From f21a47f2454ec6d0798a92f1780c040fc4e8d242 Mon Sep 17 00:00:00 2001
From: Will Kahn-Greene <willkg@mozilla.com>
Date: Fri, 6 Oct 2023 13:09:14 -0400
Subject: [PATCH] Fix entities inside a tags in linkification (#704)

---
 bleach/linkifier.py   |  2 +-
 tests/test_linkify.py | 10 ++++++++++
 2 files changed, 11 insertions(+), 1 deletion(-)

diff --git a/bleach/linkifier.py b/bleach/linkifier.py
index 679d7ead..54a432e1 100644
--- a/bleach/linkifier.py
+++ b/bleach/linkifier.py
@@ -591,7 +591,7 @@ def __iter__(self):
                     in_a = False
                     token_buffer = []
                 else:
-                    token_buffer.append(token)
+                    token_buffer.extend(list(self.extract_entities(token)))
                 continue
 
             if token["type"] in ["StartTag", "EmptyTag"]:
diff --git a/tests/test_linkify.py b/tests/test_linkify.py
index bd838133..4cc01a6f 100644
--- a/tests/test_linkify.py
+++ b/tests/test_linkify.py
@@ -323,6 +323,7 @@ def test_link_fragment():
 
 
 def test_link_entities_in_qs():
+    """Entities in the querystring get escaped"""
     assert (
         linkify("http://xx.com/?a=1&b=2")
         == '<a href="http://xx.com/?a=1&amp;b=2" rel="nofollow">http://xx.com/?a=1&amp;b=2</a>'
@@ -330,9 +331,18 @@ def test_link_entities_in_qs():
 
 
 def test_link_entities_in_characters_token():
+    """Entitites in a Characters token don't get escaped"""
     assert linkify("foo &nbsp; bar") == "foo &nbsp; bar"
 
 
+def test_link_entities_in_a_tag():
+    """Entitites between an a start tag and an a end tag don't get escaped"""
+    assert (
+        linkify('<a href="/">Some&nbsp;entity&rsquo;s</a>', callbacks=[])
+        == '<a href="/">Some&nbsp;entity&rsquo;s</a>'
+    )
+
+
 def test_escaped_html():
     """If I pass in escaped HTML, it should probably come out escaped."""
     s = "&lt;em&gt;strong&lt;/em&gt;"