220 26318 article Path: news.gmane.org!not-for-mail From: "'Edward Catmur' via ISO C++ Standard - Future Proposals" Newsgroups: gmane.comp.lang.c++.isocpp.proposals Subject: Re: Re: statement folding on variadic templates parameters pack Date: Mon, 13 Jun 2016 15:32:28 +0100 Lines: 173 Approved: news@gmane.org Message-ID: References: <1e3daa7c-e10a-4589-8118-26f47bc1a0df@isocpp.org> <2d8458f8-27b2-464f-a16e-fc866ccfd2cd@isocpp.org> <8dc1e665-42a3-4ef0-a2d2-3773afab170c@isocpp.org> <20160612153937.4898897.23968.12102@gmail.com> <57d8c679-f312-4b77-8210-705f5e093b7e@isocpp.org> <62812f9c-7c4e-4a97-86a3-aee457c9f4a9@isocpp.org> Reply-To: std-proposals@isocpp.org NNTP-Posting-Host: plane.gmane.org Mime-Version: 1.0 Content-Type: multipart/alternative; boundary=001a114f3f7c0d9d67053529c215 X-Trace: ger.gmane.org 1465828359 8498 80.91.229.3 (13 Jun 2016 14:32:39 GMT) X-Complaints-To: usenet@ger.gmane.org NNTP-Posting-Date: Mon, 13 Jun 2016 14:32:39 +0000 (UTC) To: std-proposals@isocpp.org Original-X-From: std-proposals+bncBDZLZTXF7UJBB7MH7O5AKGQEPFN7SLA@isocpp.org Mon Jun 13 16:32:34 2016 Return-path: Envelope-to: gclcip-std-proposals@m.gmane.org Original-Received: from mail-ob0-f197.google.com ([209.85.214.197]) by plane.gmane.org with esmtp (Exim 4.69) (envelope-from ) id 1bCSun-00031J-Sk for gclcip-std-proposals@m.gmane.org; Mon, 13 Jun 2016 16:32:34 +0200 Original-Received: by mail-ob0-f197.google.com with SMTP id r6sf401453obx.1 for ; Mon, 13 Jun 2016 07:32:33 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=isocpp-org.20150623.gappssmtp.com; s=20150623; h=mime-version:in-reply-to:references:from:date:message-id:subject:to :x-original-sender:x-original-authentication-results:reply-to :precedence:mailing-list:list-id:x-spam-checked-in-group:list-post :list-help:list-archive:list-subscribe:list-unsubscribe; bh=tXcZ/oDJOmExYlezPmYGFYlGFir69X9ZY5nqPva8lb8=; b=aSkBNGTYP+ro3LUIWmrqtYnqZYehs54rXqDMSOuUq0KcX3oHhgkzoOsXDLES5eM5S+ 75vPcB5+8k5ZEhuvmxzk4BOxHe11QODzQByCZbUeSfXL6beWa8zkx0YNVXKXu68BfA2z zTQQeN8RN5zZEiGnIQo8EQjnB8WV1AbMyJmmdPG3GhAC+Mx6dMOkzbicf+8xcw4Ff8CK MfP1mJ3UTB4XI+fWC/Ybl4xAP8Z3UbZwrOpoTsdtcLVDasNEzQnkcFiqv4g0sPmQNkIx m2CXM/GVbb0AOyFX0GCiNZ7GrdjqoiP2Zbhv1gKlM8zs91YIAWzRQrVJYLiSmykRZ242 hemA== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20130820; h=x-gm-message-state:mime-version:in-reply-to:references:from:date :message-id:subject:to:x-original-sender :x-original-authentication-results:reply-to:precedence:mailing-list :list-id:x-spam-checked-in-group:list-post:list-help:list-archive :list-subscribe:list-unsubscribe; bh=tXcZ/oDJOmExYlezPmYGFYlGFir69X9ZY5nqPva8lb8=; b=lP8eT8Mjkn0/K48b+Sx0CuBjfjqUDHRlqsASHPz0s1xHvHtZ3sqHV/nQVeru8b98MG hcFr5S7HK3OzlXYT6MfHK+LaDgJSANg2Up4+kHmtibWlq/QJbLtx3B8n+3EHginAOqXf JvojuzEINKpT18isqxKy1++KY3KBrEo/bKTwKPfdppZKNKVjN5vwctP0I5qBDYPTdAr8 RwANoCDjsy/n+VP54xGRLWQAXMTWRIWbDvYGwnCSA8DaklNh28jCYeoKGSz8+r0rQvYz 6GrOgnnX6wUHAYaAJts57lLdb15jKIBJ8/PLCBcAOpJBX4TYZ1fhOm9c9oEdr67QQVrH cA+A== X-Gm-Message-State: ALyK8tLnphnAuYud7oNeG0hj4m3N5BMWBPgXYbC5QJjBnZS99abhuYQXoPMSrXHYDo8H5Q== X-Received: by 10.157.3.76 with SMTP id 70mr6343381otv.41.1465828352955; Mon, 13 Jun 2016 07:32:32 -0700 (PDT) X-BeenThere: std-proposals@isocpp.org Original-Received: by 10.140.18.103 with SMTP id 94ls3332531qge.60.gmail; Mon, 13 Jun 2016 07:32:29 -0700 (PDT) X-Received: by 10.140.153.135 with SMTP id 129mr14907864qhz.71.1465828349353; Mon, 13 Jun 2016 07:32:29 -0700 (PDT) Original-Received: from mail-qk0-x22f.google.com (mail-qk0-x22f.google.com. [2607:f8b0:400d:c09::22f]) by mx.google.com with ESMTPS id o202si15346004qke.46.2016.06.13.07.32.29 for (version=TLS1_2 cipher=ECDHE-RSA-AES128-GCM-SHA256 bits=128/128); Mon, 13 Jun 2016 07:32:29 -0700 (PDT) Received-SPF: pass (google.com: domain of ecatmur@googlemail.com designates 2607:f8b0:400d:c09::22f as permitted sender) client-ip=2607:f8b0:400d:c09::22f; Original-Received: by mail-qk0-x22f.google.com with SMTP id a186so19889735qkf.0 for ; Mon, 13 Jun 2016 07:32:29 -0700 (PDT) X-Received: by 10.233.237.129 with SMTP id c123mr14679472qkg.136.1465828349155; Mon, 13 Jun 2016 07:32:29 -0700 (PDT) Original-Received: by 10.237.55.198 with HTTP; Mon, 13 Jun 2016 07:32:28 -0700 (PDT) In-Reply-To: X-Original-Sender: ecatmur@googlemail.com X-Original-Authentication-Results: mx.google.com; dkim=pass header.i=@googlemail.com; spf=pass (google.com: domain of ecatmur@googlemail.com designates 2607:f8b0:400d:c09::22f as permitted sender) smtp.mailfrom=ecatmur@googlemail.com; dmarc=pass (p=QUARANTINE dis=NONE) header.from=googlemail.com X-Original-From: Edward Catmur Precedence: list Mailing-list: list std-proposals@isocpp.org; contact std-proposals+owners@isocpp.org List-ID: X-Spam-Checked-In-Group: std-proposals@isocpp.org 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:26318 Archived-At: --001a114f3f7c0d9d67053529c215 Content-Type: text/plain; charset=UTF-8 On Mon, Jun 13, 2016 at 4:11 AM, Arthur O'Dwyer wrote: > On Sun, Jun 12, 2016 at 6:17 PM, Edward Catmur wrote: > >> On Sunday, 12 June 2016 20:32:22 UTC+1, Arthur O'Dwyer wrote: >>> >>> I found the docs for Boost.Switch >>> . If I get >>> the time, I'll try to write up a modern library implementation of the above >>> make_switch and post it here. I'm sure it's been done before, of >>> course, so if someone beats me to it that's great. >>> >> >> Too many times to count, I'm sure! I decided to take another go at it, >> trying to conform to the Boost.Switch interface but using modern >> techniques, using only the standard library and Boost.Preprocessor for >> iteration. My implementation is a little under 60 lines: >> https://gist.github.com/ecatmur/7f5f8b44e70f414742e4eae1efdd0ca7 >> >> Interesting! My implementation ended up at twice the length of yours, > but IMHO has a nicer user interface than Boost.Switch. > > https://gist.github.com/Quuxplusone/3f7100f72551f6a7028844b298c43508 > > The "apply" example in my case looks like this: extra boilerplate to deal > with making index sequences, but the tradeoff is that my interface allows > you to switch on things that aren't consecutive integers. > Actually, mine can switch on any integer sequence; it uses the preprocessor iteration over N to index into the integer sequence for the case label constant (using std::tuple_element, since that's the only available facility in the standard library for variadic indexing). The other tradeoff is that my Clang isn't smart enough to turn this into a > switch. It will go as far as inlining everything into an if-else chain (no > function calls), but it won't make a jump table. Yes, this kind of defeats > the purpose; but I'm holding out hope that it's a Quality of Implementation > issue rather than a fundamental flaw. > Yes, it's a bit disappointing that the free compilers are unable to turn a recursive if chain into a jump table. Using a syntactic switch has other benefits, though; it allows the compiler to automatically detect repeated cases, and also to detect omitted cases where the controlling type is an enum. > The "enum to integer" example in your gist doesn't compile on my Clang > because it is trying to instantiate std::integer_sequence where E (being > an enum) is not an integral type; I think you have to add some boilerplate > involving std::conditional_t, > std::underlying_type_t, E> and some explicit casts, at which point it > doesn't even seem worth it anymore. > Hm. Clang works fine for me in versions 3.5 upward in -std=c++14 mode. > Having to create an entity of the form foo<..., E::A, E::B, E::C> should > have been a red flag anyway, IMHO. > That sequence would usually be provided by an introspection library (BetterEnums, or whatever SG7 are up to these days), or by a code generator. -- 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/CAJnLdOY4UaYaeVGWVzNa5kwM3P-Gw8P350n7ggVDjyOEHFrC2Q%40mail.gmail.com. --001a114f3f7c0d9d67053529c215 Content-Type: text/html; charset=UTF-8 Content-Transfer-Encoding: quoted-printable
On M= on, Jun 13, 2016 at 4:11 AM, Arthur O'Dwyer <arthur.j.odwyer@g= mail.com> wrote:
On Sun, Jun 12, 2016 at 6:17 PM, Edward Catm= ur <ed@catmur.co.uk> wrote:
On Sunday, 12 June 2016 20:32:22 UTC+1, Arthur O'Dwye= r wrote:
=
I found the docs for Boost.Switch. I= f I get the time, I'll try to write up a modern library implementation = of the above make_switch=C2=A0an= d post it here. I'm sure it's been done before, of course, so if so= meone beats me to it that's great.
=

Too many times to count, I'm sure! I decided= to take another go at it, trying to conform to the Boost.Switch interface = but using modern techniques, using only the standard library and Boost.Prep= rocessor for iteration. My implementation is a little under 60 lines:=C2=A0= https://gist.github.com/ecatmur/7f5f8b44e70f414742e4eae= 1efdd0ca7=C2=A0

Interesti= ng!=C2=A0 My implementation ended up at twice the length of yours, but IMHO= has a nicer user interface than Boost.Switch.


The "apply" example i= n my case looks like this: extra boilerplate to deal with making index sequ= ences, but the tradeoff is that my interface allows you to switch on things= that aren't consecutive integers.
=

Actually, mine can switch on any integer sequence; it u= ses the preprocessor iteration over N to index into the integer sequence fo= r the case label constant (using std::tuple_element, since that's the o= nly available facility in the standard library for variadic indexing).

The other tradeoff is that= my Clang isn't smart enough to turn this into a switch. It will go as = far as inlining everything into an if-else chain (no function calls), but i= t won't make a jump table. Yes, this kind of defeats the purpose; but I= 'm holding out hope that it's a Quality of Implementation issue rat= her than a fundamental flaw.

<= /div>
Yes, it's a bit disappointing that the free compilers are una= ble to turn a recursive if chain into a jump table. Using a syntactic switc= h has other benefits, though; it allows the compiler to automatically detec= t repeated cases, and also to detect omitted cases where the controlling ty= pe is an enum.
=C2=A0
=
The "enum to integer" example in your gist doesn't compi= le on my Clang because it is trying to instantiate=C2=A0std::integer_sequence<E, ...>=C2=A0where=C2= =A0E=C2=A0(being an enum) is not= an integral type; I think you have to add some boilerplate involving=C2=A0= std::conditional_t<std::is_enum_v<= ;E>, std::underlying_type_t<E>, E>=C2=A0and some explici= t casts, at which point it doesn't even seem worth it anymore.

Hm. Clang works fine for m= e in versions 3.5 upward in -std=3Dc++14 mode.
=C2=A0
Having to create an entity of the form=C2=A0foo<..., E::A, E::B, E::C>=C2= =A0should have been a red flag anyway, IMHO.

That sequence would usually be provided by an int= rospection library (BetterEnums, or whatever SG7 are up to these days), or = by a code generator.=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 htt= ps://groups.google.com/a/isocpp.org/d/msgid/std-proposals/CAJnLdOY4UaYaeVGW= VzNa5kwM3P-Gw8P350n7ggVDjyOEHFrC2Q%40mail.gmail.com.
--001a114f3f7c0d9d67053529c215-- .