fix(markdown): stop escaped link destinations from swallowing later content (#1630)

packages/markdown_parser/rust/src/links.rs

@@ -412,10 +412,7 @@ fn extract_escaped_url(text: &str, start: usize) -> Option<UrlInfo<'_>> {
         if byte_at(text, pos) == b'>' {
             let url = &text[start..pos];
             pos += 1;
-            while pos < text.len() && byte_at(text, pos) != b')' {
-                pos += advance_one(text, pos);
-            }
-            if pos >= text.len() {
+            if pos >= text.len() || byte_at(text, pos) != b')' {
                 return None;
             }
             return Some(UrlInfo {
@@ -1410,7 +1407,34 @@ pub fn max_inline_scan() -> usize {
 
 #[cfg(test)]
 mod tests {
-    use super::has_valid_code_fence_language;
+    use super::{extract_escaped_url, has_valid_code_fence_language};
+
+    #[test]
+    fn escaped_url_accepts_paren_immediately_after_destination() {
+        let info = extract_escaped_url("https://example.com>) trailing", 0).expect("valid");
+        assert_eq!(info.url, "https://example.com");
+        assert!(info.escaped);
+        assert_eq!(info.advance_by, "https://example.com>)".len());
+    }
+
+    #[test]
+    fn escaped_url_rejects_content_between_destination_and_paren() {
+        assert!(extract_escaped_url("https://example.com> and more)", 0).is_none());
+        assert!(extract_escaped_url("https://example.com> )", 0).is_none());
+        assert!(extract_escaped_url("https://example.com>text)", 0).is_none());
+    }
+
+    #[test]
+    fn escaped_url_rejects_missing_paren() {
+        assert!(extract_escaped_url("https://example.com>", 0).is_none());
+        assert!(extract_escaped_url("https://example.com", 0).is_none());
+    }
+
+    #[test]
+    fn escaped_url_keeps_paren_inside_destination() {
+        let info = extract_escaped_url("https://example.com/a.yaml)>)", 0).expect("valid");
+        assert_eq!(info.url, "https://example.com/a.yaml)");
+    }
 
     #[test]
     fn rejects_leading_whitespace() {

packages/markdown_parser/rust/tests/fixtures.rs

@@ -259,6 +259,43 @@ fn native_parser_rejects_apostrophe_in_masked_link_authority() {
     );
 }
 
+#[test]
+fn native_parser_keeps_content_after_unterminated_escaped_destination() {
+    let flags = ParserFlags::ALLOW_MASKED_LINKS | ParserFlags::ALLOW_AUTOLINKS;
+    assert_eq!(
+        parse(
+            "a [x](<https://example.com/y)> b [e](<https://example.com/f>)",
+            flags,
+            ""
+        ),
+        json!({"nodes":[
+            {"type":"Text","content":"a [x]("},
+            {"type":"Link","url":"https://example.com/y)","escaped":true,"rawUrl":"https://example.com/y)","source":"<https://example.com/y)>"},
+            {"type":"Text","content":" b "},
+            {"type":"Link","text":{"type":"Text","content":"e"},"url":"https://example.com/f","escaped":true,"rawUrl":"https://example.com/f","source":"[e](<https://example.com/f>)"}
+        ]})
+    );
+    assert_eq!(
+        parse("[bad](<https://example.com/a> )", flags, ""),
+        json!({"nodes":[
+            {"type":"Text","content":"[bad]("},
+            {"type":"Link","url":"https://example.com/a","escaped":true,"rawUrl":"https://example.com/a","source":"<https://example.com/a>"},
+            {"type":"Text","content":" )"}
+        ]})
+    );
+}
+
+#[test]
+fn native_parser_still_accepts_well_formed_escaped_destinations() {
+    let flags = ParserFlags::ALLOW_MASKED_LINKS | ParserFlags::ALLOW_AUTOLINKS;
+    assert_eq!(
+        parse("[ok](<https://example.com/a>)", flags, ""),
+        json!({"nodes":[
+            {"type":"Link","text":{"type":"Text","content":"ok"},"url":"https://example.com/a","escaped":true,"rawUrl":"https://example.com/a","source":"[ok](<https://example.com/a>)"}
+        ]})
+    );
+}
+
 #[test]
 fn native_parser_rejects_masked_links_without_visible_text() {
     for input in [