[DomCrawler] Abstract URI-related logic

All the URI parsing logic is externalized in the AbstractUriElement class, implementing the UriElementInterface interface. The AbstractUriElement class have two abstract methods: * setNode: validate the DOMElement node according to the concrete class rules, and set $this->node. * getRawUri: get the raw URI from $this->node. The Link classs now extends AbstractUriElement. This refactor is desirable for #12429.
symfony · valeriangalliat · Feb 7, 2015 · Feb 7, 2015 · Feb 10, 2015 · Feb 20, 2015
commit 5259d279169ed4aef17dbe5087d24c9fc8d8a4a7
diff --git a/src/Symfony/Component/DomCrawler/AbstractUriElement.php b/src/Symfony/Component/DomCrawler/AbstractUriElement.php
@@ -0,0 +1,201 @@
+<?php
+
+/*
+ * This file is part of the Symfony package.
+ *
+ * (c) Fabien Potencier <fabien@symfony.com>
+ *
+ * For the full copyright and license information, please view the LICENSE
+ * file that was distributed with this source code.
+ */
+
+namespace Symfony\Component\DomCrawler;
+
+/**
+ * Any HTML element that can link to an URI.
+ *
+ * @author Fabien Potencier <fabien@symfony.com>
+ *
+ * @api
+ */
+abstract class AbstractUriElement implements UriElementInterface
+{
+    /**
+     * @var \DOMElement
+     */
+    protected $node;
+
+    /**
+     * @var string The method to use for the element URI
+     */
+    protected $method;
+
+    /**
+     * @var string The URI of the page where the element is embedded (or the base href)
+     */
+    protected $currentUri;
+
+    /**
+     * Constructor.
+     *
+     * @param \DOMElement $node       A \DOMElement instance
+     * @param string      $currentUri The URI of the page where the element is embedded (or the base href)
+     * @param string      $method     The method to use for the element URI (get by default)
+     *
+     * @throws \InvalidArgumentException if the node is not a link
+     *
+     * @api
+     */
+    public function __construct(\DOMElement $node, $currentUri, $method = 'GET')
+    {
+        if (!in_array(strtolower(substr($currentUri, 0, 4)), array('http', 'file'))) {
+            throw new \InvalidArgumentException(sprintf('Current URI must be an absolute URL ("%s").', $currentUri));
+        }
+
+        $this->setNode($node);
+        $this->method = $method ? strtoupper($method) : null;
+        $this->currentUri = $currentUri;
+    }
+
+    public function getNode()
+    {
+        return $this->node;
+    }
+
+    public function getMethod()
+    {
+        return $this->method;
+    }
+
+    public function getUri()
+    {
+        $uri = trim($this->getRawUri());
+
+        // absolute URL?
+        if (null !== parse_url($uri, PHP_URL_SCHEME)) {
+            return $uri;
+        }
+
+        // empty URI
+        if (!$uri) {
+            return $this->currentUri;
+        }
+
+        // an anchor
+        if ('#' === $uri[0]) {
+            return $this->cleanupAnchor($this->currentUri).$uri;
+        }
+
+        $baseUri = $this->cleanupUri($this->currentUri);
+
+        if ('?' === $uri[0]) {
+            return $baseUri.$uri;
+        }
+
+        // absolute URL with relative schema
+        if (0 === strpos($uri, '//')) {
+            return preg_replace('#^([^/]*)//.*$#', '$1', $baseUri).$uri;
+        }
+
+        $baseUri = preg_replace('#^(.*?//[^/]*)(?:\/.*)?$#', '$1', $baseUri);
+
+        // absolute path
+        if ('/' === $uri[0]) {
+            return $baseUri.$uri;
+        }
+
+        // relative path
+        $path = parse_url(substr($this->currentUri, strlen($baseUri)), PHP_URL_PATH);
+        $path = $this->canonicalizePath(substr($path, 0, strrpos($path, '/')).'/'.$uri);
+
+        return $baseUri.('' === $path || '/' !== $path[0] ? '/' : '').$path;
+    }
+
+    /**
+     * Returns raw URI data.
+     *
+     * @return string
+     */
+    abstract protected function getRawUri();
+
+    /**
+     * Returns the canonicalized URI path (see RFC 3986, section 5.2.4).
+     *
+     * @param string $path URI path
+     *
+     * @return string
+     */
+    protected function canonicalizePath($path)
+    {
+        if ('' === $path || '/' === $path) {
+            return $path;
+        }
+
+        if ('.' === substr($path, -1)) {
            $path = $path.'/';
+        }
+
+        $output = array();
+
+        foreach (explode('/', $path) as $segment) {
+            if ('..' === $segment) {
+                array_pop($output);
+            } elseif ('.' !== $segment) {
+                array_push($output, $segment);
+            }
+        }
+
+        return implode('/', $output);
+    }
+
+    /**
+     * Sets current \DOMElement instance.
+     *
+     * @param \DOMElement $node A \DOMElement instance
+     */
+    abstract protected function setNode(\DOMElement $node);
+
+    /**
+     * Removes the query string and the anchor from the given uri.
+     *
+     * @param string $uri The uri to clean
+     *
+     * @return string
+     */
+    private function cleanupUri($uri)
+    {
+        return $this->cleanupQuery($this->cleanupAnchor($uri));
+    }
+
+    /**
+     * Remove the query string from the uri.
+     *
+     * @param string $uri
+     *
+     * @return string
+     */
+    private function cleanupQuery($uri)
+    {
+        if (false !== $pos = strpos($uri, '?')) {
+            return substr($uri, 0, $pos);
+        }
+
+        return $uri;
+    }
+
+    /**
+     * Remove the anchor from the uri.
+     *
+     * @param string $uri
+     *
+     * @return string
+     */
+    private function cleanupAnchor($uri)
+    {
+        if (false !== $pos = strpos($uri, '#')) {
+            return substr($uri, 0, $pos);
+        }
+
+        return $uri;
+    }
+}
diff --git a/src/Symfony/Component/DomCrawler/CHANGELOG.md b/src/Symfony/Component/DomCrawler/CHANGELOG.md
@@ -1,6 +1,11 @@
 CHANGELOG
 =========
 
+2.7.0
+-----
+
+* All the URI parsing logic have been abstracted in the `AbstractUriElement` class. The `Link` class is now a child of `AbstractUriElement` which implements the new `UriElementInterface`, describing the common `getNode`, `getMethod` and `getUri` methods.
+
 2.5.0
 -----