Fixes paquettg#173 Added option depthFirstSearch for correct behavior

Author: paquettg

CHANGELOG.md

@@ -10,6 +10,8 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 ### Added
 - New `removeSmartyScripts` configuration setting. Defaults to true.
 - Added `declare(strict_types=1)` to all source files.
+- Added new option `depthFirstSearch`.
+- Deprecated option `depthFirstSearch` and marked for removal in `3.0.0`.
 
 ### Changed
 - Started using a changelog.

README.md

@@ -172,10 +172,15 @@ Preserves Line Breaks if set to `true`. If set to `false` line breaks are cleane
 
 Set this to `false` if you want to preserve whitespace inside of text nodes. It is set to `true` by default.
 
-
 **removeSmartyScripts**
 
-Set this to `false` if you want to preserve smarty sccript found in the html content. It is set to `true` by default.
+Set this to `false` if you want to preserve smarty script found in the html content. It is set to `true` by default.
+
+**depthFirstSearch**
+
+By default this is set to `false` for legacy support. Setting this to `true` will change the behavior of find to order elements by depth first. This will properly preserve the order of elements as they where in the HTML.
+
+This option is depricated and will be removed in version `3.0.0` with the new behavior being as if it was set to `true`.
 
 Static Facade
 -------------

src/PHPHtmlParser/Dom.php

@@ -237,7 +237,7 @@ public function find(string $selector, int $nth = null)
     {
         $this->isLoaded();
 
-        return $this->root->find($selector, $nth);
+        return $this->root->find($selector, $nth, $this->options->get('depthFirstSearch'));
     }
 
     /**

src/PHPHtmlParser/Dom/AbstractNode.php

@@ -429,11 +429,13 @@ public function ancestorByTag(string $tag): AbstractNode
      *
      * @param string $selector
      * @param int $nth
+     * @param bool $depthFirst
      * @return mixed
      */
-    public function find(string $selector, int $nth = null)
+    public function find(string $selector, int $nth = null, bool $depthFirst = false)
     {
         $selector = new Selector($selector, new SelectorParser());
+        $selector->setDepthFirstFind($depthFirst);
         $nodes    = $selector->find($this);
 
         if ( ! is_null($nth)) {

src/PHPHtmlParser/Options.php

@@ -14,6 +14,7 @@
  * @property bool preserveLineBreaks
  * @property bool removeDoubleSpace
  * @property bool removeSmartyScripts
+ * @property bool depthFirstSearch
  */
 class Options
 {
@@ -33,6 +34,7 @@ class Options
         'preserveLineBreaks'   => false,
         'removeDoubleSpace'    => true,
         'removeSmartyScripts'  => true,
+        'depthFirstSearch'     => false,
     ];
 
     /**

src/PHPHtmlParser/Selector/Selector.php

@@ -20,6 +20,11 @@ class Selector
      */
     protected $selectors = [];
 
+    /**
+     * @var bool
+     */
+    private $depthFirst = false;
+
     /**
      * Constructs with the selector string
      *
@@ -40,6 +45,15 @@ public function getSelectors()
         return $this->selectors;
     }
 
+    /**
+     * @param bool $status
+     * @return void
+     */
+    public function setDepthFirstFind(bool $status): void
+    {
+        $this->depthFirst = $status;
+    }
+
     /**
      * Attempts to find the selectors starting from the given
      * node object.
@@ -151,8 +165,19 @@ protected function seek(array $nodes, array $rule, array $options): array
                     if ($child instanceof InnerNode &&
                         $child->hasChildren()
                     ) {
-                        // we still want to check its children
-                        $children[] = $child;
+                        if ($this->depthFirst) {
+                            if ( ! isset($options['checkGrandChildren']) ||
+                                $options['checkGrandChildren']) {
+                                // we have a child that failed but are not leaves.
+                                $matches = $this->seek([$child], $rule, $options);
+                                foreach ($matches as $match) {
+                                    $return[] = $match;
+                                }
+                            }
+                        } else {
+                            // we still want to check its children
+                            $children[] = $child;
+                        }
                     }
                 }
 

tests/DomTest.php

@@ -429,4 +429,27 @@ public function testInfiniteLoopNotHappening()
         $metaNodes = $dom->root->find('meta');
         $this->assertEquals(4, count($metaNodes));
     }
+
+    public function testFindOrder()
+    {
+        $str = '<p><img src="http://example.com/first.jpg"></p><img src="http://example.com/second.jpg">';
+        $dom = new Dom();
+        $dom->load($str);
+        $images = $dom->find('img');
+
+        $this->assertEquals('<img src="http://example.com/second.jpg" />', (string)$images[0]);
+    }
+
+    public function testFindDepthFirstSearch()
+    {
+        $str = '<p><img src="http://example.com/first.jpg"></p><img src="http://example.com/second.jpg">';
+        $dom = new Dom();
+        $dom->setOptions([
+            'depthFirstSearch' => true,
+        ]);
+        $dom->load($str);
+        $images = $dom->find('img');
+
+        $this->assertEquals('<img src="http://example.com/first.jpg" />', (string)$images[0]);
+    }
 }