summaryrefslogtreecommitdiff
path: root/macros/latex/contrib/mathtools/empheq.dtx
diff options
context:
space:
mode:
authorNorbert Preining <norbert@preining.info>2019-09-02 13:46:59 +0900
committerNorbert Preining <norbert@preining.info>2019-09-02 13:46:59 +0900
commite0c6872cf40896c7be36b11dcc744620f10adf1d (patch)
tree60335e10d2f4354b0674ec22d7b53f0f8abee672 /macros/latex/contrib/mathtools/empheq.dtx
Initial commit
Diffstat (limited to 'macros/latex/contrib/mathtools/empheq.dtx')
-rw-r--r--macros/latex/contrib/mathtools/empheq.dtx4354
1 files changed, 4354 insertions, 0 deletions
diff --git a/macros/latex/contrib/mathtools/empheq.dtx b/macros/latex/contrib/mathtools/empheq.dtx
new file mode 100644
index 0000000000..45dace201f
--- /dev/null
+++ b/macros/latex/contrib/mathtools/empheq.dtx
@@ -0,0 +1,4354 @@
+% \iffalse meta-comment
+%
+% Copyright (C) 2002-2007 by Morten Hoegholm
+% Copyright (C) 2012 by Lars Madsen
+%
+% This work may be distributed and/or modified under the
+% conditions of the LaTeX Project Public License, either
+% version 1.3 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 1.3 or later is part of all distributions of
+% LaTeX version 2005/12/01 or later.
+%
+% This work has the LPPL maintenance status "maintained".
+%
+% This Current Maintainer of this work is
+% Lars Madsen
+%
+% This work consists of the main source file empheq.dtx
+% and the derived files
+% empheq.sty, empheq.pdf, empheq.ins, empheq.drv.
+%
+% Distribution:
+% CTAN:macros/latex/contrib/mh/empheq.dtx
+% CTAN:macros/latex/contrib/mh/empheq.pdf
+%
+% Unpacking:
+% (a) If empheq.ins is present:
+% tex empheq.ins
+% (b) Without empheq.ins:
+% tex empheq.dtx
+% (c) If you insist on using LaTeX
+% latex \let\install=y\input{empheq.dtx}
+% (quote the arguments according to the demands of your shell)
+%
+% Documentation:
+% (a) If empheq.drv is present:
+% latex empheq.drv
+% (b) Without empheq.drv:
+% latex empheq.dtx; ...
+% The class ltxdoc loads the configuration file ltxdoc.cfg
+% if available. Here you can specify further options, e.g.
+% use A4 as paper format:
+% \PassOptionsToClass{a4paper}{article}
+%
+% Programm calls to get the documentation (example):
+% pdflatex empheq.dtx
+% makeindex -s gind.ist empheq.idx
+% pdflatex empheq.dtx
+% makeindex -s gind.ist empheq.idx
+% pdflatex empheq.dtx
+%
+% Installation:
+% TDS:tex/latex/mh/empheq.sty
+% TDS:doc/latex/mh/empheq.pdf
+% TDS:source/latex/mh/empheq.dtx
+%
+%<*ignore>
+\begingroup
+ \def\x{LaTeX2e}
+\expandafter\endgroup
+\ifcase 0\ifx\install y1\fi\expandafter
+ \ifx\csname processbatchFile\endcsname\relax\else1\fi
+ \ifx\fmtname\x\else 1\fi\relax
+\else\csname fi\endcsname
+%</ignore>
+%<*install>
+\input docstrip.tex
+\Msg{************************************************************************}
+\Msg{* Installation}
+\Msg{* Package: empheq 2017/03/31 v2.15 empheq}
+\Msg{************************************************************************}
+
+\keepsilent
+\askforoverwritefalse
+
+\preamble
+
+This is a generated file.
+
+Copyright (C) 2002-2007 by Morten Hoegholm
+Copyright (C) 2012- by Lars Madsen
+
+
+This work may be distributed and/or modified under the
+conditions of the LaTeX Project Public License, either
+version 1.3 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 1.3 or later is part of all distributions of
+LaTeX version 2005/12/01 or later.
+
+This work has the LPPL maintenance status "maintained".
+
+This Current Maintainer of this work is
+Lars Madsen
+
+This work consists of the main source file empheq.dtx
+and the derived files
+ empheq.sty, empheq.pdf, empheq.ins, empheq.drv.
+
+\endpreamble
+
+\generate{%
+ \file{empheq.ins}{\from{empheq.dtx}{install}}%
+ \file{empheq.drv}{\from{empheq.dtx}{driver}}%
+ \usedir{tex/latex/mh}%
+ \file{empheq.sty}{\from{empheq.dtx}{package}}%
+}
+
+\obeyspaces
+\Msg{************************************************************************}
+\Msg{*}
+\Msg{* To finish the installation you have to move the following}
+\Msg{* file into a directory searched by TeX:}
+\Msg{*}
+\Msg{* empheq.sty}
+\Msg{*}
+\Msg{* To produce the documentation run the file `empheq.drv'}
+\Msg{* through LaTeX.}
+\Msg{*}
+\Msg{* Happy TeXing!}
+\Msg{*}
+\Msg{************************************************************************}
+
+\endbatchfile
+%</install>
+%<*ignore>
+\fi
+%</ignore>
+%<*driver>
+\NeedsTeXFormat{LaTeX2e}
+\ProvidesFile{empheq.drv}%
+ [2017/03/31 v2.15 Emphasizing equations]
+\documentclass{ltxdoc}
+\usepackage{mathpazo}
+\usepackage[overload]{empheq}
+\usepackage{xcolor,varioref,booktabs,fancybox}
+\usepackage[overload]{textcase}
+
+\pagestyle{headings}
+\hyphenation{emph-eq}
+\makeatletter
+\providecommand*\pkg[1]{\textsf{#1}}
+\providecommand*\env[1]{\texttt{#1}}
+\providecommand*\email[1]{\href{mailto:#1}{\texttt{#1}}}
+\providecommand*\mode[1]{\texttt{[#1]}}
+\providecommand*\file[1]{\texttt{#1}}
+\addtolength\marginparwidth{-30pt}
+
+
+\newcommand*\AddDescribe[3]{%
+ \@namedef{Describe#1}{\leavevmode\@bsphack\begingroup\MakePrivateLetters
+ \@nameuse{Describe@#1}}
+ \@namedef{Describe@#1}##1{\endgroup
+ \marginpar{\raggedleft\@nameuse{PrintDescribe#1}{##1}}%
+ \@nameuse{Special#1Index}{##1}\@esphack\ignorespaces}
+ \@namedef{Special#1Index}##1{\@bsphack
+ \index{##1\actualchar{\protect\ttfamily ##1}
+ (#2)\encapchar usage}%
+ \index{#3:\levelchar##1\actualchar{\protect\ttfamily ##1}\encapchar
+ usage}\@esphack}
+ \@namedef{PrintDescribe#1}##1{\strut \MacroFont ##1\ }%
+ \@namedef{Print#1Name}##1{\strut \MacroFont ##1\ }%
+}
+\AddDescribe{Color}{color}{colors}
+\AddDescribe{KeyFamily}{key family}{key families}
+\newcommand*\DescribeDelimiter[1]{%
+ \cs{#1}&\SpecialMkhIndex{empheq#1}\cs{empheq}\texttt{#1} &
+ \SpecialMkhIndex{empheqbig#1}\cs{empheqbig}\texttt{#1} &
+ $\@nameuse{#1}$\tabularnewline}
+\newcommand*\SpecialMkhIndex[1]{%
+ \@bsphack
+ \index{#1\actualchar{\protect\ttfamily \protect\bslash\space#1}
+ (` $\bslash#1$ ')\encapchar usage}%
+ \@esphack
+}
+\newcommand\TabDescribeOption[3][(default)]{%
+ \SpecialOptIndex{#2}\opt{#2} & #3 \tabularnewline
+}
+\newcommand*\TabDescribeAmsKey[2][]{%
+ \SpecialKeyIndex{#2}\env{#2} &
+ \texttt{\{#2#1\}} &
+ \SpecialKeyIndex{#2*}\env{#2*} &
+ \texttt{\{#2*#1\}} \tabularnewline
+}
+\newcommand*\TabDescribeOptKey[3][]{%
+ \SpecialKeyIndex{#2}\key{#2} &
+ \texttt{#2=}\meta{#3} & #1 \tabularnewline
+}
+
+\newenvironment{codesyntax}
+ {\par\small\addvspace{4.5ex plus 1ex}%
+ \vskip -\parskip
+ \noindent
+ \begin{tabular}{|l|}\hline\ignorespaces}%
+ {\\\hline\end{tabular}\nobreak\par\nobreak
+ \vspace{2.3ex}\vskip -\parskip\noindent\ignorespacesafterend}
+\makeatletter
+
+\newcommand*\FeatureRequest[2]{%
+ \hskip1sp
+ \marginpar{%
+ \parbox[b]{\marginparwidth}{\small\sffamily\raggedright
+ \strut Feature request by\\#1\\#2%
+ }
+ }%
+}
+
+\newcommand*\cttPosting[2]{%
+ \hskip1sp
+ \marginpar{%
+ \parbox[b]{\marginparwidth}{\small\sffamily\raggedright
+ \strut Posted on \texttt{comp.text.tex} \\#1\\#2%
+ }%
+ }%
+}
+
+\expandafter\def\expandafter\MakePrivateLetters\expandafter{%
+ \MakePrivateLetters \catcode`\_=11\relax
+}
+
+\providecommand*\SpecialOptIndex[1]{%
+ \@bsphack
+ \index{#1\actualchar{\protect\ttfamily #1}
+ (option)\encapchar usage}%
+ \index{options:\levelchar#1\actualchar{\protect\ttfamily #1}\encapchar
+ usage}\@esphack}
+\providecommand*\opt[1]{\texttt{#1}}
+
+\providecommand*\SpecialKeyIndex[1]{%
+ \@bsphack
+ \index{#1\actualchar{\protect\ttfamily #1}
+ (key)\encapchar usage}%
+ \index{keys:\levelchar#1\actualchar{\protect\ttfamily #1}\encapchar
+ usage}\@esphack}
+\providecommand*\key[1]{\textsf{#1}}
+\providecommand*\keyfam[1]{\textsf{#1}}
+
+\providecommand*\eTeX{$\m@th\varepsilon$-\TeX}
+
+\def\MTmeta#1{%
+ \ensuremath\langle
+ \ifmmode \expandafter \nfss@text \fi
+ {%
+ \meta@font@select
+ \edef\meta@hyphen@restore
+ {\hyphenchar\the\font\the\hyphenchar\font}%
+ \hyphenchar\font\m@ne
+ \language\l@nohyphenation
+ #1\/%
+ \meta@hyphen@restore
+ }\ensuremath\rangle
+ \endgroup
+}
+\makeatother
+\DeclareRobustCommand\meta{\begingroup\MakePrivateLetters\MTmeta}%
+\def\MToarg#1{{\ttfamily[}\meta{#1}{\ttfamily]}\endgroup}
+\DeclareRobustCommand\oarg{\begingroup\MakePrivateLetters\MToarg}%
+\def\MHmarg#1{{\ttfamily\char`\{}\meta{#1}{\ttfamily\char`\}}\endgroup}
+\DeclareRobustCommand\marg{\begingroup\MakePrivateLetters\MHmarg}%
+\def\MHarg#1{{\ttfamily\char`\{#1\ttfamily\char`\}}\endgroup}
+\DeclareRobustCommand\arg{\begingroup\MakePrivateLetters\MHarg}%
+\def\MHcs#1{\texttt{\char`\\#1}\endgroup}
+\DeclareRobustCommand\cs{\begingroup\MakePrivateLetters\MHcs}
+\def\MHbool#1{{\ttfamily #1}\endgroup}
+\DeclareRobustCommand\bool{\begingroup\MakePrivateLetters\MHbool}%
+
+\def\endverbatim{\if@newlist
+\leavevmode\fi\endtrivlist\vspace{-\baselineskip}}
+\expandafter\let\csname endverbatim*\endcsname =\endverbatim
+
+\let\MTtheindex\theindex
+\def\theindex{\MTtheindex\MakePrivateLetters}
+
+\newcommand*\widefbox[1]{\fbox{\hspace{1em}#1\hspace{1em}}}
+\definecolor{myblue}{rgb}{.8, .8, 1}
+\newcommand*\mybluebox[1]{%
+ \colorbox{myblue}{\hspace{1em}#1\hspace{1em}}}
+\newenvironment{important}[2][]{%
+ \setkeys{EmphEqEnv}{#2}%
+ \setkeys{EmphEqOpt}{box=\mybluebox,#1}%
+ \EmphEqMainEnv}%
+ {\endEmphEqMainEnv}
+\definecolor{shadecolor}{cmyk}{0,0,0.41,0}
+\definecolor{light-blue}{cmyk}{0.25,0,0,0}
+\newsavebox{\mysaveboxM} % M for math
+\newsavebox{\mysaveboxT} % T for text
+\newcommand*\Garybox[2][Example]{%
+ \sbox{\mysaveboxM}{#2}%
+ \sbox{\mysaveboxT}{\fcolorbox{black}{light-blue}{#1}}%
+ \sbox{\mysaveboxM}{%
+ \parbox[b][\ht\mysaveboxM+.5\ht\mysaveboxT+.5\dp\mysaveboxT][b]{%
+ \wd\mysaveboxM}{#2}%
+ }%
+ \sbox{\mysaveboxM}{%
+ \fcolorbox{black}{shadecolor}{%
+ \makebox[\linewidth-10em]{\usebox{\mysaveboxM}}%
+ }%
+ }%
+ \usebox{\mysaveboxM}%
+ \makebox[0pt][r]{%
+ \makebox[\wd\mysaveboxM][c]{%
+ \raisebox{\ht\mysaveboxM-0.5\ht\mysaveboxT
+ +0.5\dp\mysaveboxT-0.5\fboxrule}{\usebox{\mysaveboxT}}%
+ }%
+ }%
+}
+
+\newcommand\EmphEqpart{%
+ \def\partname{\EmphEqpartname}%
+ \part}
+
+\def\EmphEqpartname{Part}
+\makeatother
+
+\usepackage{graphicx}
+\usepackage[final,breaklinks=true,hyperindex=false]{hyperref}
+\renewcommand*\usage[1]{\textit{\hyperpage{#1}}}
+
+\OnlyDescription
+\begin{document}
+\DocInput{empheq.dtx}
+\end{document}
+%</driver>
+% \fi
+% \changes{v2.15}{2017/03/31}{Fixed a bug in relation to the AMS
+% classes and completed prefixing macros belonging to \pkg{mhsetup}
+% and \pkg{empheq}}
+% \changes{v2.14}{2014/07/16}{Updated maintenance and contact info}
+%
+% \changes{v2.11}{2004/08/03}{Removed interface to the obsolete empheq
+% package \pkg{empheq07}}
+%
+% \changes{v2.10}{2004/07/26}{Rewrote most of the documentation}
+%
+% \changes{v2.10}{2004/07/26}{Moved options \texttt{fixamsmath} and
+% \texttt{donotfixamsmathbugs} to \pkg{mathtools}}
+%
+% \changes{v2.10}{2004/07/26}{Added support for \pkg{showkeys}}
+%
+% \changes{v2.00}{2004/04/14}{Added options \opt{overload},
+% \opt{overload2}, \texttt{fixamsmath}, and
+% \texttt{donotfixamsmathbugs} plus descriptions of their use}
+%
+% \changes{v1.06}{2004/03/13}{Upgraded license to LPPL version 1.3}
+%
+% \changes{v1.00}{2004/02/15}{Complete rewrite of the package hence
+% no description of the changes made}
+%
+% \changes{v0.7a}{2004/02/15}{Made the old package obsolete and
+% changed its name to \pkg{empheq-obsolete}}
+%
+% \changes{v0.7}{2003/05/11}{Added further examples and changed
+% existing ones}
+% \changes{v0.6b}{2003/03/04}{Added example for changing margins
+% inside the boxtype}
+% \changes{v0.6a}{2003/01/28}{Minor updates to documentation}
+%
+% \changes{v0.5}{2002/12/30}{First experimental and very limited
+% release}
+%
+% \GetFileInfo{empheq.drv}
+% \CheckSum{2495}
+%
+% \title{The \pkg{empheq} package\thanks{This file has version number
+% \fileversion, last revised \filedate.}\\
+% Emphasizing equations in \LaTeXe}
+% \author{Morten H\o gholm, Lars Madsen}
+%
+% \date{\filedate}
+%
+% \maketitle
+%
+% \begin{abstract}
+% \noindent The \pkg{empheq} package can best be described as a visual markup
+% extension to \pkg{amsmath}. In short it offers a)~a multi line
+% equivalent of \cs{boxed} from \pkg{amsmath} and b)~a way to
+% produce arbitrary delimiters that span entire math displays.
+% \end{abstract}
+%
+% \tableofcontents
+%
+% \clearpage
+% \EmphEqpart{Basic user's guide}
+%
+% There can be little doubt that the de facto standard for
+% mathematical typesetting in \LaTeX{} is the \pkg{amsmath} package.
+% For the creation of \pkg{empheq}, a visual markup package for use
+% in math, it made perfect sense to have \pkg{amsmath} as the
+% backbone.
+%
+% The main idea of \pkg{empheq} is to maintain the familiar syntax
+% of the \pkg{amsmath} environments while still providing an easy
+% way of specifying markup instructions. This manual is plastered
+% with examples showing just how.
+%
+%
+% \section{Basic use of the package}
+%
+% So what is it \pkg{empheq} does? Well, it allows you to produce
+% displays like this:
+% \begin{subequations}
+% \begin{empheq}[
+% box=\mybluebox,
+% right={\;\makebox[.9em]{$\raisebox{-.5\totalheight+\fontdimen22\textfont2}
+% {\resizebox{!}{\EmphEqdisplayheight+\EmphEqdisplaydepth}{!}}$}},
+% left={X=Y\Rightarrow\empheqlbrace}
+% ]{alignat=3}
+% A_1&=b_1 & \qquad c_1&=d_1 & \qquad e_1&= f_1
+% \tag*{\_A\raisebox{1ex}{silly}\raisebox{-1ex}{tag}\_}\\
+% A_2&=b_2 & \qquad c_2&=d_2 & \qquad e_2&= f_2 \\
+% A_3&=b_3 & \qquad c_3&=d_3 & \qquad e_3&= f_3
+% \end{empheq}
+% \end{subequations}
+% In short \pkg{empheq} enables the user to put things on every side
+% of the display without said user having to worry about what
+% happens to the equation numbers. For example you can now have a
+% display containing multiple lines and still get the effect of the
+% \cs{boxed} command from \pkg{amsmath}.
+%
+%
+% \subsection{Using the \env{empheq} environment}
+%
+% The package defines a single environment
+% \SpecialEnvIndex{empheq}\env{empheq} and the usage is kind of straight
+% forward:
+% \begin{codesyntax}
+% |\begin{empheq}|\oarg{markup instructions}\marg{\AmS{} env_name}\\
+% \mbox{}\quad\meta{contents of \AmS{} environment}\\
+% |\end{empheq}|
+% \end{codesyntax}
+% A first minimal example file would then be something like
+% \begin{verbatim}
+% \documentclass{minimal}
+% \usepackage{empheq}
+% \begin{document}
+% \begin{empheq}{align*}
+% a&=b \tag{*}\\
+% E&=mc^2 + \int_a^a x\, dx
+% \end{empheq}
+% \end{document}
+% \end{verbatim}
+% This gives the following display:
+% \begin{empheq}{align*}
+% a&=b \tag{*}\\
+% E&=mc^2 + \int_a^a x\, dx
+% \end{empheq}
+% Maybe not the most impressive example, but as you can see the
+% contents of the environment is exactly the same as for the regular
+% \env{align*} evironment from \pkg{amsmath}. The rest of the \AmS{}
+% environments are chosen the same way by typing the name as the
+% mandatory argument of \env{empheq} with the exception of the
+% \env{alignat} environment. For this you have to specify the number
+% of columns as shown below.
+% \begin{verbatim}
+% \begin{empheq}{alignat=2}
+% a &= b &\quad c &= d \\
+% \text{this} &= \text{that} &\quad \mathit{fish}&\neq fish
+% \end{empheq}
+% \end{verbatim}
+% \begin{empheq}{alignat=2}
+% a &= b &\quad c &= d \\
+% \text{this} &= \text{that} &\quad \mathit{fish}&\neq fish
+% \end{empheq}
+% To choose the starred version of \env{alignat}, simply type
+% \arg{alignat*=2} instead in the above example.
+%
+% The supported \AmS{} environments are listed in Table
+% \vref{tab:environments}. Not supported is the standard \LaTeX{}
+% \env{eqnarray} environment as it is fundamentally
+% deficient.\footnote{See for instance
+% \url{http://www.tex.ac.uk/cgi-bin/texfaq2html?label=eqnarray}}
+%
+% \begin{table}
+% \centering
+% \begin{tabular}{llll}\toprule
+% Environment & Usage & Environment & Usage \\ \midrule
+% \TabDescribeAmsKey{equation}
+% \TabDescribeAmsKey{align}
+% \TabDescribeAmsKey{gather}
+% \TabDescribeAmsKey{flalign}
+% \TabDescribeAmsKey[=\textrm{\meta{cols}}]{alignat}
+% \TabDescribeAmsKey{multline} \bottomrule
+% \end{tabular}
+% \caption{The supported \pkg{amsmath}
+% environments}\label{tab:environments}
+% \end{table}
+%
+%
+% \subsection{Markup instructions}
+%
+% The optional argument of the \env{empheq} environment will take
+% markup instructions consisting of a $\meta{key}=\meta{value}$ list
+% of assignments. There are currently five such keys (a sixth is
+% added for naming consistency). They're shown in Table
+% \vref{tab:keys}.
+%
+% \begin{table}
+% \centering
+% \begin{tabular}{llp{13em}}\toprule
+% Key & Usage & Additional Info \tabularnewline \midrule
+% \TabDescribeOptKey{box}{box command}
+% \TabDescribeOptKey{innerbox}{box command}
+% \TabDescribeOptKey[Use \cs{text}\arg{...} if you need text
+% material.]{left}{math material}
+% \TabDescribeOptKey[Use \cs{text}\arg{...} if you need text
+% material.]{right}{math material}
+% \TabDescribeOptKey[Alias for \key{box}.]{outerbox}{box command}
+% \TabDescribeOptKey[Contents must be horizontally centered;
+% can only be used in \mode{fleqn} mode.]
+% {marginbox}{box command}
+% \bottomrule
+% \end{tabular}
+% \caption{The six keys for the optional argument of the
+% \env{empheq} environment}\label{tab:keys}
+% \end{table}
+%
+% \begin{codesyntax}
+% \SpecialKeyIndex{left}\key{left}
+% \end{codesyntax}
+% The key \key{left} is for material put on the left side of the
+% display. The material is typeset in math mode and centered
+% vertically.
+% \begin{verbatim}
+% \begin{empheq}[left=L\Rightarrow]{align}
+% a&=b\\
+% E&=mc^2 + \int_a^a x\, dx
+% \end{empheq}
+% \end{verbatim}
+% \begin{empheq}[left=L\Rightarrow]{align}
+% a&=b\\
+% E&=mc^2 + \int_a^a x\, dx
+% \end{empheq}
+%
+%
+% \begin{codesyntax}
+% \SpecialKeyIndex{right}\key{right}
+% \end{codesyntax}
+% As there is a \key{left} key it hopefully comes as no surprise
+% that there is a companion \key{right} key for typesetting material
+% on the right side of the display.
+% \begin{verbatim}
+% \begin{empheq}[right=\Leftarrow R]{align}
+% a&=b\\
+% E&=mc^2 + \int_a^a x\, dx
+% \end{empheq}
+% \end{verbatim}
+% \begin{empheq}[right=\Leftarrow R]{align}
+% a&=b\\
+% E&=mc^2 + \int_a^a x\, dx
+% \end{empheq}
+%
+% \begin{codesyntax}
+% \SpecialKeyIndex{box}\key{box}\\
+% \SpecialKeyIndex{outerbox}\key{outerbox}
+% \end{codesyntax}
+% The key \key{box} specifies the kind of box you would like to put
+% around the display. It can be any kind of box, as long as the
+% contents of the box is situated on the baseline like in a
+% \fbox{\cs{fbox}}.
+% \begin{verbatim}
+% \begin{empheq}[box=\fbox]{align}
+% a&=b\\
+% E&=mc^2 + \int_a^a x\, dx
+% \end{empheq}
+% \end{verbatim}
+% \begin{empheq}[box=\fbox]{align}
+% a&=b\\
+% E&=mc^2 + \int_a^a x\, dx
+% \end{empheq}
+% The key \key{outerbox} is an alias for \key{box} and is added for
+% naming consistency with the key described below.
+%
+% \begin{codesyntax}
+% \SpecialKeyIndex{innerbox}\key{innerbox}
+% \end{codesyntax}
+% There is also an \key{innerbox} key. It is not very interesting
+% unless you use one of the other keys.
+% \begin{verbatim}
+% \begin{empheq}[innerbox=\fbox,
+% left=L\Rightarrow]{align}
+% a&=b\\
+% E&=mc^2 + \int_a^a x\, dx
+% \end{empheq}
+% \end{verbatim}
+% \begin{empheq}[innerbox=\fbox,
+% left=L\Rightarrow]{align}
+% a&=b\\
+% E&=mc^2 + \int_a^a x\, dx
+% \end{empheq}
+%
+% \begin{codesyntax}
+% \SpecialKeyIndex{marginbox}\key{marginbox}
+% \end{codesyntax}
+% \FeatureRequest{Uwe Siart}{2003/12/08}
+% The last key is the \key{marginbox} key. If you typeset your math
+% in \mode{fleqn} mode\footnote{Sorry, but I can't show you because
+% this manual features centered math displays.} you may want the
+% math display and not the outer box to align at the left margin (or
+% rather: the indentation). If you make sure the contents of the
+% outer box is centered horizontally inside the box, \key{marginbox}
+% will align it properly. You shouldn't set both \key{box} and
+% \key{marginbox} at the same time, as this is surely not what you
+% want and the package will silently use the last one in the list.
+%
+%
+% \begin{quote}
+% \textbf{Warning}: \pkg{keyval} treats commas and equal signs as
+% separator and assignment signs which in turn means that if you
+% want to typeset either of them, they \emph{must}---and believe me,
+% it's really important---be enclosed in braces. In short:
+% \begin{description}
+% \item[{\makebox[3em][l]{Good}}] \texttt{[left=\{(A,B)=(1,0)\}]}
+% \item[{\makebox[3em][l]{Bad}}] \texttt{[left=(A,B)=(1,0)]}
+% \end{description}
+% \end{quote}
+%
+% \begin{codesyntax}
+% \SpecialUsageIndex{\empheqset}\cs{empheqset}\marg{markup instructions}
+% \end{codesyntax}
+% \FeatureRequest{Uwe Siart}{2004/01/13}
+% You can also set the keys globally\footnote{Or rather: outside the
+% scope of the \env{empheq} environment. The settings will still
+% obey scoping rules.} with the command
+% \cs{empheqset}. This means that
+% \begin{verbatim}
+% \empheqset{marginbox=\psframebox}
+% \end{verbatim}
+% will force a |box=\psframebox| in all occurrences of the
+% \env{empheq} environment, but an explicitly given \key{box} or
+% \key{marginbox} will override this setting. You can only use the
+% keys listed in this section as arguments to \cs{empheqset}; not
+% the \AmS\ environments.
+%
+%
+%
+% \subsection{What won't work in the \env{empheq} environment}
+%
+% \begin{codesyntax}
+% \SpecialUsageIndex{intertext}\cs{intertext}\marg{text}\\
+% \SpecialUsageIndex{displaybreak}\cs{displaybreak}\oarg{num}
+% \end{codesyntax}
+% Now we've seen some of the things that work in the \env{empheq}
+% environment but we also have to take note of what won't work. As
+% this package provides a way to box multi line math displays it can
+% come as no surprise that using either \cs{intertext}\footnote{And
+% thus also \cs{shortintertext} (\pkg{mathtools}).} or
+% \cs{displaybreak} inside the \env{empheq} environment makes no
+% sense. Should you however happen to try them anyway, you'll
+% experience that \cs{intertext} issues an error message and
+% \cs{displaybreak} issues a warning. The reason only \cs{intertext}
+% gives an error message is that you'll get output very different
+% from what you expect, and that is not the case (so much at least)
+% with \cs{displaybreak}. But of course, you'll never see either
+% message because you read the manual!
+% \begin{codesyntax}
+% Very long lines
+% \end{codesyntax}
+% The standard \AmS\ environments will move the equation number if a
+% line gets too close to it. Due to implementation, this feature is
+% \emph{not} carried correctly to the \env{empheq} environments, so keep
+% equation lines short then applying \env{empheq} to them.
+%
+% \subsection{Special delimiters}
+%
+% As you've seen a few pages back, it's possible to add material on
+% both sides of the math display. When doing so you often need a
+% delimiter that scales to fit the entire display, so it comes as no
+% surprise that \pkg{empheq} provides such delimiters.
+% \begin{codesyntax}
+% \cs{empheql}\meta{delim_name} \cs{empheqbigl}\meta{delim_name}\\
+% \cs{empheqr}\meta{delim_name} \cs{empheqbigr}\meta{delim_name}
+% \end{codesyntax}
+% For example one might need a large brace like the one in the
+% \env{cases} environment:
+% \begin{verbatim}
+% \begin{empheq}[left=\empheqlbrace, right=\empheqrbrace]{align}
+% E&=mc^2 \\
+% Y&= \sum_{n=1}^\infty \frac{1}{n^2}
+% \end{empheq}
+% \end{verbatim}
+% \begin{empheq}[left=\empheqlbrace, right=\empheqrbrace]{align}
+% E&=mc^2 \\
+% Y&= \sum_{n=1}^\infty \frac{1}{n^2}
+% \end{empheq}
+% The naming scheme is \cs{empheq}\meta{delimiter name} and
+% \cs{empheqbig}\meta{delimiter name}. Thus \cs{empheqrbrace}
+% produces an auto-scaling right brace with same size as the math
+% display while \cs{empheqbigrbrace} produces an even bigger version
+% that spans the inner box plus the math display inside it. For a
+% complete list of supported delimiters see Table
+% \vref{tab:delimiters}.
+%
+% \begin{table}
+% \centering
+% \begin{tabular}{lllc} \toprule
+% \multicolumn{3}{c}{\textbf{Delimiter}}\\ \cmidrule{1-3}
+% \multicolumn{1}{c}{Original} & \multicolumn{1}{c}{Normal} &
+% \multicolumn{1}{c}{Bigger} & Symbol\\ \midrule
+% \DescribeDelimiter{lbrace} \DescribeDelimiter{rbrace}
+% \DescribeDelimiter{lbrack} \DescribeDelimiter{rbrack} \midrule
+% \DescribeDelimiter{langle} \DescribeDelimiter{rangle}
+% \DescribeDelimiter{lparen} \DescribeDelimiter{rparen} \midrule
+% \DescribeDelimiter{lvert} \DescribeDelimiter{rvert}
+% \DescribeDelimiter{lVert} \DescribeDelimiter{rVert} \midrule
+% \DescribeDelimiter{lfloor} \DescribeDelimiter{rfloor}
+% \DescribeDelimiter{lceil} \DescribeDelimiter{rceil} \bottomrule
+% \end{tabular}
+% \caption{The supported auto-scaling delimiters in
+% \pkg{empheq}}\label{tab:delimiters}
+% \end{table}
+%
+%
+%
+% \section{Taking things a little further}
+%
+% So far we have covered the basic functions of \pkg{empheq}: the
+% markup instructions and the predefined auto-scaling delimiters.
+% There is more to \pkg{empheq} however, and we'll cover that in
+% this section.
+%
+% \subsection{Loading the package}
+%
+% The package has two main requirements: One is the
+% \pkg{mathtools}\footnote{By the same author and is distributed
+% with \pkg{empheq}. See its documentation for more information.}
+% package which provides \pkg{empheq} with a lot of necessary tools
+% for doing its thing. The other one is \pkg{amsmath} of course, and
+% to make the loading procedure as easy as possible, you can simply
+% substitute
+% \begin{verbatim}
+% \usepackage[leqno,fleqn,intlimits]{amsmath}
+% \end{verbatim}
+% with
+% \begin{verbatim}
+% \usepackage[leqno,fleqn,intlimits]{empheq}
+% \end{verbatim}
+% \pkg{empheq} makes sure that the \pkg{amsmath} options are passed
+% on and loaded by \pkg{amsmath}. The same goes for the options
+% provided by \pkg{mathtools}.
+%
+%
+% \subsubsection{Compatibility with the old version}
+%
+% The current version (\fileversion) of \pkg{empheq} is incompatible with
+% versions prior to and including~0.7e. If you have documents produced
+% with versions prior to~0.7e, then you have to load the package
+% \pkg{empheq07} instead. \pkg{empheq07} now exists as a separate package
+% package with its own documentation, but I strongly recommend switching
+% to the new version because it is so much better.
+%
+% \subsection{Package options}
+%
+% In addition to the \pkg{amsmath} options, \pkg{empheq} itself
+% provides a string of options listed in Table \vref{tab:options}.
+%
+% \begin{table}[ht]
+% \centering
+% \begin{tabular}{lp{21em}} \toprule
+% Option & Short description \tabularnewline \midrule
+% \TabDescribeOption[]{overload}
+% {Lets you use the \pkg{empheq} visual markup extensions
+% transparently in an existing document.}
+% \TabDescribeOption[]{overload2}
+% {A wild version of \opt{overload}. Use with care.}
+% \TabDescribeOption[]{ntheorem}
+% {A support module for users of \pkg{ntheorem} and its
+% \opt{thmmarks} and \opt{thref} options.}
+% \TabDescribeOption[]{newmultline}
+% {With v2.10 of \pkg{empheq} the syntax for \env{multline} and
+% \env{multlined} has been improved immensely.}
+% \TabDescribeOption[]{oldmultline}
+% {Support for the somewhat strange syntax for \env{multline} and
+% \env{multlined} in v2.00.}
+% \bottomrule
+% \end{tabular}
+% \caption{Supported options in the \pkg{empheq} package}
+% \label{tab:options}
+% \end{table}
+%
+%
+% \begin{codesyntax}
+% \SpecialOptIndex{ntheorem}\opt{ntheorem}
+% \end{codesyntax}
+% The \pkg{ntheorem} package is supported by means of the
+% \opt{ntheorem} option. In order for this to work properly, an
+% \pkg{amsmath} bug\footnote{See the \LaTeX{} Bugs Database
+% \url{http://www.latex-project.org/cgi-bin/ltxbugs2html} under
+% \AmS\LaTeX{} problem report 3624.} is fixed. For more information
+% on this option see \S \vref{sec:ntheorem-support}.
+%
+% \subsubsection{Using \env{multline}}
+%
+% \begin{codesyntax}
+% \SpecialOptIndex{newmultline}\opt{newmultline}\\
+% \SpecialOptIndex{oldmultline}\opt{oldmultline}
+% \end{codesyntax}
+% Those familiar with v2.00 of this package will recognize the
+% somewhat weird syntax for using \env{multline} and
+% \env{multlined}. The good news is that since v2.10 there is now a
+% much improved syntax available. The option \opt{newmultline}
+% (default) selects this better interface, while the old interface
+% is provided for compatibility reasons by the \opt{oldmultline}
+% option.
+%
+%
+% \begin{codesyntax}
+% \SpecialUsageIndex{\shoveleft}\cs{shoveleft}\oarg{dimen}\marg{arg}\\
+% \SpecialUsageIndex{\shoveright}\cs{shoveright}\oarg{dimen}\marg{arg}
+% \end{codesyntax}
+% With the new interface you also get an extended syntax for
+% \cs{shoveleft} and \cs{shoveright} as shown in the example below.
+% \begin{verbatim}
+% \begin{empheq}{multline}
+% \framebox[.65\columnwidth]{First line} \\
+% \framebox[.5\columnwidth]{Second line} \\
+% \shoveleft{L+E+F+T} \\
+% \shoveright{R+I+G+H+T} \\
+% \shoveleft[1cm]{L+E+F+T} \\
+% \shoveright[\widthof{$R+I+G+H+T$}]{R+I+G+H+T} \\
+% \framebox[.65\columnwidth]{Last line}
+% \end{empheq}
+% \end{verbatim}
+% \begin{empheq}{multline}
+% \framebox[.65\columnwidth]{First line} \\
+% \framebox[.5\columnwidth]{Second line} \\
+% \shoveleft{L+E+F+T} \\
+% \shoveright{R+I+G+H+T} \\
+% \shoveleft[1cm]{L+E+F+T} \\
+% \shoveright[\widthof{$R+I+G+H+T$}]{R+I+G+H+T} \\
+% \framebox[.65\columnwidth]{Last line}
+% \end{empheq}
+%
+%
+% There are however a few differences in the output between the
+% original \pkg{amsmath} version of \env{multline} and the one
+% \pkg{empheq} provides. In \pkg{amsmath} a centered line in
+% \env{multline} is centered on the page without taking into account
+% the \cs{multlinegap}, \cs{multlinetaggap}, and the tag width. Thus
+% \pkg{amsmath} can sometimes give you horrible output without
+% giving you any warning as shown below.
+% \begin{verbatim}
+% \begin{multline}
+% \framebox[.65\columnwidth]{First line} \\
+% \framebox[.9\columnwidth]{Loooong line} \\
+% \framebox[.65\columnwidth]{Last line} \tag{wide tag}
+% \end{multline}
+% \end{verbatim}
+% \begin{AmSmultline}
+% \framebox[.65\columnwidth]{First line} \\
+% \framebox[.9\columnwidth]{Loooong line} \\
+% \framebox[.65\columnwidth]{Last line} \tag{wide tag}
+% \end{AmSmultline}
+% In \pkg{empheq} these parameters are taken into account, so the
+% same input will produce this in \pkg{empheq}:
+% \begin{verbatim}
+% \begin{empheq}{multline}
+% \framebox[.65\columnwidth]{First line} \\
+% \framebox[.9\columnwidth]{Loooong line} \\
+% \framebox[.65\columnwidth]{Last line} \tag{wide tag}
+% \end{empheq}
+% \end{verbatim}
+% \begin{multline*}
+% \framebox[.65\columnwidth]{First line} \\
+% \framebox[.9\columnwidth]{Loooong line} \\
+% \framebox[.65\columnwidth]{Last line} \tag{wide tag}
+% \end{multline*}
+% This results in an
+% \begin{verbatim}
+% Overfull \hbox (30.03783pt too wide) in paragraph ...
+% \end{verbatim}
+% message in the log file, indicating a visual problem. I think this
+% behavior is more sensible than the one the original \env{multline}
+% environment provides.
+%
+% \begin{codesyntax}
+% \SpecialEnvIndex{MTmultlined}\cs{begin}\arg{MTmultlined}\oarg{pos}\oarg{width}
+% \meta{contents} \cs{end}\arg{MTmultlined}
+% \end{codesyntax}
+% When you choose the \opt{oldmultline} option you may still want to
+% use the much improved \env{multlined} environment defined in
+% \pkg{mathtools} but unfortunately there will be a name clash.
+% Instead you can access it under the name \env{MTmultlined}.
+%
+%
+%
+% \subsubsection{The \opt{overload} option}\label{sec:overload}
+%
+% \begin{codesyntax}
+% \SpecialOptIndex{overload}\opt{overload}\\
+% \SpecialOptIndex{overload2}\opt{overload2}
+% \end{codesyntax}
+% \FeatureRequest{Lars Madsen}{2004/03/25}
+% The \opt{overload} option redefines the original \AmS{}
+% environments so that they take an optional argument.
+% \begin{codesyntax}
+% \cs{begin}\marg{\AmS{} env_name}\oarg{markup instructions}\\
+% \meta{contents of \AmS{} environment}\\
+% \cs{end}\marg{\AmS{} env_name}
+% \end{codesyntax}
+% For example
+% \begin{verbatim}
+% \begin{gather}[box=\widefbox]
+% a = b
+% \end{gather}
+% \end{verbatim}
+% is then actually short for
+% \begin{verbatim}
+% \begin{empheq}[box=\widefbox]{gather}
+% a = b
+% \end{empheq}
+% \end{verbatim}
+% All the \AmS{} environments are supported by this option except
+% for the pseudo environment |\[ ... \]| (it's not really an
+% environment), \footnote{With the option \opt{oldmultline} the
+% \env{multline} and \env{multline*} environments aren't supported
+% either because their syntax in \pkg{empheq} then differ from their
+% syntax in \pkg{amsmath}.} because a)~you don't really need markup
+% instructions for a one line display, and b)~would you really like
+% a syntax like
+% \begin{verbatim}
+% \[[box=\fbox] a=b \]
+% \end{verbatim}
+% where its difficult to see whether or not there is a typo?
+% Choosing \opt{overload} is meant for people who don't want to
+% change their entire document into \pkg{empheq} syntax. I have no
+% problems with that; just be careful when you fiddle with
+% \cs{empheqset} as it will affect \emph{all} math displays!
+%
+%
+% There is of course a catch (well, two actually): These redefined
+% environments don't run as fast as the regular ones (about three
+% times as slow), but in this day and age I seriously doubt you'll
+% be able to tell the difference unless you have a vast number of
+% equations. The other catch is that you cannot use \cs{intertext}
+% and \cs{displaybreak} as described earlier. If you find yourself
+% wanting to use one of these features in say, an \env{align}
+% environment, you have to use the original \env{align} environment.
+% Luckily it is available if you call it like this:
+% \begin{codesyntax}
+% \cs{begin}\arg{AmS\meta{\rmfamily \AmS{} env_name}} \\
+% \texttt{~~}\meta{contents of \AmS{} environment}\\
+% \cs{end}\arg{AmS\meta{\rmfamily \AmS{} env_name}}
+% \end{codesyntax}
+% For example the original \env{align} environment could be selected
+% with
+% \begin{verbatim}
+% \begin{AmSalign}
+% ...
+% \end{AmSalign}
+% \end{verbatim}
+% These original versions with prefix \texttt{AmS} exist for all the
+% environments.
+%
+%
+% The option \opt{overload2} activates the overloading feature
+% for the pseudo environment |\[ ... \]|, although I doubt you'll
+% find it useful. Beware that this definition is fragile unless you
+% have \eTeX{} as \LaTeX{} engine.\footnote{\eTeX{} has been the
+% default engine for \LaTeX{} in most major distributions since
+% 2003.} Not surprisingly \opt{overload2} activates
+% \opt{overload}.
+%
+%
+%
+% Before you get all excited about these options, you should take
+% note of some aspects of centered math displays in \pkg{amsmath}.
+% In certain circumstances truly centering the display is not always
+% the best solution as in this example:
+% \begin{center}
+% \mbox{}\hfill
+% \framebox[.75\linewidth]{Wide math display; not adjusted}%
+% \hfill\mbox{}%
+% \makebox[0pt][r]{Wide Tag}
+% \end{center}
+% Instead centering the display in the available space works pretty
+% well:
+% \begin{gather*}
+% \framebox[.75\linewidth]{Wide math display; adjusted}%
+% \tag*{Wide Tag}
+% \end{gather*}
+% This is what \pkg{amsmath} does normally, but in the \env{gather}
+% environment it does it on a per line basis. This means that we get
+% results like
+% \begin{AmSgather*}
+% \framebox[.75\linewidth]{Wide math display; adjusted}%
+% \tag*{Wide Tag} \\
+% \framebox[.75\linewidth]{Wide math display; not adjusted} \tag{i}%
+% \end{AmSgather*}
+% whether we like it or not. Inside the \env{empheq} environment we
+% need to have the same adjustment for all lines else the boxing
+% process will not work properly, so when activating the
+% \opt{overload} option the above two-line \env{gather} will
+% instead look like this:
+% \begin{empheq}{gather*}
+% \framebox[.75\linewidth]{Wide math display; adjusted}%
+% \tag*{Wide Tag} \\
+% \framebox[.75\linewidth]{Wide math display; adjusted} \tag*{i}%
+% \end{empheq}
+% I leave it to you to choose whether or not this is better (I think
+% it is better).
+%
+%
+%
+% \subsection{A note on boxed displays}
+%
+% When browsing a 400+ pages textbook with at least twice the number
+% of displayed formulae, some of them are surely more important than
+% others. Thus the author of the book (in cooperation with the
+% designer) should make sure that such formulae are easily found
+% again i.e., they should be easily distinguishable from the rest of
+% the pack. One way of doing this is putting the formula into a box
+% which is something we've seen \pkg{empheq} being capable of. There
+% are however a few things to keep in mind:
+% \begin{itemize}
+% \item Don't overdo it. If you do it on half of them there's no
+% point in doing it at all. I don't see much reason for applying
+% this technique to more than 10\% of the formulae.
+% \item Choose the type of box carefully. You want to draw
+% attention to it so it might as well look good.
+% \end{itemize}
+% The latter point can be illustrated by defining a macro similar to
+% \cs{fbox} only with a little more space to the right and left of
+% the argument.
+% \begin{verbatim}
+% \newcommand*\widefbox[1]{\fbox{\hspace{1em}#1\hspace{1em}}}
+% \end{verbatim}
+% If we replace the \cs{fbox} from the \key{box} example before, we
+% get this display:
+% \begin{verbatim}
+% \begin{empheq}[box=\widefbox]{align}
+% a&=b\\
+% E&=mc^2 + \int_a^a x\, dx
+% \end{empheq}
+% \end{verbatim}
+% \begin{empheq}[box=\widefbox]{align}
+% a&=b\\
+% E&=mc^2 + \int_a^a x\, dx
+% \end{empheq}
+% Compare it with
+% \begin{empheq}[box=\fbox]{align}
+% a&=b\\
+% E&=mc^2 + \int_a^a x\, dx
+% \end{empheq}
+% and see which one you prefer.
+%
+% Similarly one might be tempted to use colored boxes:
+% \begin{verbatim}
+% \definecolor{myblue}{rgb}{.8, .8, 1}
+% \newcommand*\mybluebox[1]{%
+% \colorbox{myblue}{\hspace{1em}#1\hspace{1em}}}
+% \end{verbatim}
+% We know the drill by now.
+% \begin{verbatim}
+% \begin{empheq}[box=\mybluebox]{align}
+% a&=b\\
+% E&=mc^2 + \int_a^a x\, dx
+% \end{empheq}
+% \end{verbatim}
+% \begin{empheq}[box=\mybluebox]{align}
+% a&=b\\
+% E&=mc^2 + \int_a^a x\, dx
+% \end{empheq}
+% There is more on boxes later in this manual. If you're into the
+% \pkg{fancybox} package then remember to read \S
+% \vref{sec:fancybox_support}.
+%
+% \section{Support for other packages}
+%
+% With the multitude of packages for \LaTeX, it is not always easy
+% to be a PWWO package (``Plays Well With Others''), but
+% \pkg{empheq} tries really hard to do so. This section lists the
+% packages where \pkg{empheq} has to provide workarounds and they
+% can be divided into two categories.
+% \begin{description}
+% \item[Compatibility] Some packages affect \LaTeX' labelling
+% mechanism and since \pkg{empheq} internally has to turn off
+% labels and related code temporarily, hooks must be provided for
+% these packages. Examples of such packages are \pkg{hyperref} and
+% \pkg{showkeys}.
+%
+% \item[Enhancements] Other packages provide useful features that
+% for some reason may not work directly or optimally with
+% \pkg{empheq}. In these cases the problematic commands are
+% redefined so that they not only work with \pkg{empheq}, but
+% also give the same if not better output. An example of this
+% would be the \cs{shadowbox} command from \pkg{fancybox}
+% (described below).
+% \end{description}
+% %
+% \subsection{Support for \pkg{fancybox}}\label{sec:fancybox_support}
+%
+% The \pkg{fancybox} package provides various boxes and you can use
+% them with \pkg{empheq} as well. Here's \cs{ovalbox}:
+% \begin{verbatim}
+% \begin{empheq}[box=\ovalbox]{align}
+% E&=mc^2 \\
+% Y&= \sum_{n=1}^\infty \frac{1}{n^2}
+% \end{empheq}
+% \end{verbatim}
+% \begin{empheq}[box=\ovalbox]{align}
+% E&=mc^2 \\
+% Y&= \sum_{n=1}^\infty \frac{1}{n^2}
+% \end{empheq}
+%
+%
+% \begin{codesyntax}
+% \SpecialUsageIndex{\shadowbox}\SpecialUsageIndex{\shadowbox*}
+% \cs{shadowbox*}\marg{arg} \\
+% \SpecialColorIndex{shadowcolor} \texttt{shadowcolor}
+% \end{codesyntax}
+% The only problem with using the \pkg{fancybox} boxes in
+% conjunction with the \env{empheq} environment is \cs{shadowbox},
+% as this macro typesets the shadow on the baseline and for the
+% equation numbers to be placed correctly, the box command must set
+% its argument on the baseline. Thus the normal \cs{shadowbox}
+% produces \shadowbox{\meta{arg}}, but we need this instead:
+% \shadowbox*{\meta{arg}}. Therefore \pkg{empheq} will detect if
+% \pkg{fancybox} is loaded and in that case it'll enhance
+% \cs{shadowbox} in two ways:
+% \begin{itemize}
+% \item It defines a starred version \cs{shadowbox*} which
+% typesets its argument on the baseline.
+%
+% \item The color \texttt{shadowcolor} is introduced. The default
+% color is black.
+% \end{itemize}
+% \begin{verbatim}
+% \definecolor{shadowcolor}{rgb}{0,.5,.5}
+% \setlength\shadowsize{2pt}
+% Line of text for \shadowbox*{testing of \Verb|f#&%\| verbatim} and
+% showing the shadow color.
+% \end{verbatim}
+% \begin{quote}
+% \definecolor{shadowcolor}{rgb}{0,.5,.5}
+% \setlength\shadowsize{2pt}
+% Line of text for \shadowbox*{testing of \Verb|f#&%\| verbatim} and
+% showing the shadow color.
+% \end{quote}
+%
+% The point is that if you want a \cs{shadowbox} around your math
+% display then you must use the starred version:
+% \begin{verbatim}
+% \begin{empheq}[box=\shadowbox*]{align}
+% E&=mc^2 \\
+% Y&= \sum_{n=1}^\infty \frac{1}{n^2}
+% \end{empheq}
+% \end{verbatim}
+% \begin{empheq}[box=\shadowbox*]{align}
+% E&=mc^2 \\
+% Y&= \sum_{n=1}^\infty \frac{1}{n^2}
+% \end{empheq}
+%
+% See \S \vref{stunning} if you want
+% to make a fancy box yourself.
+%
+%
+% \subsection{Support for \pkg{ntheorem}}\label{sec:ntheorem-support}
+%
+% \changes{v2.13}{2012/05/09}{Added caveat}
+% \textbf{Caveat:} Due to an unfortunate interaction between \pkg{empheq} and
+% \pkg{ntheorem}, users may want to add
+% \begin{verbatim}
+% \usetagform{default}
+% \end{verbatim}
+% \emph{after} loading \pkg{ntheorem} otherwise the tags may be
+% placed wrong and any labels within the \env{empheq} environment may
+% be lost. In the future \pkg{ntheorem} should be able to test for
+% this and add \verb|\usetagform{default}| automatically.
+%
+% Users who use \verb|\usetagform| to get another tag design, should
+% make sure to postpone this configuration until \emph{after} loading
+% \pkg{ntheorem}.
+%
+% \bigskip
+%
+% \changes{v2.00}{2004/04/14}{Added description of the
+% \opt{ntheorem} option}
+% \changes{v1.05}{2004/02/16}{Added support for the end-of-theorem
+% marker mechanism of \pkg{ntheorem}}
+%
+% \begin{codesyntax}
+% \SpecialOptIndex{ntheorem}\opt{ntheorem}
+% \end{codesyntax}
+% The \pkg{ntheorem} package is supported by the means of the
+% \opt{ntheorem} option of \pkg{empheq}. This loads a set of
+% extra macros which fixes various compatibility problems between
+% \pkg{ntheorem} and \pkg{amsmath} and furthermore introduces
+% special (internal) macros for optimum positioning of
+% end-of-theorem markers, while retaining a user friendly interface.
+%
+%
+% When you want to use the automatic end-of-theorem marker mechanism
+% from \pkg{ntheorem} you sometimes run into problems as you would
+% want the marker to be placed aligned at the bottom of the math
+% display but still keeping the tags in their proper place. In
+% \mode{leqno} mode this is not that much of a problem as the tags
+% and the marker are set on either side of the math display like
+% in Figure \vref{fig:markerI}.
+% \changes{v1.05a}{2004/02/17}{Added explanatory drawings for
+% \pkg{ntheorem}}
+%
+% \begin{figure}[btp]\centering
+% \unitlength2.5pt \thicklines
+% \begin{picture}(90,30)(-30,40)
+% \put(0,50){\framebox(25,20){Math Height}}
+% \put(0,40){\framebox(25,10){Math Depth}}
+% \put(-30,50){\framebox(20,10){Tag height}}
+% \put(-30,45){\framebox(20,5){Tag depth}}
+% \put(40,40){\framebox(20,10){Marker}}
+% \multiput(26,40)(2,0){7}{\line(1,0){1}}
+% \end{picture}
+% \caption{Marker placement in \mode{leqno} mode}
+% \label{fig:markerI}
+% \end{figure}
+%
+% \begin{codesyntax}
+% \SpecialUsageIndex{\mintagvsep}\cs{mintagvsep}
+% \end{codesyntax}
+% Unfortunately things are not this easy in \mode{reqno} mode. There
+% are two possible situations as shown in Figure \vref{fig:markerII}.
+% \begin{figure}[btp]
+% \unitlength2.5pt \thicklines
+% \begin{picture}(60,45)(0,25)
+% \put(0,50){\framebox(25,20){Math Height}}
+% \put(0,25){\framebox(25,25){Math Depth}}
+% \put(40,50){\framebox(20,10){Tag height}}
+% \put(40,45){\framebox(20,5){Tag depth}}
+% \put(40,25){\framebox(20,10){Marker}}
+% \put(50,35){\vector(0,1){10}}
+% \put(50,45){\vector(0,-1){10}}
+% \put(50,39){\ \ $>\delta$}
+% \multiput(26,25)(2,0){7}{\line(1,0){1}}
+% \end{picture} \hfill\vrule\hfill
+% \begin{picture}(60,45)(0,25)
+% \put(0,50){\framebox(25,20){Math Height}}
+% \put(0,40){\framebox(25,10){Math Depth}}
+% \put(40,50){\framebox(20,10){Tag height}}
+% \put(40,45){\framebox(20,5){Tag depth}}
+% \put(40,28){\framebox(20,10){Marker}}
+% \put(50,38){\vector(0,1){7}}
+% \put(50,45){\vector(0,-1){7}}
+% \put(50,40){\ \ $=\delta$}
+% \end{picture}
+% \caption{The two possible situations for end marks in \mode{reqno}
+% mode}
+% \label{fig:markerII}
+% \end{figure}
+% If possible we want the bottom of the marker to be aligned with
+% the bottom of the math display, but at the same time we want to
+% ensure a minimum vertical separation $\delta$ between marker and
+% tag. In \pkg{empheq} this is controlled by the length parameter
+% \cs{mintagvsep} which by default is 5\,pt.
+%
+% The good news is that this is where \pkg{empheq} sets in.
+% Basically all you need to do is to use the \env{empheq}
+% environment to type set your equations inside the theorem
+% environments:
+% \begin{verbatim}
+% \begin{Theorem}
+% Some text at first and then a math display
+% \begin{empheq}{align}
+% a&=b\\
+% E&=mc^2 + \int_a^a x\, dx
+% \end{empheq}
+% \end{Theorem}
+% \end{verbatim}
+% Then the tag placement and the end-of-theorem marker will be set
+% properly (after a couple of runs as usual).
+%
+% To ensure the correct outcome remember to load the packages like
+% this:
+% \begin{verbatim}
+% \usepackage[ntheorem]{empheq} % this loads amsmath as well
+% \usepackage[thmmarks,amsmath]{ntheorem}
+% \end{verbatim}
+% Remember that if you use the \opt{overload} option you
+% can just use the regular math environments to get the desired
+% result with the end mark. The exception is the |\[ ... \]|
+% environment which will only work if you use the option
+% \opt{overload2}.
+%
+% \clearpage
+%
+% \EmphEqpart{Advanced user's guide}
+%
+% The \pkg{empheq} package has more to offer than you have seen, but
+% I found some of the functionality so dangerous it was best to hide
+% it for a little while. The commands you have encountered so far in
+% this manual all have one thing in common: They have only lower
+% case letters in their names. Now it's time to reveal those that
+% have \emph{mixed-case} names, thus implying that you should take
+% great care when using them.
+%
+% \section{Delimiters revisited}
+% Let's go back to delimiters, shall we?
+%
+% \subsection{Creating your own delimiters}
+% As a convenience for the user, this interface is extended in a
+% more general way so that it is also possible to declare delimiters
+% with the following two commands: \SpecialUsageIndex{\DeclareLeftDelimiter}
+% \SpecialUsageIndex{\DeclareRightDelimiter}
+% \begin{codesyntax}
+% \cs{DeclareLeftDelimiter}\oarg{space adjustment}\marg{delimiter}\\
+% \cs{DeclareRightDelimiter}\oarg{space adjustment}\marg{delimiter}
+% \end{codesyntax}
+% While \pkg{empheq} provides auto-scaling versions of the most
+% common delimiters, you may sometimes want some new ones. Say for
+% instance you have loaded the \pkg{stmaryrd} package and you want
+% to use the double bracket commands \cs{llbracket} and
+% \cs{rrbracket} with \pkg{empheq}. Then you simply do this:
+% \begin{verbatim}
+% \DeclareLeftDelimiter{\llbracket}
+% \DeclareRightDelimiter{\rrbracket}
+% \end{verbatim}
+% This defines the new delimiters \cs{empheqllbracket},
+% \cs{empheqbigllbracket}, \cs{empheqrrbracket} and
+% \cs{empheqbigrrbracket}.
+%
+% You can use \cs{big...} delimiters as well if you don't like the
+% automatic scaling. There are however ways to fine-tune if you
+% really want it.
+%
+% \subsection{Fine-tuning of delimiters}
+%
+% \TeX{} provides two primitives to control the scaling of
+% delimiters produced with \cs{left} and \cs{right}, namely the
+% dimension \cs{delimitershortfall} (denoted by $\delta$ here) and
+% the integer \cs{delimiterfactor} ($f$). The idea is that the
+% sub-formula inside the \cs{left}-\cs{right} pair is to be
+% vertically centered and given its height $h_1$ and its depth $h_2$
+% we want to produce a delimiter with total height $h$, where $h =
+% 2 \max(h_1,h_2)$. \TeX' rules on this is that the minimum
+% delimiter size $h_{\min}$ must meet the requirements
+% \[
+% h_{\min}\geq h \frac{f}{1000} \quad\land \quad h_{\min}\geq h-\delta
+% \]
+% \LaTeX{} sets $\cs{delimitershortfall} = \text{\the\delimitershortfall} $
+% and $\cs{delimiterfactor} = \the\delimiterfactor$, but in our case
+% we will almost always want a delimiter that spans the entire
+% \cs{left}-\cs{right} pair (the math display), thus a change of
+% these settings is needed.
+%
+% \begin{codesyntax}
+% \SpecialUsageIndex{\EmphEqdelimitershortfall}
+% \cs{EmphEqdelimitershortfall}\\
+% \SpecialUsageIndex{\EmphEqdelimiterfactor}
+% \cs{EmphEqdelimiterfactor}
+% \end{codesyntax}
+% However it is a bad idea to just change these two settings without
+% thinking of the effect it'll have on the rest of the mathematics
+% in your document. Therefore \pkg{empheq} provides the parameters
+% \cs{EmphEqdelimitershortfall} (default setting is
+% \the\EmphEqdelimitershortfall) and \cs{EmphEqdelimiterfactor}
+% (default is \the\EmphEqdelimiterfactor) to cater for this.
+%
+%
+% \subsection{Scaling material yourself}
+%
+% The attentive reader may have noticed that I still haven't
+% revealed how I managed to get the right size of that big
+% exclamation mark in the first example of this manual, so I guess
+% it's about time.
+% \begin{codesyntax}
+% \SpecialUsageIndex{\EmphEqdisplayheight}\cs{EmphEqdisplayheight} \\
+% \SpecialUsageIndex{\EmphEqdisplaydepth}\cs{EmphEqdisplaydepth}
+% \end{codesyntax}
+% The height of the math display plus the surrounding inner box is
+% given by \cs{EmphEqdisplayheight} and the depth by
+% \cs{EmphEqdisplaydepth}. These can be used when constructing
+% something that should scale to fit the display and can't be done
+% with ``auto-scaling'' commands like \cs{vrule} or \cs{left} or
+% \cs{right}. There is a catch however: If you do any horizontal
+% resizing of the material you want to scale, then you must specify
+% the width \emph{manually}:\footnote{The
+% \cs{fontdimen22}\cs{textfont2} in the \meta{length} argument of
+% \cs{raisebox} is the \emph{math axis} of the font. It is needed in
+% order to get the vertical positioning of the oversized exclamation
+% mark just right.}
+% \begin{verbatim}
+% \begin{empheq}[
+% box=\colorbox{myblue},
+% right={\;
+% \makebox[.9em]{%
+% $\raisebox{-.5\totalheight+\fontdimen22\textfont2}
+% {\resizebox{!}{%
+% \EmphEqdisplayheight+\EmphEqdisplaydepth}{!}}$%
+% }},
+% left={X=Y\Rightarrow}
+% ]{alignat=3}
+% A_1&=b_1 & \qquad c_1&=d_1 & \qquad e_1&= f_1
+% \tag*{\_A\raisebox{1ex}{silly}\raisebox{-1ex}{tag}\_}\\
+% A_2&=b_2 & \qquad c_2&=d_2 & \qquad e_2&= f_2 \\
+% A_3&=b_3 & \qquad c_3&=d_3 & \qquad e_3&= f_3
+% \end{empheq}
+% \end{verbatim}
+%
+% Here's another example. We want to be able to put a \cs{parbox}
+% with some descriptive text top-aligned on the side of the display.
+% \begin{verbatim}
+% \begin{empheq}[
+% left={\parbox[c][\EmphEqdisplayheight+\EmphEqdisplaydepth][t]
+% {4.5cm}
+% {You may find this kind of description useful.}\enspace}]
+% {align}
+% a&=\int_0^1 x\,dx +\frac{foo + bar}{baz}\\
+% E&= mc^2
+% \end{empheq}
+% \end{verbatim}
+% \begin{empheq}[
+% left={\parbox[c][\EmphEqdisplayheight+\EmphEqdisplaydepth][t]
+% {4.5cm}
+% {You may find this kind of description useful.}\enspace}]
+% {align}
+% a&=\int_0^1 x\,dx +\frac{foo + bar}{baz}\\
+% E&= mc^2
+% \end{empheq}
+%
+%
+% \section{A few short notes}
+%
+% \subsection{About \texorpdfstring{$\cs{eqref}$}{\textbackslash eqref}}
+%
+% Internally \env{empheq} separate the displayed math from
+% the corresponding equation numbers such that we can add special
+% delimiters or boxes. This is done by nullifying (or rather
+% conveniently redefining) an internal command called
+% \cs{maketag@@@}. Unfortunately, this component is also used by
+% \cs{eqref} to typeset a reference to an equation number, so in
+% earlier versions, \cs{eqref} would not work inside an \env{empheq}
+% environment.
+%
+% In the current version this should now be working, but if you ever
+% need to mess with \cs{eqref} it may be handy to know what is being
+% done. In essence we do something similar to this:
+% \begin{verbatim}
+% \let\empheqeqrefbase\textup
+% ... % next go inside empheq env
+% \renewcommand\eqref[1]{\empheqeqrefbase{%
+% \let\maketag@@@\EQsavedmaketag%
+% \tagform@{\ref{##1}}}}
+% \end{verbatim}
+% Thus if you need to alter things inside \cs{eqref} and need that to
+% work within \env{empheq} as well, you may get away with redefining
+% \cs{empheqeqrefbase}.
+%
+%
+%
+% \subsection{About changes to
+% \texorpdfstring{$\cs{baselineskip}$}{\textbackslash baselineskip}}
+%
+% Users should never mess with \cs{baselineskip} directly, it is not
+% the correct manner to get double spacing. Have a look at say the
+% \pkg{setspace} package or similar, or play with
+% \cs{baselinestretch} followed by \cs{normalsize} to initiate.
+%
+% Explanation: We use \cs{parbox} inside to place the display box and
+% the box containing the eqn numbers. To do its stuff \cs{parbox}
+% resets some internal settings, in this case, \cs{baselineskip} is
+% reset to \cs{normalbaselineskip}. So if you don't remember to reset
+% that one as well\dots
+%
+%
+% \section{Creating something new}
+%
+% You can create your own environments and boxes to go with
+% \pkg{empheq}, but there are certain things that must be fulfilled
+% to get it to work properly.
+%
+% \subsection{New \env{empheq}-like environments}
+%
+% \begin{codesyntax}
+% \SpecialEnvIndex{EmphEqMainEnv}\env{EmphEqMainEnv}\\
+% \SpecialUsageIndex{\EmphEqMainEnv}\cs{EmphEqMainEnv}\\
+% \SpecialUsageIndex{\endEmphEqMainEnv}\cs{endEmphEqMainEnv}
+% \end{codesyntax}
+% Although the real work in \pkg{empheq} is done by an environment
+% (\env{EmphEqMainEnv}), you cannot use it like that yourself when
+% you want to define an \env{empheq}-like environment. Due to
+% technical reasons (see~\cite{Ams99}), you have to use a slightly
+% different syntax:
+% \begin{codesyntax}
+% \cs{newenvironment}
+% \marg{env_name} \oarg{num} \oarg{default}\\
+% \texttt{~~}\arg{\meta{other commands} \cs{EmpheqMainEnv}}\\
+% \texttt{~~}\arg{\cs{endEmpheqMainEnv}}
+% \end{codesyntax}
+% In the above environment the \meta{other commands} part must
+% contain certain things before it'll work.
+% \begin{codesyntax}
+% \SpecialKeyFamilyIndex{EmphEqEnv}\keyfam{EmphEqEnv}\\
+% \SpecialKeyFamilyIndex{EmphEqOpt}\keyfam{EmphEqOpt}
+% \end{codesyntax}
+% Before your homemade environment will work you must set at least
+% one key. The key family \keyfam{EmphEqOpt} is for all the markup
+% instructions like \key{left} and \key{box}, while
+% \keyfam{EmphEqEnv} controls the type of \AmS{} environment. You
+% must set the \keyfam{EmphEqEnv} family before \cs{EmhpEqMainEnv}
+% else you'll get an error message.
+%
+% One example could be something like this:
+% \begin{verbatim}
+% \newenvironment{important}[2][]{%
+% \setkeys{EmphEqEnv}{#2}%
+% \setkeys{EmphEqOpt}{box=\mybluebox,#1}%
+% \EmphEqMainEnv}%
+% {\endEmphEqMainEnv}
+% \end{verbatim}
+% Thus
+% \begin{verbatim}
+% \begin{important}{gather}
+% a = b + c +d \\ e = f
+% \end{important}
+% \end{verbatim}
+% produces
+% \begin{important}{gather}
+% a = b + c +d \\ e = f
+% \end{important}
+% while
+% \begin{verbatim}
+% \begin{important}[left={A=B \Rightarrow \empheqlbrace}]{alignat=2}
+% a &= b &\quad c &= d \\
+% \text{this} &= \text{that} &\quad \mathit{fish}&\neq fish
+% \end{important}
+% \end{verbatim}
+% produces
+% \begin{important}[left={A=B \Rightarrow \empheqlbrace}]{alignat=2}
+% a &= b &\quad c &= d \\
+% \text{this} &= \text{that} &\quad \mathit{fish}&\neq fish
+% \end{important}
+%
+%
+% \subsection{Creating fancy boxes}\label{stunning}
+%
+% As a final example I will show you how to create complicated
+% displays involving (vertically) asymmetrical boxes like
+% \cs{shadowbox*}. In order to get the correct output, the contents
+% of the box must be placed on the baseline as in \fbox{\cs{fbox}}.
+% In this example we want to put a set of equations into a bright
+% yellow box and then add another box with some explanatory text at
+% the top of the yellow box making them overlap.
+%
+% First we define the colors used and allocate the boxes. We could
+% probably use scratch boxes, but this is safer.
+% \begin{verbatim}
+% \definecolor{shadecolor}{cmyk}{0,0,0.41,0}
+% \definecolor{light-blue}{cmyk}{0.25,0,0,0}
+% \newsavebox{\mysaveboxM} % M for math
+% \newsavebox{\mysaveboxT} % T for text
+% \end{verbatim}
+% Save the display body in \cs{mysaveboxM} and the text argument in
+% \cs{mysaveboxT}.
+% \begin{verbatim}
+% \newcommand*\Garybox[2][Example]{%
+% \sbox{\mysaveboxM}{#2}%
+% \sbox{\mysaveboxT}{\fcolorbox{black}{light-blue}{#1}}%
+% \end{verbatim}
+% Then typeset the math display in a \cs{parbox} where we control
+% the height and save it in \cs{mysaveboxM}.
+% \begin{verbatim}
+% \sbox{\mysaveboxM}{%
+% \parbox[b][\ht\mysaveboxM+.5\ht\mysaveboxT+.5\dp\mysaveboxT][b]{%
+% \wd\mysaveboxM}{#2}%
+% }%
+% \end{verbatim}
+% We put it into the colored box with the desired width.
+% \begin{verbatim}
+% \sbox{\mysaveboxM}{%
+% \fcolorbox{black}{shadecolor}{%
+% \makebox[\linewidth-10em]{\usebox{\mysaveboxM}}%
+% }%
+% }%
+% \end{verbatim}
+% Then finally we get to the real typesetting. We just insert the
+% math display and then make a box of zero width right next to it.
+% In that box we lift the text argument and center it at the top of
+% the display.
+% \begin{verbatim}
+% \usebox{\mysaveboxM}%
+% \makebox[0pt][r]{%
+% \makebox[\wd\mysaveboxM][c]{%
+% \raisebox{\ht\mysaveboxM-0.5\ht\mysaveboxT
+% +0.5\dp\mysaveboxT-0.5\fboxrule}{\usebox{\mysaveboxT}}%
+% }%
+% }%
+% }
+% \end{verbatim}
+% Let's see what it looks like.
+% \begin{verbatim}
+% \begin{empheq}[box=\Garybox]{align}
+% \sum \mathbf{F} &= \mathbf{0} \\
+% \sum F_{x} \,\mathbf{i} + \sum F_{y} \,\mathbf{j}
+% + \sum F_{z} \,\mathbf{k} &= \mathbf{0} \\
+% \sum F_{x} &= 0 \\ \sum F_{y} &= 0 \\ \sum F_{z} &= 0
+% \end{empheq}
+% \end{verbatim}
+% \begin{subequations}
+% \begin{empheq}[box=\Garybox]{align}
+% \sum \mathbf{F} &= \mathbf{0} \\
+% \sum F_{x} \,\mathbf{i} + \sum F_{y} \,\mathbf{j} + \sum F_{z}
+% \,\mathbf{k} &= \mathbf{0}\\
+% \sum F_{x} &= 0 \\ \sum F_{y} &= 0 \\ \sum F_{z} &= 0
+% \end{empheq}
+% If we use the optional argument of \cs{Garybox} we have to enclose
+% the entire argument of \key{box} in braces because we're in
+% \env{empheq}'s optional argument (as described in \cite[page
+% 167]{Lamp94}):
+% \begin{verbatim}
+% \begin{empheq}[box={\Garybox[Same old example again]}]{align}
+% a&=b\\
+% E&=mc^2 + \int_a^a x\, dx
+% \end{empheq}
+% \end{verbatim}
+% \begin{empheq}[box={\Garybox[Same old example again]}]{align}
+% a&=b\\
+% E&=mc^2 + \int_a^a x\, dx
+% \end{empheq}
+% \end{subequations}
+%
+%
+%
+%
+% \section{Contact information}
+%
+% Should you have any feature request, suggestions, or bug reports
+% then feel free to contact Lars Madsen at
+% \begin{quote}
+% \email{daleif@math.au.dk}
+% \end{quote}
+% Please put `empheq' in the subject line.
+%
+% \subsection*{Contributors}
+%
+% \begin{itemize}
+% \item In November 2002 Lars Madsen (\email{daleif@imf.au.dk})
+% asked for some features that wasn't readily available with
+% \pkg{amsmath} or any other package. This was the start of
+% \pkg{empheq}.
+%
+% \item Gary Gray (\email{gray@engr.psu.edu}) gave me some bug
+% reports on the old version (v0.7) which convinced me to rewrite
+% the package completely. The \cs{Garybox} example is inspired by
+% a wish from Gary.
+%
+% \item Uwe Siart (\email{uwe.siart@tum.de}) has been a thorough
+% beta-tester on v1.00 of the package.
+%
+% \item Bernard Alfonsi (\email{alfonsi@math.u-psud.fr}) reported
+% a problem with the \opt{ntheorem} option. It turned out to be a
+% bug in the \pkg{ntheorem} package itself but it's fixed in the
+% option.
+%
+% \item I have received reports on weird behavior from
+% \pkg{empheq} when used with the \pkg{color} package on some
+% dvi-viewers, which is caused by the lack of color support in
+% those dvi-viewers. Both Andrew B.~Collier (\email{colliera@ukzn.ac.za})
+% and Andr\'e M.~de Roos (\email{aroos@science.uva.nl}) have
+% notified me of this. The solution is to either upgrade you
+% dvi-viewer if possible or convert the document in question to
+% PDF or PS.
+% \end{itemize}
+% Thank you all.
+%
+%
+% \begin{thebibliography}{99}
+% \bibitem{Lamp94}
+% Leslie Lamport,
+% \emph{LaTeX User's Guide and Reference Manual}, 2nd edition,
+% Addison-Wesley, Reading, Massachusetts,
+% 1994.
+% \bibitem{Goos-etal94}
+% Michel Goossens, Frank Mittelbach, and Alexander Samarin,
+% \emph{The \LaTeX\ Companion},
+% Addison-Wesley, Reading, Massachusetts,
+% 1994.
+% \bibitem{Knut86}
+% Donald Ervin Knuth,
+% \emph{The \TeX book},
+% Addison-Wesley, Reading, Mas\-sa\-chu\-setts,
+% 1986.
+% \bibitem{Ams99}
+% American Mathematical Society and Michael Downes,
+% \emph{Technical notes on the \pkg{amsmath} package} Version 2.0,
+% 1999/10/29.
+% (Available from CTAN as file \texttt{technote.tex}.)
+% \bibitem{Ams00}
+% Frank Mittelbach, Rainer Sch\"opf, Michael Downes, and David M.~Jones,
+% \emph{The \pkg{amsmath} package} Version 2.13,
+% 2000/07/18.
+% (Available from CTAN as file \texttt{amsmath.dtx}.)
+%
+% \changes{v1.05b}{2004/02/18}{Added \pkg{fancybox} and
+% \pkg{ntheorem} to the \refname}
+% \bibitem{fancybox}
+% Timothy Van Zandt,
+% \emph{The \pkg{fancybox} package} Version 1.3,
+% 2000/09/19
+% (Available from CTAN as file \texttt{fancybox.sty}.)
+%
+% \bibitem{Raht03}
+% Sebastian Rahtz,
+% \emph{Hypertext marks in \LaTeX} Version v6.74m,
+% 2003/11/30.
+% (Available from \url{http://www.tug.org/applications/hyperref} as file
+% \texttt{hyperref.dtx}.)
+%
+% \bibitem{Carl99}
+% David Carlisle,
+% \emph{The \pkg{keyval} Package},
+% Version 1.13, 1999/03/16.
+% (Available from CTAN as file \texttt{keyval.dtx}.)
+%
+% \changes{v1.06}{2004/03/13}{Added \pkg{calc} to the \refname}
+% \bibitem{Thorup:calc:1998}
+% Kresten~Krab Thorup, Frank Jensen, and Chris Rowley.
+% \emph{The {\pkg{calc}} package}, Version 4.1b, 1998/07/07,
+% (Available from CTAN as file \texttt{calc.dtx}.)
+%
+% \bibitem{ntheorem}
+% Wolfgang May and Andreas Schlechte,
+% \emph{The \pkg{ntheorem} package} Version 1.203,
+% 2002/01/07,
+% (Available from CTAN as file \texttt{ntheorem.dtx}.)
+%
+% \changes{v2.10}{2004/07/26}{Added \pkg{showkeys} to the \refname}
+% \bibitem{Carlisle:showkeys:1997}
+% David Carlisle. \emph{The {\pkg{showkeys}} package}, v3.12,
+% 1997/06/12, (Available from CTAN as file \texttt{showkeys.dtx}.)
+%
+% \end{thebibliography}
+%
+% \StopEventually{}
+%
+% \EmphEqpart{Implementation}
+%
+% \section{Requirements, Options, and Strategy}
+% Lets start the package. We'll need the tools from \pkg{mhsetup}.
+% \begin{macrocode}
+%<*package>
+\ProvidesPackage{empheq}%
+ [2017/03/31 v2.15 Emphasizing equations]
+% \end{macrocode}
+% \changes{v2.15}{2017/03/31}{Bumped required mhsetup release}
+% \begin{macrocode}
+\RequirePackage{mhsetup}[2017/03/31]
+\MHInternalSyntaxOn
+% \end{macrocode}
+%
+%
+% \subsection{Requirements, Options, and Allocation}
+%
+% The \opt{overload} option allows the user to type
+% \begin{verbatim}
+% \begin{gather}[box=\fbox]
+% a = b
+% \end{gather}
+% \end{verbatim}
+% when they really mean
+% \begin{verbatim}
+% \begin{empheq}[box=\fbox]{gather}
+% a = b
+% \end{empheq}
+% \end{verbatim}
+% \begin{macrocode}
+\MH_new_boolean:n {overload_ams}
+\DeclareOption{overload}{
+ \MH_set_boolean_T:n {overload_ams}
+}
+\MH_new_boolean:n {overload_amsII}
+\DeclareOption{overload2}{
+ \MH_set_boolean_T:n {overload_ams}
+ \MH_set_boolean_T:n {overload_amsII}
+}
+% \end{macrocode}
+% Old crappy \env{multline} interface or the new dazzling?
+% \begin{macrocode}
+\MH_new_boolean:n {good_mult}
+\DeclareOption{oldmultline}{\MH_set_boolean_F:n {good_mult}}
+\DeclareOption{newmultline}{\MH_set_boolean_T:n {good_mult}}
+% \end{macrocode}
+%
+%
+% Load the \pkg{ntheorem} support file.
+% \begin{macrocode}
+\gdef\g@EQ@ntheorem@support@bool{01}
+\DeclareOption{ntheorem}{\gdef\g@EQ@ntheorem@support@bool{00}}
+% \end{macrocode}
+% Pass other options to \pkg{mathtools}.
+% \begin{macrocode}
+\DeclareOption*{
+ \PassOptionsToPackage{\CurrentOption}{mathtools}
+}
+\ExecuteOptions{newmultline}
+\ProcessOptions\relax
+% \end{macrocode}
+% The name of the game is \pkg{amsmath} so we'll need that of
+% course. Since v2.10 that is now taken care of in the
+% \pkg{mathtools} package which'll pass on the options to
+% \pkg{amsmath}.
+% \changes{v2.15}{2017/03/31}{Bumped required mathtools release}
+% \begin{macrocode}
+\MHInternalSyntaxOff
+\RequirePackage{mathtools}[2017/03/31]
+% \end{macrocode}
+% Activate the internal syntax.
+% \begin{macrocode}
+\MHInternalSyntaxOn
+\AtEndOfPackage{\MHInternalSyntaxOff}
+% \end{macrocode}
+% \changes{v2.00}{2004/04/14}{Changed description of naming
+% conventions. Altered \emph{all} private macros; this is not
+% mentioned at each macro as this would just be too troublesome}
+%
+% \begin{macro}{\l_EQ_equationtype_int}
+% \begin{macro}{\l_EQ_alignmentmode_int}
+% I'll use the following numbering scheme for the equation types,
+% and store the value in \cs{l_EQ_equationtype_int}:
+% \begin{multicols}{2}
+% \begin{enumerate}\setlength{\itemsep}{0pt}
+% \item align
+% \item align*
+% \item gather
+% \item gather*
+% \item flalign
+% \item flalign*
+% \item alignat
+% \item alignat*
+% \end{enumerate}
+% \end{multicols}
+% As you notice \env{equation} and \env{equation*} are missing. This
+% is because \env{gather} and \env{gather*} work the same way, only
+% they are `nicer'. Cuts down on the amount of code as well. I will
+% add an alias so that a user can use the key \key{equation}, but
+% then \key{gather} will kick in instead behind the scenes. Also the
+% \env{multline} environments are missing here, but again it's
+% because I'm faking it. See \S~\vref{sec:flalign+multline} for
+% further details. The value $0$ isn't used for any of the
+% environments and thus serves as a checkpoint (see
+% page~\pageref{ZeroCheckpoint}).
+%
+% Similarly I'll use the counter \cs{l_EQ_alignmentmode_int} to
+% store the standard options for placement of math displays and
+% equation numbering.
+% \begin{multicols}{2}
+% \begin{enumerate}\setlength{\itemsep}{0pt}
+% \item reqno
+% \item leqno
+% \item fleqn, reqno
+% \item fleqn, leqno
+% \end{enumerate}
+% \end{multicols}
+%
+% \changes{v2.15}{2017/03/31}{added MH\_ prefix}
+% \begin{macrocode}
+\newcount\l_EQ_equationtype_int
+\newcount\l_EQ_alignmentmode_int
+\if@fleqn
+ \iftagsleft@
+ \l_EQ_alignmentmode_int=4
+ \MH_else:
+ \l_EQ_alignmentmode_int=\thr@@
+ \MH_fi:
+\MH_else:
+ \iftagsleft@
+ \l_EQ_alignmentmode_int=\tw@
+ \MH_else:
+ \l_EQ_alignmentmode_int=\@ne
+ \MH_fi:
+\MH_fi:
+% \end{macrocode}
+% \end{macro}
+% \end{macro}
+% \begin{macro}{\l_EQ_totalwidth_dim}
+% \begin{macro}{\l_EQ_displaywidth_dim}
+% \begin{macro}{\l_EQ_temp_linewidth_dim}
+% \begin{macro}{\EmphEqdelimitershortfall}
+% \begin{macro}{\EmphEqdelimiterfactor}
+% \begin{macro}{\g_EQ_toptag_height_dim}
+% \begin{macro}{\g_EQ_bottomtag_depth_dim}
+% \begin{macro}{\g_EQ_toprow_height_dim}
+% \begin{macro}{\g_EQ_bottomrow_depth_dim}
+% \begin{macro}{\g_EQ_widesttag_dim}
+% Some lengths we'll need. Could possibly be scratch registers but we
+% have to make global assignments so that road is a little
+% dangerous.
+% \begin{macrocode}
+\newdimen\l_EQ_totalwidth_dim
+\newdimen\l_EQ_displaywidth_dim
+% \end{macrocode}
+% Just to make it accessible.
+% \begin{macrocode}
+\def\EmphEqdisplaywidth{\l_EQ_displaywidth_dim}
+\newdimen\l_EQ_temp_linewidth_dim
+\newdimen\EmphEqdelimitershortfall
+\newcount\EmphEqdelimiterfactor
+\EmphEqdelimiterfactor950
+\EmphEqdelimitershortfall\p@
+\newdimen\g_EQ_toptag_height_dim
+\newdimen\g_EQ_bottomtag_depth_dim
+\newdimen\g_EQ_toprow_height_dim
+\newdimen\g_EQ_bottomrow_depth_dim
+\newdimen\g_EQ_widesttag_dim
+\def\EmphEqtagwidth{\g_EQ_widesttag_dim}
+% \end{macrocode}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+%
+% \begin{macro}{\EQ_outerbox:}
+% \begin{macro}{\EQ_innerbox:}
+% \begin{macro}{\EQ_left_side:}
+% \begin{macro}{\EQ_right_side:}
+% The two boxes and the material on either side of the math display.
+% \begin{macrocode}
+\def\EQ_outerbox:{}
+\def\EQ_innerbox:{}
+\def\EQ_left_side:{}
+\def\EQ_right_side:{}
+% \end{macrocode}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+%
+% \subsection{Strategy}
+%
+% In non-\mode{fleqn} math displays are centered---even when having
+% tags. However this can look rather appalling when either the tag
+% or the math display are very wide like in the following example.
+% \begin{center}
+% \makebox[\linewidth][s]{\mbox{}\hfill \mbox{$\displaystyle
+% a+b+c+d+e+f+g+h+i+j+k+l+m+n+o$} \hfill\mbox{} \llap{(wide tag)}}
+% \end{center}
+% Not exactly pleasing to look at. \pkg{amsmath} takes care to do
+% this adjustment for us, so the user instead obtains the far better
+% \begin{gather*}
+% a+b+c+d+e+f+g+h+i+j+k+l+m+n+o \tag{wide tag}
+% \end{gather*}
+% So \pkg{amsmath} handles this quite well. That's good. It looks
+% like \pkg{amsmath} has an internal threshold that matches the
+% width of the math display against the width of the tags. But what
+% if we have a multiple line display and one of the lines causes an
+% adjustment to kick in while the others don't need it? Then we
+% cannot keep track of the complete display anymore. Furthermore we
+% may want to add material on either side of the math display making
+% the whole thing even more complicated, if not impossible to do
+% without some sort of different approach.
+%
+% Luckily we have such an approach. The idea that comes to mind is
+% to see if one can typeset the labels and the math displays
+% \emph{separately}, store them in separate boxes and then put them
+% back together again afterwards. Is this really possible? Yes.
+% Come, I'll show you.
+%
+% \subsection{The Main Environment}
+% \begin{macro}{\EQ_mathdisplay_box}
+% \begin{macro}{\EmphEqdisplayheight}
+% \begin{macro}{\EmphEqdisplaydepth}
+% \begin{macro}{\EQ_tag_box}
+% Okay, we'll start by allocating two box registers.
+% \begin{macrocode}
+\newsavebox\EQ_mathdisplay_box
+\def\EmphEqdisplayheight{\ht\EQ_mathdisplay_box}
+\def\EmphEqdisplaydepth{\dp\EQ_mathdisplay_box}
+\newsavebox\EQ_tag_box
+% \end{macrocode}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+%
+% \begin{macro}{\EQ_complete_mathdisplay:}
+% \begin{macro}{\EQ_getbody:n}
+% \begin{environment}{EmphEqMainEnv}
+% The \env{EmphEqMainEnv} is (as the name implies) the main
+% environment. I decided to go with the environment form because
+% that's what \pkg{amsmath} users are accustomed to. But the code
+% will require the contents of the environment to be handled as an
+% argument to a command, so we use \cs{collect@body} from
+% \pkg{amsmath} itself to collect it.
+% \begin{macrocode}
+\def\EQ_complete_mathdisplay:{}
+\def\EQ_getbody:n #1{%
+ \def\EQ_complete_mathdisplay:{}\def\EQ_mathbody:{#1}}
+\newenvironment{EmphEqMainEnv}{\collect@body\EQ_getbody:n}{%
+% \end{macrocode}
+% If the user didn't specify the mandatory argument we will catch it
+% because \cs{l_EQ_equationtype_int} is zero. An error is
+% issued\label{ZeroCheckpoint}.
+% \changes{v2.15}{2017/03/31}{added MH\_ prefix, twice}
+% \begin{macrocode}
+ \MH_if_num:w \l_EQ_equationtype_int=\z@ %ZeroCheckpoint
+ \EQ_upgrade_error:
+ \MH_else:
+% \end{macrocode}
+% Those that read the manual get the reward here: The package will
+% actually work!
+% \changes{v2.15}{2017/03/31}{added MH\_ prefix}
+% \begin{macrocode}
+ \EQ_post_process:
+ \ignorespacesafterend
+ \MH_fi:}
+% \end{macrocode}
+% \end{environment}
+% \end{macro}
+% \end{macro}
+% \subsection{Post Processing}
+% \begin{macro}{\EQ_post_process:}
+% I know: it almost looked easy for a while didn't it? Sadly the
+% post-processing step of the code is also by far the most
+% complicated.
+% \begin{macrocode}
+\def\EQ_post_process:{%
+% \end{macrocode}
+% First a hook where we initialize a few things like storing the
+% line width for later manipulation. Also we want to measure the
+% height of the top row and the depth of the bottom row for best
+% possible placement of the math display and the boxes around it
+% (same with the tags). Those four dimensions are globally set
+% during the code, hence this assignment. Most other stuff is local.
+% \begin{macrocode}
+ \EQ_initialize:
+% \end{macrocode}
+% Then we proceed by completing the input. We store it in the
+% control sequence \cs{EQ_complete_mathdisplay}.
+% \cs{EQ_begin_equation:} adds the start of the environment and
+% \cs{EQ_end_equation:} adds the corresponding end.
+% \begin{macrocode}
+ \EQ_begin_equation:
+% \end{macrocode}
+% In case we face a \env{multline} we add what's necessary.
+% \begin{macrocode}
+ \MH_if_boolean:nT {outer_mult}{
+ \setkeys{\MT_options_name:}{
+ firstline-afterskip=0pt,
+ lastline-preskip=0pt,
+ multlined-width=0pt,
+ }
+ \ifodd \l_EQ_alignmentmode_int
+ \setkeys{\MT_options_name:}{multlined-pos=b}
+ \MH_else:
+ \setkeys{\MT_options_name:}{multlined-pos=t}
+ \MH_fi:
+ \g@addto@macro\EQ_complete_mathdisplay:{
+ \begin{MTmultlined}\relax}
+ }
+ \g@addto@macro\EQ_complete_mathdisplay:\EQ_mathbody:
+ \MH_if_boolean:nT {outer_mult}{
+ \g@addto@macro\EQ_complete_mathdisplay:{\end{MTmultlined}}
+ }
+ \EQ_end_equation:
+% \end{macrocode}
+% Save the value of the \pkg{hyperref} link counter
+% \cs{Hy@linkcounter}.
+% \begin{macrocode}
+ \edef\EQ_hyperref_link_fint{\number\Hy@linkcounter}
+% \end{macrocode}
+% Firstly we call \cs{EQ_typeset_tags:} which suppresses all
+% output from the equations with the exception of the \cs{tag}'s.
+% This also allows us to find the widest tag which we'll need for
+% the environments that set's the math displays flush to the
+% margins, namely the \env{flalign} family. We save the counters
+% first though. As an extra treat we're also able to measure the
+% height and depth of the tags and the math display.
+% \begin{macrocode}
+ \savecounters@
+ \EQ_typeset_tags:
+% \end{macrocode}
+% \changes{v1.05d}{2004/02/26}{Removed one instance of
+% \cs{EQ_typeset_display:}}
+% Then find the widest tag. In an earlier implementation it was
+% necessary to typeset the display twice---a very costly procedure!
+% Now we save roughly 2120 lines of output from \cs{tracingall} on a
+% simple \env{gather} and 2600 lines with \env{align}. That's a
+% little more than 21\% saved\ldots
+% \changes{v2.10}{2004/07/26}{Removed \cs{find_widesttag:}}
+% \changes{v2.10}{2004/07/26}{Added check for equation number to
+% make sure the labelling/printing mechanism of \pkg{mathtools} work
+% properly}
+% Then we restore the counters and measure the extra material.
+% Save the value of the equation counter for later in case the
+% special equation numbering scheme from \pkg{mathtools} is in
+% effect.
+% \begin{macrocode}
+ \edef\l_EQ_equation_fint{\number\c@equation}
+ \restorecounters@
+% \end{macrocode}
+% Restore the value of the \pkg{hyperref} link counter
+% \cs{Hy@linkcounter}.
+% \begin{macrocode}
+ \Hy@linkcounter=\EQ_hyperref_link_fint
+ \settowidth\l_EQ_totalwidth_dim{%
+ $\EQ_left_side:\EQ_innerbox:{}\EQ_right_side:$}%
+% \end{macrocode}
+% Then call the appropriate \cs{EQ_modify_linewidth_}\meta{num}\texttt{:}
+% where \meta{num} is taken from the numbering scheme shown earlier.
+% \begin{macrocode}
+ \MH_if_boolean:nTF {outer_mult}
+ { \def\EQ_next:{_mult} }
+ { \def\EQ_next:{} }
+ \@nameuse{EQ_modify_linewidth_
+ \romannumeral\l_EQ_equationtype_int\EQ_next: :}%
+% \end{macrocode}
+% We can then typeset the display knowing that all modifications
+% have been made correctly. \cs{EQ_typeset_display:} gobbles the
+% \cs{tag}'s, which in turn is equivalent to setting the display
+% centered in non-\mode{fleqn} mode.
+% \begin{macrocode}
+ \EQ_typeset_display:
+% \end{macrocode}
+% Call the appropriate \cs{find_displaywidth_}\meta{num}\texttt{:}
+% which extracts the display width (\cs{l_EQ_displaywidth_dim}).
+% \changes{v2.15}{2017/03/31}{added EQ\_ prefix}
+% \begin{macrocode}
+ \@nameuse{EQ_find_displaywidth_
+ \romannumeral\l_EQ_equationtype_int\EQ_next: :}
+ \MH_addtolength:dn
+ \l_EQ_totalwidth_dim{\l_EQ_displaywidth_dim}
+% \end{macrocode}
+% We save the math display in a box of the correct size.
+% \changes{v2.15}{2017/03/31}{added MH\_ prefix}
+% \begin{macrocode}
+ \savebox\EQ_mathdisplay_box[\l_EQ_totalwidth_dim][c]{%
+ $\EQ_left_side:
+ \EQ_innerbox:{%
+ \MH_if_case:w \l_EQ_alignmentmode_int\MH_or:
+% \end{macrocode}
+% If the equation is centered we make the box centered
+% \begin{macrocode}
+ \makebox[\l_EQ_displaywidth_dim][c]
+ {\usebox{\EQ_mathdisplay_box}}%
+ \MH_or:
+ \makebox[\l_EQ_displaywidth_dim][c]
+ {\usebox{\EQ_mathdisplay_box}}%
+ \MH_or:
+% \end{macrocode}
+% Else we set it in a left aligned box and take care of the
+% \cs{@mathmargin} unless we're in \env{flalign} or \env{flalign*}.
+% \begin{macrocode}
+ \makebox[\l_EQ_displaywidth_dim][l]{%
+ \MH_if_num:w \l_EQ_equationtype_int=5 \MH_else:
+ \MH_if_num:w \l_EQ_equationtype_int=6 \MH_else:
+ \kern-\@mathmargin
+ \MH_fi:
+ \MH_fi:
+ \usebox{\EQ_mathdisplay_box}}%
+ \MH_or:
+ \makebox[\l_EQ_displaywidth_dim][l]{%
+ \MH_if_num:w \l_EQ_equationtype_int=5 \MH_else:
+ \MH_if_num:w \l_EQ_equationtype_int=6 \MH_else:
+ \kern-\@mathmargin
+ \MH_fi:
+ \MH_fi:
+ \usebox{\EQ_mathdisplay_box}}%
+ \MH_fi:}%
+ \EQ_right_side:$}%
+% \end{macrocode}
+% Then we're ready for typesetting. We just use an ordinary
+% |\begin{equation*}| | ... | |\end{equation*}| construction.
+% \changes{v2.00}{2004/04/14}{Used a saved definition of
+% \env{equation*} instead to allow overloading}
+% With v2.00 this changed to using a copy of the original
+% \env{equation*} environment. All this because of the
+% \opt{overload} option.
+% \begin{macrocode}
+ \begin{AmSequation*}
+ \@nameuse{EQ_place_display_
+ \romannumeral\l_EQ_equationtype_int\EQ_next: :}
+% \end{macrocode}
+% \cs{tag} must have a non-zero width argument in order to function
+% properly, so we make sure there is a tag before we try to typeset
+% it. Also make sure that the boolean |show_manual_tags| is true
+% just for this tag.
+% \changes{v2.15}{2017/03/31}{added MH\_ prefixes}
+% \begin{macrocode}
+ \MH_if_dim:w \g_EQ_widesttag_dim>\z@
+ \MH_set_boolean_T:n {show_manual_tags}
+% \end{macrocode}
+% Set the tag in a left aligned box when in \mode{leqno} mode and
+% right aligned when in \mode{reqno} mode.
+% \changes{v2.15}{2017/03/31}{added MH\_ prefixes}
+% \begin{macrocode}
+ \MH_if_case:w \l_EQ_alignmentmode_int
+ \MH_or: \EQ_place_tag:N r %#1
+ \MH_or: \EQ_place_tag:N l %#2
+ \MH_or: \EQ_place_tag:N r %#3
+ \MH_or: \EQ_place_tag:N l %#4
+ \MH_fi:
+ \MH_fi:
+ \end{AmSequation*}
+% \end{macrocode}
+% \changes{v2.11}{2004/08/03}{Use raw \TeX{} commands to speed it up}
+% Restore the equation counter.
+% \begin{macrocode}
+ \global\c@equation=\l_EQ_equation_fint\relax
+}
+% \end{macrocode}
+% \end{macro}
+%
+% \subsubsection{Initializing the Equations}
+%
+% \begin{macro}{\EQ_initialize:}
+% We save all the initializations in this hook:
+% \begin{macrocode}
+\def\EQ_initialize:{%
+ \m@th
+% \end{macrocode}
+% More support for \pkg{ntheorem}. We use special versions of
+% the tag setting commands in order to produce the correct result.
+% For instance in \mode{leqno} mode you would still want the end of
+% theorem symbol at the bottom of the math display.
+% \begin{macrocode}
+ \MH_let:NwN\SetTagPlusEndMark\EQ_tag_plus_endmark:
+ \MH_let:NwN\SetOnlyEndMark\EQ_only_endmark:
+ \l_EQ_temp_linewidth_dim\linewidth
+ \g_EQ_toprow_height_dim\z@
+ \g_EQ_bottomrow_depth_dim\z@
+ \g_EQ_toptag_height_dim\z@
+ \g_EQ_bottomtag_depth_dim\z@
+ \g_EQ_widesttag_dim\z@
+% \end{macrocode}
+% \changes{v2.11}{2004/08/03}{Added a \cs{g_EQ_latest_nonzerodepth_row_fint}}
+% \begin{macrocode}
+ \xdef\g_EQ_latest_nonzerodepth_row_fint{\z@}
+ \edef\EQ_restore_tex_delimiter:{%
+ \delimiterfactor\the\delimiterfactor
+ \delimitershortfall\the\delimitershortfall\relax}%
+ \g@addto@macro\EQ_complete_mathdisplay:
+ \EQ_restore_tex_delimiter:
+ \delimitershortfall\EmphEqdelimitershortfall
+ \delimiterfactor\EmphEqdelimiterfactor
+% \end{macrocode}
+% Disable \cs{intertext} and issue an error message. Using the
+% normal error message for an invalid \cs{intertext} would be less
+% than great because it would say
+% \begin{verbatim}
+% ! Package amsmath Error: Invalid use of \intertext.
+%
+% See the amsmath package documentation for explanation.
+% Type H <return> for immediate help.
+% ...
+% \end{verbatim}
+% That's not good enough, so we redefine things a bit. In
+% \pkg{amsmath} the command \cs{intertext@} is called inside the
+% math environments and defines \cs{intertext}. We hook into this.
+% \begin{macrocode}
+ \MH_let:NwN\intertext@\EQ_intertext:
+ \EQ_displaybreak:}
+% \end{macrocode}
+% \end{macro}
+% \begin{macro}{\EQ_intertext:}
+% Then we issue a suitable error message.
+% \changes{v2.15}{2017/03/31}{Extended to also include \cs{shortintertext}}
+% \begin{macrocode}
+\def\EQ_intertext:{%
+ \def\intertext##1{\PackageError{empheq}{%
+ You~cannot~use~\string\intertext\space inside~the~`empheq'
+ \MessageBreak
+ environment,~as~`empheq'~creates~an~unbreakable~box}\@eha}
+ \def\shortintertext##1{\PackageError{empheq}{%
+ You~cannot~use~\string\shortintertext\space inside~the~`empheq'
+ \MessageBreak
+ environment,~as~`empheq'~creates~an~unbreakable~box}\@eha}
+}
+\def\EQ_displaybreak:{%
+ \def\dspbrk@[##1]{\PackageWarning{empheq}{%
+ You~cannot~use~\string\displaybreak\space inside~the~`empheq'%
+ \MessageBreak
+ environment,~as~`empheq'~creates~an~unbreakable~box.}}}
+% \end{macrocode}
+% \end{macro}
+% \begin{macro}{\EQ_new_equationtype:nnn}
+% \begin{macro}{\EQ_new_equationtype_arg:nnn}
+% We have to inform \env{EmphEqMainEnv} which equation type we're
+% using, so we define an interface for that.
+% \begin{macrocode}
+\def\EQ_new_equationtype:nnn #1#2#3{%
+ \define@key{#1}{#2}[true]{%
+ \l_EQ_equationtype_int=#3\relax
+ \def\EQ_begin_equation:
+ {\g@addto@macro{\EQ_complete_mathdisplay:}{\begin{#2}}}%
+ \def\EQ_end_equation:
+ {\g@addto@macro{\EQ_complete_mathdisplay:}{\end{#2}}}}}
+\def\EQ_new_equationtype_arg:nnn #1#2#3{%
+ \define@key{#1}{#2}{%
+ \l_EQ_equationtype_int=#3\relax
+ \def\EQ_begin_equation:
+ {\g@addto@macro{\EQ_complete_mathdisplay:}{\begin{#2}{##1}}}%
+ \def\EQ_end_equation:
+ {\g@addto@macro{\EQ_complete_mathdisplay:}{\end{#2}}}}}
+% \end{macrocode}
+% \end{macro}
+% \end{macro}
+%
+% \subsubsection{Separating Tags and Display}
+%
+% \begin{macro}{\EQ_typeset_tags:}
+% Typesetting the tags without the display.
+% \begin{macrocode}
+\def\EQ_typeset_tags:{%
+ \sbox\EQ_tag_box{%
+ \parbox{.5\maxdimen}{%
+ \hbadness\@M
+% \end{macrocode}
+% Make sure all skips are zero ands support \pkg{ntheorem}.
+% \changes{v1.05a}{2004/02/17}{Forgot to support \pkg{ntheorem} in
+% both \cs{EQ_typeset_tags:} and \cs{EQ_typeset_display:}.}
+% \begin{macrocode}
+ \EQ_prevent_vertical_space:
+% \end{macrocode}
+% The AMS classes sets \cs{displaywidth} equal to \cs{columnwidth} in
+% \cs{everydisplay}. This is bad for us as then the reqno equation
+% numbers are not typeset at the right edge of the parbox we are
+% in. Simplest solution: Locally reset \cs{columnwidth}.
+% \changes{v2.15}{2017/03/28}{Resetting \cs{columnwidth} as a fix for
+% a setting in AMS classes}
+% \begin{macrocode}
+ \columnwidth\linewidth
+% \end{macrocode}
+% Support for \pkg{ntheorem}. This package patches \cs{endgather}
+% and \cs{endalign} with the commands \cs{TagsPlusEndmarks} and
+% \cs{RestoreTags} but we do not want them disturbing our output so
+% we disable them when typesetting the display.
+% \begin{macrocode}
+ \EQ_typeset_tags_hook:
+% \end{macrocode}
+% Then use our own versions of \cs{set@gather@field} and
+% \cs{set@field}.
+% \begin{macrocode}
+ \MH_let:NwN\set@gather@field\EQ_set_gather_field:
+ \MH_let:NwN\set@field\EQ_set_align_field:
+% \end{macrocode}
+% Next two lines are for measuring of the first and last row of a
+% display of the \env{align}, \env{alignat} or \env{flalign} family.
+% \begin{macrocode}
+ \MH_set_boolean_T:n {continue_measuring}
+ \global\g_EQ_temprow_int\m@ne
+ \MH_let:NwN \EQ_saved_maketag:n \maketag@@@
+ \def\maketag@@@##1{
+ \EQ_maketag:Nn \EQ_saved_maketag:n {##1}}
+% \end{macrocode}
+% Then typeset the display.
+% \begin{macrocode}
+ \EQ_complete_mathdisplay:}}}
+% \end{macrocode}
+% \end{macro}
+%
+% \begin{macro}{\EQ_predisplay_adjustment:}
+% \changes{v1.05c}{2004/02/25}{Added \cs{EQ_predisplay_adjustment:}}
+%
+% \changes{v2.00}{2004/04/14}{Added a missing \cs{normallineskip} in
+% case of extremely big tag height}
+%
+% For adjusting vertical spacing in the display.
+% \changes{v2.15}{2017/03/31}{added MH\_ prefix}
+% \begin{macrocode}
+\def\EQ_predisplay_adjustment:{%
+ \MH_if_dim:w \g_EQ_toprow_height_dim<\baselineskip
+ \vskip\g_EQ_toprow_height_dim
+ \vskip-\baselineskip
+ \MH_else:
+ \vskip-\normallineskip
+ \MH_fi:
+ \MH_if_dim:w \g_EQ_toptag_height_dim>\g_EQ_toprow_height_dim
+ \MH_if_dim:w \g_EQ_toptag_height_dim>\baselineskip
+ \vskip-\g_EQ_toptag_height_dim
+ \vskip\baselineskip
+ \vskip-\normallineskip
+ \MH_fi:
+ \MH_fi:}
+% \end{macrocode}
+% \end{macro}
+% \begin{macro}{\EQ_postdisplay_adjustment:}
+% \changes{v1.05c}{2004/02/25}{Added \cs{EQ_postdisplay_adjustment:}}
+% Name says it all.
+% \changes{v2.15}{2017/03/31}{added MH\_ prefixes}
+% \begin{macrocode}
+\def\EQ_postdisplay_adjustment:{%
+ \MH_if_dim:w \g_EQ_bottomrow_depth_dim<\dp\strutbox
+ \vskip\g_EQ_bottomrow_depth_dim
+ \vskip-\dp\strutbox
+ \MH_fi:
+ \MH_if_dim:w \g_EQ_bottomtag_depth_dim>\g_EQ_bottomrow_depth_dim
+ \MH_if_dim:w \g_EQ_bottomtag_depth_dim>\dp\strutbox
+ \vskip-\g_EQ_bottomtag_depth_dim
+ \vskip\dp\strutbox
+ \MH_fi:
+ \MH_fi:}
+% \end{macrocode}
+% \end{macro}
+% \begin{macro}{\EQ_typeset_display:}
+% \changes{v1.05c}{2004/02/25}{Now takes two arguments: Either both
+% empty or \cs{EQ_predisplay_adjustment:} \&
+% \cs{EQ_postdisplay_adjustment:}}
+% \changes{v1.05d}{2004/02/26}{Changed again. Uses only
+% \cs{EQ_predisplay_adjustment:} \& \cs{EQ_postdisplay_adjustment:}}
+% We typeset the display without the tags and save it in a box.
+% \begin{macrocode}
+\def\EQ_typeset_display:{%
+ \sbox\EQ_mathdisplay_box{%
+ \parbox{\l_EQ_temp_linewidth_dim}{%
+% \end{macrocode}
+% We adjust the vertical spacing if needed:
+% \begin{macrocode}
+ \EQ_predisplay_adjustment:
+ \EQ_prevent_vertical_space:
+% \end{macrocode}
+% Support for \pkg{ntheorem}. This package patches \cs{endgather}
+% and \cs{endalign} with the commands \cs{TagsPlusEndmarks} and
+% \cs{RestoreTags} but we do not want them disturbing our output so
+% we disable them when typesetting the display.
+% \begin{macrocode}
+ \EQ_typeset_display_hook:
+% \end{macrocode}
+% Instead of gobbling the tags it's better to use \cs{phantom} since
+% a tall tag will ruin the vertical spacing. We use a special
+% version of \cs{phantom} to help us with other things as well.
+% \begin{macrocode}
+ \def\maketag@@@##1{\EQ_maketag:Nn \vphantom{##1}}%
+ \EQ_complete_mathdisplay:
+% \end{macrocode}
+% More space adjustment.
+% \begin{macrocode}
+ \EQ_postdisplay_adjustment:}}}
+% \end{macrocode}
+% \end{macro}
+% We need a few little helpers:
+% \begin{macro}{\EQ_prevent_vertical_space:}
+% \changes{v2.00}{2004/04/14}{Used \cs{nointerlineskip} instead}
+% As we typeset the math display inside a \cs{parbox} the math
+% environment will add \cs{abovedisplayskip} at the top and
+% \cs{belowdisplayskip} at the bottom. Also we suppress interline
+% skip with \cs{nointerlineskip}.
+% \begin{macrocode}
+\def\EQ_prevent_vertical_space:{%
+ \nointerlineskip
+ \abovedisplayskip\z@ \belowdisplayskip\z@\relax}
+% \end{macrocode}
+% \end{macro}
+% \begin{macro}{\EQ_maketag:Nn}
+% \changes{v1.05d}{2004/02/26}{Changed name from
+% \cs{EmphEqphantom} to \cs{EmphEqMaketag}}
+% This takes care of two things: We need to store the height of the
+% top tag and the depth of the bottom tag for later, but we also
+% need to use a \cs{vphantom} of the tag as it may be taller than
+% the line of math it is attached to (say, if you are using node
+% connectors from \pkg{pstricks} or some other weird stuff).
+% \changes{v2.15}{2017/03/31}{added MH\_ prefixes}
+% \begin{macrocode}
+\def\EQ_maketag:Nn #1#2{%
+ \setboxz@h{#2}%
+ \MH_if_num:w \row@=\@ne
+ \global \g_EQ_toptag_height_dim = \ht\z@
+ \MH_fi:
+% \end{macrocode}
+% Similarly the depth of the last row is determined by measuring the
+% depth of box zero for each row and then using the last of the
+% values.
+% \begin{macrocode}
+ \global \g_EQ_bottomtag_depth_dim=\dp\z@
+% \end{macrocode}
+% \changes{v2.11}{2004/08/03}{Added a check for being in the last row.}
+% Save the row number for later as the \opt{ntheorem} option will need
+% it.
+% \changes{v2.15}{2017/03/31}{added MH\_ prefixes}
+% \begin{macrocode}
+ \xdef\g_EQ_latest_nonzerodepth_row_fint{\number\row@}
+ \@tempdima=\wd\z@
+ \MH_if_dim:w \@tempdima > \g_EQ_widesttag_dim
+ \global \g_EQ_widesttag_dim = \@tempdima
+ \MH_fi:
+ #1{#2}%
+}
+% \end{macrocode}
+% \end{macro}
+%
+% \subsubsection{Measuring \env{gather}}
+% \begin{macro}{\EQ_set_gather_field:}
+% Box zero (\cs{z@}) contains the entire line of math in a
+% \env{gather}, so instead of typesetting it we typeset the phantom
+% of it and call \cs{EQ_measure_gather:}.
+% \changes{v2.15}{2017/03/31}{added MH\_ prefix}
+% \begin{macrocode}
+\def\EQ_set_gather_field:{%
+ \iftagsleft@
+ \global\lineht@\ht\z@
+ \MH_else:
+ \global\lineht@\dp\z@
+ \MH_fi:
+ \kern\eqnshift@
+ %\boxz@ % Original amsmath version
+ \EQ_measure_gather:
+ \phantom{\boxz@}%
+ \hfil}
+% \end{macrocode}
+% \end{macro}
+% \begin{macro}{\EQ_measure_gather:}
+% Measuring the height of the individual lines in a \env{gather} is
+% fairly easy. If \cs{row@} is equal to one, we're in the first
+% line, and the height of the top row is simply given by the height
+% of box zero.
+% \changes{v2.15}{2017/03/31}{added MH\_ prefixes}
+% \begin{macrocode}
+\def\EQ_measure_gather:{%
+ \MH_if_num:w \row@=\@ne
+ \g_EQ_toprow_height_dim=\ht\z@
+ \MH_fi:
+% \end{macrocode}
+% Similarly the depth of the last row is determined by measuring the
+% depth of box zero for each row and then using the last of the
+% values.
+% \begin{macrocode}
+ \g_EQ_bottomrow_depth_dim=\dp\z@
+ \global\g_EQ_toprow_height_dim\g_EQ_toprow_height_dim
+ \global\g_EQ_bottomrow_depth_dim\g_EQ_bottomrow_depth_dim}
+% \end{macrocode}
+% \end{macro}
+%
+% \subsubsection{Measuring \env{align}, \env{alignat} and \env{flalign}}
+% \begin{macro}{\EQ_set_align_field:}
+% Same technique as for \cs{EQ_set_gather_field}. Only this time
+% we have to stop the underlying \cs{start@align} from going over
+% the measuring phase twice. That's when the conditional
+% \bool{continue_measuring} comes into the picture.
+% \changes{v2.15}{2017/03/31}{added MH\_ prefix}
+% \begin{macrocode}
+\def\EQ_set_align_field:{%
+ \column@plus
+ \iftagsleft@
+ \MH_if_dim:w \ht\z@>\lineht@
+ \global\lineht@\ht\z@
+ \MH_fi:
+ \MH_else:
+ \MH_if_dim:w \dp\z@>\lineht@
+ \global\lineht@\dp\z@
+ \MH_fi:
+ \MH_fi:
+ %\boxz@ % Original amsmath version
+ \MH_if_boolean:nT {continue_measuring}{ \EQ_measure_align: }
+ \phantom{\boxz@}}
+% \end{macrocode}
+% \end{macro}
+% \begin{macro}{\EQ_measure_align:}
+% We have to measure every single field of the \cs{align@} \ldots\@
+% Oh well, here it comes:
+% \changes{v2.15}{2017/03/31}{added MH\_ prefix}
+% \begin{macrocode}
+\newcount\g_EQ_temprow_int
+\MH_new_boolean:n {still_in_samerow}
+\MH_new_boolean:n {continue_measuring}
+\def\EQ_measure_align:{%
+ \MH_if_num:w \g_EQ_temprow_int=\row@
+ \MH_set_boolean_T:n {still_in_samerow}
+ \MH_else:
+ \MH_if_num:w \g_EQ_temprow_int>\row@
+ \global\MH_set_boolean_F:n {continue_measuring}
+ \MH_fi:
+ \MH_set_boolean_F:n {still_in_samerow}
+ \global\@tempdimc=\@tempdimb
+ \global\@tempdimb=\g_EQ_bottomrow_depth_dim
+ \MH_fi:
+ \MH_if_boolean:nTF {continue_measuring}{
+ \global\g_EQ_temprow_int=\row@
+ \@tempcnta=\row@
+ \MH_if_num:w \@tempcnta=\z@
+ \@tempdima=\g_EQ_toprow_height_dim
+ \g_EQ_toprow_height_dim=\ht\z@\relax
+ \MH_if_dim:w \g_EQ_toprow_height_dim<\@tempdima
+ \g_EQ_toprow_height_dim\@tempdima
+ \MH_fi:
+ \MH_fi:
+ \MH_if_boolean:nTF {still_in_samerow}{
+ \@tempdima=\g_EQ_bottomrow_depth_dim
+ \g_EQ_bottomrow_depth_dim=\dp\z@\relax
+ \MH_if_dim:w \g_EQ_bottomrow_depth_dim<\@tempdima
+ \g_EQ_bottomrow_depth_dim\@tempdima
+ \MH_fi:
+ }{
+ \g_EQ_bottomrow_depth_dim\dp\z@\relax
+ }
+ }{
+ \global\g_EQ_bottomrow_depth_dim=\@tempdimc
+ }
+ \global\g_EQ_toprow_height_dim\g_EQ_toprow_height_dim
+ \global\g_EQ_bottomrow_depth_dim\g_EQ_bottomrow_depth_dim}
+% \end{macrocode}
+% \end{macro}
+% \subsubsection{Typesetting the Tags}
+% \begin{macro}{\EQ_place_tag:N}
+% Fool \LaTeX\ into thinking the tag is only as wide as the widest
+% tag (remember the real width of the \cs{EQ_tag_box} is
+% $0.5$\cs{maxdimen}.
+% \begin{macrocode}
+\def\EQ_place_tag:N #1{%
+ \tag*{\makebox[\g_EQ_widesttag_dim][#1]{%
+ \EQ_vertical_tag_adjustment:n {\usebox{\EQ_tag_box}}}}}
+% \end{macrocode}
+% \end{macro}
+% \begin{macro}{\EQ_vertical_tag_adjustment:n}
+% \changes{v2.00}{2004/04/14}{Added a missing \cs{normallineskip} in
+% case of extremely big tag height}
+% This is a little tricky. We have measured the height of the top
+% row and top tag and the depth of the bottom row and bottom tag.
+% Now we have to adjust the vertical spacing so it comes out just
+% right. Remember we have separated the tags and the math display
+% but we would still like to have them line up as they should! We do
+% it all inside a group so we can use the scratch dimensions safely.
+% \changes{v2.15}{2017/03/31}{added MH\_ prefix}
+% \begin{macrocode}
+\def\EQ_vertical_tag_adjustment:n #1{{%
+ \@tempdima\z@\@tempdimb\z@
+ \@tempdimc=\dp\strutbox
+ \MH_if_dim:w \g_EQ_toprow_height_dim<\baselineskip
+ \MH_setlength:dn
+ \@tempdima{\g_EQ_toprow_height_dim-\baselineskip}
+ \MH_else:
+ \MH_setlength:dn \@tempdima{-\normallineskip}
+ \MH_fi:
+ \MH_if_dim:w \g_EQ_toptag_height_dim>\g_EQ_toprow_height_dim
+ \MH_if_dim:w \g_EQ_toptag_height_dim>\baselineskip
+ \MH_addtolength:dn \@tempdima
+ {-\g_EQ_toptag_height_dim+\baselineskip-\normallineskip}
+ \MH_fi:
+ \MH_fi:
+ \MH_if_dim:w \g_EQ_bottomrow_depth_dim<\@tempdimc
+ \MH_setlength:dn \@tempdimb
+ {\g_EQ_bottomrow_depth_dim-\@tempdimc}
+ \MH_fi:
+% \end{macrocode}
+% \changes{v2.00}{2004/04/14}{Fixed minor bug}
+% If the last condition is fulfilled we also need to cancel the
+% first condition.
+% \changes{v2.15}{2017/03/31}{added MH\_ prefix}
+% \begin{macrocode}
+ \MH_if_dim:w \g_EQ_bottomtag_depth_dim>\g_EQ_bottomrow_depth_dim
+ \MH_if_dim:w \g_EQ_bottomtag_depth_dim<\@tempdimc
+ \MH_else:
+ \MH_addtolength:dn \@tempdimb
+ {-\g_EQ_bottomtag_depth_dim+\@tempdimc}
+ \MH_fi:
+ \MH_fi:
+ \raisebox{-0.5\@tempdima+0.5\@tempdimb}{#1}}}%
+% \end{macrocode}
+% \end{macro}
+% \section{Adding the Equation Types}
+% Adding an equation type is fairly easy. The outline is this:
+% \begin{list}{}{\leftmargin2em}
+% \item[]|\EQ_new_equationtype{EmphEqEnv}|\marg{\AmS\
+% env}\marg{arabic num}\texttt{:}
+% \item[]|\def\EQ_modify_linewidth_|\meta{roman num}\texttt{:}\\
+% \hspace*{2em}\marg{Optionally alter \cs{l_EQ_temp_linewidth_dim}}
+% \item[]|\def\EQ_find_displaywidth_|\meta{roman num}\texttt{:}\\
+% \hspace*{2em}\marg{Set \cs{l_EQ_displaywidth_dim}}
+% \item[]|\def\EQ_place_display_|\meta{roman num}\texttt{:}\\
+% \hspace*{2em}\marg{Make use of \cs{EQ_use_mathdisplay:}}
+% \end{list}
+% \begin{macro}{\EQ_use_mathdisplay:}
+% A shorthand.
+% \begin{macrocode}
+\def\EQ_use_mathdisplay:{%
+ \EQ_outerbox:{\usebox{\EQ_mathdisplay_box}}}
+% \end{macrocode}
+% \end{macro}
+% \begin{macrocode}
+\def\EQ_options_name:{EmphEqOpt}
+% \end{macrocode}
+% \subsection{The \env{equation} Family}
+% Just a few aliases because \env{equation} and \env{equation*}
+% behave strangely. Much better to use \env{gather} instead as this
+% is easier to measure. Besides it's less work (for us, not the
+% compiler). ;-)
+% \begin{macrocode}
+\MH_keyval_alias_with_addon:nnnn {EmphEqEnv}{equation}{gather}
+ {\EQ_disable_linebreak:}
+\MH_keyval_alias_with_addon:nnnn {EmphEqEnv}{equation*}{gather*}
+ {\EQ_disable_linebreak:}
+% \end{macrocode}
+% \begin{macro}{\EQ_disable_linebreak:}
+% \changes{v1.05c}{2004/02/25}{Introduced
+% \cs{EQ_disable_linebreak:} to prevent multi line
+% \env{equation}'s}
+% \changes{v1.05c}{2004/02/25}{Fixed it so that it works better}
+% Disable the line breaks in \env{equation} as these are actually
+% possible now, since we're in a \env{gather}! \cs{Let@} does it in
+% the environment, so that's the place to hack.
+% \begin{macrocode}
+\def\EQ_disable_linebreak:{%
+ \MH_let:NwN \EQ_saved_Let: \Let@
+ \def\Let@{\def\\{%
+ \PackageError{empheq}{You~ cannot~ use~ \string\\~ in~ equation!}
+ {\@eha}}
+% \end{macrocode}
+% Then restore the meaning of \cs{Let@} in case there is a
+% \env{gathered} or alike coming up.
+% \begin{macrocode}
+ \MH_let:NwN \Let@ \EQ_saved_Let:
+ }%
+}
+% \end{macrocode}
+% \end{macro}
+%
+% \subsection{The \env{align} Family}
+% \begin{macrocode}
+\EQ_new_equationtype:nnn {EmphEqEnv}{align}{1}
+\EQ_new_equationtype:nnn {EmphEqEnv}{align*}{2}
+% \end{macrocode}
+% \begin{macrocode}
+\def\EQ_modify_linewidth_i:{%
+% \end{macrocode}
+% If $\cs{maxfields@}=2$ we're in a single column \env{align} and
+% thus there is no flexible space.
+% \changes{v2.15}{2017/03/31}{added MH\_ prefix}
+% \begin{macrocode}
+ \MH_if_num:w \maxfields@=\tw@
+ \MH_else:
+% \end{macrocode}
+% But if there is any flexible space we would very much like to
+% adjust it because we add material to the right and left of the
+% math display.
+% \changes{v2.15}{2017/03/31}{added MH\_ prefix}
+% \changes{v2.15}{2017/03/31}{added EQ\_ prefix}
+% \begin{macrocode}
+ \setlength{\l_EQ_temp_linewidth_dim}{%
+ \linewidth-\l_EQ_totalwidth_dim-\widthof{\EQ_outerbox:{}}}%
+ \MH_fi:}
+\def\EQ_modify_linewidth_ii:{\EQ_modify_linewidth_i:}
+\def\EQ_find_displaywidth_i:{%
+ \MH_setlength:dn \l_EQ_displaywidth_dim
+ {\l_EQ_temp_linewidth_dim-\alignsep@}
+ \MH_if_case:w \l_EQ_alignmentmode_int
+ \MH_or: % #1:
+ \MH_addtolength:dn \l_EQ_displaywidth_dim{-\alignsep@}%
+ \MH_or: % #2:
+ \MH_addtolength:dn \l_EQ_displaywidth_dim{-\alignsep@}%
+ \MH_or: % #3:
+ \MH_addtolength:dn \l_EQ_displaywidth_dim{-\@mathmargin}%
+ \MH_or: % #4:
+ \MH_addtolength:dn \l_EQ_displaywidth_dim{-\@mathmargin}%
+ \MH_fi:}
+\def\EQ_find_displaywidth_ii:{\EQ_find_displaywidth_i:}
+\def\EQ_place_display_i: {\EQ_use_mathdisplay:}
+\def\EQ_place_display_ii:{\EQ_place_display_i:}
+% \end{macrocode}
+%
+% \subsection{The \env{gather} Family}
+% This is a gem. Notice how easy it is once we got the framework for
+% it.
+% \changes{v2.15}{2017/03/31}{added MH\_ prefixes}
+% \changes{v2.15}{2017/03/31}{added EQ\_ prefix}
+% \begin{macrocode}
+\EQ_new_equationtype:nnn {EmphEqEnv}{gather}{3}
+\EQ_new_equationtype:nnn {EmphEqEnv}{gather*}{4}
+\def\EQ_modify_linewidth_iii:{}
+\def\EQ_modify_linewidth_iv:{\EQ_modify_linewidth_iii:}
+\def\EQ_find_displaywidth_iii:{%
+ \MH_setlength:dn \l_EQ_displaywidth_dim{\totwidth@}%
+ \MH_if_case:w \l_EQ_alignmentmode_int \MH_or: \MH_or: % #1 & #2:
+ \MH_or: % #3:
+ \MH_addtolength:dn \l_EQ_displaywidth_dim{-\@mathmargin}%
+ \MH_or: % #4:
+ \MH_addtolength:dn \l_EQ_displaywidth_dim{-\@mathmargin}%
+ \MH_fi:}
+\def\EQ_find_displaywidth_iv:{\EQ_find_displaywidth_iii:}
+\def\EQ_place_display_iii:{\EQ_use_mathdisplay:}
+\def\EQ_place_display_iv: {\EQ_place_display_iii:}
+% \end{macrocode}
+%
+% \subsection{The \env{flalign} Family}\label{sec:flalign+multline}
+% This was fairly easy, but then I had to take care of \ldots
+%
+% \subsubsection{The \env{multline} Family}
+% This was quite a pickle. Its implementation in \pkg{amsmath} is
+% very strange indeed; for instance the tag is typeset in the
+% measuring phase! So instead I thought I could fake it, but how
+% should I do it?
+%
+% The basic idea of \env{multline} is to typeset the first line
+% flush left, the last line flush right and the rest centered.
+% ``Aha,'' I thought, ``what if I make three columns, fool \LaTeX{}
+% into thinking the contents of the columns take no space and then
+% just put each line into a column?'' Thus the \env{flalign}
+% environment came to mind. I also distinctly remembered to have
+% seen an article about the equivalent of \cs{rlap} and \cs{llap} in
+% math mode, namely an article by Alexander R.~Perlis. This works
+% pretty well, and so does the same trick for producing a
+% \env{multlined} environment. That one uses an \env{alignedat} with
+% three columns.
+%
+% \changes{v1.05c}{2004/02/25}{Makes use of
+% \cs{MakeKeyvalAliasAddon} now}
+% Then we make \env{multline} into an alias for {flalign} but with
+% an extra switch \cs{ifEQ_multline:} set to true.
+% \begin{macrocode}
+\MH_keyval_alias_with_addon:nnnn {EmphEqEnv}{multline}
+ {flalign}{\MH_set_boolean_T:n {multline}}
+\MH_keyval_alias_with_addon:nnnn {EmphEqEnv}{multline*}
+ {flalign*}{\MH_set_boolean_T:n {multline}}
+% \end{macrocode}
+% \begin{macro}{\if_EQ_multline:}
+% \begin{macro}{\EQ_generic_multline:nnnnn}
+% We make a general command that'll provide the correct spacing and
+% then just define the user commands in terms of that single generic
+% one.
+% \begin{macrocode}
+\MH_new_boolean:n {multline}
+% \end{macrocode}
+% \begin{macro}{\EQ_multline_modify_linewidth_tag:}
+% \begin{macro}{\EQ_multline_modify_linewidth_notag:}
+% There is a big difference between the spacing in \env{multline}
+% depending on the presence of a tag. Therefore we split it up in
+% two cases. I made this choice because we need these adjustments
+% inside the adjustments for \env{flalign} and that makes it
+% incredibly difficult to see what happens.
+% \changes{v2.15}{2017/03/31}{added MH\_ prefixes}
+% \begin{macrocode}
+\def\EQ_multline_modify_linewidth_tag:{%
+ \MH_if_case:w \l_EQ_alignmentmode_int
+ \MH_or: % #1:
+ \MH_addtolength:dn \l_EQ_temp_linewidth_dim{%
+ -\g_EQ_widesttag_dim-\multlinetaggap-\multlinegap}%
+ \MH_or:
+ \MH_addtolength:dn \l_EQ_temp_linewidth_dim{%
+ -\g_EQ_widesttag_dim-\multlinetaggap-\multlinegap}%
+ \MH_or:
+ \MH_addtolength:dn \l_EQ_temp_linewidth_dim{%
+ -\@mathmargin-\g_EQ_widesttag_dim-\multlinetaggap}%
+ \MH_or:
+ \MH_addtolength:dn
+ \l_EQ_temp_linewidth_dim{-\@mathmargin-\multlinegap}%
+ \MH_fi:
+}
+\def\EQ_multline_modify_linewidth_notag:{%
+ \MH_if_case:w \l_EQ_alignmentmode_int
+ \MH_or:
+ \MH_addtolength:dn \l_EQ_temp_linewidth_dim{-2\multlinegap}
+ \MH_or:
+ \MH_addtolength:dn \l_EQ_temp_linewidth_dim{-2\multlinegap}
+ \MH_or:
+ \MH_addtolength:dn
+ \l_EQ_temp_linewidth_dim{-\@mathmargin-\multlinegap}%
+ \MH_or:
+ \MH_addtolength:dn
+ \l_EQ_temp_linewidth_dim{-\@mathmargin-\multlinegap}%
+ \MH_fi:
+}
+% \end{macrocode}
+% \end{macro}
+% \end{macro}
+%
+% \begin{environment}{MTmultlined}
+% Save the superior \env{multlined} from \pkg{mathtools} under a new
+% name.
+% \begin{macrocode}
+\MH_let:NwN \MTmultlined \multlined
+\MH_let:NwN \endMTmultlined \endmultlined
+% \end{macrocode}
+% \end{environment}
+% We make a choice here: the good \env{multline} or the bad?
+% \begin{macrocode}
+\MH_if_boolean:nTF {good_mult}{
+% \end{macrocode}
+% The good:
+% \changes{v2.15}{2017/03/31}{added MH\_ prefix}
+% \changes{v2.15}{2017/03/31}{added EQ\_ prefix}
+% \begin{macrocode}
+\MH_keyval_alias_with_addon:nnnn {EmphEqEnv}{multline}{gather}
+ {\MH_set_boolean_T:n {outer_mult}}
+\MH_keyval_alias_with_addon:nnnn {EmphEqEnv}{multline*}{gather*}
+ {\MH_set_boolean_T:n {outer_mult}}
+\def\EQ_modify_linewidth_iii_mult:{
+ \setlength{\l_EQ_temp_linewidth_dim}{%
+ \linewidth-\widthof{\EQ_outerbox:{}}}%
+ \MH_if_dim:w \g_EQ_widesttag_dim>\z@
+ \EQ_multline_modify_linewidth_tag:
+ \MH_else:
+ \EQ_multline_modify_linewidth_notag:
+ \MH_fi:
+ \MH_setlength:dn \l_MT_multwidth_dim{\l_EQ_temp_linewidth_dim}
+}
+\def\EQ_modify_linewidth_iv_mult:{\EQ_modify_linewidth_iii_mult:}
+\def\EQ_find_displaywidth_iii_mult:{%
+ \MH_setlength:dn \l_EQ_displaywidth_dim{\totwidth@}%
+ \MH_if_case:w \l_EQ_alignmentmode_int \MH_or: \MH_or: % #1 & #2:
+ \MH_or: % #3:
+ \MH_addtolength:dn \l_EQ_displaywidth_dim{-\@mathmargin}%
+ \MH_or: % #4:
+ \MH_addtolength:dn \l_EQ_displaywidth_dim{-\@mathmargin}%
+ \MH_fi:}
+\def\EQ_find_displaywidth_iv_mult:{\EQ_find_displaywidth_iii_mult:}
+\def\EQ_place_display_iii_mult:{
+ \MH_if_case:w \l_EQ_alignmentmode_int
+ \MH_or: % #1:
+ \MH_if_dim:w \g_EQ_widesttag_dim>\z@
+ \abovedisplayshortskip=\abovedisplayskip
+ \rlap{
+ \kern-.5\linewidth
+ \kern\multlinegap
+ \EQ_use_mathdisplay:
+ }
+ \MH_else:
+ \EQ_use_mathdisplay:
+ \MH_fi:
+ \MH_or: % #2:
+ \MH_if_dim:w \g_EQ_widesttag_dim>\z@
+ \llap{
+ \EQ_use_mathdisplay:
+ \kern-.5\linewidth
+ \kern\multlinegap
+ }%
+ \MH_else:
+ \EQ_use_mathdisplay:
+ \MH_fi:
+ \MH_or: % #3:
+ \EQ_use_mathdisplay:
+ \MH_or: % #4:
+ \EQ_use_mathdisplay:
+ \MH_fi:
+}
+\def\EQ_place_display_iv_mult: {\EQ_place_display_iii_mult:}
+% \end{macrocode}
+% \begin{macrocode}
+}{
+% \end{macrocode}
+% The bad
+% \begin{macrocode}
+\def\EQ_generic_multline:nnnnn #1#2#3#4#5{%
+% \end{macrocode}
+% This is a very tricky little detail. If we're in \mode{leqno} mode
+% we need just a we little width inside the \env{flalign}, so we put
+% in the smallest possible amount. Otherwise we wind up in a
+% \begin{verbatim}
+% ! Arithmetic overflow.
+% <recently read> \count@
+%
+% l.304 % \end{empheq}
+% \end{verbatim}
+% If you scroll past this error it will be typeset correctly, but
+% how many users can fix this by themselves?
+% \changes{v2.15}{2017/03/31}{added MH\_ prefix}
+% \begin{macrocode}
+ \hskip1sp#2#3{{}#5}#4%
+ \MH_if:w *#1\relax\MH_else:\notag\MH_fi:}
+% \end{macrocode}
+% \end{macro}
+% \end{macro}
+% \begin{macro}{\multleft}
+% \begin{macro}{\multcenter}
+% \begin{macro}{\multright}
+% Then the commands that use \cs{EQ_generic_multline:nnnnn}. The
+% first three are the general purpose ones that also have a starred
+% version. The other three are for normal use.
+% \begin{macrocode}
+\def\multleft{\@ifstar{%
+ \MH_if_boolean:nTF {multlined}{
+ \def\EQ_next:{\EQ_multlined_left:n}
+ }{
+ \def\EQ_next:{
+ \EQ_generic_multline:nnnnn {*}{}{\mathrlap}{&&&&&}}
+ }
+ \EQ_next:}{%
+ \MH_if_boolean:nTF {multlined}{
+ \def\EQ_next:{\EQ_multlined_left:n}%
+ }{
+ \def\EQ_next:{
+ \EQ_generic_multline:nnnnn {}{}{\mathrlap}{&&&&&}}
+ }
+ \EQ_next:}}%
+\def\multcenter{\@ifstar{%
+ \MH_if_boolean:nTF {multlined}{
+ \def\EQ_next:{\EQ_multlined_center:n}%
+ }{
+ \def\EQ_next:{
+ \EQ_generic_multline:nnnnn {*}{&&}{\mathclap}{&&&}}
+ }
+ \EQ_next:}{%
+ \MH_if_boolean:nTF {multlined}{
+ \def\EQ_next:{\EQ_multlined_center:n }%
+ }{
+ \def\EQ_next:{
+ \EQ_generic_multline:nnnnn {}{&&}{\mathclap}{&&&}}
+ }
+ \EQ_next:}}%
+\def\multright{\@ifstar{%
+ \MH_if_boolean:nTF {multlined}{
+ \def\EQ_next:{\EQ_multlined_right:n }%
+ }{
+ \def\EQ_next:{
+ \EQ_generic_multline:nnnnn {*}{&&&&&}{\mathllap}{}}
+ }
+ \EQ_next:}{%
+ \MH_if_boolean:nTF {multlined}{
+ \def\EQ_next:{\EQ_multlined_right:n }%
+ }{
+ \def\EQ_next:{
+ \EQ_generic_multline:nnnnn {}{&&&&&}{\mathllap}{}}
+ }
+ \EQ_next:}}
+% \end{macrocode}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \begin{macro}{\multfirst}
+% \begin{macro}{\multmiddle}
+% \begin{macro}{\multlast}
+% If we're using the \mode{leqno} option, the equation number should
+% be placed in the first line.
+% \changes{v2.15}{2017/03/31}{added MH\_ prefixes}
+% \begin{macrocode}
+\def\multfirst{%
+ \MH_if_boolean:nTF {multlined}{
+ \def\EQ_next:{\EQ_multlined_left:n }%
+ }{
+ \MH_if_case:w \l_EQ_alignmentmode_int
+ \MH_or: % #1:
+ \def\EQ_next:{\multleft}%
+ \MH_or: % #2:
+ \def\EQ_next:{\multleft*}%
+ \MH_or: % #3:
+ \def\EQ_next:{\multleft}%
+ \MH_or: % #4:
+ \def\EQ_next:{\multleft*}%
+ \MH_fi:
+ }
+ \EQ_next:}
+% \end{macrocode}
+% Pretend the star was never there.
+% \begin{macrocode}
+\def\multmiddle{%
+ \def\EQ_next:{%
+ \MH_if_boolean:nTF {multlined}{
+ \expandafter\EQ_multlined_center:n
+ }{
+ \expandafter\multcenter
+ }
+ }%
+ \@ifstar{\EQ_next:}{\EQ_next:}}
+% \end{macrocode}
+% If we're using the \mode{reqno} option, the equation number
+% should be placed in the last line.
+% \changes{v2.15}{2017/03/31}{added MH\_ prefixes}
+% \begin{macrocode}
+\def\multlast{%
+ \MH_if_boolean:nTF {multlined}{
+ \def\EQ_next:{\EQ_multlined_right:n }%
+ }{
+ \MH_if_case:w \l_EQ_alignmentmode_int
+ \MH_or: % #1:
+ \def\EQ_next:{\multright*}%
+ \MH_or: % #2:
+ \def\EQ_next:{\multright}%
+ \MH_or: % #3:
+ \def\EQ_next:{\multright*}%
+ \MH_or: % #4:
+ \def\EQ_next:{\multright}%
+ \MH_fi:
+ }
+ \EQ_next:}
+% \end{macrocode}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \begin{macro}{\multlinedwidth}
+% Here begins the code for \env{multlined}.
+% \begin{macrocode}
+\MH_new_boolean:n {multlined}
+\newdimen\multlinedwidth
+\MH_setlength:dn \multlinedwidth{.7\linewidth}
+% \end{macrocode}
+% \end{macro}
+% \begin{environment}{multlined}
+% \changes{v2.10}{2004/07/26}{Disallow spaces before optional
+% arguments.}
+% \begin{macro}{\EQ_multlined_I:w}
+% \begin{macro}{\EQ_multlined_II:w}
+% \begin{macro}{\EQ_multlined_III:nn}
+% Some trickery as we want to scan for two optional arguments that
+% may be interchanged.
+% \begin{macrocode}
+\renewenvironment{multlined}%
+ {\MH_nospace_ifnextchar:Nnn [ {\EQ_multlined_I:w}
+ {\EQ_multlined_I:w[c]}
+ }
+ { \endaligned }
+\def\EQ_multlined_I:w[#1]{%
+ \MH_nospace_ifnextchar:Nnn[ {\EQ_multlined_II:w[#1]}
+ {\EQ_multlined_II:w[#1][\multlinedwidth]}}
+\def\EQ_multlined_II:w[#1][#2]{%
+% \end{macrocode}
+% We tell \LaTeX\ that we're in \env{multlined}.
+% \begin{macrocode}
+ \MH_set_boolean_T:n {multlined}
+% \end{macrocode}
+% \changes{v2.10}{2004/07/26}{Fixed restoration of \cmd{\\}}
+% Restore the meaning of \cmd{\\} inside \env{multlined}, else it
+% wouldn't work in the \env{equation} environment.
+% \begin{macrocode}
+ \MH_let:NwN \Let@\MT_saved_Let:
+% \end{macrocode}
+% Then check the first optional argument and call
+% \cs{EQ_multlined_III:w}, which sets \cs{multlinewidth} if
+% specified and calls \cs{start@aligned} (an \env{alignedat} with
+% three columns).
+% \changes{v2.15}{2017/03/31}{added MH\_ prefix}
+% \begin{macrocode}
+ \MH_if:w t#1\relax \EQ_multlined_III:nn {#1}{#2}%
+ \MH_else:
+ \MH_if:w b#1\relax \EQ_multlined_III:nn {#1}{#2}%
+ \MH_else:
+ \MH_if:w c#1\relax \EQ_multlined_III:nn {#1}{#2}%
+ \MH_else:
+ \EQ_multlined_III:nn {#2}{#1}%
+ \MH_fi:
+ \MH_fi:
+ \MH_fi:
+ \mkern-\thinmuskip
+ \EQ_next:}
+\def\EQ_multlined_III:nn #1#2{%
+ \def\EQ_next:{\setlength\multlinedwidth{#2}%
+ \start@aligned{#1}{\thr@@}}}
+% \end{macrocode}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{environment}
+% \begin{macro}{\EQ_multlined_left:n}
+% \begin{macro}{\EQ_multlined_center:n}
+% \begin{macro}{\EQ_multlined_right:n}
+% The internal versions of \cs{multleft} etc.\ adapted for
+% \env{multlined}.
+% \begin{macrocode}
+\def\EQ_multlined_left:n #1{%
+ \mathrlap{#1}&\hskip.5\multlinedwidth&&\hskip.5\multlinedwidth&&}
+\def\EQ_multlined_center:n #1{%
+ &\hskip.5\multlinedwidth&&\mathclap{#1}\hskip.5\multlinedwidth&&}
+\def\EQ_multlined_right:n #1{%
+ &\hskip.5\multlinedwidth&&\hskip.5\multlinedwidth&&\mathllap{#1}}
+% \end{macrocode}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+%
+% End of bad multline.
+% \begin{macrocode}
+}
+% \end{macrocode}
+%
+% \subsubsection{The Code for \env{flalign} \& \env{multline}}
+% \changes{v2.15}{2017/03/31}{added MH\_ prefix}
+% \changes{v2.15}{2017/03/31}{added EQ\_ prefix}
+% \begin{macrocode}
+\EQ_new_equationtype:nnn {EmphEqEnv}{flalign}{5}
+\EQ_new_equationtype:nnn {EmphEqEnv}{flalign*}{6}
+\def\EQ_modify_linewidth_v:{%
+ \setlength{\l_EQ_temp_linewidth_dim}{%
+ \linewidth-\l_EQ_totalwidth_dim-\widthof{\EQ_outerbox:{}}}%
+ \MH_if_dim:w \g_EQ_widesttag_dim>\z@
+ \MH_if_boolean:nTF {multline}{
+ \EQ_multline_modify_linewidth_tag:
+ }{
+ \MH_if_case:w \l_EQ_alignmentmode_int
+ \MH_or: % #1:
+ \MH_addtolength:dn \l_EQ_temp_linewidth_dim
+ {-\g_EQ_widesttag_dim-\mintagsep}
+ \MH_or: % #2:
+ \MH_addtolength:dn \l_EQ_temp_linewidth_dim
+ {-\g_EQ_widesttag_dim-\mintagsep}%
+ \MH_or: % #3:
+ \MH_addtolength:dn \l_EQ_temp_linewidth_dim
+ {-\g_EQ_widesttag_dim-\mintagsep}
+ \MH_or: % #4:
+ \MH_addtolength:dn \l_EQ_temp_linewidth_dim
+ {-\@mathmargin}
+ \MH_fi:
+ }
+ \MH_else:
+ \MH_if_boolean:nT {multline}{
+ \EQ_multline_modify_linewidth_notag:
+ }
+ \MH_fi:}
+\def\EQ_modify_linewidth_vi:{\EQ_modify_linewidth_v:}
+\def\EQ_find_displaywidth_v:{%
+ \MH_setlength:dn
+ \l_EQ_displaywidth_dim{\l_EQ_temp_linewidth_dim}}
+\def\EQ_find_displaywidth_vi:{\EQ_find_displaywidth_v:}
+\def\EQ_place_display_v:{%
+ \MH_if_case:w \l_EQ_alignmentmode_int
+ \MH_or: % #1:
+ \MH_if_dim:w \g_EQ_widesttag_dim>\z@
+ \MH_if_boolean:nTF {multline}{
+ \rlap{\kern\multlinegap \kern-.5\linewidth
+ \EQ_use_mathdisplay:
+ \kern-\mintagsep \kern\multlinetaggap}%
+ }{
+ \rlap{\kern-.5\linewidth\EQ_use_mathdisplay:}%
+ }
+ \MH_else:
+ \EQ_use_mathdisplay:
+ \MH_fi:
+ \MH_or: % #2:
+ \MH_if_dim:w \g_EQ_widesttag_dim>\z@
+ \MH_if_boolean:nTF {multline}{
+ \llap{\kern-\mintagsep \kern\multlinetaggap
+ \EQ_use_mathdisplay:
+ \kern\multlinegap \kern-.5\linewidth}%
+ }{
+ \llap{\EQ_use_mathdisplay:\kern-.5\linewidth}%
+ }
+ \MH_else:
+ \EQ_use_mathdisplay:
+ \MH_fi:
+ \MH_or: % #3:
+ \MH_if_boolean:nF {multline}{ \kern-\multlinetaggap }
+ \EQ_use_mathdisplay:
+ \MH_or: % #4:
+ \MH_if_dim:w \g_EQ_widesttag_dim>\z@
+ \MH_else:
+ \MH_if_boolean:nF {multline}{ \kern-\@mathmargin }
+ \MH_fi:
+ \EQ_use_mathdisplay:
+ \MH_fi:
+ \MH_set_boolean_F:n {multline}}
+\def\EQ_place_display_vi:{\EQ_place_display_v:}
+% \end{macrocode}
+%
+% \subsection{The \env{alignat} Family}
+% This is just as easy as \env{gather}. Actually only one line is
+% different from \env{gather}!
+% \changes{v2.15}{2017/03/31}{added MH\_ prefixes}
+% \changes{v2.15}{2017/03/31}{added EQ\_ prefix}
+% \begin{macrocode}
+\EQ_new_equationtype_arg:nnn {EmphEqEnv}{alignat}{7}
+\EQ_new_equationtype_arg:nnn {EmphEqEnv}{alignat*}{8}
+\def\EQ_modify_linewidth_vii:{}
+\def\EQ_modify_linewidth_viii:{\EQ_modify_linewidth_vii:}
+\def\EQ_find_displaywidth_vii:{%
+ \MH_setlength:dn \l_EQ_displaywidth_dim{\totwidth@}%
+ \MH_if_case:w \l_EQ_alignmentmode_int
+ \MH_or: % #1:
+ \MH_or: % #2:
+ \MH_or: % #3:
+ \MH_addtolength:dn \l_EQ_displaywidth_dim{-\eqnshift@}%
+ \MH_or: % #4:
+ \MH_addtolength:dn \l_EQ_displaywidth_dim{-\@mathmargin}%
+ \MH_fi:}
+\def\EQ_find_displaywidth_viii:{\EQ_find_displaywidth_vii:}
+\def\EQ_place_display_vii:{\EQ_use_mathdisplay:}
+\def\EQ_place_display_viii:{\EQ_place_display_vii:}
+% \end{macrocode}
+% \section{The User Interface}
+% The different keys.
+% \begin{macrocode}
+\define@key{\EQ_options_name:}{box}{\def\EQ_outerbox:{#1}}
+\MH_keyval_alias:nnn {\EQ_options_name:}{outerbox}{box}
+% \end{macrocode}
+% \begin{macro}{\EQ_marginbox:nn}
+% The margin box. Uwe Siart wanted it.
+% \begin{macrocode}
+\def\EQ_marginbox:nn #1#2{%
+ \settowidth{\@tempdima}{#2}%
+ \makebox[\z@]{\hspace{\@tempdima}#1{#2}}}
+\define@key{\EQ_options_name:}{marginbox}{%
+ \def\EQ_outerbox:{\EQ_marginbox:nn {#1}}}
+% \end{macrocode}
+% \end{macro}
+% Almost self-explanatory, isn't?
+% \begin{macrocode}
+\define@key{\EQ_options_name:}{innerbox}{\def\EQ_innerbox:{#1}}
+\define@key{\EQ_options_name:}{left}{\def\EQ_left_side:{#1}}
+\define@key{\EQ_options_name:}{right}{\def\EQ_right_side:{#1}}
+% \end{macrocode}
+% \begin{macro}{\empheqset}
+% Just a user friendly way to set keys. Uwe Siart again.
+% \begin{macrocode}
+\newcommand*\empheqset[1]{\setkeys{\EQ_options_name:}{#1}}
+% \end{macrocode}
+% \end{macro}
+% The old keys from the original version are no longer supported.
+% Instead they give an error message (RM):
+% \begin{macrocode}
+\def\EQ_upgrade_error:{%
+ \PackageError{empheq}
+ {You~ are~ using~ the~ old~ syntax!\MessageBreak
+ The~ `empheq'~ environment~ takes~ a~ mandatory~
+ argument~ now.\MessageBreak
+ You~ need~ to~ exit~ and~ change~ your~ source~ file.}
+ {It~ won't~ work,~ trust me.~ Press~ `X'~ now.}}
+\define@key{\EQ_options_name:}{boxtype}{\EQ_upgrade_error:}
+\define@key{\EQ_options_name:}{Left}{\EQ_upgrade_error:}
+\define@key{\EQ_options_name:}{Right}{\EQ_upgrade_error:}
+% \end{macrocode}
+% \begin{macro}{\empheqeqrefbase}
+% This macro is used to locally redefine \cs{eqref}, it will just
+% point at \cs{textup}, the user can then redefine it if nessary.
+% \begin{macrocode}
+\let\empheqeqrefbase\textup
+% \end{macrocode}
+%
+% \end{macro}
+% \begin{environment}{empheq}
+% The environment.
+% \begin{macrocode}
+\newenvironment{empheq}[2][]{%
+% \end{macrocode}
+% \begin{macro}{\EQ_saved_maketag}
+% \changes{v2.13}{2013/02/11}{Fixing \cs{eqref}}
+% We start with a small caveat: During processing \cs{maketag@@@} is
+% redefined, enabling us to separate math and eqn numbers. But the
+% user may have used \cs{eqref} in a comment in the math, and
+% \cs{eqref} use \cs{maketag@@@}\dots{} First save the original \cs{maketag@@@}:
+% \begin{macrocode}
+ \MH_let:NwN\EQ_saved_maketag: \maketag@@@%
+% \end{macrocode}
+% Next redefine \cs{eqref} to locally bring back the original:
+% \begin{macrocode}
+ \renewcommand\eqref[1]{\empheqeqrefbase{%
+ \let\maketag@@@\EQ_saved_maketag:%
+ \tagform@{\ref{##1}}}}
+ \setkeys{EmphEqEnv}{#2}\setkeys{\EQ_options_name:}{#1}%
+ \EmphEqMainEnv}{\endEmphEqMainEnv}
+% \end{macrocode}
+% \end{macro}
+% \end{environment}
+% \begin{macro}{\DeclareLeftDelimiter}
+% \begin{macro}{\EQ_declare_left_delimiter:nnnn}
+% \begin{macro}{\DeclareRightDelimiter}
+% \begin{macro}{\EQ_declare_right_delimiter:nnnn}
+% The delimiter interface.
+% \changes{v0.7}{2003/05/11}{Introduced delimiter interface}
+% \changes{v1.05d}{2004/02/26}{Changed the definitions.}
+% We now use the \cs{EQ_tag_box} to calculate the height but
+% adjust the vertical spacing. All this to prevent running
+% \cs{EQ_typeset_display:} twice.
+% \begin{macrocode}
+\newcommand*\DeclareLeftDelimiter[2][]{%
+ \EQ_declare_left_delimiter:nnnn {#1}{#2}{}{}%
+ \EQ_declare_left_delimiter:nnnn {#1}{#2}{big}{\EQ_innerbox:}}
+\def\EQ_declare_left_delimiter:nnnn #1#2#3#4{%
+ \@namedef{empheq#3\expandafter\@gobble\string#2}{%
+ \left#2%
+ \mkern-\thinmuskip
+ \vphantom{#4{\parbox{.5\maxdimen}{%
+ \EQ_predisplay_adjustment:
+ \EQ_prevent_vertical_space:
+ \usebox{\EQ_tag_box}%
+ \EQ_postdisplay_adjustment:}}}%
+ \right.%
+ \kern-\nulldelimiterspace #1}}
+\newcommand*\DeclareRightDelimiter[2][]{%
+ \EQ_declare_right_delimiter:nnnn {#1}{#2}{}{}%
+ \EQ_declare_right_delimiter:nnnn {#1}{#2}{big}{\EQ_innerbox:}}
+\def\EQ_declare_right_delimiter:nnnn #1#2#3#4{%
+ \@namedef{empheq#3\expandafter\@gobble\string#2}{%
+ \kern-\nulldelimiterspace
+ #1%
+ \left.%
+ \vphantom{#4{\parbox{.5\maxdimen}{%
+ \EQ_predisplay_adjustment:
+ \EQ_prevent_vertical_space:
+ \usebox{\EQ_tag_box}%
+ \EQ_postdisplay_adjustment:}}}%
+ \mkern-\thinmuskip
+ \right#2}}
+% \end{macrocode}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% The actual declarations.
+% \begin{macrocode}
+\DeclareLeftDelimiter{\lbrace}
+\DeclareRightDelimiter{\rbrace}
+\DeclareLeftDelimiter{\lbrack}
+\DeclareRightDelimiter{\rbrack}
+\DeclareLeftDelimiter{\langle}
+\DeclareRightDelimiter{\rangle}
+\DeclareLeftDelimiter{\lvert}
+\DeclareRightDelimiter{\rvert}
+\DeclareLeftDelimiter{\lVert}
+\DeclareRightDelimiter{\rVert}
+\DeclareLeftDelimiter{\lfloor}
+\DeclareRightDelimiter{\rfloor}
+\DeclareLeftDelimiter{\lceil}
+\DeclareRightDelimiter{\rceil}
+\DeclareLeftDelimiter{\lparen}
+\DeclareRightDelimiter{\rparen}
+% \end{macrocode}
+%
+% \section{Support for Other Packages}
+%
+% Support for various other packages is found in this section. The
+% \pkg{ntheorem} package now has a dedicated section (\S
+% \vref{sec:impl:ntheorem}), as it now a module.
+%
+%
+%
+% \begin{macrocode}
+\def\EQ_typeset_tags_hook:{}
+\def\EQ_typeset_display_hook:{}
+% \end{macrocode}
+%
+% \subsection{\pkg{showkeys}}
+% \changes{v2.10}{2004/07/26}{Added support for the \pkg{showkeys}
+% package}
+% We just disable \pkg{showkeys} when typesetting the labels.
+% \begin{macrocode}
+\g@addto@macro\EQ_typeset_display_hook:{
+ \def\SK@@label#1>#2\SK@{}
+}
+% \end{macrocode}
+% If we overload the \AmS{} equation types we should still be able
+% to use \pkg{showkeys}.
+% \begin{macrocode}
+\@namedef{SK@AmSequation} {\SK@equationtrue}
+\@namedef{SK@AmSequation*} {\SK@equationtrue}
+\@namedef{SK@AmSalign} {\SK@equationtrue}
+\@namedef{SK@AmSalign*} {\SK@equationtrue}
+\@namedef{SK@AmSalignat} {\SK@equationtrue}
+\@namedef{SK@AmSalignat*} {\SK@equationtrue}
+\@namedef{SK@AmSgather} {\SK@equationtrue}
+\@namedef{SK@AmSgather*} {\SK@equationtrue}
+\@namedef{SK@AmSmultline} {\SK@equationtrue}
+\@namedef{SK@AmSmultline*} {\SK@equationtrue}
+\@namedef{SK@AmSflalign} {\SK@equationtrue}
+\@namedef{SK@AmSflalign*} {\SK@equationtrue}
+% \end{macrocode}
+%
+% \subsection{\pkg{hyperref}}
+%
+%
+% A little trick to avoid the (in)famous warning
+% \begin{verbatim}
+% ! pdfTeX warning (ext4): destination with the same identifier
+% (name{equation.1a }) has been already used, duplicate ignored
+% <to be read again>
+% \endgroup \set@typeset@protect
+% \end{verbatim}
+% This doesn't require \pkg{hyperref}.
+% \begin{macrocode}
+\AtBeginDocument{
+ \@ifundefined{Hy@linkcounter}{\newcount\Hy@linkcounter}{}
+}
+\g@addto@macro\EQ_typeset_display_hook:{
+ \MH_let:NwN \hyper@refstepcounter\@gobble
+}
+% \end{macrocode}
+%
+% \subsection{\pkg{nccmath}}
+%
+% Must disable this macro, otherwise the spacing is wrong.
+% \begin{macrocode}
+ \g@addto@macro\EQ_typeset_display_hook:{
+ \let\NCC@ignorepar\relax
+ }
+ \g@addto@macro\EQ_typeset_tags_hook:{
+ \let\NCC@ignorepar\relax
+ }
+% \end{macrocode}
+%
+% \subsection{Basic \pkg{ntheorem} support}
+% \begin{macrocode}
+\g@addto@macro\EQ_typeset_tags_hook:{
+ \MH_let:NwN\TagsPlusEndmarks\relax
+ \MH_let:NwN\RestoreTags\relax
+}
+\g@addto@macro\EQ_typeset_display_hook:{
+ \MH_let:NwN\TagsPlusEndmarks\relax
+ \MH_let:NwN\RestoreTags\relax
+}
+% \end{macrocode}
+% \subsection{Basic \LaTeXe{} support}
+% Gobble the \cs{ltx@label} and disable most error messages.
+% \changes{v2.00}{2004/04/14}{Make a special label-gobbling command
+% to support \pkg{ntheorem} even more}
+% \begin{macrocode}
+\g@addto@macro\EQ_typeset_display_hook:{
+ \MH_let:NwN\ltx@label\MT_gobblelabel:w
+ \MH_let:NwN \GenericError \@gobblefour
+}
+% \end{macrode}
+% \subsection{Support for \pkg{fancybox}}
+% First check if \pkg{fancybox} is loaded and then see if
+% \pkg{color} is loaded. If the latter is loaded we define
+% \texttt{shadowcolor} else we let \cs{textcolor} gobble the first
+% argument (the color).
+% \begin{macrocode}
+\AtBeginDocument{%
+ \@ifpackageloaded{fancybox}{%
+ \@ifpackageloaded{color}{\definecolor{shadowcolor}{rgb}{0,0,0}}%
+ {\newcommand*\textcolor[2]{\normalcolor{#2}}}%
+% \end{macrocode}
+% \begin{macro}{\shadowbox}
+% Then define the \cs{shadowbox}. Use a conditional to let us know if
+% it was starred or not.
+% \begin{macrocode}
+ \MH_new_boolean:n {shadowbox_star}
+ \renewcommand*\shadowbox{%
+ \@ifstar{\MH_set_boolean_T:n {shadowbox_star}
+ \VerbBox\EQ_shadowbox:n}
+ {\MH_set_boolean_F:n {shadowbox_star}
+ \VerbBox\EQ_shadowbox:n}}
+% \end{macrocode}
+% \end{macro}
+% \begin{macro}{\EQ_shadowbox:n}
+% The internal version of \cs{shadowbox}.
+% \begin{macrocode}
+ \def\EQ_shadowbox:n #1{%
+ \sbox{\z@}{\fbox{#1}}%
+% \end{macrocode}
+% The difference between the two versions is that \cs{shadowbox}
+% aligns the bottom of the shadow at the baseline whereas
+% \cs{shadowbox*} aligns the argument on the baseline, so for the
+% starred version the trick is to raise the box by the depth of the
+% argument plus the size of the shadow.
+% \begin{macrocode}
+ \MH_if_boolean:nTF {shadowbox_star}{
+ \MH_setlength:dn \@tempdima{\z@}%
+ }{
+ \MH_setlength:dn \@tempdima{\dp\z@+\shadowsize}%
+ }
+ \raisebox{\@tempdima}{%
+ \makebox[\z@][l]{%
+ \kern\shadowsize \textcolor{shadowcolor}{%
+ \rule[-\dp\z@-\shadowsize]{\wd\z@}{\shadowsize}}%
+ }%
+ \usebox\z@\textcolor{shadowcolor}{%
+ \rule[-\dp\z@-\shadowsize]{\shadowsize}{\ht\z@+\dp\z@}}}}%
+ }{}%
+}
+% \end{macrocode}
+% \end{macro}
+%
+%
+% \section{The \opt{overload} option}
+% \begin{macro}{\EQ_overload_ams_begin:n}
+% A helper for saving snapshots:
+% \begin{macrocode}
+\def\EQ_overload_ams_begin:n #1{%
+ \MH_let:cc {AmS#1}{#1}%
+ \MH_let:cc {AmS#1*}{#1*}
+}
+% \end{macrocode}
+% \end{macro}
+% \begin{macro}{\EQ_overload_ams_end:n}
+% All but one of the aligned variations must be set to the original
+% \begin{macrocode}
+\def\EQ_overload_ams_end:n #1{%
+ \MH_let:cc {endAmS#1}{endAmSalign}
+}
+% \end{macrocode}
+% \end{macro}
+% \begin{environment}{AmSequation}
+% \begin{environment}{AmSequation*}
+% \changes{v2.10}{2004/07/26}{Moved redefinitions to
+% \cs{AtBeginDocument} to give \pkg{hyperref} a chance}
+% Before we start redefining environments, we'll need a basic
+% \env{equation} environment that is not redefined.
+% \begin{macrocode}
+\AtBeginDocument{
+ \EQ_overload_ams_begin:n {equation}
+ \MH_let:cc {endAmSequation}{endequation}
+ \MH_let:cc {endAmSequation*}{endequation*}
+% \end{macrocode}
+% \end{environment}
+% \end{environment}
+% Lets start overloading.
+% \begin{macrocode}
+ \MH_if_boolean:nT {overload_ams}{
+% \end{macrocode}
+% \begin{macro}{\EQ_renew_ams_environment:n}
+% \changes{v2.10}{2004/07/26}{Disallow spaces before the optional
+% argument.}
+% Then we can renew the environments. Syntax is
+% \cs{begin}\marg{\AmS{} env}\oarg{markup}. Exception to this rule
+% is \env{alignat} (see further down).
+% \begin{macrocode}
+ \def\EQ_renew_ams_environment:n #1{%
+ \MaybeMHPrecedingSpacesOff
+ \renewenvironment{#1}[1][]{%
+% \end{macrocode}
+% \changes{v2.13}{2013/02/11}{Adding \cs{eqref} fix}
+% Remember to add \cs{eqref} fix to the overloaded envs as well.
+% \begin{macrocode}
+ \MH_let:NwN\EQ_saved_maketag: \maketag@@@%
+ \renewcommand\eqref[1]{\empheqeqrefbase{%
+ \let\maketag@@@\EQ_saved_maketag:%
+ \tagform@{\ref{####1}}}}
+ \setkeys{\EQ_options_name:}{##1}%
+ \setkeys{EmphEqEnv}{AmS#1}%
+ \EmphEqMainEnv
+ }{\endEmphEqMainEnv}
+ \MHPrecedingSpacesOn
+ }
+% \end{macrocode}
+% \end{macro}
+% Add equation type: \emph{Must} have same number as original type!
+% \begin{macrocode}
+ \EQ_new_equationtype:nnn {EmphEqEnv}{AmSalign} {1}
+ \EQ_new_equationtype:nnn {EmphEqEnv}{AmSalign*} {2}
+ \EQ_new_equationtype:nnn {EmphEqEnv}{AmSgather} {3}
+ \EQ_new_equationtype:nnn {EmphEqEnv}{AmSgather*} {4}
+ \EQ_new_equationtype:nnn {EmphEqEnv}{AmSflalign} {5}
+ \EQ_new_equationtype:nnn {EmphEqEnv}{AmSflalign*}{6}
+ \EQ_new_equationtype_arg:nnn {EmphEqEnv}{AmSalignat} {7}
+ \EQ_new_equationtype_arg:nnn {EmphEqEnv}{AmSalignat*}{8}
+% \end{macrocode}
+% Then overload by saving the snapshots of the old definitions. We
+% do it via \cs{AtBeginDocument}, because a package like
+% \pkg{ntheorem} redefines \cs{endgather} etc.
+%
+% First the \env{gather} and \env{equation} families.
+% \begin{macrocode}
+ \EQ_overload_ams_begin:n {gather}
+ \MH_let:cc {endAmSgather}{endgather}
+ \MH_let:cc {endAmSgather*}{endAmSgather}
+ \EQ_renew_ams_environment:n {gather}
+ \EQ_renew_ams_environment:n {gather*}
+ \EQ_overload_ams_begin:n {equation}
+ \MH_keyval_alias_with_addon:nnnn {EmphEqEnv} {AmSequation}
+ {AmSgather}{\EQ_disable_linebreak:}
+ \MH_keyval_alias_with_addon:nnnn {EmphEqEnv}{AmSequation*}
+ {AmSgather*}{\EQ_disable_linebreak:}
+ \EQ_renew_ams_environment:n {equation}
+ \EQ_renew_ams_environment:n {equation*}
+% \end{macrocode}
+% The \env{align} family.
+% \begin{macrocode}
+ \EQ_overload_ams_begin:n {align}
+ \MH_let:cc {endAmSalign}{endalign}
+ \EQ_overload_ams_end:n {align*}
+ \EQ_renew_ams_environment:n {align}
+ \EQ_renew_ams_environment:n {align*}
+% \end{macrocode}
+% The \env{flalign} family.
+% \begin{macrocode}
+ \EQ_overload_ams_begin:n {flalign}
+ \EQ_overload_ams_end:n {flalign}
+ \EQ_overload_ams_end:n {flalign*}
+ \EQ_renew_ams_environment:n {flalign}
+ \EQ_renew_ams_environment:n {flalign*}
+% \end{macrocode}
+% The \env{alignat} family.
+% \begin{macrocode}
+ \EQ_overload_ams_begin:n {alignat}
+ \EQ_overload_ams_end:n {alignat}
+ \EQ_overload_ams_end:n {alignat*}
+% \end{macrocode}
+% \begin{macro}{\EQ_alignat_optional:w}
+% \changes{v2.10}{2004/07/26}{Disallow spaces before the optional
+% argument.}
+% \begin{environment}{alignat}
+% \begin{environment}{alignat*}
+% This one is a little different because we want a syntax like
+% \cs{begin\{alignat\}}\linebreak[0]\marg{cols}\oarg{markup}, so we
+% have to make a separate command for testing the trailing optional
+% argument.
+% \begin{macrocode}
+ \MaybeMHPrecedingSpacesOff
+ \newcommand*\EQ_alignat_optional:w [1][]{
+ \setkeys{\EQ_options_name:}{#1}
+ \EmphEqMainEnv}
+ \MHPrecedingSpacesOn
+% \end{macrocode}
+% \changes{v2.13}{2013/02/11}{Adding \cs{eqref} fix}
+% Remember to add \cs{eqref} fix to the overloaded envs as well.
+% \begin{macrocode}
+ \renewenvironment{alignat}[1]{
+ \MH_let:NwN\EQ_saved_maketag: \maketag@@@%
+ \renewcommand\eqref[1]{\empheqeqrefbase{%
+ \let\maketag@@@\EQ_saved_maketag:%
+ \tagform@{\ref{##1}}}}
+ \setkeys{EmphEqEnv}{AmSalignat=#1}
+ \EQ_alignat_optional:w}
+ {\endEmphEqMainEnv}
+ \renewenvironment{alignat*}[1]{%
+ \MH_let:NwN\EQ_saved_maketag: \maketag@@@%
+ \renewcommand\eqref[1]{\empheqeqrefbase{%
+ \let\maketag@@@\EQ_saved_maketag:%
+ \tagform@{\ref{##1}}}}
+ \setkeys{EmphEqEnv}{AmSalignat*=#1}%
+ \EQ_alignat_optional:w}%
+ {\endEmphEqMainEnv}
+% \end{macrocode}
+% \end{environment}
+% \end{environment}
+% \end{macro}
+% \begin{macro}{\[}
+% \begin{macro}{\]}
+% Then restore \cs{[} and \cs{]}.
+% \begin{macrocode}
+ \DeclareRobustCommand{\[}{\begin{AmSequation*}}
+ \DeclareRobustCommand{\]}{\end{AmSequation*}}
+% \end{macrocode}
+% \end{macro}
+% \end{macro}
+% Finally a string of aliases so that we can blend \opt{overload}
+% syntax with regular syntax.
+% \begin{macrocode}
+ \MH_keyval_alias:nnn {EmphEqEnv}{gather} {AmSgather}
+ \MH_keyval_alias:nnn {EmphEqEnv}{gather*} {AmSgather*}
+ \MH_keyval_alias:nnn {EmphEqEnv}{align} {AmSalign}
+ \MH_keyval_alias:nnn {EmphEqEnv}{align*} {AmSalign*}
+ \MH_keyval_alias:nnn {EmphEqEnv}{flalign} {AmSflalign}
+ \MH_keyval_alias:nnn {EmphEqEnv}{flalign*}{AmSflalign*}
+ \MH_keyval_alias:nnn {EmphEqEnv}{alignat} {AmSalignat}
+ \MH_keyval_alias:nnn {EmphEqEnv}{alignat*}{AmSalignat*}
+ \MH_if_boolean:nT {good_mult}{
+ \EQ_overload_ams_begin:n {multline}
+ \MH_let:cc {endAmSmultline}{endmultline}
+ \MH_let:cc {endAmSmultline*}{endAmSmultline}
+ \MH_keyval_alias_with_addon:nnnn {EmphEqEnv}{AmSmultline}
+ {AmSgather}{\MH_set_boolean_T:n {outer_mult}}
+ \MH_keyval_alias_with_addon:nnnn {EmphEqEnv}{AmSmultline*}
+ {AmSgather*}{\MH_set_boolean_T:n {outer_mult}}
+ \EQ_renew_ams_environment:n {multline}
+ \EQ_renew_ams_environment:n {multline*}
+ }
+% \end{macrocode}
+% End of \opt{overload}. We go into \opt{overload2}\ldots This
+% is dangerous, but will probably work. The macro \cs{[} is now
+% fragile except if you're running \eTeX{} as engine for \LaTeX.
+% \begin{macrocode}
+ \MH_if_boolean:nT {overload_amsII}{
+ \MH_protected:\def\[#1\]{\begin{equation*} #1\end{equation*}}
+% \end{macrocode}
+% Then we add the \env{multline} family.
+% \begin{macrocode}
+ \MH_if_boolean:nF {good_mult}{
+ \EQ_overload_ams_begin:n {multline}
+ \MH_keyval_alias_with_addon:nnnn {EmphEqEnv}{AmSmultline}
+ {AmSflalign}{\MH_set_boolean_T:n {multline}}
+ \MH_keyval_alias_with_addon:nnnn {EmphEqEnv}{AmSmultline*}
+ {AmSflalign*}{\MH_set_boolean_T:n {multline}}
+ \EQ_renew_ams_environment:n {multline}
+ \EQ_renew_ams_environment:n {multline*}
+ }
+ }
+ }
+}
+% \end{macrocode}
+% End of \cs{AtBeginDocument} additions.
+%
+% If we need to load \pkg{ntheorem} support this is where it
+% happens. Otherwise we just stop here.
+% \begin{macrocode}
+\if\g@EQ@ntheorem@support@bool
+\else
+ \expandafter\endinput
+\fi
+% \end{macrocode}
+%
+% \section{Patching Things up for \pkg{ntheorem}}
+% \label{sec:impl:ntheorem}
+%
+% Although \pkg{ntheorem} provides a lot of useful features it
+% does however ruin things in the \pkg{amsmath} environments,
+% because of the alternative syntax of its \cs{label} command.
+% \begin{macro}{\EQ_default_raisetag}
+% \begin{macro}{\raisetag}
+% \begin{macro}{\raise@tag}
+% \changes{v1.06}{2004/03/13}{Fixed \pkg{amsmath} bug \# 3624}
+% This corrects a bug\footnote{See
+% \url{http://www.latex-project.org/cgi-bin/ltxbugs2html?pr=amslatex/3624}}
+% in \pkg{amsmath} that could cause tags to cross page boundaries
+% without warning. The \cs{nobreak} is the addition.
+% \changes{v2.15}{2017/03/31}{added MH\_ prefix}
+% \changes{v2.15}{2017/03/31}{added EQ\_ prefix}
+% \begin{macrocode}
+\def\EQ_default_raisetag:{\nobreak}
+\def\raisetag#1{
+ \skip@#1\relax
+ \xdef\raise@tag{\EQ_default_raisetag:
+ \vskip\iftagsleft@\MH_else:-\MH_fi:\the\skip@\relax}}
+\MH_let:NwN\raise@tag\EQ_default_raisetag:
+% \end{macrocode}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+% \begin{macro}{\displ@y@}
+% Patch it so that it doesn't set \cs{raise@tag} to \cs{@empty}
+% globally.
+% \changes{v2.15}{2017/03/31}{added EQ\_ prefix}
+% \begin{macrocode}
+\def\displ@y@{\@display@init{%
+ \global\column@\z@ \global\dspbrk@lvl\m@ne
+ \global\tag@false \global\MH_let:NwN\raise@tag\EQ_default_raisetag:
+}}
+% \end{macrocode}
+% \end{macro}
+% \begin{macro}{\MT_gobblelabel:w}
+% \changes{v2.00}{2004/04/14}{Added \cs{MT_gobblelabel:w}}
+% \begin{macro}{\EQ_gobbleoptional:w}
+% \changes{v2.00}{2004/04/14}{Added \cs{EQ_gobbleoptional:w}}
+% Macros for gobbling the `unorthodox' \cs{label} syntax of
+% \pkg{ntheorem}. Also overrides the normal \cs{label}-gobbling.
+% \begin{macrocode}
+\def\MT_gobblelabel:w #1{\@ifnextchar[{\EQ_gobbleoptional:w}{}}
+\def\EQ_gobbleoptional:w [#1]{}
+% \end{macrocode}
+% \end{macro}
+% \end{macro}
+% \subsection{Fixing \env{gather}}
+% The measuring phase of \env{gather} is patched by letting
+% \cs{label} be a copy of the internal \cs{MT_gobblelabel:w}
+% instead of \cs{@gobble}.
+% \begin{macrocode}
+\def\gmeasure@#1{%
+ \begingroup
+ \measuring@true
+ \totwidth@\z@
+ \global\MH_let:NwN\tag@lengths\@empty
+ \savecounters@
+ \setbox\@ne\vbox{%
+ \everycr{\noalign{\global\tag@false
+% \end{macrocode}
+% The patch for \cs{raisetag}.
+% \changes{v2.15}{2017/03/31}{added EQ\_ prefix}
+% \begin{macrocode}
+ \global\MH_let:NwN\raise@tag\EQ_default_raisetag:
+ \global\column@\z@}}%
+% \end{macrocode}
+% The patch for \cs{label}.
+% \changes{v2.15}{2017/03/31}{added MH\_ prefix}
+% \begin{macrocode}
+ \MH_let:NwN\label\MT_gobblelabel:w
+ \halign{%
+ \setboxz@h{$\m@th\displaystyle{##}$}%
+ \MH_if_dim:w \wdz@>\totwidth@
+ \global\totwidth@\wdz@
+ \MH_fi:
+ &\setboxz@h{\strut@{##}}%
+ \savetaglength@
+ \crcr
+ #1%
+ \math@cr@@@
+ }%
+ }%
+ \restorecounters@
+ \if@fleqn
+ \global\advance\totwidth@\@mathmargin
+ \MH_fi:
+ \iftagsleft@
+ \MH_if_dim:w \totwidth@>\displaywidth
+ \global\MH_let:NwN\gdisplaywidth@\totwidth@
+ \MH_else:
+ \global\MH_let:NwN\gdisplaywidth@\displaywidth
+ \MH_fi:
+ \MH_fi:
+ \endgroup
+}
+% \end{macrocode}
+%
+% \subsection{Fixing \env{align}, \env{alignat}, and \env{flalign}}
+% Just copy the definition of \cs{measure@} from \pkg{amsmath} and
+% change one (1!) line like we did for \env{gather}.
+% \begin{macrocode}
+\def\measure@#1{%
+ \begingroup
+ \measuring@true
+ \global\eqnshift@\z@
+ \global\alignsep@\z@
+ \global\MH_let:NwN\tag@lengths\@empty
+ \global\MH_let:NwN\field@lengths\@empty
+ \savecounters@
+ \global\setbox0\vbox{%
+ \MH_let:NwN\math@cr@@@\math@cr@@@align@measure
+ \everycr{\noalign{\global\tag@false
+% \end{macrocode}
+% The patch for \cs{raisetag}.
+% \changes{v2.15}{2017/03/31}{added EQ\_ prefix}
+% \begin{macrocode}
+ \global\MH_let:NwN\raise@tag\EQ_default_raisetag:
+ \global\column@\z@}}%
+% \end{macrocode}
+% The patch.
+% \changes{v2.15}{2017/03/31}{added MH\_ prefix}
+% \begin{macrocode}
+ \MH_let:NwN\label\MT_gobblelabel:w
+ \global\row@\z@
+ \tabskip\z@
+ \halign{\span\align@preamble\crcr
+ #1%
+ \math@cr@@@
+ \global\column@\z@
+ \add@amps\maxfields@\cr
+ }%
+ }%
+ \restorecounters@
+ \ifodd\maxfields@
+ \global\advance\maxfields@\@ne
+ \MH_fi:
+ \MH_if_num:w \xatlevel@=\tw@
+ \MH_if_num:w \maxfields@<\thr@@
+ \MH_let:NwN\xatlevel@\z@
+ \MH_fi:
+ \MH_fi:
+ \setbox\z@\vbox{%
+ \unvbox\z@ \unpenalty \global\setbox\@ne\lastbox
+ }%
+ \global\totwidth@\wd\@ne
+ \if@fleqn \global\advance\totwidth@\@mathmargin \MH_fi:
+ \global\MH_let:NwN\maxcolumn@widths\@empty
+ \begingroup
+ \MH_let:NwN\MH_or:\relax
+ \loop
+ \global\setbox\@ne\hbox{%
+ \unhbox\@ne \unskip \global\setbox\thr@@\lastbox
+ }%
+ \ifhbox\thr@@
+ \xdef\maxcolumn@widths{ \MH_or: \the\wd\thr@@ \maxcolumn@widths}%
+ \repeat
+ \endgroup
+ \dimen@\displaywidth
+ \advance\dimen@-\totwidth@
+ \MH_if_case:w \xatlevel@
+ \global\alignsep@\z@
+ \MH_let:NwN\minalignsep\z@
+ \@tempcntb\z@
+ \if@fleqn
+ \@tempcnta\@ne
+ \global\eqnshift@\@mathmargin
+ \MH_else:
+ \@tempcnta\tw@
+ \global\eqnshift@\dimen@
+ \global\divide\eqnshift@\@tempcnta
+ \MH_fi:
+ \MH_or:
+ \@tempcntb\maxfields@
+ \divide\@tempcntb\tw@
+ \@tempcnta\@tempcntb
+ \advance\@tempcntb\m@ne
+ \if@fleqn
+ \global\eqnshift@\@mathmargin
+ \global\alignsep@\dimen@
+ \global\divide\alignsep@\@tempcnta
+ \MH_else:
+ \global\advance\@tempcnta\@ne
+ \global\eqnshift@\dimen@
+ \global\divide\eqnshift@\@tempcnta
+ \global\alignsep@\eqnshift@
+ \MH_fi:
+ \MH_or:
+ \@tempcntb\maxfields@
+ \divide\@tempcntb\tw@
+ \global\advance\@tempcntb\m@ne
+ \global\@tempcnta\@tempcntb
+ \global\eqnshift@\z@
+ \global\alignsep@\dimen@
+ \if@fleqn
+ \global\advance\alignsep@\@mathmargin\relax
+ \MH_fi:
+ \global\divide\alignsep@\@tempcntb
+ \MH_fi:
+ \MH_if_dim:w \alignsep@<\minalignsep\relax
+ \global\alignsep@\minalignsep\relax
+ \MH_if_dim:w \eqnshift@>\z@
+ \if@fleqn\MH_else:
+ \global\eqnshift@\displaywidth
+ \global\advance\eqnshift@-\totwidth@
+ \global\advance\eqnshift@-\@tempcntb\alignsep@
+ \global\divide\eqnshift@\tw@
+ \MH_fi:
+ \MH_fi:
+ \MH_fi:
+ \MH_if_dim:w \eqnshift@<\z@
+ \global\eqnshift@\z@
+ \MH_fi:
+ \calc@shift@align
+ \global\tagshift@\totwidth@
+ \global\advance\tagshift@\@tempcntb\alignsep@
+ \if@fleqn
+ \MH_if_num:w \xatlevel@=\tw@
+ \global\advance\tagshift@-\@mathmargin\relax
+ \MH_fi:
+ \MH_else:
+ \global\advance\tagshift@\eqnshift@
+ \MH_fi:
+ \iftagsleft@ \MH_else:
+ \global\advance\tagshift@-\displaywidth
+ \MH_fi:
+ \dimen@\minalignsep\relax
+ \global\advance\totwidth@\@tempcntb\dimen@
+ \MH_if_dim:w \totwidth@>\displaywidth
+ \global\MH_let:NwN\displaywidth@\totwidth@
+ \MH_else:
+ \global\MH_let:NwN\displaywidth@\displaywidth
+ \MH_fi:
+ \endgroup
+}
+% \end{macrocode}
+%
+% \subsection{Fixing \env{multline}}
+% \begin{macro}{\EQ_multline_labelhack:n}
+% \begin{macro}{\EQ_multline_labelhack_opt:nw}
+% As \pkg{ntheorem} defines \cs{label} to have the syntax
+% \cs{label}\marg{label}\oarg{name}, we need to make sure that it
+% has that inside \env{multline} as well. Especially since the
+% measuring phase of \env{multline} \emph{defines} \cs{label} as
+% well.
+% \begin{macrocode}
+\def\EQ_multline_labelhack:n #1{%
+ \begingroup\measuring@false\label@in@display{#1}\endgroup}
+\def\EQ_multline_labelhack_opt:nw #1[#2]{%
+ \begingroup\measuring@false\label@in@display{#1}[#2]\endgroup}
+% \end{macrocode}
+% \end{macro}
+% \end{macro}
+% \begin{macro}{\multline@}
+% Patching \env{multline} takes more work because it does its
+% processing differently, so we need to apply hooks twice: One in
+% the measuring phase \cs{mmeasure@} where \cs{label} is redefined
+% and one in the production phase \cs{multline@} where the ordinary
+% \cs{label} is gobbled (we use \cs{MT_gobblelabel:w} as
+% before). First the typesetting phase.
+% \begin{macrocode}
+\def\multline@#1{%
+ \Let@
+ \@display@init{\global\advance\row@\@ne \global\dspbrk@lvl\m@ne}%
+ \chardef\dspbrk@context\z@
+ \restore@math@cr
+ \MH_let:NwN\tag\tag@in@align
+% \end{macrocode}
+% The patch for \cs{raisetag}.
+% \changes{v2.15}{2017/03/31}{added EQ\_ prefix}
+% \begin{macrocode}
+ \global\tag@false \global\MH_let:NwN\raise@tag\EQ_default_raisetag:
+ \mmeasure@{#1}%
+ \MH_let:NwN\tag\gobble@tag
+% \end{macrocode}
+% The patch.
+% \changes{v2.15}{2017/03/31}{added MH\_ prefixes}
+% \begin{macrocode}
+ \MH_let:NwN\label\MT_gobblelabel:w
+ \tabskip \if@fleqn \@mathmargin \MH_else: \z@skip \MH_fi:
+ \totwidth@\displaywidth
+ \if@fleqn
+ \advance\totwidth@-\@mathmargin
+ \MH_fi:
+ \halign\bgroup
+ \hbox to\totwidth@{%
+ \if@fleqn
+ \hskip \@centering \relax
+ \MH_else:
+ \hfil
+ \MH_fi:
+ \strut@
+ $\m@th\displaystyle{}##\endmultline@math%$
+ \hfil
+ }%
+ \crcr
+ \if@fleqn
+ \hskip-\@mathmargin
+ \def\multline@indent{\hskip\@mathmargin}% put it back
+ \MH_else:
+ \hfilneg
+ \def\multline@indent{\hskip\multlinegap}%
+ \MH_fi:
+ \iftagsleft@
+ \iftag@
+ \begingroup
+ \ifshifttag@
+ \rlap{\vbox{%
+ \normalbaselines
+ \hbox{%
+ \strut@
+ \make@display@tag
+ }%
+ \vbox to\lineht@{}%
+ \raise@tag
+ }}%
+ \multline@indent
+ \MH_else:
+ \setbox\z@\hbox{\make@display@tag}%
+ \dimen@\@mathmargin \advance\dimen@-\wd\z@
+ \MH_if_dim:w \dimen@<\multlinetaggap
+ \dimen@\multlinetaggap
+ \MH_fi:
+ \box\z@ \hskip\dimen@\relax
+ \MH_fi:
+ \endgroup
+ \MH_else:
+ \multline@indent
+ \MH_fi:
+ \MH_else:
+ \multline@indent
+ \MH_fi:
+ #1%
+}
+% \end{macrocode}
+% \end{macro}
+% \begin{macro}{\mmeasure@}
+% Then redefine the label correctly.
+% \begin{macrocode}
+\def\mmeasure@#1{%
+ \begingroup
+ \measuring@true
+% \end{macrocode}
+% The patch.
+% \changes{v2.15}{2017/03/31}{added MH\_ prefixes}
+% \begin{macrocode}
+ \def\label##1{%
+ \@ifnextchar[
+ {\EQ_multline_labelhack_opt:nw {##1}}
+ {\EQ_multline_labelhack:n {##1}}
+ }
+ \def\math@cr@@@{\cr}%
+ \MH_let:NwN\shoveleft\@iden \MH_let:NwN\shoveright\@iden
+ \savecounters@
+ \global\row@\z@
+ \setbox\@ne\vbox{%
+ \global\MH_let:NwN\df@tag\@empty
+ \halign{%
+ \setboxz@h{\@lign$\m@th\displaystyle{}##$}%
+ \iftagsleft@
+ \MH_if_num:w \row@=\@ne
+ \global\totwidth@\wdz@
+ \global\lineht@\ht\z@
+ \MH_fi:
+ \MH_else:
+ \global\totwidth@\wdz@
+ \global\lineht@\dp\z@
+ \MH_fi:
+ \crcr
+ #1%
+ \crcr
+ }%
+ }%
+ \MH_if_meaning:NN
+ \df@tag\@empty
+ \MH_else:
+ \global\tag@true
+ \MH_fi:
+ \if@eqnsw\global\tag@true\MH_fi:
+ \iftag@
+ \setboxz@h{%
+ \if@eqnsw
+ \stepcounter{equation}%
+ \tagform@\theequation
+ \MH_else:
+ \df@tag
+ \MH_fi:
+ }%
+ \global\tagwidth@\wdz@
+ \dimen@\totwidth@
+ \advance\dimen@\tagwidth@
+ \advance\dimen@\multlinetaggap
+ \iftagsleft@\MH_else:
+ \if@fleqn
+ \advance\dimen@\@mathmargin
+ \MH_fi:
+ \MH_fi:
+ \MH_if_dim:w \dimen@>\displaywidth
+ \global\shifttag@true
+ \MH_else:
+ \global\shifttag@false
+ \MH_fi:
+ \MH_fi:
+ \restorecounters@
+ \endgroup
+}
+% \end{macrocode}
+% \end{macro}
+%
+% \subsection{Setting End Marks Correctly}
+%
+% \begin{macro}{\mintagvsep}
+% \changes{v1.05}{2004/02/16}{Introduced \cs{mintagvsep} to
+% support the \pkg{ntheorem} package}
+% \begin{macro}{\EQ_only_endmark:}
+% \changes{v1.05}{2004/02/16}{Introduced \cs{EQ_only_endmark:}
+% to support the \pkg{ntheorem} package}
+% \changes{v2.11}{2004/08/03}{Make sure the tag is printed even with the
+% \pkg{mathtools} key \key{showonlyrefs} activated}
+% \begin{macro}{\EQ_tag_plus_endmark:}
+% \changes{v1.05}{2004/02/16}{Introduced
+% \cs{EQ_tag_plus_endmark:} to support the \pkg{ntheorem} package}
+% \pkg{ntheorem} provides an extension to the theorem environment
+% and also supports automatic end of theorem marks. To do so it must
+% redefine parts of the \pkg{amsmath} code and it eventually ends up
+% in problems when the tags and the marker are all typeset at the
+% right margin.
+%
+% There are basically two different possibilities:
+% \begin{center}\hfill
+% \unitlength2.5pt \thicklines
+% \begin{picture}(60,45)(0,25)
+% \put(0,50){\framebox(25,20){Math Height}}
+% \put(0,25){\framebox(25,25){Math Depth}}
+% \put(40,50){\framebox(20,10){Tag height}}
+% \put(40,45){\framebox(20,5){Tag depth}}
+% \put(40,25){\framebox(20,10){Marker}}
+% \put(50,35){\vector(0,1){10}}
+% \put(50,45){\vector(0,-1){10}}
+% \put(50,39){\ \ $\delta>0$}
+% \multiput(26,25)(2,0){7}{\line(1,0){1}}
+% \end{picture} \hfill\vrule\hfill
+% \begin{picture}(60,45)(0,25)
+% \put(0,50){\framebox(25,20){Math Height}}
+% \put(0,40){\framebox(25,10){Math Depth}}
+% \put(40,50){\framebox(20,10){Tag height}}
+% \put(40,45){\framebox(20,5){Tag depth}}
+% \put(40,28){\framebox(20,10){Marker}}
+% \put(50,38){\vector(0,1){7}}
+% \put(50,45){\vector(0,-1){7}}
+% \put(50,40){\ \ $\delta=0$}
+% \end{picture}\hfill\mbox{}
+% \end{center}
+% Let's look closer at this. $h_M$ is the height of the end of
+% theorem marker, which we'll try to place so that its baseline
+% aligns with the bottom of the math display. $d_D$ is the depth of
+% the the last line of the math display including the depth of the
+% surrounding boxes while $d_T$ is the depth of the tag. Also we let
+% $m$ denote the minimum vertical distance between tag and marker
+% and $\delta$ is some positive length that may be added to $m$
+% depending on the other values. To ensure a ``proper'' placement of
+% the marker we simply test the inequality
+% \begin{gather}
+% h_M + d_T + m > d_D \label{eq:tag_placement}
+% \end{gather}
+% If the test is true we're done and just have to set the marker so
+% that the top of it is placed at a distance of $m$ from the bottom
+% of the tag. If it's not we can safely put the marker at the bottom
+% of the display.
+% \begin{macrocode}
+\newlength\mintagvsep
+\mintagvsep5pt
+\def\EQ_only_endmark:{%
+ \MH_set_boolean_T:n {show_manual_tags}
+ \global\tag@true
+ \iftagsleft@
+ \gdef\df@tag{%
+ \hbox to \displaywidth{%
+ \hss
+% \end{macrocode}
+% If the tags are set to the left we just have to lower the end mark
+% by the total depth of the display including its surrounding box.
+% Also we have to take care of \cs{normallineskip}.
+% \changes{v2.15}{2017/03/31}{added MH\_ prefixes}
+% \begin{macrocode}
+ \raisebox{-\normallineskip
+ -\depthof{\EQ_outerbox:{}}
+ -\heightof{\EQ_outerbox:{}}
+ -\dp\EQ_mathdisplay_box
+ -\ht\EQ_mathdisplay_box}
+ [0pt][0pt]{\PotEndMark{\maketag@@@}}}}
+ \MH_else:
+ \gdef\df@tag{%
+ \raisebox{-\dp\EQ_mathdisplay_box
+ -\depthof{\EQ_outerbox:{}}}
+ {\PotEndMark{\maketag@@@}}}%
+ \MH_fi:}
+% \end{macrocode}
+% \changes{v2.00}{2004/04/14}{Fixed another bug in placement of the
+% marker}
+% \changes{v2.10}{2004/07/26}{Fixed horizontal spacing in \env{equation}}
+% In v2.00 I fixed another bug I had overlooked. This is getting
+% embarrassing. The outer box must be taken into account as well.
+% \begin{macrocode}
+\def\EQ_tag_plus_endmark:{%
+ \iftagsleft@
+ \gdef\maketag@@@##1{%
+% \end{macrocode}
+% The following fixes a bug when there is a wide tag in an
+% \env{equation} environment and was reported by Bernard
+% Alfonsi.\footnote{\email{alfonsi@math.u-psud.fr}} The problem
+% arises when \pkg{ntheorem} tries to typeset a tag and an end mark,
+% because the tag is set at the left margin and the end mark at the
+% right margin and \pkg{ntheorem} over-simplifies this and forgets
+% to take the displayed math into account! The right solution is to
+% wrap the tag in a box of the correct size, namely |\widthof{##1}|
+% and then push the end mark to the right margin. Then the displayed
+% math is set correcly.
+% \changes{v2.15}{2017/03/31}{added MH\_ prefixes}
+% \begin{macrocode}
+ \hfuzz\displaywidth
+ \makebox[\widthof{##1}][l]{%
+ \hbox to \displaywidth{\m@th\normalfont##1\hss
+ \raisebox{-\dp\EQ_mathdisplay_box
+ -\depthof{\EQ_outerbox:{}}}
+ {\PotEndMark{\hss}}}}}%
+ \MH_else:
+ \gdef\maketag@@@##1{%
+ \hbox{\m@th\normalfont##1\llap{\hss\PotEndMark{%
+ \settoheight{\@tempdima}{\@nameuse{\InTheoType Symbol}}%
+ \MH_addtolength:dn \@tempdima
+ {\g_EQ_bottomtag_depth_dim+\mintagvsep}
+ \setlength{\@tempdimb}
+ {\g_EQ_bottomrow_depth_dim+\depthof{\EQ_outerbox:{}}
+ +\depthof{\EQ_innerbox:{}}}
+ \MH_if_dim:w \@tempdima>\@tempdimb
+ \MH_setlength:dn \@tempdima{-\dp\EQ_mathdisplay_box
+ -\@tempdima+\g_EQ_bottomrow_depth_dim}
+ \addtolength{\@tempdima}{\depthof{\EQ_innerbox:{}}}%
+ \def\EQ_next:{\raisebox{\@tempdima}}%
+ \MH_else:
+% \end{macrocode}
+% \changes{v1.06}{2004/03/13}{Fixed bug in placement of the marker
+% and re-arranged some code}
+% In v1.06 I fixed a stupid bug regarding placement of the marker.
+% If I had done a proper test I would have discovered it
+% immediately!
+% \changes{v2.15}{2017/03/31}{added MH\_ prefixes}
+% \begin{macrocode}
+ \def\EQ_next:{
+ \raisebox{-\dp\EQ_mathdisplay_box
+ -\depthof{\EQ_outerbox:{}}}}%
+ \MH_fi:\EQ_next:}}}}%
+ \MH_fi:}
+% \end{macrocode}
+% \end{macro}
+% \end{macro}
+% \end{macro}
+%
+% \begin{macro}{\MT_nonumber:}
+% \begin{macrocode}
+\MH_let:NwN \MT_nonumber: \nonumber
+% \end{macrocode}
+% \end{macro}
+% \begin{macro}{\nonumber}
+% Redefine \cs{nonumber} to reset depth of bottom tag.
+% \changes{v2.15}{2017/03/31}{added MH\_ prefixes}
+% \begin{macrocode}
+\def\g_EQ_latest_nonzerodepth_row_fint{\z@}
+\renewcommand*\nonumber{
+ \@tempcnta=\row@
+ \MH_if_num:w \l_EQ_equationtype_int=\thr@@
+ \advance\@tempcnta\m@ne
+ \MH_else:
+ \MH_if_num:w \l_EQ_equationtype_int=4\relax
+ \advance\@tempcnta\m@ne
+ \MH_fi:
+ \MH_fi:
+ \MH_if_num:w \g_EQ_latest_nonzerodepth_row_fint=\@tempcnta
+ \MH_else:
+ \global\g_EQ_bottomtag_depth_dim=-\baselineskip
+ \MH_fi:
+ \MT_nonumber:
+}
+% \end{macrocode}
+% \end{macro}
+%
+% End of \pkg{ntheorem} patches and package.
+% \begin{macrocode}
+%</package>
+% \end{macrocode}
+% Finale
+\endinput