NetNews Usenet Archive 1993 #3/NN_1993_3.iso/spool/comp/lang/cplus/19953

Text File | 1993-01-28 | 4.3 KB | 111 lines

Newsgroups: comp.lang.c++
Path: sparky!uunet!digex.com!intercon!udel!bogus.sura.net!howland.reston.ans.net!spool.mu.edu!sdd.hp.com!saimiri.primate.wisc.edu!ames!ads.com!saturn!doug
From: [e-mail] (Doug Morgan)
Subject: Re: Question About Code In ARM
In-Reply-To: [e-mail]'s message of Tue, 26 Jan 1993 22:07:53 GMT
Message-ID: <[e-mail]>
Sender: [e-mail] (USENET News)
Organization: Advanced Decision Systems, Mountain View, CA 94043, +1 (415)
    960-7300
References: <[e-mail]>
    <[e-mail]>
    <[e-mail]> <[e-mail]>
Date: Wed, 27 Jan 1993 17:49:38 GMT
Lines: 95

In article <[e-mail]> [e-mail] (John MAX Skaller) writes:
   In article <[e-mail]> [e-mail] (Doug Morgan) writes:
   >In article <[e-mail]> [e-mail] (John MAX Skaller) writes:
   >   In article <[e-mail]> [e-mail] (Denys Chen) writes:
   >   >(In another word, can we remove ", public virtual W" ?)
   >   >                                 ^^^^^^^^^^^^^^^^^^
   >
   >       Depends. The requirement "you can only initialise an
   >   immediate base"...
   >
   >... In any case, someone (I think it was Stephen Clamage) said
   >the (quasi-implied) restriction has been tossed from the latest
   >version of the proposed standard.

       Better not be tossed out: its an important restriction.

Tossed out was too strong.  Modified to allow mem-initialization of
any virtual base class would probably be more accurate.  Here's the
article I was remembering.  I think it means the restriction on being
able to mem-initialize only immediate base classes is tossed out in
the case of virtual base classes.

>From [e-mail] Mon Jan 11 20:54:30 1993
>Newsgroups: comp.lang.c++
>From: [e-mail] (Steve Clamage)
>Subject: Re: Restatement: Virtual base class & reference initialisation
>Organization: TauMetric Corporation
>Date: Mon, 11 Jan 1993 19:02:33 GMT
>
>[e-mail] (A. J. Piper) writes:
>
>|My original question on this did not include an example which is probably why I
>|was misunderstood. The scenario is this:
>|
>|class A {
>|        const int& a;
>|        A(const int& i) : a(i) {}
>|        A() : a(0) {}
>|};
>|
>|class B : virtual public A {
>|        B(const int &i) : A(i) {}
>|};
>|
>|class C : public B {
>|        C(const int& i) : B(i) {}
>|};
>|
>|The reason I asked is that if you create an object of type C then B's
>|initialisation of A is ignored and the default constructor is used instead ...
>
>That's correct.
>
>|So how do I initialise the reference ? Well cfront
>|3.0.1 allows me to access the constructor of A directly from the constructor
>|of C so I can write:
>|
>|class C : public B {
>|        C(const int& i) : B(i), A(i) {}
>|};
>|
>|Which solves my original question.
>
>Right.  This area of the language changed, and older compilers might
>not work this way.  Relatively recent compilers should support this
>code, and so should all future compilers.
>--
>
>Steve Clamage, TauMetric Corp, [e-mail]

   >
   >Hopefully the sections on virtual base classes for the final standard
   >will be heavily edited by someone with a nack for logical expression
   >in English.  Those sections are really a mess in the ARM.

       Even better, the rule for virtual base initialisation
   could be improved so you dont have to init every VB in the complete
   object .. a real mess when you're mixing 20 VB together.

Very much agreed.  C++ could do with a general upgrade to support
large scale use of virtual base classes.  Current C++ facilities are
very clumsy for handling problems much past the toy level.

   --
   ;----------------------------------------------------------------------
       JOHN (MAX) SKALLER,         [e-mail]
       Maxtal Pty Ltd, 6 MacKay St ASHFIELD, NSW 2131, AUSTRALIA
   ;------ SCIENTIFIC AND ENGINEERING SOFTWARE ---ph:  2 799 8223 --------

Doug
--------------------------------------------------------------------
Doug Morgan, [e-mail], [phone]
Advanced Decision Systems (a division of Booz-Allen & Hamilton Inc.)
1500 Plymouth St., Mountain View, CA 94043-1230
--------------------------------------------------------------------