Changeset 46894 for trunk

Timestamp:

12/12/2019 05:51:35 PM (5 years ago)

Author:

whyisjake

Message:

Prevent stored XSS through wp_targeted_link_rel().

Props: vortfu, whyisjake, peterwilsoncc, xknown, SergeyBiryukov, flaviozavan.

Location:

trunk

Files:

• src/wp-includes/formatting.php (modified) (3 diffs)
• tests/phpunit/tests/formatting/WPTargetedLinkRel.php (modified) (4 diffs)

trunk/src/wp-includes/formatting.php

@@ -3149 +3149 @@
 function wp_targeted_link_rel( $text ) {
     // Don't run (more expensive) regex if no links with targets.
-    if ( stripos( $text, 'target' ) !== false && stripos( $text, '<a ' ) !== false ) {
-        if ( ! is_serialized( $text ) ) {
-            $text = preg_replace_callback( '|<a\s([^>]*target\s*=[^>]*)>|i', 'wp_targeted_link_rel_callback', $text );
+    if ( stripos( $text, 'target' ) === false || stripos( $text, '<a ' ) === false || is_serialized( $text ) ) {
+        return $text;
+    }
+
+    $script_and_style_regex = '/<(script|style).*?<\/\\1>/si';
+
+    preg_match_all( $script_and_style_regex, $text, $matches );
+    $extra_parts = $matches[0];
+    $html_parts  = preg_split( $script_and_style_regex, $text );
+
+    foreach ( $html_parts as &$part ) {
+        $part = preg_replace_callback( '|<a\s([^>]*target\s*=[^>]*)>|i', 'wp_targeted_link_rel_callback', $part );
+    }
+
+    $text = '';
+    for ( $i = 0; $i < count( $html_parts ); $i++ ) {
+        $text .= $html_parts[ $i ];
+        if ( isset( $extra_parts[ $i ] ) ) {
+            $text .= $extra_parts[ $i ];
         }
     }
@@ -3170 +3186 @@
  */
 function wp_targeted_link_rel_callback( $matches ) {
-    $link_html = $matches[1];
-    $rel_match = array();
+    $link_html          = $matches[1];
+    $original_link_html = $link_html;
+
+    // Consider the html escaped if there are no unescaped quotes
+    $is_escaped = ! preg_match( '/(^|[^\\\\])[\'"]/', $link_html );
+    if ( $is_escaped ) {
+        // Replace only the quotes so that they are parsable by wp_kses_hair, leave the rest as is
+        $link_html = preg_replace( '/\\\\([\'"])/', '$1', $link_html );
+    }
+
+    $atts = wp_kses_hair( $link_html, wp_allowed_protocols() );
 
     /**
@@ -3183 +3208 @@
     $rel = apply_filters( 'wp_targeted_link_rel', 'noopener noreferrer', $link_html );
 
-    // Avoid additional regex if the filter removes rel values.
-    if ( ! $rel ) {
-        return "<a $link_html>";
-    }
-
-    // Value with delimiters, spaces around are optional.
-    $attr_regex = '|rel\s*=\s*?(\\\\{0,1}["\'])(.*?)\\1|i';
-    preg_match( $attr_regex, $link_html, $rel_match );
-
-    if ( empty( $rel_match[0] ) ) {
-        // No delimiters, try with a single value and spaces, because `rel =  va"lue` is totally fine...
-        $attr_regex = '|rel\s*=(\s*)([^\s]*)|i';
-        preg_match( $attr_regex, $link_html, $rel_match );
-    }
-
-    if ( ! empty( $rel_match[0] ) ) {
-        $parts     = preg_split( '|\s+|', strtolower( $rel_match[2] ) );
-        $parts     = array_map( 'esc_attr', $parts );
-        $needed    = explode( ' ', $rel );
-        $parts     = array_unique( array_merge( $parts, $needed ) );
-        $delimiter = trim( $rel_match[1] ) ? $rel_match[1] : '"';
-        $rel       = 'rel=' . $delimiter . trim( implode( ' ', $parts ) ) . $delimiter;
-        $link_html = str_replace( $rel_match[0], $rel, $link_html );
-    } elseif ( preg_match( '|target\s*=\s*?\\\\"|', $link_html ) ) {
-        $link_html .= " rel=\\\"$rel\\\"";
-    } elseif ( preg_match( '#(target|href)\s*=\s*?\'#', $link_html ) ) {
-        $link_html .= " rel='$rel'";
-    } else {
-        $link_html .= " rel=\"$rel\"";
+    // Return early if no rel values to be added or if no actual target attribute
+    if ( ! $rel || ! isset( $atts['target'] ) ) {
+        return "<a $original_link_html>";
+    }
+
+    if ( isset( $atts['rel'] ) ) {
+        $all_parts = preg_split( '/\s/', "{$atts['rel']['value']} $rel", -1, PREG_SPLIT_NO_EMPTY );
+        $rel       = implode( ' ', array_unique( $all_parts ) );
+    }
+
+    $atts['rel']['whole'] = 'rel="' . esc_attr( $rel ) . '"';
+    $link_html            = join( ' ', array_column( $atts, 'whole' ) );
+
+    if ( $is_escaped ) {
+        $link_html = preg_replace( '/[\'"]/', '\\\\$0', $link_html );
     }
 

trunk/tests/phpunit/tests/formatting/WPTargetedLinkRel.php

@@ -39 +39 @@
     public function test_rel_with_single_quote_delimiter() {
         $content  = '<p>Links: <a href="/" rel=\'existing values\' target="_blank">Existing rel</a></p>';
-        $expected = '<p>Links: <a href="/" rel=\'existing values noopener noreferrer\' target="_blank">Existing rel</a></p>';
+        $expected = '<p>Links: <a href="/" rel="existing values noopener noreferrer" target="_blank">Existing rel</a></p>';
         $this->assertEquals( $expected, wp_targeted_link_rel( $content ) );
     }
@@ -52 +52 @@
         $content  = '<p>Links: <a href="/" rel = existing target="_blank">Existing rel</a></p>';
         $expected = '<p>Links: <a href="/" rel="existing noopener noreferrer" target="_blank">Existing rel</a></p>';
-        $this->assertEquals( $expected, wp_targeted_link_rel( $content ) );
-    }
-
-    public function test_rel_value_spaced_and_no_delimiter_and_values_to_escape() {
-        $content  = '<p>Links: <a href="/" rel = existing"value target="_blank">Existing rel</a></p>';
-        $expected = '<p>Links: <a href="/" rel="existing&quot;value noopener noreferrer" target="_blank">Existing rel</a></p>';
         $this->assertEquals( $expected, wp_targeted_link_rel( $content ) );
     }
@@ -115 +109 @@
 
     /**
-     * Ensure correct quotes are used when relation attribute (rel) is missing.
+     * Ensure the content of style and script tags are not processed
      *
      * @ticket 47244
      */
-    public function test_wp_targeted_link_rel_should_use_correct_quotes() {
-        $content  = '<p>Links: <a href=\'\/\' target=\'_blank\'>No rel<\/a><\/p>';
-        $expected = '<p>Links: <a href=\'\/\' target=\'_blank\' rel=\'noopener noreferrer\'>No rel<\/a><\/p>';
-        $this->assertEquals( $expected, wp_targeted_link_rel( $content ) );
-
-        $content  = '<p>Links: <a href=\'\/\' target=_blank>No rel<\/a><\/p>';
-        $expected = '<p>Links: <a href=\'\/\' target=_blank rel=\'noopener noreferrer\'>No rel<\/a><\/p>';
+    public function test_wp_targeted_link_rel_skips_style_and_scripts() {
+        $content  = '<style><a href="/" target=a></style><p>Links: <script>console.log("<a href=\'/\' target=a>hi</a>");</script><script>alert(1);</script>here <a href="/" target=_blank>aq</a></p><script>console.log("<a href=\'last\' target=\'_blank\'")</script>';
+        $expected = '<style><a href="/" target=a></style><p>Links: <script>console.log("<a href=\'/\' target=a>hi</a>");</script><script>alert(1);</script>here <a href="/" target="_blank" rel="noopener noreferrer">aq</a></p><script>console.log("<a href=\'last\' target=\'_blank\'")</script>';
         $this->assertEquals( $expected, wp_targeted_link_rel( $content ) );
     }
@@ -140 +130 @@
     }
 
+    public function test_wp_targeted_link_rel_tab_separated_values_are_split() {
+        $content  = "<p>Links: <a href=\"/\" target=\"_blank\" rel=\"ugc\t\tnoopener\t\">No rel</a></p>";
+        $expected = '<p>Links: <a href="/" target="_blank" rel="ugc noopener noreferrer">No rel</a></p>';
+        $this->assertEquals( $expected, wp_targeted_link_rel( $content ) );
+    }
+
 }