comp.lang.ada
 help / color / mirror / Atom feed
From: "Rodrigo García" <rodrigo.garcia@epfl.ch>
Subject: Re: Problem Eliminating constructors
Date: Wed, 22 Jan 2003 16:26:34 +0100
Date: 2003-01-22T16:26:34+01:00	[thread overview]
Message-ID: <3E2EB82A.7080400@epfl.ch> (raw)
In-Reply-To: 3e2d9f6e$0$33923$bed64819@news.gradwell.net

Victor Porton wrote:
> In article <3E2D3245.2060903@epfl.ch>,
> 	Rodrigo Garc�a <rodrigo.garcia@epfl.ch> writes:
> 
>>>>Victor Porton wrote:
>>>>
>>>
>>>Oh, Rodrigo, I oversimplified the real situation. Initialization with
>>>just aggregates is possible only in the example, really I need namely
>>>constructor functions.
>>
>>   I see... Have you tried with controlled types?
> 
> 
> Do you mean to use type discriminants by Initialize procedure?

   Yes, that's the idea.

> I suspect this is a size overhead in current compilers (Gnat 3.14/3.15 
> for me).

   Well, I implemented our simple example with and without controlled 
types and it is not a significant overhead.

   In my opinion, the use of controlled types is the only "standard" 
(and the cleanest) way that Ada proposes for constructors/destructors.

Rodrigo




  reply	other threads:[~2003-01-22 15:26 UTC|newest]

Thread overview: 15+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2003-01-17  0:19 Problem Eliminating constructors Victor Porton
2003-01-17 15:21 ` John English
2003-01-17 21:10 ` Victor Porton
2003-01-18 15:19 ` Simon Wright
2003-01-20 11:39 ` Rodrigo García
2003-01-20 21:07 ` Victor Porton
2003-01-21  9:44   ` Dmitry A. Kazakov
2003-01-21 19:01     ` Martin Krischik
2003-01-21 11:43   ` Rodrigo García
2003-01-21 15:03   ` Stephen Leake
2003-01-21 13:09 ` Victor Porton
2003-01-22 15:26   ` Rodrigo García [this message]
2003-01-21 18:38 ` Victor Porton
2003-01-23 17:08   ` Stephen Leake
  -- strict thread matches above, loose matches on Subject: below --
2003-01-17  6:54 Grein, Christoph
replies disabled

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox