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=ham
	autolearn_force=no version=3.4.4
X-Google-Thread: 103376,ab43ecf7e1a10743,start
X-Google-Attributes: gid103376,public
X-Google-Language: ENGLISH,ASCII-7-bit
Path: 
 g2news1.google.com!news2.google.com!fu-berlin.de!uni-berlin.de!not-for-mail
From: "Nick Roberts" <nick.roberts@acm.org>
Newsgroups: comp.lang.ada
Subject: Zero_Fill pragma
Date: Tue, 20 Jul 2004 17:22:27 +0100
Message-ID: <opsbf2fpb0p4pfvb@bram-2>
Mime-Version: 1.0
Content-Type: text/plain; format=flowed; delsp=yes; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
X-Trace: news.uni-berlin.de woEv6oE6XzKbEZNnWaPHRwLZTJfXF9lubqUrgbYQpu+6OCjV4=
User-Agent: Opera M2/7.51 (Win32, build 3798)
Xref: g2news1.google.com comp.lang.ada:2279
Date: 2004-07-20T17:22:27+01:00
List-Id: <comp.lang.ada>

Am I too late to make a proposal for a new pragma for
the Ada 2005 Amendment?

It would have the syntax:

    pragma Zero_Fill(simple_name);

It would be applicable to an object or a subtype, and it
would operate on all objects of the given subtype, or on
the given object. The object must be of a composite type.

It would have the effect of causing all the bits in the
object which are unused to have a default initialisation
of 0.

I think the definition of 'in the object' would be any
bit numbered 0 to n-1, where n is the object's size (as
the Size attribute would report).

I think the definition of 'unused' would be any bit
which is not in any of the object's components. If the
place of the component is defined by a record
representation clause, that will define the bits used
by the component, otherwise the bits used by a component
will be chosen by the compiler.

For records, the effect can be achieved by introducing
'dummy' components, but this is a pain and can degrade
readability. For arrays, the effect cannot be achieved
by standard means (other than perhaps adding components
to the array's component subtype, if it is a record
subtype, to bulk the record up to the size of an
integral number of allocation units, and then using
pragma Pack).

Either way, I think the pragma would be useful in real
programs, and that it would be useful for the pragma to
be defined by the language standard. I'm a little
surprised existing compilers don't provide something
like this.

-- 
Nick Roberts