% \iffalse meta-comment
%
% Copyright (C) 2004 by Robert J Lee <latex@rjlee.homelinux.org>
% --------------------------------------------------------------
%
% This file may be distributed and/or modified under the
% conditions of the LaTeX Project Public License, either version 1.2
% of this license or (at your option) any later version.
% The latest version of this license is in:
%
% http://www.latex-project.org/lppl.txt
%
% and version 1.2 or later is part of all distributions of LaTeX
% version 1999/12/01 or later.
%
% \fi
%
% \iffalse
%<*driver>
\ProvidesFile{rjlpshap.dtx}
%</driver>
%<package> \NeedsTeXFormat{LaTeX2e}[1999/12/01]
%<package> \ProvidesPackage{rjlpshap}
%<*package>
[2004/11/05 v1.0 .dtx rjlpshap file]
%</package>
%
%<*driver>
\documentclass{ltxdoc}
\EnableCrossrefs
\CodelineIndex
\RecordChanges
\usepackage{lipsum}
\begin{document}
\DocInput{rjlpshap.dtx}
\end{document}
%</driver>
% \fi
%
% \CheckSum{0}
%
% \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 \~}
%
%
% \changes{v1.0}{2004/11/05}{Initial version}
%
% \GetFileInfo{rjlpshap.dtx}
%
% \DoNotIndex{\newcommand,\newenvironment}
%
%
% \title{The \textsf{rjlpshap} class\thanks{This document
% corresponds to \textsf{rjlpshap}~\fileversion, dated \filedate.}}
% \author{Robert J Lee \\ \texttt{latex@rjlee.homelinux.org}}
%
% \maketitle
%
% \section{Introduction}
%
% This package provides low-level helper macros and environments. It
% is intended for authors of \LaTeX\ packages, who wish to
% programmatically change the shape of paragraphs. It overcomes several
% difficulties with \TeX's \texttt{$\backslash$parshape} register:
%
% \begin{enumerate}
% \item It is not possible to re-use a length register when issuing a
% \texttt{$\backslash$parshape} command; this makes it difficult to
% calculate the shape of a paragraph in \LaTeX\ using some
% algorithms that re-use one or more variables (registers).
% \item The number of lines must be known \textit{before} the lengths
% of each line, which can lead to difficult calculations.
% \item The format of the \texttt{$\backslash$parshape} command is
% very strict and evaluated early on by \TeX, and so it doesn't
% easily allow for parameter-replacement, making it difficult to use
% any calculations at all.
% \end{enumerate}
%
% It differs from the \texttt{shapepar} package, in that it provides
% no actual calculations for paragraph shapes. Instead, it provides a
% framework by which paragraph shapes can be calculated in a single
% run of the document by the author (or package writer).
%
% The package provides both environments and commands by which the
% author can accumulate any number of lengths and then use them in a
% \texttt{$\backslash$parshape} command injected into the document,
% without the need to re-run \LaTeX.
%
% \pagebreak[4]
% \section{Usage}
%
% There are two modes of operation: environment mode and command
% mode. Environment mode has the advantage that it does not
% require the number of lengths to be known in advance, while command
% mode provides various options for passing pre-calculated lengths.
%
% \subsection{Environment Mode}
%
% Environment mode uses a single environment,
% \texttt{parshapecollect}, which starts and ends before the paragraph
% of text. The contents of the environment are a \LaTeX\ program,
% supplied by the user, that calculates the lengths of each line.
%
% \DescribeEnv{parshapecollect} The \texttt{parshapecollect}
% environment goes at the start of a paragraph, and the entire
% environment defines a single \texttt{$\backslash$parshape} command.
% Any even number of lengths are collected, and at the end of the
% environment, the \texttt{$\backslash$parshape} command is
% automatically generated and inserted into the document.
%
% \DescribeMacro{\parshapelenout} This macro takes one mandatory
% parameter, which is the next length in the
% \texttt{$\backslash$parshape} command. Note that the parameter is
% expanded: the value of the length at the time
% \texttt{$\backslash$parshapelenout} is called is the value which is
% used --- this means the same length command can be re-used
% repeatedly. The value may be a length register, and the same
% register may be re-used after changing its contents, for example:
%
% \begin{quote}
% \begin{verbatim}
% \newlength{\foo}
% \begin{parshapecollect}
% \setlength{\foo}{1in}\parshapelenout{\foo}
% \setlength{\foo}{3in}\parshapelenout{\foo}
% \setlength{\foo}{0.5in}\parshapelenout{\foo}
% \setlength{\foo}{3.5in}\parshapelenout{\foo}
% \setlength{\foo}{0in}\parshapelenout{\foo}
% \setlength{\foo}{4in}\parshapelenout{\foo}
% \end{parshapecollect}
% \lipsum[1]
% \end{verbatim}
% Produces:
%
% \parshape 3 1in 4in 0.5in 4.5in 0in 5in
% \lipsum[1]
% \end{quote}
%
% \DescribeMacro{\parshapearrlenout} This macro is the same as
% \texttt{$\backslash$parshapelenout}, but allows the programmer to
% pass the length as the value from an array (see the
% \texttt{arrayjob} package). This is intended for situations where
% pre-calculated lengths are to be used alongside calculations; to
% pass an array of lengths the various command modes may be
% easier. The parameters to the macro are the name of the array and
% the name of a counter containing the index.
%
% For example, the above paragraph could have been created
% using:
% \begin{quote}
% \begin{verbatim}
% \newarray{foo} \newcounter{ctr}
% \begin{parshapecollect}
% \foo(1)={1in}\setcounter{ctr}{1}\parshapearrlenout{foo}{ctr}
% \foo(2)={3in}\setcounter{ctr}{2}\parshapearrlenout{foo}{ctr}
% \foo(3)={0.5in}\setcounter{ctr}{3}\parshapearrlenout{foo}{ctr}
% \foo(1)={3.5in}\setcounter{ctr}{1}\parshapearrlenout{foo}{ctr}
% \foo(2)={0in}\setcounter{ctr}{2}\parshapearrlenout{foo}{ctr}
% \foo(2)={4in}\parshapearrlenout{foo}{ctr}
% \end{parshapecollect}
% \lipsum[1]
% \end{verbatim}
% \end{quote}
%
% \subsection{Command Mode}
% Command mode enables the author to use calculated shapes by
% outputting the $\backslash$parshape command as a single operation at
% the end of the calculation.
% \DescribeMacro{\parshapeary}
% This macro takes one parameter, being the name of an array. The
% contents of the array are simply expanded and passed directly to
% \texttt{$\backslash$parshape} --- the first parameter is the number
% of lines, then for each line the left margin and line width are
% supplied in subsequent elements.
%
% For example:
% \begin{quote}
% \begin{verbatim}
% \newcommand{\ltest}{50pt}
% \newlength{\test}\setlength{\test}{300pt}
% \newarray{sizes}\readarray{sizes}{2&0em&\the\test&\ltest&\test}
% \parshapeary{sizes}
% \lipsum[1]
% \end{verbatim}
% produces:
%
% \parshape 2 0em 300pt 50pt 300pt \lipsum[1]
% \end{quote}
%
%
% \DescribeMacro{\parshapearray} This macro takes two parameters. The
% first is also the name of an array. The only difference to
% \texttt{$\backslash$parshapeary} is that instead of passing the
% width of the line, the width of the right margin is passed
% instead. This means the macro needs to know the width of the line.
%
% The second parameter is the length of the line. This is usually
% $\backslash$columnwidth, passing any other value will effectively
% change the right margin. ($\backslash$columnwidth is often equal to
% $\backslash$textwidth, but in multi-column mode,
% $\backslash$columnwidth is more likely to be correct, unless you
% want to overstrike the text with text in another column --- so
% $\backslash$columnwidth is the recommended length to use).
%
% In a similar vein to the previous example:
% \begin{quote}
% \begin{verbatim}
%\newcommand{\ltest}{50pt}
%\newlength{\test}\setlength{\test}{75pt}
%\newlength{\cwidth}\setlength{\cwidth}{\columnwidth}
%\addtolength{\cwidth}{-50pt}
%
%\newarray{sizes}\readarray{sizes}{2&0em&\the\test&\ltest&\test}
%\parshapearray{sizes}{\cwidth}
%\lipsum[1]
%
% \end{verbatim}
% which produces:
%
% \parshape 2 0.0pt
%220.0pt
%50.0pt
%170.0pt
% \lipsum[1]
% \end{quote}
%
% \DescribeMacro{\Parshapearray}
% This macro also takes two parameters, and is used to combine
% multiple shapes, where the width of the paragraph is not consistent
% throughout the paragraph.
%
% The first parameter is the name of an array, in the same format as
% the first parameter to \texttt{$\backslash$parshapeary}.
%
% The second parameter is the name of a second array. The first
% element is the number of lines in this second array, and subsequent
% elements are the widths of lines (excluding margins) in that array
% --- starting from the first line. As usual, if an insufficient
% number of lines are passed, the last line will be re-used.
%
% The actual width of a line will be the total width excluding margins
% (as per the second parameter), minus the left margin (taken from the
% first parameter, even indexes), minus the right margin (as per the
% first parameter, odd indexes).
%
% In this way, it is possible to set a shaped paragraph that allows
% for non-straight margins or non-rectangular paper --- for example, to
% typeset a list item in such a way as to allow for a cut-out on the
% page.
%
% For a larger example, consider that the paper has a sloped shape
% that extends 1em wider for every line for the first 15 lines
% only. The author wishes to set a text in an quadrilateral, with
% margins offset $\frac{1}{8}\verb!"!$ to the right for each line, but
% extending equally to follow the width of the paper.
% \begin{quote}
% \begin{verbatim}
% \newarray{quad}
% \readarray{quad}{17 & 0.000in & 2.000in & 0.125in & 1.875in &
% 0.250in & 1.750in & 0.375in & 1.625in & 0.500in & 1.500in &
% 0.625in & 1.375in & 0.750in & 1.250in & 0.875in & 1.125in &
% 1.000in & 1.000in & 1.125in & 0.875in & 1.250in & 0.750in &
% 1.375in & 0.675in & 1.500in & 0.500in & 1.625in & 0.375in &
% 1.750in & 0.250in & 1.875in & 0.125in & 2.000in & 0.000in}
% \newarray{papershape}
% \readarray{papershape}{15& 25em & 26em & 27em & 28em & 29em & 30em & 31em
% & 32em & 33em & 34em & 35em & 36em & 37em & 38em & 39em}
% \Parshapearray{quad}{papershape} \lipsum[1]
% \end{verbatim}
% Produces:
%
% \parshape 17
% 0.0pt
% 105.46039pt
% 9.03374pt
% 115.46042pt
% 18.06749pt
% 125.46043pt
% 27.10124pt
% 135.46045pt
% 36.135pt
% 145.46045pt
% 45.16875pt
% 155.46046pt
% 54.2025pt
% 165.46048pt
% 63.23624pt
% 175.46051pt
% 72.26999pt
% 185.46053pt
% 81.30374pt
% 195.46054pt
% 90.3375pt
% 205.46054pt
% 99.37125pt
% 211.84685pt
% 108.405pt
% 225.46057pt
% 117.43874pt
% 235.4606pt
% 126.47249pt
% 245.46062pt
% 135.50624pt
% 235.4606pt
% 144.54pt
% 235.46059pt \lipsum[1]
% \end{quote}
%
% \section{Prerequisites}
% This package requires the \texttt{arrayjob} package from
% CTAN\footnote{http://tug.ctan.org/cgi-bin/ctanPackageInformation.py?id=arrayjob}. Arrays
% are used as a convenient mechanism to pass values to functions,
% although they may not be used in environment mode.
% \section{Limitations and Possible Improvements}
% \begin{itemize}
% \item Usage of this package can be slow, because two temporary files
% are used for each $\backslash$parshape. Theoretically, only one
% should be needed and so it should be possible to improve this.
% \item There is no support for combining arrays, other than to change
% the paper size itself. Extra commands could be provided to
% calculate the intersection of two cutouts, or to add margins
% together.
% \item In a future version, it may be possible to avoid temporary
% files completely by using token registers, or
% $\backslash$aftergroup, with some form of string processing.
% \end{itemize}
% \section{Package Name}
% The somewhat obscure name ``rjlpshap'' was chosen following the
% guidelines given in \textit{\LaTeXe\ for Class and Package
% Writers}\footnote{\texttt{http://www.tex.ac.uk/tex-archive/macros/latex/doc/clsguide.pdf}}~(section~2.7.3,
% ``Make it Portable''). The common prefix used by these packages
% is~''rjl--'', which was chosen as the author's initials as he is
% not writing these packages as part of any organisation.
% The postfix of ``--pshap'' was chosen as a contraction of
% ``parshape'', the \TeX\ command to which this package is a front
% end. The name was contracted due to the 5-character length
% restriction. It was also considered that that the obvious package
% name of ``parshape'' would be best avoided to avoid confusion with
% the existing shapepar\footnote{\texttt{http://www.ctan.org/tex-archive/macros/latex/contrib/shapepar/}} package, and also because parshape is a
% standard \TeX\ command --- the only requirement that exists is to
% avoid the name of a standard package, but avoiding the name of a
% standard macro also seems like a good idea as this leaves the name
% free for use as a standard package name in the future.
% \section{Other References}
% The following references were essential in producing this package:
% \begin{itemize}
% \item The \TeX Book\footnote{ISBN:~978-0201134483}
% \item \TeX Live package
% contributions\footnote{http://tug.org/texlive/pkgcontrib.html}
% \item TDS Guidelines\footnote{http://dante.ctan.org/TDS-guidelines.html}
% \end{itemize}
% \StopEventually{\PrintChanges\PrintIndex}
%
% \section{Implementation}
%
% Arrayjob: we need to pass arrays of lengths and length-commands,
% so build on the established array handling functionality.
% \begin{macrocode}
\RequirePackage{arrayjob}
% \end{macrocode}
% Forloop: we need to iterate over the arrays that are passed in
% \begin{macrocode}
\RequirePackage{forloop}
% \end{macrocode}
% Internal commands use @ as a letter, not a special character, so
% switch catcodes
% \begin{macrocode}
\makeatletter
% \end{macrocode}
% \subsection{Environment mode}
% \begin{environment}{parshapecollect}
% Environment to handle writing the file, cleaning up etc.
%
% Note the importance of the % at the end of each line here: we're
% in horizontal mode so any space characters get expanded and added
% into the start of the paragraph. Any spaces here are output before
% $\backslash$parshape itself, and so will be visible in the output
% page. The $\backslash$ignorespaces is a last line of defence
% against any extra spaces inadvertently introduced by the user.
%
% \begin{macrocode}
\newenvironment{parshapecollect}{%
\immediate\openout\rjlpshap@writer=\jobname.parshape%
\setcounter{rjlpshap@linecount}{0}%
\setboolean{rjlpshap@isodd}{false}%
\ignorespaces
}{%
\immediate\closeout\rjlpshap@writer%
% \end{macrocode}
% Because the environment forms a group, we need to use aftergroup to
% delay the $\backslash$parshape command. Otherwise the group will end first and
% $\backslash$parshape will be discarded.
% \begin{macrocode}
\rjlpshap@doparshape%
\aftergroup\input\aftergroup{\aftergroup\jobname\aftergroup%
.\aftergroup%
p\aftergroup%
a\aftergroup%
r\aftergroup%
s\aftergroup%
h\aftergroup%
a\aftergroup%
p\aftergroup%
e\aftergroup%
t\aftergroup%
m\aftergroup%
p\aftergroup}\ignorespaces
}
% \end{macrocode}
% \end{environment}
% \begin{macro}{\rjlpshap@doparshape}
% Internal method to handle the final processing of the built
% up $\backslash$parshape command. The lengths have been written to
% the first temporary file, so write the entire $\backslash$parshape
% command to the second one, ready to be input.
% \begin{macrocode}
\newcommand{\rjlpshap@doparshape}{%
\immediate\openout\rjlpshap@finalwriter=\jobname.parshapetmp%
\immediate\write\rjlpshap@finalwriter{\noexpand\parshape \arabic{rjlpshap@linecount}}%
\openin\rjlpshap@reader=\jobname.parshape%
\setboolean{rjlpshap@loop}{true}%
\whiledo{\boolean{rjlpshap@loop}}{%
\read\rjlpshap@reader to\rjlpshap@temp%
% NB: \ifeof matches \loop...\repeat
\ifeof\rjlpshap@reader%
\setboolean{rjlpshap@loop}{false}%
\else
\immediate\write\rjlpshap@finalwriter{\rjlpshap@temp }%
\fi
}%
\immediate\closeout\rjlpshap@finalwriter%
\immediate\closein\rjlpshap@reader%
\let\rjlpshap@temp=\relax%
}
% \end{macrocode}
% \end{macro}
% \begin{macro}{\parshapelenout}
% Used inside a parshapecollect environment, writes length \#1 to
% the $\backslash$parshape command file. (If used anywhere else,
% it'll just write to the console since the file will not be open).
% \begin{macrocode}
\newcommand{\parshapelenout}[1]{%
\setlength{\rjlpshap@tempwidth}{#1}%
\immediate\write\rjlpshap@writer{\the\rjlpshap@tempwidth}%
\ifthenelse{\boolean{rjlpshap@isodd}}{%
\setboolean{rjlpshap@isodd}{false}%
}{%
\stepcounter{rjlpshap@linecount}%
\setboolean{rjlpshap@isodd}{true}%
}%
\ignorespaces
}
% \end{macrocode}
% \end{macro}
% \begin{macro}{\parshapearrlenout}
% Used inside a parshapecollect environment, writes a length to the
% $\backslash$parshape command file. As $\backslash$parshapelenout,
% except that the element is passed as the member of an array.
%
% Length is taken from array with name \#1, element at index \#2
% \begin{macrocode}
\newcommand{\parshapearrlenout}[2]{%
\testarray{#1}(\arabic{#2})%
\parshapelenout{\temp@macro}%
\ignorespaces
}
% \end{macrocode}
% \end{macro}
% \subsection{Command Mode}
% \begin{macro}{\parshapeary}
% As $\backslash$parshape, but takes one argument being an array
% name defined in arrayjob form. The array contains the arguments,
% starting with the count at index 1, then one length per array
% element.
% \begin{macrocode}
\newcommand{\parshapeary}[1]{%
\testarray{#1}(1)%
\edef\rjlpshap@max{\temp@macro}%
\begin{parshapecollect}%
\forloop{rjlpshap@ctr}{1}{\not \value{rjlpshap@ctr} > \rjlpshap@max }{%
\setcounter{rjlpshap@ctr2}{\value{rjlpshap@ctr}}%
\addtocounter{rjlpshap@ctr2}{\value{rjlpshap@ctr}}%
\parshapearrlenout{#1}{rjlpshap@ctr2}%
\stepcounter{rjlpshap@ctr2}%
\parshapearrlenout{#1}{rjlpshap@ctr2}%
}%
\end{parshapecollect}%
\typeout{parshapeary end}%
\ignorespaces
}
% \end{macrocode}
% \end{macro}
% \begin{macro}{\parshapearray}
% As $\backslash$parshapeary, but the second parameter is the name
% of a single-parameter macro taking a number and returning the
% width of the line on that line
% \begin{macrocode}
\newcommand{\parshapearray}[2]{%
\begin{parshapecollect}%
\testarray{#1}(1)%
\edef\rjlpshap@max{\temp@macro}%
\forloop{rjlpshap@ctr}{1}{\not \value{rjlpshap@ctr} > \rjlpshap@max }{%
\setcounter{rjlpshap@ctr2}{\value{rjlpshap@ctr}}%
\addtocounter{rjlpshap@ctr2}{\value{rjlpshap@ctr}}%
\edef\rjlpshap@pos{\arabic{rjlpshap@ctr2}}%
\testarray{#1}(\rjlpshap@pos)%
\setlength{\rjlpshap@tempwidth}{\temp@macro}%
\setlength{\rjlpshap@templinewidth}{#2}%
\addtolength{\rjlpshap@templinewidth}{-1\rjlpshap@tempwidth}%
\parshapelenout{\rjlpshap@tempwidth}%
\stepcounter{rjlpshap@ctr2}%
\edef\rjlpshap@pos{\arabic{rjlpshap@ctr2}}%
\testarray{#1}(\rjlpshap@pos)%
\setlength{\rjlpshap@tempwidth}{\temp@macro}%
\addtolength{\rjlpshap@templinewidth}{-1\rjlpshap@tempwidth}%
\parshapelenout{\rjlpshap@templinewidth}%
}%
\end{parshapecollect}%
}
% \end{macrocode}
% \end{macro}
% \begin{macro}{\Parshapearray}
% First parameter is as $\backslash$parshapea\{r,rra\}y, but each pair
% is left and right margin rather than left margin and text width,
% while the second tracks the page width.
%
% \textit{i.e.}
%
% The first parameter (an array) is the shape of the paragraph
% relative to the page.
% \begin{itemize}
% \item first element is the number of pairs of lengths in the array
% \item subsequent arguments are pairs of lengths, left margin width then
% right margin width.
% \end{itemize}
%
% The second parameter is the overall text width. This is given in the
% following format:
% \begin{itemize}
% \item first element is the number lengths
% \item subsequent elements are one width per parameter
% \end{itemize}
% \begin{macrocode}
\newcommand{\Parshapearray}[2]{%
\testarray{#1}(1)\edef\rjlpshap@maxA{\temp@macro}%
\testarray{#2}(1)\edef\rjlpshap@maxB{\temp@macro}%
\ifthenelse{\rjlpshap@maxA > \rjlpshap@maxB} {\edef\rjlpshap@max{\rjlpshap@maxA}}{\edef\rjlpshap@max{\rjlpshap@maxB}}%
\begin{parshapecollect}%
\forloop{rjlpshap@ctr}{1}{\not \value{rjlpshap@ctr} > \rjlpshap@max}{%
\ifthenelse{\value{rjlpshap@ctr} > \rjlpshap@maxA}{%
\setcounter{rjlpshap@ctr2}{\rjlpshap@maxA}%
\addtocounter{rjlpshap@ctr2}{\rjlpshap@maxA}%
}{%
\setcounter{rjlpshap@ctr2}{\value{rjlpshap@ctr}}%
\addtocounter{rjlpshap@ctr2}{\value{rjlpshap@ctr}}%
}%
% \end{macrocode}
% left margin from \#1:
% \begin{macrocode}
\parshapearrlenout{#1}{rjlpshap@ctr2}%
% \end{macrocode}
% get the line width
% \begin{macrocode}
\ifthenelse{\value{rjlpshap@ctr} > \rjlpshap@maxB}{%
\edef\rjlpshap@pos{\rjlpshap@maxB}%
\setcounter{rjlpshap@ctr3}{\rjlpshap@maxB}%
\typeout{A pos=\rjlpshap@pos\space of \rjlpshap@maxB}%
\testarray{#2}(\arabic{rjlpshap@ctr3})%
\typeout{temp=\temp@macro}%
\setlength{\rjlpshap@templinewidth}{\temp@macro}%
}{%
\stepcounter{rjlpshap@ctr}%
\edef\rjlpshap@pos{\arabic{rjlpshap@ctr}}%
\typeout{B pos=\rjlpshap@pos}%
\addtocounter{rjlpshap@ctr}{-1}%
\testarray{#2}(\rjlpshap@pos)%
\setlength{\rjlpshap@templinewidth}{\temp@macro}%
}%
\typeout{width=\the\rjlpshap@templinewidth}%
% \end{macrocode}
% subtract the left margin from the line width
% \begin{macrocode}
\addtolength{\rjlpshap@templinewidth}{-1\rjlpshap@tempwidth}%
% \end{macrocode}
% subtract the right margin from the line width
% \begin{macrocode}
\stepcounter{rjlpshap@ctr2}%
\edef\rjlpshap@pos{\arabic{rjlpshap@ctr2}}%
\testarray{#1}(\rjlpshap@pos)%
\setlength{\rjlpshap@tempwidth}{\temp@macro}%
\addtolength{\rjlpshap@templinewidth}{-1\rjlpshap@tempwidth}%
\parshapelenout{\rjlpshap@templinewidth}%
}%
\end{parshapecollect}%
}
% \end{macrocode}
% \end{macro}
% \subsection{Internal Definitions}
% rjlpshap@ctr: outermost loop counter:
% \begin{macrocode}
\newcounter{rjlpshap@ctr}
% \end{macrocode}
% rjlpshap@ctr2: used for index calculations internally:
% \begin{macrocode}
\newcounter{rjlpshap@ctr2}
% \end{macrocode}
% rjlpshap@ctr3: used for index calculations internally:
% \begin{macrocode}
\newcounter{rjlpshap@ctr3}
% \end{macrocode}
% rjlpshap@tempwidth: used as a register to read widths from arrays,
% and for calculations
% \begin{macrocode}
\newlength{\rjlpshap@tempwidth}
% \end{macrocode}
% rjlpshap@templinewidth: used as a register to calculate the line
% width when subtracting margins:
% \begin{macrocode}
\newlength{\rjlpshap@templinewidth}
% \end{macrocode}
% boolean used for breaking out of loops
% \begin{macrocode}
\newboolean{rjlpshap@loop}
% \end{macrocode}
% boolean used to track if the number of lengths output is odd or even
% \begin{macrocode}
\newboolean{rjlpshap@isodd}
% \end{macrocode}
% Filehandles needed for reading and writing temporary files
% \begin{macrocode}
\newwrite\rjlpshap@writer % holds \parshape lengths
\newwrite\rjlpshap@finalwriter % holds entire \parshape command
\newread\rjlpshap@reader % used to read writer to transfer to finalwriter
% \end{macrocode}
% rjlpshap@linecount: counts the number of lines read in the environment
% \begin{macrocode}
\newcounter{rjlpshap@linecount}
% \end{macrocode}
% Finally, undo the effect of $\backslash$makeatletter, and make @ be
% a special character again.
% \begin{macrocode}
\makeatother
% \end{macrocode}
% \Finale
\endinput