The RM is not entirely clear on convention handling in a number of cases, and in particular, it is not clear on the convention to be given to anonymous access types in general, and in particular what is to be done for the case of anonymous access-to-subprogram.
In GNAT, we decide that if an explicit Convention is applied to an object or component, and its type is such an anonymous type, then the convention will apply to this anonymous type as well. This seems to make sense since it is anomolous in any case to have a different convention for an object and its type, and there is clearly no way to explicitly specify a convention for an anonymous type, since it doesn’t have a name to specify!
Furthermore, we decide that if a convention is applied to a record type, then this convention is inherited by any of its components that are of an anonymous access type which do not have an explicitly specified convention.
The following program shows these conventions in action:
package ConvComp is
type Foo is range 1 .. 10;
type T1 is record
A : access function (X : Foo) return Integer;
B : Integer;
end record;
pragma Convention (C, T1);
type T2 is record
A : access function (X : Foo) return Integer;
pragma Convention (C, A);
B : Integer;
end record;
pragma Convention (COBOL, T2);
type T3 is record
A : access function (X : Foo) return Integer;
pragma Convention (COBOL, A);
B : Integer;
end record;
pragma Convention (C, T3);
type T4 is record
A : access function (X : Foo) return Integer;
B : Integer;
end record;
pragma Convention (COBOL, T4);
function F (X : Foo) return Integer;
pragma Convention (C, F);
function F (X : Foo) return Integer is (13);
TV1 : T1 := (F'Access, 12); -- OK
TV2 : T2 := (F'Access, 13); -- OK
TV3 : T3 := (F'Access, 13); -- ERROR
|
>>> subprogram "F" has wrong convention
>>> does not match access to subprogram declared at line 17
38. TV4 : T4 := (F'Access, 13); -- ERROR
|
>>> subprogram "F" has wrong convention
>>> does not match access to subprogram declared at line 24
39. end ConvComp;