is disposed when the closing brace associated with the using statement is reached.

In both cases, the compiler generates the call to Dispose(). The compiler generates an error if the expression in the using statement is not disposable.

Static local functions

You can now add the static modifier to local functions to ensure that local function doesn't capture (reference) any variables from the enclosing scope. Doing so generates CS8421, "A static local function can't contain a reference to <variable>."

Consider the following code. The local function LocalFunction accesses the variable y, declared in the enclosing scope (the method M). Therefore, LocalFunction can't be declared with the static modifier:

int M()
{
    int y;
    LocalFunction();
    return y;

    void LocalFunction() => y = 0;
}

The following code contains a static local function. It can be static because it doesn't access any variables in the enclosing scope:

int M()
{
    int y = 5;
    int x = 7;
    return Add(x, y);

    static int Add(int left, int right) => left + right;
}

Disposable ref structs

A struct declared with the ref modifier may not implement any interfaces and so cannot implement IDisposable. Therefore, to enable a ref struct to be disposed, it must have an accessible void Dispose() method. This also applies to readonly ref struct declarations.

Nullable reference types

Inside a nullable annotation context, any variable of a reference type is considered to be a nonnullable reference type. If you want to indicate that a variable may be null, you must append the type name with the ? to declare the variable as a nullable reference type.

For nonnullable reference types, the compiler uses flow analysis to ensure that local variables are initialized to a non-null value when declared. Fields must be initialized during construction. The compiler generates a warning if the variable is not set by a call to any of the available constructors or by an initializer. Furthermore, nonnullable reference types can't be assigned a value that could be null.

Nullable reference types aren't checked to ensure they aren't assigned or initialized to null. However, the compiler uses flow analysis to ensure that any variable of a nullable reference type is checked against null before it's accessed or assigned to a nonnullable reference type.

You can learn more about the feature in the overview of nullable reference types. Try it yourself in a new application in this nullable reference types tutorial. Learn about the steps to migrate an existing codebase to make use of nullable reference types in the migrating an application to use nullable reference types tutorial.

Asynchronous streams

Starting with C# 8.0, you can create and consume streams asynchronously. A method that returns an asynchronous stream has three properties:

  1. It's declared with the async modifier.
  2. It returns an IAsyncEnumerable<T>.
  3. The method contains yield return statements to return successive elements in the asynchronous stream.

Consuming an asynchronous stream requires you to add the await keyword before the foreach keyword when you enumerate the elements of the stream. Adding the await keyword requires the method that enumerates the asynchronous stream to be declared with the async modifier and to return a type allowed for an async method. Typically that means returning a Task or Task<TResult>. It can also be a ValueTask or ValueTask<TResult>. A method can both consume and produce an asynchronous stream, which means it would return an IAsyncEnumerable<T>. The following code generates a sequence from 0 to 19, waiting 100 ms between generating each number:

public static async System.Collections.Generic.IAsyncEnumerable<int> GenerateSequence()
{
    for (int i = 0; i < 20; i++)
    {
        await Task.Delay(100);
        yield return i;
    }
}

You would enumerate the sequence using the await foreach statement:

await foreach (var number in GenerateSequence())
{
    Console.WriteLine(number);
}

You can try asynchronous streams yourself in our tutorial on creating and consuming async streams.

Indices and ranges

Indices and ranges provide a succinct syntax for accessing single elements or ranges in a sequence.

This language support relies on two new types, and two new operators:

  • System.Index represents an index into a sequence.
  • The index from end operator ^, which specifies that an index is relative to the end of the sequence.
  • System.Range represents a sub range of a sequence.
  • The range operator .., which specifies the start and end of a range as its operands.

Let's start with the rules for indexes. Consider an array sequence. The 0 index is the same as sequence[0]. The ^0 index is the same as sequence[sequence.Length]. Note that sequence[^0] does throw an exception, just as sequence[sequence.Length] does. For any number n, the index ^n is the same as sequence.Length - n.

A range specifies the start and end of a range. The start of the range is inclusive, but the end of the range is exclusive, meaning the start is included in the range but the end is not included in the range. The range [0..^0] represents the entire range, just as [0..sequence.Length] represents the entire range.

Let's look at a few examples. Consider the following array, annotated with its index from the start and from the end:

var words = new string[]
{
                // index from start    index from end
    "The",      // 0                   ^9
    "quick",    // 1                   ^8
    "brown",    // 2                   ^7
    "fox",      // 3                   ^6
    "jumped",   // 4                   ^5
    "over",     // 5                   ^4
    "the",      // 6                   ^3
    "lazy",     // 7                   ^2
    "dog"       // 8                   ^1
};              // 9 (or words.Length) ^0

You can retrieve the last word with the ^1 index:

Console.WriteLine($"The last word is {words[^1]}");
// writes "dog"

The following code creates a subrange with the words "quick", "brown", and "fox". It includes words[1] through words[3]. The element words[4] is not in the range.

var quickBrownFox = words[1..4];

The following code creates a subrange with "lazy" and "dog". It includes words[^2] and words[^1]. The end index words[^0] is not included:

var lazyDog = words[^2..^0];

The following examples create ranges that are open ended for the start, end, or both:

var allWords = words[..]; // contains "The" through "dog".
var firstPhrase = words[..4]; // contains "The" through "fox"
var lastPhrase = words[6..]; // contains "the", "lazy" and "dog"

You can also declare ranges as variables:

Range phrase = 1..4;

The range can then be used inside the [ and ] characters:

var text = words[phrase];

Not only arrays support indices and ranges. You also can use indices and ranges with string, Span<T>, or ReadOnlySpan<T>. For more information, see Type support for indices and ranges.

You can explore more about indices and ranges in the tutorial on indices and ranges.

Null-coalescing assignment

C# 8.0 introduces the null-coalescing assignment operator ??=. You can use the ??= operator to assign the value of its right-hand operand to its left-hand operand only if the left-hand operand evaluates to null.

List<int> numbers = null;
int? i = null;

numbers ??= new List<int>();
numbers.Add(i ??= 17);
numbers.Add(i ??= 20);

Console.WriteLine(string.Join(" ", numbers));  // output: 17 17
Console.WriteLine(i);  // output: 17

For more information, see the ?? and ??= operators article.

Unmanaged constructed types

In C# 7.3 and earlier, a constructed type (a type that includes at least one type argument) cannot be an unmanaged type. Starting with C# 8.0, a constructed value type is unmanaged if it contains fields of unmanaged types only.

For example, given the following definition of the generic Coords<T> type:

public struct Coords<T>
{
    public T X;
    public T Y;
}

the Coords<int> type is an unmanaged type in C# 8.0 and later. Like for any unmanaged type, you can create a pointer to a variable of this type or allocate a block of memory on the stack for instances of this type:

Span<Coords<int>> coordinates = stackalloc[]
{
    new Coords<int> { X = 0, Y = 0 },
    new Coords<int> { X = 0, Y = 3 },
    new Coords<int> { X = 4, Y = 0 }
};

For more information, see Unmanaged types.

stackalloc in nested expressions

Starting with C# 8.0, if the result of a stackalloc expression is of the System.Span<T> or System.ReadOnlySpan<T> type, you can use the stackalloc expression in other expressions:

Span<int> numbers = stackalloc[] { 1, 2, 3, 4, 5, 6 };
var ind = numbers.IndexOfAny(stackalloc[] { 2, 4, 6 ,8 });
Console.WriteLine(ind);  // output: 1

Enhancement of interpolated verbatim strings

Order of the $ and @ tokens in interpolated verbatim strings can be any: both $@"..." and @$"..." are valid interpolated verbatim strings. In earlier C# versions, the $ token must appear before the @ token.