Re: [boost] Iterator adaptors return a value from operator[]?

From: Anthony Williams (anthwil_at_[hidden])
Date: 2002-08-19 03:49:59

David Abrahams writes:
 > 8c0_at_[hidden]><006a01c24528$8b790dd0$1d00a8c0_at_pdimov2><15709.1022.799000.496818_at_gargle.gargle.HOWL><006101c24619$40c34160$6501a8c0_at_[hidden]>
 > 	<005501c2461f$671a1bd0$1d00a8c0_at_pdimov2>
 > Subject: Re: [boost] Iterator adaptors return a value from operator[]?
 > Date: Sat, 17 Aug 2002 15:25:01 -0400
 > MIME-Version: 1.0
 > Content-Type: text/plain;
 > 	charset="Windows-1252"
 > Content-Transfer-Encoding: 7bit
 > X-Priority: 3
 > X-MSMail-Priority: Normal
 > X-Mailer: Microsoft Outlook Express 6.00.2600.0000
 > X-MIMEOLE: Produced By Microsoft MimeOLE V6.00.2600.0000
 > X-BeenThere: boost_at_[hidden]
 > X-Mailman-Version: 2.1b1
 > Precedence: bulk
 > List-Help: <mailto:boost-request_at_[hidden]?subject=help>
 > List-Post: <mailto:boost_at_[hidden]>
 > List-Subscribe: <http://lists.boost.org/mailman/listinfo.cgi/boost>,
 > 	<mailto:boost-request_at_[hidden]?subject=subscribe>
 > List-Id: Boost mailing list <boost.lists.boost.org>
 > List-Unsubscribe: <http://lists.boost.org/mailman/listinfo.cgi/boost>,
 > 	<mailto:boost-request_at_[hidden]?subject=unsubscribe>
 > List-Archive: http://lists.boost.org/MailArchives/boost/
 > Errors-To: boost-bounces_at_[hidden]
 > X-SMTP-HELO: heart-of-gold.osl.iu.edu
 > X-SMTP-MAIL-FROM: boost-bounces_at_[hidden]
 > X-SMTP-RCPT-TO: anthwil_at_[hidden]
 > X-SMTP-PEER-INFO: heart-of-gold.osl.iu.edu [129.79.245.244]
 > 
 > From: "Peter Dimov" <pdimov_at_[hidden]>
 > 
 > 
 > > From: "David Abrahams" <dave_at_[hidden]>
 > > >
 > > > We've thought about it, and haven't come up with an answer. The details
 > > > (like the fact that operator[] might be overloaded, which pretty much
 > > rules
 > > > out the approach above) are non-trivial. If /you/ can fill in the
 > details,
 > > > I'm interested. I've already tried and failed.
 > >
 > > For non-const iterators,
 > >
 > > yes test(value_type &);
 > > no test(...);
 > >
 > > test(a[i]);
 > >
 > > For const iterators, the technique doesn't work, but return by value
 > works
 > > adequately there.
 > 
 > That's not legal in the case where a[i] is a non-POD value type, is it? I
 > thought that passing a non-POD to a ... argument invokes undefined
 > behavior.

Only if you actually make the call...

sizeof(test(a[i])) is well-defined.

Anthony