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=-0.3 required=5.0 tests=BAYES_00,
	REPLYTO_WITHOUT_TO_CC autolearn=no autolearn_force=no version=3.4.4
X-Google-Thread: 103376,b563623c98f6fb66
X-Google-Attributes: gid103376,public
X-Google-Language: ENGLISH,ASCII-7-bit
Path: 
 g2news1.google.com!news4.google.com!news.germany.com!feeder.ecngs.de!ecngs!feeder2.ecngs.de!npeer.de.kpn-eurorings.net!npeer1.kpn.DE!newsfeed.arcor.de!newsspool1.arcor-online.net!news.arcor.de.POSTED!not-for-mail
From: "Dmitry A. Kazakov" <mailbox@dmitry-kazakov.de>
Subject: Re: GNAT GPL 2007 Problems with generics
Newsgroups: comp.lang.ada
User-Agent: 40tude_Dialog/2.0.15.1
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
Reply-To: mailbox@dmitry-kazakov.de
Organization: cbb software GmbH
References: <14839k5x0otsk.mv0vz7fjh95y$.dlg@40tude.net>
Date: Tue, 15 May 2007 22:43:33 +0200
Message-ID: <v6bw5ol9ahi3$.7kw2rmunw18x.dlg@40tude.net>
NNTP-Posting-Date: 15 May 2007 22:43:12 CEST
NNTP-Posting-Host: 0ffee452.newsspool1.arcor-online.net
X-Trace: 
 DXC=O@mOhiS:\\9kUFX=Y?aLP;ic==]BZ:af>4Fo<]lROoR1Fl8W>\BH3Y2N=gZ0LSBW:1DNcfSJ;bb[5IRnRBaCd<M>CRN^2`SW0J11hf;PM?@3E7
X-Complaints-To: usenet-abuse@arcor.de
Xref: g2news1.google.com comp.lang.ada:15801
Date: 2007-05-15T22:43:12+02:00
List-Id: <comp.lang.ada>

On Tue, 15 May 2007 16:14:37 +0200, Dmitry A. Kazakov wrote:

> Does anybody know anything about the following problem. When a generic
> package A instantiates another generic package B and refers to a subprogram
> C from its body, then under certain circumstances (I cannot figure out
> when), everything compiles, but then fails in gnatlink on "undefined
> reference to xxxx" From the cryptic name xxxx, it seems that the module it
> fails to find is the subprogram B.C (an instance in the instance of A)
> 
> The name xxxx has the pattern:
> 
> <name_of_instance_of_A>__<name_of_instance_of_B>__<name of C>__2X

OK, I have figured out the problem. The bug is related to the visibility
rules. It appears when C refers to a parameter of the type T declared in
some fourth package D. In the specification of B it does this as C (...; X
: T;...). In the body it does it as C (...; X : D.T; ...).

The problem disappears when both become X : T. [ As far as I can remember,
they had to be different to work around another bug in some earlier version
of GNAT. ]
 
-- 
Regards,
Dmitry A. Kazakov
http://www.dmitry-kazakov.de