public inbox for gcc-help@gcc.gnu.org
 help / color / mirror / Atom feed
From: Jonathan Wakely <jwakely.gcc@gmail.com>
To: Parmenides <mobile.parmenides@gmail.com>
Cc: gcc-help <gcc-help@gcc.gnu.org>
Subject: Re: Why can not use reference in operator function?
Date: Sun, 24 Nov 2013 18:19:00 -0000	[thread overview]
Message-ID: <CAH6eHdSdnUSiKeLEFaOMojHoVKMA01g1Huj3+2AEzrWLPmJqyg@mail.gmail.com> (raw)
In-Reply-To: <CAOXENUhNJbPC8F3ZYADpE970qOMxb4vOio+3c6LfV86hPV1evQ@mail.gmail.com>

On 23 November 2013 16:54, Parmenides <mobile.parmenides@gmail.com> wrote:
> Hi,
>
> I have the following code:
>
> #include <iostream>
> #include <cstring>
> using namespace std;
>
> class Int{
> int x;
>
> public:
> Int(int i = 0)
> {
> x = i;
> }
>
> ~Int()
> {
> }
>
> void show()
> {
> cout << x << endl;
> }
>
> friend Int operator+(Int &a, Int &b)  // Because 'b' is a reference to
> object rather than an object
> {
> return Int(a.x+b.x);
> }
> };
>
> int main()
> {
> Int i(3), j;
> j = i + Int(6);   // This can not call constructor.

This creates a temporary object and temporaries cannot bind to
non-const references.

> j.show();
>
> return 0;
> }
>
> Gcc issues error message, but VC++ 2010 compile it successfully. I

This is a well-known VC++ bug, it allows temporaries to bind to
non-const references. That does not conform to the C++ standard.

> tried to modify
>
>                 friend Int operator+(Int &a, Int &b)
>
> to
>
>                 friend Int operator+(Int a, Int b)
>
> both compiler can get it pass. Why Gcc does not want to convert a
> 'int' to a 'Int' object in the  '+' operator function when its second
> parameter is a reference to object rather than an objetc?

Because a temporary cannot bind to a non-const reference.

> I further modified
>
>                 j = i + Int(6);
>
> to
>
>                j = i + 6;
>
> both compiler can get it pass again. Therefore, I think it seem that
> Gcc encourage programmers to use implict conversion like 'i + 6'
> rather that explict conversion like 'i + Int(6)'. Is this right?

No, GCC encourages you to use const references if you want to bind to
temporaries, as required by the C++ standard.

Your operator should have been declared

                friend Int operator+(const Int& a, const Int& b)


> If
> so, does this practice comfore to the C++ standard?

G++ conforms to the standard in this regard.

  reply	other threads:[~2013-11-23 17:03 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2013-11-23 17:03 Parmenides
2013-11-24 18:19 ` Jonathan Wakely [this message]
     [not found]   ` <CAOXENUgy1FadKfQd6uv88qoYPUMR3z9sx083up3cWkbCNTFwcg@mail.gmail.com>
     [not found]     ` <CAH6eHdTHpCNtMb0EOrs034mifZ9j-r=CwyYgkTHM2J1uN6dmVg@mail.gmail.com>
2013-11-26 13:23       ` Parmenides
2013-11-26 13:41         ` Jonathan Wakely

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=CAH6eHdSdnUSiKeLEFaOMojHoVKMA01g1Huj3+2AEzrWLPmJqyg@mail.gmail.com \
    --to=jwakely.gcc@gmail.com \
    --cc=gcc-help@gcc.gnu.org \
    --cc=mobile.parmenides@gmail.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).