220 34537 article Path: news.gmane.org!.POSTED!not-for-mail From: Niall Douglas Newsgroups: gmane.comp.lang.c++.isocpp.proposals Subject: Re: base and self built-in keywords Date: Mon, 2 Oct 2017 07:29:03 -0700 (PDT) Lines: 201 Approved: news@gmane.org Message-ID: References: <9dec3353-3142-43d1-928f-b88ac74951bf@isocpp.org> <63cb64a8-5add-45ff-a856-364d37042b44@isocpp.org> Reply-To: std-proposals@isocpp.org NNTP-Posting-Host: blaine.gmane.org Mime-Version: 1.0 Content-Type: multipart/mixed; boundary="----=_Part_5247_125774284.1506954543345" X-Trace: blaine.gmane.org 1506954546 19183 195.159.176.226 (2 Oct 2017 14:29:06 GMT) X-Complaints-To: usenet@blaine.gmane.org NNTP-Posting-Date: Mon, 2 Oct 2017 14:29:06 +0000 (UTC) Cc: sergeikrivonos@gmail.com To: ISO C++ Standard - Future Proposals Original-X-From: std-proposals+bncBDGKFT5YZADRBME2ZHHAKGQEVTALFEI@isocpp.org Mon Oct 02 16:29:00 2017 Return-path: Envelope-to: gclcip-std-proposals@m.gmane.org Original-Received: from mail-qk0-f199.google.com ([209.85.220.199]) by blaine.gmane.org with esmtp (Exim 4.84_2) (envelope-from ) id 1dz1iM-0004LR-8B for gclcip-std-proposals@m.gmane.org; Mon, 02 Oct 2017 16:28:58 +0200 Original-Received: by mail-qk0-f199.google.com with SMTP id c67sf4437510qkg.18 for ; Mon, 02 Oct 2017 07:29:06 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=isocpp-org.20150623.gappssmtp.com; s=20150623; h=date:from:to:cc:message-id:in-reply-to:references:subject :mime-version:x-original-sender:reply-to:precedence:mailing-list :list-id:list-post:list-help:list-archive:list-subscribe :list-unsubscribe; bh=brMWeH34nKZoSuUhs4T8xgAcYa9od6PHlfB0sVQXLRI=; b=rI3ETNTgzdXc3nFVtfznpics1kX5j9mJwihDWQXxwqMMJzKATg5HiTv3mXtIWweELl ONM3VU+8+W4I751a6TWnZ2PHM9lu7PauRt2reftK9TLnoXrrAb/uFxLlzWGshVfrrtcf i88kD5bDzcaLxM0w2A7dxz84ayA7JLcuKvpjFOVPsP4g4dTlbXHO3o7UEVnXZdMc71rH W+UPjBoKwvcCdyVBZ+bn31ipW2v+v3htJw/1X+sBiD9ATwdEgKIGoq9zoNh4FgWwgslr d7pTAdMFP/25DJslZZ16megZ2cIs/7uaNgeo84ixXKFsRpcsLSFqYfF/qGYCQv2FTznn A84A== DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20161025; h=date:from:to:cc:message-id:in-reply-to:references:subject :mime-version:x-original-sender:reply-to:precedence:mailing-list :list-id:list-post:list-help:list-archive:list-subscribe :list-unsubscribe; bh=brMWeH34nKZoSuUhs4T8xgAcYa9od6PHlfB0sVQXLRI=; b=Hzk8ElGLYVTsIiT7pGGXRGv56lN2Jz1DF9wBo2NYxor/lW0sbIq1iKQg1FWKAEXgua aExOPFaZbrV7ylAG8mRBgF4x5TMGxvvh8xLOHjSkzYE/BGY1BMdeK4UjdQWw4Q0F2exS DI1gVl6+mG0BV8/tOaqV7ALWUNMHPiLYaRxFDfUyOfhi4INQKKcvXbmIDDO0xsDFlLKd 56gxScS68SAszOS6VmKbs5JOJpL8M6WpK2nrFV6sWxWQSyyeswFgxnVEfO2NqxU7bLoY yTITMUXMcAIEupe3RrTAZPkiM+6z1uo4Fi7ltetWRfntLs0ZXfm6lxQKJRPHAtLQ5eot 34qg== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:date:from:to:cc:message-id:in-reply-to :references:subject:mime-version:x-original-sender:reply-to :precedence:mailing-list:list-id:x-spam-checked-in-group:list-post :list-help:list-archive:list-subscribe:list-unsubscribe; bh=brMWeH34nKZoSuUhs4T8xgAcYa9od6PHlfB0sVQXLRI=; b=IwGc0YtZ/a0Z2MwwjhMltZ3mVwg+3o2d0BblTP18OPlqB70xff5hNZDK3U45IFPvE4 9h+cJRcQQhurexmal9ppq4QsvwgV++MsuKQUilMHmPFkxVQHvaq5+DYw69l/UWCJKDSt UzY0DLIIhZ2DLLn1UZ2l6ArIbGMtSS1a51Dnnuymo5KBjhk+CkKBiJrg2XLYaChltE6l z1k4mMvXkJRE325n4pYRspBVdI6r559IeaWdmT4x2sDsiGkk09yOH07EFkEqnkCDa0NA n945jgOCDl0MYbOnlN6Bn4+E3YOjlSgHLkoLXd+ozpQxRWfOnPwChKzeoJGWN059ecUP 1QGA== X-Gm-Message-State: AHPjjUg5CvPZ3btd28KUFJAvek0bVsYI8TV0zENMLP3/o6fKKZDdnPMn E28ZKJKKoR+QZGe3ee/7a4c5EQ== X-Google-Smtp-Source: AOwi7QAqzdbJqX0SvPwk35PYGgEWwMjzwsZi576OV/ewm3Y7DLczfP1Ep9ME2hy8FFeIV5mnGiZSLA== X-Received: by 10.176.17.234 with SMTP id q42mr3306332uac.21.1506954545511; Mon, 02 Oct 2017 07:29:05 -0700 (PDT) X-BeenThere: std-proposals@isocpp.org Original-Received: by 10.31.114.203 with SMTP id n194ls3542860vkc.0.gmail; Mon, 02 Oct 2017 07:29:03 -0700 (PDT) X-Received: by 10.31.199.2 with SMTP id x2mr674160vkf.17.1506954543867; Mon, 02 Oct 2017 07:29:03 -0700 (PDT) In-Reply-To: <63cb64a8-5add-45ff-a856-364d37042b44@isocpp.org> X-Original-Sender: nialldouglas14@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:34537 Archived-At: ------=_Part_5247_125774284.1506954543345 Content-Type: multipart/alternative; boundary="----=_Part_5248_680863879.1506954543346" ------=_Part_5248_680863879.1506954543346 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable > > >>> >>> Python supports multiple inheritance too and I'd suggest you clone thei= r=20 >> super() function rather than invent anything new. >> > > Oh dear. Literally anything would be better than Python's super() functio= n. > Python invents a sort of linearization of the multiple-inheritance tree (= the=20 > MRO ) so that you can= =20 > manually call super-class __init__ routines (or not) as you choose; i.e.,= =20 > Python does not have the concept of "constructor" as we understand it in= =20 > C++. It's more like Objective-C's init routines. > A more accurate explanation is that super() would return the native vtable= =20 for the selected inherited class. =20 > In my experience it's basically impossible to use correctly; if you're=20 > doing multiple inheritance in Python you're *probably* doing something=20 > wrong whether you're aware of it or not. > (In Python 2 super() requires repeating the name of your own class anyway= ,=20 > where repetition of class names was the original problem OP wanted to=20 > solve. But I see that the repetition has been dropped from Python 3 =E2= =80=94 cool! This=20 > tutorial=20 > = =20 > seems fair and balanced to me, and covers the syntax in both dialects of= =20 > Python.) > Multiple inheritance is exactly as useful in Python as in C++. Usually best= =20 sparingly, but when it's useful, it's useful, plus the same metaprogramming= =20 tricks are available in Python's multiple inheritance as in C++. And I've= =20 seen no trouble in using it correctly, you can have recursive nestings of= =20 eval() specialised fragments assembled together and it all works fine,=20 including super() to traverse the hierarchy. Python's really great. Such a deep well of evil power available to the=20 programmer. =20 > This blog post =20 > points out something I'd never consciously considered before: Python has = no=20 > non-virtual inheritance (I think?).=20 > More that virtual inheritance is the default for all member functions. You= =20 need to explicitly ask to call a non-virtualised implementation, which is= =20 via super() actually. =20 > > Anyway, I don't think there's any problem to be solved here. Certainly I= =20 > don't think C++ should introduce any special syntax for the "I'th base=20 > class." If you want something like that, you probably want Herb Sutter's= =20 > metaclasses (or better) anyway. > Reserving self and base for Herb's metaclasses, now that is something I can= =20 agree 110% with. Niall --=20 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 e= mail to std-proposals+unsubscribe@isocpp.org. To post to this group, send email to std-proposals@isocpp.org. To view this discussion on the web visit https://groups.google.com/a/isocpp= ..org/d/msgid/std-proposals/c93b4603-ff34-495f-bcd5-c844e8ee0eec%40isocpp.or= g. ------=_Part_5248_680863879.1506954543346 Content-Type: text/html; charset="UTF-8" Content-Transfer-Encoding: quoted-printable


Python supports m= ultiple inheritance too and I'd suggest you clone their super() functio= n rather than invent anything new.

<= div>Oh dear. Literally anything would be better than Python's super() f= unction.
Python invents a sort of linearization of the multiple-i= nheritance tree (the MRO) so that you can manually call super-class= __init__ routines (or not) as you choose; i.e., Python does not have the c= oncept of "constructor" as we understand it in C++. It's more= like Objective-C's init routines.

A more accurate explanation is that super() would return the native= vtable for the selected inherited class.
=C2=A0
In my experience i= t's basically impossible to use correctly; if you're doing multiple= inheritance in Python you're probably doing something wrong whe= ther you're aware of it or not.
(In Python 2=C2=A0super() req= uires repeating the name of your own class anyway, where repetition of clas= s names was the original problem OP wanted to solve. But I see that the rep= etition has been dropped from Python 3 =E2=80=94 cool! This tutorial seems fair and balanced to = me, and covers the syntax in both dialects of Python.)

Multiple inheritance is exactly as useful in Python= as in C++. Usually best sparingly, but when it's useful, it's usef= ul, plus the same metaprogramming tricks are available in Python's mult= iple inheritance as in C++. And I've seen no trouble in using it correc= tly, you can have recursive nestings of eval() specialised fragments assemb= led together and it all works fine, including super() to traverse the hiera= rchy.

Python's really great. Such a deep well = of evil power available to the programmer.
=C2=A0
This blog post points out something I'd never consciously consi= dered before: Python has no non-virtual inheritance (I think?).

More that virtual inheritance is the defa= ult for all member functions. You need to explicitly ask to call a non-virt= ualised implementation, which is via super() actually.
=C2=A0

Anyway, I don't think there's any problem to be solved her= e. Certainly I don't think C++ should introduce any special syntax for = the "I'th base class." If you want something like that, you p= robably want Herb Sutter's metaclasses (or better) anyway.
<= /blockquote>

Reserving self and base for Herb's meta= classes, now that is something I can agree 110% with.

<= div>Niall

--
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.
To view this discussion on the web visit https://groups.google.= com/a/isocpp.org/d/msgid/std-proposals/c93b4603-ff34-495f-bcd5-c844e8ee0eec= %40isocpp.org.
------=_Part_5248_680863879.1506954543346-- ------=_Part_5247_125774284.1506954543345-- .