[ruby/cgi] Escape/unescape unclosed tags as well

ruby/cgi@cd1eb08076

Co-authored-by: Nobuyoshi Nakada <nobu@ruby-lang.org>

Author: 2 people

lib/cgi/util.rb

@@ -184,7 +184,7 @@ 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($&)
       end
     else
@@ -204,7 +204,7 @@ def escapeElement(string, *elements)
   def unescapeElement(string, *elements)
     elements = elements[0] if elements[0].kind_of?(Array)
     unless elements.empty?
-      string.gsub(/&lt;\/?(?:#{elements.join("|")})(?!\w)(?:.|\n)*?&gt;/i) do
+      string.gsub(/&lt;\/?(?:#{elements.join("|")})\b(?>[^&]+|&(?![gl]t;)\w+;)*(?:&gt;)?/im) do
         unescapeHTML($&)
       end
     else

test/cgi/test_cgi_util.rb

@@ -269,6 +269,14 @@ 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
 
 
@@ -277,6 +285,16 @@ 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