From 4754d407aa26be2b608c59a073313b4aeda98683 Mon Sep 17 00:00:00 2001
From: "Edward Z. Yang" <ezyang@mit.edu>
Date: Thu, 28 Oct 2010 17:24:07 +0100
Subject: [PATCH] Fix removal of id with DirectLex by preserving armor.

Signed-off-by: Edward Z. Yang <ezyang@mit.edu>
---
 NEWS                                             |  2 ++
 library/HTMLPurifier/Strategy/MakeWellFormed.php |  4 ++--
 library/HTMLPurifier/Token/Tag.php               |  3 ++-
 tests/HTMLPurifier/HTMLT/id-img.htmlt            |  8 ++++++++
 tests/HTMLPurifier/LexerTest.php                 | 26 ++++++++++++++++++++++++
 5 files changed, 40 insertions(+), 3 deletions(-)
 create mode 100644 tests/HTMLPurifier/HTMLT/id-img.htmlt

diff --git a/NEWS b/NEWS
index 247c6799..7a2d4236 100644
--- a/NEWS
+++ b/NEWS
@@ -15,6 +15,8 @@ NEWS ( CHANGELOG and HISTORY )                                     HTMLPurifier
 - Make removal of conditional IE comments ungreedy; thanks Bernd
   for reporting.
 - Escape CDATA before removing Internet Explorer comments.
+- Fix removal of id attributes under certain conditions by ensuring
+  armor attributes are preserved when recreating tags.
 
 4.2.0, released 2010-09-15
 ! Added %Core.RemoveProcessingInstructions, which lets you remove
diff --git a/library/HTMLPurifier/Strategy/MakeWellFormed.php b/library/HTMLPurifier/Strategy/MakeWellFormed.php
index c7365840..d3f01578 100644
--- a/library/HTMLPurifier/Strategy/MakeWellFormed.php
+++ b/library/HTMLPurifier/Strategy/MakeWellFormed.php
@@ -193,12 +193,12 @@ class HTMLPurifier_Strategy_MakeWellFormed extends HTMLPurifier_Strategy
             $ok = false;
             if ($type === 'empty' && $token instanceof HTMLPurifier_Token_Start) {
                 // claims to be a start tag but is empty
-                $token = new HTMLPurifier_Token_Empty($token->name, $token->attr);
+                $token = new HTMLPurifier_Token_Empty($token->name, $token->attr, $token->line, $token->col, $token->armor);
                 $ok = true;
             } elseif ($type && $type !== 'empty' && $token instanceof HTMLPurifier_Token_Empty) {
                 // claims to be empty but really is a start tag
                 $this->swap(new HTMLPurifier_Token_End($token->name));
-                $this->insertBefore(new HTMLPurifier_Token_Start($token->name, $token->attr));
+                $this->insertBefore(new HTMLPurifier_Token_Start($token->name, $token->attr, $token->line, $token->col, $token->armor));
                 // punt (since we had to modify the input stream in a non-trivial way)
                 $reprocess = true;
                 continue;
diff --git a/library/HTMLPurifier/Token/Tag.php b/library/HTMLPurifier/Token/Tag.php
index 798be028..f4d8f640 100644
--- a/library/HTMLPurifier/Token/Tag.php
+++ b/library/HTMLPurifier/Token/Tag.php
@@ -33,7 +33,7 @@ class HTMLPurifier_Token_Tag extends HTMLPurifier_Token
      * @param $name String name.
      * @param $attr Associative array of attributes.
      */
-    public function __construct($name, $attr = array(), $line = null, $col = null) {
+    public function __construct($name, $attr = array(), $line = null, $col = null, $armor = array()) {
         $this->name = ctype_lower($name) ? $name : strtolower($name);
         foreach ($attr as $key => $value) {
             // normalization only necessary when key is not lowercase
@@ -50,6 +50,7 @@ class HTMLPurifier_Token_Tag extends HTMLPurifier_Token
         $this->attr = $attr;
         $this->line = $line;
         $this->col  = $col;
+        $this->armor = $armor;
     }
 }
 
diff --git a/tests/HTMLPurifier/HTMLT/id-img.htmlt b/tests/HTMLPurifier/HTMLT/id-img.htmlt
new file mode 100644
index 00000000..12755bea
--- /dev/null
+++ b/tests/HTMLPurifier/HTMLT/id-img.htmlt
@@ -0,0 +1,8 @@
+--INI--
+Attr.EnableID = true
+Core.LexerImpl = DirectLex
+--HTML--
+<img src="img_11775.jpg" alt="[Img #11775]" id="EMBEDDED_IMG_11775" >
+--EXPECT--
+<img src="img_11775.jpg" alt="[Img #11775]" id="EMBEDDED_IMG_11775" />
+--# vim: et sw=4 sts=4
diff --git a/tests/HTMLPurifier/LexerTest.php b/tests/HTMLPurifier/LexerTest.php
index abe1eb87..f2e63ecc 100644
--- a/tests/HTMLPurifier/LexerTest.php
+++ b/tests/HTMLPurifier/LexerTest.php
@@ -754,6 +754,32 @@ div {}
         );
     }
 
+    function test_tokenizeHTML_imgTag() {
+        $this->assertTokenization(
+            '<img src="img_11775.jpg" alt="[Img #11775]" id="EMBEDDED_IMG_11775" >',
+            array(
+                new HTMLPurifier_Token_Empty('img',
+                    array(
+                        'src' => 'img_11775.jpg',
+                        'alt' => '[Img #11775]',
+                        'id' => 'EMBEDDED_IMG_11775',
+                    )
+                )
+            ),
+            array(
+                'DirectLex' => array(
+                        new HTMLPurifier_Token_Start('img',
+                            array(
+                                'src' => 'img_11775.jpg',
+                                'alt' => '[Img #11775]',
+                                'id' => 'EMBEDDED_IMG_11775',
+                            )
+                        )
+                    ),
+                )
+        );
+    }
+
 
     /*
 
-- 
2.11.4.GIT