From e94df546c33ad62e5b404447776f4fcc113f8de1 Mon Sep 17 00:00:00 2001
From: Hugo Sales <hugo@hsal.es>
Date: Sun, 25 Apr 2021 21:14:35 +0000
Subject: [PATCH] [Posting] Extract and store URLs from note content. Introduce
 'AttachmentStoreNew' event

---
 components/Posting/Posting.php | 29 +++++++++++++++++++++-------
 src/Core/GSFile.php            | 35 +++++++++++++++++++++++-----------
 2 files changed, 46 insertions(+), 18 deletions(-)

diff --git a/components/Posting/Posting.php b/components/Posting/Posting.php
index 313a18448a..e791792ef6 100644
--- a/components/Posting/Posting.php
+++ b/components/Posting/Posting.php
@@ -41,6 +41,13 @@ use Symfony\Component\Form\Extension\Core\Type\TextareaType;
 
 class Posting extends Component
 {
+    /**
+     * "Perfect URL Regex", courtesy of https://urlregex.com/
+     */
+    const URL_REGEX = <<<END
+%(?:(?:https?|ftp)://)(?:\\S+(?::\\S*)?@|\\d{1,3}(?:\\.\\d{1,3}){3}|(?:(?:[a-z\\d\\x{00a1}-\\x{ffff}]+-?)*[a-z\\d\\x{00a1}-\\x{ffff}]+)(?:\\.(?:[a-z\\d\\x{00a1}-\\x{ffff}]+-?)*[a-z\\d\\x{00a1}-\\x{ffff}]+)*(?:\\.[a-z\\x{00a1}-\\x{ffff}]{2,6}))(?::\\d+)?(?:[^\\s]*)?%iu
+END;
+
     /**
      * HTML render event handler responsible for adding and handling
      * the result of adding the note submission form, only if a user is logged in
@@ -96,24 +103,32 @@ class Posting extends Component
      */
     public static function storeNote(int $actor_id, ?string $content, array $attachments, bool $is_local, ?int $reply_to = null, ?int $repeat_of = null)
     {
-        $note = Note::create([
+        $content = Security::sanitize($content);
+        $note    = Note::create([
             'gsactor_id' => $actor_id,
-            'content'    => Security::sanitize($content),
+            'content'    => $content,
             'is_local'   => $is_local,
             'reply_to'   => $reply_to,
             'repeat_of'  => $repeat_of,
         ]);
+
         $processed_attachments = [];
         foreach ($attachments as $f) {
-            $na = GSFile::validateAndStoreAttachment(
+            $processed_attachments[] = GSFile::validateAndStoreAttachment(
                 $f, Common::config('attachments', 'dir'),
-                Security::sanitize($title = $f->getClientOriginalName()),
-                $is_local = true, $actor_id
+                Security::sanitize($f->getClientOriginalName()),
+                is_local: true, actor_id: $actor_id
             );
-            $processed_attachments[] = $na;
-            DB::persist($na);
         }
+
+        $matched_urls = [];
+        preg_match_all(self::URL_REGEX, $content, $matched_urls, PREG_SET_ORDER);
+        foreach ($matched_urls as $match) {
+            $processed_attachments[] = GSFile::validateAndStoreURL($url);
+        }
+
         DB::persist($note);
+
         // Need file and note ids for the next step
         DB::flush();
         if ($processed_attachments != []) {
diff --git a/src/Core/GSFile.php b/src/Core/GSFile.php
index 5e8b9c3f87..60928a8908 100644
--- a/src/Core/GSFile.php
+++ b/src/Core/GSFile.php
@@ -57,20 +57,33 @@ class GSFile
             'is_local'   => $is_local,
         ]);
         $sfile->move($dest_dir, $hash);
+        DB::persist($attachment);
+        Event::handle('AttachmentStoreNew', [&$attachment]);
         return $attachment;
     }
 
     /**
-     * Perform file validation (checks and normalization) and store the given file
+     * Create an attachment for the given URL, fetching the mimetype
+     *
+     * @throws \InvalidArgumentException
      */
-    public static function validateAndStoreAttachmentThumbnail(SymfonyFile $sfile,
-                                                      string $dest_dir,
-                                                      ?string $title = null,
-                                                      bool $is_local = true,
-                                                      int $actor_id = null): Attachment//Thumbnail
+    public static function validateAndStoreURL(string $url): Attachment
     {
-        $attachment = self::validateAndStoreAttachment($sfile,$dest_dir,$title,$is_local,$actor_id);
-        return $attachment;
+        if (Common::isValidHttpUrl($url)) {
+            HTTPClient::head($url);
+            $headers    = $head->getHeaders();
+            $headers    = array_change_key_case($headers, CASE_LOWER);
+            $attachment = Attachment::create([
+                'remote_url'      => $match[0],
+                'remote_url_hash' => hash('sha256', $match[0]),
+                'mimetype'        => $headers['content-type'],
+            ]);
+            DB::persist($attachment);
+            Event::handle('AttachmentStoreNew', [&$at]);
+            return $attachment;
+        } else {
+            throw new \InvalidArgumentException();
+        }
     }
 
     /**
@@ -159,7 +172,7 @@ class GSFile
      */
     public static function mimetypeMajor(string $mime)
     {
-        return explode('/', self::mimeBare($mime))[0];
+        return explode('/', self::mimetypeBare($mime))[0];
     }
 
     /**
@@ -167,13 +180,13 @@ class GSFile
      */
     public static function mimetypeMinor(string $mime)
     {
-        return explode('/', self::mimeBare($mime))[1];
+        return explode('/', self::mimetypeBare($mime))[1];
     }
 
     /**
      *  Get only the mimetype and not additional info (separated from bare mime with semi-colon)
      */
-    public static function mimeBare(string $mimetype)
+    public static function mimetypeBare(string $mimetype)
     {
         $mimetype = mb_strtolower($mimetype);
         if (($semicolon = mb_strpos($mimetype, ';')) !== false) {