References: <5366C875.9040709@googlemail.com>
            <536751EB.5010207@latex-project.org>
User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/24.4.50 (gnu/linux)
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Message-ID:  <87ha54tx4q.fsf@fencepost.gnu.org>
Date:         Mon, 5 May 2014 11:04:05 +0200
Reply-To: Mailing list for the LaTeX3 project
              <LATEX-L@LISTSERV.UNI-HEIDELBERG.DE>
Sender: Mailing list for the LaTeX3 project <LATEX-L@LISTSERV.UNI-HEIDELBERG.DE>
From: David Kastrup <dak@GNU.ORG>
Subject: Re: \__dim_strip_bp:n
To: LATEX-L@LISTSERV.UNI-HEIDELBERG.DE
In-Reply-To:  <536751EB.5010207@latex-project.org> (Frank Mittelbach's message
              of "Mon, 5 May 2014 10:55:07 +0200")
Precedence: list
Content-Transfer-Encoding: quoted-printable
Envelope-To: <rainer.schoepf@GMX.NET>
Status: R

Frank Mittelbach <frank.mittelbach@LATEX-PROJECT.ORG> writes:

> Am 05.05.2014 01:08, schrieb Heiko Oberdiek:
>> Improved version with higher precision:
>>
>>    \cs_new:Npn \__dim_strip_bp:n #1
>>      {
>>        \__dim_strip_pt:n
>>          {
>>            \__dim_eval:w ( #1 ) * 800 / 803 \__dim_eval_end:
>>          }
>>      }
>>
>> * 800/803 is the integer fraction for 72/72.27
>
>
> perhaps I'm completely off the mark, but isn't this introducing a
> restriction on the values that can be entered? (ie generating an
> overflow by doing *800 first)

You are completely off the mark.  Let me reinstate what you cut from
Heiko's posting:

>> Of course, there will always be rounding errors, but the code
>> can be improved by a scaling operation inside e-TeX's \dimexpr,
>> from "The e-TeX manual":
>>=20
>> | The arithmetic operations are performed individually, except
>> | for =E2=80=98scaling=E2=80=99 operations (a multiplication immediate=
ly followed
>> | by a division) which are performed as one combined operation
>> | with a 64-bit product as intermediate value.

--=20
David Kastrup