From: "Grein, Christoph" <christoph.grein@eurocopter.com>
Subject: Re: Ada2005 temp solo child
Date: Mon, 16 Sep 2002 07:03:48 +0200 (MET DST)
Date: 2002-09-16T07:03:48+02:00 [thread overview]
Message-ID: <mailman.1032152942.7118.comp.lang.ada@ada.eu.org> (raw)
> ... The compiler has to look for the
> declarations in the parent (in order to determine if the child name is
> legal). Ada makes you declare almost everything before use. Why would
> this case be different? It doesn't feel much different to me than
> saying:
>
> Int := 10;
>
> and complaining that the compiler shouldn't need a declaration of Int.
Another point to consider:
Ada95 removed unneeded library bodies. They were a source of problems when the
spec was recompiled and the body forgotten.
Your proposal would add a new source of errors of this kind. Imagine there is a
parent that is somehow lost. The compiler would not know that there should be a
parent.
> That's what empty parent packages are for. It's only three lines of code!
So where is the problem in declaring empty packages?
next reply other threads:[~2002-09-16 5:03 UTC|newest]
Thread overview: 12+ messages / expand[flat|nested] mbox.gz Atom feed top
2002-09-16 5:03 Grein, Christoph [this message]
2002-09-17 14:06 ` Ada2005 temp solo child Peter Hermann
-- strict thread matches above, loose matches on Subject: below --
2002-09-16 4:19 Grein, Christoph
2002-09-13 8:59 Ada2005 temp solo child (was: Re: The Dreaded "Missing Subunits" Grein, Christoph
2002-09-13 10:24 ` Ada2005 temp solo child ica2ph
2002-09-13 10:40 ` Preben Randhol
2002-09-13 11:41 ` Peter Hermann
2002-09-13 13:50 ` Dr. Michael Paus
2002-09-13 15:29 ` Preben Randhol
2002-09-13 21:08 ` Randy Brukardt
2002-09-13 21:00 ` Stephen Leake
2002-09-13 16:06 ` Robert A Duff
2002-09-13 16:28 ` Preben Randhol
replies disabled
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox