Merge pull request #407 from kivikakk/bw-wikilinks

Add support for wikilinks format

examples/s-expr.rs

@@ -86,6 +86,8 @@ fn dump(source: &str) -> io::Result<()> {
         .multiline_block_quotes(true)
         .math_dollars(true)
         .math_code(true)
+        .wikilinks_title_after_pipe(true)
+        .wikilinks_title_before_pipe(true)
         .build()
         .unwrap();
 

fuzz/fuzz_targets/all_options.rs

@@ -23,7 +23,9 @@ fuzz_target!(|s: &str| {
     extension.math_code = true;
     extension.front_matter_delimiter = Some("---".to_string());
     extension.shortcodes = true;
-
+    extension.wikilinks_title_after_pipe = true;
+    extension.wikilinks_title_before_pipe = true;
+
     let mut parse = ParseOptions::default();
     parse.smart = true;
     parse.default_info_string = Some("rust".to_string());

fuzz/fuzz_targets/quadratic.rs

@@ -196,6 +196,8 @@ struct FuzzExtensionOptions {
     math_dollars: bool,
     math_code: bool,
     shortcodes: bool,
+    wikilinks_title_after_pipe: bool,
+    wikilinks_title_before_pipe: bool,
 }
 
 impl FuzzExtensionOptions {
@@ -213,6 +215,8 @@ impl FuzzExtensionOptions {
         extension.math_dollars = self.math_dollars;
         extension.math_code = self.math_code;
         extension.shortcodes = self.shortcodes;
+        extension.wikilinks_title_after_pipe = self.wikilinks_title_after_pipe;
+        extension.wikilinks_title_before_pipe = self.wikilinks_title_before_pipe;
         extension.front_matter_delimiter = None;
         extension.header_ids = None;
         extension

script/cibuild

@@ -40,6 +40,10 @@ python3 spec_tests.py --no-normalize --spec ../../../src/tests/fixtures/math_dol
   || failed=1
 python3 spec_tests.py --no-normalize --spec ../../../src/tests/fixtures/math_code.md "$PROGRAM_ARG -e math-code" \
   || failed=1
+python3 spec_tests.py --no-normalize --spec ../../../src/tests/fixtures/wikilinks_title_after_pipe.md "$PROGRAM_ARG -e wikilinks-title-after-pipe" \
+  || failed=1
+python3 spec_tests.py --no-normalize --spec ../../../src/tests/fixtures/wikilinks_title_before_pipe.md "$PROGRAM_ARG -e wikilinks-title-before-pipe" \
+  || failed=1
 
 python3 spec_tests.py --no-normalize --spec regression.txt "$PROGRAM_ARG" \
 	|| failed=1

src/cm.rs

@@ -2,7 +2,7 @@ use crate::ctype::{isalpha, isdigit, ispunct, isspace};
 use crate::nodes::TableAlignment;
 use crate::nodes::{
     AstNode, ListDelimType, ListType, NodeCodeBlock, NodeHeading, NodeHtmlBlock, NodeLink,
-    NodeMath, NodeTable, NodeValue,
+    NodeMath, NodeTable, NodeValue, NodeWikiLink,
 };
 #[cfg(feature = "shortcodes")]
 use crate::parser::shortcodes::NodeShortCode;
@@ -385,6 +385,7 @@ impl<'a, 'o> CommonMarkFormatter<'a, 'o> {
                 // noop - automatic escaping is already being done
             }
             NodeValue::Math(ref math) => self.format_math(math, allow_wrap, entering),
+            NodeValue::WikiLink(ref nl) => return self.format_wikilink(nl, entering),
         };
         true
     }
@@ -689,6 +690,24 @@ impl<'a, 'o> CommonMarkFormatter<'a, 'o> {
         true
     }
 
+    fn format_wikilink(&mut self, nl: &NodeWikiLink, entering: bool) -> bool {
+        if entering {
+            write!(self, "[[").unwrap();
+            if self.options.extension.wikilinks_title_after_pipe {
+                self.output(nl.url.as_bytes(), false, Escaping::Url);
+                write!(self, "|").unwrap();
+            }
+        } else {
+            if self.options.extension.wikilinks_title_before_pipe {
+                write!(self, "|").unwrap();
+                self.output(nl.url.as_bytes(), false, Escaping::Url);
+            }
+            write!(self, "]]").unwrap();
+        }
+
+        true
+    }
+
     fn format_image(&mut self, nl: &NodeLink, allow_wrap: bool, entering: bool) {
         if entering {
             write!(self, "![").unwrap();

src/html.rs

@@ -1038,6 +1038,21 @@ impl<'o> HtmlFormatter<'o> {
                     self.render_math_inline(node, literal, display_math, dollar_math)?;
                 }
             }
+            NodeValue::WikiLink(ref nl) => {
+                if entering {
+                    self.output.write_all(b"<a")?;
+                    self.render_sourcepos(node)?;
+                    self.output.write_all(b" href=\"")?;
+                    let url = nl.url.as_bytes();
+                    if self.options.render.unsafe_ || !dangerous_url(url) {
+                        self.escape_href(url)?;
+                    }
+                    self.output.write_all(b"\" data-wikilink=\"true")?;
+                    self.output.write_all(b"\">")?;
+                } else {
+                    self.output.write_all(b"</a>")?;
+                }
+            }
         }
         Ok(false)
     }

src/main.rs

@@ -159,6 +159,8 @@ enum Extension {
     MultilineBlockQuotes,
     MathDollars,
     MathCode,
+    WikilinksTitleAfterPipe,
+    WikilinksTitleBeforePipe,
 }
 
 #[derive(Clone, Copy, Debug, ValueEnum)]
@@ -238,6 +240,8 @@ fn main() -> Result<(), Box<dyn Error>> {
         .multiline_block_quotes(exts.contains(&Extension::MultilineBlockQuotes))
         .math_dollars(exts.contains(&Extension::MathDollars))
         .math_code(exts.contains(&Extension::MathCode))
+        .wikilinks_title_after_pipe(exts.contains(&Extension::WikilinksTitleAfterPipe))
+        .wikilinks_title_before_pipe(exts.contains(&Extension::WikilinksTitleBeforePipe))
         .front_matter_delimiter(cli.front_matter_delimiter);
 
     #[cfg(feature = "shortcodes")]

src/nodes.rs

@@ -182,6 +182,9 @@ pub enum NodeValue {
 
     /// **Inline**.  A character that has been [escaped](https://github.github.com/gfm/#backslash-escapes)
     Escaped,
+
+    /// **Inline**.  A wikilink to some URL.
+    WikiLink(NodeWikiLink),
 }
 
 /// Alignment of a single table cell.
@@ -253,6 +256,13 @@ pub struct NodeLink {
     pub title: String,
 }
 
+/// The details of a wikilink's destination.
+#[derive(Debug, Clone, PartialEq, Eq)]
+pub struct NodeWikiLink {
+    /// The URL for the link destination.
+    pub url: String,
+}
+
 /// The metadata of a list; the kind of list, the delimiter used and so on.
 #[derive(Debug, Default, Clone, Copy, PartialEq, Eq)]
 pub struct NodeList {
@@ -489,6 +499,7 @@ impl NodeValue {
             NodeValue::MultilineBlockQuote(_) => "multiline_block_quote",
             NodeValue::Escaped => "escaped",
             NodeValue::Math(..) => "math",
+            NodeValue::WikiLink(..) => "wikilink",
         }
     }
 }
@@ -639,7 +650,8 @@ pub fn can_contain_type<'a>(node: &'a AstNode<'a>, child: &NodeValue) -> bool {
         | NodeValue::Emph
         | NodeValue::Strong
         | NodeValue::Link(..)
-        | NodeValue::Image(..) => !child.block(),
+        | NodeValue::Image(..)
+        | NodeValue::WikiLink(..) => !child.block(),
 
         NodeValue::Table(..) => matches!(*child, NodeValue::TableRow(..)),
 
@@ -657,6 +669,7 @@ pub fn can_contain_type<'a>(node: &'a AstNode<'a>, child: &NodeValue) -> bool {
                 | NodeValue::Strikethrough
                 | NodeValue::HtmlInline(..)
                 | NodeValue::Math(..)
+                | NodeValue::WikiLink(..)
         ),
 
         #[cfg(feature = "shortcodes")]
@@ -672,6 +685,7 @@ pub fn can_contain_type<'a>(node: &'a AstNode<'a>, child: &NodeValue) -> bool {
                 | NodeValue::Strikethrough
                 | NodeValue::HtmlInline(..)
                 | NodeValue::Math(..)
+                | NodeValue::WikiLink(..)
         ),
 
         NodeValue::MultilineBlockQuote(_) => {

src/parser/inlines.rs

@@ -2,7 +2,8 @@ use crate::arena_tree::Node;
 use crate::ctype::{isdigit, ispunct, isspace};
 use crate::entity;
 use crate::nodes::{
-    Ast, AstNode, NodeCode, NodeFootnoteReference, NodeLink, NodeMath, NodeValue, Sourcepos,
+    Ast, AstNode, NodeCode, NodeFootnoteReference, NodeLink, NodeMath, NodeValue, NodeWikiLink,
+    Sourcepos,
 };
 #[cfg(feature = "shortcodes")]
 use crate::parser::shortcodes::NodeShortCode;
@@ -105,6 +106,12 @@ struct Bracket<'a> {
     bracket_after: bool,
 }
 
+#[derive(Clone, Copy)]
+struct WikilinkComponents<'i> {
+    url: &'i [u8],
+    link_label: Option<(&'i [u8], usize, usize)>,
+}
+
 impl<'a, 'r, 'o, 'd, 'i, 'c, 'subj> Subject<'a, 'r, 'o, 'd, 'i, 'c, 'subj> {
     pub fn new(
         arena: &'a Arena<AstNode<'a>>,
@@ -183,11 +190,30 @@ impl<'a, 'r, 'o, 'd, 'i, 'c, 'subj> Subject<'a, 'r, 'o, 'd, 'i, 'c, 'subj> {
             '.' => Some(self.handle_period()),
             '[' => {
                 self.pos += 1;
-                let inl =
-                    self.make_inline(NodeValue::Text("[".to_string()), self.pos - 1, self.pos - 1);
-                self.push_bracket(false, inl);
-                self.within_brackets = true;
-                Some(inl)
+
+                let mut wikilink_inl = None;
+
+                if (self.options.extension.wikilinks_title_after_pipe
+                    || self.options.extension.wikilinks_title_before_pipe)
+                    && !self.within_brackets
+                    && self.peek_char() == Some(&(b'['))
+                {
+                    wikilink_inl = self.handle_wikilink();
+                }
+
+                if wikilink_inl.is_none() {
+                    let inl = self.make_inline(
+                        NodeValue::Text("[".to_string()),
+                        self.pos - 1,
+                        self.pos - 1,
+                    );
+                    self.push_bracket(false, inl);
+                    self.within_brackets = true;
+
+                    Some(inl)
+                } else {
+                    wikilink_inl
+                }
             }
             ']' => {
                 self.within_brackets = false;
@@ -1548,6 +1574,127 @@ impl<'a, 'r, 'o, 'd, 'i, 'c, 'subj> Subject<'a, 'r, 'o, 'd, 'i, 'c, 'subj> {
         }
     }
 
+    // Handles wikilink syntax
+    //   [[link text|url]]
+    //   [[url|link text]]
+    pub fn handle_wikilink(&mut self) -> Option<&'a AstNode<'a>> {
+        let startpos = self.pos;
+        let component = self.wikilink_url_link_label()?;
+        let url_clean = strings::clean_url(component.url);
+        let (link_label, link_label_start_column, link_label_end_column) =
+            match component.link_label {
+                Some((label, sc, ec)) => (entity::unescape_html(label), sc, ec),
+                None => (
+                    entity::unescape_html(component.url),
+                    startpos + 1,
+                    self.pos - 3,
+                ),
+            };
+
+        let nl = NodeWikiLink {
+            url: String::from_utf8(url_clean).unwrap(),
+        };
+        let inl = self.make_inline(NodeValue::WikiLink(nl), startpos - 1, self.pos - 1);
+        inl.append(self.make_inline(
+            NodeValue::Text(String::from_utf8(link_label).unwrap()),
+            link_label_start_column,
+            link_label_end_column,
+        ));
+
+        Some(inl)
+    }
+
+    fn wikilink_url_link_label(&mut self) -> Option<WikilinkComponents<'i>> {
+        let left_startpos = self.pos;
+
+        if self.peek_char() != Some(&(b'[')) {
+            return None;
+        }
+
+        let found_left = self.wikilink_component();
+
+        if !found_left {
+            self.pos = left_startpos;
+            return None;
+        }
+
+        let left = strings::trim_slice(&self.input[left_startpos + 1..self.pos]);
+
+        if self.peek_char() == Some(&(b']')) && self.peek_char_n(1) == Some(&(b']')) {
+            self.pos += 2;
+            return Some(WikilinkComponents {
+                url: left,
+                link_label: None,
+            });
+        } else if self.peek_char() != Some(&(b'|')) {
+            self.pos = left_startpos;
+            return None;
+        }
+
+        let right_startpos = self.pos;
+        let found_right = self.wikilink_component();
+
+        if !found_right {
+            self.pos = left_startpos;
+            return None;
+        }
+
+        let right = strings::trim_slice(&self.input[right_startpos + 1..self.pos]);
+
+        if self.peek_char() == Some(&(b']')) && self.peek_char_n(1) == Some(&(b']')) {
+            self.pos += 2;
+
+            if self.options.extension.wikilinks_title_after_pipe {
+                Some(WikilinkComponents {
+                    url: left,
+                    link_label: Some((right, right_startpos + 1, self.pos - 3)),
+                })
+            } else {
+                Some(WikilinkComponents {
+                    url: right,
+                    link_label: Some((left, left_startpos + 1, right_startpos - 1)),
+                })
+            }
+        } else {
+            self.pos = left_startpos;
+            None
+        }
+    }
+
+    // Locates the edge of a wikilink component (link label or url), and sets the
+    // self.pos to it's end if it's found.
+    fn wikilink_component(&mut self) -> bool {
+        let startpos = self.pos;
+
+        if self.peek_char() != Some(&(b'[')) && self.peek_char() != Some(&(b'|')) {
+            return false;
+        }
+
+        self.pos += 1;
+
+        let mut length = 0;
+        let mut c = 0;
+        while unwrap_into_copy(self.peek_char(), &mut c) && c != b'[' && c != b']' && c != b'|' {
+            if c == b'\\' {
+                self.pos += 1;
+                length += 1;
+                if self.peek_char().map_or(false, |&c| ispunct(c)) {
+                    self.pos += 1;
+                    length += 1;
+                }
+            } else {
+                self.pos += 1;
+                length += 1;
+            }
+            if length > MAX_LINK_LABEL_LENGTH {
+                self.pos = startpos;
+                return false;
+            }
+        }
+
+        true
+    }
+
     pub fn spnl(&mut self) {
         self.skip_spaces();
         if self.skip_line_end() {