From 725d6f699f589d9b5454e189d33292027532984d Mon Sep 17 00:00:00 2001
From: Hiroshi SHIBATA <hsbt@ruby-lang.org>
Date: Fri, 21 Feb 2025 15:53:31 +0900
Subject: [PATCH] Escape/unescape unclosed tags as well

Co-authored-by: Nobuyoshi Nakada <nobu@ruby-lang.org>
---
 lib/cgi/util.rb           |  4 ++--
 test/cgi/test_cgi_util.rb | 18 ++++++++++++++++++
 2 files changed, 20 insertions(+), 2 deletions(-)

diff --git a/lib/cgi/util.rb b/lib/cgi/util.rb
index aab8b000cb..5ff8ba58eb 100644
--- a/lib/cgi/util.rb
+++ b/lib/cgi/util.rb
@@ -140,7 +140,7 @@ def unescapeHTML(string)
   def escapeElement(string, *elements)
     elements = elements[0] if elements[0].kind_of?(Array)
     unless elements.empty?
-      string.gsub(/<\/?(?:#{elements.join("|")})(?!\w)(?:.|\n)*?>/i) do
+      string.gsub(/<\/?(?:#{elements.join("|")})\b[^<>]*+>?/im) do
         CGI.escapeHTML($&)
       end
     else
@@ -160,7 +160,7 @@ def escapeElement(string, *elements)
   def unescapeElement(string, *elements)
     elements = elements[0] if elements[0].kind_of?(Array)
     unless elements.empty?
-      string.gsub(/&lt;\/?(?:#{elements.join("|")})(?!\w)(?:.|\n)*?&gt;/i) do
+      string.gsub(/&lt;\/?(?:#{elements.join("|")})\b(?>[^&]+|&(?![gl]t;)\w+;)*(?:&gt;)?/im) do
         unescapeHTML($&)
       end
     else
diff --git a/test/cgi/test_cgi_util.rb b/test/cgi/test_cgi_util.rb
index b7bb7b8eae..e93be474b3 100644
--- a/test/cgi/test_cgi_util.rb
+++ b/test/cgi/test_cgi_util.rb
@@ -181,6 +181,14 @@ def test_cgi_escapeElement
     assert_equal("<BR>&lt;A HREF=&quot;url&quot;&gt;&lt;/A&gt;", escapeElement('<BR><A HREF="url"></A>', ["A", "IMG"]))
     assert_equal("<BR>&lt;A HREF=&quot;url&quot;&gt;&lt;/A&gt;", escape_element('<BR><A HREF="url"></A>', "A", "IMG"))
     assert_equal("<BR>&lt;A HREF=&quot;url&quot;&gt;&lt;/A&gt;", escape_element('<BR><A HREF="url"></A>', ["A", "IMG"]))
+
+    assert_equal("&lt;A &lt;A HREF=&quot;url&quot;&gt;&lt;/A&gt;", escapeElement('<A <A HREF="url"></A>', "A", "IMG"))
+    assert_equal("&lt;A &lt;A HREF=&quot;url&quot;&gt;&lt;/A&gt;", escapeElement('<A <A HREF="url"></A>', ["A", "IMG"]))
+    assert_equal("&lt;A &lt;A HREF=&quot;url&quot;&gt;&lt;/A&gt;", escape_element('<A <A HREF="url"></A>', "A", "IMG"))
+    assert_equal("&lt;A &lt;A HREF=&quot;url&quot;&gt;&lt;/A&gt;", escape_element('<A <A HREF="url"></A>', ["A", "IMG"]))
+
+    assert_equal("&lt;A &lt;A ", escapeElement('<A <A ', "A", "IMG"))
+    assert_equal("&lt;A &lt;A ", escapeElement('<A <A ', ["A", "IMG"]))
   end
 
 
@@ -189,5 +197,15 @@ def test_cgi_unescapeElement
     assert_equal('&lt;BR&gt;<A HREF="url"></A>', unescapeElement(escapeHTML('<BR><A HREF="url"></A>'), ["A", "IMG"]))
     assert_equal('&lt;BR&gt;<A HREF="url"></A>', unescape_element(escapeHTML('<BR><A HREF="url"></A>'), "A", "IMG"))
     assert_equal('&lt;BR&gt;<A HREF="url"></A>', unescape_element(escapeHTML('<BR><A HREF="url"></A>'), ["A", "IMG"]))
+
+    assert_equal('<A <A HREF="url"></A>', unescapeElement(escapeHTML('<A <A HREF="url"></A>'), "A", "IMG"))
+    assert_equal('<A <A HREF="url"></A>', unescapeElement(escapeHTML('<A <A HREF="url"></A>'), ["A", "IMG"]))
+    assert_equal('<A <A HREF="url"></A>', unescape_element(escapeHTML('<A <A HREF="url"></A>'), "A", "IMG"))
+    assert_equal('<A <A HREF="url"></A>', unescape_element(escapeHTML('<A <A HREF="url"></A>'), ["A", "IMG"]))
+
+    assert_equal('<A <A ', unescapeElement(escapeHTML('<A <A '), "A", "IMG"))
+    assert_equal('<A <A ', unescapeElement(escapeHTML('<A <A '), ["A", "IMG"]))
+    assert_equal('<A <A ', unescape_element(escapeHTML('<A <A '), "A", "IMG"))
+    assert_equal('<A <A ', unescape_element(escapeHTML('<A <A '), ["A", "IMG"]))
   end
 end