X-Git-Url: https://git.mxchange.org/?a=blobdiff_plain;f=classes%2FFile.php;h=1a29ea5107e537dec350188c4b7b323e523b2e76;hb=be0c10e8f6f9c60b1cb840cda9ac6f42c8289738;hp=b5bcec404168b8bbf9fbabfc3d58b69d6257ef93;hpb=adf896bc12f7578196f3a93301e6777ef08174db;p=quix0rs-gnu-social.git diff --git a/classes/File.php b/classes/File.php index b5bcec4041..1a29ea5107 100644 --- a/classes/File.php +++ b/classes/File.php @@ -26,29 +26,36 @@ class File extends Managed_DataObject { public $__table = 'file'; // table name public $id; // int(4) primary_key not_null - public $url; // varchar(255) unique_key + public $urlhash; // varchar(64) unique_key + public $url; // text + public $filehash; // varchar(64) indexed public $mimetype; // varchar(50) public $size; // int(4) - public $title; // varchar(255) + public $title; // varchar(191) not 255 because utf8mb4 takes more space public $date; // int(4) public $protected; // int(4) - public $filename; // varchar(255) + public $filename; // varchar(191) not 255 because utf8mb4 takes more space public $width; // int(4) public $height; // int(4) public $modified; // timestamp() not_null default_CURRENT_TIMESTAMP + const URLHASH_ALG = 'sha256'; + const FILEHASH_ALG = 'sha256'; + public static function schemaDef() { return array( 'fields' => array( 'id' => array('type' => 'serial', 'not null' => true), - 'url' => array('type' => 'varchar', 'length' => 255, 'description' => 'destination URL after following redirections'), + 'urlhash' => array('type' => 'varchar', 'length' => 64, 'not null' => true, 'description' => 'sha256 of destination URL (url field)'), + 'url' => array('type' => 'text', 'description' => 'destination URL after following possible redirections'), + 'filehash' => array('type' => 'varchar', 'length' => 64, 'not null' => false, 'description' => 'sha256 of the file contents, only for locally stored files of course'), 'mimetype' => array('type' => 'varchar', 'length' => 50, 'description' => 'mime type of resource'), 'size' => array('type' => 'int', 'description' => 'size of resource when available'), - 'title' => array('type' => 'varchar', 'length' => 255, 'description' => 'title of resource when available'), + 'title' => array('type' => 'varchar', 'length' => 191, 'description' => 'title of resource when available'), 'date' => array('type' => 'int', 'description' => 'date of resource according to http query'), 'protected' => array('type' => 'int', 'description' => 'true when URL is private (needs login)'), - 'filename' => array('type' => 'varchar', 'length' => 255, 'description' => 'if a local file, name of the file'), + 'filename' => array('type' => 'varchar', 'length' => 191, 'description' => 'if a local file, name of the file'), 'width' => array('type' => 'int', 'description' => 'width in pixels, if it can be described as such and data is available'), 'height' => array('type' => 'int', 'description' => 'height in pixels, if it can be described as such and data is available'), @@ -56,7 +63,10 @@ class File extends Managed_DataObject ), 'primary key' => array('id'), 'unique keys' => array( - 'file_url_key' => array('url'), + 'file_urlhash_key' => array('urlhash'), + ), + 'indexes' => array( + 'file_filehash_idx' => array('filehash'), ), ); } @@ -77,10 +87,11 @@ class File extends Managed_DataObject // I don't know why we have to keep doing this but I'm adding this last check to avoid // uniqueness bugs. - $file = File::getKV('url', $given_url); + $file = File::getKV('urlhash', self::hashurl($given_url)); if (!$file instanceof File) { $file = new File; + $file->urlhash = self::hashurl($given_url); $file->url = $given_url; if (!empty($redir_data['protected'])) $file->protected = $redir_data['protected']; if (!empty($redir_data['title'])) $file->title = $redir_data['title']; @@ -105,14 +116,14 @@ class File extends Managed_DataObject * * @fixme refactor this mess, it's gotten pretty scary. * @param string $given_url the URL we're looking at - * @param int $notice_id (optional) + * @param Notice $notice (optional) * @param bool $followRedirects defaults to true * * @return mixed File on success, -1 on some errors * * @throws ServerException on failure */ - public static function processNew($given_url, $notice_id=null, $followRedirects=true) { + public static function processNew($given_url, Notice $notice=null, $followRedirects=true) { if (empty($given_url)) { throw new ServerException('No given URL to process'); } @@ -122,51 +133,56 @@ class File extends Managed_DataObject throw new ServerException('No canonical URL from given URL to process'); } - $file = File::getKV('url', $given_url); - if (!$file instanceof File) { + $file = null; + + try { + $file = File::getByUrl($given_url); + } catch (NoResultException $e) { // First check if we have a lookup trace for this URL already - $file_redir = File_redirection::getKV('url', $given_url); - if ($file_redir instanceof File_redirection) { + try { + $file_redir = File_redirection::getByUrl($given_url); $file = File::getKV('id', $file_redir->file_id); if (!$file instanceof File) { // File did not exist, let's clean up the File_redirection entry $file_redir->delete(); } + } catch (NoResultException $e) { + // We just wanted to doublecheck whether a File_thumbnail we might've had + // actually referenced an existing File object. } + } - // If we still don't have a File object, let's create one now! - if (!$file instanceof File) { - // @fixme for new URLs this also looks up non-redirect data - // such as target content type, size, etc, which we need - // for File::saveNew(); so we call it even if not following - // new redirects. - $redir_data = File_redirection::where($given_url); - if (is_array($redir_data)) { - $redir_url = $redir_data['url']; - } elseif (is_string($redir_data)) { - $redir_url = $redir_data; - $redir_data = array(); - } else { - // TRANS: Server exception thrown when a URL cannot be processed. - throw new ServerException(sprintf(_("Cannot process URL '%s'"), $given_url)); - } + // If we still don't have a File object, let's create one now! + if (!$file instanceof File) { + // @fixme for new URLs this also looks up non-redirect data + // such as target content type, size, etc, which we need + // for File::saveNew(); so we call it even if not following + // new redirects. + $redir_data = File_redirection::where($given_url); + if (is_array($redir_data)) { + $redir_url = $redir_data['url']; + } elseif (is_string($redir_data)) { + $redir_url = $redir_data; + $redir_data = array(); + } else { + // TRANS: Server exception thrown when a URL cannot be processed. + throw new ServerException(sprintf(_("Cannot process URL '%s'"), $given_url)); + } - // TODO: max field length - if ($redir_url === $given_url || strlen($redir_url) > 255 || !$followRedirects) { - // Save the File object based on our lookup trace - $file = File::saveNew($redir_data, $given_url); - } else { - // This seems kind of messed up... for now skipping this part - // if we're already under a redirect, so we don't go into - // horrible infinite loops if we've been given an unstable - // redirect (where the final destination of the first request - // doesn't match what we get when we ask for it again). - // - // Seen in the wild with clojure.org, which redirects through - // wikispaces for auth and appends session data in the URL params. - $file = self::processNew($redir_url, $notice_id, /*followRedirects*/false); - File_redirection::saveNew($redir_data, $file->id, $given_url); - } + if ($redir_url === $given_url || !$followRedirects) { + // Save the File object based on our lookup trace + $file = File::saveNew($redir_data, $given_url); + } else { + // This seems kind of messed up... for now skipping this part + // if we're already under a redirect, so we don't go into + // horrible infinite loops if we've been given an unstable + // redirect (where the final destination of the first request + // doesn't match what we get when we ask for it again). + // + // Seen in the wild with clojure.org, which redirects through + // wikispaces for auth and appends session data in the URL params. + $file = self::processNew($redir_url, $notice, /*followRedirects*/false); + File_redirection::saveNew($redir_data, $file->id, $given_url); } if (!$file instanceof File) { @@ -177,8 +193,8 @@ class File extends Managed_DataObject } } - if (!empty($notice_id)) { - File_to_post::processNew($file->id, $notice_id); + if ($notice instanceof Notice) { + File_to_post::processNew($file, $notice); } return $file; } @@ -233,19 +249,23 @@ class File extends Managed_DataObject return true; } + public function getFilename() + { + if (!self::validFilename($this->filename)) { + // TRANS: Client exception thrown if a file upload does not have a valid name. + throw new ClientException(_("Invalid filename.")); + } + return $this->filename; + } + // where should the file go? static function filename(Profile $profile, $origname, $mimetype) { - try { - $ext = common_supported_mime_to_ext($mimetype); - } catch (Exception $e) { - // We don't support this mimetype, but let's guess the extension - $ext = substr(strrchr($mimetype, '/'), 1); - } + $ext = self::guessMimeExtension($mimetype); // Normalize and make the original filename more URL friendly. - $origname = basename($origname, $ext); + $origname = basename($origname, ".$ext"); if (class_exists('Normalizer')) { // http://php.net/manual/en/class.normalizer.php // http://www.unicode.org/reports/tr15/ @@ -253,7 +273,7 @@ class File extends Managed_DataObject } $origname = preg_replace('/[^A-Za-z0-9\.\_]/', '_', $origname); - $nickname = $profile->nickname; + $nickname = $profile->getNickname(); $datestamp = strftime('%Y%m%d', time()); do { // generate new random strings until we don't run into a filename collision. @@ -263,6 +283,17 @@ class File extends Managed_DataObject return $filename; } + static function guessMimeExtension($mimetype) + { + try { + $ext = common_supported_mime_to_ext($mimetype); + } catch (Exception $e) { + // We don't support this mimetype, but let's guess the extension + $ext = substr(strrchr($mimetype, '/'), 1); + } + return strtolower($ext); + } + /** * Validation for as-saved base filenames */ @@ -303,7 +334,7 @@ class File extends Managed_DataObject } - if (StatusNet::isHTTPS()) { + if (GNUsocial::useHTTPS()) { $sslserver = common_config('attachments', 'sslserver'); @@ -352,30 +383,21 @@ class File extends Managed_DataObject function getEnclosure(){ $enclosure = (object) array(); - $enclosure->title=$this->title; - $enclosure->url=$this->url; - $enclosure->title=$this->title; - $enclosure->date=$this->date; - $enclosure->modified=$this->modified; - $enclosure->size=$this->size; - $enclosure->mimetype=$this->mimetype; - - if (!isset($this->filename)) { - $notEnclosureMimeTypes = array(null,'text/html','application/xhtml+xml'); - $mimetype = $this->mimetype; - if($mimetype != null){ - $mimetype = strtolower($this->mimetype); - } - $semicolon = strpos($mimetype,';'); - if($semicolon){ - $mimetype = substr($mimetype,0,$semicolon); - } - if (in_array($mimetype, $notEnclosureMimeTypes)) { - Event::handle('FileEnclosureMetadata', array($this, &$enclosure)); - } + foreach (array('title', 'url', 'date', 'modified', 'size', 'mimetype') as $key) { + $enclosure->$key = $this->$key; + } + + $needMoreMetadataMimetypes = array(null, 'application/xhtml+xml'); + + if (!isset($this->filename) && in_array(common_bare_mime($enclosure->mimetype), $needMoreMetadataMimetypes)) { + // This fetches enclosure metadata for non-local links with unset/HTML mimetypes, + // which may be enriched through oEmbed or similar (implemented as plugins) + Event::handle('FileEnclosureMetadata', array($this, &$enclosure)); } - if (empty($enclosure->mimetype)) { - // This means we don't know what it is, so it can't be an enclosure! + if (empty($enclosure->mimetype) || in_array(common_bare_mime($enclosure->mimetype), $needMoreMetadataMimetypes)) { + // This means we either don't know what it is, so it can't + // be shown as an enclosure, or it is an HTML link which + // does not link to a resource with further metadata. throw new ServerException('Unknown enclosure mimetype, not enough metadata'); } return $enclosure; @@ -390,67 +412,92 @@ class File extends Managed_DataObject * @param $crop bool Crop to the max-values' aspect ratio * * @return File_thumbnail + * + * @throws UseFileAsThumbnailException if the file is considered an image itself and should be itself as thumbnail + * @throws UnsupportedMediaException if, despite trying, we can't understand how to make a thumbnail for this format + * @throws ServerException on various other errors */ - public function getThumbnail($width=null, $height=null, $crop=false) + public function getThumbnail($width=null, $height=null, $crop=false, $force_still=true) { - if (intval($this->width) < 1 || intval($this->height) < 1) { - // Old files may have 0 until migrated with scripts/upgrade.php - // For any legitimately unrepresentable ones, we could generate our - // own image (like a square with MIME type in text) - throw new UnsupportedMediaException('No image geometry available.'); + // Get some more information about this file through our ImageFile class + $image = ImageFile::fromFileObject($this); + if ($image->animated && !common_config('thumbnail', 'animated')) { + // null means "always use file as thumbnail" + // false means you get choice between frozen frame or original when calling getThumbnail + if (is_null(common_config('thumbnail', 'animated')) || !$force_still) { + throw new UseFileAsThumbnailException($this->id); + } } - if ($width === null) { - $width = common_config('thumbnail', 'width'); - $height = common_config('thumbnail', 'height'); - $crop = common_config('thumbnail', 'crop'); - } + return $image->getFileThumbnail($width, $height, $crop); + } - if ($height === null) { - $height = $width; - $crop = true; + public function getPath() + { + $filepath = self::path($this->filename); + if (!file_exists($filepath)) { + throw new FileNotFoundException($filepath); } + return $filepath; + } - // Get proper aspect ratio width and height before lookup - // We have to do it through an ImageFile object because of orientation etc. - // Only other solution would've been to rotate + rewrite uploaded files. - $image = ImageFile::fromFileObject($this); - list($width, $height, $x, $y, $w2, $h2) = - $image->scaleToFit($width, $height, $crop); - - $params = array('file_id'=> $this->id, - 'width' => $width, - 'height' => $height); - $thumb = File_thumbnail::pkeyGet($params); - if ($thumb instanceof File_thumbnail) { - return $thumb; + public function getUrl() + { + if (!empty($this->filename)) { + // A locally stored file, so let's generate a URL for our instance. + $url = self::url($this->filename); + if (self::hashurl($url) !== $this->urlhash) { + // For indexing purposes, in case we do a lookup on the 'url' field. + // also we're fixing possible changes from http to https, or paths + $this->updateUrl($url); + } + return $url; } - // throws exception on failure to generate thumbnail - $outname = "thumb-{$width}x{$height}-" . $this->filename; - $outpath = self::path($outname); - - $image->resizeTo($outpath, array('width'=>$width, 'height'=>$height, - 'x'=>$x, 'y'=>$y, - 'w'=>$w2, 'h'=>$h2)); + // No local filename available, return the URL we have stored + return $this->url; + } - // Avoid deleting the original - if ($image->getPath() != $this->getPath()) { - $image->unlink(); + static public function getByUrl($url) + { + $file = new File(); + $file->urlhash = self::hashurl($url); + if (!$file->find(true)) { + throw new NoResultException($file); } - return File_thumbnail::saveThumbnail($this->id, - self::url($outname), - $width, $height, - $outname); + return $file; } - public function getPath() + /** + * @param string $hashstr String of (preferrably lower case) hexadecimal characters, same as result of 'hash_file(...)' + */ + static public function getByHash($hashstr, $alg=File::FILEHASH_ALG) { - return self::path($this->filename); + $file = new File(); + $file->filehash = strtolower($hashstr); + if (!$file->find(true)) { + throw new NoResultException($file); + } + return $file; } - public function getUrl() + + public function updateUrl($url) { - return $this->url; + $file = File::getKV('urlhash', self::hashurl($url)); + if ($file instanceof File) { + throw new ServerException('URL already exists in DB'); + } + $sql = 'UPDATE %1$s SET urlhash=%2$s, url=%3$s WHERE urlhash=%4$s;'; + $result = $this->query(sprintf($sql, $this->tableName(), + $this->_quote((string)self::hashurl($url)), + $this->_quote((string)$url), + $this->_quote((string)$this->urlhash))); + if ($result === false) { + common_log_db_error($this, 'UPDATE', __FILE__); + throw new ServerException("Could not UPDATE {$this->tableName()}.url"); + } + + return $result; } /** @@ -463,9 +510,9 @@ class File extends Managed_DataObject function blowCache($last=false) { - self::blow('file:notice-ids:%s', $this->url); + self::blow('file:notice-ids:%s', $this->id); if ($last) { - self::blow('file:notice-ids:%s;last', $this->url); + self::blow('file:notice-ids:%s;last', $this->id); } self::blow('file:notice-count:%d', $this->id); } @@ -531,9 +578,107 @@ class File extends Managed_DataObject $thumbs->delete(); } } + + $f2p = new File_to_post(); + $f2p->file_id = $this->id; + if ($f2p->find()) { + while ($f2p->fetch()) { + $f2p->delete(); + } + } } // And finally remove the entry from the database return parent::delete($useWhere); } + + public function getTitle() + { + $title = $this->title ?: $this->filename; + + return $title ?: null; + } + + static public function hashurl($url) + { + if (empty($url)) { + throw new Exception('No URL provided to hash algorithm.'); + } + return hash(self::URLHASH_ALG, $url); + } + + static public function beforeSchemaUpdate() + { + $table = strtolower(get_called_class()); + $schema = Schema::get(); + $schemadef = $schema->getTableDef($table); + + // 2015-02-19 We have to upgrade our table definitions to have the urlhash field populated + if (isset($schemadef['fields']['urlhash']) && isset($schemadef['unique keys']['file_urlhash_key'])) { + // We already have the urlhash field, so no need to migrate it. + return; + } + echo "\nFound old $table table, upgrading it to contain 'urlhash' field..."; + + $file = new File(); + $file->query(sprintf('SELECT id, LEFT(url, 191) AS shortenedurl, COUNT(*) AS c FROM %1$s WHERE LENGTH(url)>191 GROUP BY shortenedurl HAVING c > 1', $schema->quoteIdentifier($table))); + print "\nFound {$file->N} URLs with too long entries in file table\n"; + while ($file->fetch()) { + // We've got a URL that is too long for our future file table + // so we'll cut it. We could save the original URL, but there is + // no guarantee it is complete anyway since the previous max was 255 chars. + $dupfile = new File(); + // First we find file entries that would be duplicates of this when shortened + // ... and we'll just throw the dupes out the window for now! It's already so borken. + $dupfile->query(sprintf('SELECT * FROM file WHERE LEFT(url, 191) = "%1$s"', $file->shortenedurl)); + // Leave one of the URLs in the database by using ->find(true) (fetches first entry) + if ($dupfile->find(true)) { + print "\nShortening url entry for $table id: {$file->id} ["; + $orig = clone($dupfile); + $dupfile->url = $file->shortenedurl; // make sure it's only 191 chars from now on + $dupfile->update($orig); + print "\nDeleting duplicate entries of too long URL on $table id: {$file->id} ["; + // only start deleting with this fetch. + while($dupfile->fetch()) { + print "."; + $dupfile->delete(); + } + print "]\n"; + } else { + print "\nWarning! URL suddenly disappeared from database: {$file->url}\n"; + } + } + echo "...and now all the non-duplicates which are longer than 191 characters...\n"; + $file->query('UPDATE file SET url=LEFT(url, 191) WHERE LENGTH(url)>191'); + + echo "\n...now running hacky pre-schemaupdate change for $table:"; + // We have to create a urlhash that is _not_ the primary key, + // transfer data and THEN run checkSchema + $schemadef['fields']['urlhash'] = array ( + 'type' => 'varchar', + 'length' => 64, + 'not null' => false, // this is because when adding column, all entries will _be_ NULL! + 'description' => 'sha256 of destination URL (url field)', + ); + $schemadef['fields']['url'] = array ( + 'type' => 'text', + 'description' => 'destination URL after following possible redirections', + ); + unset($schemadef['unique keys']); + $schema->ensureTable($table, $schemadef); + echo "DONE.\n"; + + $classname = ucfirst($table); + $tablefix = new $classname; + // urlhash is hash('sha256', $url) in the File table + echo "Updating urlhash fields in $table table..."; + // Maybe very MySQL specific :( + $tablefix->query(sprintf('UPDATE %1$s SET %2$s=%3$s;', + $schema->quoteIdentifier($table), + 'urlhash', + // The line below is "result of sha256 on column `url`" + 'SHA2(url, 256)')); + echo "DONE.\n"; + echo "Resuming core schema upgrade..."; + } }