This README is the one indexed from the repository at its latest ref, not from the tag for this version.

Shortcodes

What it is

This is a parser for the shortcode spec as explained in the Hugo docs and used in Hugo and Nikola. Approximately.

I am implementing this for Nicolino.

It probably won't be 100% identical, but I'll try to make it as close as practical.

• Implemented in Ragel + C for performance
• The C parser allocates no memory: all positions and lengths are references to pieces of the input, and nothing is copied or heap-allocated during parsing.
• The Crystal binding allocates only the result objects: one copy of each reported string, and nothing else.

What works

• Detect shortcodes with names
• Standalone and matched shortcodes
• Capture data between tags in matched shortcodes
• Self-closing shortcodes
• Inline shortcodes
• Markdown and Verbatim shortcodes
• Capture arguments with and without names
• Capture values with and without single/double quotes, including escaped quotes
• Unquoted values support letters, numbers and unambiguous punctuation: - _ . / : @ + ~ # (e.g. src=images/photo-1.png). Anything else, or values with spaces, can be quoted. A trailing / in an unquoted value is ambiguous with the self-closing delimiters, so quote it (path="dir/").
• UTF-8 strings AFAICT (arg and shortcode names only support ascii tho)

Things you should know

• Names and unquoted values are ASCII-only. UTF-8 works fine in the text around shortcodes, in matched content and in quoted values.
• Text that looks like a shortcode but isn't one (e.g. {{% foo >}}) passes through untouched rather than producing an error, because it may be content, not a broken shortcode. Only unambiguous errors (closing a shortcode that was never opened, mismatched bracket styles in a complete shortcode) are reported.
• Only the first 10 errors are reported. Fix them in order, the rest will surface on the next pass.
• More than 100 shortcodes per document or 100 arguments per shortcode is an error; anything past the limit is dropped.
• In a Shortcode, position counts codepoints, while len and the sizes of whole and data are in bytes.

Nested shortcodes should be handled in a higher layer by recursive calls to the parser where it makes sense.

Building

You need Ragel 7 and a C compiler. The generated shortcodes.c in the repo is built with ragel 7; older versions generate different (buggy) code, so don't regenerate with 6.x.

Ragel is used to generate shortcodes.c out of shortcodes.rl. As a convenience there is a generated shortcodes.c in the repo,

Then:

  make

To run tests:

  crystal spec

Installation

1. Add the dependency to your shard.yml:

   dependencies:
     shortcodes:
       github: ralsina/shortcodes

2. Run shards install

Usage in Crystal

require "shortcodes"

Usage in C

Just copy shortcodes.c and shortcodes.h into your project and have fun.

sc_result result;
parse(input, len, &result);

The caller provides the sc_result: it is ~160KB, and returning it by value used to make optimizing compilers explode compile times dramatically. See docs/compile-time-pitfall.md for the full story.