Object-oriented Software in Ada 95

Dynamic memory allocation 223
15.5.2 Hidden vs. visible storage in a class
The main benefit of this approach is that a client of the class does not need to recompile their code when the
storage structure of the class is changed. The client code only needs to be relinked with the new implementation
code. This would usually occur when a new improved class library is provided by a software supplier. Naturally,
this assumes that the interface with the library stays the same.
The pros and cons of the two approaches are:
Criteria Hidden storage Visible storage
Compilation efficiency Fewer resources required,
as only a recompile of the
class and then a re-link
need to be performed.
Greater as all units that use
the class need to be
recompiled.
Run-time efficiency
Client access to data
components of an object.
Worse as there is the
dynamic storage allocation
overhead.
None
No extra run-time
overhead.
None
Note:
The extra cost of re-compiling and re-linking all units of a program may be marginal when compared
with just re-linking.
15.6 Access value of a function
The access value of a function or procedure may also be taken. This allows a function or procedure to be passed as
a parameter to another function or procedure. For example, the procedure apply applies the function passed as a
parameter to all elements of the array. The implementation of this function is shown in the fragment of code
below:
type P_Fun is access function(Item:in Float) return Float;
type Vector is array ( Integer range ) of Float;
procedure Apply( F:in P_Fun; To:in out Vector ) is
begin
for I in To'range loop
To(I) := F( To(I) );
end loop;
end Apply;
Note:
The de-referencing is done automatically when the function f is called. This could have been done
explicitly with F.all( To(I) ). This explicit de-referencing is, however, required if the called
function or procedure has no parameters.
The first parameter to the procedure apply can be any function which has the signature:
function(Item:in Float) return Float;
Two such functions are:
function Square( F:in Float ) return Float is
begin
return F * F;
end Square;
function Cube( F:in Float ) return Float is
begin
return F * F * F;
end Cube;
© M A Smith - May not be reproduced without permission