From mboxrd@z Thu Jan  1 00:00:00 1970
X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on polar.synack.me
X-Spam-Level: 
X-Spam-Status: No, score=-1.3 required=5.0 tests=BAYES_00,INVALID_MSGID
	autolearn=no autolearn_force=no version=3.4.4
X-Google-Language: ENGLISH,ASCII-7-bit
X-Google-Thread: 103376,dd2ec8d5747df63b,start
X-Google-Attributes: gid103376,public
From: mr101@ccsr.cam.ac.uk (Michael Roe)
Subject: Arrays of strings and Annex E
Date: 1999/03/30
Message-ID: <7dqkmu$h6o$1@pegasus.csx.cam.ac.uk>#1/1
X-Deja-AN: 460749028
Sender: mr101@kamloops.ccsr.cam.ac.uk (Michael Roe)
Organization: University of Cambridge, England
Newsgroups: comp.lang.ada
Date: 1999-03-30T00:00:00+00:00
List-Id: <comp.lang.ada>


Recently, I needed a write a server that return a variable length list of
variable-length strings as a result of a remote procedure call. If I had
been using ASN.1 and OSI protocols, I would have written something like
this:

   ListOfNames ::= SEQUENCE OF IA5String

Instead of ASN.1, I was using Ada and Annex E.
So I wrote something like this:

   with Ada.Strings.Unbounded;

   package String_RCI is

      pragma Remote_Call_Interface (String_RCI);

      type List_Of_Names is array (Integer range <>) of
         Ada.Strings.Unbounded.Unbounded_String;

      function Get_List return List_Of_Names;

   end String_RCI;


Unfortunately, this doesn't work: Unbounded_String is a private type which
can't be passed across a remote call interface in this way. I could have
used ordinary strings, but then all of the strings would have to be the
same length. For example:

   type List_Of_Names is array (Integer range <>) of 
      String (1 .. Name_Length);

I could have written something which replicates most of the
functionality of Ada.Strings.Unbounded, but in a way which can be passed
across a remote call interface. It's a shame that the language-defined
string types like Ada.Strings.Unbounded weren't designed with Annex E in
mind.

Mike

PS. At least, I *think* this is a problem with the language definition,
rather than a misfeature of a particular implementation; if I thought
this was a bug in GNAT or GLADE I'd have reported it in the usual way
rather than discussing it in comp.lang.ada.