Merge pull request #648 from Kuuuube/subtext

Add support for Discord style subtext

Author: kivikakk

examples/s-expr.rs

@@ -85,6 +85,7 @@ fn dump(source: &str) -> io::Result<()> {
         .math_code(true)
         .wikilinks_title_after_pipe(true)
         .wikilinks_title_before_pipe(true)
+        .subtext(true)
         .build();
 
     let opts = Options {

fuzz/fuzz_targets/all_options.rs

@@ -33,6 +33,7 @@ fuzz_target!(|s: &str| {
         image_url_rewriter: Some(Arc::new(url_rewriter)),
         link_url_rewriter: Some(Arc::new(url_rewriter)),
         cjk_friendly_emphasis: true,
+        subtext: true,
     };
 
     let cb = |link_ref: options::BrokenLinkReference| {

src/cm.rs

@@ -420,6 +420,7 @@ impl<'a, 'o, 'c> CommonMarkFormatter<'a, 'o, 'c> {
             NodeValue::SpoileredText => self.format_spoiler()?,
             NodeValue::EscapedTag(ref net) => self.format_escaped_tag(net)?,
             NodeValue::Alert(ref alert) => self.format_alert(alert, entering)?,
+            NodeValue::Subtext => self.format_subtext(entering)?,
         };
         Ok(true)
     }
@@ -998,6 +999,18 @@ impl<'a, 'o, 'c> CommonMarkFormatter<'a, 'o, 'c> {
         }
         Ok(())
     }
+
+    fn format_subtext(&mut self, entering: bool) -> fmt::Result {
+        if entering {
+            write!(self, "-# ")?;
+            self.begin_content = true;
+            self.no_linebreaks = true;
+        } else {
+            self.no_linebreaks = false;
+            self.blankline();
+        }
+        Ok(())
+    }
 }
 
 fn longest_byte_sequence(buffer: &[u8], ch: u8) -> usize {

src/html.rs

@@ -464,6 +464,7 @@ pub fn format_node_default<'a, T>(
         NodeValue::Superscript => render_superscript(context, node, entering),
         NodeValue::Underline => render_underline(context, node, entering),
         NodeValue::WikiLink(ref nwl) => render_wiki_link(context, node, entering, nwl),
+        NodeValue::Subtext => render_subtext(context, node, entering),
     }
 }
 
@@ -1443,6 +1444,23 @@ fn render_subscript<'a, T>(
     Ok(ChildRendering::HTML)
 }
 
+fn render_subtext<'a, T>(
+    context: &mut Context<T>,
+    node: Node<'a>,
+    entering: bool,
+) -> Result<ChildRendering, fmt::Error> {
+    if entering {
+        context.cr()?;
+        context.write_str("<p><sub")?;
+        render_sourcepos(context, node)?;
+        context.write_str(">")?;
+    } else {
+        writeln!(context, "</sub></p>")?;
+    }
+
+    Ok(ChildRendering::HTML)
+}
+
 fn render_superscript<'a, T>(
     context: &mut Context<T>,
     node: Node<'a>,

src/main.rs

@@ -195,6 +195,7 @@ enum Extension {
     Greentext,
     Alerts,
     CjkFriendlyEmphasis,
+    Subtext,
 }
 
 #[derive(Clone, Copy, Debug, ValueEnum)]
@@ -282,7 +283,8 @@ fn main() -> Result<(), Box<dyn Error>> {
         .greentext(exts.contains(&Extension::Greentext))
         .alerts(exts.contains(&Extension::Alerts))
         .maybe_front_matter_delimiter(cli.front_matter_delimiter)
-        .cjk_friendly_emphasis(exts.contains(&Extension::CjkFriendlyEmphasis));
+        .cjk_friendly_emphasis(exts.contains(&Extension::CjkFriendlyEmphasis))
+        .subtext(exts.contains(&Extension::Subtext));
 
     #[cfg(feature = "shortcodes")]
     let extension = extension.shortcodes(cli.gemoji);

src/nodes.rs

@@ -225,6 +225,10 @@ pub enum NodeValue {
     /// **Block**. GitHub style alert boxes which uses a modified blockquote syntax.
     /// Enabled with the `alerts` option.
     Alert(Box<NodeAlert>),
+
+    /// **Block**. Block scoped subscript that acts similar to a header.
+    /// Enabled with `subtext` option.
+    Subtext,
 }
 
 /// Alignment of a single table cell.
@@ -561,14 +565,18 @@ impl NodeValue {
                 | NodeValue::TaskItem(..)
                 | NodeValue::MultilineBlockQuote(_)
                 | NodeValue::Alert(_)
+                | NodeValue::Subtext
         )
     }
 
     /// Whether the type the node is of can contain inline nodes.
     pub fn contains_inlines(&self) -> bool {
         matches!(
             *self,
-            NodeValue::Paragraph | NodeValue::Heading(..) | NodeValue::TableCell
+            NodeValue::Paragraph
+                | NodeValue::Heading(..)
+                | NodeValue::TableCell
+                | NodeValue::Subtext
         )
     }
 
@@ -595,7 +603,10 @@ impl NodeValue {
     pub(crate) fn accepts_lines(&self) -> bool {
         matches!(
             *self,
-            NodeValue::Paragraph | NodeValue::Heading(..) | NodeValue::CodeBlock(..)
+            NodeValue::Paragraph
+                | NodeValue::Heading(..)
+                | NodeValue::CodeBlock(..)
+                | NodeValue::Subtext
         )
     }
 
@@ -644,6 +655,7 @@ impl NodeValue {
             NodeValue::SpoileredText => "spoiler",
             NodeValue::EscapedTag(_) => "escaped_tag",
             NodeValue::Alert(_) => "alert",
+            NodeValue::Subtext => "subtext",
         }
     }
 }
@@ -888,6 +900,7 @@ impl<'a> arena_tree::Node<'a, RefCell<Ast>> {
             | NodeValue::SpoileredText
             | NodeValue::Underline
             | NodeValue::Subscript
+            | NodeValue::Subtext
             // XXX: this is quite a hack: the EscapedTag _contains_ whatever was
             // possibly going to fall into the spoiler. This should be fixed in
             // inlines.

src/parser/mod.rs

@@ -364,7 +364,10 @@ where
                         break;
                     }
                 }
-                NodeValue::Heading(..) | NodeValue::TableRow(..) | NodeValue::TableCell => {
+                NodeValue::Heading(..)
+                | NodeValue::TableRow(..)
+                | NodeValue::TableCell
+                | NodeValue::Subtext => {
                     break;
                 }
                 NodeValue::FootnoteDefinition(..) => {
@@ -613,6 +616,7 @@ where
                     || self.handle_multiline_blockquote(container, line)
                     || self.handle_blockquote(container, line)
                     || self.handle_atx_heading(container, line)
+                    || self.handle_atx_subtext(container, line)
                     || self.handle_code_fence(container, line)
                     || self.handle_html_block(container, line)
                     || self.handle_setext_heading(container, line)
@@ -782,6 +786,26 @@ where
         scanners::atx_heading_start(&line[self.first_nonspace..])
     }
 
+    fn handle_atx_subtext(&mut self, container: &mut Node<'a>, line: &str) -> bool {
+        let Some(matched) = self.detect_atx_subtext(line) else {
+            return false;
+        };
+
+        let heading_startpos = self.first_nonspace;
+        let offset = self.offset;
+        self.advance_offset(line, heading_startpos + matched - offset, false);
+        *container = self.add_child(container, NodeValue::Subtext, heading_startpos + 1);
+
+        let container_ast = &mut container.data_mut();
+        container_ast.value = NodeValue::Subtext;
+
+        true
+    }
+
+    fn detect_atx_subtext(&self, line: &str) -> Option<usize> {
+        scanners::atx_subtext_start(&line[self.first_nonspace..])
+    }
+
     fn handle_code_fence(&mut self, container: &mut Node<'a>, line: &str) -> bool {
         let Some(matched) = self.detect_code_fence(line) else {
             return false;
@@ -1328,7 +1352,10 @@ where
 
         container.data_mut().last_line_blank = self.blank
             && match container.data().value {
-                NodeValue::BlockQuote | NodeValue::Heading(..) | NodeValue::ThematicBreak => false,
+                NodeValue::BlockQuote
+                | NodeValue::Heading(..)
+                | NodeValue::ThematicBreak
+                | NodeValue::Subtext => false,
                 NodeValue::CodeBlock(ref ncb) => !ncb.fenced,
                 NodeValue::Item(..) => {
                     container.first_child().is_some()

src/parser/options.rs

@@ -509,6 +509,22 @@ pub struct Extension<'c> {
     /// ```
     #[cfg_attr(feature = "bon", builder(default))]
     pub cjk_friendly_emphasis: bool,
+
+    /// Enables block scoped subscript that acts similar to a header.
+    ///
+    /// ```md
+    /// -# subtext
+    /// ```
+    ///
+    /// ```rust
+    /// # use comrak::{markdown_to_html, Options};
+    /// let mut options = Options::default();
+    /// options.extension.subscript = true;
+    ///
+    /// assert_eq!(markdown_to_html("-# subtext", &options),
+    ///           "<p><sub>subtext</sub></p>\n");
+    /// ```
+    pub subtext: bool,
 }
 
 impl<'c> Extension<'c> {

src/scanners.re

@@ -70,6 +70,16 @@ pub fn atx_heading_start(s: &str) -> Option<usize> {
 */
 }
 
+pub fn atx_subtext_start(s: &str) -> Option<usize> {
+    let mut cursor = 0;
+    let mut marker = 0;
+    let len = s.len();
+/*!re2c
+    [-][#] ([ \t]+|[\r\n])  { return Some(cursor); }
+    * { return None; }
+*/
+}
+
 pub fn html_block_end_1(s: &str) -> bool {
     let mut cursor = 0;
     let mut marker = 0;