Merge pull request #359 from digitalmoksha/bw-multiline-blockquote

Add a multiline blockquote extension

README.md

@@ -97,7 +97,7 @@ Options:
           Multiple extensions can be delimited with ",", e.g. --extension strikethrough,table
           
           [possible values: strikethrough, tagfilter, table, autolink, tasklist, superscript,
-          footnotes, description-lists]
+          footnotes, description-lists, multiline-block-quotes]
 
   -t, --to <FORMAT>
           Specify output format

examples/s-expr.rs

@@ -83,6 +83,7 @@ fn dump(source: &str) -> io::Result<()> {
         .superscript(true)
         .footnotes(true)
         .description_lists(true)
+        .multiline_block_quotes(true)
         .build()
         .unwrap();
 

fuzz/fuzz_targets/all_options.rs

@@ -18,6 +18,7 @@ fuzz_target!(|s: &str| {
     extension.header_ids = Some("user-content-".to_string());
     extension.footnotes = true;
     extension.description_lists = true;
+    extension.multiline_block_quotes = true;
     extension.front_matter_delimiter = Some("---".to_string());
     extension.shortcodes = true;
 

fuzz/fuzz_targets/quadratic.rs

@@ -192,6 +192,7 @@ struct FuzzExtensionOptions {
     superscript: bool,
     footnotes: bool,
     description_lists: bool,
+    multiline_block_quotes: bool,
     shortcodes: bool,
 }
 
@@ -206,6 +207,7 @@ impl FuzzExtensionOptions {
         extension.superscript = self.superscript;
         extension.footnotes = self.footnotes;
         extension.description_lists = self.description_lists;
+        extension.multiline_block_quotes = self.multiline_block_quotes;
         extension.shortcodes = self.shortcodes;
         extension.front_matter_delimiter = None;
         extension.header_ids = None;

script/cibuild

@@ -34,6 +34,8 @@ if [ x"$SPEC" = "xtrue" ]; then
 	# python3 roundtrip_tests.py --spec extensions-table-prefer-style-attributes.txt "$PROGRAM_ARG --table-prefer-style-attributes" --extensions "table strikethrough autolink tagfilter footnotes tasklist" || failed=1
 	python3 roundtrip_tests.py --spec extensions-full-info-string.txt "$PROGRAM_ARG --full-info-string" \
 		|| failed=1
+	python3 spec_tests.py --no-normalize --spec ../../../src/tests/fixtures/multiline_blockquote.txt "$PROGRAM_ARG -e multiline-block-quotes" \
+		|| failed=1
 
 	python3 spec_tests.py --no-normalize --spec regression.txt "$PROGRAM_ARG" \
 		|| failed=1

src/cm.rs

@@ -377,6 +377,7 @@ impl<'a, 'o> CommonMarkFormatter<'a, 'o> {
             NodeValue::FootnoteReference(ref nfr) => {
                 self.format_footnote_reference(nfr.name.as_bytes(), entering)
             }
+            NodeValue::MultilineBlockQuote(..) => self.format_block_quote(entering),
         };
         true
     }

src/html.rs

@@ -993,6 +993,17 @@ impl<'o> HtmlFormatter<'o> {
                     self.output.write_all(b"</li>\n")?;
                 }
             }
+            NodeValue::MultilineBlockQuote(_) => {
+                if entering {
+                    self.cr()?;
+                    self.output.write_all(b"<blockquote")?;
+                    self.render_sourcepos(node)?;
+                    self.output.write_all(b">\n")?;
+                } else {
+                    self.cr()?;
+                    self.output.write_all(b"</blockquote>\n")?;
+                }
+            }
         }
         Ok(false)
     }

src/main.rs

@@ -146,6 +146,7 @@ enum Extension {
     Superscript,
     Footnotes,
     DescriptionLists,
+    MultilineBlockQuotes,
 }
 
 #[derive(Clone, Copy, Debug, ValueEnum)]
@@ -210,6 +211,7 @@ fn main() -> Result<(), Box<dyn Error>> {
         .header_ids(cli.header_ids)
         .footnotes(exts.contains(&Extension::Footnotes))
         .description_lists(exts.contains(&Extension::DescriptionLists))
+        .multiline_block_quotes(exts.contains(&Extension::MultilineBlockQuotes))
         .front_matter_delimiter(cli.front_matter_delimiter);
 
     #[cfg(feature = "shortcodes")]

src/nodes.rs

@@ -7,6 +7,8 @@ use std::convert::TryFrom;
 #[cfg(feature = "shortcodes")]
 use crate::parser::shortcodes::NodeShortCode;
 
+use crate::parser::multiline_block_quote::NodeMultilineBlockQuote;
+
 /// The core AST node enum.
 #[derive(Debug, Clone, PartialEq, Eq)]
 pub enum NodeValue {
@@ -151,6 +153,19 @@ pub enum NodeValue {
     #[cfg(feature = "shortcodes")]
     /// **Inline**. An Emoji character generated from a shortcode. Enable with feature "shortcodes".
     ShortCode(NodeShortCode),
+
+    /// **Block**. A [multiline block quote](https://github.github.com/gfm/#block-quotes).  Spans multiple
+    /// lines and contains other **blocks**.
+    ///
+    /// ``` md
+    /// >>>
+    /// A paragraph.
+    ///
+    /// - item one
+    /// - item two
+    /// >>>
+    /// ```
+    MultilineBlockQuote(NodeMultilineBlockQuote),
 }
 
 /// Alignment of a single table cell.
@@ -391,6 +406,7 @@ impl NodeValue {
                 | NodeValue::TableRow(..)
                 | NodeValue::TableCell
                 | NodeValue::TaskItem(..)
+                | NodeValue::MultilineBlockQuote(_)
         )
     }
 
@@ -464,6 +480,7 @@ impl NodeValue {
             NodeValue::FootnoteReference(..) => "footnote_reference",
             #[cfg(feature = "shortcodes")]
             NodeValue::ShortCode(_) => "shortcode",
+            NodeValue::MultilineBlockQuote(_) => "multiline_block_quote",
         }
     }
 }
@@ -647,6 +664,10 @@ pub fn can_contain_type<'a>(node: &'a AstNode<'a>, child: &NodeValue) -> bool {
                 | NodeValue::HtmlInline(..)
         ),
 
+        NodeValue::MultilineBlockQuote(_) => {
+            child.block() && !matches!(*child, NodeValue::Item(..) | NodeValue::TaskItem(..))
+        }
+
         _ => false,
     }
 }

src/parser/mod.rs

@@ -4,6 +4,8 @@ mod inlines;
 pub mod shortcodes;
 mod table;
 
+pub mod multiline_block_quote;
+
 use crate::adapters::SyntaxHighlighterAdapter;
 use crate::arena_tree::Node;
 use crate::ctype::{isdigit, isspace};
@@ -25,6 +27,7 @@ use std::str;
 use typed_arena::Arena;
 
 use crate::adapters::HeadingAdapter;
+use crate::parser::multiline_block_quote::NodeMultilineBlockQuote;
 
 use self::inlines::RefMap;
 
@@ -337,6 +340,31 @@ pub struct ExtensionOptions {
     /// ```
     pub front_matter_delimiter: Option<String>,
 
+    /// Enables the multiline block quote extension.
+    ///
+    /// Place `>>>` before and after text to make it into
+    /// a block quote.
+    ///
+    /// ``` md
+    /// Paragraph one
+    ///
+    /// >>>
+    /// Paragraph two
+    ///
+    /// - one
+    /// - two
+    /// >>>
+    /// ```
+    ///
+    /// ```
+    /// # use comrak::{markdown_to_html, Options};
+    /// let mut options = Options::default();
+    /// options.extension.multiline_block_quotes = true;
+    /// assert_eq!(markdown_to_html(">>>\nparagraph\n>>>", &options),
+    ///            "<blockquote>\n<p>paragraph</p>\n</blockquote>\n");
+    /// ```
+    pub multiline_block_quotes: bool,
+
     #[cfg(feature = "shortcodes")]
     #[cfg_attr(docsrs, doc(cfg(feature = "shortcodes")))]
     /// Phrases wrapped inside of ':' blocks will be replaced with emojis.
@@ -963,6 +991,16 @@ impl<'a, 'o, 'c> Parser<'a, 'o, 'c> {
                         return (false, container, should_continue);
                     }
                 }
+                NodeValue::MultilineBlockQuote(..) => {
+                    if !self.parse_multiline_block_quote_prefix(
+                        line,
+                        container,
+                        ast,
+                        &mut should_continue,
+                    ) {
+                        return (false, container, should_continue);
+                    }
+                }
                 _ => {}
             }
         }
@@ -985,7 +1023,26 @@ impl<'a, 'o, 'c> Parser<'a, 'o, 'c> {
             self.find_first_nonspace(line);
             let indented = self.indent >= CODE_INDENT;
 
-            if !indented && line[self.first_nonspace] == b'>' {
+            if !indented
+                && self.options.extension.multiline_block_quotes
+                && unwrap_into(
+                    scanners::open_multiline_block_quote_fence(&line[self.first_nonspace..]),
+                    &mut matched,
+                )
+            {
+                let first_nonspace = self.first_nonspace;
+                let offset = self.offset;
+                let nmbc = NodeMultilineBlockQuote {
+                    fence_length: matched,
+                    fence_offset: first_nonspace - offset,
+                };
+                *container = self.add_child(
+                    container,
+                    NodeValue::MultilineBlockQuote(nmbc),
+                    self.first_nonspace + 1,
+                );
+                self.advance_offset(line, first_nonspace + matched - offset, false);
+            } else if !indented && line[self.first_nonspace] == b'>' {
                 let blockquote_startpos = self.first_nonspace;
 
                 let offset = self.first_nonspace + 1 - self.offset;
@@ -1444,6 +1501,51 @@ impl<'a, 'o, 'c> Parser<'a, 'o, 'c> {
         }
     }
 
+    fn parse_multiline_block_quote_prefix(
+        &mut self,
+        line: &[u8],
+        container: &'a AstNode<'a>,
+        ast: &mut Ast,
+        should_continue: &mut bool,
+    ) -> bool {
+        let (fence_length, fence_offset) = match ast.value {
+            NodeValue::MultilineBlockQuote(ref node_value) => {
+                (node_value.fence_length, node_value.fence_offset)
+            }
+            _ => unreachable!(),
+        };
+
+        let matched = if self.indent <= 3 && line[self.first_nonspace] == b'>' {
+            scanners::close_multiline_block_quote_fence(&line[self.first_nonspace..]).unwrap_or(0)
+        } else {
+            0
+        };
+
+        if matched >= fence_length {
+            *should_continue = false;
+            self.advance_offset(line, matched, false);
+
+            // The last child, like an indented codeblock, could be left open.
+            // Make sure it's finalized.
+            if nodes::last_child_is_open(container) {
+                let child = container.last_child().unwrap();
+                let child_ast = &mut *child.data.borrow_mut();
+
+                self.finalize_borrowed(child, child_ast).unwrap();
+            }
+
+            self.current = self.finalize_borrowed(container, ast).unwrap();
+            return false;
+        }
+
+        let mut i = fence_offset;
+        while i > 0 && strings::is_space_or_tab(line[self.offset]) {
+            self.advance_offset(line, 1, true);
+            i -= 1;
+        }
+        true
+    }
+
     fn add_child(
         &mut self,
         mut parent: &'a AstNode<'a>,
@@ -1484,6 +1586,7 @@ impl<'a, 'o, 'c> Parser<'a, 'o, 'c> {
                     container.first_child().is_some()
                         || container.data.borrow().sourcepos.start.line != self.line_number
                 }
+                NodeValue::MultilineBlockQuote(..) => false,
                 _ => true,
             };
 
@@ -1664,6 +1767,7 @@ impl<'a, 'o, 'c> Parser<'a, 'o, 'c> {
             NodeValue::Document => true,
             NodeValue::CodeBlock(ref ncb) => ncb.fenced,
             NodeValue::Heading(ref nh) => nh.setext,
+            NodeValue::MultilineBlockQuote(..) => true,
             _ => false,
         } {
             ast.sourcepos.end = (self.line_number, self.curline_end_col).into();

src/parser/multiline_block_quote.rs

@@ -0,0 +1,9 @@
+/// The metadata of a multiline blockquote.
+#[derive(Debug, Clone, Copy, PartialEq, Eq)]
+pub struct NodeMultilineBlockQuote {
+    /// The length of the fence.
+    pub fence_length: usize,
+
+    /// The indentation level of the fence marker.
+    pub fence_offset: usize,
+}

src/scanners.re

@@ -376,6 +376,28 @@ pub fn shortcode(s: &[u8]) -> Option<usize> {
 */
 }
 
+pub fn open_multiline_block_quote_fence(s: &[u8]) -> Option<usize> {
+    let mut cursor = 0;
+    let mut marker = 0;
+    let mut ctxmarker = 0;
+    let len = s.len();
+/*!re2c
+    [>]{3,} / [ \t]*[\r\n] { return Some(cursor); }
+    * { return None; }
+*/
+}
+
+pub fn close_multiline_block_quote_fence(s: &[u8]) -> Option<usize> {
+    let mut cursor = 0;
+    let mut marker = 0;
+    let mut ctxmarker = 0;
+    let len = s.len();
+/*!re2c
+    [>]{3,} / [ \t]*[\r\n] { return Some(cursor); }
+    * { return None; }
+*/
+}
+
 // Returns both the length of the match, and the tasklist character.
 pub fn tasklist(s: &[u8]) -> Option<(usize, u8)> {
     let mut cursor = 0;