Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!usc!rpi!zaphod.mps.ohio-state.edu!mips!sgi!shinobu!odin!sgihub!dragon!xanadu.wpd.sgi.com!pal
From: pal@xanadu.wpd.sgi.com (Anil Pal)
Newsgroups: comp.lang.c++
Subject: Re: Constructor question
Message-ID: <1991Apr15.164435.17867@dragon.wpd.sgi.com>
Date: 15 Apr 91 16:44:35 GMT
References: <1991Apr2.110623.22219@and.cs.liv.ac.uk> <20164@alice.att.com> <17400@sunquest.UUCP> <20204@alice.att.com>
Sender: news@dragon.wpd.sgi.com (CNews Account)
Reply-To: pal@wpd.sgi.com
Organization: Silicon Graphics, Inc.
Lines: 35
Cc: ark@alice.att.com

In article <20204@alice.att.com>, ark@alice.att.com (Andrew Koenig) writes:
|> 
|> The original question was whether it was possible for one constructor
|> of a class to call another constructor to help it construct the
|> same object; the answer to that question is `no.'

How about calling operator new using the placement syntax (and placing the
object at "this")?

Thus:

MyClass::MyClass(/* some parameters */)
{
    // first call basic constructor
    new (this) MyClass(/* other parameters */);

    // Then do stuff specific to this constructor
    ...

}

On reading the ARM, it appears that one must define an appropriate
operator new in order to use placement syntax.  This is fairly trivial
in this case:

    void*	operator new(size_t, void* p)	{ return p; }

I had originally thought that the pre-defined global operator new
included placement syntax, but the ARM indicates that I am mistaken.
Is there a reason why the definition above could (or should) not be
included in the language?

-- 
Anil A. Pal,	Silicon Graphics, Inc.
pal@wpd.sgi.com	(415)-335-7279