AsciiDoc User Guide

by Stuart Rackham
<srackham@methods.co.nz>
v6.0.3, 20 April 2005

AsciiDoc is a text document format for writing short documents, articles, books and UNIX man pages. AsciiDoc files can be translated to HTML (with or without stylesheets), DocBook and LinuxDoc markup using the asciidoc(1) command. AsciiDoc is highly configurable: both the AsciiDoc source file syntax and the backend output markups (which can be almost any type of SGML/XML markup) can be customized and extended by the user.

1. Introduction

This is a large rambling document, it really needs to be factored into a Tutorial, FAQ, Quick Reference and Formal Reference.

If you're new to AsciiDoc read this section and the Getting Started section and take a look at the example AsciiDoc *.txt source files in the distribution examples and doc directories.

Plain text is the most universal electronic document format, no matter what computing environment you use, you can always read and write plain text documentation. But for many applications plain text is not a viable presentation format. HTML, PDF and roff (roff is used for man pages) are the most widely used Unix presentation formats. DocBook is rapidly becoming the most popular Unix documentation markup format (DocBook is easily translated to HTML, PDF and many other presentation formats).

AsciiDoc is a plain text human readable/writable document format that can be translated directly to DocBook and HTML using the asciidoc(1) command. You can then either use asciidoc(1) generated HTML directly or run asciidoc(1) DocBook output through your favorite DocBook toolchain to produce PDF, HTML, RTF and even HTML Help presentation formats.

The AsciiDoc format is a useful presentation format in it's own right: AsciiDoc files are unencumbered by markup and is easily viewed, proofed and edited.

AsciiDoc is light weight: it consists of a single Python script and a bunch of configuration files. Apart from asciidoc(1) and a Python interpreter, no other programs are required to convert AsciiDoc text files to DocBook, HTML and LinuxDoc. See Example AsciiDoc Documents below.

You write an AsciiDoc document the same way you would write a normal text document, there are no markup tags or weird notations. Built-in AsciiDoc formatting rules have been kept to a minimum and are fairly obvious.

Text markup conventions tend to be a matter of (often strong) personal preference: if the default syntax is not to your liking you can define your own by editing the text based asciidoc(1) configuration files. You can create other backend formats to translate AsciiDoc documents to almost any SGML/XML markup.

asciidoc(1) comes with a set of configuration files to translate AsciiDoc files to HTML (articles, books, man pages), DocBook (articles, books, man pages) or LinuxDoc (articles).

My AsciiDoc Itch

DocBook has emerged as the defacto standard Open Source documentation format. But DocBook is a complex language, the marked up text is difficult to read and even more difficult to write directly, and I found I was spending far to much time consulting reference manuals and fixing syntax errors than actually writing the documentation.

2. Getting Started

AsciiDoc is written in Python so you need a Python interpreter (version 2.3 or later) to execute asciidoc(1). Python is installed by the default configurations of most FreeBSD and Linux distributions. You can download Python from the official Python website http://www.python.org.

Debian Users

There's an AsciiDoc Debian package available. Just install it using your favorite Debian installer and then skip to the next section.

The tarball contains the executable asciidoc.py script, configuration files, examples and documentation.

Test out asciidoc by changing to the AsciiDoc application directory and converting the User Guide document (./doc/asciidoc.txt) to HTML (./doc/asciidoc.html):

$ ./asciidoc.py -b html doc/asciidoc.txt   # Plain HTML
$ ./asciidoc.py -b css doc/asciidoc.txt    # HTML with CSS stylesheets

I use .txt file extensions for AsciiDoc document files.

If you want to avoid having to explicitly specify the location of asciidoc every time you run it you'll need to make sure that asciidoc.py can be found in your search path, you can do this by putting a symbolic link to asciidoc in a directory in your search path. For example (assuming ~/bin is in $PATH):

$ ln -s ~/bin/asciidoc-6.0.3/asciidoc.py ~/bin/asciidoc

You can now execute asciidoc(1) by typing asciidoc.

2.1. Example AsciiDoc Documents

Examples of AsciiDoc markup are scattered throughout this document but the best way to quickly get a feel for AsciiDoc is to view the AsciiDoc web site and/or distributed examples:

3. AsciiDoc Backends

The asciidoc(1) command translates an AsciiDoc formatted file to the backend format specified by the -b command-line option. asciidoc(1) itself little intrinsic knowledge of backend formats, all translation rules are contained in user customizable cascading configuration files.

AsciiDoc ships with the following predefined backend output formats:

3.1. docbook

AsciiDoc generates DocBook article, book and refentry documents (corresponding to the asciidoc(1) article, book and manpage document types).

As of version 4.2 the -b docbook command-line option switched from outputting DocBook SGML to the newer DocBook XML. You can still produce DocBook SGML using the -b docbook-sgml command-line option.

The DocBook DTD restricts the allowable AsciiDoc syntax:

3.2. xhtml

The asciidoc(1) -b xhtml command-line option produces plain unstyled XHTML 1.0 markup that can be viewed on any modern web browser.

3.3. html

A minor variation of xhtml — the asciidoc(1) -b html command-line option produces HTML 4 markup.

3.4. css

The asciidoc(1) -b css command-line option produces XHTML 1.0 conformant output styled with linked CSS2 stylesheets.

The main.css, article.css and manpage.css stylesheets (located in the ./stylesheets/ directory) style both screen and printed outputs. Edit the stylesheets to customize your document's appearance.

NOTE:

Browser CSS support varies from browser to browser. The examples work well on IE6, Firefox 1.0 and up, Mozilla 1.1 and up, Opera 7 and Konqueror 3.01 but have not been tested on other browsers. All browsers have their quirks. When it comes to implementing CSS, of those tested I found Mozilla and Firefox to be the most conformant.

3.5. css-embedded

The asciidoc(1) -b css-embedded command-line option produces XHTML 1.0 conformant output styled with embedded CSS2 stylesheets. Similar to the css backend but translated files do not rely on external stylesheets.

3.6. linuxdoc

The LinuxDoc DTD restricts the allowable AsciiDoc syntax:

4. Converting to other Presentation Formats

DocBook documents are not designed to be viewed directly. FreeBSD and most Linux distributions come with conversion tools (collectively called a tool chain) for converting DocBook files to presentation formats such as Postscript, HTML, PDF, DVI, roff (the native man page format), HTMLHelp, JavaHelp and text.

4.1. DocBook Conversion

DocBook Tool Chains

DocBook files are validated, parsed and translated by a combination of applications collectively called a DocBook tool chain. The core function of a tool chain is to read the DocBook markup (produced by AsciiDoc) and transform it to a presentation format (for example HTML, PDF, HTML Help) using a set of rules and stylesheets.

A wide range of user output format requirements coupled with a choice of available tools and stylesheets results in many valid tool chain combinations.

Assembling a working DocBook tool chain requires a fairly detailed understanding of how the tools work together. Popular Linux distributions come with both toolchains and toolchain wrapper scripts.

Wrapper scripts tie together the various toolchain commands and simplify the underlying complexity. Recent Redhat/Fedora distributions have shipped fully configured toolchains along with both the xmlto(1) and the jw(1) toolchain wrappers (which process DocBook XML and DocBook SGML documents using XSL and DSSSL stylesheets respectively). xmlto(1) and related tools are standard Cygwin packages and are recommended for Microsoft Windows users.

Immediately below you'll find examples using toolchain wrapper scripts — explicit use of DocBook toolchain commands can be found in the Processing DocBook Files section.

The default xmlto(1) and jw(1) outputs are quite plain (compared to the distributed AsciiDoc HTML and PDF documentation files). The Processing DocBook Files section explains how you can generate nicely styled output using custom DocBook XSL Stylesheets drivers. This is the best route for generating PDF outputs.

WARNING:

I don't recommend using xmlto(1) for producing PDF — the output is not bookmarked and callouts generate LaTeX errors.

If you use jw(1) don't forget to specify the docbook-sgml backend, not the docbook backend. This example converts the asciidoc.txt AsciiDoc User Guide to PDF format:

$ asciidoc -b docbook-sgml asciidoc.txt
$ jw -b pdf asciidoc.sgml

To convert the asciidoc.1.txt AsciiDoc manpage document to native man page groff(1) man macro package format:

$ asciidoc -d manpage -b docbook asciidoc.1.txt
$ xmlto man asciidoc.1.xml

To view the man page file as it would be displayed by the man(1) command:

$ groff -mandoc -Tascii asciidoc.1 | less

To print a high quality man page to a postscript printer:

$ groff -mandoc -Tps asciidoc.1 | lpr

You can also produce HTML from DocBook files:

$ asciidoc -b docbook asciidoc.txt
$ xmlto html-nochunks asciidoc.xml      # Single HTML file.
$ xmlto html -o chunked asciidoc.xml    # Chunked HTML file.

Why Generate HTML via DocBook?

AsciiDoc produces nicely styled HTML directly without requiring a DocBook toolchain so why would you want generate it via DocBook?

Plain and CSS styled HTML output directly from AsciiDoc is still very useful in the following situations:

4.2. LinuxDoc Conversion

LinuxDoc is an SGML documentation markup language originally created by Matt Welsh to write Linux documentation. LinuxDoc does a good job of marking up small and medium sized text based documents. Its strength is its simplicity and ease of use. Nowadays LinuxDoc has been largely superseded by DocBook.

NOTE:

LinuxDoc limitations place restrictions on the use of a number of AsciiDoc elements.

There are a number of Open Source applications available to convert the LinuxDoc SGML markup to various presentation formats, here are a couple of examples:

Create a single HTML file with a detailed table of contents using the linuxdoc(1) command that comes with the Linuxdoc-Tools package.

$ linuxdoc -B html -s 0 -T 2 mydocument.sgml

Generate a set of linked HTML files using the sgmlfmt(1) from the FreeBSD sgmlformat package:

$ sgmlfmt -f html mydocument.sgml

Create a PDF file using the sgmlfmt(1) from the FreeBSD sgmlformat package:

$ sgmlfmt -f ps mydocument.sgml
$ ps2pdf asciidoc.ps

4.3. Generating Plain Text Files

AsciiDoc does not have a text backend (for most applications AsciiDoc source text is fine), however you can convert asciidoc(1) generated HTML and DocBook files to text.

You can use the lynx(1) web browser to convert AsciiDoc generated HTML to text. You'll find an asciidoc2text.sh shell script included in the AsciiDoc distribution examples/asciidoc2text directory which, together with the asciidoc2text.conf configuration file, automates text file generation with a single command. For example:

$ ./examples/asciidoc2text/asciidoc2text.sh test.txt >test.text

You can also use jw(1) and xmlto(1) toolchain commands to convert DocBook to text.

WARNING:

By default both jw(1) and xmlto(1) commands create files with a .txt extension when generating text files. This will overwrite your AsciiDoc *.txt source files (use the -o command-line option to output to an alternative directory).

5. AsciiDoc Document Types

There are three types of AsciiDoc documents: article, book and manpage. All document types share the same AsciiDoc format with some minor variations.

Use the asciidoc(1) -d option to specify the AsciiDoc document type (defaults to article type).

5.1. article

Used for short documents, articles and general documentation. See the article.txt example AsciiDoc article in the distribution ./doc directory.

5.2. book

Books share the same format as articles; in addition there is the option to add level 0 sections to divide a book into multiple parts.

Book documents will normally be used to produce DocBook output since DocBook processors can automatically generate footnotes, table of contents, list of tables, list of figures, list of examples and indexes.

AsciiDoc markup supports all the standard DocBook frontmatter and backmatter special sections (dedication, preface, bibliography, glossary, index, colophon) plus footnotes and index entries.

Example book documents

Book
The book.txt file in the AsciiDoc distribution ./doc directory.
Multi-part book
The book-multi.txt file in the AsciiDoc distribution ./doc directory.

5.3. manpage

Used to generate UNIX manual pages. AsciiDoc manpage documents observe special header title and section naming conventions — see the Manpage Documents section for details.

The asciidoc(1) man page (asciidoc.1.txt in the AsciiDoc distribution ./doc directory) is an example of an AsciiDoc man page file.

6. Document Structure

An AsciiDoc document consists of a series of block elements starting with an optional document Header, followed by an optional Preamble, followed by zero or more document Sections.

Almost any combination of zero or more elements constitutes a valid AsciiDoc document: documents can range from a single sentence to a multi-part book.

6.1. Block Elements

Block elements consist of one or more lines of text and may contain other block elements.

The AsciiDoc block structure can be informally summarized
[This is a rough structural guide, not a rigorous syntax definition]
as follows:

Document      ::= (Header?,Preamble?,Section*)
Header        ::= (Title,(AuthorLine,RevisionLine?)?)
AuthorLine    ::= (FirstName,(MiddleName?,LastName)?,EmailAddress?)
RevisionLine  ::= (Revision?,Date)
Preamble      ::= (SectionBody)
Section       ::= (Title,SectionBody?,(Section)*)
SectionBody   ::= ((BlockTitle?,
                  (Paragraph|DelimitedBlock|List|Table|BlockMacro))+)
List          ::= (BulletedList|NumberedList|LabeledList|CalloutList)
BulletedList  ::= (ListItem)+
NumberedList  ::= (ListItem)+
CalloutList   ::= (ListItem)+
LabeledList   ::= (ItemLabel+,ListItem)+
ListItem      ::= (ItemText,((List|ListParagraph)?,ItemContinuation?)*)
Table         ::= (Ruler,TableHeader?,TableBody,TableFooter?)
TableHeader   ::= (TableRow+,TableUnderline)
TableFooter   ::= (TableRow+,TableUnderline)
TableBody     ::= (TableRow+,TableUnderline)
TableRow      ::= (TableData+)

6.2. Header

The Header is optional but and starts on first line of the document beginning with a document title. Immediately following the title are optional Author and Revision lines.

The author line contains the author's name optionally followed by the author's email address. The author's name consists of a first name followed by optional middle and last names separated by white space. The email address is last and must be enclosed in angle <> brackets. Author names cannot contain angle <> bracket characters.

The optional document header revision line should immediately follow the author line. The revision line can be one of two formats:

  1. A an alphanumeric document revision number followed by a date:

    • The revision number and date must be separated by a comma.

    • The revision number is optional but must contain at least one numeric character.

    • Any non-numeric characters preceding the first numeric character will be dropped.

  2. An RCS $Id$ marker.

The document heading is separated from the remainder of the document by one or more blank lines.

Here's an example AsciiDoc document header:

Writing Documentation using AsciiDoc
====================================
Stuart Rackham <srackham@methods.co.nz>
v2.0, February 2003

You can override or set header parameters by passing revision, data, email, author, authorinitials, firstname and lastname attributes using the asciidoc(1) -a command-line option. For example:

$ asciidoc -b docbook -a date=2004/07/27 article.txt

Attributes can also be added to the header for substitution in the header template with Attribute Entry elements.

6.3. Preamble

The Preamble is an optional untitled section body between the document Header and the first Section title. The Preamble should only be included in article documents.

6.4. Sections

AsciiDoc supports five section levels which corresponding to document levels 0 to 4 (although only book documents are allowed to contain level 0 sections). Section levels are delineated by the section title underlines.

A section consists of a section title followed by an optional section body.

Sections are translated using configuration file markup templates. To determine which configuration file section to use AsciiDoc first searches for section titles in the [specialsections] configuration entries, if not found it looks for the name [sect<level>].

You can the -n command-line option to auto-number HTML outputs (DocBook line numbering is handled automatically by the DocBook toolchain commands).

6.4.1. Special Sections

In addition to nested sections documents generally have frontmatter and backmatter sections with special semantic significance, for example: preface, bibliography, table of contents, index.

AsciiDoc configuration files can have a [specialsections] section which specifies special section titles and the corresponding backend markup.

[specialsections] entries are formatted like:

<pattern>=<name>

<pattern> is a Python regular expression and <name> is the name of a configuration file markup template section. If the <pattern> matches an AsciiDoc document section title then the backend output is marked up using the <name> markup template. The {title} attribute value is set to the value of the matched regular expression group named title, if there is no title group {title} is set to the the whole of the section title.

The special section names in the default [specialsections] section are:

Preface                    (book documents only)
Abstract                   (article documents only)
Dedication                 (book documents only)
Glossary
Bibliography|References
Colophon                   (book documents only)
Index
Appendix [A-Z][:.] <title>

6.5. Inline Elements

Inline document elements occur within block elements; inline elements can begin and end anywhere within a line but cannot span multiple lines.

Inline elements are used to markup character formatting and various types of text substitution. Inline elements and inline element syntax is defined in the asciidoc(1) configuration files.

Here is a list of AsciiDoc inline elements in the (default) order in which they are processed:

Special characters
These character sequences escape special characters used by the backend markup (typically "<", ">", and "&"). See [specialcharacters] configuration file sections.
Quotes
Characters that markup words and phrases; usually for character formatting. See [quotes] configuration file sections.
Special Words
Word or word phrase patterns singled out for markup without the need for further annotation. See [specialwords] configuration file sections.
Replacements
Each Replacement defines a word or word phrase pattern to search for along with corresponding replacement text. See [replacements] configuration file sections.
Attributes
Document attribute names enclosed in braces (attribute references) are replaced by the corresponding attribute value.
Inline Macros
Inline macros are replaced by the contents of parameterized configuration file sections.

7. Document Processing

The AsciiDoc source document is read and processed as follows:

  1. The document Header is parsed, header parameter values are substituted into the configuration file [header] template section which is then written to the output file.

  2. Each document Section is processed and it's constituent elements parsed and translated to the output file.

  3. The configuration file [footer] template section is substituted and written to the output file.

When a block element is encountered asciidoc(1) determines the type of block by checking in the following order (first to last): BlockTitles, (section) Titles, BlockMacros, Lists, DelimitedBlocks, Tables, AttributeEntrys, AttributeLists, Paragraphs.

The default paragraph definition [paradef-default] is last element to be checked.

Knowing the parsing order will help you devise unambiguous macro, list and block syntax rules.

Inline substitutions within block elements are performed in the following default order:

  1. Special characters

  2. Quotes

  3. Special words

  4. Replacements

  5. Attributes

  6. Inline Macros

The substitutions and substitution order performed on Title, Paragraph and DelimitedBlock elements is determined by configuration file parameters.

8. Text Formatting

8.1. Quoted Text

Words and phrases can be formatted by enclosing the text with predefined quoting characters:

Emphasized text
Word phrases 'enclosed in single quote characters' (acute accents) are emphasized.
Strong text
Word phrases *enclosed in asterisk characters* are rendered in a strong font (usually bold).
Monospaced text
Word phrases `enclosed in backtick characters` (grave accents) are rendered in a monospaced font.

Quoting characters can be changed and new quoting markup syntax defined by editing asciidoc(1) configuration files. See the Configuration Files section for details.

Quoted text properties

8.2. Superscripts and Subscripts

Put carets on either side of the text to be superscripted, put tildes on either side of text to be subscripted. For example, the following line:

e^{amp}#960;i^+1 = 0. H~2~O and x^10^. Some ^super text^
and ~some sub text~

Is rendered like:

eπi+1 = 0. H2O and x10. Some super text and some sub text

If you want to display caret (^) or tilde (~) characters you need to ensure only one per line otherwise they'll be misinterpreted as superscripting and subscripting.

Superscripts and subscripts are implemented as Replacements substitutions.

8.3. Line Breaks (HTML)

A plus character preceded by at least one space character at the end of a line forces a line break. It generates an HTML line break (<br/>) tag. Line breaks are ignored when outputting to DocBook since it has no line break element.

8.4. Rulers (HTML)

A line of three or more apostrophe characters will generate an HTML ruler (<hr/>) tag. Ignored when generating non-HTML output formats.

8.5. Tabs

By default tab characters input files will translated to 8 spaces. Tab expansion is set with the tabsize entry in the configuration file [miscellaneous] section and can be overridden in the include block macro by setting a tabsize attribute in the macro's attribute list. For example:

include::addendum.txt[tabsize=2]

The tab size can also be set using the -a command-line option, for example -a tabsize=4

8.6. Replacements

The following replacements are defined in the default AsciiDoc configuration:

(C) copyright, (TM) trademark, (R) registered trademark,
-- em dash, ... ellipsis.

Is rendered as:

© copyright, ™ trademark, ® registered trademark, — em dash, … ellipsis.

The Configuration Files section explains how to configure your own replacements.

8.7. Special Words

Words defined in [specialwords] configuration file sections are automatically marked up without having to be explicitly notated.

The Configuration Files section explains how to add and replace special words.

9. Titles

Document and section titles consist of one or two lines.

9.1. Two line titles

A two line title consists of a title line, starting hard against the left margin, and an underline. Section underlines consist a repeated character pairs spanning the width of the preceding title (give or take up to three characters):

The default title underlines for each of the document levels are:

Level 0 (top level):     ======================
Level 1:                 ----------------------
Level 2:                 ~~~~~~~~~~~~~~~~~~~~~~
Level 3:                 ^^^^^^^^^^^^^^^^^^^^^^
Level 4 (bottom level):  ++++++++++++++++++++++

Examples:

Level One Section Title
-----------------------
Level 2 Subsection Title
~~~~~~~~~~~~~~~~~~~~~~~~

9.2. One line titles

One line titles consist of a line starting with one or more equals characters (the exact number specified the section level) followed by a space followed by the section title. Here are some examples:

= Document Title
== Section level 1 title (top level section)
=== Section level 2 title
==== Section level 3 title
===== Section level 4 title

The syntax can be changed by editing the configuration file [titles] section sect0sect4 entries.

10. BlockTitles

A BlockTitle element is a single line beginning with a period followed by a title. The title is applied to the next Paragraph, DelimitedBlock, List, Table or BlockMacro.For example:

.Notes
- Note 1.
- Note 2.

is rendered as:

Notes

11. BlockId Element

A BlockId is a single line block element containing a unique identifier enclosed in double square brackets. It is used to assign an identifier to the ensuing block element for use by referring links. For example:

[[chapter-titles]]
Chapter titles can be ...

The preceding example identifies the following paragraph so it can be linked from other location, for example with <<chapter-titles,chapter titles>>.

BlockId elements can be applied to Title, Paragraph, List, DelimitedBlock and BlockMacro elements. The BlockId element is really just an AttributeList with a special syntax which sets the {id} attribute for substitution in the subsequent block's markup template.

12. Paragraphs

Paragraphs are terminated by a blank line, the end of file, or the start of a DelimitedBlock.

Paragraph types are defined in configuration file [paradef*] sections. AsciiDoc ships with the following predefined paragraph types:

12.1. Default Paragraph

A Default paragraph ([paradef-default]) consists of one or more non-blank lines of text. The first line must start hard against the left margin (no intervening white space). The processing expectation of the default paragraph type is that of a normal paragraph of text.

12.2. Literal Paragraph

An Literal paragraph ([paradef-literal]) consists of one or more lines of text, where the first line is indented by one or more or space or tab characters. Literal paragraphs are rendered verbatim in a monospaced font usually without any distinguishing background or border. There is no text formatting or substitutions within Literal paragraphs apart from Special Characters and Callouts. For example:

Consul necessitatibus per id,
consetetur, eu pro everti postulant
homero verear ea mea,
qui. Movet blandit mea at,
interesset at has, eu nec.

12.3. Admonition Paragraphs

Tip, Note, Important, Warning and Caution paragraph definitions support the corresponding DocBook admonishment elements, just write a normal paragraph but place NOTE:, TIP:, IMPORTANT:, WARNING: or CAUTION: as the first word of the paragraph. For example:

NOTE: This is an example note.

Renders:

NOTE:

This is an example note.

TIP:

If your admonition is more than a single paragraph use one of the admonition variants of the ExampleBlock.

12.4. Paragraph Definitions

Paragraph translation is controlled by [paradef*] configuration file section entries. Users can define new types of paragraphs and modify the behavior of existing types by editing AsciiDoc configuration files.

Here is the shipped Default paragraph definition:

[paradef-default]
delimiter=(?P<text>\S.*)
template=paragraph

The Default paragraph definition has a couple of special properties:

  1. It must exist and be defined in a configuration file section named [paradef-default].

  2. Irrespective of its position in the configuration files the default paragraph definition is always processed last.

Available paragraph definition entries:

delimiter
A Python regular expression that matches the first line of a paragraph. This expression must contain the named group text which matches the text on the first line. Paragraphs are terminated by a blank line, the end of file, or the start of a DelimitedBlock.
template
The name of the configuration file section (markup template) that will envelope the translated paragraph contents. The pipe character in the section body is substituted for the paragraph text.
options
The only allowable option is listelement. The listelement option specifies that paragraphs of this type will be part of preceding list items.
presubs, postsubs
A comma separated list of the substitutions that are performed on the block contents. Allowed values: specialcharacters, quotes, specialwords, replacements, macros, attributes. If a filter has been specified the presubs and postsubs substitutions are performed before and after the filter is run respectively. The substitutions are processed in the order in which they are listed and can appear more than once.
filter
This optional entry specifies an executable shell command for processing paragraph text. The filter command can contain references to attributes (document attribute names enclosed in braces) which are expanded prior to filter execution.

Paragraph processing proceeds as follows:

  1. The paragraph text is aligned to the left margin.

  2. Optional presubs inline substitutions are performed on the paragraph text.

  3. If a filter command is specified it is executed and the paragraph text piped to it's standard input; the filter output replaces the paragraph text.

  4. Optional postsubs inline substitutions are performed on the paragraph text.

  5. The paragraph text is enveloped by the paragraph template and written to the output file.

13. Lists

List types

List behavior

13.1. Bulleted and Numbered Lists

Bulleted list items start with a dash followed by a space or tab character. Bulleted list syntaxes are:

- List item.
* List item.

Numbered list items start with an optional number or letter followed by a period followed by a space or tab character. List numbering is optional. Numbered list syntaxes are:

.  Integer numbered list item.
1. Integer numbered list item with optional numbering.
.. Lowercase letter numbered list item.
a. Lowercase letter numbered list item with optional numbering.

Here are some examples:

- Lorem ipsum dolor sit amet, consectetuer adipiscing elit.
  * Fusce euismod commodo velit.
  * Qui in magna commodo, est labitur dolorum an. Est ne magna primis
    adolescens. Sit munere ponderum dignissim et. Minim luptatum et
    vel.
  * Vivamus fringilla mi eu lacus.
  * Donec eget arcu bibendum nunc consequat lobortis.
- Nulla porttitor vulputate libero.
  . Fusce euismod commodo velit.
  . Vivamus fringilla mi eu lacus.
    .. Fusce euismod commodo velit.
    .. Vivamus fringilla mi eu lacus.
  . Donec eget arcu bibendum nunc consequat lobortis.
- Praesent eget purus quis magna eleifend eleifend.
  1. Fusce euismod commodo velit.
    a. Fusce euismod commodo velit.
    b. Vivamus fringilla mi eu lacus.
    c. Donec eget arcu bibendum nunc consequat lobortis.
  2. Vivamus fringilla mi eu lacus.
  3. Donec eget arcu bibendum nunc consequat lobortis.
  4. Nam fermentum mattis ante.

Which render as:

13.2. Vertical Labeled Lists

Labeled list items consist of one or more text labels followed the text of the list item.

An item label begins a line with an alphanumeric character hard against the left margin and ends with a double colon :: or semi-colon ;;.

The list item text consists of one or more lines of text starting on the line immediately following the label and can be followed by nested List or ListParagraph elements. Item text can be optionally indented.

Here are some examples:

Lorem::
  Fusce euismod commodo velit.

  Fusce euismod commodo velit.

Ipsum::
  Vivamus fringilla mi eu lacus.
  * Vivamus fringilla mi eu lacus.
  * Donec eget arcu bibendum nunc consequat lobortis.
Dolor::
  Donec eget arcu bibendum nunc consequat lobortis.
  'Suspendisse';;
    A massa id sem aliquam auctor.
  'Morbi';;
    Pretium nulla vel lorem.
  'In';;
    Dictum mauris in urna.

Which render as:

Lorem
Fusce euismod commodo velit.
Fusce euismod commodo velit.
Ipsum
Vivamus fringilla mi eu lacus.
  • Vivamus fringilla mi eu lacus.

  • Donec eget arcu bibendum nunc consequat lobortis.

Dolor
Donec eget arcu bibendum nunc consequat lobortis.
Suspendisse
A massa id sem aliquam auctor.
Morbi
Pretium nulla vel lorem.
In
Dictum mauris in urna.

13.3. Horizontal Labeled Lists

Horizontal labeled lists differ from vertical labeled lists in that the label and the list item sit side-by-side as opposed to the item under the label. Item text must begin on the same line as the label. For example:

TIP:

Used vertical labeled lists in preference to horizontal labeled lists — current PDF rendering tools do not make a good job of determining the relative column widths.

Here are some examples:

*Lorem*:: Fusce euismod commodo velit.
  Qui in magna commodo, est labitur dolorum an. Est ne magna primis
  adolescens.

  Fusce euismod commodo velit.

*Ipsum*:: Vivamus fringilla mi eu lacus.
  * Vivamus fringilla mi eu lacus.
  * Donec eget arcu bibendum nunc consequat lobortis.
*Dolor*:: Donec eget arcu bibendum nunc consequat lobortis.
  Sit munere ponderum dignissim et. Minim luptatum et vel.
  'Suspendisse';; A massa id sem aliquam auctor.
  'Morbi';;       Pretium nulla vel lorem.
  'In';;          Dictum mauris in urna.

Which render as:

Lorem Fusce euismod commodo velit. Qui in magna commodo, est labitur dolorum an. Est ne magna primis adolescens.
Fusce euismod commodo velit.
Ipsum Vivamus fringilla mi eu lacus.
  • Vivamus fringilla mi eu lacus.

  • Donec eget arcu bibendum nunc consequat lobortis.

Dolor Donec eget arcu bibendum nunc consequat lobortis. Sit munere ponderum dignissim et. Minim luptatum et vel.
Suspendisse A massa id sem aliquam auctor.
Morbi Pretium nulla vel lorem.
In Dictum mauris in urna.

13.4. Question and Answer Lists

AsciiDoc comes pre-configured with a labeled list (?? label delimiter) for generating question and answer (Q&A) lists. Example:

Question one??
        Answer one.
Question two??
        Answer two.

13.5. Glossary Lists

AsciiDoc comes pre-configured with a labeled list (:- label delimiter) for generating glossary lists. Example:

A glossary term:-
    The corresponding definition.
A second glossary term:-
    The corresponding definition.

For working examples see the article.txt and book.txt documents in the AsciiDoc ./doc distribution directory.

NOTE:

Glossary lists must be located in a glossary section to generate valid DocBook output.

13.6. Bibliography Lists

AsciiDoc comes with a predefined itemized list (+ item bullet) for generating bibliography entries. Example:

+ [[[taoup]]] Eric Steven Raymond. 'The Art of Unix
  Programming'. Addison-Wesley. ISBN 0-13-142901-9.
+ [[[walsh-muellner]]] Norman Walsh & Leonard Muellner.
  'DocBook - The Definitive Guide'. O'Reilly & Associates.
  1999. ISBN 1-56592-580-7.

The [[[<reference>]]] syntax is a bibliography entry anchor, it generates an anchor named <reference> and additionally displays [<reference>] at the anchor position. For example [[[taoup]]] generates an anchor named taoup that displays [taoup] at the anchor position. Cite the reference from elsewhere your document using <<taoup>>, this displays a hyperlink ([taoup]) to the corresponding bibliography entry anchor.

For working examples see the article.txt and book.txt documents in the AsciiDoc ./doc distribution directory.

NOTE:

Bibliography lists must be located in a bibliography section to generate valid DocBook output.

13.7. List Item Continuation

To include subsequent block elements in list items (in addition to implicitly included nested lists and Literal paragraphs) place a separator line containing a single plus character between the list item and the ensuing element. Multiple block elements (excluding section Titles and BlockTitles) may be included in a list item using this technique. For example:

Here's an example of list item continuation:

1. List item one.
+
List item one continued with a second paragraph followed by an
Indented block.
+
.................
$ ls *.sh
$ mv *.sh ~/tmp
.................
+
List item one continued with a third paragraph.

2. List item two.

   List item two literal paragraph (no continuation required).

-  Nested list (item one).

   Nested list literal paragraph (no continuation required).
+
Nested list appended list item one paragraph

-  Nested list item two.

Renders:

  1. List item one.

    List item one continued with a second paragraph followed by a Listing block.

    $ ls *.sh
    $ mv *.sh ~/tmp
    

    List item one continued with a third paragraph.

  2. List item two.

    List item two literal paragraph (no continuation required).
    
    • Nested list (item one).

      Nested list literal paragraph (no continuation required).
      

      Nested list appended list item one paragraph

    • Nested list item two.

13.8. List Continuation Block

A List Continuation block is a special delimited block which is functionally equivalent to List Item Continuation except that list items contained within the block do not require explicit + list item continuation lines:

The List Continuation Block is an experimental feature and is useful for lists with long multi-element list items. The alternative List Item Continuation is arguably uglier but is more obvious when reading AsciiDoc source — if you don't have strong feelings either way use explicit list continuation.

Example:

.List Block
--
1. List item one.

List item one continued with a second paragraph followed by an
Indented block.

.................
$ ls *.sh
$ mv *.sh ~/tmp
.................

List item one continued with a third paragraph.

2. List item two.

This paragraph is part of list item 2.
--

13.9. List Definitions

List behavior and syntax is determined by [listdef*] configuration file sections. The user can change existing list behavior and add new list types by editing configuration files.

List definition sections are characterized by the following entries:

type
This is either bulleted,numbered,labeled or callout.
delimiter
A Python regular expression that matches the first line of a list element entry. This expression must contain the named group text which matches text in the first line.
subs
A comma separated list of the substitutions that are performed on list item text and terms. Allowed values: specialcharacters, quotes, specialwords, replacements, macros, attributes, default, none.
listtag
The name of the tag that envelopes the List.
itemtag
The name of the tag that envelopes the ListItem.
texttag
The name of the tag that envelopes the text of first block element in a list item.
entrytag
The name of the tag that envelopes a labeled list entry.
labeltag
The name of the tag that envelopes a variable list term.

The tag entries map the AsciiDoc list structure to backend HTML/SGML/XML markup; see the shipped AsciiDoc .conf configuration files for examples.

14. Delimited Blocks

Delimited blocks are blocks of text enveloped by leading and trailing delimiter lines (normally a series of three or more repeated characters). The behavior of Delimited Blocks is specified by entries in configuration file [blockdef*] sections.

14.1. Predefined Delimited Blocks

AsciiDoc ships with a number of predefined DelimitedBlocks (see the asciidoc.conf configuration file in the asciidoc(1) program directory):

Predefined delimited block underlines:

CommentBlock:  //////////////////////////
BackendBlock:  ++++++++++++++++++++++++++
ListingBlock:  --------------------------
LiteralBlock:  ..........................
SidebarBlock:  **************************
QuoteBlock:    __________________________

Table: Default DelimitedBlock substitutions

Backend Listing Literal Sidebar Quote
Callouts No Yes Yes No No
Attributes Yes No No Yes Yes
Inline Macros Yes No No Yes Yes
Quotes No No No Yes Yes
Replacements No No No Yes Yes
Special chars No Yes Yes Yes Yes
Special words No No No Yes Yes

14.2. Listing Blocks

ListingBlocks are rendered verbatim in a monospaced font, they retain line and whitespace formatting and often distinguished by a background or border. There is no text formatting or substitutions within Listing blocks apart from Special Characters and Callouts. Listing blocks are often used for code and file listings.

Here's an example:

--------------------------------------
#include <stdio.h>
int main() {
        printf("Hello World!\n");
        exit(0);
}
--------------------------------------

Which will be rendered like:

#include <stdio.h>

int main() {
        printf("Hello World!\n");
        exit(0);
}

14.3. Literal Blocks

LiteralBlocks behave just like LiteralParagraphs except you don't have to indent the contents.

LiteralBlocks can be used to resolve list ambiguity. If the following list was indented it would be processed as an ordered list (not an indented paragraph):

....................
1. Item 1
2. Item 2
....................

Renders:

1. Item 1
2. Item 2

14.4. SidebarBlocks

A sidebar is a short piece of text presented outside the narrative flow of the main text. The sidebar is normally presented inside a bordered box to set it apart from the main text.

The sidebar body is treated like a normal section body.

Here's an example:

.An Example Sidebar
************************************************
Any AsciiDoc SectionBody element (apart from
SidebarBlocks) can be placed inside a sidebar.
************************************************

Which will be rendered like:

An Example Sidebar

Any AsciiDoc SectionBody element (apart from SidebarBlocks) can be placed inside a sidebar.

Sidebar elements are not supported by the LinuxDoc format and sidebar text appears as part of the main document flow.

14.5. Comment Blocks

CommentBlocks are not processed; they are useful for annotations and for excluding new or outdated content that you don't want displayed. Here's and example:

//////////////////////////////////////////
CommentBlock contents are not processed by
asciidoc(1).
//////////////////////////////////////////

See also Comment Lines.

14.6. Backend Blocks

BackendBlocks are for backend specific markup, text is only subject to attribute and macro substitution. BackendBlock content will generally be backend specific. Here's an example:

++++++++++++++++++++++++++++++++++++++
<table border="1"><tr>
  <td>Cell 1</td>
  <td>Cell 2</td>
</tr></table>
++++++++++++++++++++++++++++++++++++++

14.7. Quote Blocks

QuoteBlocks are used for quoted passages of text. attribution and citetitle named attributes specify the author and source of the quote (they are equivalent to positional attribute list entries 1 and 2 respectively). Both attributes are optional and the block body is treated like a SectionBody. For example:

[Bertrand Russell, The World of Mathematics (1956)]
____________________________________________________________________
A good notation has subtlety and suggestiveness which at times makes
it almost seem like a live teacher.
____________________________________________________________________

Which is rendered as:

A good notation has subtlety and suggestiveness which at times makes it almost seem like a live teacher.

The World of Mathematics (1956)
— Bertrand Russell

In this example unquoted positional attributes have been used, the following quoted positional and named attributes are equivalent (if the attribute list contained commas then quoting would have been mandatory):

["Bertrand Russell","The World of Mathematics (1956)"]
[attribution="Bertrand Russell",citetitle="The World of Mathematics (1956)"]

14.8. Example Blocks

ExampleBlocks encapsulate the DocBook Example element and are used, well, for examples. DocBook processors automatically number examples and generate a list of examples backmatter section.

Example blocks are delimited by lines of equals characters and you can put any block elements apart from Titles, BlockTitles and Sidebars) inside an example block. AsciiDoc example blocks allow variant delimited blocks.

As a point of interest, AsciiDoc automatically wraps the following titled (preceded by a TitleBlock) AsciiDoc elements in a DocBook example element: LiteralParagraph, LiteralBlock, ListingBlock.

14.9. Variant Blocks

Variant blocks provide a mechanism for varying the behavior of a delimited block. If the block's definition has been assigned the variants option and it's attribute list includes attribute {1} then it will be used as a prefix for the block's markup template name — the name of the markup template is a lowercased {1}block.

The ExampleBlock is configured as variant block and can be used to generate admonition blocks (admonitions requiring more than just a simple admonition paragraph) by prefixing an AttributeList and setting the first attribute entry to NOTE, TIP, WARNING, IMPORTANT or CAUTION.

The following example uses the markup template configuration file section named [noteblock]:

[NOTE]
.An example note block
=====================================================================
Qui in magna commodo, est labitur dolorum an. Est ne magna primis
adolescens.

. Fusce euismod commodo velit.
. Vivamus fringilla mi eu lacus.
  .. Fusce euismod commodo velit.
  .. Vivamus fringilla mi eu lacus.
. Donec eget arcu bibendum
  nunc consequat lobortis.
=====================================================================

Renders:

NOTE: An example note block

Qui in magna commodo, est labitur dolorum an. Est ne magna primis adolescens.

  1. Fusce euismod commodo velit.

  2. Vivamus fringilla mi eu lacus.

    1. Fusce euismod commodo velit.

    2. Vivamus fringilla mi eu lacus.

  3. Donec eget arcu bibendum nunc consequat lobortis.

14.10. Block Definitions

As an example, here is the default configuration file ListingBlock definition:

[blockdef-listing]
delimiter=^-{3,}$
template=listingblock
presubs=specialcharacters,callouts
delimiter
A Python regular expression that matches the leading and trailing block underlines. In the above example a line starting with at least three tilde characters.
template
The name of the configuration file markup template section that will envelope the block contents. The pipe | character is substituted for the block contents.
options
Allowed values are section, skip and variants.
  • If the section option is set the block contents are processed as a SectionBody.

  • The skip option causes the block to be treated as a comment (see CommentBlocks). presubs, postsubs and filter entries are not used when section or skip options are set.

  • The variants option enables variant blocks.

presubs, postsubs
A comma separated list of the substitutions that are performed on the block contents. Allowed values: specialcharacters, quotes, specialwords, replacements, macros, attributes, default, none. If a filter has been specified the presubs and postsubs substitutions are performed before and after the filter is run respectively. The substitutions are processed in the order in which they are listed and can appear more than once.
filter
This optional entry specifies an executable shell command for processing block content. The filter command can contain attribute references.

DelimitedBlock processing proceeds as follows:

  1. Optional presubs substitutions are performed on the block contents.

  2. If a filter is specified it is executed and the block's contents piped to its standard input. The filter output replaces the block contents.

  3. Optional postsubs substitutions are performed on the block contents.

  4. The block contents is enveloped by the block's markup template and written to the output file.

TIP:

Attribute expansion is performed on the block filter command before it is executed, this is useful for passing arguments to the filter.

15. Footnotes

The shipped AsciiDoc configuration includes the footnote:[<text>] inline macro for generating footnotes. The footnote text can span multiple lines. Example footnote:

footnote:[An example footnote.]

Which renders
[An example footnote.]
.

Footnotes are primarily useful when generating DocBook output — DocBook conversion programs render footnote outside the primary text flow.

16. Indexes

The shipped AsciiDoc configuration includes the inline macros for generating index entries.

indexterm:[<primary>,<secondary>,<tertiary>]
++<primary>,<secondary>,<tertiary>++
This inline macro generates an index term (the <secondary> and <tertiary> attributes are optional). For example indexterm:[Tigers,Big cats] (or, using the alternative syntax ++Tigers,Big cats++. Index terms that have secondary and tertiary entries also generate separate index terms for the secondary and tertiary entries. The index terms appear in the index, not the primary text flow.
indexterm2:[<primary>]
+<primary>+
This inline macro generates an index term that appears in both the index and the primary text flow. The <primary> should not be padded to the left or right with white space characters.

Here are some index entry examples taken from the example article.txt and book.txt documents in the AsciiDoc ./doc distribution directory.

And now for something completely different: +monkeys+, lions and
tigers (Bengal and Siberian) using the alternative syntax index
entries.
++Big cats,Lions++
++Big cats,Tigers,Bengal Tiger++
++Big cats,Tigers,Siberian Tiger++
Note that multi-entry terms generate separate index entries.

NOTE:

Index entries only really make sense if you are generating DocBook markup — DocBook conversion programs automatically generate an index at the point an Index section appears in source document (see the book.txt example document in the distribution ./doc directory).

17. Callouts

Callouts are a mechanism for annotating verbatim text (source code, computer output and user input for example). Callout markers are placed inside the annotated text while the actual annotations are presented in a callout list after the annotated text. Here's an example:

.MS-DOS directory listing
.....................................................
10/17/97   9:04         <DIR>    bin
10/16/97  14:11         <DIR>    DOS            <1>
10/16/97  14:40         <DIR>    Program Files
10/16/97  14:46         <DIR>    TEMP
10/17/97   9:04         <DIR>    tmp
10/16/97  14:37         <DIR>    WINNT
10/16/97  14:25             119  AUTOEXEC.BAT   <2>
 2/13/94   6:21          54,619  COMMAND.COM    <2>
10/16/97  14:25             115  CONFIG.SYS     <2>
11/16/97  17:17      61,865,984  pagefile.sys
 2/13/94   6:21           9,349  WINA20.386     <3>
.....................................................

<1> This directory holds MS-DOS.
<2> System startup code for DOS.
<3> Some sort of Windows 3.1 hack.

Which renders:

Example: MS-DOS directory listing

10/17/97   9:04         <DIR>    bin
10/16/97  14:11         <DIR>    DOS            (1)
10/16/97  14:40         <DIR>    Program Files
10/16/97  14:46         <DIR>    TEMP
10/17/97   9:04         <DIR>    tmp
10/16/97  14:37         <DIR>    WINNT
10/16/97  14:25             119  AUTOEXEC.BAT   (2)
 2/13/94   6:21          54,619  COMMAND.COM    (2)
10/16/97  14:25             115  CONFIG.SYS     (2)
11/16/97  17:17      61,865,984  pagefile.sys
 2/13/94   6:21           9,349  WINA20.386     (3)
  1. This directory holds MS-DOS.

  2. System startup code for DOS.

  3. Some sort of Windows 3.1 hack.

Explanation:

17.1. Implementation Notes

Callout marks are generated by the callout inline macro while callout lists are generated using the callout list definition. The callout macro and callout list are special in that they work together. The callout inline macro is not enabled by the normal macros substitutions option, instead it has it's own callouts substitution option.

The following attributes are available during inline callout macro substitution:

{index}
The callout list item index inside the angle brackets.
{coid}
An identifier formatted like CO<listnumber>-<index> that uniquely identifies the callout mark. For example CO2-4 identifies the fourth callout mark in the second set of callout marks.

The {coids} attribute can be used during callout list item substitution — it is a space delimited list of callout IDs that refer to the explanatory list item.

18. Macros

Macros are a mechanism for substituting parameterized text into output documents.

Macros have a name, a single target argument and an attribute list. The default syntax is <name>:<target>[<attributelist>] for inline macros and <name>::<target>[<attributelist>] for block macros. Here are some examples:

http://www.methods.co.nz/asciidoc/index.html[Asciidoc home page]
include::chapt1.txt[tabsize=2]
mailto:srackham@methods.co.nz[]

Macro behavior

18.1. Inline Macros

Inline Macros occur in an inline element context. Predefined Inline macros include URL, image and link macros.

18.1.1. URLs

Standard http, https, ftp, file and mailto URLs are rendered using predefined inline macros.

The default AsciiDoc inline macro syntax is very similar to a URL: all you need to do is append an attribute list containing an optional caption immediately following the URL. If no text is inside the list the URL itself supplies the displayed text.

Here are some examples:

http://www.methods.co.nz/asciidoc/[The AsciiDoc home page]
mailto:joe.bloggs@foobar.com[email Joe Bloggs]
mailto:joe.bloggs@foobar.com[]

Which are rendered:

The AsciiDoc home page

email Joe Bloggs

joe.bloggs@foobar.com

TIP:

If the <target> has space characters they should be replaced by %20. For example large%20image.png.

18.1.2. Internal Cross References

Two AsciiDoc inline macros are provided for creating hypertext links within an AsciiDoc document. You can use either the standard macro syntax or the (preferred) alternative.

anchor

Used to specify hypertext link targets:

[[<id>,<xreflabel>]]
anchor:<id>[<xreflabel>]

The <id> is a unique identifier that must begin with a letter. The optional <xreflabel> is the text to be displayed by xref macros with no captions that refer to this anchor. The <xreflabel> is only really useful when generating DocBook output. Example:

[[X1]]

You may have noticed that the syntax of this inline element is the same as that of the BlockId block element, this is no coincidence since they do roughly the same job.

xref

Creates a hypertext link to a document anchor.

<<<id>,<caption>>>
xref:<id>[<caption>]

The <id> refers to an existing anchor <id>. The optional <caption> is the link's displayed text. If <caption> is not specified then the <id>, enclosed in square brackets, is displayed. Example:

<<X15,attribute lists>>

18.1.3. Linking to Local Documents

Hypertext links to files on the local filesystem are specified using the link inline macro.

link:<target>[<caption>]

The link macro generates relative URLs. The link macro <target> is the target file name (relative to the file system location of the referring document). The optional <caption> is the link's displayed text. If <caption> is not specified then <target> is displayed. Example:

link:downloads/foo.zip[download foo.zip]

You can use the <filename>#<id> syntax to refer to an anchor within a target document but this usually only makes sense when targeting HTML documents.

Images can serve as hyperlinks using the image macro.

18.1.4. Images

Inline images are inserted into the output document using the image macro. The inline syntax is:

image:<target>[<attributes>]

The contents of the image file <target> is displayed. To display the image it's file format must be supported by the target backend application. HTML and DocBook applications normally support PNG or JPG files.

Image elements are not supported by the LinuxDoc format and the alternative text in macro attribute 1 (if specified) is displayed instead.

<target> file name paths are relative to the location of the referring document.

Image macro attributes

18.2. Block Macros

A Block macro reference must be contained in a single line separated either side by a blank line or a block delimiter.

Block macros behave just like Inline macros, with the following differences:

18.2.1. Images

Formal images are inserted into the output document using the image macro. The syntax is:

image::<target>[<attributes>]

In all respects, apart from context, the use of the block image macro is exactly the same as it's inline counterpart.

The image can be titled by preceding the image macro with a BlockTitle. DocBook processors can normally be configured to include titled images in an automatically generated List of Figures.

For example:

.Main circuit board
image::images/layout.png[J14P main circuit board]

18.2.2. Comment Lines

Single lines starting with two forward slashes hard up against the left margin are treated as comments and are stripped from the output. Comment lines have been implemented as a block macro and are only valid in a block context — they are not treated as comments inside paragraphs or delimited blocks. For example:

// This is a comment.

See also Comment Blocks.

18.3. System Macros

System macros are block macros that perform a predefined task which is hardwired into the asciidoc(1) program.

18.3.1. Include Macros

These system macros include the contents of a named file in the source document; it's as if the included file were part of the parent document.

There are two include macros: include which allows nested include macros and include1 which does not allow nested includes.

Example: Include macro examples

  include::chapter1.txt[tabsize=4]

  +++++++++++++++++++++++
  include1::table6.html[]
  +++++++++++++++++++++++

Include macro behavior

18.3.2. Conditional Inclusion Macros

Lines of text in the source document can be selectively included or excluded from processing based on the the existence (or not) of a document attribute. There are two forms of conditional inclusion macro usage, the first includes document text between the ifdef and endif macros if a document attribute is defined:

ifdef::<attribute>[]
:
endif::<attribute>[]

The second for includes document text between the ifndef and endif macros if the attribute is not defined:

ifndef::<attribute>[]
:
endif::<attribute>[]

<attribute> is an attribute name which is optional in the trailing endif macro.

Take a look at the *.conf configuration files in the asciidoc(1) program directory for examples.

18.3.3. eval, sys and sys2 System Macros

These system macros exhibit the same behavior as their same named system attribute references. The difference is that they are expanded globally, not just in an inline attribute context.

The following example displays a directory listing as a literal block:

--------------------
sys::[ls -ltr *.txt]
--------------------

18.4. Macro Definitions

Each entry in the configuration [macros] section is a macro definition which can take one of the following forms:

<pattern>=<name>
Inline macro definition.
<pattern>=#<name>
Block macro definition.
<pattern>=+<name>
System macro definition.
<pattern>
Delete the existing macro with this <pattern>.

<pattern> is a Python regular expression and <name> is the name of a markup template. If <name> is omitted then it is the value of the named regular expression match group named name.

Here's what happens during macro substitution

19. Tables

Tables are the most complex AsciiDoc elements and this section is quite long.
[The current table syntax is overly complicated and unwieldy to edit, a more usable syntax will appear in future versions of AsciiDoc.]

NOTE:

AsciiDoc generates nice HTML tables, but the current crop of commonly deployed DocBook stylesheets render tables with varying degrees of success. Use tables only when really necessary.

19.1. Example Tables

The following annotated examples are all you'll need to start creating your own tables.

The only non-obvious thing you'll need to remember are the column stop characters:

Simple table:

`---`---
1   2
3   4
5   6
--------

Output:

1 2
3 4
5 6

Table with title, header and footer:

.An example table
[grid="all"]
'---------.--------------
Column 1   Column 2
-------------------------
1          Item 1
2          Item 2
3          Item 3
-------------------------
6         Three items
-------------------------

Output:

Table: An example table

Column 1 Column 2
6 Three items
1 Item 1
2 Item 2
3 Item 3

Four columns totaling 15% of the pagewidth, CSV data:

[frame="all"]
````~15
1,2,3,4
a,b,c,d
A,B,C,D
~~~~~~~~

Output:

1 2 3 4
a b c d
A B C D

A table with a numeric ruler and externally sourced CSV data:

[frame="all", grid="all"]
.15`20`25`20`~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
ID,Customer Name,Contact Name,Customer Address,Phone
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
include::customers.csv[]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Renders:

ID Customer Name Contact Name Customer Address Phone
AROUT Around the Horn Thomas Hardy 120 Hanover Sq.
London
(171) 555-7788
BERGS Berglunds snabbkop Christina Berglund Berguvsvagen 8
Lulea
0921-12 34 65
BLAUS Blauer See Delikatessen Hanna Moos Forsterstr. 57
Mannheim
0621-08460
BLONP Blondel pere et fils Frederique Citeaux 24, place Kleber
Strasbourg
88.60.15.31
BOLID Bolido Comidas preparadas Martin Sommer C/ Araquil, 67
Madrid
(91) 555 22 82
BONAP Bon app' Laurence Lebihan 12, rue des Bouchers
Marseille
91.24.45.40
BOTTM Bottom-Dollar Markets Elizabeth Lincoln 23 Tsawassen Blvd.
Tsawassen
(604) 555-4729
BSBEV B's Beverages Victoria Ashworth Fauntleroy Circus
London
(171) 555-1212
CACTU Cactus Comidas para llevar Patricio Simpson Cerrito 333
Buenos Aires
(1) 135-5555

19.2. AsciiDoc Table Block Elements

This sub-section details the AsciiDoc source file table format.

Table  ::= (Ruler,Header?,Body,Footer?)
Header ::= (Row+,Underline)
Footer ::= (Row+,Underline)
Body   ::= (Row+,Underline)
Row    ::= (Data+)

A table is terminated when the table underline is followed by a blank line or an end of file. Table underlines which separate table headers, bodies and footers should not be followed by a blank line.

19.2.1. Ruler

The first line of the table is called the Ruler. The Ruler specifies which configuration file table definition to use, column widths, column alignments and the overall table width.

There are two ruler formats:

Character ruler
The column widths are specified by the number of table fill characters between column stop characters.
Numeric ruler
The column widths are specified numerically. If a column width is omitted the previous width is used. In the degenerate case of no widths being specified columns are allocated equal widths.

The ruler format can be summarized as:

ruler ::= ((colstop,(colwidth,fillchar+)?)+, fillchar+, tablewidth?

19.2.2. Attribute List

The following optional table attributes can be specified in a table's AttributeList:

separator
The default DSV table form colon separator can be changed using the separator attribute. For example: [separator="|"].
frame
Defines the table border and can take the following values: topbot (top and bottom), all (all sides), none and sides (left and right sides). The default value is topbot.
grid
Defines which ruler lines are drawn between table rows and columns. The grid attribute value can be any of the following values: none, cols, rows and all. The default value is none. For example [frame="all", grid="none"].

You can also use an AttributeList to override the following table definition and ruler parameters: format, subs, tablewidth.

19.2.3. Underline

A table Underline consists of a line of three or more fillchar characters which are end delimiters for table header, footer and body sections.

19.2.4. Markup Attributes

The following attributes are automatically available inside table tag and markup templates.

cols
The number of columns in the table.
colalign
Column alignment assumes one of three values (left, right or center). The value is determined by the corresponding ruler column stop character (only valid inside colspec tags).
colwidth
The output column widths are calculated integers (only valid inside colspec tags).
format
The table definition format value (can be overridden with attribute list entry).
subs
The table definition subs value (can be overridden with attribute list entry).
tablewidth
The ruler tablewidth value (can be overridden with attribute list entry).
pagewidth
The pagewidth miscellaneous configuration option.
pageunits
The pageunits miscellaneous configuration option.

The colwidth value is calculated as (N is the ruler column width number and M is the sum of the ruler column widths):

( N / M ) * pagewidth

If the ruler tablewidth was specified the column width is multiplied again by this value.

There is one exception: character rulers that have no pagewidth specified. In this case the colwidth value is calculated as (where N is the column character width measured on the table ruler):

( N / textwidth ) * pagewidth

19.2.5. Row and Data Elements

Each table row consists of a line of text containing the same number of Data items as there are columns in the table,

Lines ending in a backslash character are continued on the next line.

Each Data item is an AsciiDoc substitutable string. The substitutions performed are specified by the subs table definition entry. Data cannot contain AsciiDoc block elements.

The format of the row is determined by the table definition format value:

fixed
Row data items are assigned by chopping the row up at ruler column width boundaries.
csv
Data items are assigned the parsed CSV (Comma Separated Values) data.
dsv
The DSV (Delimiter Separated Values) format is a common UNIX tabular text file format.
  • The separator character is a colon (although this can be set to any letter using the separator table attribute).

  • Common C-style backslash escapes are supported.

  • Blank lines are skipped.

19.3. Table Configuration File Definitions

Read on if you want to modify existing table behavior or create your own table definitions.

A table definition consists of two configuration file sections:

  1. A [tabledef-*] section specifying the AsciiDoc source syntax and content format plus the table's backend header, body and footer markup.

  2. A table markup template section (the template entry from the table definition section) which assembles header, body and footer parts into a complete table.

The easiest way to get a feel for how it all fits together is to take a look at the distributed AsciiDoc table definitions in the asciidoc.conf and backend configuration files.

Backend independent [tabledef-*] section entries are:

fillchar
A single character that fills source table ruler and underline lines.
subs
A list of AsciiDoc substitution options specifying the substitutions, and substitution order, performed on table data items. If undefined defaults to the default AsciiDoc substitutions.
format
The source row data format (fixed, csv or dsv).

Backend specific [tabledef-*] section entries are:

section
The name of the table template section.
comspec
The table comspec tag definition.
headrow, footrow, bodyrow
Table header, footer and body row tag definitions. headrow and footrow table definition entries default to row if undefined.
headdata, footdata, bodydata
Table header, footer and body data tag definitions. headdata and footdata table definition entries default to data if undefined.

The following attributes are available to the table markup template:

comspecs
Expands to N substituted comspec tags where N is the number of columns.
headrows, footrows, bodyrows
These references expand to sets of substituted header, footer and body rows as defined by the corresponding row and data tag definitions.

In addition tables are affected by the following [miscellaneous] configuration file entries:

textwidth
The page width (in characters) of the source text. This setting is compared to the the table ruler width when calculating the relative size of character ruler tables on the output page.
pagewidth
This integer value is the printable width of the output media. Used to calculate colwidth and tablewidth substitution values.
pageunits
The units of width in output markup width attribute values.

Table definition behavior

20. Manpage Documents

Sooner or later, if you program for a UNIX environment, you're going to have to write a man page.

By observing a couple of additional conventions you can compose AsciiDoc files that will translate to a DocBook refentry (man page) document. The resulting DocBook file can then be translated to the native roff man page format (or other formats).

For example, the asciidoc.1.txt file in the AsciiDoc distribution ./doc directory was used to generate both asciidoc.1.css-embedded.html HTML file and (via the xmlto(1) command) the asciidoc.1 roff formatted asciidoc(1) man page.

To find out more about man pages view the man(7) manpage (man 7 man command).

20.1. Document Heading

The document Header is mandatory. The title line contains the man page name followed immediately by the manual section number in brackets, for example ASCIIDOC(1). The title name should not contain white space and the manual section number is a single digit optionally followed by a single character.

20.2. The NAME Section

The first manpage section is mandatory and must be called NAME and contain a single paragraph (usually a single line) consisting of a list of one or more comma separated command name(s) separated from the command purpose by a dash character. The dash must have at least one white space character on either side. For example:

printf, fprintf, sprintf - print formatted output

20.3. The SYNOPSIS Section

The second manpage section is mandatory and must be called SYNOPSIS.

21. Configuration Files

AsciiDoc source file syntax and output file markup is largely controlled by a set of cascading, text based, configuration files. At runtime The AsciiDoc default configuration files are combined with optional document and user specific configuration files.

21.1. Configuration File Format

Configuration files contain named sections. Each section begins with a section name in square brackets []. The section body consists of the lines of text between adjacent section headings.

TIP:

When creating custom configuration files you only need to include the sections and entries that differ from the default configuration.

TIP:

The best way to learn about configuration files is to read the default configuration files in the asciidoc(1) program directory along with generated backend output and the backend markup spec. You view configuration file processing by turning on the asciidoc(1) -v command-line option.

21.2. Markup Template Sections

Markup template sections supply backend markup for translating AsciiDoc elements. Since the text is normally backend dependent you'll find these sections in the backend specific configuration files. A markup template section body can contain:

The document content placeholder is a single | character and is replaced by text from the source document. Use the {brvbar} attribute reference if you need a literal | character.

21.3. Special Sections

AsciiDoc reserves the following predefined, or Special, section names for specific purposes:

miscellaneous
Configuration options that don't belong anywhere else.
attributes
Named substitution values.
specialcharacters
Special characters reserved by the backend markup.
tags
Markup tag definitions used by AsciiDoc markup mechanisms.
quotes
Definitions for inline character formatting quoting.
specialwords
Lists of words and phrases singled out for special markup.
replacements
Find and replace substitution definitions.
specialsections
Used to single out section names for special markup.
macros
Macro syntax definitions. titles: Heading, section and block title parameters.
paradef*
Paragraph element definitions.
blockdef*
DelimitedBlock element definitions.
listdef*
List element definitions.
tabledef*
Table element definitions.

Each line of text in a Special section is a section entry. Section entries can take the following forms:

name=value
The entry value is set to value.
name=
The entry value is set to a zero length string.
name
The entry is undefined (deleted from the configuration).

Section entry behavior

21.3.1. Miscellaneous

The optional [miscellaneous] section specifies the following name=value options:

newline
Output file line termination characters. Can include any valid Python string escape sequences. The default value is \r\n (carriage return, line feed). Should not be quoted or contain explicit spaces (use \x20 instead). For example:
$ asciidoc -a 'newline=\n' -b docbook mydoc.txt
outfilesuffix
The default extension for the output file, for example outfilesuffix=.html. Defaults to backend name.
tabsize
The number of spaces to expand tab characters, for example tabsize=4. Defaults to 8. A tabsize of zero suppresses tab expansion (useful when piping included files through block filters). Included files can override this option with the include macro tabsize attribute list entry.
textwidth, pagewidth, pageunits
These global table related options are documented in the Table Configuration File Definitions sub-section.

NOTE:

[miscellaneous] configuration file entries can be set using the asciidoc(1) -a command-line option and perform the usual attribute substitutions.

21.3.2. Titles

sectiontitle
Two line section title pattern. The entry value is a Python regular expression containing the named group title.
underlines
A comma separated list of document and section title underline character pairs starting with the Header underline and ending with Section level 4 underline. The default setting is:
underlines="==","--","~~","^^","++"
sect0…sect4
One line section title patterns. The entry value is a Python regular expression containing the named group title.
blocktitle
BlockTitle line title pattern. The entry value is a Python regular expression containing the named group title. The title group's matching value is used as the following element's title.
underlines="==","--","~~","^^","++"
subs
A comma separated list of substitutions that are performed on document Header and Section titles. The default value for this entry is specialcharacters,quotes,replacements,attributes,macros.

21.3.3. Tags

The [tags] section contains AsciiDoc tag definitions (one per line). Tags are used to translate AsciiDoc elements to backend markup.

An AsciiDoc tag definition is formatted like <tagname>=<starttag>|<endtag>. For example:

emphasis=<em>|</em>

In this example asciidoc(1) replaces the | character with the emphasized text from the AsciiDoc input file and writes the result to the output file.

Use the {brvbar} attribute reference if you need to include a | pipe character inside tag text.

21.3.4. Attributes Section

The optional [attributes] section contains attribute entries.

If the attribute value requires leading or trailing spaces then the text text should be enclosed in double-quote (") characters.

To delete a attribute insert a name only entry in a downstream configuration file or use the asciidoc(1) -a ^name command-line option (the attribute name is prefixed with a ^ character to delete it).

21.3.5. Special Characters

The [specialcharacters] section specifies how to translate each of the characters reserved by the backend markup. Each translation is specified on a single line formatted like:

special_character=translated_characters

Special characters are normally confined to those that resolve markup ambiguity (in the case of SGML/XML markups the ampersand, less than and greater than characters). For example:

<=&lt;

When special character substitution is enabled all occurrences of < will be replaced by &lt;.

21.3.6. Quoted Text

The [quotes] section defines the text formatting markup characters that are used to envelope words and word phrases. Each section entry value has a corresponding tag entry from the [tags] section. The entry name defines the one (or more) characters that quote the text. In this example a double underscore defines underlined HTML markup:

[quotes]
__=underline
[tags]
underline=<u>|</u>

You can specify the left and right quote strings separately by separating them with a | character, for example:

[quotes]
((|))=gui

If you set the tag to none then a blank string will be substituted for the quoted text which has the effect of dropping the quote from the output document.

Quoted text behavior

21.3.7. Special Words

The [specialwords] section is used to single out words and phrases that you want to consistently highlight in some way throughout your document without having to repeatedly specify the markup. The name of each entry corresponds to a markup template section and the entry value consists of a list of words and phrases to be highlighted. For example:

[specialwords]
strongwords=NOTE: TODO:
[strongwords]
<strong>{words}</strong>

The examples specifies that any occurrence of NOTE: or TODO: should appear in a bold font.

Words and word phrases are treated as Python regular expressions: for example, the word ^NOTE: would only match NOTE: if appeared at the start of a line.

AsciiDoc comes with three built-in Special Word types: emphasizedwords, monospacedwords and strongwords. Each type has a corresponding markup template section. Edit the configuration files to customize existing Special Words and to add new ones.

Special word behavior

21.3.8. Replacements

[replacements] configuration file entries specify find and replace text formatted like:

find_pattern=replacement_text

The find text can be a Python regular expression; the replace text can contain Python regular expression group references.

Use Replacement shortcuts for often used macro references, for example:

\.NEW\.=image:smallnew.gif[New!]

Replacement behavior

21.4. Configuration File Names and Locations

Configuration files have a .conf file name extension; they are loaded implicitly (using predefined file names and locations) or explicitly (using the asciidoc(1) -f command-line option).

Implicit configuration files are loaded from the following directories in the following order:

  1. The /etc/asciidoc directory (if it exists).

  2. The asciidoc(1) program directory.

  3. The .asciidoc directory in the user's home directory (if it exists).

  4. The AsciiDoc source file directory.

The following implicit configuration files from each of the above locations are loaded in the following order:

  1. asciidoc.conf

  2. <backend>.conf

  3. <backend>-<doctype>.conf

Where <backend> and <doctype> are values specified by the asciidoc(1) -b and -d command-line options.

Finally, configuration files named like the source file will be automatically loaded if they are found in the source file directory. For example if the source file is mydoc.txt and the -b html option is used then asciidoc(1) will look for mydoc.conf and mydoc-html.conf in that order.

Implicit configuration files that don't exist will be silently skipped.

The user can explicitly specify additional configuration files using the asciidoc(1) -f command-line option. The -f option can be specified multiple times, in which case configuration files will be processed in the order they appear on the command-line.

For example, when we translate our AsciiDoc document mydoc.txt with:

$ asciidoc -b xhtml -f extra.conf mydoc.txt

Configuration files (those that exist) will be processed in the following order:

  1. First default global configuration files from the asciidoc program directory are loaded:

    asciidoc.conf
    xhtml.conf
    xhtml-article.conf
    
  2. Then, from the users home ~/.asciidoc directory. This is were you put customization specific to your own asciidoc documents:

    asciidoc.conf
    xhtml.conf
    xhtml-article.conf
    
  3. Next from the source document project directory (the first three apply to all documents in the directory, the last two are specific to the mydoc.txt document):

    asciidoc.conf
    xhtml.conf
    xhtml-article.conf
    mydoc.conf
    mydoc-xhtml.conf
    
  4. Finally the file specified by the -f command-line option is loaded:

    extra.conf
    

TIP:

Use the asciidoc(1) -v command-line option to see which configuration files are loaded and the order of they are loaded.

22. Document Attributes

A document attribute is comprised of a name and a textual value and is used for textual substitution in AsciiDoc documents and configuration files. An attribute reference (an attribute name enclosed in braces) is replaced by it's their corresponding attribute value.

There are four sources of document attributes (from highest to lowest precedence):

Within each of these divisions the last processed entry takes precedence.

IMPORTANT:

If an attribute is not defined then the line containing the attribute reference is dropped and is not written to the output file. This property is used extensively in AsciiDoc configuration files to facilitate conditional markup generation.

23. Attribute Entries

The AttributeEntry block element allows document attributes to be assigned within an AsciiDoc document. Attribute entries are added to the global document attributes dictionary. The attribute name/value syntax is a single line like:

:<name>: <value>

For example:

:Author Initials: JB

This will set an attribute reference {authorinitials} to the value JB in the current document.

AttributeEntry properties

NOTE:

The author attribute as a special case, it also sets the firstname attribute and (if specified) surname and authorinitials attributes.

Here's another example:

AsciiDoc User Manual
====================
:Author:    Stuart Rackham
:Email:     srackham@methods.co.nz
:Date:      April 23, 2004
:Revision:  5.1.1
:Key words: linux, ralink, debian, wireless
:Revision history:

Which create these attributes:

{author}, {firstname}, {surname}, {authorinitials}, {email},
{date}, {revision}, {keywords}, {revisionhistory}

The preceding example is equivalent to the standard AsciiDoc two line document header. Actually it's a little bit different with the addition of the {keywords} and {revisionhistory}
[The existence of a {revisionhistory} attribute causes a revision history file (if it exists) to be included in DocBook outputs. If a file named like {docname}-revhistory.xml exists in the document's directory then it will be added to the DocBook header (see the ./doc/asciidoc-revhistory.xml example that comes with the AsciiDoc distribution).]
attributes.

TIP:

To delete an attribute using the document Attribute element — set it to an undefined attribute, for example:

:Delete me: {undefined-attribute}

24. Attribute Lists

An attribute list is a comma separated list of attribute values. The entire list is enclosed in square brackets. Attribute lists are used to pass parameters to macros and block elements.

The list consists of zero or more positional attribute values followed by zero or more named attribute values. Here are three examples:

[Hello]
[Bertrand Russell, The World of Mathematics (1956)]
["22 times", backcolor="#0e0e0e", options="noborders,wide"]

Attribute list properties

TIP:

To view processed attribute list values use the asciidoc(1) -v command-line option.

24.1. Macro Attribute lists

All macros calls are suffixed with an attribute list. The list may be empty but it cannot be omitted. List entries are used to pass attribute values for macro substitution.

24.2. AttributeList Element

An attribute list on a line by itself constitutes an AttributeList block element, it's function is to parameterize the following block element. The list attributes are passed to the next block element for markup template substitution.

List attributes are available in Title, Paragraph, DelimitedBlock, List and Table element markup templates.

25. Intrinsic Attributes

Intrinsic attributes are created automatically from document header parameters, asciidoc(1) command-line arguments, environment parameters along with attributes defined in the default configuration files. Here's the list of predefined intrinsic attributes:

{asciidoc-version}    the version of asciidoc(1)
{asciidoc-dir}        the asciidoc(1) application directory
{user-dir}            the ~/.asciidoc directory (if it exists)
{authorinitials}      author initials (from document header)
{author}              author's full name ({firstname} {middlename}
                      {lastname})
{authored}            empty string '' if {author} or {email} defined,
                      otherwise undefined.
{date}                document date (from document header)
{doctitle}            document title (from document header)
{email}               author's email address (from document header)
{firstname}           author first name (from document header)
{lastname}            author last name (from document header)
{localdate}           the current date
{localtime}           the current time
{manname}             manpage name (defined in NAME section)
{manpurpose}          manpage (defined in NAME section)
{mantitle}            manpage title (from document header)
{manvolnum}           manpage volume number (1..8) (from document header)
{middlename}          author middle name (from document header)
{revision}            document revision number (from document header)
{title}               section title (defined titled element substitution
                      sections)
{sectnum}             section number (defined in section titles
                      markup template sections)
{amp}                 ampersand (&) character
{lt}                  less than (<) character
{gt}                  greater than (>) character
{brvbar}              broken vertical bar (|) character
{empty}               empty string ''
{infile}              input file name
{outfile}             output file name
{docdir}              document directory name (no trailing separator)
{docname}             document file name without extension
{doctype}             document type specified by `-d` option
{filetype}            output file name file extension
{backend}             document backend specified by `-b` option
{backend-<backend>}   empty string ''
{<backend>-<doctype>} empty string ''
{doctype-<doctype>}   empty string ''
{filetype-<fileext>}  empty string ''
{basebackend}         html, css, docbook or linuxdoc
{basebackend-<base>}  empty string ''
{imagesdir}           directory containing admonition icons (HTML
                      backend; if undefined defaults to images)
{stylesdir}           directory containing CSS stylesheets (CSS
                      backends; if undefined defaults to .)

The entries that translate to blank strings are designed to be used for conditional text inclusion (remember that if an undefined attribute is referenced then the containing line will be dropped from the output). You can also use the ifdef, ifndef and endif System macros for conditional inclusion.
[Conditional inclusion using ifdef and ifndef macros differs from attribute conditional inclusion in that the former occurs when the file is read while the latter occurs when the contents are written.]

26. Attribute References

An attribute references is an attribute name (possibly followed by an expression) enclosed in braces. When an attribute reference is encountered it is evaluated and replaced by its corresponding text value. If there is no corresponding text value the attribute is said to be undefined and the line containing the attribute is dropped.

There are three types of attribute reference: Simple, Conditional and System.

Attribute reference behavior

26.1. Simple Attributes References

Simple attribute references take the form {<name>}. If the attribute name is defined its text value is substituted otherwise the line containing the reference is dropped from the output.

26.2. Conditional Attribute References

Conditional attribute references return text that is predicated on the existence of an attribute. Conditional attribute references take the following forms:

{<name>=<value>}
<value> is substituted if the attribute <name> undefined otherwise it's value is substituted. <value> can contain simple attribute references.
{<name>?<value>}
<value> is substituted if the attribute <name> is defined otherwise an empty string is substituted. <value> can contain simple attribute references.
{<name>!<value>}
<value> is substituted if the attribute <name> is undefined otherwise an empty string is substituted. <value> can contain simple attribute references.
{<name>#<value>}
<value> is substituted if the attribute <name> is defined otherwise the undefined attribute entry causes the containing line to be dropped. <value> can contain simple attribute references.
{<name>%<value>}
<value> is substituted if the attribute <name> is not defined otherwise the containing line is dropped. <value> can contain simple attribute references.

26.3. System Attribute References

System attribute references generate the attribute text value by executing a predefined action parameterized by a single argument. The syntax is {<action>:<argument>}.

{eval:<expression>}
Substitutes the result of the Python <expression>. If <expression> evaluates to None the line containing the reference is dropped from the output file.
{include:<filename>}
Substitutes contents of the file named <filename>.
  • The included file is read at the time of attribute substitution not when the file is read (as is the case with include block macros) — this can be significant when dealing with configuration files.

  • If the file does not exist a warning is emitted and the line containing the reference is dropped from the output file.

  • Tabs are expanded based on the current tabsize.

{sys:<command>}
Substitutes the stdout generated by the execution of the shell <command>.
{sys2:<command>}
Substitutes the stdout and stderr generated by the execution of the shell <command>.

System reference behavior

27. Filters

Filters are external shell commands used to process Paragraph and DelimitedBlock content; they are specified in configuration file Paragraph and DelimitedBlock definitions.

There's nothing special about the filters, they're just standard UNIX filters: they read text from the standard input, process it, and write it to the standard output.

Attribute substitution is performed on the filter command prior to execution — attributes can be used to pass parameters from the AsciiDoc source document to the filter.

NOTE:

Filter functionality is currently only available on POSIX platforms (this includes Cygwin).

27.1. Filter Search Paths

If the filter command does not specify a directory path then asciidoc(1) searches for the command:

27.2. Filter Configuration Files

Since filters are normally part of new Paragraph or DelimitedBlock definitions they are usually accompanied by a configuration file.

asciidoc(1) auto-loads all .conf files found in the user's $HOME/.asciidoc/filters directory and the asciidoc(1) ./filters subdirectory.

27.3. Code Filter

AsciiDoc comes with a simple minded code-filter for highlighting source code keywords and comments.
[The example code filter shipped with AsciiDoc is provided as an example filter and is by no means meant to be a production quality syntax highlighter]
You can find the code-filter in the AsciiDoc distribution ./filters subdirectory (read the ./filters/code-filter-readme.txt file for instructions).

The following example highlights Python keywords in the block's content:

.Code filter example
[python]
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
''' A multi-line
    comment.'''
def sub_word(mo):
        ''' Single line comment.'''
        word = mo.group('word')       # Inline comment
        if word in keywords[language]:
                return quote + word + quote
        else:
                return word
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Outputs:

Example: Code filter example

''' A multi-line
    comment.'''
def sub_word(mo):
        ''' Single line comment.'''
        word = mo.group('word') # Inline comment
        if word in keywords[language]:
                return quote + word + quote
        else:
                return word

28. Tips and Tricks

28.1. Know Your Editor

Writing AsciiDoc documents will be a whole lot more pleasant if you know your favorite text editor. Learn how to indent and reformat text blocks, paragraphs, lists and sentences. Tips for vim users follow.

28.2. Vim Commands for Formatting AsciiDoc

The Vim text editor's gq command is great for reformatting and indenting AsciiDoc paragraphs and lists.

TIP:

The Vim website (http://www.vim.org) has a wealth of resources, including scripts for automated spell checking and ASCII Art drawing.

28.2.1. Text Wrap Paragraphs

Use the vim :gq command to reformat paragraphs. Setting the textwidth sets the right text wrap margin; for example:

:set textwidth=70

To reformat a paragraph:

  1. Position the cursor at the start of the paragraph.

  2. Type gq}.

Execute :help gq command to read about the vim gq command.

TIP:

Put set commands in your ~/.vimrc file so you don't have to enter them manually.

28.2.2. Format Lists

The :gq command can also be used to format bulleted and numbered lists. First you need to:

  1. Set the textwidth right wrap margin.

  2. Set the formatoptions n flag to enable numbered list reformatting (this flag also requires the autoindent option be set).

  3. Add fb:*,fb:.,fb:+,fb:> to the comments option to assist the Vim :gq command reformat the AsciiDoc bulleted and numbered lists (in the example the C style comments middle part (mb:*) has been dropped to avoid ambiguity). Run the vim :help format-comments command for more about reformatting).

For example:

:set textwidth=70 formatoptions=tcqn autoindent
:set comments=s1:/*,ex:*/,://,b:#,:%,fb:-,fb:*,fb:.,fb:+,fb:>

Now you can format simple lists that use dash, asterisk, period and plus bullets along with numbered ordered lists:

  1. Position the cursor at the start of the list.

  2. Type gq}.

TIP:

Assign the gq} command to the Q key with the :nnoremap Q gq} command or put it in your ~/.vimrc file to so it's always available.

Here's how I setup my .vimrc file:

nnoremap Q gq}

autocmd BufRead,BufNewFile *.txt,README,TODO,CHANGELOG,NOTES
        \ setlocal autoindent expandtab tabstop=8 softtabstop=2 shiftwidth=2
        \ textwidth=70 wrap formatoptions=tcqn
        \ comments=s1:/*,ex:*/,://,b:#,:%,:XCOMM,fb:-,fb:*,fb:+,fb:.,fb:>

28.2.3. Indent Paragraphs

Indent whole paragraphs by indenting the fist line with the desired indent and then executing the gq} command.

28.3. Troubleshooting

28.4. Gotchas

Misinterpreted text formatting
If text in your document is incorrectly interpreted as formatting instructions you can suppress formatting by placing a backslash character immediately in front of the leading quote character(s). For example in the following line the backslash prevents text between the two asterisks from being output in a strong (bold) font:
Add `\*.cs` files and `*.resx` files.
Overlapping text formatting
Overlapping text formatting will generate illegal overlapping markup tags which will result in downstream XML parsing errors. Here's an example:
Some *strong markup 'that overlaps* emphasized markup'.
Ambiguous underlines
A DelimitedBlock can immediately follow paragraph without an intervening blank line, but be careful, a single line paragraph underline may be misinterpreted as a section title underline resulting in a "closing block delimiter expected" error.
Ambiguous ordered list items
Lines beginning with numbers at the end of sentences will be interpreted as ordered list items. The following example (incorrectly) begins a new list with item number 1999:
He was last sighted in
1999. Since then things have moved on.

The list item out of sequence warning makes is unlikely that this problem will go unnoticed.

Escaping inside DSV table data
Delimiter separated text uses C style backslash escape sequences. If you want to enter a backslash (for example, to escape AsciiDoc text formatting or an inline macro) you need to escape it by entering two backslashes.
Special characters in attribute values
Special character substitution precedes attribute substitution so if attribute values contain special characters you may, depending on the substitution context, need to substitute the special characters yourself. For example:
$ asciidoc -a 'companyname=Bill &amp; Ben' -b html mydoc.txt
Macro attribute lists
If named attribute list entries are used then positional attribute values must be quoted. For example:
["Desktop screenshot",width=32]
CSS rule ambiguity
One liner embedded CSS rules of the form selector {property: value} will be interpreted as a glossary references unless a space is inserted after the opening brace.

28.5. Combining Separate Documents

You have a number of stand-alone AsciiDoc documents that you want to process as a single document. Simply processing them with a series of include macros won't work, because instead of starting at level 1 the section levels of the combined document start at level 0 (the document title level).

The solution is to redefine the title underlines so that document and section titles are pushed down one level.

  1. Push the standard title underlines down one level by defining a new level 0 underline in a custom configuration file. For example combined.conf:

    [titles]
    underlines="__","==","--","~~","^^"
    
  2. Create a top level wrapper document. For example combined.txt:

     Combined Document Title
     _______________________
    
     include::document1.txt[]
    
     include::document2.txt[]
    
     include::document3.txt[]
    
  3. Process the wrapper document. For example:

    $ asciidoc -b html -f combined.conf combined.txt
    

Actually the -f is unnecessary as asciidoc(1) automatically looks for a same-named .conf file.

28.6. Processing Document Sections Separately

You have divided your AsciiDoc document into separate files (one per top level section) which are combined and processed with the following top level document:

 Combined Document Title
 =======================
 Joe Bloggs
 v1.0, 12-Aug-03

 include::section1.txt[]

 include::section2.txt[]

 include::section3.txt[]

You also want to process the section files as separate documents. This is easy because asciidoc(1) will quite happily process section1.txt, section2.txt and section3.txt separately.

If you want to promote the section levels up one level, so the document is processed just like a stand-alone document, then pop the section underline definition up one level:

[titles]
underlines="--","~~","^^","++","__"

The last "__" underline is a dummy that won't actually be used but is necessary to legitimize the underline definition.

This is just the reverse of the technique used for combining separate documents explained in the previous section.

28.7. Processing Document Chunks

asciidoc(1) can be used as a filter, so you can pipe chunks of text through it. For example:

$ echo 'Hello *World!*' | asciidoc -s -b html -
<p>
Hello <strong>World!</strong>
</p>

The -s command-line option suppresses header and footer output and is useful if the processed output is to be included in another file.

28.8. Badges in HTML Page Footers

See the badges.conf file in the AsciiDoc distribution examples\ directory.

28.9. Pretty Printing AsciiDoc Output

If the indentation and layout of the asciidoc(1) output is not to your liking you can:

  1. Change the indentation and layout of configuration file markup template sections. The {empty} glossary entry is useful for outputting trailing blank lines in markup templates.

  2. Or use Dave Raggett's excellent HTML Tidy program to tidy asciidoc(1) output. Examples:

    $ asciidoc -b docbook -o - mydoc.txt | tidy -indent -xml >mydoc.xml
    $ asciidoc -b css-embedded -o - mydoc.txt | tidy -indent >mydoc.html
    

HTML Tidy can be downloaded from http://tidy.sourceforge.net/

The Mozilla and Firefox web browsers also do a good job of displaying XML content in a collapsible outline format.

28.10. Supporting Minor DocBook DTD Variations

The distribution docbook-sgml.conf file illustrates how to support minor DTD variations.

28.11. Shipping Stand-alone AsciiDoc Source

Reproducing presentation documents from some else's source has one major problem: unless your configuration files are the same creator's you won't get the same output.

The solution is to create a single backend specific composite configuration file using the asciidoc(1) -c command-line option. You then ship this file along with the AsciiDoc source document plus the asciidoc.py script. The only end user requirement is that they have Python installed. This example creates a composite HTML configuration file for mydoc.txt:

$ asciidoc -c -b xhtml mydoc.txt > mydoc-html.conf

Ship mydoc.txt, mydoc-html.conf, and asciidoc.py. With these three files (and a Python interpreter) the recipient can regenerate the HMTL output:

$ ./asciidoc.py -e -b xhtml mydoc.txt

The -e option excludes the use of any existing implicit configuration files, ensuring that only entries from the mydoc-html.conf configuration are used.

28.12. Inserting Blank Space

Adjust your style sheets to add the correct separation between block elements. Inserting blank paragraphs containing a single non-breaking space character {nbsp} works but is an ad hoc solution compared to using style sheets.

28.13. Closing Open Sections

You can close off section tags up to level N by calling the eval::[Section.setlevel(N)] system macro. This is useful if you want to a section composed of raw markup. The following example includes a DocBook glossary division at the top section level (level 0):

  ifdef::backend-docbook[]

  eval::[Section.setlevel(0)]

  +++++++++++++++++++++++++++++++
  <glossary>
    <title>Glossary</title>
    <glossdiv>
    ...
    </glossdiv>
  </glossary>
  +++++++++++++++++++++++++++++++
  endif::backend-docbook[]

29. Processing DocBook Files

For reasons outlined previously documents with an article or book type structure will usually be processed using AsciiDoc's DocBook output. The distributed AsciiDoc User Guide plus the example article and book template documents have been generated in this way.

The toolchain processing steps are:

  1. Convert AsciiDoc (*.txt) documents to DocBook XML (*.xml) using AsciiDoc.

  2. Convert DocBook XML documents to HTML, XSL-FO or HTML Help source files using DocBook XSL Stylesheets and an XML parser.

  3. Convert the XSL-FO file to PDF using FOP and the HTML Help source files to an HTML Help (*.chm) file using the Microsoft HTML Help Compiler.

29.1. Toolchain Components

Here are the commands and packages I use to generate the AsciiDoc HTML, PDF and HTML Help documentation files:

AsciiDoc
Converts AsciiDoc (*.txt) files to DocBook XML (*.xml) files.
DocBook XSL Stylesheets
This package contains a set of XSL stylesheets for converting DocBook XML documents to HTML, XSL-FO and HTML Help source (see the DocBook XSL Stylesheets section.
xsltproc
xsltproc is a command line XML parser for applying XSLT stylesheets (in our case the DocBook XSL Stylesheets) to XML documents. It is part of libxslt, the XSLT C library for GNOME (see http://www.xmlsoft.org).
FOP
The Apache Formatting Objects Processor converts XSL-FO (*.fo) files to PDF files (see the FOP section.
Microsoft Help Compiler
The Microsoft HTML Help Compiler (hhc.exe) is a command-line tool that converts HTML Help source files to a single HTML Help (*.chm) file. It runs on MS Windows platforms and can be downloaded from http://www.microsoft.com.

29.2. AsciiDoc XSL Drivers

You will have noticed that the distributed PDF, HTML and HTML Help documentation files (for example ./doc/asciidoc.html) are not the plain outputs produced using the default DocBook XSL Stylesheets configuration. This is because they have been processed using customized DocBook XSL Stylesheet drivers.

You'll find these DocBook XSL drivers in the distribution ./doc directory. The examples which follow are executed from the distribution ./doc directory:

common.xsl
Shared driver parameters. This file is not used directly but is included in all the following drivers.
chunked.xsl
Generate chunked XHTML (separate HTML pages for each document section) in the ./doc/chunked directory. For example:
$ python ../asciidoc.py -b docbook asciidoc.txt
$ xsltproc --nonet chunked.xsl asciidoc.xml
fo.xsl
Generate XSL Formatting Object (*.fo) files for subsequent PDF file generation using FOP. For example:
$ python ../asciidoc.py -b docbook article.txt
$ xsltproc --nonet fo.xsl article.xml > article.fo
$ fop.sh article.fo article.pdf
htmlhelp.xsl
Generate Microsoft HTML Help source files for the MS HTML Help Compiler in the ./doc/htmlhelp directory. See the article at http://www.codeproject.com/winhelp/docbook_howto.asp. This example is run on MS Windows from a Cygwin shell prompt:
$ python ../asciidoc.py -b docbook asciidoc.txt
$ xsltproc --nonet htmlhelp.xsl asciidoc.xml
$ c:/Program\ Files/HTML\ Help\ Workshop/hhc.exe htmlhelp.hhp
$ mv htmlhelp.chm asciidoc.chm
manpages.xsl
Generate a roff(1) format UNIX man page from a DocBook XML refentry document. This example generates an asciidoc.1 man page file:
$ python ../asciidoc.py -d manpage -b docbook asciidoc.1.txt
$ xsltproc --nonet manpages.xsl asciidoc.1.xml
xhtml.xsl
Convert a DocBook XML file to a single XHTML file. For example:
$ python ../asciidoc.py -b docbook asciidoc.txt
$ xsltproc --nonet xhtml.xsl asciidoc.xml > asciidoc.html

If you want to see how the complete documentation set is processed take a look at the A-A-P script ./doc/main.aap.

30. DocBook XSL Stylesheets

The DocBook XSL Stylesheets package contains a set of XSL stylesheets for converting DocBook XML documents to HTML, XSL-FO and HTML Help source (see http://sourceforge.net/projects/docbook/). It is used in conjunction with an XML parser such as xsltproc.

30.1. Installing DocBook XSL Stylesheets on Linux

These instructions were carried out on Fedora Core 1 and may differ on other Linux distributions.

Extract the distribution and set it as the default DocBook XSL Stylesheets with a symbolic link:

$ su
# cd /usr/share/sgml/docbook
# tar -xzf ~srackham/tmp/docbook-xsl-1.67.2.tar.gz
# rm xsl-stylesheets
# ln -s docbook-xsl-1.67.2 xsl-stylesheets

NOTE:

The assumption is made that you have an up to date version of libxslt (which includes the xsltproc(1) command) from http://xmlsoft.org/XSLT/. To check enter xsltproc —version at the command prompt.

Don't hardwire file locations into XSLT drivers

There's no need to have elements like this in your XSLT drivers:

<xsl:import href="C:/bin/docbook-xsl-1.65.1/xhtml/docbook.xsl"/>

It makes them machine and release dependent. It's much better to add rewrite statements to your machine's XML catalog. This allows you to use generic URL stylesheet locations, for example:

<xsl:import href=
http://docbook.sourceforge.net/release/xsl/current/fo/docbook.xsl"/>

Here are the relevant rewrite statements from the /etc/xml/catalog file on a Fedora Linux workstation (the catalog location may differ on other Linux distributions):

<rewriteSystem systemIdStartString=
"http://docbook.sourceforge.net/release/xsl/current"
rewritePrefix=
"file:///usr/share/sgml/docbook/xsl-stylesheets"/>
<rewriteURI uriStartString=
"http://docbook.sourceforge.net/release/xsl/current"
rewritePrefix=
"file:///usr/share/sgml/docbook/xsl-stylesheets"/>

31. FOP

XSL Stylesheets can be used to generate FO (Formatting Object) files, which in turn can be used to produce PDF files using the Apache Formatting Object Processor program (FOP). More the FOP home page is at http://xml.apache.org/fop/.

Reasons to us FOP:

  1. You can produce PDF on both Windows and POSIX platforms.

  2. The PDF quality is on a par with that produced by jw(1).

  3. PDF files are about half the size of those produced by jw(1) and friends.

  4. Processes images, table of contents and images and inserts PDF Bookmarks and active hypertext links.

  5. Uses DocBook XML (no need to produce DocBook SGML).

As of version 0.20.5 installation and configuration of FOP is a manual process. You also need a working Java Runtime to run FOP.

31.1. Installing FOP on Windows

  1. Download latest FOP distribution from http://xml.apache.org/fop/.

  2. Unzip to C:\bin.

  3. Edit the distribution fop.bat file and put it in the search PATH:

    set LOCAL_FOP_HOME=C:\bin\fop-0.20.5
    
  4. Download the JIMI image processing library from http://java.sun.com/products/jimi/.

  5. Extract the JimiProClasses.jar library from the JIMI distribution and copy to the FOP ./lib directory.

  6. Edit the distribution fop.bat file again and add the JIMI library to LOCALCLASSPATH:

    set LOCALCLASSPATH=%LOCALCLASSPATH%;%LIBDIR%\JimiProClasses.jar
    
  7. You should now be able to run FOP from a DOS prompt — execute it without arguments to get a list of command options:

    > fop.bat
    

31.2. Installing FOP on Linux

Here's how I installed FOP on Fedora Core 1:

  1. Download latest FOP distribution from http://xml.apache.org/fop/.

  2. Install the FOP distribution:

    $ su
    # cd /usr/local/lib
    # unzip ~srackham/tmp/fop-0.20.5-bin.zip
    # cp /usr/local/lib/fop-0.20.5/fop.sh /usr/local/bin
    # chmod +x /usr/local/bin/fop.sh
    
  3. Edit the FOP start script fop.sh adding this line to the start of the script:

    FOP_HOME=/usr/local/lib/fop-0.20.5
    
  4. Download the JIMI image processing library from http://java.sun.com/products/jimi/.

  5. Extract the JimiProClasses.jar library from the JIMI distribution and copy to the FOP lib directory.

    # cp ~srackham/tmp/JimiProClasses.jar /usr/local/lib/fop-0.20.5/lib/
    
  6. You should now be able to run FOP from a DOS prompt — execute it without arguments to get a list of command options:

    $ fop.sh
    

31.3. Installing Java on Windows

First check that Java is not already installed:

  1. Open a DOS Command Prompt window.

  2. Enter this command:

    java -version
    

You should see something like this:

java version "1.4.2_01"
Java(TM) 2 Runtime Environment, Standard Edition (build 1.4.2_01-b06)
Java HotSpot(TM) Client VM (build 1.4.2_01-b06, mixed mode)

If you don't Java is not installed and you need to:

  1. Download the Java Runtime (JRE) for Windows from http://java.sun.com.

  2. Install using the instructions on the download page.

31.4. Installing Java on Linux

Check Java is not already installed by entering the following command:

$ java -version

You should see something like this:

java version "1.4.2_01"
Java(TM) 2 Runtime Environment, Standard Edition (build 1.4.2_01-b06)
Java HotSpot(TM) Client VM (build 1.4.2_01-b06, mixed mode)

If you don't Java is not installed and you need to download the Sun Java Runtime (JRE) for Linux from http://java.sun.com.

Here's how I installed the RPM version of the JRE on Fedora Core 1:

$ ./j2re-1_4_2_05-linux-i586-rpm.bin
$ su
# rpm -vih j2re-1_4_2_05-linux-i586.rpm
# vi /etc/profile.d/java.sh
# chmod +x /etc/profile.d/java.sh
^D
$ . /etc/profile.d/java.sh
$ java -version
java version "1.4.2_05"
Java(TM) 2 Runtime Environment, Standard Edition (build
1.4.2_05-b04)
Java HotSpot(TM) Client VM (build 1.4.2_05-b04, mixed mode)
$

The following two lines are entered into the /etc/profile.d/java.sh file:

export PATH=$PATH:/usr/java/j2re1.4.2_05/bin/
export JAVA_HOME=/usr/java/j2re1.4.2_05

32. XML and Character Sets

The default XML character set UTF-8 is used when AsciiDoc generates DocBook files (but you can change it by changing the xmldecl entry in the [attributes] section of the docbook.conf file or by composing your own configuration file [header] section).

If you're familiar with HTML there are many predefined character entities that you will have taken for granted — for example the non-breaking space character &nbsp;. XML has only five predefined named character entities: &amp;, &lt;, &gt;, &quot; and &apos;. Any others (for example &nbsp;) have to be either defined or included.

TIP:

If you get an undefined entity error when processing DocBook files you'll may find that you've used and HTML character entity that's not been included. An easy (although inelegant) fix is to use the character's character code instead of it's symbolic name (for example use &#160; instead of &nbsp;).

If your system has been configured with an XML catalog you may find a number of entity sets are already automatically included — if you're using Fedora Linux take a look at the global /etc/xml/catalog file.

32.1. PDF Fonts

The Adobe PDF Specification states that the following 14 fonts should be available to every PDF reader: Helvetica (normal, bold, italic, bold italic), Times (normal, bold, italic, bold italic), Courier (normal, bold, italic, bold italic), Symbol and ZapfDingbats. Non-standard fonts should be embedded in the distributed document.

33. Glossary

Block element
An AsciiDoc block element is a document entity composed of one or more whole lines which translates to a block of output lines.
Formal element
An AsciiDoc block element that has a BlockTitle. Formal elements are normally listed in front or back matter, for example lists of tables, examples and figures.
Inline element
AsciiDoc inline elements occur within block element textual content, they perform formatting and substitution tasks.
Verbatim element
The word verbatim indicates that white space and line breaks in the source document are to be preserved in the output document.