public $url; // text
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
return array(
'fields' => array(
'id' => array('type' => 'serial', 'not null' => true),
- 'urlhash' => array('type' => 'varchar', 'length' => 64, 'description' => 'sha256 of destination URL (url field)'),
+ '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 redirections'),
'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'),
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, $notice_id, /*followRedirects*/false);
+ File_redirection::saveNew($redir_data, $file->id, $given_url);
}
if (!$file instanceof File) {
if (!empty($this->filename)) {
// A locally stored file, so let's generate a URL for our instance.
$url = self::url($this->filename);
- if (self::urlhash($url) !== $this->urlhash) {
+ if (self::hashurl($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 $this->url;
}
+ static public function getByUrl($url)
+ {
+ $file = new File();
+ $file->urlhash = self::hashurl($url);
+ if (!$file->find(true)) {
+ throw new NoResultException($file);
+ }
+ return $file;
+ }
+
public function updateUrl($url)
{
$file = File::getKV('urlhash', self::hashurl($url));
}
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...";
+ // 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,
+ 'description' => 'sha256 of destination URL after following redirections',
+ );
+ $schema->ensureTable($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...";
+ }
}