Issue #2897254 by Jo Fitzgerald, heddn, rakesh.gectcr, maxocub, masipila, phenaproxima, larowlan: URLs without http:// are broken after migration from d6 or d7

2017-10-23 13:48:49 +01:00 · 2017-10-23 13:48:49 +01:00 · df1dd607ad
parent df974e5b19
commit df1dd607ad
7 changed files with 239 additions and 82 deletions
--- a/core/modules/link/src/Plugin/migrate/field/d6/LinkField.php
+++ b/core/modules/link/src/Plugin/migrate/field/d6/LinkField.php
@ -39,7 +39,7 @@ class LinkField extends FieldPluginBase {
   */
  public function processFieldValues(MigrationInterface $migration, $field_name, $data) {
    $process = [
-      'plugin' => 'd6_field_link',
+      'plugin' => 'field_link',
      'source' => $field_name,
    ];
    $migration->mergeProcessOfProperty($field_name, $process);
--- a/core/modules/link/src/Plugin/migrate/process/FieldLink.php
+++ b/core/modules/link/src/Plugin/migrate/process/FieldLink.php
@ -0,0 +1,133 @@
+<?php
+
+namespace Drupal\link\Plugin\migrate\process;
+
+use Drupal\migrate\Plugin\MigrationInterface;
+use Drupal\migrate\MigrateExecutableInterface;
+use Drupal\migrate\ProcessPluginBase;
+use Drupal\migrate\Row;
+
+/**
+ * Transform a pre-Drupal 8 formatted link for use in Drupal 8.
+ *
+ * Previous to Drupal 8, URLs didn't need to have a URI scheme assigned. The
+ * contrib link module would auto-prefix the URL with a URI scheme. A link in
+ * Drupal 8 has more validation and external links must include the URI scheme.
+ * All external URIs need to be converted to use a URI scheme.
+ *
+ * Available configuration keys
+ * - uri_scheme: (optional) The URI scheme prefix to use for URLs without a
+ *   scheme. Defaults to 'http://', which was the default in Drupal 6 and
+ *   Drupal 7.
+ *
+ * Examples:
+ *
+ * Consider a link field migration, where you want to use https:// as the
+ * prefix:
+ *
+ * @code
+ * process:
+ *   field_link:
+ *     plugin: field_link
+ *     uri_scheme: 'https://'
+ *     source: field_link
+ * @endcode
+ *
+ * @MigrateProcessPlugin(
+ *   id = "field_link"
+ * )
+ */
+class FieldLink extends ProcessPluginBase {
+
+  /**
+   * {@inheritdoc}
+   */
+  public function __construct(array $configuration, $plugin_id, $plugin_definition, MigrationInterface $migration) {
+    $configuration += ['uri_scheme' => 'http://'];
+    parent::__construct($configuration, $plugin_id, $plugin_definition);
+  }
+
+  /**
+   * Turn a Drupal 6/7 URI into a Drupal 8-compatible format.
+   *
+   * @param string $uri
+   *   The 'url' value from Drupal 6/7.
+   *
+   * @return string
+   *   The Drupal 8-compatible URI.
+   *
+   * @see \Drupal\link\Plugin\Field\FieldWidget\LinkWidget::getUserEnteredStringAsUri()
+   */
+  protected function canonicalizeUri($uri) {
+    // If we already have a scheme, we're fine.
+    if (empty($uri) || parse_url($uri, PHP_URL_SCHEME)) {
+      return $uri;
+    }
+
+    // Remove the <front> component of the URL.
+    if (strpos($uri, '<front>') === 0) {
+      $uri = substr($uri, strlen('<front>'));
+    }
+    else {
+      // List of unicode-encoded characters that were allowed in URLs,
+      // according to link module in Drupal 7. Every character between &#x00BF;
+      // and &#x00FF; (except × &#x00D7; and ÷ &#x00F7;) with the addition of
+      // &#x0152;, &#x0153; and &#x0178;.
+      // @see http://cgit.drupalcode.org/link/tree/link.module?h=7.x-1.5-beta2#n1382
+      $link_ichars = '¿ÀÁÂÃÄÅÆÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖØÙÚÛÜÝÞßàáâãäåæçèéêëìíîïðñòóôõöøùúûüýþÿŒœŸ';
+
+      // Pattern specific to internal links.
+      $internal_pattern = "/^(?:[a-z0-9" . $link_ichars . "_\-+\[\] ]+)";
+
+      $directories = "(?:\/[a-z0-9" . $link_ichars . "_\-\.~+%=&,$'#!():;*@\[\]]*)*";
+      // Yes, four backslashes == a single backslash.
+      $query = "(?:\/?\?([?a-z0-9" . $link_ichars . "+_|\-\.~\/\\\\%=&,$'():;*@\[\]{} ]*))";
+      $anchor = "(?:#[a-z0-9" . $link_ichars . "_\-\.~+%=&,$'():;*@\[\]\/\?]*)";
+
+      // The rest of the path for a standard URL.
+      $end = $directories . '?' . $query . '?' . $anchor . '?' . '$/i';
+
+      if (!preg_match($internal_pattern . $end, $uri)) {
+        $link_domains = '[a-z][a-z0-9-]{1,62}';
+
+        // Starting a parenthesis group with (?: means that it is grouped, but is not captured
+        $authentication = "(?:(?:(?:[\w\.\-\+!$&'\(\)*\+,;=" . $link_ichars . "]|%[0-9a-f]{2})+(?::(?:[\w" . $link_ichars . "\.\-\+%!$&'\(\)*\+,;=]|%[0-9a-f]{2})*)?)?@)";
+        $domain = '(?:(?:[a-z0-9' . $link_ichars . ']([a-z0-9' . $link_ichars . '\-_\[\]])*)(\.(([a-z0-9' . $link_ichars . '\-_\[\]])+\.)*(' . $link_domains . '|[a-z]{2}))?)';
+        $ipv4 = '(?:[0-9]{1,3}(\.[0-9]{1,3}){3})';
+        $ipv6 = '(?:[0-9a-fA-F]{1,4}(\:[0-9a-fA-F]{1,4}){7})';
+        $port = '(?::([0-9]{1,5}))';
+
+        // Pattern specific to external links.
+        $external_pattern = '/^' . $authentication . '?(' . $domain . '|' . $ipv4 . '|' . $ipv6 . ' |localhost)' . $port . '?';
+        if (preg_match($external_pattern . $end, $uri)) {
+          return $this->configuration['uri_scheme'] . $uri;
+        }
+      }
+    }
+
+    // Add the internal: scheme and ensure a leading slash.
+    return 'internal:/' . ltrim($uri, '/');
+  }
+
+  /**
+   * {@inheritdoc}
+   */
+  public function transform($value, MigrateExecutableInterface $migrate_executable, Row $row, $destination_property) {
+    $attributes = unserialize($value['attributes']);
+    // Drupal 6/7 link attributes might be double serialized.
+    if (!is_array($attributes)) {
+      $attributes = unserialize($attributes);
+    }
+
+    if (!$attributes) {
+      $attributes = [];
+    }
+
+    // Massage the values into the correct form for the link.
+    $route['uri'] = $this->canonicalizeUri($value['url']);
+    $route['options']['attributes'] = $attributes;
+    $route['title'] = $value['title'];
+    return $route;
+  }
+
+}
--- a/core/modules/link/src/Plugin/migrate/process/d6/CckLink.php
+++ b/core/modules/link/src/Plugin/migrate/process/d6/CckLink.php
@ -2,9 +2,9 @@

 namespace Drupal\link\Plugin\migrate\process\d6;

-@trigger_error('CckLink is deprecated in Drupal 8.3.x and will be removed before
-Drupal 9.0.x. Use \Drupal\link\Plugin\migrate\process\d6\FieldLink instead.',
-E_USER_DEPRECATED);
+use Drupal\link\Plugin\migrate\process\FieldLink;
+
+@trigger_error('CckLink is deprecated in Drupal 8.3.x and will be removed before Drupal 9.0.x. Use \Drupal\link\Plugin\migrate\process\FieldLink instead.', E_USER_DEPRECATED);

 /**
 * @MigrateProcessPlugin(
@ -12,6 +12,6 @@ E_USER_DEPRECATED);
 * )
 *
 * @deprecated in Drupal 8.3.x, to be removed before Drupal 9.0.x. Use
- * \Drupal\link\Plugin\migrate\process\d6\FieldLink instead.
+ * \Drupal\link\Plugin\migrate\process\FieldLink instead.
 */
 class CckLink extends FieldLink {}
--- a/core/modules/link/src/Plugin/migrate/process/d6/FieldLink.php
+++ b/core/modules/link/src/Plugin/migrate/process/d6/FieldLink.php
@ -2,85 +2,16 @@

 namespace Drupal\link\Plugin\migrate\process\d6;

-use Drupal\Core\Plugin\ContainerFactoryPluginInterface;
-use Drupal\migrate\Plugin\MigrationInterface;
-use Drupal\migrate\MigrateExecutableInterface;
-use Drupal\migrate\ProcessPluginBase;
-use Drupal\migrate\Row;
-use Symfony\Component\DependencyInjection\ContainerInterface;
+use Drupal\link\Plugin\migrate\process\FieldLink as GeneralPurposeFieldLink;
+
+@trigger_error('FieldLink is deprecated in Drupal 8.4.x and will be removed before Drupal 9.0.x. Use \Drupal\link\Plugin\migrate\process\FieldLink instead.', E_USER_DEPRECATED);

 /**
 * @MigrateProcessPlugin(
 *   id = "d6_field_link"
 * )
+ *
+ * @deprecated in Drupal 8.4.x, to be removed before Drupal 9.0.x. Use
+ * \Drupal\link\Plugin\migrate\process\FieldLink instead.
 */
-class FieldLink extends ProcessPluginBase implements ContainerFactoryPluginInterface {
-
-  /**
-   * {@inheritdoc}
-   */
-  public function __construct(array $configuration, $plugin_id, $plugin_definition, MigrationInterface $migration) {
-    parent::__construct($configuration, $plugin_id, $plugin_definition);
-    $this->migration = $migration;
-  }
-
-  /**
-   * {@inheritdoc}
-   */
-  public static function create(ContainerInterface $container, array $configuration, $plugin_id, $plugin_definition, MigrationInterface $migration = NULL) {
-    return new static(
-      $configuration,
-      $plugin_id,
-      $plugin_definition,
-      $migration
-    );
-  }
-
-  /**
-   * Turn a Drupal 6 URI into a Drupal 8-compatible format.
-   *
-   * @param string $uri
-   *   The 'url' value from Drupal 6.
-   *
-   * @return string
-   *   The Drupal 8-compatible URI.
-   *
-   * @see \Drupal\link\Plugin\Field\FieldWidget\LinkWidget::getUserEnteredStringAsUri()
-   */
-  protected function canonicalizeUri($uri) {
-    // If we already have a scheme, we're fine.
-    if (empty($uri) || !is_null(parse_url($uri, PHP_URL_SCHEME))) {
-      return $uri;
-    }
-
-    // Remove the <front> component of the URL.
-    if (strpos($uri, '<front>') === 0) {
-      $uri = substr($uri, strlen('<front>'));
-    }
-
-    // Add the internal: scheme and ensure a leading slash.
-    return 'internal:/' . ltrim($uri, '/');
-  }
-
-  /**
-   * {@inheritdoc}
-   */
-  public function transform($value, MigrateExecutableInterface $migrate_executable, Row $row, $destination_property) {
-    $attributes = unserialize($value['attributes']);
-    // Drupal 6 link attributes might be double serialized.
-    if (!is_array($attributes)) {
-      $attributes = unserialize($attributes);
-    }
-
-    if (!$attributes) {
-      $attributes = [];
-    }
-
-    // Massage the values into the correct form for the link.
-    $route['uri'] = $this->canonicalizeUri($value['url']);
-    $route['options']['attributes'] = $attributes;
-    $route['title'] = $value['title'];
-    return $route;
-  }
-
-}
+class FieldLink extends GeneralPurposeFieldLink {}
--- a/core/modules/link/tests/src/Unit/Plugin/migrate/field/d6/LinkFieldTest.php
+++ b/core/modules/link/tests/src/Unit/Plugin/migrate/field/d6/LinkFieldTest.php
@ -50,7 +50,7 @@ class LinkFieldTest extends UnitTestCase {
    $this->plugin->processFieldValues($this->migration, 'somefieldname', []);

    $expected = [
-      'plugin' => 'd6_field_link',
+      'plugin' => 'field_link',
      'source' => 'somefieldname',
    ];
    $this->assertSame($expected, $this->migration->getProcess());
--- a/core/modules/link/tests/src/Unit/Plugin/migrate/process/FieldLinkTest.php
+++ b/core/modules/link/tests/src/Unit/Plugin/migrate/process/FieldLinkTest.php
@ -0,0 +1,92 @@
+<?php
+
+namespace Drupal\Tests\link\Unit\Plugin\migrate\process;
+
+use Drupal\link\Plugin\migrate\process\FieldLink;
+use Drupal\migrate\MigrateExecutableInterface;
+use Drupal\migrate\Plugin\MigrationInterface;
+use Drupal\migrate\Row;
+use Drupal\Tests\UnitTestCase;
+
+/**
+ * @group Link
+ */
+class FieldLinkTest extends UnitTestCase {
+
+  /**
+   * Test the url transformations in the FieldLink process plugin.
+   *
+   * @dataProvider canonicalizeUriDataProvider
+   */
+  public function testCanonicalizeUri($url, $expected, $configuration = []) {
+    $link_plugin = new FieldLink($configuration, '', [], $this->getMock(MigrationInterface::class));
+    $transformed = $link_plugin->transform([
+      'url' => $url,
+      'title' => '',
+      'attributes' => serialize([]),
+    ], $this->getMock(MigrateExecutableInterface::class), $this->getMockBuilder(Row::class)->disableOriginalConstructor()->getMock(), NULL);
+    $this->assertEquals($expected, $transformed['uri']);
+  }
+
+  /**
+   * Data provider for testCanonicalizeUri.
+   */
+  public function canonicalizeUriDataProvider() {
+    return [
+      'Simple front-page' => [
+        '<front>',
+        'internal:/',
+      ],
+      'Front page with query' => [
+        '<front>?query=1',
+        'internal:/?query=1',
+      ],
+      'No leading forward slash' => [
+        'node/10',
+        'internal:/node/10',
+      ],
+      'Leading forward slash' => [
+        '/node/10',
+        'internal:/node/10',
+      ],
+      'Existing scheme' => [
+        'scheme:test',
+        'scheme:test',
+      ],
+      'Absolute URL with protocol prefix' => [
+        'http://www.google.com',
+        'http://www.google.com',
+      ],
+      'Absolute URL without protocol prefix' => [
+        'www.yahoo.com',
+        'http://www.yahoo.com',
+      ],
+      'Absolute URL without protocol prefix nor www' => [
+        'yahoo.com',
+        'https://yahoo.com',
+        ['uri_scheme' => 'https://'],
+      ],
+      'Absolute URL with non-standard characters' => [
+        'http://www.ßÀÑÐ¥ƒå¢ë.com',
+        'http://www.ßÀÑÐ¥ƒå¢ë.com',
+      ],
+      'Absolute URL with non-standard characters, without protocol prefix' => [
+        'www.ÐØÑ¢åþë.com',
+        'http://www.ÐØÑ¢åþë.com',
+      ],
+      'Absolute URL with non-standard top level domain' => [
+        'http://www.example.xxx',
+        'http://www.example.xxx',
+      ],
+      'Internal link with fragment' => [
+        '/node/10#top',
+        'internal:/node/10#top',
+      ],
+      'External link with fragment' => [
+        'http://www.example.com/page#links',
+        'http://www.example.com/page#links',
+      ],
+    ];
+  }
+
+}
--- a/core/modules/link/tests/src/Unit/Plugin/migrate/process/d6/FieldLinkTest.php
+++ b/core/modules/link/tests/src/Unit/Plugin/migrate/process/d6/FieldLinkTest.php
@ -7,6 +7,7 @@ use Drupal\Tests\UnitTestCase;

 /**
 * @group Link
+ * @group legacy
 */
 class FieldLinkTest extends UnitTestCase {