3 * GNU social - a federating social network
5 * Abstraction for files
7 * LICENCE: This program is free software: you can redistribute it and/or modify
8 * it under the terms of the GNU Affero General Public License as published by
9 * the Free Software Foundation, either version 3 of the License, or
10 * (at your option) any later version.
12 * This program is distributed in the hope that it will be useful,
13 * but WITHOUT ANY WARRANTY; without even the implied warranty of
14 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
15 * GNU Affero General Public License for more details.
17 * You should have received a copy of the GNU Affero General Public License
18 * along with this program. If not, see <http://www.gnu.org/licenses/>.
22 * @author Mikael Nordfeldth <mmn@hethane.se>
23 * @author Miguel Dantas <biodantas@gmail.com>
24 * @copyright 2008-2009, 2019 Free Software Foundation http://fsf.org
25 * @license http://www.fsf.org/licensing/licenses/agpl-3.0.html GNU Affero General Public License version 3.0
26 * @link https://www.gnu.org/software/social/
29 defined('GNUSOCIAL') || die();
32 * Table Definition for file
34 class File extends Managed_DataObject
36 public $__table = 'file'; // table name
37 public $id; // int(4) primary_key not_null
38 public $urlhash; // varchar(64) unique_key
40 public $filehash; // varchar(64) indexed
41 public $mimetype; // varchar(50)
42 public $size; // int(4)
43 public $title; // text()
44 public $date; // int(4)
45 public $protected; // int(4)
46 public $filename; // text()
47 public $width; // int(4)
48 public $height; // int(4)
49 public $modified; // datetime() not_null default_CURRENT_TIMESTAMP
51 const URLHASH_ALG = 'sha256';
52 const FILEHASH_ALG = 'sha256';
54 public static function schemaDef()
58 'id' => array('type' => 'serial', 'not null' => true),
59 'urlhash' => array('type' => 'varchar', 'length' => 64, 'not null' => true, 'description' => 'sha256 of destination URL (url field)'),
60 'url' => array('type' => 'text', 'description' => 'destination URL after following possible redirections'),
61 'filehash' => array('type' => 'varchar', 'length' => 64, 'not null' => false, 'description' => 'sha256 of the file contents, only for locally stored files of course'),
62 'mimetype' => array('type' => 'varchar', 'length' => 50, 'description' => 'mime type of resource'),
63 'size' => array('type' => 'int', 'description' => 'size of resource when available'),
64 'title' => array('type' => 'text', 'description' => 'title of resource when available'),
65 'date' => array('type' => 'int', 'description' => 'date of resource according to http query'),
66 'protected' => array('type' => 'int', 'description' => 'true when URL is private (needs login)'),
67 'filename' => array('type' => 'text', 'description' => 'if file is stored locally (too) this is the filename'),
68 'width' => array('type' => 'int', 'description' => 'width in pixels, if it can be described as such and data is available'),
69 'height' => array('type' => 'int', 'description' => 'height in pixels, if it can be described as such and data is available'),
70 'modified' => array('type' => 'datetime', 'not null' => true, 'default' => 'CURRENT_TIMESTAMP', 'description' => 'date this record was modified'),
72 'primary key' => array('id'),
73 'unique keys' => array(
74 'file_urlhash_key' => array('urlhash'),
77 'file_filehash_idx' => array('filehash'),
82 public static function isProtected($url)
84 $protected_urls_exps = array(
85 'https://www.facebook.com/login.php',
86 common_path('main/login')
89 foreach ($protected_urls_exps as $protected_url_exp) {
90 if (preg_match('!^'.preg_quote($protected_url_exp).'(.*)$!i', $url) === 1) {
99 * Save a new file record.
101 * @param array $redir_data lookup data eg from File_redirection::where()
102 * @param string $given_url
104 * @throws ServerException
106 public static function saveNew(array $redir_data, $given_url)
110 // I don't know why we have to keep doing this but we run a last check to avoid
112 $file = File::getByUrl($given_url);
114 } catch (NoResultException $e) {
115 // We don't have the file's URL since before, so let's continue.
118 // if the given url is an local attachment url and the id already exists, don't
119 // save a new file record. This should never happen, but let's make it foolproof
120 // FIXME: how about attachments servers?
121 $u = parse_url($given_url);
122 if (isset($u['host']) && $u['host'] === common_config('site', 'server')) {
124 // Skip the / in the beginning or $r->map won't match
126 $args = $r->map(mb_substr($u['path'], 1));
127 if ($args['action'] === 'attachment') {
129 if (!empty($args['attachment'])) {
130 return File::getByID($args['attachment']);
131 } elseif ($args['filehash']) {
132 return File::getByHash($args['filehash']);
134 } catch (NoResultException $e) {
135 // apparently this link goes to us, but is _not_ an existing attachment (File) ID?
138 } catch (Exception $e) {
139 // Some other exception was thrown from $r->map, likely a
140 // ClientException (404) because of some malformed link to
141 // our own instance. It's still a valid URL however, so we
142 // won't abort anything... I noticed this when linking:
143 // https://social.umeahackerspace.se/mmn/foaf' (notice the
144 // apostrophe in the end, making it unrecognizable for our
146 // That specific issue (the apostrophe being part of a link
147 // is something that may or may not have been fixed since,
148 // in lib/util.php in common_replace_urls_callback().
153 $file->url = $given_url;
154 if (!empty($redir_data['protected'])) {
155 $file->protected = $redir_data['protected'];
157 if (!empty($redir_data['title'])) {
158 $file->title = $redir_data['title'];
160 if (!empty($redir_data['type'])) {
161 $file->mimetype = $redir_data['type'];
163 if (!empty($redir_data['size'])) {
164 $file->size = intval($redir_data['size']);
166 if (isset($redir_data['time']) && $redir_data['time'] > 0) {
167 $file->date = intval($redir_data['time']);
173 public function saveFile()
175 $this->urlhash = self::hashurl($this->url);
177 if (!Event::handle('StartFileSaveNew', array(&$this))) {
178 throw new ServerException('File not saved due to an aborted StartFileSaveNew event.');
181 $this->id = $this->insert();
183 if ($this->id === false) {
184 throw new ServerException('File/URL metadata could not be saved to the database.');
187 Event::handle('EndFileSaveNew', array($this));
191 * Go look at a URL and possibly save data about it if it's new:
192 * - follow redirect chains and store them in file_redirection
193 * - if a thumbnail is available, save it in file_thumbnail
194 * - save file record with basic info
195 * - optionally save a file_to_post record
196 * - return the File object with the full reference
198 * @param string $given_url the URL we're looking at
199 * @param Notice $notice (optional)
200 * @param bool $followRedirects defaults to true
202 * @return mixed File on success, -1 on some errors
204 * @throws ServerException on failure
206 public static function processNew($given_url, Notice $notice=null, $followRedirects=true)
208 if (empty($given_url)) {
209 throw new ServerException('No given URL to process');
212 $given_url = File_redirection::_canonUrl($given_url);
213 if (empty($given_url)) {
214 throw new ServerException('No canonical URL from given URL to process');
217 $redir = File_redirection::where($given_url);
219 $file = $redir->getFile();
220 } catch (EmptyPkeyValueException $e) {
221 common_log(LOG_ERR, 'File_redirection::where gave object with empty file_id for given_url '._ve($given_url));
222 throw new ServerException('URL processing failed without new File object');
223 } catch (NoResultException $e) {
224 // This should not happen
225 common_log(LOG_ERR, 'File_redirection after discovery could still not return a File object.');
226 throw new ServerException('URL processing failed without new File object');
229 if ($notice instanceof Notice) {
230 File_to_post::processNew($file, $notice);
236 public static function respectsQuota(Profile $scoped, $fileSize)
238 if ($fileSize > common_config('attachments', 'file_quota')) {
239 // TRANS: Message used to be inserted as %2$s in the text "No file may
240 // TRANS: be larger than %1$d byte and the file you sent was %2$s.".
241 // TRANS: %1$d is the number of bytes of an uploaded file.
242 $fileSizeText = sprintf(_m('%1$d byte', '%1$d bytes', $fileSize), $fileSize);
244 $fileQuota = common_config('attachments', 'file_quota');
245 // TRANS: Message given if an upload is larger than the configured maximum.
246 // TRANS: %1$d (used for plural) is the byte limit for uploads,
247 // TRANS: %2$s is the proper form of "n bytes". This is the only ways to have
248 // TRANS: gettext support multiple plurals in the same message, unfortunately...
249 throw new ClientException(
252 'No file may be larger than %1$d byte and the file you sent was %2$s. Try to upload a smaller version.',
253 'No file may be larger than %1$d bytes and the file you sent was %2$s. Try to upload a smaller version.',
264 $query = "select sum(size) as total from file join file_to_post on file_to_post.file_id = file.id join notice on file_to_post.post_id = notice.id where profile_id = {$scoped->id} and file.url like '%/notice/%/file'";
265 $file->query($query);
267 $total = $file->total + $fileSize;
268 if ($total > common_config('attachments', 'user_quota')) {
269 // TRANS: Message given if an upload would exceed user quota.
270 // TRANS: %d (number) is the user quota in bytes and is used for plural.
271 throw new ClientException(
274 'A file this large would exceed your user quota of %d byte.',
275 'A file this large would exceed your user quota of %d bytes.',
276 common_config('attachments', 'user_quota')
278 common_config('attachments', 'user_quota')
282 $query .= ' AND EXTRACT(month FROM file.modified) = EXTRACT(month FROM now()) and EXTRACT(year FROM file.modified) = EXTRACT(year FROM now())';
283 $file->query($query);
285 $total = $file->total + $fileSize;
286 if ($total > common_config('attachments', 'monthly_quota')) {
287 // TRANS: Message given id an upload would exceed a user's monthly quota.
288 // TRANS: $d (number) is the monthly user quota in bytes and is used for plural.
289 throw new ClientException(
292 'A file this large would exceed your monthly quota of %d byte.',
293 'A file this large would exceed your monthly quota of %d bytes.',
294 common_config('attachments', 'monthly_quota')
296 common_config('attachments', 'monthly_quota')
303 public function getFilename()
305 return self::tryFilename($this->filename);
308 public function getSize()
310 return intval($this->size);
313 // where should the file go?
315 public static function filename(Profile $profile, $origname, $mimetype)
317 $ext = self::guessMimeExtension($mimetype, $origname);
319 // Normalize and make the original filename more URL friendly.
320 $origname = basename($origname, ".$ext");
321 if (class_exists('Normalizer')) {
322 // http://php.net/manual/en/class.normalizer.php
323 // http://www.unicode.org/reports/tr15/
324 $origname = Normalizer::normalize($origname, Normalizer::FORM_KC);
326 $origname = preg_replace('/[^A-Za-z0-9\.\_]/', '_', $origname);
328 $nickname = $profile->getNickname();
329 $datestamp = strftime('%Y%m%d', time());
331 // generate new random strings until we don't run into a filename collision.
332 $random = strtolower(common_confirmation_code(16));
333 $filename = "$nickname-$datestamp-$origname-$random.$ext";
334 } while (file_exists(self::path($filename)));
339 * @param string $filename
340 * @return string|bool Value from the 'extblacklist' array, in the config
342 public static function getSafeExtension(string $filename) {
343 if (preg_match('/^.+?\.([A-Za-z0-9]+)$/', $filename, $matches) === 1) {
344 // we matched on a file extension, so let's see if it means something.
345 $ext = mb_strtolower($matches[1]);
346 $blacklist = common_config('attachments', 'extblacklist');
347 // If we got an extension from $filename we want to check if it's in a blacklist
348 // so we avoid people uploading restricted files
349 if (array_key_exists($ext, $blacklist)) {
350 if (!is_string($blacklist[$ext])) {
354 // return a safe replacement extension ('php' => 'phps' for example)
355 return $blacklist[$ext];
357 // the attachment extension based on its filename was not blacklisted so it's ok to use it
367 * @param $mimetype string The mimetype we've discovered for this file.
368 * @param $filename string An optional filename which we can use on failure.
369 * @return mixed|string
370 * @throws ClientException
372 public static function guessMimeExtension($mimetype, $filename=null)
375 // first see if we know the extension for our mimetype
376 $ext = common_supported_mime_to_ext($mimetype);
379 } catch (UnknownMimeExtensionException $e) {
380 // We don't know the extension for this mimetype, but let's guess.
381 // If we can't recognize the extension from the MIME, we try
382 // to guess based on filename, if one was supplied.
383 if (!is_null($filename)) {
384 $ext = getSafeExtension($filename);
385 if ($ext === false) {
386 // we don't have a safe replacement extension
387 throw new ClientException(_('Blacklisted file extension.'));
392 } catch (Exception $e) {
393 common_log(LOG_INFO, 'Problem when figuring out extension for mimetype: '._ve($e));
396 // If nothing else has given us a result, try to extract it from
397 // the mimetype value (this turns .jpg to .jpeg for example...)
399 // Will get jpeg from image/jpeg as well as json from application/jrd+json
400 if (!preg_match('/[\/+-\.]([a-z0-9]+)/', mb_strtolower($mimetype), $matches)) {
401 throw new Exception("Malformed mimetype: {$mimetype}");
403 return mb_strtolower($matches[1]);
407 * Validation for as-saved base filenames
411 public static function validFilename($filename)
413 return preg_match('/^[A-Za-z0-9._-]+$/', $filename);
416 public static function tryFilename($filename)
418 if (!self::validFilename($filename)) {
419 throw new InvalidFilenameException($filename);
421 // if successful, return the filename for easy if-statementing
428 * @throws InvalidFilenameException
430 public static function path($filename)
432 self::tryFilename($filename);
434 $dir = common_config('attachments', 'dir');
436 if (!in_array($dir[mb_strlen($dir)-1], ['/', '\\'])) {
437 $dir .= DIRECTORY_SEPARATOR;
440 return $dir . $filename;
443 public static function url($filename)
445 self::tryFilename($filename);
447 if (common_config('site', 'private')) {
448 return common_local_url(
450 array('filename' => $filename)
454 if (GNUsocial::useHTTPS()) {
455 $sslserver = common_config('attachments', 'sslserver');
457 if (empty($sslserver)) {
458 // XXX: this assumes that background dir == site dir + /file/
459 // not true if there's another server
460 if (is_string(common_config('site', 'sslserver')) &&
461 mb_strlen(common_config('site', 'sslserver')) > 0) {
462 $server = common_config('site', 'sslserver');
463 } elseif (common_config('site', 'server')) {
464 $server = common_config('site', 'server');
466 $path = common_config('site', 'path') . '/file/';
468 $server = $sslserver;
469 $path = common_config('attachments', 'sslpath');
471 $path = common_config('attachments', 'path');
477 $path = common_config('attachments', 'path');
478 $server = common_config('attachments', 'server');
480 if (empty($server)) {
481 $server = common_config('site', 'server');
484 $ssl = common_config('attachments', 'ssl');
486 $protocol = ($ssl) ? 'https' : 'http';
489 if ($path[strlen($path)-1] != '/') {
493 if ($path[0] != '/') {
497 return $protocol.'://'.$server.$path.$filename;
500 public static $_enclosures = array();
502 public function getEnclosure()
504 if (isset(self::$_enclosures[$this->getID()])) {
505 return self::$_enclosures[$this->getID()];
508 $enclosure = (object) array();
509 foreach (array('title', 'url', 'date', 'modified', 'size', 'mimetype', 'width', 'height') as $key) {
510 if ($this->$key !== '') {
511 $enclosure->$key = $this->$key;
515 $needMoreMetadataMimetypes = array(null, 'application/xhtml+xml', 'text/html');
517 if (!isset($this->filename) && in_array(common_bare_mime($enclosure->mimetype), $needMoreMetadataMimetypes)) {
518 // This fetches enclosure metadata for non-local links with unset/HTML mimetypes,
519 // which may be enriched through oEmbed or similar (implemented as plugins)
520 Event::handle('FileEnclosureMetadata', array($this, &$enclosure));
522 if (empty($enclosure->mimetype)) {
523 // This means we either don't know what it is, so it can't
524 // be shown as an enclosure, or it is an HTML link which
525 // does not link to a resource with further metadata.
526 throw new ServerException('Unknown enclosure mimetype, not enough metadata');
529 self::$_enclosures[$this->getID()] = $enclosure;
533 public function hasThumbnail()
536 $this->getThumbnail();
537 } catch (Exception $e) {
544 * Get the attachment's thumbnail record, if any.
545 * Make sure you supply proper 'int' typed variables (or null).
547 * @param $width int Max width of thumbnail in pixels. (if null, use common_config values)
548 * @param $height int Max height of thumbnail in pixels. (if null, square-crop to $width)
549 * @param $crop bool Crop to the max-values' aspect ratio
550 * @param $force_still bool Don't allow fallback to showing original (such as animated GIF)
551 * @param $upscale mixed Whether or not to scale smaller images up to larger thumbnail sizes. (null = site default)
553 * @return File_thumbnail
555 * @throws UseFileAsThumbnailException if the file is considered an image itself and should be itself as thumbnail
556 * @throws UnsupportedMediaException if, despite trying, we can't understand how to make a thumbnail for this format
557 * @throws ServerException on various other errors
559 public function getThumbnail($width=null, $height=null, $crop=false, $force_still=true, $upscale=null)
561 // Get some more information about this file through our ImageFile class
562 $image = ImageFile::fromFileObject($this);
563 if ($image->animated && !common_config('thumbnail', 'animated')) {
564 // null means "always use file as thumbnail"
565 // false means you get choice between frozen frame or original when calling getThumbnail
566 if (is_null(common_config('thumbnail', 'animated')) || !$force_still) {
568 // remote files with animated GIFs as thumbnails will match this
569 return File_thumbnail::byFile($this);
570 } catch (NoResultException $e) {
571 // and if it's not a remote file, it'll be safe to use the locally stored File
572 throw new UseFileAsThumbnailException($this);
577 return $image->getFileThumbnail(
581 !is_null($upscale) ? $upscale : common_config('thumbnail', 'upscale')
585 public function getPath()
587 $filepath = self::path($this->filename);
588 if (!file_exists($filepath)) {
589 throw new FileNotFoundException($filepath);
595 * Returns the path to either a file, or it's thumbnail if the file doesn't exist.
596 * This is useful in case the original file is deleted, or, as is the case for Embed
597 * thumbnails, we only have a thumbnail and not a file
598 * @return string Path
599 * @throws FileNotFoundException
600 * @throws FileNotStoredLocallyException
601 * @throws InvalidFilenameException
602 * @throws ServerException
604 public function getFileOrThumbnailPath($thumbnail = null) : string
606 if (!empty($thumbnail)) {
607 return $thumbnail->getPath();
609 if (!empty($this->filename)) {
610 $filepath = self::path($this->filename);
611 if (file_exists($filepath)) {
614 throw new FileNotFoundException($filepath);
618 return File_thumbnail::byFile($this, true)->getPath();
619 } catch (NoResultException $e) {
620 // File not stored locally
621 throw new FileNotStoredLocallyException($this);
627 * Return the mime type of the thumbnail if we have it, or, if not, of the File
629 * @throws FileNotFoundException
630 * @throws NoResultException
631 * @throws ServerException
632 * @throws UnsupportedMediaException
634 public function getFileOrThumbnailMimetype($thumbnail = null) : string
636 if (!empty($thumbnail)) {
637 $filepath = $thumbnail->getPath();
638 } elseif (!empty($this->filename)) {
639 return $this->mimetype;
641 $filepath = File_thumbnail::byFile($this, true)->getPath();
644 $info = @getimagesize($filepath);
645 if ($info !== false) {
646 return $info['mime'];
648 throw new UnsupportedMediaException(_("Thumbnail is not an image."));
653 * Return the size of the thumbnail if we have it, or, if not, of the File
655 * @throws FileNotFoundException
656 * @throws NoResultException
657 * @throws ServerException
659 public function getFileOrThumbnailSize($thumbnail = null) : int
661 if (!empty($thumbnail)) {
662 return filesize($thumbnail->getPath());
663 } elseif (!empty($this->filename)) {
666 return filesize(File_thumbnail::byFile($this)->getPath());
670 public function getAttachmentUrl()
672 return common_local_url('attachment', array('attachment'=>$this->getID()));
675 public function getAttachmentDownloadUrl()
677 return common_local_url('attachment_download', array('attachment'=>$this->getID()));
680 public function getAttachmentViewUrl()
682 return common_local_url('attachment_view', array('attachment'=>$this->getID()));
686 * @param mixed $use_local true means require local, null means prefer local, false means use whatever is stored
688 * @throws FileNotStoredLocallyException
690 public function getUrl($use_local=null)
692 if ($use_local !== false) {
693 if (is_string($this->filename) || !empty($this->filename)) {
694 // A locally stored file, so let's generate a URL for our instance.
695 return $this->getAttachmentViewUrl();
698 // if the file wasn't stored locally (has filename) and we require a local URL
699 throw new FileNotStoredLocallyException($this);
704 // No local filename available, return the URL we have stored
708 public static function getByUrl($url)
711 $file->urlhash = self::hashurl($url);
712 if (!$file->find(true)) {
713 throw new NoResultException($file);
719 * @param string $hashstr String of (preferrably lower case) hexadecimal characters, same as result of 'hash_file(...)'
721 * @throws NoResultException
723 public static function getByHash($hashstr)
726 $file->filehash = strtolower($hashstr);
727 if (!$file->find(true)) {
728 throw new NoResultException($file);
733 public function updateUrl($url)
735 $file = File::getKV('urlhash', self::hashurl($url));
736 if ($file instanceof File) {
737 throw new ServerException('URL already exists in DB');
739 $sql = 'UPDATE %1$s SET urlhash=%2$s, url=%3$s WHERE urlhash=%4$s;';
740 $result = $this->query(sprintf(
743 $this->_quote((string)self::hashurl($url)),
744 $this->_quote((string)$url),
745 $this->_quote((string)$this->urlhash)
747 if ($result === false) {
748 common_log_db_error($this, 'UPDATE', __FILE__);
749 throw new ServerException("Could not UPDATE {$this->tableName()}.url");
756 * Blow the cache of notices that link to this URL
758 * @param boolean $last Whether to blow the "last" cache too
763 public function blowCache($last=false)
765 self::blow('file:notice-ids:%s', $this->id);
767 self::blow('file:notice-ids:%s;last', $this->id);
769 self::blow('file:notice-count:%d', $this->id);
773 * Stream of notices linking to this URL
775 * @param integer $offset Offset to show; default is 0
776 * @param integer $limit Limit of notices to show
777 * @param integer $since_id Since this notice
778 * @param integer $max_id Before this notice
780 * @return array ids of notices that link to this file
783 public function stream($offset=0, $limit=NOTICES_PER_PAGE, $since_id=0, $max_id=0)
785 // FIXME: Try to get the Profile::current() here in some other way to avoid mixing
786 // the current session user with possibly background/queue processing.
787 $stream = new FileNoticeStream($this, Profile::current());
788 return $stream->getNotices($offset, $limit, $since_id, $max_id);
791 public function noticeCount()
793 $cacheKey = sprintf('file:notice-count:%d', $this->id);
795 $count = self::cacheGet($cacheKey);
797 if ($count === false) {
798 $f2p = new File_to_post();
800 $f2p->file_id = $this->id;
802 $count = $f2p->count();
804 self::cacheSet($cacheKey, $count);
810 public function isLocal()
812 return !empty($this->filename);
815 public function delete($useWhere=false)
817 // Delete the file, if it exists locally
818 if (!empty($this->filename) && file_exists(self::path($this->filename))) {
819 $deleted = @unlink(self::path($this->filename));
821 common_log(LOG_ERR, sprintf('Could not unlink existing file: "%s"', self::path($this->filename)));
825 // Clear out related things in the database and filesystem, such as thumbnails
826 if (Event::handle('FileDeleteRelated', array($this))) {
827 $thumbs = new File_thumbnail();
828 $thumbs->file_id = $this->id;
829 if ($thumbs->find()) {
830 while ($thumbs->fetch()) {
835 $f2p = new File_to_post();
836 $f2p->file_id = $this->id;
838 while ($f2p->fetch()) {
844 // And finally remove the entry from the database
845 return parent::delete($useWhere);
848 public function getTitle()
850 $title = $this->title ?: MediaFile::getDisplayName($this);
852 return $title ?: null;
855 public function setTitle($title)
857 $orig = clone($this);
858 $this->title = mb_strlen($title) > 0 ? $title : null;
859 return $this->update($orig);
862 public static function hashurl($url)
865 throw new Exception('No URL provided to hash algorithm.');
867 return hash(self::URLHASH_ALG, $url);
870 public static function beforeSchemaUpdate()
872 $table = strtolower(get_called_class());
873 $schema = Schema::get();
874 $schemadef = $schema->getTableDef($table);
876 // 2015-02-19 We have to upgrade our table definitions to have the urlhash field populated
877 if (isset($schemadef['fields']['urlhash']) && isset($schemadef['unique keys']['file_urlhash_key'])) {
878 // We already have the urlhash field, so no need to migrate it.
881 echo "\nFound old $table table, upgrading it to contain 'urlhash' field...";
884 $file->query(sprintf('SELECT id, LEFT(url, 191) AS shortenedurl, COUNT(*) AS c FROM %1$s WHERE LENGTH(url)>191 GROUP BY shortenedurl HAVING c > 1', $schema->quoteIdentifier($table)));
885 print "\nFound {$file->N} URLs with too long entries in file table\n";
886 while ($file->fetch()) {
887 // We've got a URL that is too long for our future file table
888 // so we'll cut it. We could save the original URL, but there is
889 // no guarantee it is complete anyway since the previous max was 255 chars.
890 $dupfile = new File();
891 // First we find file entries that would be duplicates of this when shortened
892 // ... and we'll just throw the dupes out the window for now! It's already so borken.
893 $dupfile->query(sprintf('SELECT * FROM file WHERE LEFT(url, 191) = %1$s', $dupfile->_quote($file->shortenedurl)));
894 // Leave one of the URLs in the database by using ->find(true) (fetches first entry)
895 if ($dupfile->find(true)) {
896 print "\nShortening url entry for $table id: {$file->id} [";
897 $orig = clone($dupfile);
898 $origurl = $dupfile->url; // save for logging purposes
899 $dupfile->url = $file->shortenedurl; // make sure it's only 191 chars from now on
900 $dupfile->update($orig);
901 print "\nDeleting duplicate entries of too long URL on $table id: {$file->id} [";
902 // only start deleting with this fetch.
903 while ($dupfile->fetch()) {
904 common_log(LOG_INFO, sprintf('Deleting duplicate File entry of %1$d: %2$d (original URL: %3$s collides with these first 191 characters: %4$s', $dupfile->id, $file->id, $origurl, $file->shortenedurl));
910 print "\nWarning! URL suddenly disappeared from database: {$file->url}\n";
913 echo "...and now all the non-duplicates which are longer than 191 characters...\n";
914 $file->query('UPDATE file SET url=LEFT(url, 191) WHERE LENGTH(url)>191');
916 echo "\n...now running hacky pre-schemaupdate change for $table:";
917 // We have to create a urlhash that is _not_ the primary key,
918 // transfer data and THEN run checkSchema
919 $schemadef['fields']['urlhash'] = array(
922 'not null' => false, // this is because when adding column, all entries will _be_ NULL!
923 'description' => 'sha256 of destination URL (url field)',
925 $schemadef['fields']['url'] = array(
927 'description' => 'destination URL after following possible redirections',
929 unset($schemadef['unique keys']);
930 $schema->ensureTable($table, $schemadef);
933 $classname = ucfirst($table);
934 $tablefix = new $classname;
935 // urlhash is hash('sha256', $url) in the File table
936 echo "Updating urlhash fields in $table table...";
937 // Maybe very MySQL specific :(
938 $tablefix->query(sprintf(
939 'UPDATE %1$s SET %2$s=%3$s;',
940 $schema->quoteIdentifier($table),
942 // The line below is "result of sha256 on column `url`"
946 echo "Resuming core schema upgrade...";