Re: Question on array of strings

["Jeffrey R. Carter" <spam.jrcarter.not@spam.acm.org>]

Peter Mueller wrote:
> 
> I have some strings and depending of an enum I want to return a
> string. My code looks the following:
> 
> function Event_To_String(Event : in Events_Type) return String is
>    Ev1_Name : aliased constant String :="ev3434";
>    Ev2_Name : aliased constant String :="evEnd";
>    eventStrings : constant array (1..2) of access constant String :=
> (Ev1_Name'Access,Ev2_Name'Access);
> 
>    begin
> 
>    return EventStrings(Events_Type'Pos(Event));
> end Event_To_String;
> 
> The compiler says:
> 
> expected type "Standard.String"
> found type access to "Standard.String" defined ...
> 
> I think that my code returns the pointer to the string and not the
> string. But I don't know how to fix it.

You're correct. But first, a probable error:

'Pos for an enumeration type returns values starting at zero:

type E is (One, Two);

E'Pos (One) = 0
E'Pos (Two) = 1

Your array index range is "Integer range 1 .. 2". So, assuming Events_Type is 
the first-named subtype (the name used in the type declaration), if you ever 
called this with Events_Type'First, you would get Constraint_Error.

However, there is no reason to use Integer indices. An array may be indexed by 
any discrete subtype:

type Event_Name is array (Events_Type) of ...;

Then you can say

Eventstrings (Event)

As for your question, you dereference an access value by appending .all to it:

Eventstrings (Event).all

> Is there a better solution to create an array of strings and accessing
> them with an index?

For a discrete subtype like this, probably not. What you have is a mapping 
Events_Type => String. If you were mapping something non-discrete to String, 
than a map or similar associative data structure would be better.

It's also possible to store String values in [Un]Bounded_String objects; that 
would be better if the strings are not constant. It would also eliminate 
(explicit) use of access types and values.

-- 
Jeff Carter
"C's solution to this [variable-sized array parameters] has real
problems, and people who are complaining about safety definitely
have a point."
Dennis Ritchie
25