Merge branch 'hotfix/24-random-placeholder'

Close #24

Author: weierophinney

CHANGELOG.md

@@ -2,7 +2,7 @@
 
 All notable changes to this project will be documented in this file, in reverse chronological order by release.
 
-## 2.7.1 - TBD
+## 2.7.1 - 2018-04-09
 
 ### Added
 
@@ -30,6 +30,10 @@ All notable changes to this project will be documented in this file, in reverse
   statement for the `DOMNode` class to the (deprecated) `Zend\Dom\Query` class
   definition.
 
+- [#24](https://github.com/zendframework/zend-dom/pull/24) updates how the
+  tokenizer marks multiple words within attribute values in order to be
+  more robust.
+
 ## 2.7.0 - 2018-03-27
 
 ### Added

src/Document/Query.php

@@ -86,19 +86,23 @@ public static function cssToXpath($path)
             return implode('|', $expressions);
         }
 
+        do {
+            $placeholder = '{' . uniqid(mt_rand(), true) . '}';
+        } while (strpos($path, $placeholder) !== false);
+
         // Arbitrary attribute value contains whitespace
         $path = preg_replace_callback(
-            '/\[\S+?([\'"])((?:(?!\1)[^\\\]|\\.)*)\1\]/',
-            function ($matches) {
-                return str_replace($matches[2], preg_replace('/\s+/', '\s', $matches[2]), $matches[0]);
+            '/\[\S+?([\'"])((?!\1|\\\1).*?)\1\]/',
+            function ($matches) use ($placeholder) {
+                return str_replace($matches[2], preg_replace('/\s+/', $placeholder, $matches[2]), $matches[0]);
             },
             $path
         );
 
         $paths    = ['//'];
         $path     = preg_replace('|\s+>\s+|', '>', $path);
         $segments = preg_split('/\s+/', $path);
-        $segments = str_replace('\s', ' ', $segments);
+        $segments = str_replace($placeholder, ' ', $segments);
 
         foreach ($segments as $key => $segment) {
             $pathSegment = static::_tokenize($segment);
@@ -147,7 +151,7 @@ protected static function _tokenize($expression)
 
         // arbitrary attribute strict equality
         $expression = preg_replace_callback(
-            '/\[@?([a-z0-9_-]+)=([\'"])((?:(?!\2)[^\\\]|\\.)*)\2\]/i',
+            '/\[@?([a-z0-9_-]+)=([\'"])((?!\2|\\\2).*?)\2\]/i',
             function ($matches) {
                 return sprintf("[@%s='%s']", strtolower($matches[1]), str_replace("'", "\\'", $matches[3]));
             },
@@ -156,7 +160,7 @@ function ($matches) {
 
         // arbitrary attribute contains full word
         $expression = preg_replace_callback(
-            '/\[([a-z0-9_-]+)~=([\'"])((?:(?!\2)[^\\\]|\\.)*)\2\]/i',
+            '/\[([a-z0-9_-]+)~=([\'"])((?!\2|\\\2).*?)\2\]/i',
             function ($matches) {
                 return "[contains(concat(' ', normalize-space(@" . strtolower($matches[1]) . "), ' '), ' "
                      . $matches[3] . " ')]";
@@ -166,7 +170,7 @@ function ($matches) {
 
         // arbitrary attribute contains specified content
         $expression = preg_replace_callback(
-            '/\[([a-z0-9_-]+)\*=([\'"])((?:(?!\2)[^\\\]|\\.)*)\2\]/i',
+            '/\[([a-z0-9_-]+)\*=([\'"])((?!\2|\\\2).*?)\2\]/i',
             function ($matches) {
                 return "[contains(@" . strtolower($matches[1]) . ", '"
                      . $matches[3] . "')]";

test/Document/QueryTest.php

@@ -202,4 +202,10 @@ public function testTransformNestedAttributeSelectors($selector, $expectedXpath)
     {
         $this->assertEquals($expectedXpath, Query::cssToXpath($selector));
     }
+
+    public function testAttributeValueWithBackslash()
+    {
+        $test = Query::cssToXpath('select[name="\stop  \start"]');
+        $this->assertEquals("//select[@name='\stop \start']", $test);
+    }
 }