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,cee84258116182b5,start
X-Google-Attributes: gid103376,public
From: akonstan@news.cs.columbia.edu (Alexander V. Konstantinou)
Subject: Formal Packages definition limitation
Date: 1996/09/13
Message-ID: <51cofa$ku4@herald.cs.columbia.edu>#1/1
X-Deja-AN: 180490773
organization: Columbia University Department of Computer Science
newsgroups: comp.lang.ada
Date: 1996-09-13T00:00:00+00:00
List-Id: <comp.lang.ada>


Consider a generic procedure Iter_Swap (swaps iterators)

generic
  with package Iterators1 is new Forward_Iterators(<>);
  with package Iterators2 is new Forward_Iterators(<>);
procedure Iter_Swap(A: Iterators1.Iterator; B: Iterators2.Iterator);

Now, this definition is too general since I would like to restrict
that 
	Iterators1.Value_Type = Iterators2.Value_Type

A reasonable way to do this (you can argue about the syntax) would be :

generic
  with package Iterators1 is new Forward_Iterators(<>);
  with package Iterators2 is new Forward_Iterators
    (Value_Type=>Iterators1.Value_Type);
procedure Iter_Swap(A: Iterators1.Iterator; B: Iterators2.Iterator);

I'd want the compiler to accept any instantiation of Forward_Iterators
with the limitation that its value type should be Iterators1.Value_Type.

Alas the reference manual defines (Sec. 12.7):

    formal_package_actual_part ::= (<>) | [generic_actual_part]

leaving no room for partial generic_actual_parts.  Any ideas why this
was not allowed ?  At present I am forced to pass two conversion
functions, but that is clearly not a good solution.

Alexander V. Konstantinou
akonstan@cs.columbia.edu
-- 
--
Alexander V. Konstantinou              http://www.cs.columbia.edu/~akonstan/
akonstan@cs.columbia.edu               akonstan@acm.org