From 8f95f4ad58b96a7116083c847b247348ade279a7 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Mi=C5=A1o=20Belica?= <miso.belica@gmail.com>
Date: Fri, 21 Feb 2014 11:31:44 +0100
Subject: Ignore attributes with illegal chars in name (fixes #23)

This is neccesary because method "DOMElement::setAttribute"
throws exception for wrong names so DOM elements
can't contain these attributes.
---
 src/HTML5/Parser/DOMTreeBuilder.php |  8 +++++++-
 src/HTML5/Parser/Tokenizer.php      | 25 +++++++++++++++++++++----
 2 files changed, 28 insertions(+), 5 deletions(-)

(limited to 'src/HTML5')

diff --git a/src/HTML5/Parser/DOMTreeBuilder.php b/src/HTML5/Parser/DOMTreeBuilder.php
index cecebaf..b79c298 100644
--- a/src/HTML5/Parser/DOMTreeBuilder.php
+++ b/src/HTML5/Parser/DOMTreeBuilder.php
@@ -234,7 +234,13 @@ class DOMTreeBuilder implements EventHandler {
         $aName = Elements::normalizeMathMlAttribute($aName);
       }
 
-      $ele->setAttribute($aName, $aVal);
+      try {
+        $ele->setAttribute($aName, $aVal);
+      }
+      catch(\DOMException $e) {
+        $this->parseError("Illegal attribute name for tag $name. Ignoring: $aName");
+        continue;
+      }
 
       // This is necessary on a non-DTD schema, like HTML5.
       if ($aName == 'id') {
diff --git a/src/HTML5/Parser/Tokenizer.php b/src/HTML5/Parser/Tokenizer.php
index a79781f..f21d30b 100644
--- a/src/HTML5/Parser/Tokenizer.php
+++ b/src/HTML5/Parser/Tokenizer.php
@@ -414,16 +414,33 @@ class Tokenizer {
       $name = $this->scanner->current();
       $this->scanner->next();
     }
-    if (preg_match('/[\'\"]/', $name)) {
-    //if (strspn($name, '\'\"')) {
+
+    $isValidAttribute = TRUE;
+    // Attribute names can contain most Unicode characters for HTML5.
+    // But method "DOMElement::setAttribute" is throwing exception
+    // because of it's own internal restriction so these have to be filtered.
+    // see issue #23: https://github.com/Masterminds/html5-php/issues/23
+    // and http://www.w3.org/TR/2011/WD-html5-20110525/syntax.html#syntax-attribute-name
+    if (preg_match("/[\x1-\x2C\\/\x3B-\x40\x5B-\x5E\x60\x7B-\x7F]/u", $name)) {
       $this->parseError("Unexpected characters in attribute name: %s", $name);
+      $isValidAttribute = FALSE;
+    }
+    // There is no limitation for 1st character in HTML5.
+    // But method "DOMElement::setAttribute" is throwing exception for the
+    // characters below so they have to be filtered.
+    // see issue #23: https://github.com/Masterminds/html5-php/issues/23
+    // and http://www.w3.org/TR/2011/WD-html5-20110525/syntax.html#syntax-attribute-name
+    else if (preg_match("/^[0-9.-]/u", $name)) {
+      $this->parseError("Unexpected character at the begining of attribute name: %s", $name);
+      $isValidAttribute = FALSE;
     }
     // 8.1.2.3
     $this->scanner->whitespace();
 
     $val = $this->attributeValue();
-    //return array($name, $val);
-    $attributes[$name] = $val;
+    if($isValidAttribute) {
+      $attributes[$name] = $val;
+    }
     return TRUE;
   }
 
-- 
cgit v1.2.3