5 Statements 1 [A statement defines an action to be performed upon its execution.] 2/3 {AI95-00318-02} {AI05-0299-1} [This clause describes the general rules applicable to all statements. Some statements are discussed in later clauses: Procedure_call_statements and return statements are described in 6, " Subprograms". Entry_call_statements, requeue_statements, delay_statements, accept_statements, select_statements, and abort_statements are described in 9, "Tasks and Synchronization". Raise_statements are described in 11, " Exceptions", and code_statements in 13. The remaining forms of statements are presented in this clause.] Wording Changes from Ada 83 2.a/2 {AI95-00318-02} The description of return statements has been moved to 6.5, "Return Statements", so that it is closer to the description of subprograms. 5.1 Simple and Compound Statements - Sequences of Statements 1 [A statement is either simple or compound. A simple_statement encloses no other statement. A compound_statement can enclose simple_statements and other compound_statements.] Syntax 2/3 {AI05-0179-1} sequence_of_statements ::= statement {statement } {label} 3 statement ::= {label} simple_statement | {label} compound_statement 4/2 {AI95-00318-02} simple_statement ::= null_statement | assignment_statement | exit_statement | goto_statement | procedure_call_statement | simple_return_statement | entry_call_statement | requeue_statement | delay_statement | abort_statement | raise_statement | code_statement 5/2 {AI95-00318-02} compound_statement ::= if_statement | case_statement | loop_statement | block_statement | extended_return_statement | accept_statement | select_statement 6 null_statement ::= null; 7 label ::= <> 8 statement_identifier ::= direct_name 9 The direct_name of a statement_identifier shall be an identifier (not an operator_symbol). Name Resolution Rules 10 The direct_name of a statement_identifier shall resolve to denote its corresponding implicit declaration (see below). Legality Rules 11 Distinct identifiers shall be used for all statement_identifiers that appear in the same body, including inner block_statements but excluding inner program units. Static Semantics 12 For each statement_identifier, there is an implicit declaration (with the specified identifier) at the end of the declarative_part of the innermost block_statement or body that encloses the statement_identifier. The implicit declarations occur in the same order as the statement_identifiers occur in the source text. If a usage name denotes such an implicit declaration, the entity it denotes is the label, loop_statement, or block_statement with the given statement_identifier. 12.a Reason: We talk in terms of individual statement_identifiers here rather than in terms of the corresponding statements, since a given statement may have multiple statement_identifiers. 12.b A block_statement that has no explicit declarative_part has an implicit empty declarative_part, so this rule can safely refer to the declarative_part of a block_statement. 12.c The scope of a declaration starts at the place of the declaration itself (see 8.2). In the case of a label, loop, or block name, it follows from this rule that the scope of the implicit declaration starts before the first explicit occurrence of the corresponding name, since this occurrence is either in a statement label, a loop_statement, a block_statement, or a goto_statement. An implicit declaration in a block_statement may hide a declaration given in an outer program unit or block_statement (according to the usual rules of hiding explained in 8.3). 12.d The syntax rule for label uses statement_identifier which is a direct_name (not a defining_identifier), because labels are implicitly declared. The same applies to loop and block names. In other words, the label itself is not the defining occurrence; the implicit declaration is. 12.e We cannot consider the label to be a defining occurrence. An example that can tell the difference is this: 12.f declare -- Label Foo is implicitly declared here. begin for Foo in ... loop ... <> -- Illegal. ... end loop; end; 12.g/3 {AI05-0299-1} The label in this example is hidden from itself by the loop parameter with the same name; the example is illegal. We considered creating a new syntactic category name, separate from direct_name and selector_name, for use in the case of statement labels. However, that would confuse the rules in Clause 8, so we didn't do it. 12.1/3 {AI05-0179-1} If one or more labels end a sequence_of_statements, an implicit null_statement follows the labels before any following constructs. 12.g.1/3 Reason: The semantics of a goto_statement is defined in terms of the statement having (following) that label. Thus we ensure that every label has a following statement, which might be implicit. Dynamic Semantics 13 The execution of a null_statement has no effect. 14/2 {AI95-00318-02} A transfer of control is the run-time action of an exit_statement, return statement, goto_statement, or requeue_statement, selection of a terminate_alternative, raising of an exception, or an abort, which causes the next action performed to be one other than what would normally be expected from the other rules of the language. [As explained in 7.6.1, a transfer of control can cause the execution of constructs to be completed and then left, which may trigger finalization.] 15 The execution of a sequence_of_statements consists of the execution of the individual statements in succession until the sequence_ is completed. 15.a Ramification: It could be completed by reaching the end of it, or by a transfer of control. NOTES 16 1 A statement_identifier that appears immediately within the declarative region of a named loop_statement or an accept_statement is nevertheless implicitly declared immediately within the declarative region of the innermost enclosing body or block_statement; in other words, the expanded name for a named statement is not affected by whether the statement occurs inside or outside a named loop or an accept_statement - only nesting within block_statements is relevant to the form of its expanded name. 16.a Discussion: Each comment in the following example gives the expanded name associated with an entity declared in the task body: 16.b task body Compute is Sum : Integer := 0; -- Compute.Sum begin Outer: -- Compute.Outer for I in 1..10 loop -- Compute.Outer.I Blk: -- Compute.Blk declare Sum : Integer := 0; -- Compute.Blk.Sum begin accept Ent(I : out Integer; J : in Integer) do -- Compute.Ent.I, Compute.Ent.J Compute.Ent.I := Compute.Outer.I; Inner: -- Compute.Blk.Inner for J in 1..10 loop -- Compute.Blk.Inner.J Sum := Sum + Compute.Blk.Inner.J * Compute.Ent.J; end loop Inner; end Ent; Compute.Sum := Compute.Sum + Compute.Blk.Sum; end Blk; end loop Outer; Record_Result(Sum); end Compute; Examples 17 Examples of labeled statements: 18 <> <> <> <> null; 19 <> X := 1; Extensions to Ada 83 19.a The requeue_statement is new. Wording Changes from Ada 83 19.b We define the syntactic category statement_identifier to simplify the description. It is used for labels, loop names, and block names. We define the entity associated with the implicit declarations of statement names. 19.c Completion includes completion caused by a transfer of control, although RM83-5.1(6) did not take this view. Extensions to Ada 95 19.d/2 {AI95-00318-02} The extended_return_statement is new (simple_return_statement is merely renamed). Extensions to Ada 2005 19.e/3 {AI95-0179-1} A label can end a sequence_of_statements, eliminating the requirement for having an explicit null; statement after an ending label (a common use). 5.2 Assignment Statements 1 [An assignment_statement replaces the current value of a variable with the result of evaluating an expression.] Syntax 2 assignment_statement ::= variable_name := expression; 3 The execution of an assignment_statement includes the evaluation of the expression and the assignment of the value of the expression into the target. [An assignment operation (as opposed to an assignment_statement) is performed in other contexts as well, including object initialization and by-copy parameter passing.] The target of an assignment operation is the view of the object to which a value is being assigned; the target of an assignment_- statement is the variable denoted by the variable_name. 3.a Discussion: Don't confuse this notion of the "target" of an assignment with the notion of the "target object" of an entry call or requeue. 3.b Don't confuse the term "assignment operation" with the assignment_statement. The assignment operation is just one part of the execution of an assignment_statement. The assignment operation is also a part of the execution of various other constructs; see 7.6.1, "Completion and Finalization" for a complete list. Note that when we say, "such-and-such is assigned to so-and-so", we mean that the assignment operation is being applied, and that so-and-so is the target of the assignment operation. Name Resolution Rules 4/2 {AI95-00287-01} The variable_name of an assignment_statement is expected to be of any type. The expected type for the expression is the type of the target. 4.a Implementation Note: An assignment_statement as a whole is a "complete context," so if the variable_name of an assignment_statement is overloaded, the expression can be used to help disambiguate it. For example: 4.b type P1 is access R1; type P2 is access R2; 4.c function F return P1; function F return P2; 4.d X : R1; begin F.all := X; -- Right hand side helps resolve left hand side Legality Rules 5/2 {AI95-00287-01} The target [denoted by the variable_name] shall be a variable of a nonlimited type. 6 If the target is of a tagged class-wide type T'Class, then the expression shall either be dynamically tagged, or of type T and tag-indeterminate (see 3.9.2). 6.a Reason: This is consistent with the general rule that a single dispatching operation shall not have both dynamically tagged and statically tagged operands. Note that for an object initialization (as opposed to the assignment_statement), a statically tagged initialization expression is permitted, since there is no chance for confusion (or Tag_Check failure). Also, in an object initialization, tag-indeterminate expressions of any type covered by T'Class would be allowed, but with an assignment_statement, that might not work if the tag of the target was for a type that didn't have one of the dispatching operations in the tag-indeterminate expression. Dynamic Semantics 7 For the execution of an assignment_statement, the variable_name and the expression are first evaluated in an arbitrary order. 7.a Ramification: Other rules of the language may require that the bounds of the variable be determined prior to evaluating the expression, but that does not necessarily require evaluation of the variable_name, as pointed out by the ACID. 8 When the type of the target is class-wide: 9 * If the expression is tag-indeterminate (see 3.9.2), then the controlling tag value for the expression is the tag of the target; 9.a Ramification: See 3.9.2, " Dispatching Operations of Tagged Types". 10 * Otherwise [(the expression is dynamically tagged)], a check is made that the tag of the value of the expression is the same as that of the target; if this check fails, Constraint_Error is raised. 11 The value of the expression is converted to the subtype of the target. [The conversion might raise an exception (see 4.6).] 11.a Ramification: 4.6, "Type Conversions" defines what actions and checks are associated with subtype conversion. For non-array subtypes, it is just a constraint check presuming the types match. For array subtypes, it checks the lengths and slides if the target is constrained. "Sliding" means the array doesn't have to have the same bounds, so long as it is the same length. 12 In cases involving controlled types, the target is finalized, and an anonymous object might be used as an intermediate in the assignment, as described in 7.6.1, "Completion and Finalization". In any case, the converted value of the expression is then assigned to the target, which consists of the following two steps: 12.a To be honest: 7.6.1 actually says that finalization happens always, but unless controlled types are involved, this finalization during an assignment_statement does nothing. 13 * The value of the target becomes the converted value. 14/3 * {AI05-0299-1} If any part of the target is controlled, its value is adjusted as explained in subclause 7.6. 14.a Ramification: If any parts of the object are controlled, abort is deferred during the assignment operation itself, but not during the rest of the execution of an assignment_statement. NOTES 15 2 The tag of an object never changes; in particular, an assignment_statement does not change the tag of the target. 16/2 This paragraph was deleted.{AI95-00363-01} 16.a Ramification: The implicit subtype conversion described above for assignment_statements is performed only for the value of the right-hand side expression as a whole; it is not performed for subcomponents of the value. 16.b The determination of the type of the variable of an assignment_statement may require consideration of the expression if the variable name can be interpreted as the name of a variable designated by the access value returned by a function call, and similarly, as a component or slice of such a variable (see 8.6 , "The Context of Overload Resolution"). Examples 17 Examples of assignment statements: 18 Value := Max_Value - 1; Shade := Blue; 19 Next_Frame(F)(M, N) := 2.5; -- see 4.1.1 U := Dot_Product(V, W); -- see 6.3 20/4 {AI12-0056-1} Writer := (Status => Open, Unit => Printer, Line_Count => 60); -- see 3.8.1 Next.all := (72074, null, Head); -- see 3.10.1 21 Examples involving scalar subtype conversions: 22 I, J : Integer range 1 .. 10 := 5; K : Integer range 1 .. 20 := 15; ... 23 I := J; -- identical ranges K := J; -- compatible ranges J := K; -- will raise Constraint_Error if K > 10 24 Examples involving array subtype conversions: 25 A : String(1 .. 31); B : String(3 .. 33); ... 26 A := B; -- same number of components 27 A(1 .. 9) := "tar sauce"; A(4 .. 12) := A(1 .. 9); -- A(1 .. 12) = "tartar sauce" NOTES 28 3 Notes on the examples: Assignment_statements are allowed even in the case of overlapping slices of the same array, because the variable_name and expression are both evaluated before copying the value into the variable. In the above example, an implementation yielding A(1 .. 12) = "tartartartar" would be incorrect. Extensions to Ada 83 28.a We now allow user-defined finalization and value adjustment actions as part of assignment_statements (see 7.6, " Assignment and Finalization"). Wording Changes from Ada 83 28.b The special case of array assignment is subsumed by the concept of a subtype conversion, which is applied for all kinds of types, not just arrays. For arrays it provides "sliding". For numeric types it provides conversion of a value of a universal type to the specific type of the target. For other types, it generally has no run-time effect, other than a constraint check. 28.c We now cover in a general way in 3.7.2 the erroneous execution possible due to changing the value of a discriminant when the variable in an assignment_statement is a subcomponent that depends on discriminants. Incompatibilities With Ada 95 28.d/2 {AI95-00287-01} The change of the limited check from a resolution rule to a legality rule is not quite upward compatible. For example 28.e/2 type AccNonLim is access NonLim; function Foo (Arg : in Integer) return AccNonLim; type AccLim is access Lim; function Foo (Arg : in Integer) return AccLim; Foo(2).all := Foo(1).all; 28.f/2 where NonLim is a nonlimited type and Lim is a limited type. The assignment is legal in Ada 95 (only the first Foo would be considered), and is ambiguous in Ada 2005. We made the change because we want limited types to be as similar to nonlimited types as possible. Limited expressions are now allowed in all other contexts (with a similar incompatibility), and it would be odd if assignments had different resolution rules (which would eliminate ambiguities in some cases). Moreover, examples like this one are rare, as they depend on assigning into overloaded function calls. 5.3 If Statements 1 [An if_statement selects for execution at most one of the enclosed sequences_of_statements, depending on the (truth) value of one or more corresponding conditions.] Syntax 2 if_statement ::= if condition then sequence_of_statements {elsif condition then sequence_of_statements} [else sequence_of_statements] end if; Paragraphs 3 and 4 were deleted. Dynamic Semantics 5/3 {AI05-0264-1} For the execution of an if_statement, the condition specified after if, and any conditions specified after elsif, are evaluated in succession (treating a final else as elsif True then), until one evaluates to True or all conditions are evaluated and yield False. If a condition evaluates to True, then the corresponding sequence_of_statements is executed; otherwise, none of them is executed. 5.a Ramification: The part about all evaluating to False can't happen if there is an else, since that is herein considered equivalent to elsif True then. Examples 6 Examples of if statements: 7 if Month = December and Day = 31 then Month := January; Day := 1; Year := Year + 1; end if; 8 if Line_Too_Short then raise Layout_Error; elsif Line_Full then New_Line; Put(Item); else Put(Item); end if; 9 if My_Car.Owner.Vehicle /= My_Car then -- see 3.10.1 Report ("Incorrect data"); end if; Wording Changes from Ada 2005 9.a/3 {AI05-0147-1} Moved definition of condition to 4.5.7 in order to eliminate a forward reference. 5.4 Case Statements 1 [A case_statement selects for execution one of a number of alternative sequences_of_statements; the chosen alternative is defined by the value of an expression.] Syntax 2/3 {AI05-0188-1} case_statement ::= case selecting_expression is case_statement_alternative {case_statement_alternative} end case; 3 case_statement_alternative ::= when discrete_choice_list => sequence_of_statements Name Resolution Rules 4/3 {AI05-0188-1} The selecting_expression is expected to be of any discrete type. The expected type for each discrete_choice is the type of the selecting_expression. Legality Rules 5/3 {AI05-0153-3} The choice_expressions, subtype_indications, and ranges given as discrete_choices of a case_statement shall be static. [A discrete_choice others, if present, shall appear alone and in the last discrete_choice_list.] 6/3 {AI05-0188-1} {AI05-0240-1} The possible values of the selecting_expression shall be covered (see 3.8.1) as follows: 6.a/3 Discussion: {AI05-0240-1} The meaning of "covered" here and in the following rules is that of the term "cover a value" that is defined in 3.8.1. 7/4 * {AI05-0003-1} {AI05-0153-3} {AI05-0188-1} {AI05-0262-1} {AI12-0071-1} If the selecting_expression is a name [(including a type_conversion, qualified_expression, or function_call)] having a static and constrained nominal subtype, then each non-others discrete_choice shall cover only values in that subtype that satisfy its predicates (see 3.2.4), and each value of that subtype that satisfies its predicates shall be covered by some discrete_choice [(either explicitly or by others)]. 7.a Ramification: Although not official names of objects, a value conversion still has a defined nominal subtype, namely its target subtype. See 4.6. 8/3 * {AI05-0188-1} If the type of the selecting_expression is root_integer, universal_integer, or a descendant of a formal scalar type, then the case_statement shall have an others discrete_choice. 8.a Reason: This is because the base range is implementation defined for root_integer and universal_integer, and not known statically in the case of a formal scalar type. 9/3 * {AI05-0188-1} Otherwise, each value of the base range of the type of the selecting_expression shall be covered [(either explicitly or by others)]. 10 Two distinct discrete_choices of a case_statement shall not cover the same value. 10.a/3 Ramification: {AI05-0188-1} The goal of these coverage rules is that any possible value of the selecting_expression of a case_statement should be covered by exactly one discrete_choice of the case_statement, and that this should be checked at compile time. The goal is achieved in most cases, but there are two minor loopholes: 10.b * If the expression reads an object with an invalid representation (e.g. an uninitialized object), then the value can be outside the covered range. This can happen for static constrained subtypes, as well as nonstatic or unconstrained subtypes. It cannot, however, happen if the case_statement has the discrete_choice others, because others covers all values, even those outside the subtype. 10.c/3 * {AI95-00114-01} {AI05-0188-1} If the compiler chooses to represent the value of an expression of an unconstrained subtype in a way that includes values outside the bounds of the subtype, then those values can be outside the covered range. For example, if X: Integer := Integer'Last;, and the case selecting_expression is X+1, then the implementation might choose to produce the correct value, which is outside the bounds of Integer. (It might raise Constraint_Error instead.) This case can only happen for nongeneric subtypes that are either unconstrained or nonstatic (or both). It can only happen if there is no others discrete_choice. 10.d In the uninitialized variable case, the value might be anything; hence, any alternative can be chosen, or Constraint_Error can be raised. (We intend to prevent, however, jumping to random memory locations and the like.) In the out-of-range case, the behavior is more sensible: if there is an others, then the implementation may choose to raise Constraint_Error on the evaluation of the expression (as usual), or it may choose to correctly evaluate the expression and therefore choose the others alternative. Otherwise (no others), Constraint_Error is raised either way - on the expression evaluation, or for the case_statement itself. 10.e For an enumeration type with a discontiguous set of internal codes (see 13.4), the only way to get values in between the proper values is via an object with an invalid representation; there is no "out-of-range" situation that can produce them. Dynamic Semantics 11/3 {AI05-0188-1} For the execution of a case_statement the selecting_expression is first evaluated. 12/3 {AI05-0188-1} If the value of the selecting_expression is covered by the discrete_choice_list of some case_statement_alternative, then the sequence_of_- statements of the _alternative is executed. 13 Otherwise (the value is not covered by any discrete_choice_list, perhaps due to being outside the base range), Constraint_Error is raised. 13.a Ramification: In this case, the value is outside the base range of its type, or is an invalid representation. NOTES 14 4 The execution of a case_statement chooses one and only one alternative. Qualification of the expression of a case_statement by a static subtype can often be used to limit the number of choices that need be given explicitly. Examples 15 Examples of case statements: 16 case Sensor is when Elevation => Record_Elevation(Sensor_Value); when Azimuth => Record_Azimuth (Sensor_Value); when Distance => Record_Distance (Sensor_Value); when others => null; end case; 17 case Today is when Mon => Compute_Initial_Balance; when Fri => Compute_Closing_Balance; when Tue .. Thu => Generate_Report(Today); when Sat .. Sun => null; end case; 18 case Bin_Number(Count) is when 1 => Update_Bin(1); when 2 => Update_Bin(2); when 3 | 4 => Empty_Bin(1); Empty_Bin(2); when others => raise Error; end case; Incompatibilities With Ada 83 18.a.1/1 In Ada 95, function_calls and type_conversions are names, whereas in Ada 83, they were expressions. Therefore, if the expression of a case_statement is a function_call or type_conversion, and the result subtype is static, it is illegal to specify a choice outside the bounds of the subtype. For this case in Ada 83 choices only are required to be in the base range of the type. 18.a.2/1 In addition, the rule about which choices must be covered is unchanged in Ada 95. Therefore, for a case_statement whose expression is a function_call or type_conversion, Ada 83 required covering all choices in the base range, while Ada 95 only requires covering choices in the bounds of the subtype. If the case_statement does not include an others discrete_choice, then a legal Ada 83 case_statement will be illegal in Ada 95 if the bounds of the subtype are different than the bounds of the base type. Extensions to Ada 83 18.a In Ada 83, the expression in a case_statement is not allowed to be of a generic formal type. This restriction is removed in Ada 95; an others discrete_choice is required instead. 18.b In Ada 95, a function call is the name of an object; this was not true in Ada 83 (see 4.1, "Names"). This change makes the following case_statement legal: 18.c subtype S is Integer range 1..2; function F return S; case F is when 1 => ...; when 2 => ...; -- No others needed. end case; 18.d/3 {AI05-0005-1} Note that the result subtype given in a function renaming_declaration is ignored; for a case_statement whose expression calls a such a function, the full coverage rules are checked using the result subtype of the original function. Note that predefined operators such as "+" have an unconstrained result subtype (see 4.5.1). Note that generic formal functions do not have static result subtypes. Note that the result subtype of an inherited subprogram need not correspond to any nameable subtype; there is still a perfectly good result subtype, though. Wording Changes from Ada 83 18.e Ada 83 forgot to say what happens for "legally" out-of-bounds values. 18.f We take advantage of rules and terms (e.g. cover a value) defined for discrete_choices and discrete_choice_lists in 3.8.1, " Variant Parts and Discrete Choices". 18.g In the Name Resolution Rule for the case expression, we no longer need RM83-5.4(3)'s "which must be determinable independently of the context in which the expression occurs, but using the fact that the expression must be of a discrete type," because the expression is now a complete context. See 8.6, " The Context of Overload Resolution". 18.h Since type_conversions are now defined as names, their coverage rule is now covered under the general rule for names, rather than being separated out along with qualified_expressions. Wording Changes from Ada 2005 18.i/3 {AI05-0003-1} Rewording to reflect that a qualified_expression is now a name. 18.j/3 {AI05-0153-3} Revised for changes to discrete_choices made to allow static predicates (see 3.2.4) as case choices (see 3.8.1). 18.k/3 {AI05-0188-1} Added the selecting_ prefix to make this wording consistent with case_expression, and to clarify which expression is being talked about in the wording. Wording Changes from Ada 2012 18.l/4 {AI12-0071-1} Corrigendum: Updated wording of case coverage to use the new term "satisfies the predicates" (see 3.2.4). 5.5 Loop Statements 1 [A loop_statement includes a sequence_of_statements that is to be executed repeatedly, zero or more times.] Syntax 2 loop_statement ::= [loop_statement_identifier:] [iteration_scheme] loop sequence_of_statements end loop [loop_identifier]; 3/3 {AI05-0139-2} iteration_scheme ::= while condition | for loop_parameter_specification | for iterator_specification 4 loop_parameter_specification ::= defining_identifier in [reverse] discrete_subtype_definition 5 If a loop_statement has a loop_statement_identifier, then the identifier shall be repeated after the end loop; otherwise, there shall not be an identifier after the end loop. Static Semantics 6 A loop_parameter_specification declares a loop parameter, which is an object whose subtype is that defined by the discrete_subtype_definition. Dynamic Semantics 7 For the execution of a loop_statement, the sequence_of_statements is executed repeatedly, zero or more times, until the loop_statement is complete. The loop_statement is complete when a transfer of control occurs that transfers control out of the loop, or, in the case of an iteration_scheme, as specified below. 8 For the execution of a loop_statement with a while iteration_scheme, the condition is evaluated before each execution of the sequence_of_statements; if the value of the condition is True, the sequence_of_statements is executed; if False, the execution of the loop_statement is complete. 9/4 {AI05-0139-2} {AI05-0262-1} {AI12-0071-1} For the execution of a loop_statement with the iteration_scheme being for loop_parameter_specification, the loop_parameter_specification is first elaborated. This elaboration creates the loop parameter and elaborates the discrete_subtype_definition. If the discrete_subtype_definition defines a subtype with a null range, the execution of the loop_statement is complete. Otherwise, the sequence_of_statements is executed once for each value of the discrete subtype defined by the discrete_subtype_definition that satisfies the predicates of the subtype (or until the loop is left as a consequence of a transfer of control). Prior to each such iteration, the corresponding value of the discrete subtype is assigned to the loop parameter. These values are assigned in increasing order unless the reserved word reverse is present, in which case the values are assigned in decreasing order. 9.a Ramification: The order of creating the loop parameter and evaluating the discrete_subtype_definition doesn't matter, since the creation of the loop parameter has no side effects (other than possibly raising Storage_Error, but anything can do that). 9.b/3 {AI05-0262-1} The predicate (if any) necessarily has to be a static predicate as a dynamic predicate is explicitly disallowed - see 3.2.4. 9.c/3 Reason: {AI05-0262-1} If there is a predicate, the loop still visits the values in the order of the underlying base type; the order of the values in the predicate is irrelevant. This is the case so that the following loops have the same sequence of calls and parameters on procedure Call for any subtype S: 9.d for I in S loop Call (I); end loop; 9.e for I in S'Base loop if I in S then Call (I); end if; end loop; 9.1/3 {AI05-0262-1} [For details about the execution of a loop_statement with the iteration_scheme being for iterator_specification, see 5.5.2.] NOTES 10 5 A loop parameter is a constant; it cannot be updated within the sequence_of_statements of the loop (see 3.3). 11 6 An object_declaration should not be given for a loop parameter, since the loop parameter is automatically declared by the loop_parameter_specification. The scope of a loop parameter extends from the loop_parameter_specification to the end of the loop_statement, and the visibility rules are such that a loop parameter is only visible within the sequence_of_statements of the loop. 11.a Implementation Note: An implementation could give a warning if a variable is hidden by a loop_parameter_specification. 12 7 The discrete_subtype_definition of a for loop is elaborated just once. Use of the reserved word reverse does not alter the discrete subtype defined, so that the following iteration_schemes are not equivalent; the first has a null range. 13 for J in reverse 1 .. 0 for J in 0 .. 1 13.a Ramification: If a loop_parameter_specification has a static discrete range, the subtype of the loop parameter is static. Examples 14 Example of a loop statement without an iteration scheme: 15 loop Get(Current_Character); exit when Current_Character = '*'; end loop; 16 Example of a loop statement with a while iteration scheme: 17 while Bid(N).Price < Cut_Off.Price loop Record_Bid(Bid(N).Price); N := N + 1; end loop; 18 Example of a loop statement with a for iteration scheme: 19 for J in Buffer'Range loop -- works even with a null range if Buffer(J) /= Space then Put(Buffer(J)); end if; end loop; 20 Example of a loop statement with a name: 21 Summation: while Next /= Head loop -- see 3.10.1 Sum := Sum + Next.Value; Next := Next.Succ; end loop Summation; Wording Changes from Ada 83 21.a The constant-ness of loop parameters is specified in 3.3, " Objects and Named Numbers". Wording Changes from Ada 2005 21.b/3 {AI05-0139-2} {AI05-0262-1} {AI05-0299-1} Generalized iterator_specifications are allowed in for loops; these are documented as an extension in the appropriate subclause. Wording Changes from Ada 2012 21.c/4 {AI12-0071-1} Corrigendum: Updated wording of loop execution to use the new term "satisfies the predicates" (see 3.2.4). 5.5.1 User-Defined Iterator Types Static Semantics 1/3 {AI05-0139-2} The following language-defined generic library package exists: 2/3 generic type Cursor; with function Has_Element (Position : Cursor) return Boolean; package Ada.Iterator_Interfaces is pragma Pure (Iterator_Interfaces); 3/3 type Forward_Iterator is limited interface; function First (Object : Forward_Iterator) return Cursor is abstract; function Next (Object : Forward_Iterator; Position : Cursor) return Cursor is abstract; 4/3 type Reversible_Iterator is limited interface and Forward_Iterator; function Last (Object : Reversible_Iterator) return Cursor is abstract; function Previous (Object : Reversible_Iterator; Position : Cursor) return Cursor is abstract; 5/3 end Ada.Iterator_Interfaces; 6/3 {AI05-0139-2} An iterator type is a type descended from the Forward_Iterator interface from some instance of Ada.Iterator_Interfaces. A reversible iterator type is a type descended from the Reversible_Iterator interface from some instance of Ada.Iterator_Interfaces. An iterator object is an object of an iterator type. A reversible iterator object is an object of a reversible iterator type. The formal subtype Cursor from the associated instance of Ada.Iterator_Interfaces is the iteration cursor subtype for the iterator type. 7/3 {AI05-0139-2} {AI05-0292-1} The following type-related operational aspects may be specified for an indexable container type T (see 4.1.6): 8/3 Default_Iterator This aspect is specified by a name that denotes exactly one function declared immediately within the same declaration list in which T is declared, whose first parameter is of type T or T'Class or an access parameter whose designated type is type T or T'Class, whose other parameters, if any, have default expressions, and whose result type is an iterator type. This function is the default iterator function for T. Its result subtype is the default iterator subtype for T. The iteration cursor subtype for the default iterator subtype is the default cursor subtype for T. 8.a/3 Aspect Description for Default_Iterator: Default iterator to be used in for loops. 9/3 Iterator_Element This aspect is specified by a name that denotes a subtype. This is the default element subtype for T. 9.a/3 Aspect Description for Iterator_Element: Element type to be used for user-defined iterators. 10/3 These aspects are inherited by descendants of type T (including T'Class). 11/3 {AI05-0139-2} {AI05-0292-1} An iterable container type is an indexable container type with specified Default_Iterator and Iterator_Element aspects. A reversible iterable container type is an iterable container type with the default iterator type being a reversible iterator type. An iterable container object is an object of an iterable container type. A reversible iterable container object is an object of a reversible iterable container type. 11.a.1/3 Glossary entry: An iterable container type is one that has user-defined behavior for iteration, via the Default_Iterator and Iterator_Element aspects. 11.1/4 {AI12-0138-1} The Default_Iterator and Iterator_Element aspects are nonoverridable (see 13.1.1). 11.a/4 Reason: This ensures that all descendants of an iterable container type have aspects with the same properties. This prevents generic contract problems with formal derived types. Legality Rules 12/3 {AI05-0139-2} {AI05-0292-1} The Constant_Indexing aspect (if any) of an iterable container type T shall denote exactly one function with the following properties: 13/3 * the result type of the function is covered by the default element type of T or is a reference type (see 4.1.5) with an access discriminant designating a type covered by the default element type of T; 14/3 * the type of the second parameter of the function covers the default cursor type for T; 15/3 * if there are more than two parameters, the additional parameters all have default expressions. 16/3 This function (if any) is the default constant indexing function for T. 16.a/3 Ramification: This does not mean that Constant_Indexing has to designate only one subprogram, only that there is only one routine that meets all of these properties. There can be other routines designated by Constant_Indexing, but they cannot have the profile described above. For instance, map containers have a version of Constant_Indexing that takes a key instead of a cursor; this is allowed. 17/3 {AI05-0139-2} {AI05-0292-1} The Variable_Indexing aspect (if any) of an iterable container type T shall denote exactly one function with the following properties: 18/3 * the result type of the function is a reference type (see 4.1.5) with an access discriminant designating a type covered by the default element type of T; 19/3 * the type of the second parameter of the function covers the default cursor type for T; 20/3 * if there are more than two parameters, the additional parameters all have default expressions. 21/3 This function (if any) is the default variable indexing function for T. Extensions to Ada 2005 21.a/3 {AI05-0139-2} User-defined iterator types are new in Ada 2012. Incompatibilities With Ada 2012 21.b/4 {AI12-0138-1} Corrigendum: Defined Default_Iterator and Iterator_Element to be nonoveridable, which makes redefinitions and hiding of these aspects illegal. It's possible that some program could violate one of these new restrictions, but in most cases this can easily be worked around by using overriding rather than redefinition. 5.5.2 Generalized Loop Iteration 1/3 {AI05-0139-2} Generalized forms of loop iteration are provided by an iterator_specification. Syntax 2/3 {AI05-0139-2} {AI05-0292-1} iterator_specification ::= defining_identifier in [reverse] iterator_name | defining_identifier [: subtype_indication ] of [reverse] iterable_name Name Resolution Rules 3/3 {AI05-0139-2} {AI05-0292-1} For the first form of iterator_specification, called a generalized iterator, the expected type for the iterator_name is any iterator type. For the second form of iterator_specification, the expected type for the iterable_name is any array or iterable container type. If the iterable_name denotes an array object, the iterator_specification is called an array component iterator; otherwise it is called a container element iterator. 3.a.1/3 Glossary entry: An iterator is a construct that is used to loop over the elements of an array or container. Iterators may be user defined, and may perform arbitrary computations to access elements from a container. Legality Rules 4/3 {AI05-0139-2} If the reserved word reverse appears, the iterator_specification is a reverse iterator; otherwise it is a forward iterator. In a reverse generalized iterator, the iterator_name shall be of a reversible iterator type. In a reverse container element iterator, the default iterator type for the type of the iterable_name shall be a reversible iterator type. 5/4 {AI05-0139-2} {AI12-0151-1} The subtype defined by the subtype_indication, if any, of an array component iterator shall statically match the component subtype of the type of the iterable_name. The subtype defined by the subtype_indication, if any, of a container element iterator shall statically match the default element subtype for the type of the iterable_name. 6/3 {AI05-0139-2} In a container element iterator whose iterable_name has type T, if the iterable_name denotes a constant or the Variable_Indexing aspect is not specified for T, then the Constant_Indexing aspect shall be specified for T. 6.1/4 {AI12-0047-1} The iterator_name or iterable_name of an iterator_specification shall not denote a subcomponent that depends on discriminants of an object whose nominal subtype is unconstrained, unless the object is known to be constrained. 6.a/4 Reason: This is the same rule that applies to renames; it serves the same purpose of preventing the object from disappearing while the iterator is still using it. 6.2/4 {AI12-0120-1} A container element iterator is illegal if the call of the default iterator function that creates the loop iterator (see below) is illegal. 6.b/4 Ramification: This can happen if the parameter to the default iterator function is in out and the iterable_name is a constant. The wording applies to any reason that the call would be illegal, as it's possible that one of the default parameters would be illegal, or that some accessibility check would fail. 6.3/4 {AI12-0120-1} A generalized iterator is illegal if the iteration cursor subtype of the iterator_name is a limited type at the point of the generalized iterator. A container element iterator is illegal if the default cursor subtype of the type of the iterable_name is a limited type at the point of the container element iterator. 6.c/4 Reason: If the cursor type is limited, the assignment to the loop parameter for a generalized iterator would be illegal. The same is true for a container element iterator. We have to say "at the point of the iterator" as the limitedness of a type can change due to visibility. Static Semantics 7/3 {AI05-0139-2} {AI05-0269-1} {AI05-0292-1} An iterator_specification declares a loop parameter. In a generalized iterator, the nominal subtype of the loop parameter is the iteration cursor subtype. In an array component iterator or a container element iterator, if a subtype_indication is present, it determines the nominal subtype of the loop parameter. In an array component iterator, if a subtype_indication is not present, the nominal subtype of the loop parameter is the component subtype of the type of the iterable_name. In a container element iterator, if a subtype_indication is not present, the nominal subtype of the loop parameter is the default element subtype for the type of the iterable_name. 8/3 {AI05-0139-2} {AI05-0292-1} In a generalized iterator, the loop parameter is a constant. In an array component iterator, the loop parameter is a constant if the iterable_name denotes a constant; otherwise it denotes a variable. In a container element iterator, the loop parameter is a constant if the iterable_name denotes a constant, or if the Variable_Indexing aspect is not specified for the type of the iterable_name; otherwise it is a variable. 8.a/4 Ramification: {AI12-0093-1} The loop parameter of a generalized iterator has the same accessibility as the loop statement. This means that the loop parameter object is finalized when the loop statement is left. (It also may be finalized as part of assigning a new value to the loop parameter.) For array component iterators and container element iterators, the loop parameter directly denotes an element of the array or container and has the accessibility of the associated array or container. Dynamic Semantics 9/3 {AI05-0139-2} For the execution of a loop_statement with an iterator_specification, the iterator_specification is first elaborated. This elaboration elaborates the subtype_indication, if any. 10/3 {AI05-0139-2} For a generalized iterator, the loop parameter is created, the iterator_name is evaluated, and the denoted iterator object becomes the loop iterator. In a forward generalized iterator, the operation First of the iterator type is called on the loop iterator, to produce the initial value for the loop parameter. If the result of calling Has_Element on the initial value is False, then the execution of the loop_statement is complete. Otherwise, the sequence_of_statements is executed and then the Next operation of the iterator type is called with the loop iterator and the current value of the loop parameter to produce the next value to be assigned to the loop parameter. This repeats until the result of calling Has_Element on the loop parameter is False, or the loop is left as a consequence of a transfer of control. For a reverse generalized iterator, the operations Last and Previous are called rather than First and Next. 10.a/4 Ramification: {AI12-0093-1} The loop parameter of a generalized iterator is a variable of which the user only has a constant view. It follows the normal rules for a variable of its nominal subtype. In particular, if the nominal subtype is indefinite, the variable is constrained by its initial value. Similarly, if the nominal subtype is class-wide, the variable (like all variables) has the tag of the initial value. Constraint_Error may be raised by a subsequent iteration if Next or Previous return an object with a different tag or constraint. 11/3 {AI05-0139-2} {AI05-0292-1} For an array component iterator, the iterable_name is evaluated and the denoted array object becomes the array for the loop. If the array for the loop is a null array, then the execution of the loop_statement is complete. Otherwise, the sequence_of_statements is executed with the loop parameter denoting each component of the array for the loop, using a canonical order of components, which is last dimension varying fastest (unless the array has convention Fortran, in which case it is first dimension varying fastest). For a forward array component iterator, the iteration starts with the component whose index values are each the first in their index range, and continues in the canonical order. For a reverse array component iterator, the iteration starts with the component whose index values are each the last in their index range, and continues in the reverse of the canonical order. The loop iteration proceeds until the sequence_of_statements has been executed for each component of the array for the loop, or until the loop is left as a consequence of a transfer of control. 12/3 {AI05-0139-2} {AI05-0292-1} For a container element iterator, the iterable_name is evaluated and the denoted iterable container object becomes the iterable container object for the loop. The default iterator function for the type of the iterable container object for the loop is called on the iterable container object and the result is the loop iterator. An object of the default cursor subtype is created (the loop cursor). 13/3 {AI05-0139-2} {AI05-0292-1} For a forward container element iterator, the operation First of the iterator type is called on the loop iterator, to produce the initial value for the loop cursor. If the result of calling Has_Element on the initial value is False, then the execution of the loop_statement is complete. Otherwise, the sequence_of_statements is executed with the loop parameter denoting an indexing (see 4.1.6) into the iterable container object for the loop, with the only parameter to the indexing being the current value of the loop cursor; then the Next operation of the iterator type is called with the loop iterator and the loop cursor to produce the next value to be assigned to the loop cursor. This repeats until the result of calling Has_Element on the loop cursor is False, or until the loop is left as a consequence of a transfer of control. For a reverse container element iterator, the operations Last and Previous are called rather than First and Next. If the loop parameter is a constant (see above), then the indexing uses the default constant indexing function for the type of the iterable container object for the loop; otherwise it uses the default variable indexing function. 14/4 {AI12-0120-1} Any exception propagated by the execution of a generalized iterator or container element iterator is propagated by the immediately enclosing loop statement. 14.a/4 Ramification: This text covers exceptions raised by called functions that make up the execution of the iterator as well as exceptions raised by the assignment to the loop parameter or cursor. Examples 15/3 {AI05-0269-1} -- Array component iterator example: for Element of Board loop -- See 3.6.1. Element := Element * 2.0; -- Double each element of Board, a two-dimensional array. end loop; 16/3 {AI05-0268-1} For examples of use of generalized iterators, see A.18.32 and the corresponding container packages in A.18.2 and A.18.3. Extensions to Ada 2005 16.a/3 {AI05-0139-2} Generalized forms of loop iteration are new. Incompatibilities With Ada 2012 16.b/4 {AI12-0047-1} Corrigendum: Added a rule to ensure that the object being iterated cannot be a component that could disappear before the loop completes. This could be incompatible by making a loop that was legal (and worked correctly, so long as the enclosing object is not modified during the loop) from the original Ada 2012 illegal in corrected Ada 2012. Such loops should be pretty rare, especially as these iterator forms are new to Ada 2012. 16.c/4 {AI12-0120-1} Corrigendum: Added rules to reject loops if the call to the default iterator function for a container element iterator is illegal, or if the cursor type of an iterator is limited. These are formally incompatible with original Ada 2012, but as it's unlikely that any Ada 2012 compiler ever allowed the illegal usages in an expansion of a loop (it's much more likely that they would have just caused an internal error in the compiler), this should have no effect in practice. 16.d/4 {AI12-0151-1} Corrigendum: Added a requirement that the given subtype statically match the subtype of the element or component for a component element iterator or array component iterator. Original Ada 2012 text allowed any type that covered the subtype of the element or component, but that led to questions of what the meaning was if they are different. In this case, the element is essentially a renaming of the container element, and it doesn't make sense for the constraints to be different. Ignoring explicitly defined constraints in renames is a mistake that we don't want to continue, thus we require static matching. This means that some programs might be illegal, but those programs were misleading at best, and potentially would raise unexpected exceptions because the element values might have been invalid or abnormal with respect to the declared constraint. Wording Changes from Ada 2012 16.e/4 {AI12-0120-1} Corrigendum: Added wording to specify that a loop propagates any exceptions propagated by the execution of an iterator. Since that's what naturally would happen from a macro-style expansion of the parts of an iterator, and no other interpretation makes sense given the way the rest of Ada works, we consider it so unlikely that any Ada 2012 implementation ever did anything else that we don't document this as a possible inconsistency. 5.6 Block Statements 1 [A block_statement encloses a handled_sequence_of_statements optionally preceded by a declarative_part.] Syntax 2 block_statement ::= [block_statement_identifier:] [declare declarative_part] begin handled_sequence_of_statements end [block_identifier]; 3 If a block_statement has a block_statement_identifier, then the identifier shall be repeated after the end; otherwise, there shall not be an identifier after the end. Static Semantics 4 A block_statement that has no explicit declarative_part has an implicit empty declarative_part. 4.a Ramification: Thus, other rules can always refer to the declarative_part of a block_statement. Dynamic Semantics 5 The execution of a block_statement consists of the elaboration of its declarative_part followed by the execution of its handled_sequence_of_statements. Examples 6 Example of a block statement with a local variable: 7 Swap: declare Temp : Integer; begin Temp := V; V := U; U := Temp; end Swap; 7.a Ramification: If task objects are declared within a block_statement whose execution is completed, the block_statement is not left until all its dependent tasks are terminated (see 7.6). This rule applies to completion caused by a transfer of control. 7.b Within a block_statement, the block name can be used in expanded names denoting local entities such as Swap.Temp in the above example (see 4.1.3). Wording Changes from Ada 83 7.c The syntax rule for block_statement now uses the syntactic category handled_sequence_of_statements. 5.7 Exit Statements 1 [An exit_statement is used to complete the execution of an enclosing loop_statement; the completion is conditional if the exit_statement includes a condition.] Syntax 2 exit_statement ::= exit [loop_name] [when condition]; Name Resolution Rules 3 The loop_name, if any, in an exit_statement shall resolve to denote a loop_statement. Legality Rules 4 Each exit_statement applies to a loop_statement; this is the loop_statement being exited. An exit_statement with a name is only allowed within the loop_statement denoted by the name, and applies to that loop_- statement. An exit_statement without a name is only allowed within a loop_- statement, and applies to the innermost enclosing one. An exit_statement that applies to a given loop_statement shall not appear within a body or accept_- statement, if this construct is itself enclosed by the given loop_statement. Dynamic Semantics 5 For the execution of an exit_statement, the condition, if present, is first evaluated. If the value of the condition is True, or if there is no condition, a transfer of control is done to complete the loop_statement. If the value of the condition is False, no transfer of control takes place. NOTES 6 8 Several nested loops can be exited by an exit_statement that names the outer loop. Examples 7 Examples of loops with exit statements: 8 for N in 1 .. Max_Num_Items loop Get_New_Item(New_Item); Merge_Item(New_Item, Storage_File); exit when New_Item = Terminal_Item; end loop; 9 Main_Cycle: loop -- initial statements exit Main_Cycle when Found; -- final statements end loop Main_Cycle; 5.8 Goto Statements 1 [A goto_statement specifies an explicit transfer of control from this statement to a target statement with a given label.] Syntax 2 goto_statement ::= goto label_name; Name Resolution Rules 3 The label_name shall resolve to denote a label; the statement with that label is the target statement. Legality Rules 4 The innermost sequence_of_statements that encloses the target statement shall also enclose the goto_statement. Furthermore, if a goto_statement is enclosed by an accept_statement or a body, then the target statement shall not be outside this enclosing construct. 4.a Ramification: The goto_statement can be a statement of an inner sequence_. 4.b It follows from the second rule that if the target statement is enclosed by such a construct, then the goto_statement cannot be outside. Dynamic Semantics 5 The execution of a goto_statement transfers control to the target statement, completing the execution of any compound_statement that encloses the goto_statement but does not enclose the target. NOTES 6 9 The above rules allow transfer of control to a statement of an enclosing sequence_of_statements but not the reverse. Similarly, they prohibit transfers of control such as between alternatives of a case_statement, if_statement, or select_statement; between exception_handlers; or from an exception_handler of a handled_sequence_of_statements back to its sequence_of_statements. Examples 7 Example of a loop containing a goto statement: 8 <> for I in 1 .. N-1 loop if A(I) > A(I+1) then Exchange(A(I), A(I+1)); goto Sort; end if; end loop;