From e07fadea76aa7329c4b955a59e74cb867c733706 Mon Sep 17 00:00:00 2001
From: Kevin Decherf <kevin@kdecherf.com>
Date: Thu, 6 Sep 2018 22:26:20 +0200
Subject: Refactor updateOriginUrl to include new behaviors behaviors

- Leave origin_url unchanged if difference is an ending slash
- Leave origin_url unchanged if difference is scheme
- Ignore (noop) if difference is query string or fragment

Signed-off-by: Kevin Decherf <kevin@kdecherf.com>
---
 src/Wallabag/CoreBundle/Helper/ContentProxy.php | 54 ++++++++++++++++++++-----
 1 file changed, 45 insertions(+), 9 deletions(-)

(limited to 'src/Wallabag/CoreBundle/Helper')

diff --git a/src/Wallabag/CoreBundle/Helper/ContentProxy.php b/src/Wallabag/CoreBundle/Helper/ContentProxy.php
index da0ec5a3..007ee8bb 100644
--- a/src/Wallabag/CoreBundle/Helper/ContentProxy.php
+++ b/src/Wallabag/CoreBundle/Helper/ContentProxy.php
@@ -246,15 +246,7 @@ class ContentProxy
      */
     private function stockEntry(Entry $entry, array $content)
     {
-        // When a redirection occurs while fetching an entry
-        // we move the original url in origin_url property if empty
-        // and set the entry url with the final value
-        if (!empty($content['url']) && $entry->getUrl() !== $content['url']) {
-            if (empty($entry->getOriginUrl())) {
-                $entry->setOriginUrl($entry->getUrl());
-            }
-            $entry->setUrl($content['url']);
-        }
+        $this->updateOriginUrl($entry, $content['url']);
 
         $this->setEntryDomainName($entry);
 
@@ -320,6 +312,50 @@ class ContentProxy
         }
     }
 
+    /**
+     * Update the origin_url field when a redirection occurs
+     * This field is set if it is empty and new url does not match ignore list.
+     *
+     * @param Entry  $entry
+     * @param string $url
+     */
+    private function updateOriginUrl(Entry $entry, $url)
+    {
+        if (!empty($url) && $entry->getUrl() !== $url) {
+            $parsed_entry_url = parse_url($entry->getUrl());
+            $parsed_content_url = parse_url($url);
+
+            $diff_ec = array_diff_assoc($parsed_entry_url, $parsed_content_url);
+            $diff_ce = array_diff_assoc($parsed_content_url, $parsed_entry_url);
+
+            $diff = array_merge($diff_ec, $diff_ce);
+            $diff_keys = array_keys($diff);
+            sort($diff_keys);
+
+            switch ($diff_keys) {
+                case ['path']:
+                    if (($parsed_entry_url['path'] . '/' === $parsed_content_url['path']) // diff is trailing slash, we only replace the url of the entry
+                        || ($url === urldecode($entry->getUrl()))) { // we update entry url if new url is a decoded version of it, see EntryRepository#findByUrlAndUserId
+                        $entry->setUrl($url);
+                    }
+                    break;
+                case ['scheme']:
+                    $entry->setUrl($url);
+                    break;
+                case ['fragment']:
+                case ['query']:
+                    // noop
+                    break;
+                default:
+                    if (empty($entry->getOriginUrl())) {
+                        $entry->setOriginUrl($entry->getUrl());
+                    }
+                    $entry->setUrl($url);
+                    break;
+            }
+        }
+    }
+
     /**
      * Validate that the given content has at least a title, an html and a url.
      *
-- 
cgit v1.2.3