From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-patches-return-348077-listarch-gcc-patches=gcc.gnu.org@gcc.gnu.org>
Received: (qmail 13280 invoked by alias); 2 Sep 2013 18:24:10 -0000
Mailing-List: contact gcc-patches-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Id: <gcc-patches.gcc.gnu.org>
List-Archive: <http://gcc.gnu.org/ml/gcc-patches/>
List-Post: <mailto:gcc-patches@gcc.gnu.org>
List-Help: <mailto:gcc-patches-help@gcc.gnu.org>
Sender: gcc-patches-owner@gcc.gnu.org
Received: (qmail 13263 invoked by uid 89); 2 Sep 2013 18:24:09 -0000
Received: from qmta15.emeryville.ca.mail.comcast.net (HELO qmta15.emeryville.ca.mail.comcast.net) (76.96.27.228) by sourceware.org (qpsmtpd/0.93/v0.84-503-g423c35a) with ESMTP; Mon, 02 Sep 2013 18:24:09 +0000
Authentication-Results: sourceware.org; auth=none
X-Virus-Found: No
X-Spam-SWARE-Status: No, score=-2.5 required=5.0 tests=AWL,BAYES_00,FREEMAIL_FROM,KHOP_THREADED,RCVD_IN_DNSWL_NONE,RCVD_IN_HOSTKARMA_NO,RDNS_NONE,SPF_PASS autolearn=no version=3.3.2
X-HELO: qmta15.emeryville.ca.mail.comcast.net
Received: from omta16.emeryville.ca.mail.comcast.net ([76.96.30.72])	by qmta15.emeryville.ca.mail.comcast.net with comcast	id LJ041m0041ZMdJ4AFJQ6y9; Mon, 02 Sep 2013 18:24:06 +0000
Received: from up.mrs.kithrup.com ([24.4.193.8])	by omta16.emeryville.ca.mail.comcast.net with comcast	id LJQ61m00e0BKwT48cJQ6zZ; Mon, 02 Sep 2013 18:24:07 +0000
Content-Type: text/plain; charset=us-ascii
Mime-Version: 1.0 (Mac OS X Mail 6.5 \(1508\))
Subject: Re: [RFC] Changes to the wide-int classes
From: Mike Stump <mikestump@comcast.net>
In-Reply-To: <87ob8bbpkl.fsf@talisman.default>
Date: Mon, 02 Sep 2013 18:24:00 -0000
Cc: "gcc-patches\@gcc.gnu.org Patches" <gcc-patches@gcc.gnu.org>, rguenther@suse.de, Kenneth Zadeck <zadeck@naturalbridge.com>
Content-Transfer-Encoding: quoted-printable
Message-Id: <ABBB9A8C-FAF2-47CF-BA61-D36916564086@comcast.net>
References: <87wqn0bb5q.fsf@talisman.default> <5223F9D7.1030307@naturalbridge.com> <873699CE-3E8F-4285-8C6C-6B488B92982F@comcast.net> <87ob8bbpkl.fsf@talisman.default>
To: Richard Sandiford <rdsandiford@googlemail.com>
X-IsSubscribed: yes
X-SW-Source: 2013-09/txt/msg00084.txt.bz2

On Sep 2, 2013, at 1:22 AM, Richard Sandiford <rdsandiford@googlemail.com> =
wrote:
> What I'm saying is that the assert stops plain:
>=20
>   max_wide_int x =3D wide_int_var;

Sorry for being slow.  I see the point, indeed, from the preexisting code, =
we did:

  fixed_wide_int (const wide_int_ro w) : wide_int_ro (w) {
    /* We only allow the same size in, as otherwise=20=20=20=20=20=20=20=20=
=20=20=20=20=20=20=20=20=20=20=20=20=20=20=20=20=20=20=20=20=20=20=20=20=20=
=20=20=20=20=20=20=20=20=20=20=20
       we would not know how to extend it.  */
    gcc_assert (precision =3D=3D bitsize);
  }

to disallow it dynamically.  Doing this, or pushing it further into the typ=
e system so that it is caught a compile time, is fine.

I also like the passing of sign:

  max_wide_int x =3D max_wide_int::from (wide_int_var, SIGNED);
  max_wide_int x =3D max_wide_int::from (wide_int_var, UNSIGNED);

like this.  In the previous code we did:

  static inline fixed_wide_int from_wide_int (const wide_int& w) {
    if (w.neg_p (SIGNED))
      return w.sforce_to_size (bitsize);
    return w.zforce_to_size (bitsize);
  }

which is, as you point out, dangerous.