Export: hands-on examples

• To a buffer (the capital-letter keys: h H, m M, t A, t U, l L, l B, O O, m G): the result opens in a vertical split, in a scratch buffer named like "Org HTML Export #12". Nothing is written to disk; close it with :q. All exercises in this file use these keys.
• To a file (lower-case keys: h h, m m, t a, l l, o o, …): the file is written next to this one, inside the repo (examples/19-export.html and friends, or #+EXPORT_FILE_NAME). Only a few exercises do that and they say so; delete the files afterwards (git status shows them as untracked) or set export.output_dir in your config to send them elsewhere.

Almost every exercise exports one subtree: put the cursor anywhere inside a heading called "Sample: …" and press <prefix>e s first (s toggles "export scope = subtree"), then the back-end keys. A subtree export:

• uses only the body of that heading: its children become the top-level sections of the output, and the heading itself is not a section;
• reads EXPORT_* properties from the heading's :PROPERTIES: drawer (EXPORT_TITLE, EXPORT_OPTIONS, …). They override the #+... keywords of the file, so every sample carries its own settings.

Important: #+TITLE:, #+OPTIONS:, #+HTML_HEAD: and the other settings keywords apply to the whole file, wherever they are written, even when you export only a subtree. That is why this file shows file-level keywords inside #+begin_src org blocks (those are just code) and uses EXPORT_* properties for the samples. The only live settings keywords are the few at the top of this file and the #+MACRO: lines of the macros sample.

Each "Sample: …" heading is followed by an "Exercises: …" heading with the Try: / Expect: steps. Keep the cursor inside the sample when you export: with the cursor in the Exercises heading you would export the exercise text instead.

Hello from a subtree. It has emphasis, verbatim and a link.

Try: put the cursor on the line "Hello from a subtree" in the sample above. Press <prefix>e, then s (the menu now says "export scope = subtree"), then m, then M.

Expect: a vertical split with a Markdown buffer. Its whole content is:

Hello from a **subtree**. It has *emphasis*, `verbatim` and a [link](https://neovim.io).


# A child heading

Children of the sample become the top-level headings of the output.

The comment line (# This is the body ...) is not there, and Markdown has no title block, so EXPORT_TITLE and EXPORT_AUTHOR are not visible. Close the split with :q.

Try: same subtree, <prefix>e s t A (plain ASCII to a buffer).

Expect: now the title and author are centered at the top, and the link target is listed as a note after the paragraph:

                           _________________

                            MY FIRST EXPORT

                                  Ada
                           _________________


Hello from a *subtree*. It has /emphasis/, `verbatim' and a [link].


[link] <https://neovim.io>


A child heading
===============

  Children of the sample become the top-level headings of the output.

Try: same subtree, <prefix>e s b h H (subtree, body only, HTML).

Expect: no <!DOCTYPE>, <head> or title, just the body:

<p>
Hello from a <b>subtree</b>. It has <i>emphasis</i>, <code>verbatim</code> and a <a href="https://neovim.io">link</a>.
</p>
<div id="outline-container-org6793350" class="outline-2">
<h2 id="org6793350">A child heading</h2>

(The org6793350 ids are computed from the content, so they are the same on every export and change only when you edit the text.)

Try: <prefix>e h H without s: the whole file is exported. Search the HTML buffer for My first export with /: it is not found, because EXPORT_TITLE only applies to subtree exports; the <title> is now "Export: hands-on examples", from the #+TITLE: line at the top.

Try: :Org export md subtree buffer with the cursor in the sample: the same Markdown buffer as the first exercise, without the menu.

With v on, whatever is folded at that moment is left out: fold a subtree with <Tab> and its body disappears from the output (its heading line is visible, so it stays). Useful to export "just the outline" or to hide parts temporarily.

Combined with s, only the visible part of the subtree is exported; if the subtree's own heading is folded, its body is empty (Emacs does the same).

The intro paragraph is always visible.

Try: open the sample above so you see both sections, then fold "Section one" with <Tab> on its heading. Press <prefix>e v t A (no s: the whole file, visible parts only). In the output, search with /The intro paragraph.

Expect: the fold state of the whole file decides. Around the search hit you find (section numbers depend on what else is open):

  The intro paragraph is always visible.


2.4.1 Section one
-----------------


2.4.2 Section two
-----------------

  Text of section two.

"Text of section one." is missing because it was folded.

Try: <S-Tab> until the buffer shows only the top-level headings (OVERVIEW), then <prefix>e v t A again: the output is little more than the table of contents and the numbered top-level headings.

a only concerns PDF: with it on, LaTeX compilation (l p, l P) runs in the background and a message says when the PDF is ready, so you can keep editing. With the default config it already does (export.latex.async_compile = true); a matters when you set that to false. The other formats are fast and always synchronous. :Org export pdf async is the command form.

<prefix>e # asks for an option category (default, html, latex, md, …; <Tab> completes) and inserts every option with its current value:

• without s: #+options: lines and #+title:, #+author:, #+date:, #+email:, #+language:, #+select_tags: … keywords at the cursor;
• with s: EXPORT_OPTIONS, EXPORT_TITLE, EXPORT_AUTHOR, … properties in the drawer of the current heading.

Revenue went up.

Try: cursor on "Revenue went up.", <prefix>e s t A.

Expect: a banner with the title, the subtitle, the author, the e-mail (because of email:t) and the date:

                          ____________________

                            QUARTERLY REPORT
                           Third quarter 2026

                              Ada Lovelace
                            ada@example.org
                          ____________________


                            <2026-10-01 Thu>


Revenue went up.

Try: <prefix>e s h H and look at the top of the HTML buffer.

Expect: the title in <head>, and as the first heading with the subtitle under it:

<title>Quarterly report</title>
<meta name="author" content="Ada Lovelace" />
...
<h1 class="title">Quarterly report
<br />
<span class="subtitle">Third quarter 2026</span>
</h1>

and at the end a postamble with the date, author, e-mail link and the creation time of the export (timestamp:t):

<div id="postamble" class="status">
<p class="date">Date: 2026-10-01 Thu 00:00</p>
<p class="author">Author: Ada Lovelace</p>
<p class="email">Email: <a href="mailto:ada@example.org">ada@example.org</a></p>
<p class="date">Created: 2026-09-28 Mon 15:13</p>

Try: <prefix>e s l L (LaTeX to a buffer).

Expect: the e-mail becomes a \thanks, the subtitle a second title line:

\author{Ada Lovelace\thanks{ada@example.org}}
\date{\textit{<2026-10-01 Thu>}}
\title{Quarterly report\\\medskip
\large Third quarter 2026}

Try: change email:t to email:nil in EXPORT_OPTIONS and export as ASCII again: the ada@example.org line disappears. With date:nil the date goes, with author:nil the author.

Voir la note¹.

Try: <prefix>e s t U (UTF-8 text to a buffer).

Expect: the words the exporter generates are in French:

Table des matières
──────────────────

1. Première partie


Voir la note[1].


1 Première partie
═════════════════

  Du texte.



Notes de bas de page
────────────────────

[1] Une note de bas de page.

Try: <prefix>e s t A (plain ASCII): the table of contents is now titled Sommaire, the ASCII-only form of the French dictionary.

Try: change fr to de and export again: Inhaltsverzeichnis and Fußnoten.

Intro.

Try: <prefix>e s t A.

Expect: a table of contents with two levels, numbered sections, and the third level turned into a list item (a * bullet) under 1.1:

Table of Contents
_________________

1. Chapter
.. 1. Section
2. Another chapter


Intro.


1 Chapter
=========

1.1 Section
~~~~~~~~~~~

* 1.1.1 Too deep to be a section

  This one becomes a list item.


2 Another chapter
=================

  Text.

Try: change EXPORT_OPTIONS to toc:nil num:1 H:2 title:nil author:nil and export again: no table of contents, and only the first level is numbered: 1 Chapter, then Section without a number, then * Too deep to be a section, then 2 Another chapter.

Try: H:3 instead of H:2: "Too deep to be a section" is a real (sub-sub)section again, underlined with dashes.

file_name and x^2 stay as they are, but H₂O and E = mc² are converted.

Try: <prefix>e s b m M (body only, Markdown).

Expect:

file\_name and x^2 stay as they are, but H<sub>2</sub>O and E = mc<sup>2</sup> are converted.

Try: change ^:{} to ^:t and export again: file_name becomes file<sub>name</sub> and x^2 becomes x<sup>2</sup> too. With ^:nil nothing is converted. ^:{} is the usual choice for documents full of snake_case names.

Try: <prefix>e s b m M.

Expect: keywords, priorities, cookies and tags are all there (the statistics cookie is wrapped in <code>, like Emacs does):

# TODO [#A] Write the report     :work:


# DONE Book the room     :office:


# NEXT [#C] Plan the offsite <code>[1/2]</code>

-   [X] pick a date
-   [ ] pick a place

Try: set EXPORT_OPTIONS to toc:nil num:nil title:nil author:nil tags:nil todo:nil pri:nil stat:nil (one line) and export again: the headings become plain # Write the report, # Book the room and # Plan the offsite; the checklist stays.

Try: replace tags:t todo:t pri:t by tasks:todo: only the not-done entries are exported (# TODO Write the report :work: and # NEXT Plan the offsite <code>[1/2]</code>; the priority is gone too because pri: is nil by default). tasks:done keeps only "Book the room", tasks:("NEXT") only "Plan the offsite", tasks:nil drops all three.

Try: <prefix>e s b t A.

Expect: by default only the heading, the NOTES drawer (every drawer but LOGBOOK is exported), the lone timestamp and the paragraph:

DONE Ship the release
=====================

  Remember to tag the commit.
  <2026-10-05 Mon>

  The release went out on time.

Try: one at a time, add these to EXPORT_OPTIONS and export again (each adds or removes lines between the underline and "Remember to tag"):

• p:t adds CLOSED: [2026-09-25 Fri 17:02] SCHEDULED: <2026-09-24 Thu>;
• prop:t adds VERSION: 2.1 and OWNER: Ada; prop:("OWNER") only OWNER: Ada;
• c:t alone changes nothing: the clock line is inside LOGBOOK, which is still excluded. c:t d:t (all drawers) shows CLOCK: [2026-09-24 Thu 09:00]--[2026-09-24 Thu 11:30] => 2:30;
• d:nil removes "Remember to tag the commit." (no drawers at all); d:("NOTES") exports only the NOTES drawer (same output as now);
• <:nil removes <2026-10-05 Mon>, the paragraph made only of a timestamp (an empty line is left in its place).

Roses are red, violets are blue. "Quoted" text – with an en dash — and an em dash…

Try: <prefix>e s b h H.

Expect: each line ends with <br /> (from \n:t), the quotes become curly (':t) and the dashes and dots entities (-:t):

<p>
Roses are red,<br />
violets are blue.<br />
&ldquo;Quoted&rdquo; text &ndash; with an en dash &mdash; and an em dash&hellip;<br />
</p>

Try: set EXPORT_OPTIONS to toc:nil title:nil author:nil \n:nil ':nil -:nil and export again: the three lines come back as typed, with straight quotes, --, --- and ..., and no <br />.

A bold claimα with a footnote².

fixed-width lines are dropped too

Try: <prefix>e s b m M.

Expect: only one line: the asterisks stay literal (*:nil), \alpha stays as typed (e:nil), no footnote, no table, no fixed-width line. Markdown escapes the characters that would otherwise be markup, so you see backslashes:

A \*bold\* claim\\alpha with a footnote.

Try: remove *:nil e:nil and export again. Now the emphasis is converted and the entity is written as HTML:

A **bold** claimα with a footnote.

• A heading tagged :noexport: is removed with its whole subtree.
• If any heading carries an :export: tag, only the tagged subtrees are exported (in a whole-file export the text before the first heading is kept too).
• #+EXCLUDE_TAGS: / #+SELECT_TAGS: (or the EXPORT_EXCLUDE_TAGS / EXPORT_SELECT_TAGS properties) replace those tag lists. Tags are inherited: a child of a :noexport: heading goes too.

Try: <prefix>e s b m M.

Expect: two headings only, "Kept" and "Also kept"; no "Private text":

# Kept

Public text.


# Also kept

More public text.

Intro text of the sample.

Try: <prefix>e s b m M.

Expect: only the heading "Chosen". Even the intro text is dropped: in a subtree export it belongs to the sample heading, which is not selected (Emacs does the same).

# Chosen

In.

Try: remove the keep tag from both "Chosen" headings (edit the lines, or <prefix>t on each heading and delete keep from the prompt). Now no heading is selected, so the select tags do not apply at all:

Intro text of the sample.


# Chosen

In.


# Not chosen

Out: no keep tag, and another heading has one.

"Chosen but secret" is still excluded by its secret tag. Press u to restore the tags.

Visible paragraph.

Try: <prefix>e s b m M.

Expect:

Visible paragraph.


# Finished part

Done.

Try: put the cursor on the "COMMENT Work in progress" heading and press <prefix>hC (toggle COMMENT, Emacs C-c ;): the word COMMENT goes away. Export again: "Work in progress" and its text are now exported between "Visible paragraph." and "Finished part". Press <prefix>hC again to put it back.

Try: <prefix>e s b m M.

Expect: "Old project" appears as a heading, but "Old details." is gone (arch:headline, the default):

# Current work

Live.


# Old project

Try: add arch:t to EXPORT_OPTIONS: "Old details." comes back. With arch:nil the "Old project" heading goes too.

A standalone page made from one heading.

Try: <prefix>e s h H.

Expect: <title>Sample: a subtree is its own document</title>, the meta line from EXPORT_HTML_HEAD inside <head>, and "Author: Grace" in the postamble.

Try (writes a file outside the repo): <prefix>e s m m. Because of EXPORT_FILE_NAME, the message says Exported to /tmp/org-nvim-sample.md (the extension is added for you). A relative name would be relative to this file, i.e. inside examples/.

Try: <prefix>e s b t A.

Expect: a first list with the level-1 headings (Preface without a number, 1. Setup, no Colophon), then under "1 Setup" a local list with only its children:

Table of Contents
_________________

Preface
1. Setup


Preface
=======

  No number here.


1 Setup
=======

  .. 1. Install
  .. 2. Configure


1.1 Install
~~~~~~~~~~~


1.2 Configure
~~~~~~~~~~~~~


Colophon
========

  Not in the table of contents.

Try: change #+TOC: headlines 1 to #+TOC: headlines 2: the first list now also shows .. 1. Install and .. 2. Configure under 1. Setup.

Try: <prefix>e s b h H and look for "Discussion".

Expect: the captions are numbered, the named elements and the target get ids, and the links point to them. The heading with a CUSTOM_ID uses it as its id (data); the other ids look like org plus 7 hex digits:

<h2 id="data"><span class="section-number-2">1.</span> Data</h2>
<table id="orgc141a8c" border="2" cellspacing="0" cellpadding="6" rules="groups" frame="hsides">
<caption class="t-above"><span class="table-number">Table 1:</span> Monthly sales</caption>
...
<label class="org-src-name"><span class="listing-number">Listing 1: </span>Doubling a number</label><pre class="src src-lua" id="org608ce3d"><code>return 2 * 21
...
<li><a id="org7457e97"></a>Check the totals.</li>
...
See table <a href="#orgc141a8c">1</a>, listing <a href="#org608ce3d">1</a> and section <a href="#data">1</a>. The
same section by id: <a href="#data">the data section</a>. Never skip step <a href="#org7457e97">2</a>.

A link to a <<target>> without a description shows the number of what holds the target: here the list item, so "step 2".

(The generated ids are computed from the content, so they stay the same from one export to the next.)

Try: <prefix>e s b t A: in plain text the references are just the numbers, and the table and the listing get their captions below them:

   Month  Units
  --------------
   Oct       12
   Nov       17
  Table 1: Monthly sales

  ,----
  | return 2 * 21
  `----
  Listing 1: Doubling a number
...
  See table 1, listing 1 and section 1. The same section by id: [the
  data section]. Never skip step 2.


[the data section] See section 1

(The last line is the plain-text version of a described link: a note after the section.)

Try: in the Discussion paragraph, replace [[*Data]] by [[*Dataa]] and export again: the link no longer resolves. Thanks to broken-links:mark at the top of this file you get [BROKEN LINK: *Dataa] instead of an error (see "Broken links" below). Press u.

A named note³, an inline one⁴ and the named note again³.

Try: <prefix>e s b t A.

Expect: references become [1] and [2] (the second use of sample is [1] again), with a footnotes section at the end:

A named note[1], an inline one[2] and the named note again[1].
...
Footnotes
_________

[1] The definition of the named note.

[2] Defined right here.

More on footnotes: 14-footnotes.org.

Hello, world! two before one. This is "Export: hands-on examples" by org.nvim examples, exported from 19-export.org. Step 1, step 2, step 3; again step 3. Press C-c C-e.

Try: <prefix>e s b t A.

Expect:

Hello, world!  two before one.  This is "Export: hands-on examples" by
org.nvim examples, exported from 19-export.org.  Step 1, step 2, step 3;
again step 3.  Press .


Owned part
==========

  This part is owned by Charles.

Things to notice:

• {{{title}}} and {{{author}}} are the file's #+TITLE: and #+AUTHOR:, not the sample's EXPORT_TITLE (Emacs does the same).
• The kbd macro produced nothing in plain text: it expands to an HTML export snippet (next section). With <prefix>e s b h H the line reads Press <kbd>C-c C-e</kbd>.
• {{{property(OWNER)}}} reads the heading it is written under. Text right under the exported heading reads that heading's own drawer: in the sample, {{{property(EXPORT_TITLE)}}} would give Macro demo.

Try: add a line {{{greet(Neovim\, again)}}} to the sample and export: Hello, Neovim, again! (the escaped comma is not an argument separator).

This word is highlighted in HTML, emphasized in LaTeX.

The end.

Try: <prefix>e s b h H.

Expect: the HTML parts only:

<p>
This word is <mark>highlighted</mark> in HTML, emphasized in LaTeX.
</p>
<hr class="fancy" />
<div class="note">Only in HTML.</div>
<p>
The end.
</p>

Try: <prefix>e s b t A.

Expect: no markup at all, and the ascii block:

This word is highlighted in HTML, emphasized in LaTeX.
Only in plain text.
The end.

Try: <prefix>e s b l L.

Expect: the LaTeX parts:

This word is highlighted in HTML, \emph{emphasized} in LaTeX.
\newpage
\begin{center}Only in \LaTeX.\end{center}
The end.

Neovim

• one
• two

Try: <prefix>e s h H.

Expect: the style line at the end of <head>, and in the body:

<style>.wide { width: 100%; }</style>
</head>
...
<table id="team" border="2" cellspacing="0" cellpadding="6" rules="groups" frame="hsides" class="wide">
...
<p target="_blank" title="Neovim home page">
<a href="https://neovim.io" target="_blank" title="Neovim home page">Neovim</a>
</p>

<ul class="org-ul checklist">

The :id replaced the generated table id, the :class was added to the default attributes (for lists, to the org-ul class). For a paragraph holding only a link, the attributes land on both the <p> and the <a>, exactly like Emacs. The page ends right after the content: no postamble (html-postamble:nil).

Inline math \(a^2 + b^2 = c^2\) and a display:

Try: <prefix>e s l L.

Expect: the class, its options and the extra package in the preamble:

\documentclass[a4paper]{report}
...
\usepackage{booktabs}

and in the body a table float with [h], the lr column spec and booktabs rules:

\begin{table}[h]
\caption{Results}
\centering
\begin{tabular}{lr}
\toprule
Test & Score\\
\midrule
A & 91\\
\bottomrule
\end{tabular}
\end{table}

The math is copied as it is. With the report class, level-1 headings would be \chapter instead of \section.

Try: <prefix>e s l B.

Expect: a Beamer document with the theme, a title frame, then one frame per level-1 heading of the sample:

\documentclass[presentation]{beamer}
...
\usetheme{Madrid}
\author{Ada}
\date{\today}
\title{A tiny talk}
...
\maketitle
\begin{frame}[label={sec:orgcc50823}]{First slide}
\begin{itemize}
\item point one
\item point two
\end{itemize}
\end{frame}
\begin{frame}[label={sec:orgcd668b6}]{Two columns}
\begin{columns}
\begin{column}{0.5\columnwidth}
Text on the left.
\end{column}
\begin{column}{0.5\columnwidth}
\begin{alertblock}{Right}
Careful!
\end{alertblock}
\end{column}
\end{columns}
\end{frame}

• [X] done item
• [ ] open item

print("hi")

Try: <prefix>e s b m M (plain Markdown).

Expect: Markdown has no table syntax, so the table is written as HTML; the code block is indented by four spaces:

-   [X] done item
-   [ ] open item

<table border="2" cellspacing="0" cellpadding="6" rules="groups" frame="hsides">
...
    print("hi")

Try: <prefix>e s b m G (GitHub flavoured Markdown).

Expect: task-list items, a pipe table (not padded, which GitHub does not need) and a fenced block with the language:

- [x] done item
- [ ] open item

| Tool | Needs |
|---|---|
| md | nothing |

```python
print("hi")
```

Try: <prefix>e s b t A (ASCII).

Expect:

- [X] done item
- [ ] open item

 Tool  Needs
---------------
 md    nothing

,----
| print("hi")
`----

Try: <prefix>e s b t U (UTF-8).

Expect: the same drawn with Unicode characters:

• ☑ done item
• ☐ open item

━━━━━━━━━━━━━━━
 Tool  Needs
───────────────
 md    nothing
━━━━━━━━━━━━━━━

┌────
│ print("hi")
└────

t L (Latin-1) sits in between: Latin-1 characters where they exist, ASCII otherwise.

Try: <prefix>e s b O O (Org back-end).

Expect: the subtree comes back as Org, after macros, #+INCLUDE and noexport have been processed (useful to see exactly what the other back-ends receive). Note the code indented by two spaces, like Emacs:

- [X] done item
- [ ] open item

| Tool | Needs   |
|------+---------|
| md   | nothing |

#+begin_src python :exports code
  print("hi")
#+end_src

The first lines of the bundled init file:

-- Try org.nvim without touching your own config:
--
--   nvim -u examples/minimal_init.lua examples/tutorial.org

And a section of the tutorial: Export options go in #+OPTIONS: at the top of a file, for example #+OPTIONS: toc:2 num:nil ^:{} todo:nil. #+TITLE:, #+AUTHOR: and

Try: <prefix>e s b m M.

Expect: the three comment lines of minimal_init.lua as a code block, followed by the first two lines of the tutorial's "Export settings" section:

The first lines of the bundled init file:

    -- Try org.nvim without touching your own config:
    --
    --   nvim -u examples/minimal_init.lua examples/tutorial.org

And a section of the tutorial:
Export options go in `#+OPTIONS:` at the top of a file, for example
`#+OPTIONS: toc:2 num:nil ^:{} todo:nil`. `#+TITLE:`, `#+AUTHOR:` and

#+SETUPFILE: "other.org" is related: it reads the settings keywords (#+OPTIONS, #+MACRO, #+TODO, …) of another file.

(+ 1 2)

(* 6 7)

(concat "org" "." "nvim")

(message "invisible")

(* 100 100)

stale value kept on export

The answer is (* 6 7).

Try: <prefix>e s b t A. You are asked "Evaluate this emacs-lisp code block on your system?" three times (the results block, the both block and the inline src_ call): answer y each time. (Emacs Lisp runs in emacs --batch when Emacs is installed, otherwise on org.nvim's own Lisp interpreter, which handles these simple expressions.)

Expect: code and results in the text back-end's boxes; the inline result is verbatim (`42'):

,----
| (+ 1 2)
`----

,----
| 42
`----

,----
| (concat "org" "." "nvim")
`----

,----
| org.nvim
`----

,----
| stale value kept on export
`----

The answer is `42'.

The :exports none block is absent, and the never-export block shows its old #+RESULTS: instead of 10000. The buffer is not modified: no #+RESULTS: was added under the other blocks.

Try: export again and answer n to every question: the results blocks then export nothing (they have no #+RESULTS: in the buffer), the both block only its code, and the inline call is left empty.

As shown before (Doe, John and Smith, Jane, 2020), and in a book (see Zed, Anna, 2019 p. 3). Alpha, Bob (2018) wrote a thesis.

Alpha, Bob (2018). Thesis, Univ.

Doe, John and Smith, Jane (2020). On the TeXbook and \(\alpha\) α things, Journal of Tests.

Zed, Anna (2019). A Book of Strings, ACME Press.

Try: <prefix>e s b t A.

Expect: author-year citations, and the bibliography sorted by author where #+PRINT_BIBLIOGRAPHY: is:

As shown before (Doe, John and Smith, Jane, 2020), and in a book (see
Zed, Anna, 2019 p. 3).  Alpha, Bob (2018) wrote a thesis.

Alpha, Bob (2018). /Thesis/, Univ.

Doe, John and Smith, Jane (2020). /On the TeXbook and $\alpha$ alpha
things/, Journal of Tests.

Zed, Anna (2019). /A Book of Strings/, ACME Press.

Try: <prefix>e s b h H: the same text in <p> elements, with the titles in <i> and the $\alpha$ of the .bib title as MathJax \(\alpha\).

Try: change [cite:@doe2020] to [cite/na:@doe2020] (no author): As shown before (2020). And [cite:see @zed2019 p. 3] to [cite/nb:@zed2019] (numeric): in a book (3), the entry's position in the bibliography.

This points to [BROKEN LINK: *A heading that does not exist].

Try: <prefix>e s b t A.

Expect: This points to [BROKEN LINK: *A heading that does not exist].

Try: change broken-links:mark to broken-links:nil in EXPORT_OPTIONS and export again. No buffer opens; instead:

Export failed: Org export aborted.  Unable to resolve link: "*A heading that does not exist"
See export.with_broken_links (org-export-with-broken-links)

Try: broken-links:t: the link simply disappears (This points to .).

Try (writes examples/19-export.ics): <prefix>e c f. The dispatcher's iCalendar entries always export the whole file, so the three "Evaluate?" questions of the :exports sample come up again: answer n. Open the file with :e examples/19-export.ics.

Expect: among others

BEGIN:VEVENT
DTSTART:20261007T100000
DTEND:20261007T110000
SUMMARY:Team meeting
...
BEGIN:VEVENT
DTSTART;VALUE=DATE:20261015
DTEND;VALUE=DATE:20261016
SUMMARY:DL: Submit the report

and an event for <2026-10-05 Mon> of the "Ship the release" sample. A TODO entry's deadline becomes a VTODO only with export.icalendar.include_todo; that is why "Submit the report" has no TODO keyword. Afterwards: :!rm examples/19-export.ics.

Try (writes examples/19-export.odt): <prefix>e s o o, open the file in LibreOffice (or <prefix>e s o O to open it right away), then delete it.