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
	autolearn=unavailable autolearn_force=no version=3.4.4
Path: 
 eternal-september.org!reader01.eternal-september.org!reader02.eternal-september.org!news.eternal-september.org!mx02.eternal-september.org!feeder.eternal-september.org!news.swapon.de!fu-berlin.de!newsfeed-00.mathworks.com!nntp.TheWorld.com!.POSTED!not-for-mail
From: Robert A Duff <bobduff@shell01.TheWorld.com>
Newsgroups: comp.lang.ada
Subject: Re: Multiple procedures in the same adb file?
Date: Mon, 12 Jan 2015 09:57:58 -0500
Organization: The World Public Access UNIX, Brookline, MA
Message-ID: <wcck30s2gi1.fsf@shell01.TheWorld.com>
References: <5fc3a42d-f922-4e34-ab30-59613b980fb6@googlegroups.com>
 <wcc1tn1m6yr.fsf@shell01.TheWorld.com> <chg8qqFe9apU1@mid.individual.net>
 <wcc1tn0x1j6.fsf@shell01.TheWorld.com> <m902gd$sp3$1@dont-email.me>
NNTP-Posting-Host: shell01.theworld.com
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
X-Trace: pcls7.std.com 1421074685 9121 192.74.137.71 (12 Jan 2015 14:58:05
 GMT)
X-Complaints-To: abuse@TheWorld.com
NNTP-Posting-Date: Mon, 12 Jan 2015 14:58:05 +0000 (UTC)
User-Agent: Gnus/5.1008 (Gnus v5.10.8) Emacs/21.3 (irix)
Cancel-Lock: sha1:Cc1tm0GIw4GXRTK8v6eAWYqD4Wk=
Xref: news.eternal-september.org comp.lang.ada:24550
Date: 2015-01-12T09:57:58-05:00
List-Id: <comp.lang.ada>

"J-P. Rosen" <rosen@adalog.fr> writes:

> Le 12/01/2015 01:53, Robert A Duff a écrit :
>> This happened to me:  Existing code:
>> 
>>     function Some_Root_Package.Some_Function(...) return String;
>> 
>> Now I happen to know that it always returns String(1..20).
>> And returning known-length strings is more efficient,
>> and this one is a bottleneck, so it's worth changing to
>> something like:
>> 
>>     subtype String_20 is String(1..20);
>>     function Some_Function(...) return String_20;
>> 
>> But that doesn't work without adding a new package, which breaks
>> compatibility.  And this was a widely-used library, so I couldn't do
>> that.  And String_20 really doesn't belong in Some_Root_Package,
>> nor anywhere else than the package that Some_Function is (directly) in
>> (which didn't exist!).
>
> Put the function in a package, and declare Some_Function as:
> function Some_Function (...) renames New_Package.Some_Function;
>
> (For compatibility, new programs would use directly the function in the
> package).

I started to reply, "Good idea", but then I remembered it doesn't work.
A library unit can only rename another library unit.  If it did work,
I'd still prefer to use packages from the start.

I see AARM-10.1.1(14.a) explains why we have this restriction, in part
"because they wouldn't be particularly useful".  Probably written by
somebody who doesn't like library subprograms.  ;-)

- Bob