220 27885 <72b7e733-cb85-4f47-a09b-10611f04211b@isocpp.org> article Path: news.gmane.org!.POSTED!not-for-mail From: Nicol Bolas Newsgroups: gmane.comp.lang.c++.isocpp.proposals Subject: Re: Re: Stabilize the numeric types in all platforms Date: Mon, 22 Aug 2016 21:18:34 -0700 (PDT) Lines: 167 Approved: news@gmane.org Message-ID: <72b7e733-cb85-4f47-a09b-10611f04211b@isocpp.org> References: <0d61be1e-2020-4f09-baa6-64dca463cb68@isocpp.org> Reply-To: std-proposals@isocpp.org NNTP-Posting-Host: blaine.gmane.org Mime-Version: 1.0 Content-Type: multipart/mixed; boundary="----=_Part_4814_1123196411.1471925914645" X-Trace: blaine.gmane.org 1471925921 2925 195.159.176.226 (23 Aug 2016 04:18:41 GMT) X-Complaints-To: usenet@blaine.gmane.org NNTP-Posting-Date: Tue, 23 Aug 2016 04:18:41 +0000 (UTC) To: ISO C++ Standard - Future Proposals Original-X-From: std-proposals+bncBCEKFTV6ZUMBBG45566QKGQEW6CVLPA@isocpp.org Tue Aug 23 06:18:36 2016 Return-path: Envelope-to: gclcip-std-proposals@m.gmane.org Original-Received: from mail-yb0-f199.google.com ([209.85.213.199]) by blaine.gmane.org with esmtp (Exim 4.84_2) (envelope-from ) id 1bc3Aa-0000RL-CD for gclcip-std-proposals@m.gmane.org; Tue, 23 Aug 2016 06:18:36 +0200 Original-Received: by mail-yb0-f199.google.com with SMTP id f60sf100672554ybi.3 for ; Mon, 22 Aug 2016 21:18:37 -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=ePHbf+lcTu6XrXWTQ1pC31jQyfTQuEp9fbuYwfR8HMs=; b=yVOov15rTKap9oqj8T3k2QgQaZP7IZMQh66iFxzKwAEmQoVE7dlWra+kqBnNZea95n 3Fn1FUPjkxiyDSP7ug3+8SuTAkq585xaYpComaIvHVDE2QRFj7USlT1fKfueYOwRrlOZ qVY3bgciNd02HTbs/hVU6Ucl8wkPmBL483GODWBn4YjpLmz8jp6GE1mhjbndG5wex4zT iOeHh2F3E07pNocpD1TOKbWKyVe4+UDsSCduG18/CgWwDAspbI/ELBRp+aYdYvhoUlW9 SXA7nMdv3m48dcEsDubQDmf+xLs6PnaOMLo9KxsvJD5RQWhHR46PrB+RhZ+/slpZP4iv cRmw== 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=ePHbf+lcTu6XrXWTQ1pC31jQyfTQuEp9fbuYwfR8HMs=; b=LzT13k3XckvzECis3octiCBHocJB/I+HcmsA84pTNjO7owMDPrr9BleirAN1l7q1HQ jcsZgq1vpiHUVhP+FJ7zy0AZBFcTvKmcLSKmUMV4CPsJnCzH0Me3/N1eLmR/NPNNNg8b psIbv4JTuqgGrSrvFUbXaM5TiqXtL/vJwqrC2pzNCQrGtyMyuPTVDdcVnqSn4JSHfgMM UJHViKvhnEeXaUMfYQSWwKE6TpV408RyDcnVjcHcO25CEeWWPzA32K4b7MSJAKwL49/8 xJli2ekLIpdHe8v+E4qSGmYynKgOSbUnqvatptBvGNWuZ4kNosEBg3KrpHEWP63SmsFt Holg== 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=ePHbf+lcTu6XrXWTQ1pC31jQyfTQuEp9fbuYwfR8HMs=; b=Eq49nUoju2AC/A6FtZWRbjwF5WASCa2CPZ4AC/zKQE0C6D1LE2tAJ/VEqg9xqtkJJR wgaUTtIlnRmGBr8QtH/ENIZPcJYpQhsyYsJbaZ/M8LzRO8lZZe4axqaPTF9LcGPHPUOB /sp11H8W+uJXTL4zSJKpRDHQnCKlRk7qkd01vs9EETsBo/G5tzX982Ph7hS1IVtGIQB5 gJOSWsJjG0NLIscwC03vE4oPJAykHnG/md6vxl1KWhFdFFvX3+MftkwKX5DQ5Cpqq+TU slm5ir0rHXwXv4UIekmOYt2wqPGWGvY/+qmvWRQHVLc9Kb7Rx+rKLklPO3NcJcF+y98Y Vq5w== X-Gm-Message-State: AEkooutowYRdmcMNhuOix7WrjdfJ5iwt1VrAUBWrnb7Um2q0Ruq9gXTGGqTVqx3a5log0Q== X-Received: by 10.200.37.20 with SMTP id 20mr19097654qtm.17.1471925916765; Mon, 22 Aug 2016 21:18:36 -0700 (PDT) X-BeenThere: std-proposals@isocpp.org Original-Received: by 10.107.20.68 with SMTP id 65ls4743544iou.16.gmail; Mon, 22 Aug 2016 21:18:35 -0700 (PDT) X-Received: by 10.36.238.143 with SMTP id b137mr766421iti.7.1471925915478; Mon, 22 Aug 2016 21:18:35 -0700 (PDT) In-Reply-To: X-Original-Sender: jmckesson@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:27885 Archived-At: ------=_Part_4814_1123196411.1471925914645 Content-Type: multipart/alternative; boundary="----=_Part_4815_1407464808.1471925914645" ------=_Part_4815_1407464808.1471925914645 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: quoted-printable On Monday, August 22, 2016 at 11:40:34 PM 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 U= TC+8=E4=B8=8A=E5=8D=8810:56:48=EF=BC=8CRen Industries=E5=86=99=E9=81=93=EF= =BC=9A >> >> Yes, *what if another platform is 16 bit? *How in the world are we to=20 >> change that platform so it supports 32 bit ints? >> All the currently working code would break, and for a lot of legacy=20 >> systems the actual code might be gone. >> >> Even if everyone had the source code to every program, many programs may= =20 >> *depend* on int being 16 bits. >> The code might not make sense using 32 bit integers. For example, some= =20 >> DSP's are still 16 bit; emulating 32 bits on 16 bit processors will more= =20 >> than half the speed of the code. And much code on these sorts of units i= s=20 >> time critical; they might not have that many cycles to spare. >> =20 >> > Finally, the fundamental meaning of "int" would change; int would no=20 >> longer be the natural size for the platform. So why the hell are we call= ing=20 >> it int anymore? It's a 32 bit integer, not a platform neutral integer.= =20 >> Either we need to name it more strictly (such as int32_t or something) o= r=20 >> we need to give a new name to "platform neutral" integer or "most natura= l"=20 >> integer or "fastest" integer. Maybe intnatural? natural? fixed? >> >> There is no guarantee about "natural size". > Yes, there is no guarantee about what that means. But compiler writers=20 would encounter problems if they start arbitrarily violating the=20 expectations of users of that platform. As such, `int` will generally be=20 the "natural size" for the processor in question. There have been the names for "fastest" ones: intN_fast_t/uintN_fast_t.=20 > > It is reasonable that things like intN_t cannot be mandated for every=20 > platforms by practical reasons. But why not int_leastN_t/int_fastN_t? > They *are mandated* for all platforms. The only optional ones are those=20 that use a specific bitdepth and the `intptr_t`s. Every valid C++11 implementation must provide those types. Even=20 freestanding implementations, which are permitted to only provide a subset= =20 of the full standard library, are *required*, to provide these types. =20 > The fact of keeping too many integer types without guarantees in the=20 > language specification is the true ugliness. Their co-existence also=20 > frustrates the normative type system (e.g. unsigned long vs. unsigned whe= n=20 > they are the same thing in the ABI). I prefer most of them will be=20 > eventually conditionally-supported. > > BTW, I also don't think we have good reasons to provide more wrapper type= s=20 > by the language specification. They can be unusable placeholders too=20 > easily, causing compatibility and maintenance nightmare. Why should=20 > programmers suffer from a `wchar_t` with 8-bit (yeah, damned Bionic)? > > --=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/72b7e733-cb85-4f47-a09b-10611f04211b%40isocpp.or= g. ------=_Part_4815_1407464808.1471925914645 Content-Type: text/html; charset=UTF-8 Content-Transfer-Encoding: quoted-printable
On Monday, August 22, 2016 at 11:40:34 PM UTC-4, FrankHB19= 89 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=8A=E5=8D=8810:56:48=EF=BC=8CRen Industries=E5=86=99=E9=81=93=EF=BC= =9A
Yes, what if= another platform is 16 bit? How in the world are we to change that pla= tform so it supports 32 bit ints?
All the currently working code would = break, and for a lot of legacy systems the actual code might be gone.
=

Even if everyone had the source code to every program, = many programs may depend=C2=A0on int being 16 bits.
The co= de might not make sense using 32 bit integers. For example, some DSP's = are still 16 bit; emulating 32 bits on 16 bit processors will more than hal= f the speed of the code. And much code on these sorts of units is time crit= ical; they might not have that many cycles to spare.
=C2=A0
=
Finally, the fundamental meaning of "int" would = change; int would no longer be the natural size for the platform. So why th= e hell are we calling it int anymore? It's a 32 bit integer, not a plat= form neutral integer. Either we need to name it more strictly (such as int3= 2_t or something) or we need to give a new name to "platform neutral&q= uot; integer or "most natural" integer or "fastest" int= eger. Maybe intnatural? natural? fixed?

There is no guarantee about "natural size".

Yes, there is no guarantee about what that means. = But compiler writers would encounter problems if they start arbitrarily vio= lating the expectations of users of that platform. As such, `int` will gene= rally be the "natural size" for the processor in question.
<= /div>
There have been the names for "fastest" ones: intN_fast= _t/uintN_fast_t.

It is reasonable that things like intN_= t cannot be mandated for every platforms by practical reasons. But why not = int_leastN_t/int_fastN_t?

They are mand= ated for all platforms. The only optional ones are those that use a spe= cific bitdepth and the `intptr_t`s.

Every valid C++11 implementation= must provide those types. Even freestanding implementations, which are per= mitted to only provide a subset of the full standard library, are requir= ed, to provide these types.
=C2=A0
The fact of keeping too many integer= types without guarantees in the language specification is the true uglines= s. Their co-existence also frustrates the normative type system (e.g. unsig= ned long vs. unsigned when they are the same thing in the ABI). I prefer mo= st of them will be eventually conditionally-supported.

BTW, I also d= on't think we have good reasons to provide more wrapper types by the la= nguage specification. They can be unusable placeholders too easily, causing= compatibility and maintenance nightmare. Why should programmers suffer fro= m a `wchar_t` with 8-bit (yeah, damned Bionic)?

--
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/72b7e733-cb85-4f47-a09b-10611f04211b= %40isocpp.org.
------=_Part_4815_1407464808.1471925914645-- ------=_Part_4814_1123196411.1471925914645-- .