220 35366 <5776c615-e654-4918-ba57-e77dca559456@isocpp.org> article Path: news.gmane.org!.POSTED!not-for-mail From: bastienpenava@gmail.com Newsgroups: gmane.comp.lang.c++.isocpp.proposals Subject: allowing constexpr arguments in deduction guides: a first step Date: Wed, 15 Nov 2017 00:52:28 -0800 (PST) Lines: 407 Approved: news@gmane.org Message-ID: <5776c615-e654-4918-ba57-e77dca559456@isocpp.org> Reply-To: std-proposals@isocpp.org NNTP-Posting-Host: blaine.gmane.org Mime-Version: 1.0 Content-Type: multipart/mixed; boundary="----=_Part_13586_1529819433.1510735948135" X-Trace: blaine.gmane.org 1510735948 6013 195.159.176.226 (15 Nov 2017 08:52:28 GMT) X-Complaints-To: usenet@blaine.gmane.org NNTP-Posting-Date: Wed, 15 Nov 2017 08:52:28 +0000 (UTC) To: ISO C++ Standard - Future Proposals Original-X-From: std-proposals+bncBCG3XHNI6IDRBTEAWDIAKGQESPNMNGY@isocpp.org Wed Nov 15 09:52:23 2017 Return-path: Envelope-to: gclcip-std-proposals@m.gmane.org Original-Received: from mail-vk0-f71.google.com ([209.85.213.71]) by blaine.gmane.org with esmtp (Exim 4.84_2) (envelope-from ) id 1eEtQl-0001D0-2t for gclcip-std-proposals@m.gmane.org; Wed, 15 Nov 2017 09:52:23 +0100 Original-Received: by mail-vk0-f71.google.com with SMTP id 22sf6887710vkq.20 for ; Wed, 15 Nov 2017 00:52:30 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=isocpp-org.20150623.gappssmtp.com; s=20150623; h=date:from:to:message-id:subject:mime-version:x-original-sender :reply-to:precedence:mailing-list:list-id:list-post:list-help :list-archive:list-subscribe:list-unsubscribe; bh=t5IxY+B0/S0WDZCKmkAAq2wewLycHqgCVs5mO0lMfx8=; b=V0c1d9DS1acSbcG6YqxJscZgYJMglT200Gw2BGyVXBTTvW+aajRNdAoxFAsQGaB+E0 83qSpaRNUDkNMGhBx57qdFL5NB909II2de+PAJb4BKumzLbl8A9K1r+zYlARSduF05Cq somIZgBHFa8E+i6dbQR4G2UGzaTSloCwYtr0PXmM1iA1ygpZUhjQqbPAXXo+FYiQnavF 9A3fWMa5hOnhqKdhGaH4IdC8iremXa3l3TPONVY4qAGxn3HdyWgSaLRYVOt3eJ5TUxxw RgO4cpoJqnfT3RF33pNiXa7JVRNii8nZmMCYKPlMD9TrGgFZ5672XDxDMCQ7PKubI6NB /gJQ== DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20161025; h=date:from:to:message-id:subject:mime-version:x-original-sender :reply-to:precedence:mailing-list:list-id:list-post:list-help :list-archive:list-subscribe:list-unsubscribe; bh=t5IxY+B0/S0WDZCKmkAAq2wewLycHqgCVs5mO0lMfx8=; b=Mm1mZ9lJ0FmweaCeHg+0/+y6KDLKMtP0WB7xh3rVJCvp7BqJ2xYHmZhIzTGu4OMljY VKFaQDgHP9GmL//rYQQmKejOGu124orPj1KsCv325QIV1u83ni4jpyg2TjmC35ygsWoh 4D9zPBizaeWnr5Mz0w7/CWdp/X4mVaIVfsp5Q/u+jGGgtFCSNBPV/UNog4SWSW7Osr/B 67Bdq2/KTnZ0LAzsADoB+lvZxm+hW0LzghGT7tV3VksM4gJJS1UK5iDxioIYV4o/tZZE 4lS4GNwQIiMsUgeA+IbjOGaqSfRnFy4/IDL15xWiUe9V3XdzSQek/kjVBMmJnp3wjq5b pObQ== 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:message-id: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=t5IxY+B0/S0WDZCKmkAAq2wewLycHqgCVs5mO0lMfx8=; b=VJXygueM4aQvUS1s6hO0L1j1ikWi7CYZB9M3rcRN1ejWk6hnpCzlFTY05LkBsEWJEC lM+P+aJUd/Po4TI2XUyZqHYoLphUU+5kjL7V84h64tre6vKkSQ31g2i0PYoxTbytwggu fuWnsUhRpP/ZAVCxIeBAq98246uGmvaBjCO396WG7bnTFrA9d77R/VSzYQIiPQDX6JDq i9bevzt3Cg50sWe9uHjhM65PM6yu5FR/hfk0jP5tAzhoY2kNkk9QnmAt0LEbHvxB15D0 7ECYBwwOnz3B4bU7REIkudd7n69rBzxnVBNOC1iDc1KQEp3yjz2QQZuQm2OKdRZtW5al X5Cw== X-Gm-Message-State: AJaThX7kEvKZ5bNcm4AQLkvpQzFmzvKw2nev4T/sdrzHiPtCB2ZHzr8u NYlNxuUz+YHTIPgoJuxmtOcr9w== X-Google-Smtp-Source: AGs4zMbCI+Cc22fiJx/1uoneUij2/PvfwJdIoUpazhN+IEjCC6OvKXAiLsCrNlejAdGqZ8FmjQqSqQ== X-Received: by 10.31.73.66 with SMTP id w63mr5458355vka.0.1510735950185; Wed, 15 Nov 2017 00:52:30 -0800 (PST) X-BeenThere: std-proposals@isocpp.org Original-Received: by 10.159.62.77 with SMTP id c13ls4843523uaj.18.gmail; Wed, 15 Nov 2017 00:52:28 -0800 (PST) X-Received: by 10.31.125.129 with SMTP id y123mr2024100vkc.1.1510735948653; Wed, 15 Nov 2017 00:52:28 -0800 (PST) X-Original-Sender: BastienPenava@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:35366 Archived-At: ------=_Part_13586_1529819433.1510735948135 Content-Type: multipart/alternative; boundary="----=_Part_13587_1883275120.1510735948136" ------=_Part_13587_1883275120.1510735948136 Content-Type: text/plain; charset="UTF-8" Hi, So we know that constexpr argument would be hard to implement because of the the mangling and linking of function with such a parameter. That being said a deduction guide don't have those issues as they only exist at compile-time. This means that constexpr arguments could be allowed in deduction guides. This would allow through type-value semantics to simulate constexpr arguments and generalized non-type template parameter. Here's an example: namespace std { template integral_constant(constexpr T&& x) -> integral_constant; template struct tuple { //... template constexpr decltype(auto) operator[](integral_constant) const { return std::get(*this); } //... }; } So that takes care of currently allowed non-type template parameter but what about the rest ? With lambdas! template struct constexpr_arg { template constexpr constexpr_arg(T&&) {} static constexpr decltype(auto) value = Fn(); }; template constexpr_arg(constexpr T&& x) -> constexpr_arg<+[] -> decltype(auto) { return x;}>; //lambdas in unevaluated context now allowed using runtime_arg = constexpr_arg; template constexpr_arg(T&&) -> runtime_arg template void func(constexpr_arg) //a bit ugly but no longer forces the caller to write ugly code { using type = constexpr_arg; static_assert(type::value == 3.14); } void func(runtime_arg) //runtime specialization { } int main() { func(3.14); func(new int); //calls the runtime_arg version } This would also allow to simulate generalized non-type template parameters: template struct A { static_assert(T::value == "hello"sv); }; int main() { A tmp; } -- 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. To view this discussion on the web visit https://groups.google.com/a/isocpp.org/d/msgid/std-proposals/5776c615-e654-4918-ba57-e77dca559456%40isocpp.org. ------=_Part_13587_1883275120.1510735948136 Content-Type: text/html; charset="UTF-8" Content-Transfer-Encoding: quoted-printable
Hi,

So we know that constexpr argument = would be hard to implement because of the the mangling and linking of funct= ion with such a parameter.
That being said a deduction guide don&= #39;t have those issues as they only exist at compile-time.
This = means that constexpr arguments could be allowed in deduction guides.
<= div>This would allow through type-value semantics to simulate constexpr arg= uments and generalized non-type template parameter.
Here's an= example:

namespace std
{
=C2=A0 =C2=A0template<class= T>
=C2=A0 =C2= =A0integral_constant(constexpr= T&&am= p; x) -> integral_constant<T, x>;

=C2=A0 =C2=A0template<class... <= /span>Args= >
=C2=A0 =C2=A0struct tuple
=C2=A0 =C2= =A0{
=C2=A0 =C2= =A0 =C2=A0 //...
= =C2=A0 =C2=A0 =C2=A0 template<size_t I>
= =C2=A0 =C2=A0 =C2=A0 constexpr decltype(= auto) operator[](integral_constant<I>) const { return std::get<I>(*this); }
=C2=A0 =C2=A0 =C2= =A0 //...<= /span>
=C2=A0 = =C2=A0};
<= span style=3D"color: #660;" class=3D"styled-by-prettify">}

So that takes care of currently allowed non-type template param= eter but what about the rest ? With lambdas!

template<auto Fn>
struct constexpr_arg
{
=C2=A0 =C2=A0template<class T> constexpr constexpr_arg(T&&) {}
=C2=A0 =C2=A0static constexpr decltype(aut= o)<= span style=3D"color: #000;" class=3D"styled-by-prettify"> value =3D Fn();
};

template<= class T>
constexpr_arg(constexpr T&& x) -> constexpr_arg<+[] -> declty= pe(= auto) {return x;}>; //lambdas in unevaluated context now allowed

using runtime_arg =3D constexpr_arg<nullptr>;

template<class T><= br>constexpr_arg(T&&)<= /span> -> runtime_arg

template= <auto Fn>
void func(constexpr_arg<Fn>) //a bit ugly but no longer forces the caller to write ugly code
{
=C2=A0 =C2=A0using type =3D constexpr_arg<Fn>;
=C2=A0 =C2=A0static_assert(type::value =3D=3D 3.14);
}
void func(runtime_arg) //runtime specialization
{
}

int main()
{
=C2=A0 =C2=A0func<= /span>(3.14);
=C2=A0 =C2=A0func(new int); //calls the runtime_arg version
}

This would also all= ow to simulate generalized non-type template parameters:

templ= ate<class<= span style=3D"color: #000;" class=3D"styled-by-prettify"> T>
struct A
{
=C2=A0 static_assert(T::va= lue =3D=3D= "hello"sv);
};

int main()
{
= =C2=A0 =C2=A0A<de= cltype(constexpr_arg= {"hello"sv})> tmp;
}
=C2= =A0

--
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/5776c615-e654-4918-ba57-e77dca559456= %40isocpp.org.
------=_Part_13587_1883275120.1510735948136-- ------=_Part_13586_1529819433.1510735948135-- .