% -*- coding: iso-latin-1; time-stamp-format: "%02d-%02m-%:y at %02H:%02M:%02S %Z" -*-
%<*dtx>
\def\bnedtxtimestamp  {Time-stamp: <04-11-2014 at 12:11:09 CET>}
\iffalse
%</dtx>
%<*drv>
%% ---------------------------------------------------------------
\def\bnedocdate {2014/11/04} % package bnumexpr documentation date
\def\bnepackdate{2014/10/28} % package bnumexpr date
\def\bneversion {1.1b}       % package bnumexpr version
%</drv>
%<*readme>--------------------------------------------------------

| Source:  bnumexpr.dtx
| Version: v1.1b, 2014/10/28 (doc: 2014/11/04)
| Author:  Jean-Francois Burnol
| Info:    Expressions with big integers
| License: LPPL 1.3c or later

README: [Usage], [Installation], [License]
==========================================

Usage
-----

    \usepackage{bnumexpr}

Then

    \thebnumexpr <expression with +,-,*,/,(,)> \relax

is like

    \the\numexpr <expression with +,-,*,/,(,)> \relax

with the difference of accepting or producing arbitrarily big
integers. For example,

    \thebnumexpr 2*1234567890\relax

outputs `2469135780` which would have created an arithmetic overflow in
`\numexpr` as it exceeds the maximal allowed TeX integer `2147483647`.

`\bnumexpr...\relax` is a scaled down version of `\xintiiexpr...\relax`
from package xintexpr.[^1]

- by default, bnumexpr loads xintcore [^1] for its arithmetic macros doing
  addition, subtraction, multiplication, division (and powers).

- option _allowpower_ enables `^` as power operator, for example:

        \thebnumexpr 2^31\relax % smallest integer exceeding the TeX bound

- option _bigintcalc_ loads package bigintcalc [^2] and uses its
  arithmetic macros rather than those from xintcore.

- option _l3bigint_ loads package l3bigint [^3], from the experimental
  trunk of the on-going [LaTeX3 project](http://latex-project.org).

- with option _custom_, no extra package is loaded and it is up to the
  user to define suitably expandable macros `\bnumexprAdd`,
  `\bnumexprSub`, `\bnumexprMul`, and `\bnumexprDiv` doing the basic
  arithmetic operations.

[^1]: <http://www.ctan.org/pkg/xint>

[^2]: <http://www.ctan.org/pkg/bigintcalc>

[^3]: <http://latex-project.org/svnroot/experimental/trunk/l3trial/l3bigint>,
or from <https://github.com/latex3/svn-mirror>.

Installation
------------

Obtain `bnumexpr.dtx` (and possibly, `bnumexpr.ins` and the `README`)
from CTAN:

> <http://www.ctan.org/pkg/bnumexpr>

Both `"tex bnumexpr.ins"` and `"tex bnumexpr.dtx"` extract from
`bnumexpr.dtx` the following files:

`bnumexpr.sty`
  : this is the style file.

`bnumexprReadme.md`
  : reconstitutes this README.

`bnumexpr.changes`
  : lists changes from the initial version.

`bnumexpr.tex`
  : can be used to generate the documentation:

    - with latex+dvipdfmx: `"latex bnumexpr.tex"` (thrice) then
      `"dvipdfmx bnumexpr.dvi"`.

        Ignore dvipdfmx warnings, but if the pdf file has problems with
      fonts (possibly from an old dvipdfmx), use then rather pdflatex.

    - with pdflatex: `"pdflatex bnumexpr.tex"` (thrice).

    In both cases files `bnumexprReadme.md` and `bnumexpr.changes` must
    be present in the same repertory.

without `bnumexpr.tex`:
  : `"pdflatex bnumexpr.dtx"` (thrice) extracts all files and
    simultaneously generates the pdf documentation.

Finishing the installation:

           bnumexpr.sty   --> TDS:tex/latex/bnumexpr/

           bnumexpr.dtx   --> TDS:source/latex/bnumexpr/
           bnumexpr.ins   --> TDS:source/latex/bnumexpr/

           bnumexpr.pdf   --> TDS:doc/latex/bnumexpr/
                 README   --> TDS:doc/latex/bnumexpr/

Files `bnumexpr.tex`, `bnumexpr.changes`, `bnumexprReadme.md` may be
discarded.

License
-------

Copyright (C) 2014 by Jean-Francois Burnol

| This Work may be distributed and/or modified under the
| conditions of the LaTeX Project Public License, either
| version 1.3c of this license or (at your option) any later
| version. This version of this license is in

>   <http://www.latex-project.org/lppl/lppl-1-3c.txt>

| and the latest version of this license is in

>   <http://www.latex-project.org/lppl.txt>

| and version 1.3 or later is part of all distributions of
| LaTeX version 2005/12/01 or later.

This Work has the LPPL maintenance status "maintained".

The Current Maintainer of this Work is Jean-Francois Burnol.

This Work consists of the main source file `bnumexpr.dtx`
and the derived files

    bnumexpr.sty, bnumexpr.pdf, bnumexpr.ins, bnumexpr.tex,
    bnumexpr.changes, bnumexprReadme.md

%</readme>--------------------------------------------------------
%<*!readme>
%% ---------------------------------------------------------------
%% The bnumexpr package: Expressions with big integers
%% Copyright (C) 2014 by Jean-Francois Burnol
%%
%</!readme>
%<*changes>-------------------------------------------------------
\item[1.1b (2014/10/28)]
  \begin{itemize}
  \item README converted to |markdown/pandoc| syntax,
  \item the package now loads only |xintcore|, which belongs to |xint| bundle
    version |1.1| and extracts from the earlier |xint| package the core arithmetic
    operations as used by |bnumexpr|.
  \end{itemize}

\item[1.1a (2014/09/22)]
  \begin{itemize}
  \item added |l3bigint| option to use experimental \LaTeX3
    package of the same name,

  \item added Changes and Readme sections to the documentation,

  \item better |\BNE_protect| mechanism for use of
  |\bnumexpr...\relax| inside an |\edef| (without |\bnethe|). Previous one,
  inherited from |xintexpr.sty 1.09n|, assumed that the |\.=<digits>| dummy
  control sequence encapsulating the computation result had |\relax|
  meaning. But removing this assumption was only a matter of letting
  |\BNE_protect| protect two, not one, tokens. This will be backported to
  next version of |xintexpr.sty|, naturally.
  \end{itemize}

\item[1.1 (2014/09/21)] First release. This is down-scaled from the
  (development version of) |xintexpr.sty|. Motivation came the previous day
  from a chat with \textsc{Joseph Wright} over big int status in \LaTeX3.
  The |\bnumexpr...\relax| parser can be used on top of big int macros of
  one's choice. Functionalities limited to the basic operations. I leave
  the power operator |^| as an option.
%</changes>-------------------------------------------------------
%<*drv>-----------------------------------------------------------
%% latex bnumexpr.tex (thrice) && dvipdfmx bnumexpr.dvi
%% to produce bnumexpr.pdf
%%
%% or pdflatex bnumexpr.tex (no need to change \Withdvipdfmx toggle,
%% pdf engine will be detected automatically)
%%
\NeedsTeXFormat{LaTeX2e}
\ProvidesFile{bnumexpr.tex}%
[\bnepackdate\space v\bneversion\space driver file for %
  bnumexpr documentation (jfB)]%
\PassOptionsToClass{a4paper,fontsize=11pt}{scrdoc}
\chardef\Withdvipdfmx 1 %
\chardef\NoSourceCode 0 % replace 0 by 1 for not including source code
\input bnumexpr.dtx
%%% Local Variables:
%%% mode: latex
%%% End:
%</drv>-----------------------------------------------------------
%<*ins>-----------------------------------------------------------
%% tex bnumexpr.ins will extract bnumexpr.sty from bnumexpr.dtx
%%
\input docstrip.tex
\askforoverwritefalse
\generate{\nopreamble\nopostamble
\file{bnumexprReadme.md}{\from{bnumexpr.dtx}{readme}}
\usepreamble\defaultpreamble
\usepostamble\defaultpostamble
\file{bnumexpr.changes}{\from{bnumexpr.dtx}{changes}}
\file{bnumexpr.tex}{\from{bnumexpr.dtx}{drv}}
\file{bnumexpr.sty}{\from{bnumexpr.dtx}{package}}}
\catcode32=13\relax% active space
\let =\space%
\Msg{************************************************************************}
\Msg{*}
\Msg{* To finish the installation you have to move the following}
\Msg{* files into a directory searched by TeX:}
\Msg{*}
\Msg{*     bnumexpr.sty}
\Msg{*}
\Msg{* To produce the documentation run latex thrice on bnumexpr.tex}
\Msg{* then dvipdfmx on bnumexpr.dvi. (ignore the dvipdfmx warnings)}
\Msg{*}
\Msg{* Happy TeXing!}
\Msg{*}
\Msg{************************************************************************}
\endbatchfile
%</ins>-----------------------------------------------------------
%<*dtx>
\fi % end of \iffalse block around generated files
\chardef\noetex 0
\ifx\numexpr\undefined \chardef\noetex 1 \fi
\ifnum\noetex=1 \chardef\extractfiles 0 % extract files, then stop
\else
    \ifx\ProvidesFile\undefined
      \chardef\extractfiles 0 % no LaTeX2e; etex, ... on bnumexpr.dtx
    \else % latex/pdflatex on bnumexpr.tex or on bnumexpr.dtx
      \ifx\Withdvipdfmx\undefined
        % latex run is on bnumexpr.dtx, we will extract all files
        \chardef\extractfiles 1 % 1 = extract and typeset, 2=only typeset
        \chardef\Withdvipdfmx 0 % 0 = pdflatex or latex+dvips, 1 = dvipdfmx
        \chardef\NoSourceCode 0 % 0 = include source code, 1 = do not
        \NeedsTeXFormat{LaTeX2e}%
        \PassOptionsToClass{a4paper,fontsize=11pt}{scrdoc}%
      \else % latex run is on bnumexpr.tex,
        \chardef\extractfiles 2 % no extractions
      \fi
      \ProvidesFile{bnumexpr.dtx}[bundle source (\bnedtxtimestamp)]%
    \fi
\fi
\ifnum\extractfiles<2 % extract files
\def\MessageDeFin{\newlinechar10 \let\Msg\message
\Msg{^^J}%
\Msg{********************************************************************^^J}%
\Msg{*^^J}%
\Msg{* To finish the installation you have to move the following^^J}%
\Msg{* files into a directory searched by TeX:^^J}%
\Msg{*^^J}%
\Msg{*\space\space\space\space bnumexpr.sty^^J}%
\Msg{*^^J}%
\Msg{* To produce the documentation run latex thrice on bnumexpr.tex^^J}%
\Msg{* then dvipdfmx on bnumexpr.dvi. (ignore the dvipdfmx warnings)^^J}%
\Msg{*^^J}%
\Msg{* Happy TeXing!^^J}%
\Msg{*^^J}%
\Msg{********************************************************************^^J}%
}%
\begingroup
    \input docstrip.tex
    \askforoverwritefalse
    \generate{\nopreamble\nopostamble
    \file{bnumexprReadme.md}{\from{bnumexpr.dtx}{readme}}
    \usepreamble\defaultpreamble
    \usepostamble\defaultpostamble
    \file{bnumexpr.changes}{\from{bnumexpr.dtx}{changes}}
    \file{bnumexpr.ins}{\from{bnumexpr.dtx}{ins}}
    \file{bnumexpr.tex}{\from{bnumexpr.dtx}{drv}}
    \file{bnumexpr.sty}{\from{bnumexpr.dtx}{package}}}
\endgroup
\fi % end of file extraction
\ifnum\extractfiles=0
% direct tex/etex/xetex on bnumexpr.dtx, files now extracted, stop
  \MessageDeFin\expandafter\end
\fi
% no use of docstrip to extract files if latex compilation was on bnumexpr.tex
\ifdefined\MessageDeFin\AtEndDocument{\MessageDeFin}\fi
%-------------------------------------------------------------------------------
\documentclass {scrdoc}
\ifnum\NoSourceCode=1 \OnlyDescription\fi
\usepackage{ifpdf}
\ifpdf\chardef\Withdvipdfmx 0 \fi
\makeatletter
\ifnum\Withdvipdfmx=1
   \@for\@tempa:=hyperref,bookmark,graphicx,xcolor,pict2e\do
            {\PassOptionsToPackage{dvipdfmx}\@tempa}
   %
   \PassOptionsToPackage{dvipdfm}{geometry}
   \PassOptionsToPackage{bookmarks=true}{hyperref}
   \PassOptionsToPackage{dvipdfmx-outline-open}{hyperref}
   \PassOptionsToPackage{dvipdfmx-outline-open}{bookmark}
   %
   \def\pgfsysdriver{pgfsys-dvipdfm.def}
\else
   \PassOptionsToPackage{bookmarks=true}{hyperref}
\fi
\makeatother

\pagestyle{headings}
\makeatletter
\def\buggysectionmark #1{% KOMA 3.12 as released to CTAN December 2013
    \if@twoside\expandafter\markboth\else\expandafter\markright\fi
    {\MakeMarkcase{\ifnumbered{section}{\sectionmarkformat\fi}{}#1}}{}}
\ifx\buggysectionmark\sectionmark
\def\sectionmark #1{%
    \if@twoside\expandafter\markboth\else\expandafter\markright\fi
    {\MakeMarkcase{\ifnumbered{section}{\sectionmarkformat}{}#1}}{}}
\fi
\makeatother

\usepackage[T1]{fontenc}
\usepackage[latin1]{inputenc}

\usepackage[hscale=0.66,vscale=0.75]{geometry}

\usepackage[zerostyle=a,scaled=0.95]{newtxtt}
\renewcommand\familydefault\ttdefault
\usepackage[noendash]{mathastext}
\renewcommand\familydefault\sfdefault

\usepackage{graphicx}
\usepackage[dvipsnames]{xcolor}
\definecolor{joli}{RGB}{225,95,0}
\definecolor{JOLI}{RGB}{225,95,0}
\definecolor{BLUE}{RGB}{0,0,255}
\definecolor{niceone}{RGB}{38,128,192}
\colorlet{jfverbcolor}{yellow!5}

% transféré de xint-manual.tex
\DeclareFontFamily{U}{MdSymbolC}{}
\DeclareFontShape {U}{MdSymbolC}{m}{n}{<-> MdSymbolC-Regular}{}

\makeatletter
\newbox\cdbx@SoftWrapIcon
\colorlet {softwrapicon}{blue}

\def\cdbx@SetSoftWrapBox{%
    \setbox\cdbx@SoftWrapIcon\hb@xt@\z@
    {\hb@xt@\fontdimen2\font
        {\hss{\color{softwrapicon}\usefont{U}{MdSymbolC}{m}{n}\char"97}\hss}%
     \hss}%
   }
\makeatother

\usepackage[english]{babel}

\usepackage[pdfencoding=pdfdoc]{hyperref}
\hypersetup{%
%linktoc=all,%
breaklinks=true,%
colorlinks=true,%
urlcolor=niceone,%
linkcolor=blue,%
pdfauthor={Jean-Fran\c cois Burnol},%
pdftitle={The bnumexpr package},%
pdfsubject={Arithmetic with TeX},%
pdfkeywords={Expansion, arithmetic, TeX},%
pdfstartview=FitH,%
pdfpagemode=UseOutlines}
\usepackage{bookmark}

%---- \verb, and verbatim like `environments'. \MicroFont et \MacroFont
\makeatletter
\def\MicroFont {\ttfamily \cdbx@SetSoftWrapBox }

% \MacroFont est utilisé par macrocode, mais sa définition est figée dans
% \macro@font au \begin{document}

\def\MacroFont {\ttfamily \baselineskip12pt\relax }

%--- November 4, 2014 making quotes straight.
%
% there is no hook in \macrocode after \dospecials etc. Thus I will need to
% take the risk that some future evolution of doc.sty (or perhaps scrdoc)
% invalidates the following.
%
% Actually, I should not at all rely on the doc class, I should do it all by
% myself.

\def\macrocode{\macro@code
               \frenchspacing \@vobeyspaces
               \makestarlowast\makequotesstraight
               \xmacro@code }

%--- lower * symbol in text

\def\lowast{\raisebox{-.25\height}{*}}
\catcode`* \active
\def\makestarlowast {\let*\lowast\catcode`\*\active}%
\catcode`* 12

%--- straight quotes, added (finally...) Nov 4, 2014

\begingroup\makeatletter
  \catcode`\'\active
  \catcode`\`\active
\@firstofone {\endgroup
  \def\makequotesstraight {% textcomp package has been loaded by newtxtt
     \let`\textasciigrave
     \let'\textquotesingle
     \catcode39\active
     \catcode96\active }%
}

% Tentative, Mardi 09 septembre 2014 à 22:41:28
% (not usable in footnotes)
\def\verb
% on pourrait mettre le #1 ici et économiser \@@jfverb
{%
  \relax\leavevmode\null
  \begingroup\MicroFont
  %\let\do\do@noligs  \verbatim@nolig@list % not needed here
  \let\do\@makeother \dospecials
  \makestarlowast \makequotesstraight % added here Nov 4, 2014
% I should also take care of verbatim environment, but this is not needed in
% bnumexpr.dtx 
  \@vobeyspaces \fboxsep\z@
  \@@jfverb
}%

\def\@@jfverb #1{\catcode`#1 3 \@@@jfverb }

\def\@@@jfverb #1{\ifcat\noexpand#1$% $
                      \endgroup\else
%      \penalty\z@
      \discretionary{\copy\cdbx@SoftWrapIcon}{}{}%
      \colorbox{jfverbcolor}{\strut #1}% unfortunately, impacts pdf size
      \expandafter\@@@jfverb\fi }

\makeatother

% currently, used only in (sub)-section titles from implementation part
% let's use there also sans font, not monotype

\DeclareRobustCommand\csa [1]{{\char92\detokenize{#1}}}
\newcommand\csh[1]{\texorpdfstring{\csa{#1}}{\textbackslash\detokenize{#1}}}

\usepackage{xspace}

\def\bnename
   {\texorpdfstring
          {\hyperref[sec:bnumexpr]{{\color{joli}\ttzfamily bnumexpr}}}
          {bnumexpr}%
     \xspace }%
\def\bnenameimp
   {\texorpdfstring
          {\hyperref[sec:bnumexprcode]{{\color[named]{RoyalPurple}\ttzfamily bnumexpr}}}
          {bnumexpr}%
    \xspace }%


\frenchspacing
% possible options: custom, bigintcalc, l3bigint, nocsv, notacitmul, allowpower
\usepackage[allowpower]{bnumexpr}

\usepackage{etoc}

\begin{document}
\thispagestyle{empty}
\ttzfamily
\pdfbookmark[1]{Title page}{TOP}

{%
\normalfont\Large\parindent0pt \parfillskip 0pt\relax
 \leftskip 2cm plus 1fil \rightskip 2cm plus 1fil
 The \bnename package\par
}

{\centering
  \textsc{Jean-François Burnol}\par
  \footnotesize
  jfbu (at) free (dot) fr\par
  Package version: \bneversion\ (\bnepackdate);
            documentation date: \bnedocdate.\par
  {From source file \texttt{bnumexpr.dtx}. \bnedtxtimestamp.}\par
}

\etocsetnexttocdepth{section}
\tableofcontents

\section{Readme}
\begingroup
\makeatletter\def\x{\baselineskip10pt
                    \ttfamily\settowidth\dimen@{X}%
                    %\parindent \dimexpr.5\linewidth-34\dimen@\relax
                    \parindent\z@
                    \let\do\do@noligs\verbatim@nolig@list
                    \let\do\@makeother\dospecials
                    \makestarlowast\makequotesstraight
                    %\def\par{\leavevmode\null\@@par\penalty\interlinepenalty}%
                    \def\par{\leavevmode\null\@@par\pagebreak[1]}%
                    \@vobeyspaces\obeylines
                    \noindent\kern\parindent\input bnumexprReadme.md
\endgroup }\x


\section{Introduction}
\label{sec:bnumexpr}

Package \bnename provides |\bnumexpr...\relax| which is analogous to
|\numexpr...\relax|, while allowing arbitrarily big
integers. Important items:
\begin{enumerate}
\item one must use either |\thebnumexpr| or |\bnethe\bnumexpr| to get a
  printable result, as |\bnumexpr...\relax| expands to a private format;
  however one may embed directly one |\bnumexpr...\relax| in another
  |\bnumexpr...\relax|,

\item contrarily to |\numexpr|, the |\bnumexpr| parser stops only after having
  found (and swallowed) the mandatory ending |\relax| token,

\item in particular spaces between digits do not stop |\bnumexpr|, in contrast
  with |\numexpr|:

  |\the\numexpr 3 5+79\relax| expands (in one step) to \expandafter|\the\numexpr 3 5+79\relax|

  |\thebnumexpr 3 5+79\relax| expands (in two steps) to \expandafter\expandafter\expandafter|\thebnumexpr 3 5+79\relax|

\item one may do |\edef\tmp{\bnumexpr 1+2\relax}|, and then either use |\tmp|
  in another |\bnumexpr...\relax|, or print it via |\bnethe\tmp|. The
  computation is done at the time of the |\edef| (and two expansion steps
  suffice). This is again in contrast with |\numexpr...\relax| which, without
  |\the| or |\number| as prefix would not expand in an |\edef|,

\item tacit multiplication applies in front of parenthesized sub-expressions, or
  sub |\bnumexpr...\relax| (or |\numexpr...\relax|), or in front of a |\count|
  or |\dimen| register. This may be de-activated by option |notacitmul|,

\item expressions may be comma separated. On input, spaces are ignored,
  naturally, and on output the values are comma separated with a space after
  each comma. This functionality may be turned off via option |nocsv|,

\item even with options |notacitmul| and |nocsv| the syntax is more flexible
  than with |\numexpr|: things such as
  |\bnumexpr -(1+1)\relax| are legal.
\end{enumerate}

The parser |\bnumexpr| is a scaled-down version of parser |\xintiiexpr| from
package \href{http://www.ctan.org/pkg/xint}{xintexpr}. The goal here is to
extend |\numexpr| only to the extent of accepting big integers. Thus by
default, the syntax allows |+,-,*,/|, parentheses, and also |\count| or
|\dimen| registers or variables. Option |allowpower| is a bonus which enables
|^| as power operator. It may well be that the code of the parser is in some
places quite sub-optimal as it was derived from code handling far more stuff.

The |\xintNewExpr| construct has been left out.

Package \bnename needs some underlying big integer engine to provide the
macros doing the actual computations. By default, it loads package |xintcore| (a
subset of \href{http://www.ctan.org/pkg/xint}{xint}) and its |\xintiiAdd|,
|\xintiiSub|, |\xintiiMul|, and |\xintiiDivRound| macros (also |\xintiiPow| if
option |allowpower| is made use of).
See the discussion of options |bigintcalc|, |l3bigint| and |custom| in
the next section %\autoref{sec:options}
for alternatives.

I recall from documentation of |xintexpr| that there is a potential impact on
the memory of \TeX{} (the hash table) because each arithmetic operation is
done inside a dummy |\csname...\endcsname| used as single token to move
around in one-go the possibly hundreds of digits composing a number. This
becomes apparent only if the document does (tens of) thousands of evaluations.

\section{Options}\label{sec:options}

The package does by default:
\begin{verbatim}
\RequirePackage{xintcore}
\let\bnumexprAdd\xintiiAdd
\let\bnumexprSub\xintiiSub
\let\bnumexprMul\xintiiMul
\let\bnumexprMul\xintiiDivRound
\let\bnumexprPow\xintiiPow % only if option allowpower
\end{verbatim}
% To let |/| do euclidean division (like currently in |\xintiiexpr|) it is
% thus sufficient to do |\let\bnumexprDiv\xintiiQuo| after loading the package.

Option |bigintcalc| says to not load
\href{http://www.ctan.org/pkg/xint}{xintcore}
but to use rather the macros from package
\href{http://www.ctan.org/pkg/bigintcalc}{bigintcalc}
by \textsc{Heiko Oberdiek}. Note though that |/| is mapped to |\bigintcalcDiv|
which does \emph{truncated} (not rounded) division.

Option |l3bigint| similarly says to use the macros which are provided with the
eponym package, a part of the development work of the
\href{http://latex-project.org/code.html}{\LaTeX3 project}. There is no power
operation available with this option.

Option |custom| does not load any package and leaves it up to the user to
specify the macros to be used, i.e. provide definitions for |\bnumexprAdd|,
|\bnumexprSub|, |\bnumexprMul|, |\bnumexprDiv| (and possibly |\bnumexprPow|).
These macros must be expandable, and they must allow arguments in need to be
first (`f'-) expanded. They should produce on output (big) integers with no
leading zeros, at most one minus sign and no plus sign (else the \bnename macro
used for handling the |-| prefix operator may need to be modified). They will be
expanded inside |\csname...\endcsname|. The macros from |xintcore.sty| (as well
as those of |bigintcalc.sty|) are expandable in a stronger sense (only two
expansion steps suffice). Perhaps speed gains are achievable from dropping these
stronger requirements.

Option |nocsv| makes comma separated expressions illegal.

Option |notacitmul| removes the possibility of tacit multiplication in front of
parentheses, \string\count\space registers, sub-expressions.

Option |allowpower| enables the |^| as power operator (left associative).

\section{Further commands}

The package provides |\bnumexprUsesxint|, |\bnumexprUsesbigintcalc| and
|\bnumexprUsesliiibigint|: after issuing one of these commands,
|\bnumexpr...\relax| will use the arithmetic macros from the corresponding
package. But it is up to the user to have issued the necessary |\usepackage|
or |\RequirePackage| in the preamble. It does not matter if the packages are
loaded before or after \bnename.

Recall that |xintcore| is loaded by default, but is not loaded in case of one
of the options |custom|, |bigintcalc|, |l3bigint|.


\section{Examples}

\begin{flushleft}\parindent0pt \obeylines
|\thebnumexpr 128637867168*2187917891279\relax| %
\thebnumexpr 128637867168*2187917891279\relax

|\thebnumexpr 30*(21-43*(512-67*(6133-812*2897)))\relax| %
\thebnumexpr 30*(21-43*(512-67*(6133-812*2897)))\relax

\newcount\cnta\cnta 123
\newcount\cntb \cntb 188

|\newcount\cnta \cnta 123 \newcount\cntb \cntb 188|
% |\the\numexpr \cnta*\cnta*\cnta\relax|  %
% \the\numexpr \cnta*\cnta*\cnta\relax {}
% |\the\bnumexpr \cnta*\cnta*\cnta\relax| %
% \thebnumexpr \cnta*\cnta*\cnta\relax {}
|\the\numexpr \cnta*\cnta*\cnta/\cntb+\cntb*\cntb*\cntb/\cnta\relax| %
\the\numexpr \cnta*\cnta*\cnta/\cntb+\cntb*\cntb*\cntb/\cnta\relax {}
|\thebnumexpr \cnta*\cnta*\cnta/\cntb+\cntb*\cntb*\cntb/\cnta\relax|  %
\thebnumexpr \cnta*\cnta*\cnta/\cntb+\cntb*\cntb*\cntb/\cnta\relax   {}

% |\the\numexpr \cnta*\cnta*(\cnta/\cntb)+\cntb*\cntb*(\cntb/\cnta)\relax|
% \the\numexpr \cnta*\cnta*(\cnta/\cntb)+\cntb*\cntb*(\cntb/\cnta)\relax  {}
% |\thebnumexpr \cnta*\cnta*(\cnta/\cntb)+\cntb*\cntb*(\cntb/\cnta)\relax|
% \thebnumexpr \cnta*\cnta*(\cnta/\cntb)+\cntb*\cntb*(\cntb/\cnta)\relax {}
|\the\numexpr 123/188*188\relax|, |\the\numexpr 123/(188*188)\relax|, %
|\thebnumexpr 123/188*188\relax|, |\thebnumexpr 123/(188*188)\relax|.
\the\numexpr 123/188*188\relax, \the\numexpr 123/(188*188)\relax, %
\thebnumexpr 123/188*188\relax, \thebnumexpr 123/(188*188)\relax.

\edef\tmp {\bnumexpr 121873197*123-218137917*188\relax}
|\edef\tmp {\bnumexpr 121873197*123-218137917*188\relax}\bnethe\tmp|
\bnethe\tmp {}
|\meaning\tmp|
\meaning\tmp {}
|\thebnumexpr \tmp*(173197129797-\tmp)*(2179171982-\tmp)\relax|
\thebnumexpr \tmp*(173197129797-\tmp)*(2179171982-\tmp)\relax

|\cnta \thebnumexpr 2152966419779999/987654321\relax\space|
|\the\cnta| %
\cnta \thebnumexpr 2152966419779999/987654321\relax\space \the\cnta

|\thebnumexpr 2179878*987654321-2152966419779999, 2179879*987654321-2152966419779999\relax|
\thebnumexpr  2179878*987654321-2152966419779999,2179879*987654321-2152966419779999\relax\ (there was indeed rounding of the exact quotient.)

Some example with the power operator |^|:
\ \ \ \ (requires option |allowpower|, not compatible with |l3bigint|)
|\thebnumexpr (1^10+2^10+3^10+4^10+5^10+6^10)^3\relax| %
\thebnumexpr (1^10+2^10+3^10+4^10+5^10+6^10)^3\relax

|\thebnumexpr 2^31-1,2^31,2^31+1,2^100\relax|
\thebnumexpr 2^31-1,2^31,2^31+1,2^100\relax
\end{flushleft}

\section{Changes}
%\small
\begin{description}
\input bnumexpr.changes
\end{description}

\StopEventually{\end{document}\endinput}

\newgeometry{hscale=0.75,vscale=0.75}% ATTENTION \newgeometry fait
                                % un reset de vscale si on ne le
                                % précise pas ici !!!

\MakePercentIgnore
%
% \catcode`\<=0 \catcode`\>=11 \catcode`\*=11 \catcode`\/=11
% \let</dtx>\relax
% \def<*package>{\catcode`\<=12 \catcode`\>=12 \catcode`\*=12 \catcode`\/=12 }
%</dtx>
%<*package>
%
% \section{Package \bnenameimp implementation}
% \label{sec:bnumexprcode}
%
% \localtableofcontents \bigskip
%
% Comments are sparse. Error handling by the parser is kept to a minimum; if
% something goes wrong, the offensive token gets discarded, and some undefined
% control sequence attempts to trigger writing to the log of some sort of
% informative message. It is recommended to set |\errorcontextlines| to at
% least |2| for more meaningful context.
% \subsection{Package identification and catcode setup}
%    \begin{macrocode}
\NeedsTeXFormat{LaTeX2e}%
\ProvidesPackage{bnumexpr}[2014/10/28 v1.1b Expressions with big integers (jfB)]%
\edef\BNErestorecatcodes {\catcode`\noexpand\!\the\catcode`\!
                   \catcode`\noexpand\?\the\catcode`\?
                   \catcode`\noexpand\_\the\catcode`\_
                   \catcode`\noexpand\:\the\catcode`\:
                   \catcode`\noexpand\(\the\catcode`\(
                   \catcode`\noexpand\)\the\catcode`\)
                   \catcode`\noexpand\*\the\catcode`\*
                   \catcode`\noexpand\,\the\catcode`\,\relax }%
\catcode`\! 11
\catcode`\? 11
\catcode`\_ 11
\catcode`\: 11
%    \end{macrocode}
% \subsection{Package options}
%    \begin{macrocode}
\def\BNE_tmpa {0}%
\DeclareOption {custom}{\def\BNE_tmpa {1}%
    \PackageWarningNoLine{bnumexpr}{^^J
  Option custom: package xintcore not loaded. Definitions are needed for:^^J
  \protect\bnumexprAdd, \protect\bnumexprSub,
  \protect\bnumexprMul\space and \protect\bnumexprDiv }%
}%
\DeclareOption {bigintcalc}{\def\BNE_tmpa {2}%
    \PackageWarningNoLine{bnumexpr}{^^J
    Option bigintcalc: the macros from package bigintcalc are used.^^J
    Notice that / is mapped to \protect\bigintcalcDiv\space which does truncated division}%
}%
\DeclareOption {l3bigint}{\def\BNE_tmpa {3}%
    \PackageWarningNoLine{bnumexpr}{^^J
    Option l3bigint: the macros from package l3bigint are used.^^J
    There is no power operation, currently}%
}%
\DeclareOption {nocsv}{%
    \PackageInfo{bnumexpr}{Comma separated expressions disabled}%
    \AtEndOfPackage{\expandafter\let\csname BNE_precedence_,\endcsname
                                    \undefined }%
}%
\DeclareOption {notacitmul}{%
    \PackageInfo{bnumexpr}{Tacit multiplication disabled}%
    \AtEndOfPackage{\BNE_notacitmultiplication}%
}%
\def\BNE_allowpower {0}%
\DeclareOption {allowpower}{%
    \PackageInfo{bnumexpr}{Power operator ^ authorized}%
    \def\BNE_allowpower {1}%
}%
\ProcessOptions\relax
%    \end{macrocode}
% \subsection{Mapping to an underlying big integer engine.}
% In case option |bigintcalc| is used, notice that |/| is mapped to the macro
% |\bigintcalcDiv| which does truncated division. We did not add the extra code
% for rounded division in that case.
%
% With option |l3bigint|, there is no power operation available currently.
% Furthermore the package is part of the experimental trunk of the \LaTeX3
% project hence the names of its macros could change.
%    \begin{macrocode}
\def\bnumexprUsesxint {%
    \let\bnumexprAdd\xintiiAdd
    \let\bnumexprSub\xintiiSub
    \let\bnumexprMul\xintiiMul
    \let\bnumexprDiv\xintiiDivRound
    \let\bnumexprPow\xintiiPow
}%
\def\bnumexprUsesbigintcalc {%
    \let\bnumexprAdd\bigintcalcAdd
    \let\bnumexprSub\bigintcalcSub
    \let\bnumexprMul\bigintcalcMul
    \let\bnumexprDiv\bigintcalcDiv % NOTE: THIS DOES TRUNCATED DIVISION
    \let\bnumexprPow\bigintcalcPow
}%
\def\bnumexprUsesliiibigint {%
    \let\bnumexprAdd\bigint_add:nn
    \let\bnumexprSub\bigint_sub:nn
    \let\bnumexprMul\bigint_mul:nn
    \let\bnumexprDiv\bigint_div_round:nn
    \let\bnumexprPow\bigint_pow:nn % does not exist!
}%
\if0\BNE_tmpa % Toggle to load xintcore.sty
    \RequirePackage{xintcore}%
    \bnumexprUsesxint
\fi
\if2\BNE_tmpa % Toggle to load bigintcalc.sty
    \RequirePackage{bigintcalc}%
    \bnumexprUsesbigintcalc
\fi
\if3\BNE_tmpa % Toggle to load l3bigint.sty
    \RequirePackage{l3bigint}%
    \bnumexprUsesliiibigint
\fi
%    \end{macrocode}
% \subsection{Some helper macros and constants from xint}
% These macros from xint should not change, hence overwriting them here should
% not be cause for alarm. I opted against renaming everything with |\BNE_|
% prefix rather than |\xint_|. The |\xint_dothis|/|\xint_orthat| thing is a new
% style I have adopted for expandably forking. The least probable branches
% should be specified first, for better efficiency. See examples of uses in the
% present code.
%    \begin{macrocode}
\chardef\xint_c_     0
\chardef\xint_c_i    1
\chardef\xint_c_ii   2
% \chardef\xint_c_iii  3
% \chardef\xint_c_iv   4
% \chardef\xint_c_v    5
\chardef\xint_c_vi   6
\chardef\xint_c_vii  7
\chardef\xint_c_viii 8
\chardef\xint_c_ix   9
% \chardef\xint_c_x     10
% \chardef\xint_c_xviii 18
\long\def\xint_gobble_i      #1{}%
\long\def\xint_gobble_iii    #1#2#3{}%
\long\def\xint_firstofone    #1{#1}%
\long\def\xint_firstoftwo    #1#2{#1}%
\long\def\xint_secondoftwo   #1#2{#2}%
\long\def\xint_firstofthree  #1#2#3{#1}%
\long\def\xint_secondofthree #1#2#3{#2}%
\long\def\xint_thirdofthree  #1#2#3{#3}%
\def\xint_gob_til_!   #1!{}% this ! has catcode 11
\def\xint_UDsignfork  #1-#2#3\krof {#2}%
\long\def\xint_afterfi       #1#2\fi {\fi #1}%
\long\def\xint_dothis        #1#2\xint_orthat #3{\fi #1}% new in v1.1
\let\xint_orthat             \xint_firstofone
%    \end{macrocode}
% \subsection{Encapsulation of numbers in pseudo cs names}
% We define here a |\BNE_num| to not have to invoke |\xintNum|; hence
% dependency on |xint.sty| is kept to the actual arithmetic operations. We
% only need to get rid of leading zeros as plus and minus signs have already
% been stripped off; generally speaking user input will have no leading zeros
% thus the macro is designed to go fast when it is not needed... and as
% everything happens inside a |\csname...\endcsname|, we can leave some
% trailing |\fi|'s.
%    \begin{macrocode}
\edef\BNE_lock #1!{\noexpand\expandafter\space\noexpand
                            \csname .=\noexpand\BNE_num #1\endcsname }%
\def\BNE_num #1{\if #10\expandafter\BNE_num\else
                \ifcat #1\relax 0\expandafter\expandafter\expandafter #1\else
                #1\fi\fi }%
\def\BNE_unlock   {\expandafter\BNE_unlock_a\string }%
\def\BNE_unlock_a #1.={}%
%    \end{macrocode}
% \subsection{\csh{bnumexpr}, \csh{bnethe}, \csh{thebnumexpr}, \dots}
% In the full |\xintexpr|, the final unlocking may involve post-treatment of
% the comma separated values, hence there are |_print| macros to handle the
% possibly comma separated values. Here we may just identify |_print| with
% |_unlock|.
%    \begin{macrocode}
\def\bnumexpr {\romannumeral0\bnumeval }%
\def\bnumeval {\expandafter\BNE_wrap\romannumeral0\BNE_eval }%
\def\BNE_eval {\expandafter\BNE_until_end_a\romannumeral-`0\BNE_getnext }%
\def\BNE_wrap { !\BNE_usethe\BNE_protect\BNE_unlock }%
\protected\def\BNE_usethe\BNE_protect {\BNE:missing_bnethe!}%
\def\BNE_protect\BNE_unlock {\noexpand\BNE_protect\noexpand\BNE_unlock\noexpand }%
\let\BNE_done\space
\def\thebnumexpr
              {\romannumeral-`0\expandafter\BNE_unlock\romannumeral0\BNE_eval }%
\def\bnethe #1{\romannumeral-`0\expandafter\xint_gobble_iii\romannumeral-`0#1}%
%    \end{macrocode}
% \subsection{\csh{BNE_getnext}}
% The getnext scans forward to find a number: after expansion of what comes
% next, an opening parenthesis signals a parenthesized sub-expression, a |!|
% with catcode 11 signals there was there a sub |\bnumexpr...\relax| (now
% evaluated), a minus sign is treated as a prefix operator inheriting its
% precedence level from the previous operator, a plus sign is swallowed, a
% |\count| or |\dimen| will get fetched to |\number| (in case of a count
% variable, this provides a full locked number but |\count0 1| for example is
% like |1231| if |\count0|'s value is |123|); a digit triggers the
% number scanner. After the digit scanner finishes the integer is trimmed of
% leading zeros and locked as a single token into a |\csname .=...\endcsname|.
% The flow then proceeds with |\BNE_getop| which looks for the next operator
% or possibly the end of the expression. Note: |\bnumexpr\relax| is illegal.
%    \begin{macrocode}
\def\BNE_getnext #1%
{%
    \expandafter\BNE_getnext_a\romannumeral-`0#1%
}%
\def\BNE_getnext_a #1%
{%
    \xint_gob_til_! #1\BNE_gn_foundexpr !% this ! has catcode 11
    \ifcat\relax#1% \count or \numexpr etc... token or count, dimen, skip cs
       \expandafter\BNE_gn_countetc
    \else
       \expandafter\expandafter\expandafter\BNE_gn_fork\expandafter\string
    \fi
    #1%
}%
\def\BNE_gn_foundexpr !#1\fi !{\expandafter\BNE_getop\xint_gobble_iii }%
\def\BNE_gn_countetc #1%
{%
    \ifx\count#1\else\ifx#1\dimen\else\ifx#1\numexpr\else\ifx#1\dimexpr\else
    \ifx\skip#1\else\ifx\glueexpr#1\else\ifx\fontdimen#1\else
      \BNE_gn_unpackvar
    \fi\fi\fi\fi\fi\fi\fi
    \expandafter\BNE_getnext\number #1%
}%
\def\BNE_gn_unpackvar\fi\fi\fi\fi\fi\fi\fi\expandafter
                            \BNE_getnext\number #1%
{%
  \fi\fi\fi\fi\fi\fi\fi
  \expandafter\BNE_getop\csname .=\number#1\endcsname
}%
%    \end{macrocode}
% This is quite simplified here compared to |\xintexpr|, for various reasons: we
% have dropped the |\xintNewExpr| thing, and we can treat the |(| directly as we
% don't have to get back to check if we are in an |\xintexpr|,
% |\xintfloatexpr|, etc..
%    \begin{macrocode}
\def\BNE_gn_fork #1{%
    \if#1+\xint_dothis \BNE_getnext\fi
    \if#1-\xint_dothis -\fi
    \if#1(\xint_dothis \BNE_oparen \fi
    \xint_orthat       {\BNE_scan_number #1}%
}%
%    \end{macrocode}
% \subsection{Parsing an integer}
% We gather a string of digits, plus and minus prefixes have already been
% swallowed. There might be some leading string of zeros which will have to be
% removed. In the full |\xintexpr 1.1| the situation is more involved as it
% has to recognize and accept decimal numbers, numbers in scientific notation,
% also hexadecimal numbers, function names, variable names\dots
%    \begin{macrocode}
\def\BNE_scan_number #1% this #1 has necessarily here catcode 12
{%
    \ifnum \xint_c_ix<1#1 \expandafter \BNE_scan_nbr\else
                          \expandafter \BNE_notadigit\fi #1%
}%
\def\BNE_notadigit #1{\BNE:not_a_digit! \xint_gobble_i {#1}}%
%    \end{macrocode}
% Scanning for a number. Once gathered, lock it and do |_getop|. If we hit
% against some catcode eleven |!|, this means there was a sub |\bnumexpr..\relax|.
% We then apply tacit multiplication.
%    \begin{macrocode}
\def\BNE_scan_nbr
{%
    \expandafter\BNE_getop\romannumeral-`0\expandafter
    \BNE_lock\romannumeral-`0\BNE_scan_nbr_c
}%
\def\BNE_scan_nbr_a #1%
{% careful that ! has catcode letter here
    \ifcat \relax #1\xint_dothis{!#1}\fi % stops the scan
    \ifx         !#1\xint_dothis{!*!}\fi % tacit multiplication before subexpr
    \xint_orthat {\expandafter\BNE_scan_nbr_b\string #1}%
}%
\def\BNE_scan_nbr_b #1% #1 with catcode 12
{%
    \ifnum \xint_c_ix<1#1 \expandafter\BNE_scan_nbr_c
    \else\expandafter !\fi #1%
}%
\def\BNE_scan_nbr_c #1#2%
{%
    \expandafter #1\romannumeral-`0\expandafter
                   \BNE_scan_nbr_a\romannumeral-`0#2%
}%
%    \end{macrocode}
%    \begin{macrocode}
%    \end{macrocode}
% \subsection{\csh{BNE_getop}}
% This finds the next infix operator or closing parenthesis or expression end.
% It then leaves in the token flow <precedence> <operator> <locked number>.
% The <precedence> stops expansion and ultimately gives back control to a
% |\BNE_until_<op>| command. The code here is derived from more involved
% context where the actual macro associated to the operator may vary,
% depending if we are in |\xintexpr|, |\xintfloatexpr| or |\xintiiexpr|. Here
% things are simpler but I have kept the general scheme, thus the actual macro
% to be used for the <operator> is not decided immediately (|xintexpr.sty| has
% extra things to allow multi-characters operators like |&&|).
%    \begin{macrocode}
\def\BNE_getop #1#2% this #1 is the current locked computed value
{%
    \expandafter\BNE_getop_a\expandafter #1\romannumeral-`0#2%
}%
\catcode`* 11
\def\BNE_getop_a #1#2%
{%  if a control sequence is found, must be \relax, or possibly register or
 %  variable if tacit multiplication is allowed
    \ifx \relax #2\xint_dothis\xint_firstofthree\fi
    % tacit multiplications:
    \ifcat \relax #2\xint_dothis\xint_secondofthree\fi
    \if    (#2\xint_dothis      \xint_secondofthree\fi
    \ifx   !#2\xint_dothis      \xint_secondofthree\fi
    \xint_orthat \xint_thirdofthree
    {\BNE_foundend #1}%
    {\BNE_precedence_* *#1#2}% tacit multiplication
    {\BNE_foundop #2#1}%
}%
\catcode`* 12
\def\BNE_foundend {\xint_c_ \relax }% \relax is only a place-holder here.
\def\BNE_foundop #1%
{%
    \ifcsname BNE_precedence_#1\endcsname
        \csname BNE_precedence_#1\expandafter\endcsname
        \expandafter #1%
    \else
        \BNE_notanoperator {#1}\expandafter\BNE_getop
    \fi
}%
\def\BNE_notanoperator #1{\BNE:not_an_operator! \xint_gobble_i {#1}}%
%    \end{macrocode}
% \subsection{Until macros for global expression and parenthesized sub-ones}
% The minus sign as prefix is treated here.
%    \begin{macrocode}
\catcode`) 11
\def\BNE_tmpa #1{%  #1=\BNE_op_-vi token
    \def\BNE_until_end_a ##1%
    {%
        \xint_UDsignfork
            ##1{\expandafter\BNE_until_end_a\romannumeral-`0#1}%
              -{\BNE_until_end_b ##1}%
        \krof
    }%
}\expandafter\BNE_tmpa\csname BNE_op_-vi\endcsname
\def\BNE_until_end_b #1#2%
    {%
        \ifcase #1\expandafter\BNE_done
        \or
        \xint_afterfi{\BNE:extra_)_?\expandafter
                      \BNE_until_end_a\romannumeral-`0\BNE_getop }%
        \else
        \xint_afterfi{\expandafter\BNE_until_end_a
                      \romannumeral-`0\csname BNE_op_#2\endcsname }%
        \fi
    }%
\catcode`( 11
\def\BNE_op_( {\expandafter\BNE_until_)_a\romannumeral-`0\BNE_getnext }%
\let\BNE_oparen\BNE_op_(
\catcode`( 12
\def\BNE_tmpa #1{% #1=\BNE_op_-vi
    \def\BNE_until_)_a ##1{\xint_UDsignfork
                           ##1{\expandafter \BNE_until_)_a\romannumeral-`0#1}%
                             -{\BNE_until_)_b ##1}%
                      \krof }%
}\expandafter\BNE_tmpa\csname BNE_op_-vi\endcsname
\def \BNE_until_)_b #1#2%
    {%
     \ifcase  #1\expandafter    \BNE_missing_)_? % missing ) ?
                \or\expandafter \BNE_getop       % found closing )
                \else \xint_afterfi
      {\expandafter \BNE_until_)_a\romannumeral-`0\csname BNE_op_#2\endcsname }%
        \fi
    }%
\def\BNE_missing_)_? {\BNE:missing_)_inserted \xint_c_ \BNE_done }%
\let\BNE_precedence_) \xint_c_i
\let\BNE_op_)   \BNE_getop
\catcode`) 12
%    \end{macrocode}
% \subsection{The arithmetic operators.}
% This is where the infix operators are mapped to actual macros. These macros
% must ``f-expand'' their arguments, and know how to handle then big integers
% having no leading zeros and at most a minus sign.
%    \begin{macrocode}
\def\BNE_tmpc #1#2#3#4#5#6#7%
{%
  \def #1##1% \BNE_op_<op>
  {% keep value, get next number and operator, then do until
    \expandafter #2\expandafter ##1\romannumeral-`0\expandafter\BNE_getnext }%
  \def #2##1##2% \BNE_until_<op>_a
  {\xint_UDsignfork
    ##2{\expandafter #2\expandafter ##1\romannumeral-`0#4}%
      -{#3##1##2}%
   \krof }%
  \def #3##1##2##3##4% \BNE_until_<op>_b
  {% either execute next operation now, or first do next (possibly unary)
    \ifnum ##2>#5%
    \xint_afterfi {\expandafter #2\expandafter ##1\romannumeral-`0%
      \csname BNE_op_##3\endcsname {##4}}%
    \else \xint_afterfi {\expandafter ##2\expandafter ##3%
      \csname .=#6{\BNE_unlock ##1}{\BNE_unlock ##4}\endcsname }%
    \fi }%
  \let #7#5%
}%
\def\BNE_tmpb #1#2#3%
{%
  \expandafter\BNE_tmpc
  \csname BNE_op_#1\expandafter\endcsname
  \csname BNE_until_#1_a\expandafter\endcsname
  \csname BNE_until_#1_b\expandafter\endcsname
  \csname BNE_op_-#2\expandafter\endcsname
  \csname xint_c_#2\expandafter\endcsname
  \csname #3\expandafter\endcsname
  \csname BNE_precedence_#1\endcsname
}%
\BNE_tmpb  +{vi}{bnumexprAdd}%
\BNE_tmpb  -{vi}{bnumexprSub}%
\BNE_tmpb  *{vii}{bnumexprMul}%
\BNE_tmpb  /{vii}{bnumexprDiv}%
\if1\BNE_allowpower\BNE_tmpb ^{viii}{bnumexprPow}\fi
%    \end{macrocode}
% \subsection{The minus as prefix operator of variable precedence level}
% We only need here two levels of precedence, |vi| and |vii|. If the power |^|
% operation is authorized, then one further level |viii| is needed.
%    \begin{macrocode}
\def\BNE_tmpa #1% #1=vi or vii
{%
\expandafter\BNE_tmpb
    \csname BNE_op_-#1\expandafter\endcsname
    \csname BNE_until_-#1_a\expandafter\endcsname
    \csname BNE_until_-#1_b\expandafter\endcsname
    \csname xint_c_#1\endcsname
}%
\def\BNE_tmpb #1#2#3#4%
{%
    \def #1% \BNE_op_-<level>
    {%  get next number+operator then switch to _until macro
        \expandafter #2\romannumeral-`0\BNE_getnext
    }%
    \def #2##1% \BNE_until_-<level>_a
    {\xint_UDsignfork
        ##1{\expandafter #2\romannumeral-`0#1}%
          -{#3##1}%
     \krof }%
    \def #3##1##2##3% \BNE_until_-<level>_b
    {%
        \ifnum ##1>#4%
         \xint_afterfi {\expandafter #2\romannumeral-`0%
                        \csname BNE_op_##2\endcsname {##3}}%
        \else
         \xint_afterfi {\expandafter ##1\expandafter ##2%
                        \csname .=\expandafter\BNE_Opp
                                  \romannumeral-`0\BNE_unlock ##3\endcsname }%
        \fi
    }%
}%
\BNE_tmpa {vi}%
\BNE_tmpa {vii}%
\if1\BNE_allowpower\BNE_tmpa {viii}\fi
\def\BNE_Opp #1{\if-#1\else\if0#10\else-#1\fi\fi }%
%    \end{macrocode}
% \subsection{The comma may separate expressions.}
% It suffices to treat the comma as a binary operator of precedence |ii|. We
% insert a space after the comma. The current code in |\xintexpr| does not do
% it at this stage, but only later during the final unlocking, as there is
% anyhow need for some processing for final formatting and was considered to
% be as well the opportunity to insert the space. Here, let's do it
% immediately. These spaces are not an issue when |\bnumexpr| is
% identified as a sub-expression in |\xintexpr|, for example in:
% |\xinttheiiexpr lcm(\bnumexpr 175-12,123+34,56*31\relax)\relax| (this
% example requires package |xintgcd|).
%    \begin{macrocode}
\catcode`, 11
\def\BNE_op_, #1%
{%
    \expandafter \BNE_until_,_a\expandafter #1\romannumeral-`0\BNE_getnext
}%
\def\BNE_tmpa #1{% #1 = \BNE_op_-vi
  \def\BNE_until_,_a ##1##2%
  {%
    \xint_UDsignfork
        ##2{\expandafter \BNE_until_,_a\expandafter ##1\romannumeral-`0#1}%
          -{\BNE_until_,_b ##1##2}%
     \krof }%
}\expandafter\BNE_tmpa\csname BNE_op_-vi\endcsname
\def\BNE_until_,_b #1#2#3#4%
{%
    \ifnum #2>\xint_c_ii
        \xint_afterfi {\expandafter \BNE_until_,_a
                   \expandafter #1\romannumeral-`0%
                   \csname BNE_op_#3\endcsname {#4}}%
    \else
        \xint_afterfi {\expandafter #2\expandafter #3%
                       \csname .=\BNE_unlock #1, \BNE_unlock #4\endcsname }%
    \fi
}%
\let \BNE_precedence_, \xint_c_ii
\catcode`, 12
%    \end{macrocode}
% \subsection{Disabling tacit multiplication}
%    \begin{macrocode}
\def\BNE_notacitmultiplication{%
  \def\BNE_getop_a ##1##2{%
    \ifx \relax ##2\expandafter\xint_firstoftwo\else
                   \expandafter\xint_secondoftwo\fi
    {\BNE_foundend ##1}%
    {\BNE_foundop  ##2##1}%
  }%
  \def\BNE_scan_nbr_a ##1{%
    \ifcat \relax ##1\expandafter\xint_firstoftwo\else
                     \expandafter\xint_secondoftwo\fi
    {!##1}{\expandafter\BNE_scan_nbr_b\string ##1}%
  }%
}%
%    \end{macrocode}
% \subsection{Cleanup}
%    \begin{macrocode}
\let\BNE_tmpa\relax \let\BNE_tmpb\relax \let\BNE_tmpc\relax
\BNErestorecatcodes
%    \end{macrocode}
% \MakePercentComment
%</package>
%<*dtx>

\DeleteShortVerb{\|}
\CharacterTable
 {Upper-case    \A\B\C\D\E\F\G\H\I\J\K\L\M\N\O\P\Q\R\S\T\U\V\W\X\Y\Z
  Lower-case    \a\b\c\d\e\f\g\h\i\j\k\l\m\n\o\p\q\r\s\t\u\v\w\x\y\z
  Digits        \0\1\2\3\4\5\6\7\8\9
  Exclamation   \!     Double quote  \"     Hash (number) \#
  Dollar        \$     Percent       \%     Ampersand     \&
  Acute accent  \'     Left paren    \(     Right paren   \)
  Asterisk      \*     Plus          \+     Comma         \,
  Minus         \-     Point         \.     Solidus       \/
  Colon         \:     Semicolon     \;     Less than     \<
  Equals        \=     Greater than  \>     Question mark \?
  Commercial at \@     Left bracket  \[     Backslash     \\
  Right bracket \]     Circumflex    \^     Underscore    \_
  Grave accent  \`     Left brace    \{     Vertical bar  \|
  Right brace   \}     Tilde         \~}
\CheckSum {833}
\makeatletter\check@checksum\makeatother
\Finale
%% End of file xint.dtx