Output ‘data-wikilink=“true”’ for wikilinks

This is so it’s possible to tell the difference between
a normal link and a wikilink during post-processing.

Author: digitalmoksha

src/cm.rs

@@ -367,7 +367,13 @@ impl<'a, 'o> CommonMarkFormatter<'a, 'o> {
             NodeValue::TaskItem(symbol) => self.format_task_item(symbol, node, entering),
             NodeValue::Strikethrough => self.format_strikethrough(),
             NodeValue::Superscript => self.format_superscript(),
-            NodeValue::Link(ref nl) => return self.format_link(node, nl, entering),
+            NodeValue::Link(ref nl) => {
+                if nl.wikilink {
+                    return self.format_wikilink(nl, entering);
+                } else {
+                    return self.format_link(node, nl, entering);
+                }
+            }
             NodeValue::Image(ref nl) => self.format_image(nl, allow_wrap, entering),
             #[cfg(feature = "shortcodes")]
             NodeValue::ShortCode(ref ne) => self.format_shortcode(ne, entering),
@@ -689,6 +695,24 @@ impl<'a, 'o> CommonMarkFormatter<'a, 'o> {
         true
     }
 
+    fn format_wikilink(&mut self, nl: &NodeLink, 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

@@ -826,6 +826,9 @@ impl<'o> HtmlFormatter<'o> {
                         self.output.write_all(b"\" title=\"")?;
                         self.escape(nl.title.as_bytes())?;
                     }
+                    if nl.wikilink {
+                        self.output.write_all(b"\" data-wikilink=\"true")?;
+                    }
                     self.output.write_all(b"\">")?;
                 } else {
                     self.output.write_all(b"</a>")?;

src/nodes.rs

@@ -251,6 +251,9 @@ pub struct NodeLink {
     /// Note this field is used for the `title` attribute by the HTML formatter even for images;
     /// `alt` text is supplied in the image inline text.
     pub title: String,
+
+    /// Whether this is a wikilink or not
+    pub wikilink: bool,
 }
 
 /// The metadata of a list; the kind of list, the delimiter used and so on.

src/parser/autolink.rs

@@ -122,6 +122,7 @@ fn www_match<'a>(
         NodeValue::Link(NodeLink {
             url,
             title: String::new(),
+            wikilink: false,
         }),
         (0, 1, 0, 1).into(),
     );
@@ -290,6 +291,7 @@ fn url_match<'a>(
         NodeValue::Link(NodeLink {
             url: url.clone(),
             title: String::new(),
+            wikilink: false,
         }),
         (0, 1, 0, 1).into(),
     );
@@ -398,6 +400,7 @@ fn email_match<'a>(
         NodeValue::Link(NodeLink {
             url,
             title: String::new(),
+            wikilink: false,
         }),
         (0, 1, 0, 1).into(),
     );

src/parser/inlines.rs

@@ -1491,7 +1491,11 @@ impl<'a, 'r, 'o, 'd, 'i, 'c, 'subj> Subject<'a, 'r, 'o, 'd, 'i, 'c, 'subj> {
     pub fn close_bracket_match(&mut self, is_image: bool, url: String, title: String) {
         let brackets_len = self.brackets.len();
 
-        let nl = NodeLink { url, title };
+        let nl = NodeLink {
+            url,
+            title,
+            wikilink: false,
+        };
         let inl = self.make_inline(
             if is_image {
                 NodeValue::Image(nl)
@@ -1585,6 +1589,7 @@ impl<'a, 'r, 'o, 'd, 'i, 'c, 'subj> Subject<'a, 'r, 'o, 'd, 'i, 'c, 'subj> {
         let nl = NodeLink {
             url: String::from_utf8(url_clean).unwrap(),
             title: String::new(),
+            wikilink: true,
         };
         let inl = self.make_inline(NodeValue::Link(nl), startpos - 1, self.pos - 1);
         inl.append(self.make_inline(
@@ -1724,6 +1729,7 @@ impl<'a, 'r, 'o, 'd, 'i, 'c, 'subj> Subject<'a, 'r, 'o, 'd, 'i, 'c, 'subj> {
             NodeValue::Link(NodeLink {
                 url: String::from_utf8(strings::clean_autolink(url, kind)).unwrap(),
                 title: String::new(),
+                wikilink: false,
             }),
             start_column + 1,
             end_column + 1,

src/parser/mod.rs

@@ -435,7 +435,7 @@ pub struct ExtensionOptions {
     /// let mut options = Options::default();
     /// options.extension.wikilinks_title_after_pipe = true;
     /// assert_eq!(markdown_to_html("[[url|link text]]", &options),
-    ///            "<p><a href=\"url\">link text</a></p>\n");
+    ///            "<p><a href=\"url\" data-wikilink=\"true\">link text</a></p>\n");
     /// ```
     pub wikilinks_title_after_pipe: bool,
 
@@ -450,7 +450,7 @@ pub struct ExtensionOptions {
     /// let mut options = Options::default();
     /// options.extension.wikilinks_title_before_pipe = true;
     /// assert_eq!(markdown_to_html("[[link text|url]]", &options),
-    ///            "<p><a href=\"url\">link text</a></p>\n");
+    ///            "<p><a href=\"url\" data-wikilink=\"true\">link text</a></p>\n");
     /// ```
     pub wikilinks_title_before_pipe: bool,
 }

src/tests/wikilinks.rs

@@ -1,17 +1,26 @@
 use super::*;
 
+#[test]
+fn wikilinks_escaped_chars_in_link_text() {
+    html_opts!(
+        [extension.wikilinks_title_before_pipe],
+        concat!("[[a \\<link|url]]",),
+        concat!("<p><a href=\"url\" data-wikilink=\"true\">a &lt;link</a></p>\n"),
+    );
+}
+
 #[test]
 fn wikilinks_does_not_unescape_html_entities_in_link_text() {
     html_opts!(
-        [extension.wikilinks_title_after_pipe],
+        [extension.wikilinks_title_after_pipe, render.escaped_char_spans],
         concat!("This is [[&lt;script&gt;alert(0)&lt;/script&gt;|a &lt;link]]",),
-        concat!("<p>This is <a href=\"%3Cscript%3Ealert(0)%3C/script%3E\">a &lt;link</a></p>\n"),
+        concat!("<p>This is <a href=\"%3Cscript%3Ealert(0)%3C/script%3E\" data-wikilink=\"true\">a &lt;link</a></p>\n"),
     );
 
     html_opts!(
         [extension.wikilinks_title_before_pipe],
         concat!("This is [[a &lt;link|&lt;script&gt;alert(0)&lt;/script&gt;]]",),
-        concat!("<p>This is <a href=\"%3Cscript%3Ealert(0)%3C/script%3E\">a &lt;link</a></p>\n"),
+        concat!("<p>This is <a href=\"%3Cscript%3Ealert(0)%3C/script%3E\" data-wikilink=\"true\">a &lt;link</a></p>\n"),
     );
 }
 
@@ -20,13 +29,13 @@ fn wikilinks_sanitizes_the_href_attribute_case_1() {
     html_opts!(
         [extension.wikilinks_title_after_pipe],
         concat!("[[http:\'\"injected=attribute&gt;&lt;img/src=\"0\"onerror=\"alert(0)\"&gt;https://example.com|a]]",),
-        concat!("<p><a href=\"http:&#x27;%22injected=attribute%3E%3Cimg/src=%220%22onerror=%22alert(0)%22%3Ehttps://example.com\">a</a></p>\n"),
+        concat!("<p><a href=\"http:&#x27;%22injected=attribute%3E%3Cimg/src=%220%22onerror=%22alert(0)%22%3Ehttps://example.com\" data-wikilink=\"true\">a</a></p>\n"),
     );
 
     html_opts!(
         [extension.wikilinks_title_before_pipe],
         concat!("[[a|http:\'\"injected=attribute&gt;&lt;img/src=\"0\"onerror=\"alert(0)\"&gt;https://example.com]]",),
-        concat!("<p><a href=\"http:&#x27;%22injected=attribute%3E%3Cimg/src=%220%22onerror=%22alert(0)%22%3Ehttps://example.com\">a</a></p>\n"),
+        concat!("<p><a href=\"http:&#x27;%22injected=attribute%3E%3Cimg/src=%220%22onerror=%22alert(0)%22%3Ehttps://example.com\" data-wikilink=\"true\">a</a></p>\n"),
     );
 }
 
@@ -35,13 +44,13 @@ fn wikilinks_sanitizes_the_href_attribute_case_2() {
     html_opts!(
         [extension.wikilinks_title_after_pipe],
         concat!("<i>[[\'\"&gt;&lt;svg&gt;&lt;i/class=gl-show-field-errors&gt;&lt;input/title=\"&lt;script&gt;alert(0)&lt;/script&gt;\"/&gt;&lt;/svg&gt;https://example.com|a]]",),
-        concat!("<p><!-- raw HTML omitted --><a href=\"&#x27;%22%3E%3Csvg%3E%3Ci/class=gl-show-field-errors%3E%3Cinput/title=%22%3Cscript%3Ealert(0)%3C/script%3E%22/%3E%3C/svg%3Ehttps://example.com\">a</a></p>\n"),
+        concat!("<p><!-- raw HTML omitted --><a href=\"&#x27;%22%3E%3Csvg%3E%3Ci/class=gl-show-field-errors%3E%3Cinput/title=%22%3Cscript%3Ealert(0)%3C/script%3E%22/%3E%3C/svg%3Ehttps://example.com\" data-wikilink=\"true\">a</a></p>\n"),
     );
 
     html_opts!(
         [extension.wikilinks_title_before_pipe],
         concat!("<i>[[a|\'\"&gt;&lt;svg&gt;&lt;i/class=gl-show-field-errors&gt;&lt;input/title=\"&lt;script&gt;alert(0)&lt;/script&gt;\"/&gt;&lt;/svg&gt;https://example.com]]",),
-        concat!("<p><!-- raw HTML omitted --><a href=\"&#x27;%22%3E%3Csvg%3E%3Ci/class=gl-show-field-errors%3E%3Cinput/title=%22%3Cscript%3Ealert(0)%3C/script%3E%22/%3E%3C/svg%3Ehttps://example.com\">a</a></p>\n"),
+        concat!("<p><!-- raw HTML omitted --><a href=\"&#x27;%22%3E%3Csvg%3E%3Ci/class=gl-show-field-errors%3E%3Cinput/title=%22%3Cscript%3Ealert(0)%3C/script%3E%22/%3E%3C/svg%3Ehttps://example.com\" data-wikilink=\"true\">a</a></p>\n"),
     );
 }