fix: keep emphasis in the text of a rejected reference link

src/Lexer.ts

@@ -300,6 +300,40 @@ export class _Lexer<ParserOutput = string, RendererOutput = string> {
     return tokens;
   }
 
+  /**
+   * Does this link text hold a link already? An image does not count: an image
+   * may hold a link, a link may not.
+   */
+  private linkInText(text: string, links: string[]): boolean {
+    if (!text.includes('[')) {
+      return false;
+    }
+
+    const linkRule = this.tokenizer.rules.inline.link;
+    for (const match of text.matchAll(this.tokenizer.rules.inline.blockSkip)) {
+      // blockSkip also matches code spans and html, and the `!` of an image is
+      // left out of the match, so read the character before it.
+      if (linkRule.test(match[0]) && text.charAt(match.index - 1) !== '!') {
+        return true;
+      }
+    }
+
+    for (const match of text.matchAll(this.tokenizer.rules.inline.reflinkSearch)) {
+      const match0 = match[0];
+      const refStart = match0.lastIndexOf('[');
+      if (match0.charAt(0) === '!' || !links.includes(match0.slice(refStart + 1, -1))) {
+        continue;
+      }
+      // a candidate holding a link is not a link either, so it does not count
+      if (refStart > 1 && this.linkInText(match0.slice(1, refStart - 1), links)) {
+        continue;
+      }
+      return true;
+    }
+
+    return false;
+  }
+
   /**
    * Lexing/Compiling
    */
@@ -312,10 +346,27 @@ export class _Lexer<ParserOutput = string, RendererOutput = string> {
     if (this.tokens.links) {
       const links = Object.keys(this.tokens.links);
       if (links.length > 0) {
-        maskedSrc = maskedSrc.replace(this.tokenizer.rules.inline.reflinkSearch, match0 =>
-          links.includes(match0.slice(match0.lastIndexOf('[') + 1, -1))
-            ? '[' + 'a'.repeat(match0.length - 2) + ']'
-            : match0);
+        const reflinkSearch = this.tokenizer.rules.inline.reflinkSearch;
+        const maskReflink = (match0: string): string => {
+          const refStart = match0.lastIndexOf('[');
+          if (!links.includes(match0.slice(refStart + 1, -1))) {
+            return match0;
+          }
+          // CommonMark: "Links may not contain other links, at any level of
+          // nesting." A candidate whose text already holds one never becomes a
+          // link, so flattening the whole span would hide the emphasis that
+          // does still apply inside it. Mask the links it holds instead.
+          // Images are exempt: their text is flattened into an alt attribute.
+          if (refStart > 1 && match0.charAt(0) !== '!') {
+            const text = match0.slice(1, refStart - 1);
+            if (this.linkInText(text, links)) {
+              return '[' + text.replace(reflinkSearch, maskReflink)
+                + '][' + 'a'.repeat(match0.length - refStart - 2) + ']';
+            }
+          }
+          return '[' + 'a'.repeat(match0.length - 2) + ']';
+        };
+        maskedSrc = maskedSrc.replace(reflinkSearch, maskReflink);
       }
     }
 

test/specs/commonmark/commonmark.0.31.2.json

@@ -4269,8 +4269,7 @@
     "example": 533,
     "start_line": 8053,
     "end_line": 8059,
-    "section": "Links",
-    "shouldFail": true
+    "section": "Links"
   },
   {
     "markdown": "*[foo*][ref]\n\n[ref]: /uri\n",

test/specs/gfm/commonmark.0.31.2.json

@@ -4269,8 +4269,7 @@
     "example": 533,
     "start_line": 8053,
     "end_line": 8059,
-    "section": "Links",
-    "shouldFail": true
+    "section": "Links"
   },
   {
     "markdown": "*[foo*][ref]\n\n[ref]: /uri\n",

test/specs/new/emphasis_in_link_text_with_link.html

@@ -0,0 +1,4 @@
+<p>[foo <em>bar <a href="/url">baz</a> qux</em>]<a href="/uri">ref</a></p>
+<p>[foo <em>bar <a href="/uri">baz</a> qux</em>]<a href="/uri">ref</a></p>
+<p><a href="/uri">foo <em>bar <img src="/img.png" alt="baz"> qux</em></a></p>
+<p>*foo <a href="/uri">bar*</a></p>

test/specs/new/emphasis_in_link_text_with_link.md

@@ -0,0 +1,9 @@
+[foo *bar [baz](/url) qux*][ref]
+
+[foo _bar [baz][ref] qux_][ref]
+
+[foo *bar ![baz](/img.png) qux*][ref]
+
+*foo [bar*][ref]
+
+[ref]: /uri