3 * StatusNet - the distributed open-source microblogging tool
4 * Copyright (C) 2008, 2009, StatusNet, Inc.
6 * This program is free software: you can redistribute it and/or modify
7 * it under the terms of the GNU Affero General Public License as published by
8 * the Free Software Foundation, either version 3 of the License, or
9 * (at your option) any later version.
11 * This program is distributed in the hope that it will be useful,
12 * but WITHOUT ANY WARRANTY; without even the implied warranty of
13 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14 * GNU Affero General Public License for more details.
16 * You should have received a copy of the GNU Affero General Public License
17 * along with this program. If not, see <http://www.gnu.org/licenses/>.
20 if (!defined('GNUSOCIAL')) { exit(1); }
23 * Table Definition for file
25 class File extends Managed_DataObject
27 public $__table = 'file'; // table name
28 public $id; // int(4) primary_key not_null
29 public $urlhash; // varchar(64) unique_key
31 public $filehash; // varchar(64) indexed
32 public $mimetype; // varchar(50)
33 public $size; // int(4)
34 public $title; // varchar(191) not 255 because utf8mb4 takes more space
35 public $date; // int(4)
36 public $protected; // int(4)
37 public $filename; // varchar(191) not 255 because utf8mb4 takes more space
38 public $width; // int(4)
39 public $height; // int(4)
40 public $modified; // timestamp() not_null default_CURRENT_TIMESTAMP
42 const URLHASH_ALG = 'sha256';
43 const FILEHASH_ALG = 'sha256';
45 public static function schemaDef()
49 'id' => array('type' => 'serial', 'not null' => true),
50 'urlhash' => array('type' => 'varchar', 'length' => 64, 'not null' => true, 'description' => 'sha256 of destination URL (url field)'),
51 'url' => array('type' => 'text', 'description' => 'destination URL after following possible redirections'),
52 'filehash' => array('type' => 'varchar', 'length' => 64, 'not null' => false, 'description' => 'sha256 of the file contents, only for locally stored files of course'),
53 'mimetype' => array('type' => 'varchar', 'length' => 50, 'description' => 'mime type of resource'),
54 'size' => array('type' => 'int', 'description' => 'size of resource when available'),
55 'title' => array('type' => 'varchar', 'length' => 191, 'description' => 'title of resource when available'),
56 'date' => array('type' => 'int', 'description' => 'date of resource according to http query'),
57 'protected' => array('type' => 'int', 'description' => 'true when URL is private (needs login)'),
58 'filename' => array('type' => 'varchar', 'length' => 191, 'description' => 'if a local file, name of the file'),
59 'width' => array('type' => 'int', 'description' => 'width in pixels, if it can be described as such and data is available'),
60 'height' => array('type' => 'int', 'description' => 'height in pixels, if it can be described as such and data is available'),
62 'modified' => array('type' => 'timestamp', 'not null' => true, 'description' => 'date this record was modified'),
64 'primary key' => array('id'),
65 'unique keys' => array(
66 'file_urlhash_key' => array('urlhash'),
69 'file_filehash_idx' => array('filehash'),
74 function isProtected($url) {
75 return 'http://www.facebook.com/login.php' === $url;
79 * Save a new file record.
81 * @param array $redir_data lookup data eg from File_redirection::where()
82 * @param string $given_url
85 public static function saveNew(array $redir_data, $given_url)
89 // I don't know why we have to keep doing this but we run a last check to avoid
91 $file = File::getByUrl($given_url);
93 } catch (NoResultException $e) {
94 // We don't have the file's URL since before, so let's continue.
97 if (!Event::handle('StartFileSaveNew', array(&$redir_data, $given_url))) {
98 throw new ServerException('File not saved due to an aborted StartFileSaveNew event.');
102 $file->urlhash = self::hashurl($given_url);
103 $file->url = $given_url;
104 if (!empty($redir_data['protected'])) $file->protected = $redir_data['protected'];
105 if (!empty($redir_data['title'])) $file->title = $redir_data['title'];
106 if (!empty($redir_data['type'])) $file->mimetype = $redir_data['type'];
107 if (!empty($redir_data['size'])) $file->size = intval($redir_data['size']);
108 if (isset($redir_data['time']) && $redir_data['time'] > 0) $file->date = intval($redir_data['time']);
109 $file_id = $file->insert();
111 if ($file_id === false) {
112 throw new ServerException('File/URL metadata could not be saved to the database.');
115 Event::handle('EndFileSaveNew', array($file, $redir_data, $given_url));
120 * Go look at a URL and possibly save data about it if it's new:
121 * - follow redirect chains and store them in file_redirection
122 * - if a thumbnail is available, save it in file_thumbnail
123 * - save file record with basic info
124 * - optionally save a file_to_post record
125 * - return the File object with the full reference
127 * @fixme refactor this mess, it's gotten pretty scary.
128 * @param string $given_url the URL we're looking at
129 * @param Notice $notice (optional)
130 * @param bool $followRedirects defaults to true
132 * @return mixed File on success, -1 on some errors
134 * @throws ServerException on failure
136 public static function processNew($given_url, Notice $notice=null, $followRedirects=true) {
137 if (empty($given_url)) {
138 throw new ServerException('No given URL to process');
141 $given_url = File_redirection::_canonUrl($given_url);
142 if (empty($given_url)) {
143 throw new ServerException('No canonical URL from given URL to process');
149 $file = File::getByUrl($given_url);
150 } catch (NoResultException $e) {
151 // First check if we have a lookup trace for this URL already
153 $file_redir = File_redirection::getByUrl($given_url);
154 $file = File::getKV('id', $file_redir->file_id);
155 if (!$file instanceof File) {
156 // File did not exist, let's clean up the File_redirection entry
157 $file_redir->delete();
159 } catch (NoResultException $e) {
160 // We just wanted to doublecheck whether a File_thumbnail we might've had
161 // actually referenced an existing File object.
165 // If we still don't have a File object, let's create one now!
166 if (!$file instanceof File) {
167 // @fixme for new URLs this also looks up non-redirect data
168 // such as target content type, size, etc, which we need
169 // for File::saveNew(); so we call it even if not following
171 $redir_data = File_redirection::where($given_url);
172 if (is_array($redir_data)) {
173 $redir_url = $redir_data['url'];
174 } elseif (is_string($redir_data)) {
175 $redir_url = $redir_data;
176 $redir_data = array();
178 // TRANS: Server exception thrown when a URL cannot be processed.
179 throw new ServerException(sprintf(_("Cannot process URL '%s'"), $given_url));
182 if ($redir_url === $given_url || !$followRedirects) {
183 // Save the File object based on our lookup trace
184 $file = File::saveNew($redir_data, $given_url);
186 // This seems kind of messed up... for now skipping this part
187 // if we're already under a redirect, so we don't go into
188 // horrible infinite loops if we've been given an unstable
189 // redirect (where the final destination of the first request
190 // doesn't match what we get when we ask for it again).
192 // Seen in the wild with clojure.org, which redirects through
193 // wikispaces for auth and appends session data in the URL params.
194 $file = self::processNew($redir_url, $notice, /*followRedirects*/false);
195 File_redirection::saveNew($redir_data, $file->id, $given_url);
198 if (!$file instanceof File) {
199 // This should only happen if File::saveNew somehow did not return a File object,
200 // though we have an assert for that in case the event there might've gone wrong.
201 // If anything else goes wrong, there should've been an exception thrown.
202 throw new ServerException('URL processing failed without new File object');
206 if ($notice instanceof Notice) {
207 File_to_post::processNew($file, $notice);
212 public static function respectsQuota(Profile $scoped, $fileSize) {
213 if ($fileSize > common_config('attachments', 'file_quota')) {
214 // TRANS: Message used to be inserted as %2$s in the text "No file may
215 // TRANS: be larger than %1$d byte and the file you sent was %2$s.".
216 // TRANS: %1$d is the number of bytes of an uploaded file.
217 $fileSizeText = sprintf(_m('%1$d byte','%1$d bytes',$fileSize),$fileSize);
219 $fileQuota = common_config('attachments', 'file_quota');
220 // TRANS: Message given if an upload is larger than the configured maximum.
221 // TRANS: %1$d (used for plural) is the byte limit for uploads,
222 // TRANS: %2$s is the proper form of "n bytes". This is the only ways to have
223 // TRANS: gettext support multiple plurals in the same message, unfortunately...
224 throw new ClientException(
225 sprintf(_m('No file may be larger than %1$d byte and the file you sent was %2$s. Try to upload a smaller version.',
226 'No file may be larger than %1$d bytes and the file you sent was %2$s. Try to upload a smaller version.',
228 $fileQuota, $fileSizeText));
233 $query = "select sum(size) as total from file join file_to_post on file_to_post.file_id = file.id join notice on file_to_post.post_id = notice.id where profile_id = {$scoped->id} and file.url like '%/notice/%/file'";
234 $file->query($query);
236 $total = $file->total + $fileSize;
237 if ($total > common_config('attachments', 'user_quota')) {
238 // TRANS: Message given if an upload would exceed user quota.
239 // TRANS: %d (number) is the user quota in bytes and is used for plural.
240 throw new ClientException(
241 sprintf(_m('A file this large would exceed your user quota of %d byte.',
242 'A file this large would exceed your user quota of %d bytes.',
243 common_config('attachments', 'user_quota')),
244 common_config('attachments', 'user_quota')));
246 $query .= ' AND EXTRACT(month FROM file.modified) = EXTRACT(month FROM now()) and EXTRACT(year FROM file.modified) = EXTRACT(year FROM now())';
247 $file->query($query);
249 $total = $file->total + $fileSize;
250 if ($total > common_config('attachments', 'monthly_quota')) {
251 // TRANS: Message given id an upload would exceed a user's monthly quota.
252 // TRANS: $d (number) is the monthly user quota in bytes and is used for plural.
253 throw new ClientException(
254 sprintf(_m('A file this large would exceed your monthly quota of %d byte.',
255 'A file this large would exceed your monthly quota of %d bytes.',
256 common_config('attachments', 'monthly_quota')),
257 common_config('attachments', 'monthly_quota')));
262 public function getFilename()
264 if (!self::validFilename($this->filename)) {
265 // TRANS: Client exception thrown if a file upload does not have a valid name.
266 throw new ClientException(_("Invalid filename."));
268 return $this->filename;
271 // where should the file go?
273 static function filename(Profile $profile, $origname, $mimetype)
275 $ext = self::guessMimeExtension($mimetype);
277 // Normalize and make the original filename more URL friendly.
278 $origname = basename($origname, ".$ext");
279 if (class_exists('Normalizer')) {
280 // http://php.net/manual/en/class.normalizer.php
281 // http://www.unicode.org/reports/tr15/
282 $origname = Normalizer::normalize($origname, Normalizer::FORM_KC);
284 $origname = preg_replace('/[^A-Za-z0-9\.\_]/', '_', $origname);
286 $nickname = $profile->getNickname();
287 $datestamp = strftime('%Y%m%d', time());
289 // generate new random strings until we don't run into a filename collision.
290 $random = strtolower(common_confirmation_code(16));
291 $filename = "$nickname-$datestamp-$origname-$random.$ext";
292 } while (file_exists(self::path($filename)));
296 static function guessMimeExtension($mimetype)
299 $ext = common_supported_mime_to_ext($mimetype);
300 } catch (Exception $e) {
301 // We don't support this mimetype, but let's guess the extension
302 $ext = substr(strrchr($mimetype, '/'), 1);
304 return strtolower($ext);
308 * Validation for as-saved base filenames
310 static function validFilename($filename)
312 return preg_match('/^[A-Za-z0-9._-]+$/', $filename);
316 * @throws ClientException on invalid filename
318 static function path($filename)
320 if (!self::validFilename($filename)) {
321 // TRANS: Client exception thrown if a file upload does not have a valid name.
322 throw new ClientException(_("Invalid filename."));
324 $dir = common_config('attachments', 'dir');
326 if ($dir[strlen($dir)-1] != '/') {
330 return $dir . $filename;
333 static function url($filename)
335 if (!self::validFilename($filename)) {
336 // TRANS: Client exception thrown if a file upload does not have a valid name.
337 throw new ClientException(_("Invalid filename."));
340 if (common_config('site','private')) {
342 return common_local_url('getfile',
343 array('filename' => $filename));
347 if (GNUsocial::useHTTPS()) {
349 $sslserver = common_config('attachments', 'sslserver');
351 if (empty($sslserver)) {
352 // XXX: this assumes that background dir == site dir + /file/
353 // not true if there's another server
354 if (is_string(common_config('site', 'sslserver')) &&
355 mb_strlen(common_config('site', 'sslserver')) > 0) {
356 $server = common_config('site', 'sslserver');
357 } else if (common_config('site', 'server')) {
358 $server = common_config('site', 'server');
360 $path = common_config('site', 'path') . '/file/';
362 $server = $sslserver;
363 $path = common_config('attachments', 'sslpath');
365 $path = common_config('attachments', 'path');
371 $path = common_config('attachments', 'path');
372 $server = common_config('attachments', 'server');
374 if (empty($server)) {
375 $server = common_config('site', 'server');
378 $ssl = common_config('attachments', 'ssl');
380 $protocol = ($ssl) ? 'https' : 'http';
383 if ($path[strlen($path)-1] != '/') {
387 if ($path[0] != '/') {
391 return $protocol.'://'.$server.$path.$filename;
394 function getEnclosure(){
395 $enclosure = (object) array();
396 foreach (array('title', 'url', 'date', 'modified', 'size', 'mimetype') as $key) {
397 $enclosure->$key = $this->$key;
400 $needMoreMetadataMimetypes = array(null, 'application/xhtml+xml');
402 if (!isset($this->filename) && in_array(common_bare_mime($enclosure->mimetype), $needMoreMetadataMimetypes)) {
403 // This fetches enclosure metadata for non-local links with unset/HTML mimetypes,
404 // which may be enriched through oEmbed or similar (implemented as plugins)
405 Event::handle('FileEnclosureMetadata', array($this, &$enclosure));
407 if (empty($enclosure->mimetype) || in_array(common_bare_mime($enclosure->mimetype), $needMoreMetadataMimetypes)) {
408 // This means we either don't know what it is, so it can't
409 // be shown as an enclosure, or it is an HTML link which
410 // does not link to a resource with further metadata.
411 throw new ServerException('Unknown enclosure mimetype, not enough metadata');
417 * Get the attachment's thumbnail record, if any.
418 * Make sure you supply proper 'int' typed variables (or null).
420 * @param $width int Max width of thumbnail in pixels. (if null, use common_config values)
421 * @param $height int Max height of thumbnail in pixels. (if null, square-crop to $width)
422 * @param $crop bool Crop to the max-values' aspect ratio
424 * @return File_thumbnail
426 * @throws UseFileAsThumbnailException if the file is considered an image itself and should be itself as thumbnail
427 * @throws UnsupportedMediaException if, despite trying, we can't understand how to make a thumbnail for this format
428 * @throws ServerException on various other errors
430 public function getThumbnail($width=null, $height=null, $crop=false, $force_still=true)
432 // Get some more information about this file through our ImageFile class
433 $image = ImageFile::fromFileObject($this);
434 if ($image->animated && !common_config('thumbnail', 'animated')) {
435 // null means "always use file as thumbnail"
436 // false means you get choice between frozen frame or original when calling getThumbnail
437 if (is_null(common_config('thumbnail', 'animated')) || !$force_still) {
438 throw new UseFileAsThumbnailException($this->id);
442 return $image->getFileThumbnail($width, $height, $crop);
445 public function getPath()
447 $filepath = self::path($this->filename);
448 if (!file_exists($filepath)) {
449 throw new FileNotFoundException($filepath);
454 public function getUrl()
456 if (!empty($this->filename)) {
457 // A locally stored file, so let's generate a URL for our instance.
458 $url = self::url($this->filename);
459 if (self::hashurl($url) !== $this->urlhash) {
460 // For indexing purposes, in case we do a lookup on the 'url' field.
461 // also we're fixing possible changes from http to https, or paths
462 $this->updateUrl($url);
467 // No local filename available, return the URL we have stored
471 static public function getByUrl($url)
474 $file->urlhash = self::hashurl($url);
475 if (!$file->find(true)) {
476 throw new NoResultException($file);
482 * @param string $hashstr String of (preferrably lower case) hexadecimal characters, same as result of 'hash_file(...)'
484 static public function getByHash($hashstr)
487 $file->filehash = strtolower($hashstr);
488 if (!$file->find(true)) {
489 throw new NoResultException($file);
494 public function updateUrl($url)
496 $file = File::getKV('urlhash', self::hashurl($url));
497 if ($file instanceof File) {
498 throw new ServerException('URL already exists in DB');
500 $sql = 'UPDATE %1$s SET urlhash=%2$s, url=%3$s WHERE urlhash=%4$s;';
501 $result = $this->query(sprintf($sql, $this->tableName(),
502 $this->_quote((string)self::hashurl($url)),
503 $this->_quote((string)$url),
504 $this->_quote((string)$this->urlhash)));
505 if ($result === false) {
506 common_log_db_error($this, 'UPDATE', __FILE__);
507 throw new ServerException("Could not UPDATE {$this->tableName()}.url");
514 * Blow the cache of notices that link to this URL
516 * @param boolean $last Whether to blow the "last" cache too
521 function blowCache($last=false)
523 self::blow('file:notice-ids:%s', $this->id);
525 self::blow('file:notice-ids:%s;last', $this->id);
527 self::blow('file:notice-count:%d', $this->id);
531 * Stream of notices linking to this URL
533 * @param integer $offset Offset to show; default is 0
534 * @param integer $limit Limit of notices to show
535 * @param integer $since_id Since this notice
536 * @param integer $max_id Before this notice
538 * @return array ids of notices that link to this file
541 function stream($offset=0, $limit=NOTICES_PER_PAGE, $since_id=0, $max_id=0)
543 $stream = new FileNoticeStream($this);
544 return $stream->getNotices($offset, $limit, $since_id, $max_id);
547 function noticeCount()
549 $cacheKey = sprintf('file:notice-count:%d', $this->id);
551 $count = self::cacheGet($cacheKey);
553 if ($count === false) {
555 $f2p = new File_to_post();
557 $f2p->file_id = $this->id;
559 $count = $f2p->count();
561 self::cacheSet($cacheKey, $count);
567 public function isLocal()
569 return !empty($this->filename);
572 public function delete($useWhere=false)
574 // Delete the file, if it exists locally
575 if (!empty($this->filename) && file_exists(self::path($this->filename))) {
576 $deleted = @unlink(self::path($this->filename));
578 common_log(LOG_ERR, sprintf('Could not unlink existing file: "%s"', self::path($this->filename)));
582 // Clear out related things in the database and filesystem, such as thumbnails
583 if (Event::handle('FileDeleteRelated', array($this))) {
584 $thumbs = new File_thumbnail();
585 $thumbs->file_id = $this->id;
586 if ($thumbs->find()) {
587 while ($thumbs->fetch()) {
592 $f2p = new File_to_post();
593 $f2p->file_id = $this->id;
595 while ($f2p->fetch()) {
601 // And finally remove the entry from the database
602 return parent::delete($useWhere);
605 public function getTitle()
607 $title = $this->title ?: $this->filename;
609 return $title ?: null;
612 static public function hashurl($url)
615 throw new Exception('No URL provided to hash algorithm.');
617 return hash(self::URLHASH_ALG, $url);
620 static public function beforeSchemaUpdate()
622 $table = strtolower(get_called_class());
623 $schema = Schema::get();
624 $schemadef = $schema->getTableDef($table);
626 // 2015-02-19 We have to upgrade our table definitions to have the urlhash field populated
627 if (isset($schemadef['fields']['urlhash']) && isset($schemadef['unique keys']['file_urlhash_key'])) {
628 // We already have the urlhash field, so no need to migrate it.
631 echo "\nFound old $table table, upgrading it to contain 'urlhash' field...";
634 $file->query(sprintf('SELECT id, LEFT(url, 191) AS shortenedurl, COUNT(*) AS c FROM %1$s WHERE LENGTH(url)>191 GROUP BY shortenedurl HAVING c > 1', $schema->quoteIdentifier($table)));
635 print "\nFound {$file->N} URLs with too long entries in file table\n";
636 while ($file->fetch()) {
637 // We've got a URL that is too long for our future file table
638 // so we'll cut it. We could save the original URL, but there is
639 // no guarantee it is complete anyway since the previous max was 255 chars.
640 $dupfile = new File();
641 // First we find file entries that would be duplicates of this when shortened
642 // ... and we'll just throw the dupes out the window for now! It's already so borken.
643 $dupfile->query(sprintf('SELECT * FROM file WHERE LEFT(url, 191) = "%1$s"', $file->shortenedurl));
644 // Leave one of the URLs in the database by using ->find(true) (fetches first entry)
645 if ($dupfile->find(true)) {
646 print "\nShortening url entry for $table id: {$file->id} [";
647 $orig = clone($dupfile);
648 $dupfile->url = $file->shortenedurl; // make sure it's only 191 chars from now on
649 $dupfile->update($orig);
650 print "\nDeleting duplicate entries of too long URL on $table id: {$file->id} [";
651 // only start deleting with this fetch.
652 while($dupfile->fetch()) {
658 print "\nWarning! URL suddenly disappeared from database: {$file->url}\n";
661 echo "...and now all the non-duplicates which are longer than 191 characters...\n";
662 $file->query('UPDATE file SET url=LEFT(url, 191) WHERE LENGTH(url)>191');
664 echo "\n...now running hacky pre-schemaupdate change for $table:";
665 // We have to create a urlhash that is _not_ the primary key,
666 // transfer data and THEN run checkSchema
667 $schemadef['fields']['urlhash'] = array (
670 'not null' => false, // this is because when adding column, all entries will _be_ NULL!
671 'description' => 'sha256 of destination URL (url field)',
673 $schemadef['fields']['url'] = array (
675 'description' => 'destination URL after following possible redirections',
677 unset($schemadef['unique keys']);
678 $schema->ensureTable($table, $schemadef);
681 $classname = ucfirst($table);
682 $tablefix = new $classname;
683 // urlhash is hash('sha256', $url) in the File table
684 echo "Updating urlhash fields in $table table...";
685 // Maybe very MySQL specific :(
686 $tablefix->query(sprintf('UPDATE %1$s SET %2$s=%3$s;',
687 $schema->quoteIdentifier($table),
689 // The line below is "result of sha256 on column `url`"
692 echo "Resuming core schema upgrade...";