220 5319 <03a23bfd-10ab-40db-a7f9-3cbd177a2d52@isocpp.org> article Path: news.gmane.org!not-for-mail From: =?UTF-8?Q?Micha=C5=82_Dominiak?= Newsgroups: gmane.comp.lang.c++.isocpp.proposals Subject: Re: Allowing any integral type as an argument for user defined literal operators Date: Sat, 6 Jul 2013 14:15:48 -0700 (PDT) Lines: 133 Approved: news@gmane.org Message-ID: <03a23bfd-10ab-40db-a7f9-3cbd177a2d52@isocpp.org> References: <80016e3a-2dc8-4887-b59b-a8aff20439e0@isocpp.org> Reply-To: std-proposals@isocpp.org NNTP-Posting-Host: plane.gmane.org Mime-Version: 1.0 Content-Type: multipart/alternative; boundary="----=_Part_3275_1565986.1373145348862" X-Trace: ger.gmane.org 1373145352 25868 80.91.229.3 (6 Jul 2013 21:15:52 GMT) X-Complaints-To: usenet@ger.gmane.org NNTP-Posting-Date: Sat, 6 Jul 2013 21:15:52 +0000 (UTC) To: std-proposals@isocpp.org Original-X-From: std-proposals+bncBCLPTLXGRAFRBBMS4KHAKGQE3EUXQGA@isocpp.org Sat Jul 06 23:15:54 2013 Return-path: Envelope-to: gclcip-std-proposals@m.gmane.org Original-Received: from mail-vc0-f199.google.com ([209.85.220.199]) by plane.gmane.org with esmtp (Exim 4.69) (envelope-from ) id 1UvZpe-0005cm-W2 for gclcip-std-proposals@m.gmane.org; Sat, 06 Jul 2013 23:15:51 +0200 Original-Received: by mail-vc0-f199.google.com with SMTP id gf11sf4002149vcb.6 for ; Sat, 06 Jul 2013 14:15:49 -0700 (PDT) X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20120113; h=x-beenthere:date:from:to:message-id:in-reply-to:references:subject :mime-version:x-original-sender:reply-to:precedence:mailing-list :list-id:x-google-group-id:list-post:list-help:list-archive :list-subscribe:list-unsubscribe:content-type; bh=biZkiiJpw1I3EnKpxZ6Ui8iuWOVaQwSJm4e0V9GK/mU=; b=LtThFPIAaNx44UArZrXArgWJOu+5Gm07O+OWU8rf7KjFEkmH5JGvsar8p/6EUzcsGc bN1+iFW+Amet2dBEqbgq8xFAPayCKJ7S0VzOVHBT79TZfVWHID/mpV9Yh0soMWkmGrHH D/i+VZPdRreU005LqKfnqSHNk6HXze22Yu11jzPcPqSgifTBmeBdg+jgNX/juuEna27h rRjSOjzhmIXsuWclhg6nhvYB0foZKLP6LUChGHiTLANznnNnZwTQmNWoZxqWFBWsTYA9 CV83vI/ljtF4hDHVElXUtfyz5+G+7LEKOWijyhJWaofNJkuioIEgokKT2aJbthop6DRA 1i9Q== X-Received: by 10.224.55.200 with SMTP id v8mr18192548qag.7.1373145349784; Sat, 06 Jul 2013 14:15:49 -0700 (PDT) X-BeenThere: std-proposals@isocpp.org Original-Received: by 10.49.12.169 with SMTP id z9ls1463233qeb.20.gmail; Sat, 06 Jul 2013 14:15:49 -0700 (PDT) X-Received: by 10.49.62.3 with SMTP id u3mr353818qer.26.1373145349196; Sat, 06 Jul 2013 14:15:49 -0700 (PDT) In-Reply-To: X-Original-Sender: griwes@griwes.info 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:5319 Archived-At: ------=_Part_3275_1565986.1373145348862 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: quoted-printable On Saturday, 6 July 2013 23:02:04 UTC+2, Ville Voutilainen wrote: > > > > > On 6 July 2013 23:26, Micha=C5=82 Dominiak >wrote: > >> Today I came across a very silly issue: you cannot do `uint64_t=20 >> operator""_foo(uint64_t)`, because `uint64_t` is a typedef for `unsigned= =20 >> long`, not `unsigned long long` on both Clang and GCC on Linux (despite = the=20 >> fact that `sizeof(unsigned long) =3D=3D sizeof(unsigned long long)`). Yo= u have=20 >> to type the long type name, just because the typedef isn't for the exact= =20 >> type required by the standard. I hope we all agree this is a silly=20 >> situation. >> >> So, I'd like to propose allowing `operator""`s to take arguments of any= =20 >> integral types. >> > > It's not supposed to take any integral types. The operator takes an=20 > unsigned type because any negation > happens outside it. Since we already have such a restriction for the type= =20 > a literal operator accepts, it's > not a far stretch to have a further limitation. I wouldn't mind it=20 > accepting any unsigned integral type, > but I don't see it a big deal that you have to have a certain signature= =20 > for a literal operator. > Ok, s/any/any unsigned/.=20 =20 > Not allowing multiple kinds of integral types as literal operator=20 > parameters also avoid having > to deal with overload resolution for the literal operators, and not=20 > allowing narrow types avoids > having to deal with narrowing conversions. > =20 No-one says allowing any unsigned argument type would allow overloading (I= =20 think it would be a counter-feature, really). And since you know all the=20 values passed to the operator at compile time, no narrowing would ever=20 happen - just slap the user with an error about the value being to big for= =20 given type. --=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. Visit this group at http://groups.google.com/a/isocpp.org/group/std-proposa= ls/. ------=_Part_3275_1565986.1373145348862 Content-Type: text/html; charset=UTF-8 Content-Transfer-Encoding: quoted-printable

On Saturday, 6 July 2013 23:02:04 UTC+2, Ville Voutilainen wrote:<= blockquote class=3D"gmail_quote" style=3D"margin: 0;margin-left: 0.8ex;bord= er-left: 1px #ccc solid;padding-left: 1ex;">


<= br>
On 6 July 2013 23:26, Micha=C5=82 Dominiak <gri...@griwes.info> wrote:
Today I came across a very silly issue: you = cannot do `uint64_t operator""_foo(uint64_t)`, because `uint64_t` is a type= def for `unsigned long`, not `unsigned long long` on both Clang and GCC on = Linux (despite the fact that `sizeof(unsigned long) =3D=3D sizeof(unsigned = long long)`). You have to type the long type name, just because the typedef= isn't for the exact type required by the standard. I hope we all agree thi= s is a silly situation.

So, I'd like to propose allowing `operator""`s to take argum= ents of any integral types.

It's not = supposed to take any integral types. The operator takes an unsigned type be= cause any negation
happens outside it. Since we already have such a restriction for= the type a literal operator accepts, it's
not a far stretch to have a f= urther limitation. I wouldn't mind it accepting any unsigned integral type,=
but I don't see it a big deal that you have to have a certain signature for= a literal operator.

Ok, s/any/any unsigned/. 
 
Not allowing multiple kinds of integral types as literal= operator parameters also avoid having
to deal with overload resolution for the literal operators, and not al= lowing narrow types avoids
having to deal with narrowing conversions.
 
No-one says= allowing any unsigned argument type would allow overloading (I think it wo= uld be a counter-feature, really). And since you know all the values passed= to the operator at compile time, no narrowing would ever happen - just sla= p the user with an error about the value being to big for given type.

--
 
---
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-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/<= /a>.
 
 
------=_Part_3275_1565986.1373145348862-- .