220 28050 <4fcf4f5e-4a0a-49b9-b62e-ac9569858b11@isocpp.org> article Path: news.gmane.org!.POSTED!not-for-mail From: FrankHB1989 Newsgroups: gmane.comp.lang.c++.isocpp.proposals Subject: Re: Re: Stabilize the numeric types in all platforms Date: Tue, 30 Aug 2016 04:48:42 -0700 (PDT) Lines: 208 Approved: news@gmane.org Message-ID: <4fcf4f5e-4a0a-49b9-b62e-ac9569858b11@isocpp.org> References: <72b7e733-cb85-4f47-a09b-10611f04211b@isocpp.org> <18a39611-8088-4e04-be0a-bb1b7c485683@isocpp.org> <1706311.trY4MaExHO@tjmaciei-mobl1> <36958399-1282-4cef-a12b-0439e176aeda@isocpp.org> <616f8567-2214-4424-a1e3-85ce70e5c99b@isocpp.org> Reply-To: std-proposals@isocpp.org NNTP-Posting-Host: blaine.gmane.org Mime-Version: 1.0 Content-Type: multipart/mixed; boundary="----=_Part_1782_591487521.1472557722340" X-Trace: blaine.gmane.org 1472557731 25718 195.159.176.226 (30 Aug 2016 11:48:51 GMT) X-Complaints-To: usenet@blaine.gmane.org NNTP-Posting-Date: Tue, 30 Aug 2016 11:48:51 +0000 (UTC) To: ISO C++ Standard - Future Proposals Original-X-From: std-proposals+bncBCTJVBPG3QIBBG7FSW7AKGQEUMM3BVY@isocpp.org Tue Aug 30 13:48:46 2016 Return-path: Envelope-to: gclcip-std-proposals@m.gmane.org Original-Received: from mail-pf0-f199.google.com ([209.85.192.199]) by blaine.gmane.org with esmtp (Exim 4.84_2) (envelope-from ) id 1behX2-00065G-3I for gclcip-std-proposals@m.gmane.org; Tue, 30 Aug 2016 13:48:44 +0200 Original-Received: by mail-pf0-f199.google.com with SMTP id w128sf39310245pfd.3 for ; Tue, 30 Aug 2016 04:48:45 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=isocpp-org.20150623.gappssmtp.com; s=20150623; h=date:from:to: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=EpxvnBxzyceB3oB5hRByl3Duwq+SbepECeAb4DJ3FUM=; b=NsWPVk5pO3k/LsL8xTg2lNkRKgIsU2iasFscNm6vtMS9IPDs5ALgleOdmlq0PFM7K/ WUrpMzON6NT/j3/EwWUEe+3N+KcE3DDcmhAps/MKrR6tnwrTbhchgk5sQW4NFzl2pTv4 DU9lz9GOkoe1estthppEcxsR9c3W6voMb5bCjyXKuQzmw4jKh8mzXndP6OdF3xLon8je aCiAZ9ukNa6N2piQgUlFe3HN5P0kcKELS1lZFSswsfXS0Bgn/zDVDJzvX5M+BrAtedrZ MYIodOiYSU2ikiHu3ipdaxH2BFeRnOWH8bJbe+5rC+lcgtCMqPvzVvFSakSpuyITkciN R6Fg== DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20120113; h=date:from:to: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=EpxvnBxzyceB3oB5hRByl3Duwq+SbepECeAb4DJ3FUM=; b=rZ/6BfscKgc/PGqukJ6fbgRnxpwoAske4tNdA+7m+SPfdjAA0RByBPy8GmoqT+bzWh FEweTRqu4PrvcYZiy5JnDFjGmGrUTPXgK1K1zSOGqWlOq4J04PcWxV1uL2ivlT5i3htX GHZNsRbqgehgKDkPoOV7x+D7P7RqqXA3tYxZYXA4Ng3ly8wZUO8O2wUYwZL0euCSGenV SrqhB4vDnjp6/v5wgCFYNuNPvdsReFFSOsek9CxKUzX5vp+6OmnVpXPBC5klhSGprbWl cRngZ09/jn2TG6AP6QeE5hEIIA43HoNczSYkNrxvFnzs58AFMsaXehxW5KKr62Lz0dPH 48CQ== 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: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=EpxvnBxzyceB3oB5hRByl3Duwq+SbepECeAb4DJ3FUM=; b=QvcHA4y+iaDmlsuSXbhGIrjJQSQ/bjS2WqV07TPDnecdc03DBKM8QlLCPNcCDAzd+S Zh31sv14NwU6JlhWLBYYG9Ygt8MZ7zp6LEXRW58fkHCYpp2LnhBk4KRxatVH1dv4+AVm nMh5E0F3TtaXXeTIvSi5du76PCzTeup/n5cK6CKvntIpnPYEqO9cV6Rvv/e2Pi2qCleh 84mgYu2yO+OtvJKh1MbjlOSHhQwQgA54RnhPKzt0V8j6oVDEgzJgZ1IZMts8TXSz9HRd ejvQAbmPVJaJg8mb2IfSDhOq0Bl1+2iXrZDGZ0wEsYj4MEGNfGqxID8uztKpJxNrMZH2 G7uA== X-Gm-Message-State: AE9vXwMKxvzyAM1iLDhc8SOkKu8PjQXpTXNgm1lbmvL5Bl0xVwdKbhD5LHhdNDavH+E8DQ== X-Received: by 10.66.246.197 with SMTP id xy5mr1481758pac.47.1472557724594; Tue, 30 Aug 2016 04:48:44 -0700 (PDT) X-BeenThere: std-proposals@isocpp.org Original-Received: by 10.36.242.68 with SMTP id j65ls1138046ith.8.gmail; Tue, 30 Aug 2016 04:48:43 -0700 (PDT) X-Received: by 10.36.225.69 with SMTP id n66mr94317ith.3.1472557723348; Tue, 30 Aug 2016 04:48:43 -0700 (PDT) In-Reply-To: <616f8567-2214-4424-a1e3-85ce70e5c99b@isocpp.org> X-Original-Sender: frankhb1989@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:28050 Archived-At: ------=_Part_1782_591487521.1472557722340 Content-Type: multipart/alternative; boundary="----=_Part_1783_582962063.1472557722340" ------=_Part_1783_582962063.1472557722340 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: quoted-printable =20 =E5=9C=A8 2016=E5=B9=B48=E6=9C=8823=E6=97=A5=E6=98=9F=E6=9C=9F=E4=BA=8C UTC= +8=E4=B8=8B=E5=8D=889:33:09=EF=BC=8CNicol Bolas=E5=86=99=E9=81=93=EF=BC=9A > > On Tuesday, August 23, 2016 at 7:31:20 AM UTC-4, FrankHB1989 wrote: >> >> =E5=9C=A8 2016=E5=B9=B48=E6=9C=8823=E6=97=A5=E6=98=9F=E6=9C=9F=E4=BA=8C = UTC+8=E4=B8=8B=E5=8D=8812:49:54=EF=BC=8CThiago Macieira=E5=86=99=E9=81=93= =EF=BC=9A >>> >>> On segunda-feira, 22 de agosto de 2016 21:36:12 PDT FrankHB1989 wrote:= =20 >>> > OK, I should have said, they are not mandated *by the core language*,= =20 >>> or,=20 >>> > not guaranteed both built-in and portable.=20 >>> >>> They are mandated in a header that complements the core language. Just= =20 >>> like=20 >>> , and , those types cannot be=20 >>> implemented by anyone except the compiler vendor.=20 >>> >>> Therefore, for all intents and purposes, the minimum-width and fast=20 >>> integer=20 >>> types are guaranteed to be portable.=20 >>> >>> The exact-width ones aren't guaranteed to be portable because their=20 >>> concept=20 >>> isn't portable. Some machines don't have 8-, 16-, 32- and 64-bit types= =20 >>> (they=20 >>> could be multiples of 9 bits). Also note how int32_t is required to use= =20 >>> two's=20 >>> complement representation, which again excludes some machines.=20 >>> >>> Good point about 2's complements. However, all of these are what I have= =20 >> already known. I still wonder why these (at least partially) width-aware= =20 >> types are not in the core language while traditional fundamental integer= =20 >> types are still in it. >> > > ... why does it matter if it's a reserved keyword or not? It's *there*;= =20 > it's available on every platform that provides C++11 or better. You can u= se=20 > it to solve the problem you have with types not having a relatively fixed= =20 > size. > > I don't think "to solve the problem you have with types not having a=20 relatively fixed size" useful in reality. Overflow of signed integers leads= =20 to UB. Wrapped behavior on unsigned integers will cause unexpected result= =20 too easily. Without a concrete limit of the range, they are dangerous (if= =20 not that useless). So we need to know the size sooner or later. When size= =20 is needed, they are *almost always suboptimal* choices. With such types we= =20 actually have some implicit weak guarantees, e.g. an INT_MAX should be=20 greater than 32767. That's too subtle on readability. And I find nothing=20 reasonable to encourage such use. Then keeping them would cause=20 teachability problems, probably. Well, I admit I am not motivated enough to push the change forward=20 currently since it does not benefit much at once. But getting consensus on= =20 some points of design should be good. The placement of something in the standard library or core language is not= =20 > based on some evaluation of the *merit* of the feature. > Please also note "it works" should not be a normal reason in a process of= =20 standardization. There will be great risks to simply pile features without= =20 enough alternative choices considered carefully, and this will probably=20 waste time a lot in future (e.g. `export` and dynamic exception=20 specification). To do the "right thing" pedantically is unlikely to be a=20 premature optimization in such cases than in ordinary projects. There should be at least one topic concerned here: the return type of=20 `std::uncaught_exceptions`.=20 --=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/4fcf4f5e-4a0a-49b9-b62e-ac9569858b11%40isocpp.or= g. ------=_Part_1783_582962063.1472557722340 Content-Type: text/html; charset=UTF-8 Content-Transfer-Encoding: quoted-printable
=C2=A0
=E5=9C=A8 2016=E5=B9=B48=E6=9C=8823=E6=97=A5=E6= =98=9F=E6=9C=9F=E4=BA=8C UTC+8=E4=B8=8B=E5=8D=889:33:09=EF=BC=8CNicol Bolas= =E5=86=99=E9=81=93=EF=BC=9A
On Tuesday, August 23, 2016 at 7:31:20 AM UTC-4, FrankHB1989 = wrote:
=E5=9C=A8 20= 16=E5=B9=B48=E6=9C=8823=E6=97=A5=E6=98=9F=E6=9C=9F=E4=BA=8C UTC+8=E4=B8=8B= =E5=8D=8812:49:54=EF=BC=8CThiago Macieira=E5=86=99=E9=81=93=EF=BC=9AOn segunda-feira, 22 de agosto de 2016 21:= 36:12 PDT FrankHB1989 wrote:
> OK, I should have said, they are not mandated *by the core languag= e*, or,=20
> not guaranteed both built-in and portable.

They are mandated in a header that complements the core language. Just = like=20
<intializer_list>, <type_traits> and <limits>, those = types cannot be=20
implemented by anyone except the compiler vendor.

Therefore, for all intents and purposes, the minimum-width and fast int= eger=20
types are guaranteed to be portable.

The exact-width ones aren't guaranteed to be portable because their= concept=20
isn't portable. Some machines don't have 8-, 16-, 32- and 64-bi= t types (they=20
could be multiples of 9 bits). Also note how int32_t is required to use= two's=20
complement representation, which again excludes some machines.

Good point about 2's complements. However, all of= these are what I have already known. I still wonder why these (at least pa= rtially) width-aware types are not in the core language while traditional f= undamental integer types are still in it.

= .... why does it matter if it's a reserved keyword or not? It's t= here; it's available on every platform that provides C++11 or bette= r. You can use it to solve the problem you have with types not having a rel= atively fixed size.

I don't think = "to solve the problem you have with types not having a=20 relatively fixed size" useful in reality. Overflow of signed integers= =20 leads to UB. Wrapped behavior on unsigned integers will cause unexpected result too easily. Without a concrete limit of the range, they are=20 dangerous (if not that useless). So we need to know the size sooner or=20 later. When size is needed, they are almost always suboptimal=20 choices. With such types we actually have some implicit weak guarantees, e.g. an INT_MAX should be greater than 32767. That's too subtle on=20 readability. And I find nothing reasonable to encourage such=20 use. Then keeping them would cause teachability problems, probably.

= Well, I admit I am not motivated enough to push the change forward=20 currently since it does not benefit much at once. But getting consensus=20 on some points of design should be good.

The placement of something in= the standard library or core language is not based on some evaluation of t= he merit of the feature.

Please= also note "it works" should=20 not be a normal reason in a process of standardization. There will be great= risks to simply pile features without enough alternative choices considere= d carefully, and this will probably waste time a lot in future (e.g. `expor= t` and dynamic exception specification). To do the "right thing" = pedantically is unlikely to be a premature optimization in such cases than = in ordinary projects.

There should be at least one topic concerned h= ere: the return type of `std::uncaught_exceptions`.



<= /div>

--
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/4fcf4f5e-4a0a-49b9-b62e-ac9569858b11= %40isocpp.org.
------=_Part_1783_582962063.1472557722340-- ------=_Part_1782_591487521.1472557722340-- .