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=-1.9 required=5.0 tests=BAYES_00,FREEMAIL_FROM
	autolearn=unavailable autolearn_force=no version=3.4.4
X-Received: by 2002:ac8:758a:: with SMTP id
 s10mr14316316qtq.283.1581419707671;
        Tue, 11 Feb 2020 03:15:07 -0800 (PST)
X-Received: by 2002:a05:6808:4c7:: with SMTP id
 a7mr2529417oie.83.1581419707406;
 Tue, 11 Feb 2020 03:15:07 -0800 (PST)
Path: 
 eternal-september.org!reader01.eternal-september.org!feeder.eternal-september.org!news.gegeweb.eu!gegeweb.org!usenet-fr.net!proxad.net!feeder1-2.proxad.net!209.85.160.216.MISMATCH!news-out.google.com!nntp.google.com!postnews.google.com!google-groups.googlegroups.com!not-for-mail
Newsgroups: comp.lang.ada
Date: Tue, 11 Feb 2020 03:15:07 -0800 (PST)
Complaints-To: groups-abuse@google.com
Injection-Info: google-groups.googlegroups.com; posting-host=173.225.51.130;
 posting-account=ShYTIAoAAABytvcS76ZrG9GdaV-nXYKy
NNTP-Posting-Host: 173.225.51.130
User-Agent: G2/1.0
MIME-Version: 1.0
Message-ID: <7b1e3204-6a95-4d33-a967-a458566208f2@googlegroups.com>
Subject: Generic Parent Specification
From: sbelmont700@gmail.com
Injection-Date: Tue, 11 Feb 2020 11:15:07 +0000
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Xref: reader01.eternal-september.org comp.lang.ada:58028
Date: 2020-02-11T03:15:07-08:00
List-Id: <comp.lang.ada>

Is there any incantation for generics that allows you to specify that one a=
ctual generic package must be a child of another?  Suppose a generic packag=
e along with a child, and then a third generic that withs in them both, e.g=
:

generic
package X is
  type T1 is new Integer;
end X;

generic
package X.Y is
  subtype T2 is T1;
end X.Y;

generic
  with package K1 is new X (<>);
  with package K2 is new X.Y (<>);
package P is
  o1 : K2.T2 :=3D 0;
  o2 : K1.T1 :=3D o1; -- error!
end P;


The error seems sensible enough, since the actual for K1 might not actually=
 be a "real" parent of the actual for K2, and the types might indeed be inc=
ompatible.  But in cases where K2 is a legit child of K1, the conversion ou=
ght to be fine.  Conceptually I want something like this:

generic
  with package K1 is new X (<>);
  with package K2 is new K1.Y (<>);  -- K2 must be a instantiation of X.Y f=
rom K1

but either I can't find the right spell to do this, or it's not possible.  =
Any ideas?

-sb