Supported Markdown features

• Backticks. Text inside ‘`backticks`’ will be put inside ‘<code></code>’. Text surrounded by triple backticks (‘```’) on the lines by themselves will be put inside ‘<pre></pre>’. Any “less-than” characters inside backticks will be converted to ‘<’ and “ampersand” characters to ‘&’. Any text following the first triple backticks until the end of the line will be ignored. This is to account for language highlighting specification, which isn't yet supported.
• Blockquotes. Starting the line with ‘>’ will surround it with a ‘<blockquote>’ tag. Multiple lines are supported.
• Bold/italic. Text inside ‘*asterisks*’ or ‘_underlines_’ will be put inside ‘<em></em>’. Text inside ‘**double asterisks**’ or ‘__double underlines__’ will be put inside ‘<strong></strong>’. ‘***More than two***’ of the ‘___same symbol___’ should be avoided.
• Break marks. To address combinations of lists and “raw” tags, which are problematic to mix together, slweb supports explicitly marking a tag with a “break mark”. This currently means that when such a tag is introduced, any running list item and list will be closed prior to outputting the tag. Tags with break marks are introduced by prepending the exclamation mark to the tag name. For example, without a break mark, the input:

  - Some list
  {break-here}

  produces the following output:

  <ul><li><p>Some list
  <break-here></li></ul>

  When changed to:

  - Some list
  {!break-here}

  slweb will output:

  <ul><li><p>Some list
  </li></ul>
  <break-here>

  Same for end tags:

  {dl}{dt}{p}Introduction{/dt}
  {dd}
  - One
  - Two
  {/dd}{/dl}

  produces:

  <dl><dt><p>Introduction</dt>
  <dd>
  <ul><li><p>One
  </li>
  <li><p>Two
  </dd></dl></li></ul>

  while

  {dl}{dt}{p}Introduction{/dt}
  {dd}
  - One
  - Two
  {/!dd}{/dl}

  produces:

  <dl><dt><p>Introduction</dt>
  <dd>
  <ul><li><p>One
  </li>
  <li><p>Two
  </li></ul>
  </dd></dl>

• Footnotes (regular and inline). Two-part regular footnotes can be added in a manner similar to links (see Links). Inline footnotes are also supported.
  • Regular footnotes have two mandatory parts: first, the footnote mark is represented by ‘[^footnoteid]’, where footnoteid is the footnote identifier. Second, the text of the footnote is represented by ‘[^footnoteid]: sometext’, with sometext being the text of the footnote. Footnote text construct needs to begin at the start of a line, and it is the most practical to have it placed near the bottom of the file, similar to normal links.

    Combined, the input:

    This is a text[^first].
    
    [^first]: With a footnote!

    gives (some lines are wrapped for this manual):

    <p>This is a text<a href="#footnote-1" id="footnote-text-1">
    <sup>1</sup></a>.
    
    <hr>
    <p id="footnote-1"><a href="#footnote-text-1">1.</a>
    With a footnote!

  • Inline footnotes can be added by using the construct ‘^[footnotetext]’. For example, input

    Inline footnote^[Footnote text] in a paragraph.

    will produce

    Inline footnote<a href="#inline-footnote-1"
    id="inline-footnote-text-1"><sup>1</sup></a> in a
    paragraph.
    <hr>
    <p id="inline-footnote-1"><a href="#inline-footnote-text-1">
    1.</a> Footnote text

    The horizontal rule and the list of footnotes is output only once, before the end of HTML document body. Rule and the list of footnotes can be surrounded by a div with the “footnotes” id by setting the YAML variable add-footnote-div to “1” (see add-footnote-div).

    Inline and regular footnotes can be used at the same time, but they don't share the numbering. This doesn't affect functionality, but it does affect presentation, as some footnote numbers will overlap. As a result, whenever slweb encounters both footnote types in the same document, a warning will be issued to stderr.

• Headings. A line starting with ‘#’ followed by space will be put inside ‘<h?></h?>’, where ‘?’ stands for 1-4, depending on the number of hash signs. Tags will have id attributes set to ‘heading-?’, with ‘?’ set to the number representing its position within the list of all headings.
• Horizontal rules. Three dashes at the start of the line will produce a ‘<hr>’ in the output. As this Markdown feature clashes with YAML block boundaries, it will work only if YAML block is present in the input and before the three dashes for the horizontal rule. Any text following the three dashes on the same line will be ignored.

  Alternatively, three asterisks (‘***’) can be used instead of dashes, avoiding the necessity for a YAML block preceding the asterisks.

• Images. Similar to links (see Links), images can be added by using:

  ![Some title](/path/to/image.png)

  Which will produce:

  <figure>
  <a href="/path/to/image.png" title="Some title" class="image"
    target="_blank"><img src="/path/to/image.png" alt="Some title"
    width="100" height="50" ></a><figcaption>Some title</figcaption>
  </figure>

  The attributes width and height will be determined automatically using the command identify(1) (if present), provided the src attribute contains a valid file path. (Also see image-file-prefix).

  As with links, a form similar to the “regular form” of links can also be used, using the image id instead of the direct URL.

  If the additional link or ‘<figure>’/‘<figcaption>’ tags are not desirable, they can be turned off by setting add-image-links and add-figcaption to “0”.

• Keyboard tags. Text inside ‘||double bars||’ will be put inside ‘<kbd></kbd>’ tags. As with backticks, any “less-than” character will be converted to ‘<’.
• Line breaks. Two spaces followed by a newline will become ‘<br>’.
• Links.
  • Inline links. The construct ‘[A link](https://example.com)’ will be converted into ‘<a href="https://example.com">A link</a>’.

    Special case is the form

    [=somemacro Link title](https://anything)
    which prepends the body of a macro somemacro into the ‘<a></a>’ tag (here broken into multiple lines for clarity):

    <a href="https://anything">
    <!-- contents of somemacro -->
    Link title</a>

    This can, for example, be used to add SVG icons to links. See Macros.

    Also, the form ‘[(Link title)](http://asite.com)’ will surround the link title (text between ‘<a></a>’ tags) with ‘<span></span>’, like so:

    <a href="http://asite.com"><span>Link title</span></a>

    This allows for separate styling of link text. It can be combined with the macro-form:

    [=somemacro (Link title)](http://asite.com)

    It will prepend the body of a macro somemacro outside of the ‘<span></span>’:

    <a href="http://asite.com">
    <!-- contents of somemacro -->
    <span>Link title</span></a>

    Exception is the form

    [Some text (parenthesized)](https://somelink)
    , which won't output ‘<span>’ tags, and will simply include the parentheses in the output instead. This form doesn't apply to macro-form
    [=macro Some (text)](http://link)
    , which will always output ‘<span>’ tags.
  • Regular links. Everything said about inline links applies to regular links, with the exception that instead of the parenthesized URL, link text inside brackets should be followed by link id (different than the id attribute!) inside brackets, and a separate definition of that link id is needed, usually near the end of input.

    For example:

    Here's a link to [my website][mw].
    
    [mw]: https://mysite.com

    will produce:

    <p>Here's a link to <a href="https://mysite.com">my website</a>.

    This can help reduce the amount of code in the text of the page.

• Lists. Lines starting with a dash (‘-’), an asterisk (‘*’) or lowercase “o”, followed by a space, will start an unordered list (if used for the first time) and start a list item. Input:

  Colors are:
  
  - Red
  - Green
  - Blue
  
  and so on.

  will produce:

  <p>Colors are:
  
  <ul>
  <li><p>Red
  </li>
  <li><p>Green
  </li>
  <li><p>Blue
  </li></ul>
  <p>And so on.

  Indenting a line after a blank line with two spaces will create a paragraph within the list item, otherwise it will end the list:

  - This is a text inside a list item.
  
      Remarkably, this paragraph is as well.
  - This is already a second item.
  
  This is after the list.

  will produce:

  <ul><li><p>This is a text inside a list item.
  <p>Remarkably, this paragraph is as well.
  </li>
  <li><p>This is already a second item.</li></ul>
  <p>This is after the list.

  Lists can't be nested, and the characters inside an existing list which would otherwise start a list shall just be inserted as-is in the output.

• Non-breaking hyphen. Tilde followed by a hyphen (‘~-’) will insert ‘‑’ in the output.
• Non-breaking space. Two consecutive tildes (‘~~’) will produce ‘ ’ in the output.
• Paragraphs. Text surrounded by newlines will be prepended by ‘<p>’ tag. See Known limitations.
• Strikethrough. Text surrounded by tildes (‘~’) will be put inside ‘<s></s>’.
• Tables. Lines starting with vertical bars (‘|’) will be transformed into HTML tables. The first such line is treated as a header row, second as the alignment specifier (currently ignored), and the rest regular rows. Table cells and header cells are specified by additional bar characters. For example, input:

  | Month | Jan | Feb | Mar | Apr |
  |-------|-----|-----|-----|-----|
  | Qty   | 2.35| 1.2 |  8  |  15 |

  will be transformed into:

  <table>
  <thead>
  <tr><th> Month </th><th> Jan </th><th> Feb </th><th> Mar </th>
  <th> Apr </th></tr>
  </thead>
  <tbody>
  <tr><td> Qty   </td><td> 2.35</td><td> 1.2 </td><td>  8  </td>
  <td>  15 </td></tr>
  </tbody>
  </table>

• Partial tables are a modification of Markdown tables, allowing them to be combined with the TSV or CSV directive. In order for tables to work with templating, they need two additional lines at the top and the bottom, to mark the start and the end of the table. All the lines in partial tables need to start with ‘|’, with only the character following at-mark being different:

  |@\------|-----|-----|-----|-----|
  {tsv "../tsv/sales" 1}
  |@#  $#1 | $#2 | $#3 | $#4 | $#5 |
  {/tsv}
  |@-------|-----|-----|-----|-----|
  {tsv "../tsv/sales"}
  |@    $1 | $2  | $3  | $4  | $5  |
  {/tsv}
  |@/------|-----|-----|-----|-----|

  The rest of the line following the third character in top, bottom and lines separating header from body is ignored by the parser and is included in this example only for aesthetic purposes. See TSV/CSV templating.

• Zero width space. Tilde followed by an exclamation mark (‘~!’) will produce ‘​’ in the output.

Math mode

Math mode is supported through KaTeX. With katex installed, anything between dollar signs (‘$’) will be transformed into MathML and HTML markup. To generate display math, use double dollar signs (‘$$’). The text between the dollar signs in both cases should be LaTeX source code, and is passed to katex. The KaTeX stylesheet is not included, and needs to be included separately through the stylesheet or inline-stylesheet YAML variables (see stylesheet, inline-stylesheet).

Directives

• Brand “watermark”
  . Directive ‘{made-by}’ results in a div to be output at the end of the document, with the id made-by containing the “watermark” text and a link to the slweb home page.
• General tags. Directive ‘{sometag}{/sometag}’ will be transformed into ‘<sometag></sometag>’.
• Class attributes. Directive ‘{tag.myclass mysecondclass}{/tag}’ will be transformed into ‘<tag class="myclass mysecondclass"></tag>’. Only one class attribute is permitted per tag directive (subsequent dots will be inserted as part of the class attribute), but you can include multiple classes by separating them with a space, just like in HTML, and you can have both a class attribute and an id attribute per tag.

  A variation of this is to use a special form ‘{.myclass}{/.myclass}’, which will be transformed into ‘<div class="myclass"></div>’.

• Id attributes. Directive ‘{tag#myid}{/tag}’ will be transformed into ‘<tag id="myid"></tag>’. Only one id attribute is permitted per tag directive (subsequent hash signs will be included as part of the id attribute), and you can have both an id attribute and a class attribute per tag.

  A variation of this is to use a special form ‘{#myid}{/#myid}’, which will be transformed into ‘<div id="myid"></div>’.

• Includes. Directive ‘{include somefile }’ will fork, parse somefile.slw (related to basedir) and output the resulting HTML as if the option --body-only was specified. All YAML variables will be preserved.
• Macros. Macros can be declared using the directive ‘{=!macroname}{/=!macroname}’, where macroname is the name of the macro. Anything between those two tags will then become the body of a macro. Whenever ‘{=macroname}’ is subsequently encountered in the file, it will be replaced by the macro body. Macro definitions can't be nested nor can they contain macro calls, and doing so will produce an error. Tags won't be processed within the body of a macro.
• Previous Git commit information. Directive ‘{git-log}’ is converted into a div with the id ‘git-log’ containing the information about the previous commit (as having information about the current commit would be impossible).
• Subdirectory inclusion (blogging directive). The directive

  {incdir "dirname" [num] [=macroname] [listonly]}

  (num, macroname and the literal ‘listonly’ are optional) will be expanded as follows:
  1. A ‘<ul class="incdir">’ tag will be inserted into the document instead of the directive.
  2. For every subdirectory of dirname, relative to current directory and up to num (if present) or 5 (if omitted) total subdirectories, a ‘<li>’ tag will be inserted into the ‘ul’ tag.

     If the parameter listonly was specified, a ‘<li>’ tag will instead be inserted for every .slw file within each of the subdirectories, sorted in reverse lexicographical order. In this case, each ‘<li>’ will contain a permalink with a title preceded by a timestamp formatted according to list_timestamp_format. Title and date are read from each .slw file, as YAML variables title and date. For this case, processing of the incdir directive stops here.

  3. A ‘<details>’ tag will be inserted into each ‘<li>’ tag.
  4. Inside of the ‘<details>’ tag, a ‘<summary>’ tag will be inserted with the name of the subdirectory. If macroname is present, the name of the subdirectory will be prepended with the body of a macro macroname (for example, this can be used to include custom SVG markup as arrows).
  5. After ‘<summary>’ tag, a ‘<div>’ tag will be inserted into ‘<details>’, containing the concatenated output from processing each of the .slw files inside that subdirectory in reverse lexicographical order, in a manner similar to the include directive. (See Includes and date, ext-in-permalink, permalink-url variables.) The output of each file will be surrounded by an ‘<article>’ tag.

  If the included file has the variable incdir-only-summary set to “1”, only the text within the summary marks (‘/-’ and ‘-/’) will be output, and if the variable incdir-footer-permalink-text is also set, ‘<footer>’ element with a permalink to the article will be added after it. Example:

  (posts/blog-post.slw)

  ---
  incdir-only-summary: 1
  ---
  
  /-This article is great!-/ The quick brown fox jumps over the lazy sleeping dog.
  The quick brown fox jumps over the lazy sleeping dog. The quick brown fox jumps
  over the lazy sleeping dog. The quick brown fox jumps over the lazy sleeping
  dog. The quick brown fox jumps over the lazy sleeping dog. The quick brown fox
  jumps over the lazy sleeping dog.

  (blog-index.slw)

  ---
  incdir-footer-permalink-text: Click here!
  ---
  {incdir "."}

  will produce

  <!DOCTYPE html>
  <html lang="en">
  <head>
  <title></title>
  <meta charset="utf-8">
  <meta name="viewport" content="width=device-width, initial-scale=1">
  <meta name="generator" content="slweb">
  </head>
  <body>
  <ul class="incdir">
  <li>
  <details open>
  <summary>posts</summary>
  <div>
  <article>
  This article is great!<footer>
  <a href="././posts/blog-post">Click here!</a>
  </footer>
  </article>
  </div>
  </details>
  </li>
  </ul>
  </body>
  </html>

  for blog-index.slw, and

  <!DOCTYPE html>
  <html lang="en">
  <head>
  <title></title>
  <meta charset="utf-8">
  <meta name="viewport" content="width=device-width, initial-scale=1">
  <meta name="generator" content="slweb">
  </head>
  <body>
  
  <p>This article is great! The quick brown fox jumps over the lazy sleeping dog.
  The quick brown fox jumps over the lazy sleeping dog. The quick brown fox jumps
  over the lazy sleeping dog. The quick brown fox jumps over the lazy sleeping
  dog. The quick brown fox jumps over the lazy sleeping dog. The quick brown fox
  jumps over the lazy sleeping dog.
  </body>
  </html>

  for posts/blog-post.slw.

• TSV/CSV templating. Directive

  {tsv "tsvfile" [iter]}{/tsv}

  marks a template. Whatever is between ‘{tsv}’ and ‘{/tsv}’ will be processed by slweb and collected as a template. Afterwards, file tsvfile.tsv will be read and for each of its lines (up to iter, if present) the collected template will be output, substituting each occurrence of a register mark ‘$n’ (where n is between 1 and 9, inclusive) with the corresponding TSV field of the read line.

  All of this applies to the similar ‘{csv}’ directive for CSV files. In the case of CSV files, they need to use semicolons (‘;’) or commas (‘,’) as delimiters (or set csv-delimiter as a YAML variable in the calling .slw file) and double quotation marks (‘"’) as field boundaries. First line in the TSV file is parsed as a header and associated with register marks ‘$#1’ to ‘$#9’. The symbol ‘$’ is represented as ‘$$’. Consequently, math modes (both inline and display) cannot be used within templates.

  Parts of the template can be conditionally rendered by using the construct:

  $?n
  <!-- code to include if $n nonempty -->
  $?!
  <!-- code to include if $n empty -->
  $?/

  where n is between 1 and 9, inclusive. For example, having a file sales.tsv:

  Item	Q1 2020	Q2 2020	Q3 2020	Q4 2020
  Toothpick	15.2	12		10
  Teapot	1.2		5	3.5

  We can write

  {tsv "sales"}
  {.sales-segment}
  $#1 {q}$1{/q} sales by quarter for the last year in
  thousands of units sold (for $$) were as follows:
  $?2{.q1}$2{/.q1}$?!{.q1 na}N/A{/.q1}$?/
  $?3{.q2}$3{/.q2}$?!{.q2 na}N/A{/.q2}$?/
  $?4{.q3}$4{/.q3}$?!{.q3 na}N/A{/.q3}$?/
  $?5{.q4}$5{/.q4}$?!{.q4 na}N/A{/.q4}$?/
  {/.sales-segment}
  {/tsv}

  to get:

  <div class="sales-segment">
  Item <q>Toothpick</q> sales by quarter for the last year in
  thousands of units sold (for $) were as follows:
  <div class="q1">15.2</div>
  <div class="q2">12</div>
  <div class="q3 na">N/A</div>
  <div class="q4">10</div>
  </div>
  
  <div class="sales-segment">
  Item <q>Teapot</q> sales by quarter for the last year in
  thousands of units sold (for $) were as follows:
  <div class="q1">1.2</div>
  <div class="q2 na">N/A</div>
  <div class="q3">5</div>
  <div class="q4">3.5</div>
  </div>

  TSV directives can't be nested and doing so will produce an error.

  The directives ‘{csv-count "csvfile"}’ and ‘{tsv-count "tsvfile"}’ output the number of “records” in the file csvfile.csv or tsvfile.tsv (number of lines minus one for the header row). No newline is output following the number.

Special YAML variables

• add-article-header. If set to “1”, title, date and header text at the beginning of the output body will be enclosed in a ‘<header>’ tag.
• add-figcaption. If set to “0”, ‘<figure>’ and ‘<figcaption>’ tags around images will not be added (otherwise they will by default, see Images).
• add-footnote-div. If set to “1”, list of footnotes preceded by a horizontal rule will be surrounded with a div having the id “footnotes” (see Footnotes).
• add-image-links. If set to “0”, links around images will not be added (otherwise they will by default, see Images).
• author. If set, the contents of this variable will be added to the start of the output body (inside a ‘<header>’ tag if it is set), surrounded by ‘<address></address>’.
• canonical. Contents of this variable will be added as a value of a ‘rel’ attribute for the ‘<link rel="canonical">’ tag.
• csv-delimiter. Changes the delimiter used for .csv file parsing. By default, this will be a semicolon  (‘;’). See TSV/CSV templating.
• date. If present, this variable, assumed to be in ISO 8601 UTC datetime format, will be parsed to determine the timestamp in permalinks generated by the incdir directive. The actual values used depend on the source-level constants article_timestamp_format (for incdir without the argument listonly) and list_timestamp_format (for incdir with the argument listonly).
• ext-in-permalink. If set to “0”, permalinks generated by the incdir directive will not have the extension included in the href attribute. For example, instead of blog/2020/august.html, the resulting permalink will have blog/2020/august.
• favicon-url. If present, this URL will be used as a favicon URL instead of the default, /favicon.ico.
• feed, feed-type, feed-desc. If all are present, contents of those variables will be added as href, type and title attributes (respectively) to the ‘<link rel="alternate">’ tag in the page's ‘<head>’.
• header-text. The contents of this variable will be added to the start of the output body (inside a ‘<header>’ tag if it is set), preceeded by ‘<p>’.
• image-file-prefix. If present, the value of src attribute will be appended to the contents of this variable, with path separator added as necessary, when determining the actual file path to pass to the identify(1) command. Otherwise, src is appended to the base directory, specified by the parameter -d (or --basedir), and defaulting to “.”.
• incdir-footer-permalink-text. If set, and the variable incdir-only-summary is set to “1”, when the file where this variable is set is included by the ‘{incdir}’ directive, ‘<footer>’ tag containing the permalink to the calling file will be appended to the text between the summary marks ‘/-’ and ‘-/’.
• incdir-only-summary. If set to “1”, when the file where this variable is set is included by the ‘{incdir}’ directive, only the text within the summary marks ‘/-’ and ‘-/’ will be output.
• inline-stylesheet. The contents of this variable will be treated as a CSS file name to be included inline in the header using the ‘<style>’ tag. You can add more than one inline-stylesheet declaration per .slw file.
• lang. Contents of this variable will be used for the lang attribute of the ‘<html>’ tag.
• link-prefix. This variable specifies the prefix which will be prepended to relative URLs in links. Final slash needs to be included if link-prefix is a complete path.
• meta. Contents of this variable is treated as a filename specifying the TSV file containing a list of values to be inserted into meta tags. The first row of this TSV file is treated as the header row and ignored. First column in this TSV file contains the meta variable names, and the second column contains meta variable values. For each row, a ‘<meta name="firstcol" content="secondcol">’ is inserted, where firstcol is the first column, and secondcol is the second column. If the values in the second column are surrounded with percent signs (‘%’), the value between them is treated as the name of the YAML variable to insert. For example, having the file m.tsv:

  Name	Content
  og:url	%canonical%

  and

  ---
  canonical: https://example.com/
  meta: m.tsv
  ---

  in the .slw file, a

  <meta name="og:url" content="https://example.com/">

  line will be added to the output.

• permalink-url. If present, this URL will completely replace the link used in the href attribute of permalinks generated by the incdir directive. This variable is more useful in individual .slw files inside the subdirectories of the dirname provided to the incdir directive.
• site-desc. The contents of this variable will be inserted as the value of the content attribute of the ‘<meta name="description">’ tag.
• site-name. The contents of this variable will be inserted inside the ‘<title></title>’ tags.
• stylesheet. The contents of this variable will be treated as a CSS file name to be included using the ‘<link rel="stylesheet">’ tag. You can add more than one stylesheet declaration per .slw file.
• title. If present, contents of this variable will be prepended to the body (inside a ‘<header>’ tag if it is set) as a heading with the level determined by the title-heading-level variable, defaulting to 2. (‘title: Some title’ becomes ‘<h2>Some title</h2>’).
• title-heading-level. See title.

Special macros

• permalink. If present, the body of this macro will be inserted into permalinks generated by the incdir directive, similar to the macro-form of links.

Known limitations

The limitation of Markdown is that there is no clear way to determine where the paragraph inside a tag should begin and end. Before the addition of Break marks and the change of behavior of slweb to not output ending ‘</p>’ tags, without adding blank lines or using the ‘{p}{/p}’ notation, ‘<p></p>’ tags in the output would be intertwined together with other tags. For example, the code

{tag}
First paragraph

Second paragraph
{/tag}

would produce

<tag>
First paragraph

<p>Second paragraph
</tag></p>

One could suppress the paragraph starting/ending tags by prepending blank lines with a backslash, like this:

{tag}
First paragraph
\
Second paragraph
{/tag}

which would produce

<tag>
First paragraph

Second paragraph
</tag>

Adding blank lines was another option if paragraphs were desired:

{tag}

First paragraph

Second paragraph

{/tag}

would produce

<tag>
<p>First paragraph</p>

<p>Second paragraph</p>
</tag>