संपादित करें

इसके माध्यम से साझा किया गया


13 Statements

13.1 General

C# provides a variety of statements.

Note: Most of these statements will be familiar to developers who have programmed in C and C++. end note

statement
    : labeled_statement
    | declaration_statement
    | embedded_statement
    ;

embedded_statement
    : block
    | empty_statement
    | expression_statement
    | selection_statement
    | iteration_statement
    | jump_statement
    | try_statement
    | checked_statement
    | unchecked_statement
    | lock_statement
    | using_statement
    | yield_statement
    | unsafe_statement   // unsafe code support
    | fixed_statement    // unsafe code support
    ;

unsafe_statement (§23.2) and fixed_statement (§23.7) are only available in unsafe code (§23).

The embedded_statement nonterminal is used for statements that appear within other statements. The use of embedded_statement rather than statement excludes the use of declaration statements and labeled statements in these contexts.

Example: The code

void F(bool b)
{
   if (b)
      int i = 44;
}

results in a compile-time error because an if statement requires an embedded_statement rather than a statement for its if branch. If this code were permitted, then the variable i would be declared, but it could never be used. Note, however, that by placing i’s declaration in a block, the example is valid.

end example

13.2 End points and reachability

Every statement has an end point. In intuitive terms, the end point of a statement is the location that immediately follows the statement. The execution rules for composite statements (statements that contain embedded statements) specify the action that is taken when control reaches the end point of an embedded statement.

Example: When control reaches the end point of a statement in a block, control is transferred to the next statement in the block. end example

If a statement can possibly be reached by execution, the statement is said to be reachable. Conversely, if there is no possibility that a statement will be executed, the statement is said to be unreachable.

Example: In the following code

void F()
{
    Console.WriteLine("reachable");
    goto Label;
    Console.WriteLine("unreachable");
  Label:
    Console.WriteLine("reachable");
}

the second invocation of Console.WriteLine is unreachable because there is no possibility that the statement will be executed.

end example

A warning is reported if a statement other than throw_statement, block, or empty_statement is unreachable. It is specifically not an error for a statement to be unreachable.

Note: To determine whether a particular statement or end point is reachable, the compiler performs flow analysis according to the reachability rules defined for each statement. The flow analysis takes into account the values of constant expressions (§12.23) that control the behavior of statements, but the possible values of non-constant expressions are not considered. In other words, for purposes of control flow analysis, a non-constant expression of a given type is considered to have any possible value of that type.

In the example

void F()
{
    const int i = 1;
    if (i == 2)
        Console.WriteLine("unreachable");
}

the Boolean expression of the if statement is a constant expression because both operands of the == operator are constants. As the constant expression is evaluated at compile-time, producing the value false, the Console.WriteLine invocation is considered unreachable. However, if i is changed to be a local variable

void F()
{
    int i = 1;
    if (i == 2)
        Console.WriteLine("reachable");
}

the Console.WriteLine invocation is considered reachable, even though, in reality, it will never be executed.

end note

The block of a function member or an anonymous function is always considered reachable. By successively evaluating the reachability rules of each statement in a block, the reachability of any given statement can be determined.

Example: In the following code

void F(int x)
{
    Console.WriteLine("start");
    if (x < 0)
        Console.WriteLine("negative");
}

the reachability of the second Console.WriteLine is determined as follows:

  • The first Console.WriteLine expression statement is reachable because the block of the F method is reachable (§13.3).
  • The end point of the first Console.WriteLine expression statement is reachable because that statement is reachable (§13.7 and §13.3).
  • The if statement is reachable because the end point of the first Console.WriteLine expression statement is reachable (§13.7 and §13.3).
  • The second Console.WriteLine expression statement is reachable because the Boolean expression of the if statement does not have the constant value false.

end example

There are two situations in which it is a compile-time error for the end point of a statement to be reachable:

  • Because the switch statement does not permit a switch section to “fall through” to the next switch section, it is a compile-time error for the end point of the statement list of a switch section to be reachable. If this error occurs, it is typically an indication that a break statement is missing.

  • It is a compile-time error for the end point of the block of a function member or an anonymous function that computes a value to be reachable. If this error occurs, it typically is an indication that a return statement is missing (§13.10.5).

13.3 Blocks

13.3.1 General

A block permits multiple statements to be written in contexts where a single statement is allowed.

block
    : '{' statement_list? '}'
    ;

A block consists of an optional statement_list (§13.3.2), enclosed in braces. If the statement list is omitted, the block is said to be empty.

A block may contain declaration statements (§13.6). The scope of a local variable or constant declared in a block is the block.

A block is executed as follows:

  • If the block is empty, control is transferred to the end point of the block.
  • If the block is not empty, control is transferred to the statement list. When and if control reaches the end point of the statement list, control is transferred to the end point of the block.

The statement list of a block is reachable if the block itself is reachable.

The end point of a block is reachable if the block is empty or if the end point of the statement list is reachable.

A block that contains one or more yield statements (§13.15) is called an iterator block. Iterator blocks are used to implement function members as iterators (§15.14). Some additional restrictions apply to iterator blocks:

  • It is a compile-time error for a return statement to appear in an iterator block (but yield return statements are permitted).
  • It is a compile-time error for an iterator block to contain an unsafe context (§23.2). An iterator block always defines a safe context, even when its declaration is nested in an unsafe context.

13.3.2 Statement lists

A statement list consists of one or more statements written in sequence. Statement lists occur in blocks (§13.3) and in switch_blocks (§13.8.3).

statement_list
    : statement+
    ;

A statement list is executed by transferring control to the first statement. When and if control reaches the end point of a statement, control is transferred to the next statement. When and if control reaches the end point of the last statement, control is transferred to the end point of the statement list.

A statement in a statement list is reachable if at least one of the following is true:

  • The statement is the first statement and the statement list itself is reachable.
  • The end point of the preceding statement is reachable.
  • The statement is a labeled statement and the label is referenced by a reachable goto statement.

The end point of a statement list is reachable if the end point of the last statement in the list is reachable.

13.4 The empty statement

An empty_statement does nothing.

empty_statement
    : ';'
    ;

An empty statement is used when there are no operations to perform in a context where a statement is required.

Execution of an empty statement simply transfers control to the end point of the statement. Thus, the end point of an empty statement is reachable if the empty statement is reachable.

Example: An empty statement can be used when writing a while statement with a null body:

bool ProcessMessage() {...}
void ProcessMessages()
{
    while (ProcessMessage())
        ;
}

Also, an empty statement can be used to declare a label just before the closing “}” of a block:

void F(bool done)
{
    ...
    if (done)
    {
        goto exit;
    }
    ...
  exit:
    ;
}

end example

13.5 Labeled statements

A labeled_statement permits a statement to be prefixed by a label. Labeled statements are permitted in blocks, but are not permitted as embedded statements.

labeled_statement
    : identifier ':' statement
    ;

A labeled statement declares a label with the name given by the identifier. The scope of a label is the whole block in which the label is declared, including any nested blocks. It is a compile-time error for two labels with the same name to have overlapping scopes.

A label can be referenced from goto statements (§13.10.4) within the scope of the label.

Note: This means that goto statements can transfer control within blocks and out of blocks, but never into blocks. end note

Labels have their own declaration space and do not interfere with other identifiers.

Example: The example

int F(int x)
{
    if (x >= 0)
    {
        goto x;
    }
    x = -x;
  x:
    return x;
}

is valid and uses the name x as both a parameter and a label.

end example

Execution of a labeled statement corresponds exactly to execution of the statement following the label.

In addition to the reachability provided by normal flow of control, a labeled statement is reachable if the label is referenced by a reachable goto statement, unless the goto statement is inside the try block or a catch block of a try_statement that includes a finally block whose end point is unreachable, and the labeled statement is outside the try_statement.

13.6 Declaration statements

13.6.1 General

A declaration_statement declares one or more local variables, one or more local constants, or a local function. Declaration statements are permitted in blocks and switch blocks, but are not permitted as embedded statements.

declaration_statement
    : local_variable_declaration ';'
    | local_constant_declaration ';'
    | local_function_declaration
    ;

A local variable is declared using a local_variable_declaration (§13.6.2). A local constant is declared using a local_constant_declaration (§13.6.3). A local function is declared using a local_function_declaration (§13.6.4).

The declared names are introduced into the nearest enclosing declaration space (§7.3).

13.6.2 Local variable declarations

13.6.2.1 General

A local_variable_declaration declares one or more local variables.

local_variable_declaration
    : implicitly_typed_local_variable_declaration
    | explicitly_typed_local_variable_declaration
    | ref_local_variable_declaration
    ;

Local variable declarations fall into one of the three categories: implicitly typed, explicitly typed, and ref local.

Implicitly typed declarations contain the contextual keyword (§6.4.4) var resulting in a syntactic ambiguity between the three categories which is resolved as follows:

  • If there is no type named var in scope and the input matches implicitly_typed_local_variable_declaration then it is chosen;
  • Otherwise if a type named var is in scope then implicitly_typed_local_variable_declaration is not considered as a possible match.

Within a local_variable_declaration each variable is introduced by a declarator, which is one of implicitly_typed_local_variable_declarator, explicitly_typed_local_variable_declarator or ref_local_variable_declarator for implicitly typed, explicitly typed and ref local variables respectively. The declarator defines the name (identifier) and initial value, if any, of the introduced variable.

If there are multiple declarators in a declaration then they are processed, including any initializing expressions, in order left to right (§9.4.4.5).

Note: For a local_variable_declaration not occuring as a for_initializer (§13.9.4) or resource_acquisition (§13.14) this left to right order is equivalent to each declarator being within a separate local_variable_declaration. For example:

void F()
{
    int x = 1, y, z = x * 2;
}

is equivalent to:

void F()
{
    int x = 1;
    int y;
    int z = x * 2;
}

end note

The value of a local variable is obtained in an expression using a simple_name (§12.8.4). A local variable shall be definitely assigned (§9.4) at each location where its value is obtained. Each local variable introduced by a local_variable_declaration is initially unassigned (§9.4.3). If a declarator has an initializing expression then the introduced local variable is classified as assigned at the end of the declarator (§9.4.4.5).

The scope of a local variable introduced by a local_variable_declaration is defined as follows (§7.7):

  • If the declaration occurs as a for_initializer then the scope is the for_initializer, for_condition, for_iterator, and embedded_statement (§13.9.4);
  • If the declaration occurs as a resource_acquisition then the scope is the outermost block of the semantically equivalent expansion of the using_statement (§13.14);
  • Otherwise the scope is the block in which the declaration occurs.

It is an error to refer to a local variable by name in a textual position that precedes its declarator, or within any initializing expression within its declarator. Within the scope of a local variable, it is a compile-time error to declare another local variable, local function or constant with the same name.

The ref-safe-context (§9.7.2) of a ref local variable is the ref-safe-context of its initializing variable_reference. The ref-safe-context of non-ref local variables is declaration-block.

13.6.2.2 Implicitly typed local variable declarations

implicitly_typed_local_variable_declaration
    : 'var' implicitly_typed_local_variable_declarator
    | ref_kind 'var' ref_local_variable_declarator
    ;

implicitly_typed_local_variable_declarator
    : identifier '=' expression
    ;

An implicity_typed_local_variable_declaration introduces a single local variable, identifier. The expression or variable_reference shall have a compile-time type, T. The first alternative declares a variable with an initial value of expression; its type is T? when T is a non-nullable reference type, otherwise its type is T. The second alternative declares a ref variable with an initial value of ref variable_reference; its type is ref T? when T is a non-nullable reference type, otherwise its type is ref T.

Example:

var i = 5;
var s = "Hello";
var d = 1.0;
var numbers = new int[] {1, 2, 3};
var orders = new Dictionary<int,Order>();
ref var j = ref i;
ref readonly var k = ref i;

The implicitly typed local variable declarations above are precisely equivalent to the following explicitly typed declarations:

int i = 5;
string s = "Hello";
double d = 1.0;
int[] numbers = new int[] {1, 2, 3};
Dictionary<int,Order> orders = new Dictionary<int,Order>();
ref int j = ref i;
ref readonly int k = ref i;

The following are incorrect implicitly typed local variable declarations:

var x;                  // Error, no initializer to infer type from
var y = {1, 2, 3};      // Error, array initializer not permitted
var z = null;           // Error, null does not have a type
var u = x => x + 1;     // Error, anonymous functions do not have a type
var v = v++;            // Error, initializer cannot refer to v itself

end example

13.6.2.3 Explicitly typed local variable declarations

explicitly_typed_local_variable_declaration
    : type explicitly_typed_local_variable_declarators
    ;

explicitly_typed_local_variable_declarators
    : explicitly_typed_local_variable_declarator
      (',' explicitly_typed_local_variable_declarator)*
    ;

explicitly_typed_local_variable_declarator
    : identifier ('=' local_variable_initializer)?
    ;

local_variable_initializer
    : expression
    | array_initializer
    ;

An explicity_typed_local_variable_declaration introduces one or more local variables with the specified type.

If a local_variable_initializer is present then its type shall be appropriate according to the rules of simple assignment (§12.21.2) or array initialization (§17.7) and its value is assigned as the initial value of the variable.

13.6.2.4 Ref local variable declarations

ref_local_variable_declaration
    : ref_kind type ref_local_variable_declarators
    ;

ref_local_variable_declarators
    : ref_local_variable_declarator (',' ref_local_variable_declarator)*
    ;

ref_local_variable_declarator
    : identifier '=' 'ref' variable_reference
    ;

The initializing variable_reference shall have type type and meet the same requirements as for a ref assignment (§12.21.3).

If ref_kind is ref readonly, the identifier(s) being declared are references to variables that are treated as read-only. Otherwise, if ref_kind is ref, the identifier(s) being declared are references to variables that shall be writable.

It is a compile-time error to declare a ref local variable, or a variable of a ref struct type, within a method declared with the method_modifier async, or within an iterator (§15.14).

13.6.3 Local constant declarations

A local_constant_declaration declares one or more local constants.

local_constant_declaration
    : 'const' type constant_declarators
    ;

constant_declarators
    : constant_declarator (',' constant_declarator)*
    ;

constant_declarator
    : identifier '=' constant_expression
    ;

The type of a local_constant_declaration specifies the type of the constants introduced by the declaration. The type is followed by a list of constant_declarators, each of which introduces a new constant. A constant_declarator consists of an identifier that names the constant, followed by an “=” token, followed by a constant_expression (§12.23) that gives the value of the constant.

The type and constant_expression of a local constant declaration shall follow the same rules as those of a constant member declaration (§15.4).

The value of a local constant is obtained in an expression using a simple_name (§12.8.4).

The scope of a local constant is the block in which the declaration occurs. It is an error to refer to a local constant in a textual position that precedes the end of its constant_declarator.

A local constant declaration that declares multiple constants is equivalent to multiple declarations of single constants with the same type.

13.6.4 Local function declarations

A local_function_declaration declares a local function.

local_function_declaration
    : local_function_modifier* return_type local_function_header
      local_function_body
    | ref_local_function_modifier* ref_kind ref_return_type
      local_function_header ref_local_function_body
    ;

local_function_header
    : identifier '(' parameter_list? ')'
    | identifier type_parameter_list '(' parameter_list? ')'
      type_parameter_constraints_clause*
    ;

local_function_modifier
    : ref_local_function_modifier
    | 'async'
    ;

ref_local_function_modifier
    : 'static'
    | unsafe_modifier   // unsafe code support
    ;

local_function_body
    : block
    | '=>' null_conditional_invocation_expression ';'
    | '=>' expression ';'
    ;

ref_local_function_body
    : block
    | '=>' 'ref' variable_reference ';'
    ;

Grammar note: When recognising a local_function_body if both the null_conditional_invocation_expression and expression alternatives are applicable then the former shall be chosen. (§15.6.1)

Example: There are two common use cases for local functions: iterator methods and async methods. In iterator methods, any exceptions are observed only when calling code that enumerates the returned sequence. In async methods, any exceptions are only observed when the returned Task is awaited. The following example demonstrates separating parameter validation from the iterator implementation using a local function:

public static IEnumerable<char> AlphabetSubset(char start, char end)
{
    if (start < 'a' || start > 'z')
    {
        throw new ArgumentOutOfRangeException(paramName: nameof(start),
            message: "start must be a letter");
    }
    if (end < 'a' || end > 'z')
    {
        throw new ArgumentOutOfRangeException(paramName: nameof(end),
            message: "end must be a letter");
    }
    if (end <= start)
    {
        throw new ArgumentException(
            $"{nameof(end)} must be greater than {nameof(start)}");
    }
    return AlphabetSubsetImplementation();

    IEnumerable<char> AlphabetSubsetImplementation()
    {
        for (var c = start; c < end; c++)
        {
            yield return c;
        }
    }
}

end example

Unless specified otherwise below, the semantics of all grammar elements is the same as for method_declaration (§15.6.1), read in the context of a local function instead of a method.

The identifier of a local_function_declaration shall be unique in its declared block scope, including any enclosing local variable declaration spaces. One consequence of this is that overloaded local_function_declarations are not allowed.

A local_function_declaration may include one async (§15.15) modifier and one unsafe (§23.1) modifier. If the declaration includes the async modifier then the return type shall be void or a «TaskType» type (§15.15.1). If the declaration includes the static modifier, the function is a static local function; otherwise, it is a non-static local function. It is a compile-time error for type_parameter_list or parameter_list to contain attributes. If the local function is declared in an unsafe context (§23.2), the local function may include unsafe code, even if the local function declaration doesn’t include the unsafe modifier.

A local function is declared at block scope. A non-static local function may capture variables from the enclosing scope while a static local function shall not (so it has no access to enclosing locals, parameters, non-static local functions, or this). It is a compile-time error if a captured variable is read by the body of a non-static local function but is not definitely assigned before each call to the function. The compiler shall determine which variables are definitely assigned on return (§9.4.4.33).

When the type of this is a struct type, it is a compile-time error for the body of a local function to access this. This is true whether the access is explicit (as in this.x) or implicit (as in x where x is an instance member of the struct). This rule only prohibits such access and does not affect whether member lookup results in a member of the struct.

It is a compile-time error for the body of the local function to contain a goto statement, a break statement, or a continue statement whose target is outside the body of the local function.

Note: the above rules for this and goto mirror the rules for anonymous functions in §12.19.3. end note

A local function may be called from a lexical point prior to its declaration. However, it is a compile-time error for the function to be declared lexically prior to the declaration of a variable used in the local function (§7.7).

It is a compile-time error for a local function to declare a parameter, type parameter or local variable with the same name as one declared in any enclosing local variable declaration space.

Local function bodies are always reachable. The endpoint of a local function declaration is reachable if the beginning point of the local function declaration is reachable.

Example: In the following example, the body of L is reachable even though the beginning point of L is not reachable. Because the beginning point of L isn’t reachable, the statement following the endpoint of L is not reachable:

class C
{
    int M()
    {
        L();
        return 1;

        // Beginning of L is not reachable
        int L()
        {
            // The body of L is reachable
            return 2;
        }
        // Not reachable, because beginning point of L is not reachable
        return 3;
    }
}

In other words, the location of a local function declaration doesn’t affect the reachability of any statements in the containing function. end example

If the type of the argument to a local function is dynamic, the function to be called shall be resolved at compile time, not runtime.

A local function shall not be used in an expression tree.

A static local function

  • May reference static members, type parameters, constant definitions and static local functions from the enclosing scope.
  • Shall not reference this or base nor instance members from an implicit this reference, nor local variables, parameters, or non-static local functions from the enclosing scope. However, all these are permitted in a nameof() expression.

13.7 Expression statements

An expression_statement evaluates a given expression. The value computed by the expression, if any, is discarded.

expression_statement
    : statement_expression ';'
    ;

statement_expression
    : null_conditional_invocation_expression
    | invocation_expression
    | object_creation_expression
    | assignment
    | post_increment_expression
    | post_decrement_expression
    | pre_increment_expression
    | pre_decrement_expression
    | await_expression
    ;

Not all expressions are permitted as statements.

Note: In particular, expressions such as x + y and x == 1, that merely compute a value (which will be discarded), are not permitted as statements. end note

Execution of an expression_statement evaluates the contained expression and then transfers control to the end point of the expression_statement. The end point of an expression_statement is reachable if that expression_statement is reachable.

13.8 Selection statements

13.8.1 General

Selection statements select one of a number of possible statements for execution based on the value of some expression.

selection_statement
    : if_statement
    | switch_statement
    ;

13.8.2 The if statement

The if statement selects a statement for execution based on the value of a Boolean expression.

if_statement
    : 'if' '(' boolean_expression ')' embedded_statement
    | 'if' '(' boolean_expression ')' embedded_statement
      'else' embedded_statement
    ;

An else part is associated with the lexically nearest preceding if that is allowed by the syntax.

Example: Thus, an if statement of the form

if (x) if (y) F(); else G();

is equivalent to

if (x)
{
    if (y)
    {
        F();
    }
    else
    {
        G();
    }
}

end example

An if statement is executed as follows:

  • The boolean_expression (§12.24) is evaluated.
  • If the Boolean expression yields true, control is transferred to the first embedded statement. When and if control reaches the end point of that statement, control is transferred to the end point of the if statement.
  • If the Boolean expression yields false and if an else part is present, control is transferred to the second embedded statement. When and if control reaches the end point of that statement, control is transferred to the end point of the if statement.
  • If the Boolean expression yields false and if an else part is not present, control is transferred to the end point of the if statement.

The first embedded statement of an if statement is reachable if the if statement is reachable and the Boolean expression does not have the constant value false.

The second embedded statement of an if statement, if present, is reachable if the if statement is reachable and the Boolean expression does not have the constant value true.

The end point of an if statement is reachable if the end point of at least one of its embedded statements is reachable. In addition, the end point of an if statement with no else part is reachable if the if statement is reachable and the Boolean expression does not have the constant value true.

13.8.3 The switch statement

The switch statement selects for execution a statement list having an associated switch label that corresponds to the value of the switch expression.

switch_statement
    : 'switch' '(' expression ')' switch_block
    ;

switch_block
    : '{' switch_section* '}'
    ;

switch_section
    : switch_label+ statement_list
    ;

switch_label
    : 'case' pattern case_guard?  ':'
    | 'default' ':'
    ;

case_guard
    : 'when' expression
    ;

A switch_statement consists of the keyword switch, followed by a parenthesized expression (called the switch expression), followed by a switch_block. The switch_block consists of zero or more switch_sections, enclosed in braces. Each switch_section consists of one or more switch_labels followed by a statement_list (§13.3.2). Each switch_label containing case has an associated pattern (§11) against which the value of the switch expression is tested. If case_guard is present, its expression shall be implicitly convertible to the type bool and that expression is evaluated as an additional condition for the case to be considered satisfied.

The governing type of a switch statement is established by the switch expression.

  • If the type of the switch expression is sbyte, byte, short, ushort, int, uint, long, ulong, char, bool, string, or an enum_type, or if it is the nullable value type corresponding to one of these types, then that is the governing type of the switch statement.
  • Otherwise, if exactly one user-defined implicit conversion exists from the type of the switch expression to one of the following possible governing types: sbyte, byte, short, ushort, int, uint, long, ulong, char, string, or, a nullable value type corresponding to one of those types, then the converted type is the governing type of the switch statement.
  • Otherwise, the governing type of the switch statement is the type of the switch expression. It is an error if no such type exists.

There can be at most one default label in a switch statement.

It is an error if the pattern of any switch label is not applicable (§11.2.1) to the type of the input expression.

It is an error if the pattern of any switch label is subsumed by (§11.3) the set of patterns of earlier switch labels of the switch statement that do not have a case guard or whose case guard is a constant expression with the value true.

Example:

switch (shape)
{
    case var x:
        break;
    case var _: // error: pattern subsumed, as previous case always matches
        break;
    default:
        break;  // warning: unreachable, all possible values already handled.
}

end example

A switch statement is executed as follows:

  • The switch expression is evaluated and converted to the governing type.
  • Control is transferred according to the value of the converted switch expression:
    • The lexically first pattern in the set of case labels in the same switch statement that matches the value of the switch expression, and for which the guard expression is either absent or evaluates to true, causes control to be transferred to the statement list following the matched case label.
    • Otherwise, if a default label is present, control is transferred to the statement list following the default label.
    • Otherwise, control is transferred to the end point of the switch statement.

Note: The order in which patterns are matched at runtime is not defined. A compiler is permitted (but not required) to match patterns out of order, and to reuse the results of already matched patterns to compute the result of matching of other patterns. Nevertheless, the compiler is required to determine the lexically first pattern that matches the expression and for which the guard clause is either absent or evaluates to true. end note

If the end point of the statement list of a switch section is reachable, a compile-time error occurs. This is known as the “no fall through” rule.

Example: The example

switch (i)
{
    case 0:
        CaseZero();
        break;
    case 1:
        CaseOne();
        break;
    default:
        CaseOthers();
        break;
}

is valid because no switch section has a reachable end point. Unlike C and C++, execution of a switch section is not permitted to “fall through” to the next switch section, and the example

switch (i)
{
    case 0:
        CaseZero();
    case 1:
        CaseZeroOrOne();
    default:
        CaseAny();
}

results in a compile-time error. When execution of a switch section is to be followed by execution of another switch section, an explicit goto case or goto default statement shall be used:

switch (i)
{
    case 0:
        CaseZero();
        goto case 1;
    case 1:
        CaseZeroOrOne();
        goto default;
    default:
        CaseAny();
        break;
}

end example

Multiple labels are permitted in a switch_section.

Example: The example

switch (i)
{
    case 0:
        CaseZero();
        break;
    case 1:
        CaseOne();
        break;
    case 2:
    default:
        CaseTwo();
        break;
}

is valid. The example does not violate the “no fall through” rule because the labels case 2: and default: are part of the same switch_section.

end example

Note: The “no fall through” rule prevents a common class of bugs that occur in C and C++ when break statements are accidentally omitted. For example, the sections of the switch statement above can be reversed without affecting the behavior of the statement:

switch (i)
{
    default:
        CaseAny();
        break;
    case 1:
        CaseZeroOrOne();
        goto default;
    case 0:
        CaseZero();
        goto case 1;
}

end note

Note: The statement list of a switch section typically ends in a break, goto case, or goto default statement, but any construct that renders the end point of the statement list unreachable is permitted. For example, a while statement controlled by the Boolean expression true is known to never reach its end point. Likewise, a throw or return statement always transfers control elsewhere and never reaches its end point. Thus, the following example is valid:

switch (i)
{
     case 0:
         while (true)
         {
             F();
         }
     case 1:
         throw new ArgumentException();
     case 2:
         return;
}

end note

Example: The governing type of a switch statement can be the type string. For example:

void DoCommand(string command)
{
    switch (command.ToLower())
    {
        case "run":
            DoRun();
            break;
        case "save":
            DoSave();
            break;
        case "quit":
            DoQuit();
            break;
        default:
            InvalidCommand(command);
            break;
    }
}

end example

Note: Like the string equality operators (§12.12.8), the switch statement is case sensitive and will execute a given switch section only if the switch expression string exactly matches a case label constant. end note When the governing type of a switch statement is string or a nullable value type, the value null is permitted as a case label constant.

The statement_lists of a switch_block may contain declaration statements (§13.6). The scope of a local variable or constant declared in a switch block is the switch block.

A switch label is reachable if at least one of the following is true:

  • The switch expression is a constant value and either
    • the label is a case whose pattern would match (§11.2.1) that value, and label’s guard is either absent or not a constant expression with the value false; or
    • it is a default label, and no switch section contains a case label whose pattern would match that value, and whose guard is either absent or a constant expression with the value true.
  • The switch expression is not a constant value and either
    • the label is a case without a guard or with a guard whose value is not the constant false; or
    • it is a default label and
      • the set of patterns appearing among the cases of the switch statement that do not have guards or have guards whose value is the constant true, is not exhaustive (§11.4) for the switch governing type; or
      • the switch governing type is a nullable type and the set of patterns appearing among the cases of the switch statement that do not have guards or have guards whose value is the constant true does not contain a pattern that would match the value null.
  • The switch label is referenced by a reachable goto case or goto default statement.

The statement list of a given switch section is reachable if the switch statement is reachable and the switch section contains a reachable switch label.

The end point of a switch statement is reachable if the switch statement is reachable and at least one of the following is true:

  • The switch statement contains a reachable break statement that exits the switch statement.
  • No default label is present and either
    • The switch expression is a non-constant value, and the set of patterns appearing among the cases of the switch statement that do not have guards or have guards whose value is the constant true, is not exhaustive (§11.4) for the switch governing type.
    • The switch expression is a non-constant value of a nullable type, and no pattern appearing among the cases of the switch statement that do not have guards or have guards whose value is the constant true would match the value null.
    • The switch expression is a constant value and no case label without a guard or whose guard is the constant true would match that value.

Example: The following code shows a succinct use of the when clause:

static object CreateShape(string shapeDescription)
{
   switch (shapeDescription)
   {
        case "circle":
            return new Circle(2);
        …
        case var o when string.IsNullOrWhiteSpace(o):
            return null;
        default:
            return "invalid shape description";
    }
}

The var case matches null, the empty string, or any string that contains only white space. end example

13.9 Iteration statements

13.9.1 General

Iteration statements repeatedly execute an embedded statement.

iteration_statement
    : while_statement
    | do_statement
    | for_statement
    | foreach_statement
    ;

13.9.2 The while statement

The while statement conditionally executes an embedded statement zero or more times.

while_statement
    : 'while' '(' boolean_expression ')' embedded_statement
    ;

A while statement is executed as follows:

  • The boolean_expression (§12.24) is evaluated.
  • If the Boolean expression yields true, control is transferred to the embedded statement. When and if control reaches the end point of the embedded statement (possibly from execution of a continue statement), control is transferred to the beginning of the while statement.
  • If the Boolean expression yields false, control is transferred to the end point of the while statement.

Within the embedded statement of a while statement, a break statement (§13.10.2) may be used to transfer control to the end point of the while statement (thus ending iteration of the embedded statement), and a continue statement (§13.10.3) may be used to transfer control to the end point of the embedded statement (thus performing another iteration of the while statement).

The embedded statement of a while statement is reachable if the while statement is reachable and the Boolean expression does not have the constant value false.

The end point of a while statement is reachable if at least one of the following is true:

  • The while statement contains a reachable break statement that exits the while statement.
  • The while statement is reachable and the Boolean expression does not have the constant value true.

13.9.3 The do statement

The do statement conditionally executes an embedded statement one or more times.

do_statement
    : 'do' embedded_statement 'while' '(' boolean_expression ')' ';'
    ;

A do statement is executed as follows:

  • Control is transferred to the embedded statement.
  • When and if control reaches the end point of the embedded statement (possibly from execution of a continue statement), the boolean_expression (§12.24) is evaluated. If the Boolean expression yields true, control is transferred to the beginning of the do statement. Otherwise, control is transferred to the end point of the do statement.

Within the embedded statement of a do statement, a break statement (§13.10.2) may be used to transfer control to the end point of the do statement (thus ending iteration of the embedded statement), and a continue statement (§13.10.3) may be used to transfer control to the end point of the embedded statement (thus performing another iteration of the do statement).

The embedded statement of a do statement is reachable if the do statement is reachable.

The end point of a do statement is reachable if at least one of the following is true:

  • The do statement contains a reachable break statement that exits the do statement.
  • The end point of the embedded statement is reachable and the Boolean expression does not have the constant value true.

13.9.4 The for statement

The for statement evaluates a sequence of initialization expressions and then, while a condition is true, repeatedly executes an embedded statement and evaluates a sequence of iteration expressions.

for_statement
    : 'for' '(' for_initializer? ';' for_condition? ';' for_iterator? ')'
      embedded_statement
    ;

for_initializer
    : local_variable_declaration
    | statement_expression_list
    ;

for_condition
    : boolean_expression
    ;

for_iterator
    : statement_expression_list
    ;

statement_expression_list
    : statement_expression (',' statement_expression)*
    ;

The for_initializer, if present, consists of either a local_variable_declaration (§13.6.2) or a list of statement_expressions (§13.7) separated by commas. The scope of a local variable declared by a for_initializer is the for_initializer, for_condition, for_iterator, and embedded_statement.

The for_condition, if present, shall be a boolean_expression (§12.24).

The for_iterator, if present, consists of a list of statement_expressions (§13.7) separated by commas.

A for statement is executed as follows:

  • If a for_initializer is present, the variable initializers or statement expressions are executed in the order they are written. This step is only performed once.
  • If a for_condition is present, it is evaluated.
  • If the for_condition is not present or if the evaluation yields true, control is transferred to the embedded statement. When and if control reaches the end point of the embedded statement (possibly from execution of a continue statement), the expressions of the for_iterator, if any, are evaluated in sequence, and then another iteration is performed, starting with evaluation of the for_condition in the step above.
  • If the for_condition is present and the evaluation yields false, control is transferred to the end point of the for statement.

Within the embedded statement of a for statement, a break statement (§13.10.2) may be used to transfer control to the end point of the for statement (thus ending iteration of the embedded statement), and a continue statement (§13.10.3) may be used to transfer control to the end point of the embedded statement (thus executing the for_iterator and performing another iteration of the for statement, starting with the for_condition).

The embedded statement of a for statement is reachable if one of the following is true:

  • The for statement is reachable and no for_condition is present.
  • The for statement is reachable and a for_condition is present and does not have the constant value false.

The end point of a for statement is reachable if at least one of the following is true:

  • The for statement contains a reachable break statement that exits the for statement.
  • The for statement is reachable and a for_condition is present and does not have the constant value true.

13.9.5 The foreach statement

The foreach statement enumerates the elements of a collection, executing an embedded statement for each element of the collection.

foreach_statement
    : 'foreach' '(' ref_kind? local_variable_type identifier 'in' 
      expression ')' embedded_statement
    ;

The local_variable_type and identifier of a foreach statement declare the iteration variable of the statement. If the var identifier is given as the local_variable_type, and no type named var is in scope, the iteration variable is said to be an implicitly typed iteration variable, and its type is taken to be the element type of the foreach statement, as specified below.

If the foreach_statement contains both or neither ref and readonly, the iteration variable denotes a variable that is treated as read-only. Otherwise, if foreach_statement contains ref without readonly, the iteration variable denotes a variable that shall be writable.

The iteration variable corresponds to a local variable with a scope that extends over the embedded statement. During execution of a foreach statement, the iteration variable represents the collection element for which an iteration is currently being performed. If the iteration variable denotes a read-only variable, a compile-time error occurs if the embedded statement attempts to modify it (via assignment or the ++ and -- operators) or pass it as a reference or output parameter.

In the following, for brevity, IEnumerable, IEnumerator, IEnumerable<T> and IEnumerator<T> refer to the corresponding types in the namespaces System.Collections and System.Collections.Generic.

The compile-time processing of a foreach statement first determines the collection type, enumerator type and iteration type of the expression. This determination proceeds as follows:

  • If the type X of expression is an array type then there is an implicit reference conversion from X to the IEnumerable interface (since System.Array implements this interface). The collection type is the IEnumerable interface, the enumerator type is the IEnumerator interface and the iteration type is the element type of the array type X.
  • If the type X of expression is dynamic then there is an implicit conversion from expression to the IEnumerable interface (§10.2.10). The collection type is the IEnumerable interface and the enumerator type is the IEnumerator interface. If the var identifier is given as the local_variable_type then the iteration type is dynamic, otherwise it is object.
  • Otherwise, determine whether the type X has an appropriate GetEnumerator method:
    • Perform member lookup on the type X with identifier GetEnumerator and no type arguments. If the member lookup does not produce a match, or it produces an ambiguity, or produces a match that is not a method group, check for an enumerable interface as described below. It is recommended that a warning be issued if member lookup produces anything except a method group or no match.
    • Perform overload resolution using the resulting method group and an empty argument list. If overload resolution results in no applicable methods, results in an ambiguity, or results in a single best method but that method is either static or not public, check for an enumerable interface as described below. It is recommended that a warning be issued if overload resolution produces anything except an unambiguous public instance method or no applicable methods.
    • If the return type E of the GetEnumerator method is not a class, struct or interface type, an error is produced and no further steps are taken.
    • Member lookup is performed on E with the identifier Current and no type arguments. If the member lookup produces no match, the result is an error, or the result is anything except a public instance property that permits reading, an error is produced and no further steps are taken.
    • Member lookup is performed on E with the identifier MoveNext and no type arguments. If the member lookup produces no match, the result is an error, or the result is anything except a method group, an error is produced and no further steps are taken.
    • Overload resolution is performed on the method group with an empty argument list. If overload resolution results in no applicable methods, results in an ambiguity, or results in a single best method but that method is either static or not public, or its return type is not bool, an error is produced and no further steps are taken.
    • The collection type is X, the enumerator type is E, and the iteration type is the type of the Current property. The Current property may include the ref modifier, in which case, the expression returned is a variable_reference (§9.5) that is optionally read-only.
  • Otherwise, check for an enumerable interface:
    • If among all the types Tᵢ for which there is an implicit conversion from X to IEnumerable<Tᵢ>, there is a unique type T such that T is not dynamic and for all the other Tᵢ there is an implicit conversion from IEnumerable<T> to IEnumerable<Tᵢ>, then the collection type is the interface IEnumerable<T>, the enumerator type is the interface IEnumerator<T>, and the iteration type is T.
    • Otherwise, if there is more than one such type T, then an error is produced and no further steps are taken.
    • Otherwise, if there is an implicit conversion from X to the System.Collections.IEnumerable interface, then the collection type is this interface, the enumerator type is the interface System.Collections.IEnumerator, and the iteration type is object.
    • Otherwise, an error is produced and no further steps are taken.

The above steps, if successful, unambiguously produce a collection type C, enumerator type E and iteration type T, ref T, or ref readonly T. A foreach statement of the form

foreach (V v in x) «embedded_statement»

is then equivalent to:

{
    E e = ((C)(x)).GetEnumerator();
    try
    {
        while (e.MoveNext())
        {
            V v = (V)(T)e.Current;
            «embedded_statement»
        }
    }
    finally
    {
        ... // Dispose e
    }
}

The variable e is not visible to or accessible to the expression x or the embedded statement or any other source code of the program. The variable v is read-only in the embedded statement. If there is not an explicit conversion (§10.3) from T (the iteration type) to V (the local_variable_type in the foreach statement), an error is produced and no further steps are taken.

When the iteration variable is a reference variable (§9.7), a foreach statement of the form

foreach (ref V v in x) «embedded_statement»

is then equivalent to:

{
    E e = ((C)(x)).GetEnumerator();
    try
    {
        while (e.MoveNext())
        {
            ref V v = ref e.Current;
            «embedded_statement»
        }
    }
    finally
    {
        ... // Dispose e
    }
}

The variable e is not visible or accessible to the expression x or the embedded statement or any other source code of the program. The reference variable v is read-write in the embedded statement, but v shall not be ref-reassigned (§12.21.3). If there is not an identity conversion (§10.2.2) from T (the iteration type) to V (the local_variable_type in the foreach statement), an error is produced and no further steps are taken.

A foreach statement of the form foreach (ref readonly V v in x) «embedded_statement» has a similar equivalent form, but the reference variable v is ref readonly in the embedded statement, and therefore cannot be ref-reassigned or reassigned.

Note: If x has the value null, a System.NullReferenceException is thrown at run-time. end note

An implementation is permitted to implement a given foreach_statement differently; e.g., for performance reasons, as long as the behavior is consistent with the above expansion.

The placement of v inside the while loop is important for how it is captured (§12.19.6.2) by any anonymous function occurring in the embedded_statement.

Example:

int[] values = { 7, 9, 13 };
Action f = null;
foreach (var value in values)
{
    if (f == null)
    {
        f = () => Console.WriteLine("First value: " + value);
    }
}
f();

If v in the expanded form were declared outside of the while loop, it would be shared among all iterations, and its value after the for loop would be the final value, 13, which is what the invocation of f would print. Instead, because each iteration has its own variable v, the one captured by f in the first iteration will continue to hold the value 7, which is what will be printed. (Note that earlier versions of C# declared v outside of the while loop.)

end example

The body of the finally block is constructed according to the following steps:

  • If there is an implicit conversion from E to the System.IDisposable interface, then

    • If E is a non-nullable value type then the finally clause is expanded to the semantic equivalent of:

      finally
      {
          ((System.IDisposable)e).Dispose();
      }
      
    • Otherwise the finally clause is expanded to the semantic equivalent of:

      finally
      {
          System.IDisposable d = e as System.IDisposable;
          if (d != null)
          {
              d.Dispose();
          }
      }
      

      except that if E is a value type, or a type parameter instantiated to a value type, then the conversion of e to System.IDisposable shall not cause boxing to occur.

  • Otherwise, if E is a sealed type, the finally clause is expanded to an empty block:

    finally {}
    
  • Otherwise, the finally clause is expanded to:

    finally
    {
        System.IDisposable d = e as System.IDisposable;
        if (d != null)
        {
            d.Dispose();
        }
    }
    

The local variable d is not visible to or accessible to any user code. In particular, it does not conflict with any other variable whose scope includes the finally block.

The order in which foreach traverses the elements of an array, is as follows: For single-dimensional arrays elements are traversed in increasing index order, starting with index 0 and ending with index Length – 1. For multi-dimensional arrays, elements are traversed such that the indices of the rightmost dimension are increased first, then the next left dimension, and so on to the left.

Example: The following example prints out each value in a two-dimensional array, in element order:

class Test
{
    static void Main()
    {
        double[,] values =
        {
            {1.2, 2.3, 3.4, 4.5},
            {5.6, 6.7, 7.8, 8.9}
        };
        foreach (double elementValue in values)
        {
            Console.Write($"{elementValue} ");
        }
        Console.WriteLine();
    }
}

The output produced is as follows:

1.2 2.3 3.4 4.5 5.6 6.7 7.8 8.9

end example

Example: In the following example

int[] numbers = { 1, 3, 5, 7, 9 };
foreach (var n in numbers)
{
    Console.WriteLine(n);
}

the type of n is inferred to be int, the iteration type of numbers.

end example

13.10 Jump statements

13.10.1 General

Jump statements unconditionally transfer control.

jump_statement
    : break_statement
    | continue_statement
    | goto_statement
    | return_statement
    | throw_statement
    ;

The location to which a jump statement transfers control is called the target of the jump statement.

When a jump statement occurs within a block, and the target of that jump statement is outside that block, the jump statement is said to exit the block. While a jump statement can transfer control out of a block, it can never transfer control into a block.

Execution of jump statements is complicated by the presence of intervening try statements. In the absence of such try statements, a jump statement unconditionally transfers control from the jump statement to its target. In the presence of such intervening try statements, execution is more complex. If the jump statement exits one or more try blocks with associated finally blocks, control is initially transferred to the finally block of the innermost try statement. When and if control reaches the end point of a finally block, control is transferred to the finally block of the next enclosing try statement. This process is repeated until the finally blocks of all intervening try statements have been executed.

Example: In the following code

class Test
{
    static void Main()
    {
        while (true)
        {
            try
            {
                try
                {
                    Console.WriteLine("Before break");
                    break;
                }
                finally
                {
                    Console.WriteLine("Innermost finally block");
                }
            }
            finally
            {
                Console.WriteLine("Outermost finally block");
            }
        }
        Console.WriteLine("After break");
    }
}

the finally blocks associated with two try statements are executed before control is transferred to the target of the jump statement. The output produced is as follows:

Before break
Innermost finally block
Outermost finally block
After break

end example

13.10.2 The break statement

The break statement exits the nearest enclosing switch, while, do, for, or foreach statement.

break_statement
    : 'break' ';'
    ;

The target of a break statement is the end point of the nearest enclosing switch, while, do, for, or foreach statement. If a break statement is not enclosed by a switch, while, do, for, or foreach statement, a compile-time error occurs.

When multiple switch, while, do, for, or foreach statements are nested within each other, a break statement applies only to the innermost statement. To transfer control across multiple nesting levels, a goto statement (§13.10.4) shall be used.

A break statement cannot exit a finally block (§13.11). When a break statement occurs within a finally block, the target of the break statement shall be within the same finally block; otherwise a compile-time error occurs.

A break statement is executed as follows:

  • If the break statement exits one or more try blocks with associated finally blocks, control is initially transferred to the finally block of the innermost try statement. When and if control reaches the end point of a finally block, control is transferred to the finally block of the next enclosing try statement. This process is repeated until the finally blocks of all intervening try statements have been executed.
  • Control is transferred to the target of the break statement.

Because a break statement unconditionally transfers control elsewhere, the end point of a break statement is never reachable.

13.10.3 The continue statement

The continue statement starts a new iteration of the nearest enclosing while, do, for, or foreach statement.

continue_statement
    : 'continue' ';'
    ;

The target of a continue statement is the end point of the embedded statement of the nearest enclosing while, do, for, or foreach statement. If a continue statement is not enclosed by a while, do, for, or foreach statement, a compile-time error occurs.

When multiple while, do, for, or foreach statements are nested within each other, a continue statement applies only to the innermost statement. To transfer control across multiple nesting levels, a goto statement (§13.10.4) shall be used.

A continue statement cannot exit a finally block (§13.11). When a continue statement occurs within a finally block, the target of the continue statement shall be within the same finally block; otherwise a compile-time error occurs.

A continue statement is executed as follows:

  • If the continue statement exits one or more try blocks with associated finally blocks, control is initially transferred to the finally block of the innermost try statement. When and if control reaches the end point of a finally block, control is transferred to the finally block of the next enclosing try statement. This process is repeated until the finally blocks of all intervening try statements have been executed.
  • Control is transferred to the target of the continue statement.

Because a continue statement unconditionally transfers control elsewhere, the end point of a continue statement is never reachable.

13.10.4 The goto statement

The goto statement transfers control to a statement that is marked by a label.

goto_statement
    : 'goto' identifier ';'
    | 'goto' 'case' constant_expression ';'
    | 'goto' 'default' ';'
    ;

The target of a goto identifier statement is the labeled statement with the given label. If a label with the given name does not exist in the current function member, or if the goto statement is not within the scope of the label, a compile-time error occurs.

Note: This rule permits the use of a goto statement to transfer control out of a nested scope, but not into a nested scope. In the example

class Test
{
    static void Main(string[] args)
    {
        string[,] table =
        {
            {"Red", "Blue", "Green"},
            {"Monday", "Wednesday", "Friday"}
        };
        foreach (string str in args)
        {
            int row, colm;
            for (row = 0; row <= 1; ++row)
            {
                for (colm = 0; colm <= 2; ++colm)
                {
                    if (str == table[row,colm])
                    {
                        goto done;
                    }
                }
            }
            Console.WriteLine($"{str} not found");
            continue;
          done:
            Console.WriteLine($"Found {str} at [{row}][{colm}]");
        }
    }
}

a goto statement is used to transfer control out of a nested scope.

end note

The target of a goto case statement is the statement list in the immediately enclosing switch statement (§13.8.3) which contains a case label with a constant pattern of the given constant value and no guard. If the goto case statement is not enclosed by a switch statement, if the nearest enclosing switch statement does not contain such a case, or if the constant_expression is not implicitly convertible (§10.2) to the governing type of the nearest enclosing switch statement, a compile-time error occurs.

The target of a goto default statement is the statement list in the immediately enclosing switch statement (§13.8.3), which contains a default label. If the goto default statement is not enclosed by a switch statement, or if the nearest enclosing switch statement does not contain a default label, a compile-time error occurs.

A goto statement cannot exit a finally block (§13.11). When a goto statement occurs within a finally block, the target of the goto statement shall be within the same finally block, or otherwise a compile-time error occurs.

A goto statement is executed as follows:

  • If the goto statement exits one or more try blocks with associated finally blocks, control is initially transferred to the finally block of the innermost try statement. When and if control reaches the end point of a finally block, control is transferred to the finally block of the next enclosing try statement. This process is repeated until the finally blocks of all intervening try statements have been executed.
  • Control is transferred to the target of the goto statement.

Because a goto statement unconditionally transfers control elsewhere, the end point of a goto statement is never reachable.

13.10.5 The return statement

The return statement returns control to the current caller of the function member in which the return statement appears, optionally returning a value or a variable_reference (§9.5).

return_statement
    : 'return' ';'
    | 'return' expression ';'
    | 'return' 'ref' variable_reference ';'
    ;

A return_statement without expression is called a return-no-value; one containing ref expression is called a return-by-ref; and one containing only expression is called a return-by-value.

It is a compile-time error to use a return-no-value from a method declared as being returns-by-value or returns-by-ref (§15.6.1).

It is a compile-time error to use a return-by-ref from a method declared as being returns-no-value or returns-by-value.

It is a compile-time error to use a return-by-value from a method declared as being returns-no-value or returns-by-ref.

It is a compile-time error to use a return-by-ref if expression is not a variable_reference or is a reference to a variable whose ref-safe-context is not caller-context (§9.7.2).

It is a compile-time error to use a return-by-ref from a method declared with the method_modifier async.

A function member is said to compute a value if it is a method with a returns-by-value method (§15.6.11), a returns-by-value get accessor of a property or indexer, or a user-defined operator. Function members that are returns-no-value do not compute a value and are methods with the effective return type void, set accessors of properties and indexers, add and remove accessors of events, instance constructors, static constructors and finalizers. Function members that are returns-by-ref do not compute a value.

For a return-by-value, an implicit conversion (§10.2) shall exist from the type of expression to the effective return type (§15.6.11) of the containing function member. For a return-by-ref, an identity conversion (§10.2.2) shall exist between the type of expression and the effective return type of the containing function member.

return statements can also be used in the body of anonymous function expressions (§12.19), and participate in determining which conversions exist for those functions (§10.7.1).

It is a compile-time error for a return statement to appear in a finally block (§13.11).

A return statement is executed as follows:

  • For a return-by-value, expression is evaluated and its value is converted to the effective return type of the containing function by an implicit conversion. The result of the conversion becomes the result value produced by the function. For a return-by-ref, the expression is evaluated, and the result shall be classified as a variable. If the enclosing method’s return-by-ref includes readonly, the resulting variable is read-only.
  • If the return statement is enclosed by one or more try or catch blocks with associated finally blocks, control is initially transferred to the finally block of the innermost try statement. When and if control reaches the end point of a finally block, control is transferred to the finally block of the next enclosing try statement. This process is repeated until the finally blocks of all enclosing try statements have been executed.
  • If the containing function is not an async function, control is returned to the caller of the containing function along with the result value, if any.
  • If the containing function is an async function, control is returned to the current caller, and the result value, if any, is recorded in the return task as described in (§15.15.3).

Because a return statement unconditionally transfers control elsewhere, the end point of a return statement is never reachable.

13.10.6 The throw statement

The throw statement throws an exception.

throw_statement
    : 'throw' expression? ';'
    ;

A throw statement with an expression throws an exception produced by evaluating the expression. The expression shall be implicitly convertible to System.Exception, and the result of evaluating the expression is converted to System.Exception before being thrown. If the result of the conversion is null, a System.NullReferenceException is thrown instead.

A throw statement with no expression can be used only in a catch block, in which case, that statement re-throws the exception that is currently being handled by that catch block.

Because a throw statement unconditionally transfers control elsewhere, the end point of a throw statement is never reachable.

When an exception is thrown, control is transferred to the first catch clause in an enclosing try statement that can handle the exception. The process that takes place from the point of the exception being thrown to the point of transferring control to a suitable exception handler is known as exception propagation. Propagation of an exception consists of repeatedly evaluating the following steps until a catch clause that matches the exception is found. In this description, the throw point is initially the location at which the exception is thrown. This behavior is specified in (§21.4).

  • In the current function member, each try statement that encloses the throw point is examined. For each statement S, starting with the innermost try statement and ending with the outermost try statement, the following steps are evaluated:

    • If the try block of S encloses the throw point and if S has one or more catch clauses, the catch clauses are examined in order of appearance to locate a suitable handler for the exception. The first catch clause that specifies an exception type T (or a type parameter that at run-time denotes an exception type T) such that the run-time type of E derives from T is considered a match. If the clause contains an exception filter, the exception object is assigned to the exception variable, and the exception filter is evaluated. When a catch clause contains an exception filter, that catch clause is considered a match if the exception filter evaluates to true. A general catch (§13.11) clause is considered a match for any exception type. If a matching catch clause is located, the exception propagation is completed by transferring control to the block of that catch clause.
    • Otherwise, if the try block or a catch block of S encloses the throw point and if S has a finally block, control is transferred to the finally block. If the finally block throws another exception, processing of the current exception is terminated. Otherwise, when control reaches the end point of the finally block, processing of the current exception is continued.
  • If an exception handler was not located in the current function invocation, the function invocation is terminated, and one of the following occurs:

    • If the current function is non-async, the steps above are repeated for the caller of the function with a throw point corresponding to the statement from which the function member was invoked.

    • If the current function is async and task-returning, the exception is recorded in the return task, which is put into a faulted or canceled state as described in §15.15.3.

    • If the current function is async and void-returning, the synchronization context of the current thread is notified as described in §15.15.4.

  • If the exception processing terminates all function member invocations in the current thread, indicating that the thread has no handler for the exception, then the thread is itself terminated. The impact of such termination is implementation-defined.

13.11 The try statement

The try statement provides a mechanism for catching exceptions that occur during execution of a block. Furthermore, the try statement provides the ability to specify a block of code that is always executed when control leaves the try statement.

try_statement
    : 'try' block catch_clauses
    | 'try' block catch_clauses? finally_clause
    ;

catch_clauses
    : specific_catch_clause+
    | specific_catch_clause* general_catch_clause
    ;

specific_catch_clause
    : 'catch' exception_specifier exception_filter? block
    | 'catch' exception_filter block
    ;

exception_specifier
    : '(' type identifier? ')'
    ;

exception_filter
    : 'when' '(' boolean_expression ')'
    ;

general_catch_clause
    : 'catch' block
    ;

finally_clause
    : 'finally' block
    ;

A try_statement consists of the keyword try followed by a block, then zero or more catch_clauses, then an optional finally_clause. There shall be at least one catch_clause or a finally_clause.

In an exception_specifier the type, or its effective base class if it is a type_parameter, shall be System.Exception or a type that derives from it.

When a catch clause specifies both a class_type and an identifier, an exception variable of the given name and type is declared. The exception variable is introduced into the declaration space of the specific_catch_clause (§7.3). During execution of the exception_filter and catch block, the exception variable represents the exception currently being handled. For purposes of definite assignment checking, the exception variable is considered definitely assigned in its entire scope.

Unless a catch clause includes an exception variable name, it is impossible to access the exception object in the filter and catch block.

A catch clause that specifies neither an exception type nor an exception variable name is called a general catch clause. A try statement can only have one general catch clause, and, if one is present, it shall be the last catch clause.

Note: Some programming languages might support exceptions that are not representable as an object derived from System.Exception, although such exceptions could never be generated by C# code. A general catch clause might be used to catch such exceptions. Thus, a general catch clause is semantically different from one that specifies the type System.Exception, in that the former might also catch exceptions from other languages. end note

In order to locate a handler for an exception, catch clauses are examined in lexical order. If a catch clause specifies a type but no exception filter, it is a compile-time error for a later catch clause of the same try statement to specify a type that is the same as, or is derived from, that type.

Note: Without this restriction, it would be possible to write unreachable catch clauses. end note

Within a catch block, a throw statement (§13.10.6) with no expression can be used to re-throw the exception that was caught by the catch block. Assignments to an exception variable do not alter the exception that is re-thrown.

Example: In the following code

class Test
{
    static void F()
    {
        try
        {
            G();
        }
        catch (Exception e)
        {
            Console.WriteLine("Exception in F: " + e.Message);
            e = new Exception("F");
            throw; // re-throw
        }
    }

    static void G() => throw new Exception("G");

    static void Main()
    {
        try
        {
            F();
        }
        catch (Exception e)
        {
            Console.WriteLine("Exception in Main: " + e.Message);
        }
    }
}

the method F catches an exception, writes some diagnostic information to the console, alters the exception variable, and re-throws the exception. The exception that is re-thrown is the original exception, so the output produced is:

Exception in F: G
Exception in Main: G

If the first catch block had thrown e instead of rethrowing the current exception, the output produced would be as follows:

Exception in F: G
Exception in Main: F

end example

It is a compile-time error for a break, continue, or goto statement to transfer control out of a finally block. When a break, continue, or goto statement occurs in a finally block, the target of the statement shall be within the same finally block, or otherwise a compile-time error occurs.

It is a compile-time error for a return statement to occur in a finally block.

When execution reaches a try statement, control is transferred to the try block. If control reaches the end point of the try block without an exception being propagated, control is transferred to the finally block if one exists. If no finally block exists, control is transferred to the end point of the try statement.

If an exception has been propagated, the catch clauses, if any, are examined in lexical order seeking the first match for the exception. The search for a matching catch clause continues with all enclosing blocks as described in §13.10.6. A catch clause is a match if the exception type matches any exception_specifier and any exception_filter is true. A catch clause without an exception_specifier matches any exception type. The exception type matches the exception_specifier when the exception_specifier specifies the exception type or a base type of the exception type. If the clause contains an exception filter, the exception object is assigned to the exception variable, and the exception filter is evaluated.

If an exception has been propagated and a matching catch clause is found, control is transferred to the first matching catch block. If control reaches the end point of the catch block without an exception being propagated, control is transferred to the finally block if one exists. If no finally block exists, control is transferred to the end point of the try statement. If an exception has been propagated from the catch block, control transfers to the finally block if one exists. The exception is propagated to the next enclosing try statement.

If an exception has been propagated, and no matching catch clause is found, control transfers to the finally block, if it exists. The exception is propagated to the next enclosing try statement.

The statements of a finally block are always executed when control leaves a try statement. This is true whether the control transfer occurs as a result of normal execution, as a result of executing a break, continue, goto, or return statement, or as a result of propagating an exception out of the try statement. If control reaches the end point of the finally block without an exception being propagated, control is transferred to the end point of the try statement.

If an exception is thrown during execution of a finally block, and is not caught within the same finally block, the exception is propagated to the next enclosing try statement. If another exception was in the process of being propagated, that exception is lost. The process of propagating an exception is discussed further in the description of the throw statement (§13.10.6).

Example: In the following code

public class Test
{
    static void Main()
    {
        try
        {
            Method();
        }
        catch (Exception ex) when (ExceptionFilter(ex))
        {
            Console.WriteLine("Catch");
        }

        bool ExceptionFilter(Exception ex)
        {
            Console.WriteLine("Filter");
            return true;
        }
    }

    static void Method()
    {
        try
        {
            throw new ArgumentException();
        }
        finally
        {
            Console.WriteLine("Finally");
        }
    }
}

the method Method throws an exception. The first action is to examine the enclosing catch clauses, executing any exception filters. Then, the finally clause in Method executes before control transfers to the enclosing matching catch clause. The resulting output is:

Filter
Finally
Catch

end example

The try block of a try statement is reachable if the try statement is reachable.

A catch block of a try statement is reachable if the try statement is reachable.

The finally block of a try statement is reachable if the try statement is reachable.

The end point of a try statement is reachable if both of the following are true:

  • The end point of the try block is reachable or the end point of at least one catch block is reachable.
  • If a finally block is present, the end point of the finally block is reachable.

13.12 The checked and unchecked statements

The checked and unchecked statements are used to control the overflow-checking context for integral-type arithmetic operations and conversions.

checked_statement
    : 'checked' block
    ;

unchecked_statement
    : 'unchecked' block
    ;

The checked statement causes all expressions in the block to be evaluated in a checked context, and the unchecked statement causes all expressions in the block to be evaluated in an unchecked context.

The checked and unchecked statements are precisely equivalent to the checked and unchecked operators (§12.8.20), except that they operate on blocks instead of expressions.

13.13 The lock statement

The lock statement obtains the mutual-exclusion lock for a given object, executes a statement, and then releases the lock.

lock_statement
    : 'lock' '(' expression ')' embedded_statement
    ;

The expression of a lock statement shall denote a value of a type known to be a reference. No implicit boxing conversion (§10.2.9) is ever performed for the expression of a lock statement, and thus it is a compile-time error for the expression to denote a value of a value_type.

A lock statement of the form

lock (x)

where x is an expression of a reference_type, is precisely equivalent to:

bool __lockWasTaken = false;
try
{
    System.Threading.Monitor.Enter(x, ref __lockWasTaken);
    ...
}
finally
{
    if (__lockWasTaken)
    {
        System.Threading.Monitor.Exit(x);
    }
}

except that x is only evaluated once.

While a mutual-exclusion lock is held, code executing in the same execution thread can also obtain and release the lock. However, code executing in other threads is blocked from obtaining the lock until the lock is released.

13.14 The using statement

The using statement obtains one or more resources, executes a statement, and then disposes of the resource.

using_statement
    : 'using' '(' resource_acquisition ')' embedded_statement
    ;

resource_acquisition
    : local_variable_declaration
    | expression
    ;

A resource is a class or struct that implements the System.IDisposable interface, which includes a single parameterless method named Dispose. Code that is using a resource can call Dispose to indicate that the resource is no longer needed.

If the form of resource_acquisition is local_variable_declaration then the type of the local_variable_declaration shall be either dynamic or a type that can be implicitly converted to System.IDisposable. If the form of resource_acquisition is expression then this expression shall be implicitly convertible to System.IDisposable.

Local variables declared in a resource_acquisition are read-only, and shall include an initializer. A compile-time error occurs if the embedded statement attempts to modify these local variables (via assignment or the ++ and -- operators), take the address of them, or pass them as reference or output parameters.

A using statement is translated into three parts: acquisition, usage, and disposal. Usage of the resource is implicitly enclosed in a try statement that includes a finally clause. This finally clause disposes of the resource. If a null resource is acquired, then no call to Dispose is made, and no exception is thrown. If the resource is of type dynamic it is dynamically converted through an implicit dynamic conversion (§10.2.10) to IDisposable during acquisition in order to ensure that the conversion is successful before the usage and disposal.

A using statement of the form

using (ResourceType resource = «expression» ) «statement»

corresponds to one of three possible expansions. When ResourceType is a non-nullable value type or a type parameter with the value type constraint (§15.2.5), the expansion is semantically equivalent to

{
    ResourceType resource = «expression»;
    try
    {
        «statement»;
    }
    finally
    {
        ((IDisposable)resource).Dispose();
    }
}

except that the cast of resource to System.IDisposable shall not cause boxing to occur.

Otherwise, when ResourceType is dynamic, the expansion is

{
    ResourceType resource = «expression»;
    IDisposable d = resource;
    try
    {
        «statement»;
    }
    finally
    {
        if (d != null)
        {
            d.Dispose();
        }
    }
}

Otherwise, the expansion is

{
    ResourceType resource = «expression»;
    try
    {
        «statement»;
    }
    finally
    {
        IDisposable d = (IDisposable)resource;
        if (d != null)
        {
            d.Dispose();
        }
    }
}

In any expansion, the resource variable is read-only in the embedded statement, and the d variable is inaccessible in, and invisible to, the embedded statement.

An implementation is permitted to implement a given using_statement differently, e.g., for performance reasons, as long as the behavior is consistent with the above expansion.

A using statement of the form:

using («expression») «statement»

has the same three possible expansions. In this case ResourceType is implicitly the compile-time type of the expression, if it has one. Otherwise the interface IDisposable itself is used as the ResourceType. The resource variable is inaccessible in, and invisible to, the embedded statement.

When a resource_acquisition takes the form of a local_variable_declaration, it is possible to acquire multiple resources of a given type. A using statement of the form

using (ResourceType r1 = e1, r2 = e2, ..., rN = eN) «statement»

is precisely equivalent to a sequence of nested using statements:

using (ResourceType r1 = e1)
using (ResourceType r2 = e2)
...
using (ResourceType rN = eN)
«statement»

Example: The example below creates a file named log.txt and writes two lines of text to the file. The example then opens that same file for reading and copies the contained lines of text to the console.

class Test
{
    static void Main()
    {
        using (TextWriter w = File.CreateText("log.txt"))
        {
            w.WriteLine("This is line one");
            w.WriteLine("This is line two");
        }
        using (TextReader r = File.OpenText("log.txt"))
        {
            string s;
            while ((s = r.ReadLine()) != null)
            {
                Console.WriteLine(s);
            }
        }
    }
}

Since the TextWriter and TextReader classes implement the IDisposable interface, the example can use using statements to ensure that the underlying file is properly closed following the write or read operations.

end example

13.15 The yield statement

The yield statement is used in an iterator block (§13.3) to yield a value to the enumerator object (§15.14.5) or enumerable object (§15.14.6) of an iterator or to signal the end of the iteration.

yield_statement
    : 'yield' 'return' expression ';'
    | 'yield' 'break' ';'
    ;

yield is a contextual keyword (§6.4.4) and has special meaning only when used immediately before a return or break keyword.

There are several restrictions on where a yield statement can appear, as described in the following.

  • It is a compile-time error for a yield statement (of either form) to appear outside a method_body, operator_body, or accessor_body.
  • It is a compile-time error for a yield statement (of either form) to appear inside an anonymous function.
  • It is a compile-time error for a yield statement (of either form) to appear in the finally clause of a try statement.
  • It is a compile-time error for a yield return statement to appear anywhere in a try statement that contains any catch_clauses.

Example: The following example shows some valid and invalid uses of yield statements.

delegate IEnumerable<int> D();

IEnumerator<int> GetEnumerator()
{
    try
    {
        yield return 1; // Ok
        yield break;    // Ok
    }
    finally
    {
        yield return 2; // Error, yield in finally
        yield break;    // Error, yield in finally
    }
    try
    {
        yield return 3; // Error, yield return in try/catch
        yield break;    // Ok
    }
    catch
    {
        yield return 4; // Error, yield return in try/catch
        yield break;    // Ok
    }
    D d = delegate
    {
        yield return 5; // Error, yield in an anonymous function
    };
}

int MyMethod()
{
    yield return 1;     // Error, wrong return type for an iterator block
}

end example

An implicit conversion (§10.2) shall exist from the type of the expression in the yield return statement to the yield type (§15.14.4) of the iterator.

A yield return statement is executed as follows:

  • The expression given in the statement is evaluated, implicitly converted to the yield type, and assigned to the Current property of the enumerator object.
  • Execution of the iterator block is suspended. If the yield return statement is within one or more try blocks, the associated finally blocks are not executed at this time.
  • The MoveNext method of the enumerator object returns true to its caller, indicating that the enumerator object successfully advanced to the next item.

The next call to the enumerator object’s MoveNext method resumes execution of the iterator block from where it was last suspended.

A yield break statement is executed as follows:

  • If the yield break statement is enclosed by one or more try blocks with associated finally blocks, control is initially transferred to the finally block of the innermost try statement. When and if control reaches the end point of a finally block, control is transferred to the finally block of the next enclosing try statement. This process is repeated until the finally blocks of all enclosing try statements have been executed.
  • Control is returned to the caller of the iterator block. This is either the MoveNext method or Dispose method of the enumerator object.

Because a yield break statement unconditionally transfers control elsewhere, the end point of a yield break statement is never reachable.