220 17114 <04122f67-f4cd-457c-902e-589091f90063@isocpp.org> article Path: news.gmane.org!not-for-mail From: jgottman6@gmail.com Newsgroups: gmane.comp.lang.c++.isocpp.proposals Subject: Re: virtual constexpr fields Date: Fri, 20 Mar 2015 18:39:26 -0700 (PDT) Lines: 150 Approved: news@gmane.org Message-ID: <04122f67-f4cd-457c-902e-589091f90063@isocpp.org> References: <532355906.19321.1426710347371.JavaMail.yahoo@mail.yahoo.com> Reply-To: std-proposals@isocpp.org NNTP-Posting-Host: plane.gmane.org Mime-Version: 1.0 Content-Type: multipart/mixed; boundary="----=_Part_1232_861976290.1426901966993" X-Trace: ger.gmane.org 1426901979 16354 80.91.229.3 (21 Mar 2015 01:39:39 GMT) X-Complaints-To: usenet@ger.gmane.org NNTP-Posting-Date: Sat, 21 Mar 2015 01:39:39 +0000 (UTC) Cc: c++std-ext@accu.org, ehsan@mozilla.com, botond_ballo@yahoo.ca, botond_ballo@yahoo.ca To: std-proposals@isocpp.org Original-X-From: std-proposals+bncBD7LHGWIQEPRBT4XWOUAKGQEQ5D6ZOY@isocpp.org Sat Mar 21 02:39:38 2015 Return-path: Envelope-to: gclcip-std-proposals@m.gmane.org Original-Received: from mail-ob0-f200.google.com ([209.85.214.200]) by plane.gmane.org with esmtp (Exim 4.69) (envelope-from ) id 1YZ8Nv-0002Ss-1q for gclcip-std-proposals@m.gmane.org; Sat, 21 Mar 2015 02:39:31 +0100 Original-Received: by obcva2 with SMTP id va2sf156212431obc.0 for ; Fri, 20 Mar 2015 18:39:29 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113; h=date:from:to:cc:message-id:in-reply-to:references:subject :mime-version:content-type:x-original-sender:reply-to:precedence :mailing-list:list-id:list-post:list-help:list-archive :list-subscribe:list-unsubscribe; bh=MIZT8S7tHTArt/wgdW1SkHNVBD90+VW0aXVKomSCDbs=; b=vOFdzNIdaNz5Pn/X3fg0aaYm8tiGttAVKhir1UJDdwEBrRhjEsGzrTqj19HE/d49MD s0GnNQc3Sdvn90zjkLyqVX1MnE2hXpjZuOVZCBXx1irYgZxOn39HwLCc7tGsGvmr2qBh Zbl3AeY/iZIRlUNI8/uEwHAw0k3m/tgG8LjnWXFsaLxrvzevfu/dMfWswsGunxdSjXGw SVyw1M3/4uCjHzpY1Mm/c39ZnlyHQoXyr6y0CwXUPt98d1PVzIFQ0x2xHvhVs91YJ142 +C5fQmmrfQoN0RTOTvZAVfHtI5VrqF7XTQ1IVTz6qs/ulXeMhA5ufn0R8v3bvNrrkEWV cOqg== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20130820; h=x-gm-message-state:date:from:to:cc:message-id:in-reply-to :references:subject:mime-version:content-type:x-original-sender :reply-to:precedence:mailing-list:list-id:list-post:list-help :list-archive:list-subscribe:list-unsubscribe; bh=MIZT8S7tHTArt/wgdW1SkHNVBD90+VW0aXVKomSCDbs=; b=O7OuZ/dlf0SfjpjUXFXzsxvVERyDyFdAO+dl2bufqZdvcxwkQ089RVFkpc1FsoRp/0 WoK3DXTQN5qErd+Kb8LV+pjjKtBTBNi8+5q5MMl8jaS8jtvh6sgIyexwGs8a5sd34oy/ +LZL09+CM5/Hf/2qNsoTSPa6kTGjsKH1cCvvHMEg1mY6CD065u2frO04eMozLR3NCawh FBAs4S7+PJEgwlumCFqcZzugG2hJ26i4/bH8l48TxJK7zHUGUUrjus2F9lM1BtM1FYah aAGnkti6bNx4TkimKMpo+jwgY1I0d1kF5vq1SiX3RalpKkmM3icX+pLT79zoARfX0JwI Qbfw== X-Gm-Message-State: ALoCoQkNI+PHFdJYazDruJwc6jUcE9Sz2RsLBNIOCayYdLDUxdKJYzXfoyKPo+s1XWync4TAxQbI X-Received: by 10.182.126.168 with SMTP id mz8mr96213525obb.9.1426901969692; Fri, 20 Mar 2015 18:39:29 -0700 (PDT) X-BeenThere: std-proposals@isocpp.org Original-Received: by 10.182.68.103 with SMTP id v7ls939283obt.57.gmail; Fri, 20 Mar 2015 18:39:27 -0700 (PDT) X-Received: by 10.182.131.135 with SMTP id om7mr709505obb.7.1426901967727; Fri, 20 Mar 2015 18:39:27 -0700 (PDT) In-Reply-To: <532355906.19321.1426710347371.JavaMail.yahoo@mail.yahoo.com> X-Original-Sender: jgottman6@gmail.com Precedence: list Mailing-list: list std-proposals@isocpp.org; contact std-proposals+owners@isocpp.org List-ID: X-Google-Group-Id: 399137483710 List-Post: , List-Help: , List-Archive: List-Subscribe: , List-Unsubscribe: , Xref: news.gmane.org gmane.comp.lang.c++.isocpp.proposals:17114 Archived-At: ------=_Part_1232_861976290.1426901966993 Content-Type: multipart/alternative; boundary="----=_Part_1233_310823368.1426901966993" ------=_Part_1233_310823368.1426901966993 Content-Type: text/plain; charset=UTF-8 Another, possibly simpler way of achieving the same effect is to weaken the rule that virtual functions can't be constexpr. If we say that *final *virtual functions can be final, then we can define Shoe::price as struct Shoe : Merchandise { constexpr int price() override final { return 10; } This would be perfectly safe, because the compiler could always compute Shoe::price() without having to look at a vtable. Joe Gottman On Wednesday, March 18, 2015 at 4:25:51 PM UTC-4, Botond Ballo wrote: > > There are cases where you would like to associate constants with a given > type. Here is how you can currently do it in C++: > > struct Merchandise { > virtual int price() { return 0; } > }; > > struct Shoe : Merchandise { > virtual int price() { return 10; > } > > If you could do it this way instead: > > struct Merchandise { > virtual constexpr int price = 0; > }; > > struct Shoe : Merchandise { > virtual constexpr int price = 10; > } > > then the compiler could store the constant value directly in the vtable. > This would eliminate the double indirection involved in dereferencing a > function vtable entry, the cost of running that function at runtime, and > the code size of the function in the compiled binary. > > Virtual fields would be limited to being constexpr, so that the value to > be placed in the vtable, like other values in the vtable, can be determined > at compile time. > > Has this, or something similar, ever been proposed? Is there any interest > in it? > > -- Botond and Ehsan > > -- --- You received this message because you are subscribed to the Google Groups "ISO C++ Standard - Future Proposals" group. To unsubscribe from this group and stop receiving emails from it, send an email to std-proposals+unsubscribe@isocpp.org. To post to this group, send email to std-proposals@isocpp.org. Visit this group at http://groups.google.com/a/isocpp.org/group/std-proposals/. ------=_Part_1233_310823368.1426901966993 Content-Type: text/html; charset=UTF-8 Content-Transfer-Encoding: quoted-printable
Another, possibly simpler way of achieving the same effect= is to weaken the rule that virtual functions can't be constexpr.  If = we say that final virtual functions can be final, then we can define= Shoe::price as


struct Shoe : Merchandise {
    constexpr int price() override final { return 10;=
}

This would be perfectly safe, because the compiler could alw= ays compute Shoe::price() without having to look at a vtable.

Joe Go= ttman



On Wednesday, March 18, 2015 at 4:25:51 PM UTC-4, Boto= nd Ballo wrote:
There are cases where you would like to associate constants with a given = type. Here is how you can currently do it in C++:

struct Merchandise= {
    virtual int price() { return 0; }
};

str= uct Shoe : Merchandise {
    virtual int price() { return= 10;
}

If you could do it this way instead:

struct Merchan= dise {
    virtual constexpr int price =3D 0;
};
struct Shoe : Merchandise {
    virtual constexpr int p= rice =3D 10;
}

then the compiler could store the constant value d= irectly in the vtable. This would eliminate the double indirection involved= in dereferencing a function vtable entry, the cost of running that functio= n at runtime, and the code size of the function in the compiled binary.
=
Virtual fields would be limited to being constexpr, so that the value t= o be placed in the vtable, like other values in the vtable, can be determin= ed at compile time.

Has this, or something similar, ever been propos= ed? Is there any interest in it?

-- Botond and Ehsan

--

---
You received this message because you are subscribed to the Google Groups &= quot;ISO C++ Standard - Future Proposals" group.
To unsubscribe from this group and stop receiving emails from it, send an e= mail to std-proposa= ls+unsubscribe@isocpp.org.
To post to this group, send email to std-proposals@isocpp.org.
Visit this group at http://groups.google.com/a/isocpp.org/group/std-proposals/<= /a>.
------=_Part_1233_310823368.1426901966993-- ------=_Part_1232_861976290.1426901966993-- .