220 25890 article Path: news.gmane.org!not-for-mail From: Brent Friedman Newsgroups: gmane.comp.lang.c++.isocpp.proposals Subject: Re: RFC - memory management algorithms Date: Tue, 10 May 2016 23:32:08 -0500 Lines: 272 Approved: news@gmane.org Message-ID: References: <3d1f34cb-5376-f565-7b31-be27c5e2da0f@f2.dion.ne.jp> Reply-To: std-proposals@isocpp.org NNTP-Posting-Host: plane.gmane.org Mime-Version: 1.0 Content-Type: multipart/alternative; boundary=001a1141ab60471fc20532898621 X-Trace: ger.gmane.org 1462941132 19150 80.91.229.3 (11 May 2016 04:32:12 GMT) X-Complaints-To: usenet@ger.gmane.org NNTP-Posting-Date: Wed, 11 May 2016 04:32:12 +0000 (UTC) To: std-proposals@isocpp.org Original-X-From: std-proposals+bncBCHLPXGUXMJBBSHLZK4QKGQEGXXC2TI@isocpp.org Wed May 11 06:32:12 2016 Return-path: Envelope-to: gclcip-std-proposals@m.gmane.org Original-Received: from mail-pf0-f199.google.com ([209.85.192.199]) by plane.gmane.org with esmtp (Exim 4.69) (envelope-from ) id 1b0Loh-0002tV-Iu for gclcip-std-proposals@m.gmane.org; Wed, 11 May 2016 06:32:11 +0200 Original-Received: by mail-pf0-f199.google.com with SMTP id 203sf63580161pfy.2 for ; Tue, 10 May 2016 21:32:11 -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:date:message-id:subject:from: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=IAPcLP72orX/EqfMoKC8EfFl74RQ3JbuPASUNANtsoM=; b=kMzZTOy7zaCkQ0o9UZ0P8+lD2/u6mDhC5geqW6VkznrJF4Hqu6Y54mq9AoG3ZH3Ulu 879tSyox/2CpTMweiMYX+Gyzh+8dgEj/TJ++JHVg4BE7cgm3+bV/V6WdDFndKoDOfR4L Q2Z94hTtEpDtpfJ0Ftl8Z3qtMPiQ8MTs04WAQvmsVt0Z+l3UVDXl160WdQFC7PMTjj6s VIj+noD5GXtRNTRSMQnUD5L+sIRidTiPSc9miHGG08TqixwmFtHHbO/S63na3/pGPfak lRYLApF33ytFVF2/gwXTSrcTtqVmZOn3QbvzLKy+3ywCrSty1SHcj7SD762/x/r7mVcW PLZw== 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:date :message-id:subject:from: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=IAPcLP72orX/EqfMoKC8EfFl74RQ3JbuPASUNANtsoM=; b=SnDU8eqjrF4ET4lYcweQPoLyLJnq5dOzz77jJY7KrGectVRGfRN8SDSS7Oy5eSybJc 7TtqMMZfBGBM64CABM2AMoYn/kS/bRGTYhd/6epBipjPZSNYIBbq3O8+ywBZRMGB6vhK 47+DFsfYWtJGZDU9VFx9UfjskBCngQbMr88QS43JizZBDPulxsItrxqXi2IS9eVtdfmK mpUW8UNBuwXEqqMF4Mnn6TmrLNYD+DPnRIYy50JSukgTrW5+PFEzooqW1XGF5662znug 3ohwpoqari5SGYIkcKhUapE9HkqPV5iORT10KPc9sf1JnStz309fZ/2pHOCMvj6Dr2gG KQVA== X-Gm-Message-State: AOPr4FV5nxLV3YhgmfZfvRTnTmovu9kPd4gf91bK+HFN1qBAaS+gXh0BMhZ8Gk1kL2VqGA== X-Received: by 10.66.157.229 with SMTP id wp5mr811644pab.35.1462941129800; Tue, 10 May 2016 21:32:09 -0700 (PDT) X-BeenThere: std-proposals@isocpp.org Original-Received: by 10.157.10.169 with SMTP id 38ls249285otq.47.gmail; Tue, 10 May 2016 21:32:08 -0700 (PDT) X-Received: by 10.202.225.66 with SMTP id y63mr656677oig.193.1462941128588; Tue, 10 May 2016 21:32:08 -0700 (PDT) Original-Received: from mail-oi0-x230.google.com (mail-oi0-x230.google.com. [2607:f8b0:4003:c06::230]) by mx.google.com with ESMTPS id 188si1994684oie.49.2016.05.10.21.32.08 for (version=TLS1_2 cipher=ECDHE-RSA-AES128-GCM-SHA256 bits=128/128); Tue, 10 May 2016 21:32:08 -0700 (PDT) Received-SPF: pass (google.com: domain of fourthgeek@gmail.com designates 2607:f8b0:4003:c06::230 as permitted sender) client-ip=2607:f8b0:4003:c06::230; Original-Received: by mail-oi0-x230.google.com with SMTP id x19so49376891oix.2 for ; Tue, 10 May 2016 21:32:08 -0700 (PDT) X-Received: by 10.157.57.200 with SMTP id y66mr560665otb.169.1462941128267; Tue, 10 May 2016 21:32:08 -0700 (PDT) Original-Received: by 10.182.98.42 with HTTP; Tue, 10 May 2016 21:32:08 -0700 (PDT) In-Reply-To: <3d1f34cb-5376-f565-7b31-be27c5e2da0f@f2.dion.ne.jp> X-Original-Sender: fourthgeek@gmail.com X-Original-Authentication-Results: mx.google.com; dkim=pass header.i=@gmail.com; spf=pass (google.com: domain of fourthgeek@gmail.com designates 2607:f8b0:4003:c06::230 as permitted sender) smtp.mailfrom=fourthgeek@gmail.com; dmarc=pass (p=NONE dis=NONE) header.from=gmail.com 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:25890 Archived-At: --001a1141ab60471fc20532898621 Content-Type: text/plain; charset=UTF-8 > > Will you be in Oulu? > Hi Patrice! Yes, I'm booked to be in Oulu for the whole week. Katsutoshi Satoda, thanks for the detailed feedback! See below: > I don't see what "move-assignment of arrays" mean: builtin arrays can't be assigned. Does it mean the behavior of a default move constructor of a class with non static data member of array type? That's unfortunate. I forgot that assignment of arrays isn't a thing. Maybe I'll change this to refer to std::array or something else. Hm. The proposed wording says: > > Remarks: If an exception is thrown there are no effects on the range > > [first,last). You're right, this was intended to be "... no effects on the range [result, advance(result, distance(first,last))) ". IE, no effects on the destination rather than no effects on the source. "[first, first + count)" is more consistent with other descriptions. See fill_n() for example. As I understand, this + syntax is enabled by specific wording. n4582 25.1.12 devotes an entire paragraph to describing what + means in this context. I went with advance to avoid copy-pasting that. Would it be better to copy-paste this paragraph from [algorithms.general]? On Sat, May 7, 2016 at 9:29 AM, Kazutoshi Satoda wrote: > On 2016/05/06 0:59 +0900, Brent Friedman wrote: > > The paper proposes the following new algorithms: > > > > uninitialized_value_construct, uninitialized_value_construct_n > > uninitialized_default_construct, uninitialized_default_construct_n > > uninitialized_move, uninitialized_move_n > > destroy, destroy_n, destroy_at > > From "III. Discussion" about uninitialized_move(): > > Our approach to this problem is to emulate existing behavior throughout > > the language and library. uninitialized_move should behave like > > move-assignment of arrays. Namely, the constructed objects in the > > desination buffer are destroyed and the exception is propagated. This > > matches the behavior of uninitialized_copy and the behavior of algorithm > > move. > ("desination" here looks a typo of "destination") > > I don't see what "move-assignment of arrays" mean: builtin arrays can't > be assigned. Does it mean the behavior of a default move constructor of > a class with non static data member of array type? > > The algorithm move() doesn't destroy the destination on an exception. So > the mention here looks inappropriate. > > The proposed wording says: > > Remarks: If an exception is thrown there are no effects on the range > > [first,last). > I couldn't see what "on the range [first,last)" is meant for. It looks > saying nothing about the destination, meaning it doesn't match the above > quoted discussion. Also, "no effects on the range [first,last)" sounds > requiring to move back to the original range before destroying the > destination. But it is impossible in general. To match the above quoted > discussion, I think the Remarks should be like this: > Remarks: If an exception is thrown there are no effects except > possibly leaving some objects in the range [first,last) moved from. > Same issue applies to uninitialized_move_n(). > > > The following are some editorial issues: > > The names of Size argument of uninitialized_move_n() are inconsistent: > - "Size s" at synopsis > - "Size count" at declaration in [uninitialized.move] > - "n" in Effects > > In Remarks on uninitialized_move_n(), > > Remarks: If an exception is thrown there are no effects on the range > > [first,advance(first,count)). > "[first, first + count)" is more consistent with other descriptions. > See fill_n() for example. > > -- > k_satoda > > -- > 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/3d1f34cb-5376-f565-7b31-be27c5e2da0f%40f2.dion.ne.jp > . > -- 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/CADbh%2BeRNz%2Bn9rFbXS%3DSKLZ-z2GM7J_tymWgti5P66hEOF_Ungg%40mail.gmail.com. --001a1141ab60471fc20532898621 Content-Type: text/html; charset=UTF-8 Content-Transfer-Encoding: quoted-printable
Will you be in Oulu?

Hi Patrice! Yes, I'm booked to be in Oulu for the whole= week.


Katsutoshi Satoda, thanks fo= r the detailed feedback! See below:
=C2=A0
= =C2=A0I don't see what "move-assi= gnment of arrays" mean: builtin arrays can't=C2=A0
=C2=A0be assigned. Do= es it mean the behavior of a default move constructor of=C2=A0
=C2=A0a class with= non static data member of array type?

That's unfortunate. I forgot that assignment of arrays isn't a th= ing. Maybe I'll change this to refer to std::array or something else. H= m.

=
The proposed wording says:
>=C2=A0 =C2=A0Remarks: If= an exception is thrown there are no effects on the range
>=C2=A0 =C2=A0 = =C2=A0[first,last).

You're right= , this was intended to be "... no effects on the range [result, advanc= e(result, distance(first,last))) ". IE, no effects on the destination = rather than no effects on the source.
=C2=A0

=C2=A0"[first, first + count)" is more consistent with= other descriptions.=C2=A0
See fill_n() for example.

As I understand, this + syntax is enabled by specific wordi= ng. n4582 25.1.12 devotes an entire paragraph to describing what + means in= this context. I went with advance to avoid copy-pasting that. Would it be = better to copy-paste this paragraph from [algorithms.general]?
=

On Sat, May= 7, 2016 at 9:29 AM, Kazutoshi Satoda <k_satoda@f2.dion.ne.jp>= wrote:
On 2016/0= 5/06 0:59 +0900, Brent Friedman wrote:
> The paper proposes the following new algorithms:
>
> uninitialized_value_construct, uninitialized_value_construct_n
> uninitialized_default_construct, uninitialized_default_construct_n
> uninitialized_move, uninitialized_move_n
> destroy, destroy_n, destroy_at

From "III. Discussion" about uninitialized_move():
> Our approach to this problem is to emulate existing behavior throughou= t
> the language and library. uninitialized_move should behave like
> move-assignment of arrays. Namely, the constructed objects in the
> desination buffer are destroyed and the exception is propagated. This<= br> > matches the behavior of uninitialized_copy and the behavior of algorit= hm
> move.
("desination" here looks a typo of "destination")

I don't see what "move-assignment of arrays" mean: builtin ar= rays can't
be assigned. Does it mean the behavior of a default move constructor of
a class with non static data member of array type?

The algorithm move() doesn't destroy the destination on an exception. S= o
the mention here looks inappropriate.

The proposed wording says:
>=C2=A0 =C2=A0Remarks: If an exception is thrown there are no effects on= the range
>=C2=A0 =C2=A0 =C2=A0[first,last).
I couldn't see what "on the range [first,last)" is meant for.= It looks
saying nothing about the destination, meaning it doesn't match the abov= e
quoted discussion. Also, "no effects on the range [first,last)" s= ounds
requiring to move back to the original range before destroying the
destination. But it is impossible in general. To match the above quoted
discussion, I think the Remarks should be like this:
=C2=A0 Remarks: If an exception is thrown there are no effects except
=C2=A0 =C2=A0 possibly leaving some objects in the range [first,last) moved= from.
Same issue applies to uninitialized_move_n().


The following are some editorial issues:

The names of Size argument of uninitialized_move_n() are inconsistent:
=C2=A0 - "Size s" at synopsis
=C2=A0 - "Size count" at declaration in [uninitialized.move]
=C2=A0 - "n" in Effects

In Remarks on uninitialized_move_n(),
>=C2=A0 =C2=A0Remarks: If an exception is thrown there are no effects on= the range
>=C2=A0 =C2=A0 =C2=A0[first,advance(first,count)).
"[first, first + count)" is more consistent with other descriptio= ns.
See fill_n() for example.

--
k_satoda

--
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-propo= sals+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/3d1f34cb-5376-f565-7b31-be27= c5e2da0f%40f2.dion.ne.jp.

--
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/CADbh%2BeRN= z%2Bn9rFbXS%3DSKLZ-z2GM7J_tymWgti5P66hEOF_Ungg%40mail.gmail.com.
--001a1141ab60471fc20532898621-- .