Introduction

Hello there, welcome to arara, the cool TeX automation tool! This chapter is actually a quick introduction to what you can (and cannot) expect from arara. For now, concepts will be informally presented and will be detailed later on, in the next chapters.

What is this tool?

Good question! arara is a TeX automation tool based on rules and directives. It is, in some aspects, similar to other well-known tools like latexmk and rubber. The key difference (and probably the selling point) might be the fact that arara aims at explicit instructions in the source code (in the form of comments) in order to determine what to do instead of relying on other resources, such as log file analysis. It is a different approach for an automation tool, and we have both advantages and disadvantages of such design. Let us use the following file hello.tex as an example:

\documentclass{article}

\begin{document}
Hello world!
\end{document}

How would one successfully compile hello.tex with latexmk and rubber, for instance? It is quite straightforward: it is just a matter of providing the file to the tool and letting it do the hard work:

$ latexmk -pdf hello.tex
$ rubber --pdf hello.tex

The mentioned tools perform an analysis on the file and decide what has to be done. However, if one tries to invoke arara on hello.tex, I am afraid nothing will be generated; the truth is, arara does not know what to do with your file, and the tool will even raise an error message complaining about this issue:

$ arara hello.tex
  __ _ _ __ __ _ _ __ __ _
 / _` | '__/ _` | '__/ _` |
| (_| | | | (_| | | | (_| |
 \__,_|_|  \__,_|_|  \__,_|

Processing "hello.tex" (size: 70 B, last modified: 12/28/2020
07:03:16), please wait.

  ERROR

It looks like no directives were found in the provided file. Make
sure to include at least one directive and try again.

Total: 0.04 seconds

Quite surprising. However, this behaviour is not wrong at all, it is completely by design: arara needs to know what you want. And for that purpose, you need to tell the tool what to do.

A very important concept
That is the major difference of arara when compared to other tools: it is not an automatic process and the tool does not employ any guesswork on its own. You are in control of your documents; arara will not do anything unless you teach it how to do a task and explicitly tell it to execute the task.

Now, how does one tell arara to do a task? That is actually the easy part, provided that you have everything up and running. We accomplish the task by adding a special comment line, hereafter known as directive, somewhere in our hello.tex file (preferably in the first lines):

% arara: pdflatex
\documentclass{article}

\begin{document}
Hello world!
\end{document}

For now, do not worry too much about the terms, we will come back to them later on, in Important concepts. It suffices to say that arara expects you to provide a list of tasks, and this is done by inserting special comments in the source file. Let us see how arara behaves with this updated code:

$ arara hello.tex
  __ _ _ __ __ _ _ __ __ _
 / _` | '__/ _` | '__/ _` |
| (_| | | | (_| | | | (_| |
 \__,_|_|  \__,_|_|  \__,_|

Processing "hello.tex" (size: 88 B, last modified: 12/28/2020
07:05:05), please wait.

(PDFLaTeX) PDFLaTeX engine .............................. SUCCESS

Total: 0.56 seconds

Hurrah, we finally got our document properly compiled with a TeX engine by the inner workings of our beloved tool, resulting in an expected hello.pdf file created using the very same system call that typical automation tools like latexmk and rubber use. Observe that arara works practically on other side of the spectrum: you need to tell it how and when to do a task.

Core concepts

When adding a directive in our source code, we are explicitly telling the tool what we want it to do, but I am afraid that is not sufficient at all. So far, arara knows what to do, but now it needs to know how the task should be done. If we want arara to run pdflatex on hello.tex, we need to have instructions telling our tool how to run that specific application. This particular sequence of instructions is referred as a rule in our context.

Note on rules
Although the core team provides a lot of rules shipped with arara out of the box, with the possibility of extending the set by adding more rules, some users might find this decision rather annoying, since other tools have most of their rules hard-coded, making the automation process even more transparent. However, since arara does not rely on a specific automation or compilation scheme, it becomes more extensible. The use of directives in the source code make the automation steps more fluent, which allows the specification of complex workflows much easier.

Despite the inherited verbosity of automation steps not being suitable for small documents, arara really shines when you have a document which needs full control of the automation process (for instance, a thesis or a manual). Complex workflows are easily tackled by our tool.

Rules and directives are the core concepts of arara: the first dictates how a task is done, and the latter is the proper instance of the associated rule on the current document, i.e, when and where the commands must be executed.

The name
We love araras!

Do you like araras? We do, specially our tool which shares the same name of this colorful bird.

The tool name was chosen as an homage to a Brazilian bird of the same name, which is a macaw. The word arara comes from the Tupian word a'rara, which means big bird (much to my chagrin, Sesame Street's iconic character Big Bird is not a macaw; according to some sources, he claims to be a golden condor). Araras are colorful, noisy, naughty and very funny. Everybody loves araras. The name seemed catchy for a tool and, in the blink of an eye, arara was quickly spread to the whole TeX world.

Now that we informally introduced rules and directives, let us take a look on how arara actually works given those two elements. The whole idea is pretty straightforward, and I promise to revisit these concepts later on in this manual for a comprehensive explanation (more precisely, in Important concepts).

First and foremost, we need to add at least one instruction in the source code to tell arara what to do. This instruction is named a directive and it will be parsed during the preparation phase. Observe that arara will tell you if no directive was found in a file, as seen in our first interaction with the tool.

An arara directive is usually defined in a line of its own, started with a comment (denoted by a percent sign in TeX and friends), followed by the word arara: and task name:

% arara: pdflatex
\documentclass{article}
...

Our example has one directive, referencing pdflatex. It is important to observe that the pdflatex identifier does not represent the command to be executed, but the name of the rule associated with that directive.

Once arara finds a directive, it will look for the associated rule. In our example, it will look for a rule named pdflatex which will evidently run the pdflatex command line application. Rules are YAML files named according to their identifiers followed by the yaml extension and follow a strict structure. This concept is covered in Important concepts.

Now, we have a queue of pairs (directive, rule) to process. For each pair, arara will map the directive to its corresponding rule, evaluate it and run the proper command. The execution chain requires that command i was successfully executed to then proceed to command i + 1, and so forth. This is also by design: arara will halt the execution if any of the commands in the queue had raised an error. How does one know if a command was successfully executed? arara checks the corresponding exit status available after a command execution. In general, a successful execution yields 0 as its exit status.

In order to decide whether a command execution is successful, arara relies on exit status checking. Typically, a command is successful if, and only if, its resulting exit status is 0 and no other value. However, we can define any value, or even forget about it and make it always return a valid status regardless of execution (for instance, in a rule that always is successful — see, for instance, the clean rule).

That is pretty much how arara works: directives in the source code are mapped to rules. These pairs are added to a queue. The queue is then executed and the status is reported. More details about the expansion process are presented in Important concepts. In short, we teach arara to do a task by providing a rule, and tell it to execute it through directives in the source code.

Operating system remarks

The application is written using the Kotlin language, so arara runs on top of a Java virtual machine, available on all the major operating systems — in some cases, you might need to install the proper virtual machine. We tried very hard to keep both code and libraries compatible with older virtual machines or from other vendors. Currently, arara is known to run on Java 8 to 18, from any vendor.

Outdated Java virtual machines
Dear reader, beware of outdated software, mainly Java virtual machines! Although arara offers support for older virtual machines, try your best to keep your software updated as frequently as possible. The legacy support exists only for historical reasons, and also due to the sheer fact that we know some people that still runs arara on very old hardware. If you are not in this particular scenario, get the latest virtual machine.

In Building arara, we provide instructions on how to build arara from sources using Gradle. Even if you use multiple operating systems, arara should behave the same, including the rules. There are helper functions available in order to provide support for system-specific rules based on the underlying operating system.

Support

If you run into any issue with arara, please let us know. We all have very active profiles in the TeX community at StackExchange, so just use the arara tag in your question and we will help you the best we can (also, take a look at their starter guide).

We also have a Matrix and Gitter chat rooms, in which we occasionally hang out. Also, if you think the report is worthy of an issue, open one in our GitLab repository.

We really hope you like our humble contribution to the TeX community. Let arara enhance your TeX experience, it will help you when you will need it the most. Enjoy the manual.