shortcodes: capture all known aliases.

src/cm.rs

@@ -752,7 +752,7 @@ impl<'a, 'o> CommonMarkFormatter<'a, 'o> {
  fn format_shortcode(&mut self, ne: &NodeShortCode, entering: bool) {
  if entering {
  write!(self, ":").unwrap();
- self.output(ne.shortcode().as_bytes(), false, Escaping::Literal);
+ self.output(ne.code.as_bytes(), false, Escaping::Literal);
  write!(self, ":").unwrap();
  }
  }

src/html.rs

@@ -853,7 +853,7 @@ impl<'o> HtmlFormatter<'o> {
  #[cfg(feature = "shortcodes")]
  NodeValue::ShortCode(ref nsc) => {
  if entering {
- self.output.write_all(nsc.emoji().as_bytes())?;
+ self.output.write_all(nsc.emoji.as_bytes())?;
  }
  }
  NodeValue::Table(..) => {

src/parser/inlines.rs

@@ -1227,7 +1227,7 @@ impl<'a, 'r, 'o, 'd, 'i, 'c, 'subj> Subject<'a, 'r, 'o, 'd, 'i, 'c, 'subj> {
  str::from_utf8_unchecked(&self.input[self.pos + 1..self.pos + 1 + matchlen - 1])
  };
 
- let nsc = NodeShortCode::try_from(shortcode).ok()?;
+ let nsc = NodeShortCode::resolve(shortcode)?;
  self.pos += 1 + matchlen;
 
  Some(self.make_inline(

src/parser/shortcodes.rs

@@ -1,37 +1,22 @@
-use std::{convert::TryFrom, str};
-
-/// The details of an inline emoji.
+/// The details of an inline "shortcode" emoji/gemoji.
+///
+/// ("gemoji" name context: https://github.com/github/gemoji)
 #[derive(Debug, Clone, PartialEq, Eq)]
-pub struct NodeShortCode(
- /// A short code that is translated into an emoji
- String,
-);
-
-impl NodeShortCode {
- /// Checks whether the input is a valid short code.
- pub fn is_valid(value: &str) -> bool {
- emojis::get_by_shortcode(value).is_some()
- }
+pub struct NodeShortCode {
+ /// The shortcode that was resolved, e.g. "rabbit".
+ pub code: String,
 
- /// Get the underlying shortcode.
- pub fn shortcode(&self) -> &str {
- &self.0
- }
-
- /// Get the emoji for this short code.
- pub fn emoji(&self) -> &'static str {
- emojis::get_by_shortcode(&self.0).unwrap().as_str()
- }
+ /// The emoji `code` resolved to, e.g. "🐰".
+ pub emoji: String,
 }
 
-impl TryFrom<&str> for NodeShortCode {
- type Error = ();
-
- fn try_from(value: &str) -> Result<Self, ()> {
- if Self::is_valid(value) {
- Ok(Self(value.into()))
- } else {
- Err(())
- }
+impl NodeShortCode {
+ /// Checks whether the input is a valid short code.
+ pub fn resolve(code: &str) -> Option<Self> {
+ let emoji = emojis::get_by_shortcode(code)?;
+ Some(NodeShortCode {
+ code: code.to_string(),
+ emoji: emoji.to_string(),
+ })
  }
 }

src/scanners.re

@@ -381,7 +381,7 @@ pub fn shortcode(s: &[u8]) -> Option<usize> {
  let mut marker = 0;
  let len = s.len();
 /*!re2c
- [A-Za-z_-]+ [:] { return Some(cursor); }
+ [A-Za-z0-9+_-]+ [:] { return Some(cursor); }
  * { return None; }
 */
 }

src/tests/api.rs

@@ -215,8 +215,9 @@ fn exercise_full_api() {
  let _: String = nl.title;
  }
  #[cfg(feature = "shortcodes")]
- nodes::NodeValue::ShortCode(ne) => {
- let _: &str = ne.shortcode();
+ nodes::NodeValue::ShortCode(nsc) => {
+ let _: String = nsc.code;
+ let _: String = nsc.emoji;
  }
  nodes::NodeValue::FootnoteReference(nfr) => {
  let _: String = nfr.name;

src/tests/shortcodes.rs

@@ -32,3 +32,14 @@ fn emojis() {
  concat!("<p>Hello, happy days! :diego:</p>\n"),
  );
 }
+
+#[test]
+fn emojis_specials() {
+ // Take a quick trip to https://raw.githubusercontent.com/github/gemoji/master/db/emoji.json
+ // with `jq -r .[].aliases[] | sort | grep -E '[^a-z_-]'` to see what else there is to see.
+ html_opts!(
+ [extension.shortcodes],
+ ":+1: :-1: :clock12::1234: :1st_place_medal: :e-mail: :non-potable_water:",
+ "<p>👍 👎 🕛🔢 🥇 📧 🚱</p>\n",
+ );
+}

src/xml.rs

@@ -256,7 +256,7 @@ impl<'o> XmlFormatter<'o> {
  #[cfg(feature = "shortcodes")]
  NodeValue::ShortCode(ref nsc) => {
  self.output.write_all(b" id=\"")?;
- self.escape(nsc.shortcode().as_bytes())?;
+ self.escape(nsc.code.as_bytes())?;
  self.output.write_all(b"\"")?;
  }
  NodeValue::Escaped => {