% \iffalse meta-comment
% Copyright (C) 2022 by Martin Vassor
% This file 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.
%
% \fi
% \iffalse
%<package> \NeedsTeXFormat{LaTeX2e}
%<package> \ProvidesPackage{messagepassing}[2023/12/13 v1.2 A package to draw message passing diagrams]
%<package> \RequirePackage{tikz}
%<package> \usetikzlibrary{quotes, calc, arrows.meta}
%<package> \RequirePackage{xparse}
%<package> \RequirePackage{float}
%
%<*driver>
\documentclass{ltxdoc}
\usepackage{messagepassing}
\usepackage[hidelinks]{hyperref}
\usepackage{xspace}
\usepackage[super]{nth}
\usepackage{subcaption}
\usepackage{listings}
\lstset{
language=[LaTeX]{TeX},
}
\EnableCrossrefs
\CodelineIndex
\RecordChanges
\begin{document}
\DocInput{messagepassing.dtx}
\end{document}
%</driver>
% \fi
% \CheckSum{0}
%
% \CharacterTable
% {Upper-case \A\B\C\D\E\F\G\H\I\J\K\L\M\N\O\P\Q\R\S\T\U\V\W\X\Y\Z
% Lower-case \a\b\c\d\e\f\g\h\i\j\k\l\m\n\o\p\q\r\s\t\u\v\w\x\y\z
% Digits \0\1\2\3\4\5\6\7\8\9
% Exclamation \! Double quote \" Hash (number) \#
% Dollar \$ Percent \% Ampersand \&
% Acute accent \' Left paren \( Right paren \)
% Asterisk \* Plus \+ Comma \,
% Minus \- Point \. Solidus \/
% Colon \: Semicolon \; Less than \<
% Equals \= Greater than \> Question mark \?
% Commercial at \@ Left bracket \[ Backslash \\
% Right bracket \] Circumflex \^ Underscore \_
% Grave accent \` Left brace \{ Vertical bar \|
% Right brace \} Tilde \~}
%
%
% \changes{v1.2}{2023/12/13}{Add optional display names in processes and message node options}
% \changes{v1.1}{2022/07/20}{Consistent spelling of "colour"}
% \changes{v1.0}{2022/02/18}{Initial version for publication}
%
% \GetFileInfo{messagepassing.sty}
%
% \title{The \textsf{messagepassing} package\thanks{This document
% corresponds to \textsf{messagepassing}~\fileversion,
% dated \filedate.}}
% \author{Martin Vassor\\ \texttt{
[email protected]}}
%
%
% \maketitle
% \tableofcontents
% \newcommand{\TikZ}{Ti\textit{k}Z\xspace}
% \section{Introduction}
% This package provides an environment and associated macros to easily draw
% message passing diagrams. For instance, Execution.~\ref{mp:ex1} shows the capabilities offered
% by the package.
%
% \begin{messagepassing}[][An example of message passing][h][mp:ex1]
% \newprocesswithlength{p}{9}
% \newprocesswithlength{q}{9}
% \newprocesswithlength{r}{5}
% \crash{r}{5}{k}
%
% \checkpoint{p}{0}{c_p^0}
% \checkpoint{q}{0}{c_q^0}
% \checkpoint{r}{0}{c_r^0}
% \checkpoint{p}{4}{c_p^1}
%
% \stateinterval{p}{0}{si_p^0}
% \stateinterval{q}{0}{si_q^0}
% \stateinterval{r}{0}{si_r^0}
%
% \sendwithstateinterval{p}{1}{q}{2}{si_q^1}
% \sendwithstateinterval{r}{3}{q}{4}{si_q^2}
% \sendwithstateintervalandname{q}{5}{p}{6}{si_p^1}{\(m\)}
% \sendoutofband{q}{5}{r}{6}{\(m_\mathrm{copy}\)}
% \stateinterval{r}{6}{si_r^0}
% \restart{r}{6}{3}
%
% \annotate{r}{8}{restart}
% \annotatexplicit{p}{8}{\(1+1=2\)}{note}
%
% \colouredbox{p}{q}{7}{9}{Protocol}
%
% \drawtimeline{9}
% \end{messagepassing}
%
% \section{Usage}
% \subsection{Loading the package}
% The package accepts two options: \verb+vertical+ and \verb+annotatevertical+.
% If the former is set, executions will be drawn with time going from top to
% bottom, instead of from left to right. Doing so, almost all
% labels\footnote{Annotations are not rotated, unless explicitly asked.} are
% rotated as well. If, \emph{in addition}, \verb+annotatevertical+ is set, then
% annotations (including names of colouredboxes) are rotated as well.
% \subsection{Basic usage}
% \subsubsection{Creating a diagram.}
% \DescribeEnv{messagepassing} A diagram can easily be created using the
% |messagepassing| environment. The syntax is:
% |\begin{messagepassing}| \oarg{tikz} \oarg{caption} \oarg{placement}
% \oarg{label}.
% The first optional argument (\emph{tikz}) contains arguments that are passed
% to the underlying tikz environment. The second argument (\emph{caption}) has
% two effect: if set, it turns the diagram into a floating figure, and the
% content of the argument is the caption of the floating figure. The third
% argument (\emph{placement}) is the placement option of the figure, the default
% is |p|. Finally, the fourth option (\emph{label}) is the label used to
% reference the figure.
%
% For instance, the diagram in Figure~\ref{mp:ex1} is created with the following
% commands:
% \begin{verbatim}
% \begin{messagepassing}[][An example of message passing][h][mp:ex1]
% % ...
% \end{messagepassing}
% \end{verbatim}
% \paragraph{Setting up the diagram.}
% When created, the diagram is empty. Before actually writing the message
% exchanges, we have to set up a few things: set whether we want a timeline (and
% if it is the case, of which length), and set the number of processes with
% their names, etc.
% \subparagraph{Creating a new process.}
% Each process is characterised by its name.
% \DescribeMacro{\newprocess} The simplest macro to create a new process is then
% |\newprocess| \marg{name} \oarg{display name}, where \marg{name} is the name
% of the process (resp. \(p\), \(q\), and \(r\) in Figure~\ref{mp:ex1}). If
% \oarg{display name} is provided, it is used as the name when rendering the
% figure, while \marg{name} is used internally to refer to the process. This
% allows you to use names that can not be used internally (e.g.\ with maths or
% things like that).
%
% In addition, we often draw a horizontal\footnote{By default, the line is vertical if
% the option |vertical| is used.} line that represent the running process.
% Although this line can be manually
% added\footnote{|processlength|\marg{process}\marg{length}
% creates a line of length \emph{length} for process \emph{process}.}, we also
% provide a simple macro that performs both actions:
% \DescribeMacro{\newprocesswithlength}
% |\newprocesswithlength|\marg{name}\oarg{display name}\marg{length}.
%
% An other alternative is to name the state in which the process starts (in
% Fig~\ref{mp:ex1}, we call those states \(si\) as \emph{state intervals}).
% Again, this can be achieved using individual commands, but we also provide
% \DescribeMacro{\newprocesswithstateinterval}
% |\newprocesswithstateinterval|\marg{name}\oarg{display name}\marg{state name}.
%
% Finally, an other way to create a process is to create a process that
% (eventually) fails, which is represented by the process' line terminating
% early with a cross. For that, we provide \DescribeMacro{\newprocesswithcrash}
% |\newprocesswithcrash|\marg{name}\oarg{display name}\marg{length}\marg{crash coordinate name}.
% The first arguments are similar to |\newprocesswithlength|, and the last one
% is used to provide a name for the coordinate where the crash occurs. This name
% can later be used to place nodes.
%
% Of course, we can imagine other combinations (e.g. a process with a length and
% a state interval). We do not provide individual commands for each combination,
% but the can be easily achieved using separate commands.
%
% As an example, the processes of Fig.~\ref{mp:ex1} are created as follows.
% \begin{verbatim}
% \newprocesswithlength{p}{9}
% \newprocesswithlength{q}{9}
% \newprocesswithlength{r}{5}
% \end{verbatim}
%
% \subparagraph{Setting up a timeline.}
% An other setup action consists in setting up (if wanted) the timeline. Notice
% that this can be done \emph{at any place} in the diagram. To do so, simply use
% the command \DescribeMacro{\drawtimeline}|\drawtimeline|\marg{length}, where
% \emph{length} is the length of the desired timeline.
% \subsubsection{Populating the run.}
% Now that we have some processes, we have to populate the diagram with some
% actions.
%
% \paragraph{Basic message.}
% The most basic action is to send a message. For that, we provide the command
% \DescribeMacro{\send} |\send|\marg{sender}\marg{send
% time}\marg{receiver}\marg{receive time}. The sender and receiver are
% identified with their names, and the sending and receiving times are given
% according to their timestamp\footnote{Notice that nothing prevents sending
% messages in the past, simply set a receiving time before the sending time.}.
%
% For instance, in Figure~\ref{fig:ex_msg}, we use |\send{p}{1}{q}{2}|.
%
% In addition, we can label the arrow with the message that is sent with
% \DescribeMacro{\sendwithname}
% |\sendwithname|\marg{sender}\marg{send time}\marg{receiver}\marg{receive
% time}\marg{label}\oarg{label options}. The \marg{label} contains the label
% that should be displayed. The package provides default positioning options for
% the label, which should be acceptable for most cases. Those positioning
% options can be overridden by \oarg{label options}, which should be \TikZ node
% options.
%
% Finally, we sometimes distinguish \emph{out-of-band} messages, e.g. messages
% that do not carry informations, but that are for instance used for metadata,
% etc.. We provide the macro
% \DescribeMacro{\sendoutofband}|\sendoutofband|\marg{sender}\marg{send time}\marg{receiver}\marg{receive
% time}\marg{label}\oarg{label options}, which behaves similarly to |\sendwithname|, but prints the
% message in an other colour.
%
% \begin{figure}
% \centering
% \begin{minipage}{.35\textwidth}
% \begin{messagepassing}
% \newprocesswithlength{p}{4}
% \newprocesswithlength{q}{4}
% \send{p}{1}{q}{2}
% \sendwithname{p}{2}{q}{3}{msg}
% \end{messagepassing}
% \end{minipage}
% ~
% \begin{minipage}{.60\textwidth}
% \iffalse
%<*example>
% \fi
% \begin{lstlisting}
\begin{messagepassing}
\newprocesswithlength{p}{4}
\newprocesswithlength{q}{4}
\send{p}{1}{q}{2}
\sendwithname{p}{2}{q}{3}{msg}
\end{messagepassing}
\end{lstlisting}
% \iffalse
%</example>
% \fi
% \end{minipage}
% \caption{A very simple protocol with a single message exchanged.}
% \label{fig:ex_msg}
% \end{figure}
%
%
% \paragraph{Process crash and restore.}
% The crash of a process can be represented using
% \DescribeMacro{\crash}
% |\crash|\marg{process name}\marg{time}\marg{crash name}. The argument \emph{process name}
% is the name of the process that crashes, and \emph{crash name} is used to give
% a name to the crash. Naming the crash is useful for coordinates (see below).
% Finally, \emph{time} specifies when the crash occurs. Notice that this does
% not modify the timeline: it simply adds a crash token at the specified
% coordinate. This means that (i) then timeline has to stop at the crash's time;
% and (ii) it has to be restarted after. To stop the timeline, simply take the
% crash into account when setting the initial timeline. To restart the timeline,
% we provide the command
% \DescribeMacro{\restart}|\restart|\marg{name}\marg{date}\marg{duration}.
% \emph{name} specifies which process is to be restarted; \emph{date} specifies
% when the process should be restarted, and \emph{duration} specifies how long
% the process shoud be alive (i.e. what is the length of the timeline) after the
% restart.
% \begin{figure}
% \centering
% \begin{minipage}{.35\textwidth}
% \begin{messagepassing}
% \newprocesswithlength{p}{4}
% \newprocesswithlength{q}{2}
% \send{p}{1}{q}{2}
% \crash{q}{2}{crash}
% \restart{q}{3}{1}
% \end{messagepassing}
% \end{minipage}
% ~
% \begin{minipage}{.60\textwidth}
% \iffalse
%<*example>
% \fi
% \begin{lstlisting}
\begin{messagepassing}
\newprocesswithlength{p}{4}
\newprocesswithlength{q}{2}
\send{p}{1}{q}{2}
\crash{q}{2}{crash}
\restart{q}{3}{1}
\end{messagepassing}
\end{lstlisting}
% \iffalse
%</example>
% \fi
% \end{minipage}
% \caption{A protocol with a crashed process.}
% \end{figure}
%
% \paragraph{Tokens on the run}
%
% The package also proposes two kinds of tokens that can be added on protocols'
% lines. The first one is a \emph{checkpoint} (i.e. a state that is saved
% somewhere) and the second is used to denote the begining of a \emph{state
% interval} (a state interval denotes a period in which a process only performs
% deterministic events). The former are denoted with a small black rectangle,
% while the later is denoted with a vertical line. Although those two tokens are
% intended for the usage mentionned above, we encourage users to use them for
% other usages if need be.
%
% A checkpoint can be added with
% \DescribeMacro{\checkpoint}|\checkpoint|\marg{process}\marg{time}\marg{name},
% where \emph{process} is the name of the process which takes a checkpoint,
% \emph{time} is the time at which the checkpoint is taken, and \emph{name}
% is the name of the checkpoint, that is printed next to it, and can be used as
% a coordinate. Notice that the name is printed in a math environment, as we
% expect most checkpoints names to be indexed, e.g.\(c_1\), \(c_2\), etc.
% To have more control on the printed name, or if the proposed name is not a
% valid coordinate name, we offer a variant
% \DescribeMacro{\checkpointspecial}|\checkpointspecial|\marg{process}\marg{time}\marg{name}\marg{label},
% where \emph{name} is the name of the coordinate of the checkpoint, and
% \emph{label} is the label to be printed. Notice that, in that case, the label
% is printed as is, i.e. not typeset as maths.
%
% A state interval can be added similarly with the command
% \DescribeMacro{\stateinterval}|\stateinterval|
% \marg{process}\marg{time}\marg{name}.
%
% For the sake of completeness, if you need the name of the coordinate and the
% displayed label to be different (e.g. if the label can not be the name of a
% coordinate, for whatever reason), we also provide the command
% \DescribeMacro{\stateintervalspecial}
% |\stateintervalspecial|\marg{process}\marg{time}\marg{name}\marg{label}, in
% which \emph{name} is the name of the created coordinate, and \emph{label} is
% the label attached to the state interval.
% \begin{figure}
% \centering
% \begin{minipage}{.35\textwidth}
% \begin{messagepassing}
% \newprocesswithlength{q}{2}
% \checkpoint{q}{1}{c_1}
% \crash{q}{2}{crash}
% \restart{q}{3}{1}
% \end{messagepassing}
% \end{minipage}
% ~
% \begin{minipage}{.60\textwidth}
% \iffalse
%<*example>
% \fi
% \begin{lstlisting}
\begin{messagepassing}
\newprocesswithlength{q}{2}
\checkpoint{q}{1}{c_1}
\crash{q}{2}{crash}
\restart{q}{3}{1}
\end{messagepassing}
\end{lstlisting}
% \iffalse
%</example>
% \fi
% \end{minipage}
% \caption{A protocol with a checkpoint.}
% \end{figure}
% \paragraph{Grey boxes}
%
% In Execution~\ref{mp:ex1}, we created a light-red box between processes \(p\)
% and \(q\), from time \(7\) to \(9\), to indicate that they perform a given
% protocol that we don't detail further. We call such boxes (which can be used for a lot
% of other purposes) \emph{colouredboxes}, and they
% can be added with \DescribeMacro{\colouredbox} |\colouredbox|\marg{first
% process}\marg{second process}\marg{start time}\marg{end time}\marg{label}.
% This creates a box that spans between \emph{first process} and \emph{second
% process}, from \emph{start time} to \emph{end time}, with the label
% \emph{label} printed.
%
% Notice that there are no technical restrictions to adding messages on top of a
% box, typically to highlight a specific part of a larger execution.
% \paragraph{Annotations}
% Finally, it is possible to add annotations on the diagram. To do so, we
% provide the macro
% \DescribeMacro{\annotate}|\annotate|\marg{process}\marg{time}\marg{text} which
% adds an annotation with \emph{text} over the timeline of the given
% \emph{process} at time \emph{time}. This also creates a coordinate at the
% annotation time, which name is the content of the annotation (i.e.\
% \emph{text}). If \emph{text} is not a valid coordinate name, then the
% alternative \DescribeMacro{\annotatexplicit}
% |\annotatexplicit|\marg{process}\marg{time}\marg{text}\marg{name} behaves
% similarly, except that the coordinate name is explicitly given in argument
% \emph{name}.
% \subsubsection{Combined commands}
% The above commands are sufficient to use all primitives offered by the
% package. In addition, we provide a lot of \emph{combined commands}, which, as
% the name suggest, have the effect of multiple \emph{simple} commands.
%
% \begin{itemize}
% \item |\newprocesswithlength|\marg{name}\marg{lifetime}: combination of
% |\newprocess|\marg{name} and |\processlength|\marg{name}\marg{lifetime}
% \item |\newprocesswithstateinterval|\marg{process name}\marg{state interval
% name}: combination of |\newprocess|\marg{process name} and
% |\stateinterval|\marg{process name}\marg{0}\marg{state interval name}
% \item |\newprocesswithcrash|\marg{process name}\marg{crash time}\marg{crash
% name}: creates a process \emph{process name} that runs until \emph{crash
% time}. The crash is named \emph{crash name}.
% \item |\sendwithstateinterval|\marg{sender}\marg{send
% time}\marg{receiver}\marg{receive time}\marg{si name}: combines |\send| and
% |\stateinterval|.
% \item |\sendwithstateintervalandname|\marg{sender}\marg{send
% time}\marg{receiver}\marg{receive time}\marg{si name}\marg{message name}:
% combines |\sendwithname| and |\stateinterval|
% \end{itemize}
% \subsection{Advanced usage}
% \subsubsection{Customising colours}
%
% Two parts of the package use colours: colouredboxes and out-of-band messages. By
% default both are shades of red. We provide commands to change that if desired.
%
% \DescribeMacro{\colouredboxcolour}|\colouredboxcolour|\marg{colour} changes the colour
% used for colouredboxes. Notice that this sets both the background colour (which is
% a light variant of the provided colour) and the text colour (which uses the
% provided colour).
%
% \DescribeMacro{\oobcolour}|\oobcolour|\marg{colour} changes the colour used for
% out-of-band messages.
%
% \subsubsection{Coordinates}
%
% \paragraph{\TikZ coordinates.}
% Message passing diagrams are drawn using \TikZ, which means that one can add
% arbitrary commands to a diagram. In addition, the package defines useful
% coordinates to refer to. Execution~\ref{mp:ex2} shows the \TikZ coordinate plan
% overlayed on top of Execution~\ref{mp:ex1}.
%
% On \TikZ \(y\)-axis processes are instanciated one unit apart from each other,
% in their declaration order.
% To keep the coordinate system simple, processes expand in the negative (e.g.
% the first process declared is at coordinate \((0, -1)\), the second at \((0,
% -2)\), etc.).
%
% The \TikZ \(x\)-axis corresponds to the time axis of the diagram. Therefore,
% e.g. coordinate \((3, -4)\) corresponds to the \nth{3} time step of the
% \nth{4} process.
%
% \paragraph{Named coordinates.}
% In addition to explicit coordinates explained above, the package names most of
% the points of interest in the diagram.
%
% \subparagraph{Coordinates of processes.}
% At each process declaration, a coordinate named after the name of the process
% is created. The coordinate corresponds to the beginning of the corresponding
% process' timeline (for instance, in Execution~\ref{mp:ex2}, we show coordinate
% \((\texttt{q})\), that corresponds to the process \(q\)).
%
% \subparagraph{Coordinate of states intervals (resp. checkpoints, resp. crashes).}
% Similarly to processes, each state interval (resp. checkpoint, resp. crashes) creates a
% coordinate, named after the name of the state interval (resp. checkpoint,
% resp. crashes), is created. The coordinates refers to the place of the state
% interval (resp. checkpoint, resp. crashes). For instance, in
% Execution~\ref{mp:ex2}, we show the coordinates \((\texttt{si\_q\^{}1})\),
% \((\texttt{c\_p\^{}1})\) and \((\texttt{k})\), that respectively correspond to the state interval
% \(si_q^1\), the checkpoint \(c_p^1\) and the crash\footnote{The names of
% crashes are not printed on the figure, although they are internally defined.} \(k\).
%
% \subparagraph{Coordinates of annotations.}
% When an annotation is created, a coordinate is created at the same place, on
% the process' timeline\footnote{Notice that, using the explicit variant of |annotate|
% (|annotatexplicit|), the name of the annotation has to be explicitly given.}
% For instance, in Execution~\ref{mp:ex1}, the annotation \(1+1=2\) is created
% with the explicit name ``\texttt{note}''. We show the corresponding coordinate in
% Execution~\ref{mp:ex2}.
%
%
% \begin{messagepassing}[][Showing \TikZ coordinates][h][mp:ex2]
% \newprocesswithlength{p}{9}
% \newprocesswithlength{q}{9}
% \newprocesswithlength{r}{5}
% \crash{r}{5}{k}
%
% \checkpoint{p}{0}{c_p^0}
% \checkpoint{q}{0}{c_q^0}
% \checkpoint{r}{0}{c_r^0}
% \checkpoint{p}{4}{c_p^1}
%
% \stateinterval{p}{0}{si_p^0}
% \stateinterval{q}{0}{si_q^0}
% \stateinterval{r}{0}{si_r^0}
%
% \sendwithstateinterval{p}{1}{q}{2}{si_q^1}
% \sendwithstateinterval{r}{3}{q}{4}{si_q^2}
% \sendwithstateintervalandname{q}{5}{p}{6}{si_p^1}{\(m\)}
% \sendoutofband{q}{5}{r}{6}{\(m_\mathrm{copy}\)}
% \stateinterval{r}{6}{si_r^0}
% \restart{r}{6}{3}
%
% \annotate{r}{8}{restart}
% \annotatexplicit{p}{8}{\(1+1=2\)}{note}
%
% \colouredbox{p}{q}{7}{9}{Protocol}
% \drawtimeline{9}
%
%
% \fill[fill=white, opacity=0.8] (-1, 0) rectangle (10, -5);
%
% \draw[dotted, gray] (-1,0) grid (10,-5);
% \draw[->, thick, red] (0, 0) -- (1, 0) node [anchor=south, text=black] {\((1, 0)\)};
% \draw[->, thick, red] (0, 0) -- (0, -1) node [anchor=east, text=black] {\((0, -1)\)};
%
% \draw[fill] (q) circle [radius=.5mm];
% \node[anchor=east] at (q) {\((\texttt{q})\)};
%
% \draw[fill] (si_q^1) circle [radius=.5mm];
% \node[anchor=east] at (si_q^1) {\((\texttt{si\_q\^{}1})\)};
%
% \draw[fill] (c_p^1) circle [radius=.5mm];
% \node[anchor=east] at (c_p^1) {\((\texttt{c\_p\^{}1})\)};
%
% \draw[fill] (k) circle [radius=.5mm];
% \node[anchor=east] at (k) {\((\texttt{k})\)};
%
% \draw[fill] (note) circle [radius=.5mm];
% \node[anchor=east] at (note) {\((\texttt{note})\)};
%
% \end{messagepassing}
% \StopEventually{}
%
% \section{Implementation}
% \begin{macrocode}
\newcounter{processnb}
\setcounter{processnb}{0}
\newcounter{maxtime}
\pgfdeclarelayer{background}
\pgfsetlayers{background,main}
\newfloat{float_messagepassing}{t b h p}{.mp}
\floatname{float_messagepassing}{Execution}
\newif\ifmp@vertical\mp@verticalfalse
\DeclareOption{vertical}{
\mp@verticaltrue
}
\newif\ifmp@annotatevertical\mp@annotateverticalfalse
\DeclareOption{annotatevertical}{
\mp@annotateverticaltrue
}
\ProcessOptions\relax
\ifmp@vertical
\newcommand{\mp@processnameanchor}{south}
\newcommand{\mp@timeticksanchor}{east}
\newcommand{\mp@messagelabelanchor}{south}
\newcommand{\mp@stateintervalanchor}{north west}
\newcommand{\mp@checkpointanchor}{east}
\newcommand{\mp@verticalrotation}{270}
\ifmp@annotatevertical
\newcommand{\mp@annotaterotation}{270}
\else
\newcommand{\mp@annotaterotation}{0}
\fi
\else
\newcommand{\mp@processnameanchor}{east}
\newcommand{\mp@timeticksanchor}{north}
\newcommand{\mp@messagelabelanchor}{west}
\newcommand{\mp@stateintervalanchor}{south west}
\newcommand{\mp@checkpointanchor}{north}
\newcommand{\mp@verticalrotation}{0}
\newcommand{\mp@annotaterotation}{0}
\fi
\newcommand{\mp@oobcolour}{red}
\newcommand{\oobcolour}[1]{
\renewcommand\mp@oobcolour{#1}
}
\newcommand{\mp@colouredboxcolour}{red}
\newcommand{\colouredboxcolour}[1]{
\renewcommand\mp@colouredboxcolour{#1}
}
\newif\iftimeline
% \end{macrocode}
%
%\begin{environment}{messagepassing}
% \begin{macrocode}
\ExplSyntaxOn
%% 1st argument: tikz arguments
%% 2nd argument: Float caption (turns in floating)
%% 3rd argument: Float placement (`p` by default)
%% 4th argument: Float label
\NewDocumentEnvironment{messagepassing} {o o o o}
{
\timelinefalse
\setcounter{processnb}{0}
\IfNoValueTF{#2} {
}{
\IfNoValueTF{#3}{
\begin{float_messagepassing}[p]
} {
\begin{float_messagepassing}[#3]
}
\begin{center}
}
\IfNoValueTF{#1}{
\begin{tikzpicture}[rotate=\mp@verticalrotation]
} {
\begin{tikzpicture}[rotate=\mp@verticalrotation, #1]
}
}{
%% Draw timeline if boolean is true
\iftimeline
\begin{pgfonlayer}{background}
\setcounter{maxtime}{\@maxtime}
\addtocounter{maxtime}{-1}
\coordinate (maxtime) at (\@maxtime, 0);
\addtocounter{processnb}{1}
\coordinate (timeline) at (0, -\value{processnb});
\draw (timeline) node [anchor=\mp@processnameanchor] {{\it time}};
\draw[->] (timeline) -- ($(timeline) + (maxtime)$);
\foreach \i in {0,...,\value{maxtime}} {
\draw ($(timeline) + (\i, 0) + (0, 0.1)$) -- ($(timeline) + (\i, 0) + (0, -0.1)$) node [anchor=\mp@timeticksanchor] {$\i$};
}
\end{pgfonlayer}
\else
\fi
\end{tikzpicture}
\IfNoValueTF{#2} {
\linebreak
} {
\end{center}
\caption{#2}
\IfNoValueTF{#4} {
}{
\label{#4}
}
\end{float_messagepassing}
}
}
\ExplSyntaxOff
% \end{macrocode}
% \end{environment}
%
% \begin{macrocode}
%% #1: name
%% #2: display name
\NewDocumentCommand{\newprocess}{m o}{
\addtocounter{processnb}{1}
\coordinate (#1) at (0, -\value{processnb});
\draw (#1) node[anchor=\mp@processnameanchor] {\IfValueTF{#2}{#2}{$#1$}};
}
% \end{macrocode}
%
% \begin{macrocode}
%% #1: name
%% #2: display name
%% #3: width
\NewDocumentCommand{\newprocesswithlength}{m o m}{
\newprocess{#1}[#2]
\processlength{#1}{#3}
}
% \end{macrocode}
%
% \begin{macrocode}
%% #1: name
%% #2: display name
%% #3: state interval name
\NewDocumentCommand{\newprocesswithstateinterval}{m o m}{
\newprocess{#1}[#2]
\stateinterval{#1}{0}{#3}
}
% \end{macrocode}
%
% \begin{macrocode}
%% #1: name
%% #2: display name
%% #3: width
%% #4: crash name
\NewDocumentCommand{\newprocesswithcrash}{m o m m}{
\newprocess{#1}[#2]{#3}
\crash{#1}{#3}{#4}
}
% \end{macrocode}
%
% \begin{macrocode}
%% #1: sender's name
%% #2: send date
%% #3: receiver's name
%% #4: receive date
\newcommand{\send}[4]{
\draw[->] (#1) +(#2, 0) -- ($ (#3) +(#4, 0) $);
}
% \end{macrocode}
%
% \begin{macrocode}
%% #1: sender's name
%% #2: send date
%% #3: receiver's name
%% #4: receive date
%% #5: message name
%% #6: message name display options
\NewDocumentCommand{\sendwithname}{m m m m m o}{
\IfValueTF{#6}{
\draw[->] (#1) +(#2, 0) -- ($ (#3) +(#4, 0) $) node[anchor=\mp@messagelabelanchor, pos=0.3, #6] {#5};
}{
\draw[->] (#1) +(#2, 0) -- ($ (#3) +(#4, 0) $) node[anchor=\mp@messagelabelanchor, pos=0.3] {#5};
}
}
% \end{macrocode}
%
% \begin{macrocode}
%% #1: process name
%% #2: process width
\newcommand{\processlength}[2]{
\draw (#1) -- +(#2, 0);
}
% \end{macrocode}
%
% \begin{macrocode}
%% #1: sender's name
%% #2: send date
%% #3: receiver's name
%% #4: receive date
%% #5: state interval name
\newcommand{\sendwithstateinterval}[5] {
\send{#1}{#2}{#3}{#4}
\stateinterval{#3}{#4}{#5}
}
% \end{macrocode}
%
% \begin{macrocode}
%% #1: sender's name
%% #2: send date
%% #3: receiver's name
%% #4: receive date
%% #5: state interval name
%% #6: message name
%% #7: message name display options
\NewDocumentCommand{\sendwithstateintervalandname}{m m m m m m o} {
\sendwithname{#1}{#2}{#3}{#4}{#6}[#7]
\stateinterval{#3}{#4}{#5}
}
% \end{macrocode}
%
% \begin{macrocode}
%% #1: sender's name
%% #2: send date
%% #3: receiver's name
%% #4: receive date
%% #5: OoB message name
%% #6: OoB message name display options
\NewDocumentCommand{\sendoutofband}{m m m m m o}{
\IfValueTF{#6}{
\draw[->, color=\mp@oobcolour] (#1) +(#2, 0) -- ($ (#3) +(#4, 0) $) node[anchor=\mp@messagelabelanchor, pos=0.3, #6] {#5};
}{
\draw[->, color=\mp@oobcolour] (#1) +(#2, 0) -- ($ (#3) +(#4, 0) $) node[anchor=\mp@messagelabelanchor, pos=0.3] {#5};
}
}
% \end{macrocode}
%
% \begin{macrocode}
%% #1: process's name
%% #2: state interval date
%% #3: state interval name
\newcommand{\stateinterval}[3] {
\stateintervalspecial{#1}{#2}{#3}{#3}
}
% \end{macrocode}
%
% \begin{macrocode}
%% #1: process's name
%% #2: state interval date
%% #3: coordinate name
%% #4: state interval label
\newcommand{\stateintervalspecial}[4] {
\coordinate (#3) at ($ (#1) +(#2, 0) $);
\draw (#3) + (0, 0.1) -- +(0, -0.1) node[anchor=\mp@stateintervalanchor] {$#4$};
}
% \end{macrocode}
%
% \begin{macrocode}
%% #1: process's name
%% #2: checkpoint date
%% #3: checkpoint name
\newcommand{\checkpoint}[3]{
\coordinate (#3) at ($ (#1) + (#2, 0) $);
\fill (#3) + (-0.05, 0.1) rectangle +(0.05, -0.1);
\draw (#3) + (0, -0.1) node[anchor=\mp@checkpointanchor] {$#3$};
}
% \end{macrocode}
%
% \begin{macrocode}
%% #1: process's name
%% #2: checkpoint date
%% #3: checkpoint coordinate name
%% #4: checkpoint label
\newcommand{\checkpointspecial}[4]{
\coordinate (#3) at ($ (#1) + (#2, 0) $);
\fill (#3) + (-0.05, 0.1) rectangle +(0.05, -0.1);
\draw (#3) + (0, -0.1) node[anchor=\mp@checkpointanchor] {#4};
}
% \end{macrocode}
%
% \begin{macrocode}
%% #1: process's name
%% #2: crash date
%% #3: crash name
\newcommand{\crash}[3]{
\coordinate (#3) at ($ (#1) + (#2, 0) $);
\draw (#3) + (-0.1, -0.1) -- +(0.1, 0.1);
\draw (#3) + (0.1, -0.1) -- +(-0.1, 0.1);
}
% \end{macrocode}
%
% \begin{macrocode}
%% #1: process's name
%% #2: restart date
%% #3: restart length
\newcommand{\restart}[3]{
\draw (#1) + (#2, 0) -- ($ (#1) + (#2, 0) + (#3, 0) $);
}
% \end{macrocode}
%
% \begin{macrocode}
%% #1: first process's name
%% #2: second process's name
%% #3: begining of the grey box
%% #4: end of the grey box
%% #5: caption
\newcommand{\colouredbox}[5]{
\begin{pgfonlayer}{background}
\fill[color=\mp@colouredboxcolour!20] ($(#1) + (#3, 0)$) rectangle ($(#2) + (#4, 0)$) node[midway, color = \mp@colouredboxcolour, rotate=\mp@annotaterotation]{#5};
\end{pgfonlayer}
}
% \end{macrocode}
%
% \begin{macrocode}
%% #1: Timeline length
\newcommand{\drawtimeline}[1]{
\timelinetrue
\def\@maxtime{#1}
}
% \end{macrocode}
%
% \begin{macrocode}
%% #1: process's name
%% #2: annotation date
%% #3: annotation
\newcommand{\annotate}[3]{
\annotatexplicit{#1}{#2}{#3}{#3}
}
% \end{macrocode}
%
% \begin{macrocode}
%% Same than annotate, but with the coordinate name provided explicitly
%% #1: process's name
%% #2: annotation date
%% #3: annotation
%% #4: coordinate name
\newcommand{\annotatexplicit}[4]{
\coordinate (#4) at ($ (#1) +(#2, 0.1) $);
\draw (#4) node[rotate=\mp@annotaterotation, anchor=south] {#3};
}
% \end{macrocode}
