path: root/support/light-latex-make/llmk.tex

% llmk: the reference manual
% Copyright 2020 Takuto ASAKURA (wtsnjp)

% +++
% latex = "xelatex"
% +++
\documentclass{llmk-doc}

% Metadata
\title{llmk: Light {\LaTeX} Make}
\author{Takuto Asakura (wtsnjp)}
\subtitle{Reference Manual}
\date{v1.1.0\quad\today}
\keywords{llmk, build-tool, toml, lua, luatex}

\begin{document}

\maketitle

\section{Overview}

Light {\LaTeX} Make (\prog{llmk}) is yet another build tool specific for
{\LaTeX} documents. Its aim is to provide a simple way to specify a workflow of
processing {\LaTeX} documents and encourage people to always explicitly show
the right workflow for each document.

The main features of \prog{llmk} are all about the above purpose. First, you
can write the workflows either in an external file \code{llmk.toml} or in a
{\LaTeX} document source in a form of magic comments. Further, multiple magic
comment formats can be used. Second, it is fully cross-platform. The only
requirement of the program is the \code{texlua} command; \prog{llmk} makes a
uniform way to describe the workflows available for nearly all {\TeX}
environments. Third, it behaves exactly the same in any environment. At this
point, \prog{llmk} intentionally does not provide any method for user
configuration. Therefore, one can guarantee that a {\LaTeX} document with an
\prog{llmk} setup, the process of typesetting the document must be reproduced
in any {\TeX} environment with the program.

\subsection{Installation}

This software is included in {\TeX} Live and MiK{\TeX} as Package
\code{light-latex-make}. If you have one of the latest distributions, you
normally don't need to install it by yourself. If you want to install the
development version, please refer to our {\README}.

\subsection{Learning \prog{llmk}}

The bundled {\README} has a general introduction for the program. If you are
new to \prog{llmk} and looking for a quick guidance, you are recommended to
read it first. Conversely, this document can be regarded as a reference manual:
it contains detailed descriptions for every feature of \prog{llmk} as much as
possible, but unsuitable for getting general ideas of its basic usage.

\begin{samepage}
All official resources are available from the repository on GitHub:
%
\begin{quote}
\url{https://github.com/wtsnjp/llmk}
\end{quote}
%
Notably, you can find some example {\LaTeX} documents with \prog{llmk} setups
in the \href{https://github.com/wtsnjp/llmk/tree/master/examples}
{\code{examples}} directory.
% TODO: mention wiki on github when some contents are gathered?
\end{samepage}

The design concept of \prog{llmk} is described in a separate TUGboat
article~\cite{asakura2020}. It will not give you practical tips, but if you are
interested in the underlying ideas of the program, it should be worth reading.
The differences from other similar tools, \eg\prog{latexmk}~\cite{latexmk} and
\prog{arara}~\cite{arara}, are also discussed in the article.

\subsection{Reporting bugs and requesting features}

If you get trouble with \prog{llmk} or think you have found a bug, please
report it by creating either an issue or a pull request on the GitHub
repository:
%
\begin{quote}
\url{https://github.com/wtsnjp/llmk/issues}
\end{quote}
%
If you do not want to use GitHub for some reasons, it is also fine to directly
send an email to the author (\email{tkt.asakura@gmail.com}).

Requests for new features are also welcome; the author cannot promise to
implement the requested features, but will happy to take them into account.
Before making a request, it is strongly recommended to read the article about
the design concept~\cite{asakura2020}.

\section{Command-line interface}

\subsection{Command usage}
\label{sec:command}

The full usage of \prog{llmk} can be summarized as follows:
%
\begin{htcode}
llmk \meta{options} \meta{files}
\end{htcode}

Herein, \meta{options} are the command-line options, that start with the hyphen
character |-|, and \meta{files} are the arguments for the |llmk| command.

\subsubsection*{Arguments \meta{files}}

You can specify the filenames of the source {\TeX} files, normally the files
with \code{.tex} or \code{.ltx} extensions, as the arguments for the \code{llmk}
command. When one or more \meta{files} are specified, \prog{llmk} will read
either the TOML field (Section~\ref{sec:toml}) or another supported magic
comment (Section~\ref{sec:magic-comment}) in each of the source files and
process it with the specified workflow in the given order.

As well as the \code{tex} command, you can omit the \code{.tex} extensions and
just give the basenames of the files for the argument; when a \meta{basename},
which must not contain any dot character, is given and the file that exactly
matches to the name does not exist, \prog{llmk} will automatically add the
\code{.tex} extension and process it like any other if the file
|\meta{basename}.tex| exists.

Note that \prog{llmk} naively pass the given filenames to invoked commands.
Filenames that contains special characters of {\TeX}, \eg |#| and |%|, are very
likely to be causes of troubles. Moreover, at this point \prog{llmk} does not
any specific features to take care of multi-byte characters: filenames
including multi-byte characters may work in some cases but can be cause of
problems\footnote{The author admits that \prog{llmk} needs to be enhanced in
this aspect: it should have better features to treat various filenames with
multi-byte characters, though the author is negative to support special
characters of {\TeX}. Suggestions and patches in this direction are especially
welcome.}. Using filenames that contain only the characters in the range of the
ASCII code, except special characters of {\TeX}, is the safest way to go at any
rate.

Alternatively, if the argument is not specified, \prog{llmk} will read the
special TOML file \code{llmk.toml} in the working directory and execute the
workflow specified in the file (see Section~\ref{sec:toml}). In case the
argument is not specified and the \code{llmk.toml} does not exist, it will
result in an error. When \prog{llmk} executes the workflow specified in the
file \code{llmk.toml}, all magic comments, including TOML fields and other
formats, in each source {\LaTeX} files specified in the \ckey{source} array
will be ignored.

\subsubsection*{Command-line options \meta{options}}
\enlargethispage{5mm}% FIXME

We have tried to implement a GNU-compatible option parser. Short options, each of
which consists of a single letter, must start with a single hyphen |-|.
Multiple short options can be specified with a single hyphen, \eg |-vs| is
equivalent to |-v -s|. Long options have to be following double hyphens |--|.
All options must be specified before the first argument. A string beginning
with a hyphen after the first argument will be treated as an argument.

When two or more options are specified, \prog{llmk} applies them in the given
order. If contradicting options are specified, \eg \sopt{q} v.s.\ \sopt{v}, the
option in the latter position wins over the former one.

\begin{clopt}{\sopt{c}, \lopt{clean}}
Removes temporary files such as \code{aux} and \code{log} files. The files
removed with this action can be customized with keys \ckey{clean\_files} and
\ckey{extra\_clean\_files}.
\end{clopt}

\begin{clopt}{\sopt{C}, \lopt{clobber}}
Removes all generated files including final PDFs. The files removed with this
action can be customized with the key \ckey{clobber\_files}.
\end{clopt}

\begin{clopt}{%
  \code{\sopt{d} \meta{category}}, \code{\lopt{debug}=\meta{category}},
  \sopt{D}, \lopt{debug}}
Activates the specified debug category; debugging messages related to the
activated category will be shown. Herein, available debug categories are:
|config|, |parser|, |run|, |fdb|, |programs|, and |all| to activate all of
these. You can repeat this option more than once to activate multiple
categories. If you specify |-D| or |--debug| without the argument
\meta{category}, it activates all available debug categories.
\end{clopt}

\begin{clopt}{\sopt{h}, \lopt{help}}
Shows a quick help message (namely a list of command-line options) and exit
successfully. When this is specified, all other options and arguments are
ignored.
\end{clopt}

\begin{clopt}{\sopt{n}, \lopt{dry-run}}
Show what would have been executed without actually invoking the commands. This
flag is useful if you want to make sure whether your configuration will work as
expected before the actual building. With option \lopt{verbose}, you can get
further detailed information.
\end{clopt}

\begin{clopt}{\sopt{q}, \lopt{quiet}}
This suppress most of the messages from the program.
\end{clopt}

\begin{clopt}{\sopt{s}, \lopt{silent}}
Silence messages from invoked programs. To be more specific, this redirects
both standard output and standard error streams to the null device.
\end{clopt}

\begin{clopt}{\sopt{v}, \lopt{verbose}}
Make \prog{llmk} to print additional information such as invoked commands
with options and arguments by the program.
\end{clopt}

\begin{clopt}{\sopt{V}, \lopt{version}}
Shows the current version of the program and exit successfully. When this is
specified, all other options and arguments are ignored.
\end{clopt}

\subsection{Exit codes}

\begin{samepage}
You can grasp whether \prog{llmk} successfully executed or not by seeing its
status code. Note that the exit codes of invoked programs are not directly
transferred as the exit code of \prog{llmk}; instead, the statuses of external
programs that failed, if any, are reported in the error messages.
%
\begin{center}
\begin{tabular}{rp{12em}rp{12em}}
\toprule
Code & Error type & Code & Error Type \\ \midrule
\code{0} & Success & \code{3} & Parser Error \\
\code{1} & General Error & \code{4} & Type Error \\
\code{2} & Invoked program failed & \\
\bottomrule
\end{tabular}
\end{center}
\end{samepage}

\section{Writing workflows in TOML format}
\label{sec:toml}

The primary configuration format for \prog{llmk} is TOML\Dash Tom's Obvious
Minimal Language~\cite{toml}. You can specify the workflows to process your
{\LaTeX} documents in the format either in the special configuration file
(\code{llmk.toml}) or in the TOML field (Section~\ref{sec:toml-where}). You
have full access to the \prog{llmk} configuration with this primary format,
while other supported magic comment formats (Section~\ref{sec:magic-comment})
have only partial access. You can read the entire TOML specification at
\href{https://toml.io/}{its website}.

\subsection{TOML in \prog{llmk}}

The configuration for \prog{llmk} written in the TOML format is read by our
built-in parser. At this point, the built-in parser supports a subset of the
TOML specification; only the data-types that necessary for the configuration
keys (Section~\ref{sec:top-level-keys}~\&~\ref{sec:keys-in-programs}) are
supported.

\begin{center}
\newcommand{\ok}{{\color{special}\checkmark}}
\begin{tabular}{llc}
\toprule
Type & Example & Supported \\ \midrule
Bare keys & \code{key} & \ok \\
Quoted keys & \code{"key"} & \\
Dotted keys & \code{tex.latex} & \ok \\ \midrule
Basic strings & \code{"str"} & \ok \\
Multi-line basic strings & & \\
Literal strings & \code{'str'} & \ok \\
Multi-line literal strings & & \\ \midrule
Integer & \code{123} & \ok \\ \midrule
Boolean & \code{true} & \ok \\ \midrule
Float & \code{3.14} & \\ \midrule
Date \& Time & \code{1979-05-27} & \\ \midrule
Array & \code{[1, 2, 3]} & \ok \\ \midrule
Table & \code{[table]} & \ok \\
Inline table & & \\
Array of tables & \code{[[fruit]]} & \\
\bottomrule
\end{tabular}
\end{center}

\subsection{Where to write}
\label{sec:toml-where}

You have two options to write the configuration for \prog{llmk} in TOML format:
(1)~creating a special file \code{llmk.toml} and (2)~writing a TOML field in
your {\LaTeX} source file. Either way, you have full access to the \prog{llmk}
configuration and specify the same workflows in (almost) the same manner.

\subsubsection*{Special file: \code{llmk.toml}}

When the \code{llmk} command is executed without any argument, the special file
\code{llmk.toml} is loaded automatically (Section~\ref{sec:command}). This
filename is fixed and cannot be customized at this point. The entire content of
the file must be a valid TOML; you can include supplemental information in the
form of TOML comment that starts with the \code{\#} character. The file must
be encoded in UTF-8 because it is required by the TOML
specification~\cite{toml}. The \ckey{source} key is \emph{required} in this
file.

\subsubsection*{TOML field}

The other way to pass the configuration in TOML format to \prog{llmk} is using
\emph{TOML fields}\Dash special comment areas in {\LaTeX} source files that are
given by comment lines containing only three or more consecutive \code{+}
characters. The following is a simple example.
%
\begin{lstlisting}[style=latex]
% +++
% # This is a sample TOML field!
% latex = "xelatex"
% +++
\documentclass{article}
\end{lstlisting}

The formal syntax of opening and closing for TOML fields is:
%
\begin{htcode}
\meta{optional spaces}%\meta{optional spaces}+++\meta{optional pluses}\meta{optional spaces}
\end{htcode}
%
where the definitions of \meta{optional spaces} and \meta{optional pluses} are
given as follows (hereafter, whitespace \code{\textvisiblespace} denotes tab
\texttt{0x09} or space \texttt{0x20}).
%
\begin{align*}
&\text{\meta{optional spaces}}
  \longrightarrow \text{\meta{empty}}
  \mid\text{\code{\textvisiblespace}\meta{optional spaces}} \\
&\text{\meta{optional pluses}}
  \longrightarrow \text{\meta{empty}}
  \mid\text{\code{+}\meta{optional pluses}}
\end{align*}

The line of opening and closing for TOML fields must include only the
characters specified in the above. In a TOML filed, you can write TOML code for
\prog{llmk} configuration in the form of {\LaTeX} comment lines.
%
\begin{htcode}
\meta{optional spaces}%\meta{optional spaces}\meta{TOML line}\meta{optional spaces}
\end{htcode}

If one or more arguments are given for the \code{llmk} command, it first looks
for a TOML field from the beginning of each file. Only the topmost field in a
file is the valid TOML field, \ie you cannot have multiple TOML fields in a
file. TOML fields have the highest priority for \prog{llmk} configuration; if a
TOML field is found in a file, other supported magic comments described in
Section~\ref{sec:magic-comment} are ignored.

Though the author recommend you to always encode your {\LaTeX} source file in
UTF-8, you can use other encodings. In any case, the TOML lines in the fields
must be consist of valid UTF-8 encoded strings. Therefore, it is recommended to
use only the characters in the range of ASCII code in your TOML field if you
chose the other encodings for some reasons.

\subsection{Available top-level keys}
\label{sec:top-level-keys}

Now we are going to look over all available TOML keys for \prog{llmk}
configuration. This section includes the full list of available top-level keys,
that are effective for an entire workflow. The detailed specification for each
program in your workflow can be given by the keys in the \ckey{programs} table,
which will be described in the next section. Only the keys shown in these two
sections are effective for \prog{llmk}; if other key names are specified, they
will be ignored. Each key requests a value of specified type; if a value of
which is not the expected type, it will result in a type error.

In the string values for some specific keys, a few \emph{format specifiers} are
available. These specifiers will be replaced to appropriate strings before
executing actions by \prog{llmk}:
%
\begin{itemize}
\item |%S|
  is replaced by the filename given to the \code{llmk} as a command-line
  argument or as an element of the \ckey{source} array.
\item |%T|
  is replaced by the target for each program.
\item |%B|
  is replaced by the basename of |%S|.
\end{itemize}

Some keys have default values; the default configuration of \prog{llmk} should
work well for typical and simple {\LaTeX} documents. Only the keys you
explicitly specify in the TOML format override the default configuration. If
you do not write a key in your configuration, the default value will be used.
In other words, you only need to write the differences from the default
configuration.

\begin{confkey}{bibtex}{type: \type{string}}[default: \code{"bibtex"}]
The command to use for the \progname{bibtex} program. Reference processing
tools that are compatible with {\BibTeX} can be specified for this key, \eg
\code{"biber"}. Internally, this key is an alias for the \ckey{command} key in
the \progname{bibtex} entry. If the \ckey{command} key is specified in the
\ckey{programs} table, this alias is ineffective.
\end{confkey}

\begin{confkey}{clean\_files}{type: \type{array of strings}}
  [default: see bellow]
The files to be removed with the cleaning action (\lopt{clean}). The format
specifiers are available for this key. The default value is:
%
\begin{htcode}
[
  "%B.aux", "%B.bbl", "%B.bcf", "%B-blx.bib", "%B.blg", "%B.fls",
  "%B.idx", "%B.ilg", "%B.ind", "%B.lof", "%B.log", "%B.lot",
  "%B.nav", "%B.out", "%B.run.xml", "%B.snm", "%B.toc", "%B.vrb"
]
\end{htcode}
\end{confkey}

\begin{confkey}{clobber\_files}{type: \type{array of strings}}
  [default: see bellow]
The files to be removed with the clobbering action (\lopt{clobber}). The format
specifiers are available for this key. The default value is:
%
\begin{htcode}
["%B.dvi", "%B.pdf", "%B.ps", "%B.synctex.gz"]
\end{htcode}
\end{confkey}

\begin{confkey}{dvipdf}{type: \type{string}}[default: \code{"dvipdfmx"}]
The command to use for the \progname{dvipdf} program. Internally, this key is
an alias for the \ckey{command} key in the \progname{dvipdf} entry. If the
\ckey{command} key is specified in the \ckey{programs} table, this alias is
ineffective.
\end{confkey}

\begin{confkey}{dvips}{type: \type{string}}[default: \code{"dvips"}]
The command to use for the \progname{dvips} program. Internally, this key is an
alias for the \ckey{command} key in the \progname{dvips} entry. If the
\ckey{command} key is specified in the \ckey{programs} table, this alias is
ineffective.
\end{confkey}

\begin{confkey}{extra\_clean\_files}{type: \type{array of strings}}
  [default: \code{[]}]
Extra files to be removed with the cleaning action (\lopt{clean}). By using
this key, you can easily add files to be removed on top of the default
\ckey{clean\_files}. The format specifiers are available for this key.
\end{confkey}

\begin{confkey}{latex}{type: \type{string}}[default: \code{"lualatex"}]
The command to use for the \progname{latex} program. Internally, this key is an
alias for the \ckey{command} key in the \progname{latex} entry. If the
\ckey{command} key is specified in the \ckey{programs} table, this alias is
ineffective.
\end{confkey}

\begin{confkey}{llmk\_version}{type: \type{string}}
You can declare the \prog{llmk} version to use with this key. This is
especially useful in consideration of compatibility. In case breaking changes
are made in the future updates and an incompatible version is declared with the
key, \prog{llmk} will fallback to the previous behavior or at least show you a
warning message. The versioning of \prog{llmk} will try to follow the semantic
versioning~\cite{semvar} and you can specify one of the versions in the
following syntax:
%
\begin{htcode}
\meta{major}.\meta{minor}.\meta{patch}
\end{htcode}
%
Optionally, you can ommit the tailing \code{.\meta{patch}} part.
\end{confkey}

\begin{confkey}{makeindex}{type: \type{string}}[default: \code{"makeindex"}]
The command to use for the \progname{makeindex} program. Internally, this key
is an alias for the \ckey{command} key in the \progname{makeindex} entry. If
the \ckey{command} key is specified in the \ckey{programs} table, this alias is
ineffective.
\end{confkey}

\begin{confkey}{max\_repeat}{type: \type{integer}}[default: \code{5}]
You can specify the maximum number of execution repetitions for each command in
your \ckey{sequence}. When processing your \ckey{sequence}, \prog{llmk} repeats
a command until \ckey{aux\_file} becomes unchanged from the former execution if
the key is specified and the corresponding auxiliary file exists. This key is to
prevent the potential infinite loop of repetition.
\end{confkey}

\begin{confkey}{programs}{type: \type{table}}
  [default: see Section~\ref{sec:default-programs}]
The table that contains the detailed configuration for each program. See
Section~\ref{sec:keys-in-programs} for the details.
\end{confkey}

\begin{confkey}{ps2pdf}{type: \type{string}}[default: \code{"ps2pdf"}]
The command to use for the \progname{ps2pdf} program. Internally, this key is
an alias for the \ckey{command} key in the \progname{ps2pdf} entry. If the
\ckey{command} key is specified in the \ckey{programs} table, this alias is
ineffective.
\end{confkey}

\begin{confkey}{sequence}{type: \type{array of strings}}
  [default: see Section~\ref{sec:default-sequence}]
The array that contains the names of programs that will be executed by
\prog{llmk}. The programs specified in this array are processed in the given
order with the setups specified in the \ckey{programs} table. Further
information about the default behavior can be found in
Section~\ref{sec:default-sequence}.
\end{confkey}

\begin{confkey}{source}{type: \type{string} or \type{array of strings}}
You can specify a {\LaTeX} source file as a \type{string} value or one or more
files as an \type{array of strings}. This key is only effective and
\emph{required} in \code{llmk.toml}. Conversely, ineffective in TOML fields.
The given filenames will be treated as well as those specified as command-line
arguments.
\end{confkey}

\subsection{Available keys in \code{programs} table}
\label{sec:keys-in-programs}

As it is described, \prog{llmk} invokes each program whose name specified in
the \ckey{sequence} array in the given order. You can control each program
execution by specifying the detailed configuration in the \ckey{programs}
table.

The \ckey{programs} table is a table of tables; each of the entries (\aka
elements) is a set of configuration consists of the keys in the following list.
The \ckey{programs} table must have corresponding entries for all names in the
\ckey{sequence} array, otherwise, it will result in an error.

As well as the top-level keys, some available keys in the \ckey{programs} table
have default values, and only the values for the keys you explicitly specified
in your configuration override. In string values for some specific keys, the
format specifiers described in Section~\ref{sec:top-level-keys} can be used in
the same way.

\begin{confkey}{args}{type: \type{string} or \type{array of strings}}
  [default: \code{["\%T"]}]
You can specify an argument to give the command as a \type{string} value or one
or more arguments as an \type{array of strings}. Each argument will be
surrounded by a pair of double quotations, \eg \code{"arg"}. The format
specifiers are available for this key.
\end{confkey}

\begin{confkey}{aux\_file}{type: \type{string}}
The auxiliary file to monitor so that to check whether rerunning for the
program is necessary or not. If this key is set and the specified file that is
non-empty (see \ckey{aux\_empty\_size}) exists, \prog{llmk} will repeat the
program execution until no change is made to the auxiliary file. This key is
originally for the auxiliary file of {\LaTeX}, but the monitoring feature
should be applicable to other programs. The format specifiers are available
for this key.
\end{confkey}

\begin{confkey}{aux\_empty\_size}{type: \type{integer}}
The empty size in bytes with this key, The auxiliary files specified with the
\ckey{aux\_file} which is smaller than the value of this key are recognized as
\emph{empty} and will be ignored. For instance, the auxiliary files of {\LaTeX},
which normally have \code{.aux} extensions, contain
|\cs{relax}\textvisiblespace\meta{line break}| (9 bytes); thus the default
value of this key for the \progname{latex} program is \code{9}.
\end{confkey}

\begin{confkey}{command}{type: \type{string}}
The name of the command to invoke. This is the only \emph{required} key for
each entry in the \ckey{programs} table.
\end{confkey}

\begin{confkey}{generated\_target}{type: \type{boolean}}
This flag is to denote whether or not the \ckey{target} is a file that should
be generated in the \ckey{sequence} execution. For instance, DVI files, PS files,
and the input files for \prog{makeindex}, which have \code{.idx} extensions are
applicable. Conversely, files that you directly edit are not generated files.
When the value is \code{true}, \prog{llmk} will consider only the target files
newer than the time that the sequence processing starts. In case the target
file is older, it will be ignored and the program will not be called.
\end{confkey}

\begin{confkey}{opts}{type: \type{string} or \type{array of strings}}
You can specify a command-line option to give the command as a \type{string}
value or one or more options as an \type{array of strings}. The format
specifiers are available for this key.
\end{confkey}

\begin{confkey}{postprocess}{type: \type{string}}
The name of a program (in the \ckey{programs} table) to be executed as
postprocess. The specified program will be called only when the current main
program is executed.
\end{confkey}

\begin{confkey}{target}{type: \type{string}}[default: \code{"\%S"}]
The target filename of the program. The existence of the target file is checked
by \prog{llmk} right before trying to execute each program, and the programs
are invoked only when it exists. This filename is also used for replacing the
value for the format specifiers. When this key is not specified, the input
filename that is given as a command-line argument, or the filename in the value
for the \ckey{source} key is used as a target.
\end{confkey}

\subsection{Default \code{programs} and \code{sequence}}

In this section, the default values of the two important data structures for
advanced control for \prog{llmk}, \ie the default configuration for the
\ckey{programs} table and the \ckey{sequence} array. As it is already described
in Section~\ref{sec:top-level-keys}, \ckey{programs} is the table to contain
detailed configuration for each program to execute by \prog{llmk}, and
\ckey{sequence} is an array to contain the names of entries in the
\ckey{programs} table in the order of execution for a workflow. The default
values of these are designed to work well for typical {\LaTeX} documents.
Therefore, you normally do not need to change the values, but you can override
any of the values by writing new values in your TOML configuration.

\subsubsection{The default \code{programs}}
\label{sec:default-programs}

The followings are the default values in the \ckey{programs} table for each
entry, \ie program, expressed in the TOML format. The default \ckey{programs}
table contains useful settings for popular tools in the ecosystem of {\LaTeX}%
\footnote{Reuquests for new settings for other programs will be considered.}.
Only some of them are used in the default \ckey{sequence} (see Section~%
\ref{sec:default-sequence}), but other entries can be easily used just by
overriding the \ckey{sequence} array.

\Program{bibtex} The entry for the {\BibTeX} program and friends, \eg Biber.
The \progname{latex} program is set as \ckey{postprocess} so that to make sure
rerunning {\LaTeX} command after this execution.
%
\begin{lstlisting}[style=toml]
[programs.bibtex]
command = "bibtex"
target = "%B.bib"
args = ["%B"]
postprocess = "latex"
\end{lstlisting}

\Program{dvipdf} The entry for the \prog{dvipdf} program and friends.
%
\begin{lstlisting}[style=toml]
[programs.dvipdf]
command = "dvipdfmx"
target = "%B.dvi"
generated_target = true
\end{lstlisting}

\Program{dvips} The entry for the \prog{dvips} program and friends.
%
\begin{lstlisting}[style=toml]
[programs.dvips]
command = "dvips"
target = "%B.dvi"
generated_target = true
\end{lstlisting}

\Program{latex} The entry for the main {\LaTeX} programs. The default value for
the \ckey{command} is \code{"lualatex"}; since \prog{llmk} runs on
\code{texlua}, the installation of {\LuaTeX} is guaranteed. That is why the
command is chosen for the default.
%
\begin{lstlisting}[style=toml]
[programs.latex]
command = "lualatex"
opts = ["-interaction=nonstopmode", "-file-line-error", "-synctex=1"]
aux_file = "%B.aux"
aux_empty_size = 9
\end{lstlisting}

\Program{makeindex} The entry for the Makeindex program and friends. The
\progname{latex} program is set as \ckey{postprocess} so that to make sure
rerunning {\LaTeX} command after this execution.
%
\begin{lstlisting}[style=toml]
[programs.makeindex]
command = "makeindex"
target = "%B.idx"
generated_target = true
postprocess = "latex"
\end{lstlisting}

\Program{ps2pdf} The entry for the \prog{ps2pdf} program and friends.
%
\begin{lstlisting}[style=toml]
[programs.ps2pdf]
command = "ps2pdf"
target = "%B.ps"
generated_target = true
\end{lstlisting}

\subsubsection{The default \code{sequence}}
\label{sec:default-sequence}

The following is the default value for the \ckey{sequence} array:
%
\begin{htcode}
["latex", "bibtex", "makeindex", "dvipdf"]
\end{htcode}

With these default settings in the \ckey{programs} table the \ckey{sequence}
array, the default behavior of \prog{llmk} can be summarized as follows.
%
\begin{enumerate}
\item It runs the {\LaTeX} command, by default {\LuaLaTeX}, against your
  {\LaTeX} source file. This execution may be repeated when the auxiliary file
  exists until no change made for the file to resolve all cross-references and
  so on.
\item If the corresponding {\BibTeX} database file, whose name matches to
  \code{\%B.bib}, exists, the {\BibTeX} program is executed. When the execution
  occurs, the \progname{latex} program is executed again right after because
  the program is set as \ckey{postprocess}.
\item Identically, if the corresponding input file for Makeindex exists, the
  program is executed. When the execution occurs, the \progname{latex} program
  is executed again right after because the program is set as \ckey{postprocess}.
\item In case the corresponding DVI file is generated in the previous steps,
  though this will not happen with the default value of \ckey{latex}, \ie
  \code{"lualatex"}, the \progname{dvipdf} program, by default \code{dvipdfmx},
  is executed to produce the final PDF file.
\end{enumerate}

The entries that are not used within the default \ckey{sequence} are to make it
easier to use other tools from the above. For instance, if you wan to use
$\text{\progname{dvips}}+\text{\progname{ps2pdf}}$ combination instead of
\progname{dvipdf}, you can just modify the value of \ckey{sequence} a bit:
%
\begin{lstlisting}[style=toml]
sequence = ["latex", "bibtex", "makeindex", "dvips", "ps2pdf"]
\end{lstlisting}

\section{Other supported formats}
\label{sec:magic-comment}

In addition to the TOML format, \prog{llmk} also supports a few other magic
comment formats that are supported in some existing tools. These features are
for user convenience, but note that the aim of \prog{llmk} is not to behave
perfectly compatible with other tools.

For \prog{llmk}, in your {\LaTeX} file, configuration in TOML format has the
highest priority: when a TOML field is found, all the other magic comments will
be ignored. The precedence of magic comment formats is as follows. Remember
that all of these will be ignored if \prog{llmk} uses the configuration in a
special file \code{llmk.toml}.
%
\begin{enumerate}
\item TOML field (Section~\ref{sec:toml})
\item {\TeX}Shop directives (Section~\ref{sec:ts-directive})
\item Shebang-like directive (Section~\ref{sec:shebang})
\end{enumerate}

\subsection{{\TeX}Shop directives}
\label{sec:ts-directive}

{\TeX}Shop~\cite{texshop}, a {\TeX}-oriented IDE, understands special
directives, which typically start with \code{\%\textvisiblespace !TEX}, to
specify a workflow for processing {\LaTeX} document, \eg which {\TeX} engine to
use. Similar directives are also supported in a few other tools such as {\TeX}%
works~\cite{texworks} and {\TeX}studio~\cite{texstudio}.

Among several variation of the directives, \prog{llmk} supports two of them.
One is the directive to specify a variant of {\TeX} engine. The formal syntax
of the directive is:
%
\bgroup
\newcommand{\OS}{\meta{optional spaces}}
\newcommand{\VS}{\textvisiblespace}
\begin{htcode}
\meta{TS prefix}TEX\VS{\OS}program\meta{equals}\meta{command}\OS
\meta{TS prefix}TeX\VS{\OS}program\meta{equals}\meta{command}\OS
\meta{TS prefix}TEX\VS{\OS}TS-program\meta{equals}\meta{command}\OS
\end{htcode}
\egroup
%
where the definitions of \meta{TS prefix} and \meta{equals} are given as
follows.
%
\bgroup
\newcommand{\OS}{\meta{optional spaces}}
\begin{align*}
&\text{\meta{TS prefix}}
  \longrightarrow \text{\code{\OS\%\OS!\OS}} \\
&\text{\meta{equals}}
  \longrightarrow \text{\code{\OS=\OS}}
\end{align*}
\egroup
%
The \meta{command} part will be passed to the \ckey{latex} key of \prog{llmk}.
The other supported {\TeX}Shop directive is that for the {\BibTeX} program. The
syntax is very similar to the first one:
%
\bgroup
\newcommand{\OS}{\meta{optional spaces}}
\newcommand{\VS}{\textvisiblespace}
\begin{htcode}
\meta{TS prefix}BIB\VS{\OS}program\meta{equals}\meta{command}\OS
\meta{TS prefix}BIB\VS{\OS}TS-program\meta{equals}\meta{command}\OS
\end{htcode}
\egroup
%
The \meta{command} part will be passed to the \ckey{bibtex} key of \prog{llmk}.
For both of the two directives, only the topmost ones are effective; the others
will be ignored. For example, the following two configuration are equivalent:\\
%
\begin{minipage}[t]{.5\textwidth}
\begin{lstlisting}[style=latex]
% !TEX TS-program = xelatex
% !BIB TS-program = biber
\documentclass{article}
\end{lstlisting}
\end{minipage}
\begin{minipage}[t]{.49\textwidth}
\begin{lstlisting}[style=latex]
% +++
% latex = "xelatex"
% bibtex = "biber"
% +++
\documentclass{article}
\end{lstlisting}
\end{minipage}

\subsection{Shebang-like directive}
\label{sec:shebang}

A few existing tools, notably the YaTeX mode for Emacs~\cite{yatex}, support
the so-called shebang-like directive. This is also supported by \prog{llmk}.
The syntax is:
%
\begin{htcode}
\meta{optional spaces}%#!\meta{optional spaces}\meta{command}\meta{optional spaces}
\end{htcode}
%
The \meta{command} part will be passed to the \ckey{latex} key of \prog{llmk}.
This directive is only effective strictly in the first line in your {\LaTeX}
source file.

For example, the following two configuration are equivalent:\\
%
\begin{minipage}[t]{.5\textwidth}
\begin{lstlisting}[style=latex]
%#!pdflatex
\documentclass{article}
\end{lstlisting}
\end{minipage}
\begin{minipage}[t]{.49\textwidth}
\begin{lstlisting}[style=latex]
% +++
% latex = "pdflatex"
% +++
\documentclass{article}
\end{lstlisting}
\end{minipage}

\section{Acknowledgements}

This project has been supported by the {\TeX} Development Fund created by the {\TeX}
Users Group (No.~29). I would like to thank all contributors and the people who
gave me advice and suggestions for new features for the \prog{llmk} project.

\section{License}

This software is released under the MIT license:
%
\begin{quotation}
\parindent=0pt
\parskip=\baselineskip
\small\ttfamily
\noindent
\printLICENSE
\end{quotation}

\subsection*{Third-party software}

\paragraph{toml.lua} Copyright 2017 Jonathan Stoler. Released under the MIT
license:
%
\begin{quote}
\url{https://github.com/jonstoler/lua-toml/blob/master/LICENSE}
\end{quote}

\begin{thebibliography}{9}
\bibitem{asakura2020}
  Takuto Asakura. \textit{The design concept for \prog{llmk}\Dash Light {\LaTeX}
  Make}. TUGboat, Volume~41, No.~2. (2020)

\bibitem{arara}
  Paulo Cereda, et al. \textit{arara\Dash The cool {\TeX} automation
  tool}. \url{https://ctan.org/pkg/arara}

\bibitem{latexmk}
  John Collins. \textit{latexmk\Dash generate {\LaTeX} document}.
  \url{https://ctan.org/pkg/latexmk}

\bibitem{yatex}
  Yuuji Hirose. \textit{YaTeX\Dash Yet Another TeX mode for Emacs}.
  \url{https://www.yatex.org/}

\bibitem{texworks}
  Jonathan Kew, Stefan L\"offler, and Charlie Sharpsteen.
  \textit{{\TeX}works\Dash lowering the entry barrier to the {\TeX} world}.
  \url{https://tug.org/texworks/}

\bibitem{texshop}
  Richard Koch, et al. \textit{{\TeX}Shop}.
  \url{https://pages.uoregon.edu/koch/texshop/}

\bibitem{semvar}
  Tom Preston-Werner. \textit{Semantic Versioning 2.0.0}.
  \url{https://semver.org/}

\bibitem{toml}
  Tom Preston-Werner. \textit{TOML: Tom's Obvious Minimal Language}.
  \url{https://toml.io/}

\bibitem{texstudio}
  Benito van der Zander, et al.
  \textit{{\TeX}studio\Dash {\LaTeX} made comfortable}.\\
  \url{https://texstudio.org/}
\end{thebibliography}

\end{document}
% vim: set spell: