gnu-social/classes/File.php

871 lines
34 KiB
PHP
Raw Normal View History

<?php
/**
* GNU social - a federating social network
*
* Abstraction for files
*
* LICENCE: This program is free software: you can redistribute it and/or modify
* it under the terms of the GNU Affero General Public License as published by
* the Free Software Foundation, either version 3 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU Affero General Public License for more details.
*
* You should have received a copy of the GNU Affero General Public License
* along with this program. If not, see <http://www.gnu.org/licenses/>.
*
* @category Files
* @package GNUsocial
* @author Mikael Nordfeldth <mmn@hethane.se>
* @author Miguel Dantas <biodantas@gmail.com>
* @copyright 2008-2009, 2019 Free Software Foundation http://fsf.org
* @license http://www.fsf.org/licensing/licenses/agpl-3.0.html GNU Affero General Public License version 3.0
* @link https://www.gnu.org/software/social/
*/
defined('GNUSOCIAL') || die();
/**
* Table Definition for file
*/
2011-08-22 22:52:02 +01:00
class File extends Managed_DataObject
{
public $__table = 'file'; // table name
public $id; // int(4) primary_key not_null
public $urlhash; // varchar(64) unique_key
public $url; // text
public $filehash; // varchar(64) indexed
public $mimetype; // varchar(50)
public $size; // int(4)
public $title; // text()
public $date; // int(4)
public $protected; // int(4)
public $filename; // text()
public $width; // int(4)
public $height; // int(4)
public $modified; // timestamp() not_null default_CURRENT_TIMESTAMP
const URLHASH_ALG = 'sha256';
const FILEHASH_ALG = 'sha256';
2011-08-22 22:52:02 +01:00
public static function schemaDef()
{
return array(
'fields' => array(
'id' => array('type' => 'serial', 'not null' => true),
'urlhash' => array('type' => 'varchar', 'length' => 64, 'not null' => true, 'description' => 'sha256 of destination URL (url field)'),
'url' => array('type' => 'text', 'description' => 'destination URL after following possible redirections'),
'filehash' => array('type' => 'varchar', 'length' => 64, 'not null' => false, 'description' => 'sha256 of the file contents, only for locally stored files of course'),
2011-08-22 22:52:02 +01:00
'mimetype' => array('type' => 'varchar', 'length' => 50, 'description' => 'mime type of resource'),
'size' => array('type' => 'int', 'description' => 'size of resource when available'),
'title' => array('type' => 'text', 'description' => 'title of resource when available'),
2011-08-22 22:52:02 +01:00
'date' => array('type' => 'int', 'description' => 'date of resource according to http query'),
'protected' => array('type' => 'int', 'description' => 'true when URL is private (needs login)'),
'filename' => array('type' => 'text', 'description' => 'if file is stored locally (too) this is the filename'),
'width' => array('type' => 'int', 'description' => 'width in pixels, if it can be described as such and data is available'),
'height' => array('type' => 'int', 'description' => 'height in pixels, if it can be described as such and data is available'),
2011-08-22 22:52:02 +01:00
'modified' => array('type' => 'timestamp', 'not null' => true, 'description' => 'date this record was modified'),
),
'primary key' => array('id'),
'unique keys' => array(
'file_urlhash_key' => array('urlhash'),
2011-08-22 22:52:02 +01:00
),
'indexes' => array(
'file_filehash_idx' => array('filehash'),
),
2011-08-22 22:52:02 +01:00
);
}
public static function isProtected($url)
{
$protected_urls_exps = array(
'https://www.facebook.com/login.php',
common_path('main/login')
);
foreach ($protected_urls_exps as $protected_url_exp) {
if (preg_match('!^'.preg_quote($protected_url_exp).'(.*)$!i', $url) === 1) {
return true;
}
}
2016-01-25 19:00:46 +00:00
return false;
}
/**
* Save a new file record.
*
* @param array $redir_data lookup data eg from File_redirection::where()
* @param string $given_url
* @return File
* @throws ServerException
*/
public static function saveNew(array $redir_data, $given_url)
{
$file = null;
try {
// I don't know why we have to keep doing this but we run a last check to avoid
// uniqueness bugs.
$file = File::getByUrl($given_url);
return $file;
} catch (NoResultException $e) {
// We don't have the file's URL since before, so let's continue.
}
// if the given url is an local attachment url and the id already exists, don't
// save a new file record. This should never happen, but let's make it foolproof
// FIXME: how about attachments servers?
$u = parse_url($given_url);
if (isset($u['host']) && $u['host'] === common_config('site', 'server')) {
$r = Router::get();
2016-02-26 13:13:46 +00:00
// Skip the / in the beginning or $r->map won't match
try {
$args = $r->map(mb_substr($u['path'], 1));
if ($args['action'] === 'attachment') {
try {
// $args['attachment'] should always be set if action===attachment, given our routing rules
$file = File::getByID($args['attachment']);
return $file;
} catch (EmptyPkeyValueException $e) {
// ...but $args['attachment'] can also be 0...
} catch (NoResultException $e) {
// apparently this link goes to us, but is _not_ an existing attachment (File) ID?
}
}
} catch (Exception $e) {
// Some other exception was thrown from $r->map, likely a
// ClientException (404) because of some malformed link to
// our own instance. It's still a valid URL however, so we
// won't abort anything... I noticed this when linking:
// https://social.umeahackerspace.se/mmn/foaf' (notice the
// apostrophe in the end, making it unrecognizable for our
// URL routing.
// That specific issue (the apostrophe being part of a link
// is something that may or may not have been fixed since,
// in lib/util.php in common_replace_urls_callback().
}
}
$file = new File;
$file->url = $given_url;
if (!empty($redir_data['protected'])) {
$file->protected = $redir_data['protected'];
}
if (!empty($redir_data['title'])) {
$file->title = $redir_data['title'];
}
if (!empty($redir_data['type'])) {
$file->mimetype = $redir_data['type'];
}
if (!empty($redir_data['size'])) {
$file->size = intval($redir_data['size']);
}
if (isset($redir_data['time']) && $redir_data['time'] > 0) {
$file->date = intval($redir_data['time']);
}
$file->saveFile();
return $file;
}
public function saveFile()
{
$this->urlhash = self::hashurl($this->url);
if (!Event::handle('StartFileSaveNew', array(&$this))) {
throw new ServerException('File not saved due to an aborted StartFileSaveNew event.');
}
$this->id = $this->insert();
if ($this->id === false) {
throw new ServerException('File/URL metadata could not be saved to the database.');
}
Event::handle('EndFileSaveNew', array($this));
}
/**
2010-11-15 19:00:42 +00:00
* Go look at a URL and possibly save data about it if it's new:
* - follow redirect chains and store them in file_redirection
* - if a thumbnail is available, save it in file_thumbnail
* - save file record with basic info
* - optionally save a file_to_post record
* - return the File object with the full reference
*
* @param string $given_url the URL we're looking at
* @param Notice $notice (optional)
2010-11-15 19:00:42 +00:00
* @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 $notice=null, $followRedirects=true)
{
if (empty($given_url)) {
throw new ServerException('No given URL to process');
}
$given_url = File_redirection::_canonUrl($given_url);
if (empty($given_url)) {
throw new ServerException('No canonical URL from given URL to process');
}
$redir = File_redirection::where($given_url);
try {
$file = $redir->getFile();
} catch (EmptyPkeyValueException $e) {
common_log(LOG_ERR, 'File_redirection::where gave object with empty file_id for given_url '._ve($given_url));
throw new ServerException('URL processing failed without new File object');
} catch (NoResultException $e) {
// This should not happen
common_log(LOG_ERR, 'File_redirection after discovery could still not return a File object.');
throw new ServerException('URL processing failed without new File object');
}
if ($notice instanceof Notice) {
File_to_post::processNew($file, $notice);
}
return $file;
}
public static function respectsQuota(Profile $scoped, $fileSize)
{
if ($fileSize > common_config('attachments', 'file_quota')) {
// TRANS: Message used to be inserted as %2$s in the text "No file may
// TRANS: be larger than %1$d byte and the file you sent was %2$s.".
// TRANS: %1$d is the number of bytes of an uploaded file.
$fileSizeText = sprintf(_m('%1$d byte', '%1$d bytes', $fileSize), $fileSize);
$fileQuota = common_config('attachments', 'file_quota');
// TRANS: Message given if an upload is larger than the configured maximum.
// TRANS: %1$d (used for plural) is the byte limit for uploads,
// TRANS: %2$s is the proper form of "n bytes". This is the only ways to have
// TRANS: gettext support multiple plurals in the same message, unfortunately...
throw new ClientException(
sprintf(
_m(
'No file may be larger than %1$d byte and the file you sent was %2$s. Try to upload a smaller version.',
'No file may be larger than %1$d bytes and the file you sent was %2$s. Try to upload a smaller version.',
$fileQuota
),
$fileQuota,
$fileSizeText
)
);
}
$file = new File;
$query = "select sum(size) as total from file join file_to_post on file_to_post.file_id = file.id join notice on file_to_post.post_id = notice.id where profile_id = {$scoped->id} and file.url like '%/notice/%/file'";
$file->query($query);
$file->fetch();
$total = $file->total + $fileSize;
if ($total > common_config('attachments', 'user_quota')) {
// TRANS: Message given if an upload would exceed user quota.
2010-11-04 17:33:39 +00:00
// TRANS: %d (number) is the user quota in bytes and is used for plural.
throw new ClientException(
sprintf(
_m(
'A file this large would exceed your user quota of %d byte.',
'A file this large would exceed your user quota of %d bytes.',
common_config('attachments', 'user_quota')
),
common_config('attachments', 'user_quota')
)
);
}
$query .= ' AND EXTRACT(month FROM file.modified) = EXTRACT(month FROM now()) and EXTRACT(year FROM file.modified) = EXTRACT(year FROM now())';
$file->query($query);
$file->fetch();
$total = $file->total + $fileSize;
if ($total > common_config('attachments', 'monthly_quota')) {
// TRANS: Message given id an upload would exceed a user's monthly quota.
2010-11-04 17:33:39 +00:00
// TRANS: $d (number) is the monthly user quota in bytes and is used for plural.
throw new ClientException(
sprintf(
_m(
'A file this large would exceed your monthly quota of %d byte.',
'A file this large would exceed your monthly quota of %d bytes.',
common_config('attachments', 'monthly_quota')
),
common_config('attachments', 'monthly_quota')
)
);
}
return true;
}
public function getFilename()
{
return self::tryFilename($this->filename);
}
2016-06-24 14:53:23 +01:00
public function getSize()
{
return intval($this->size);
}
// where should the file go?
public static function filename(Profile $profile, $origname, $mimetype)
{
$ext = self::guessMimeExtension($mimetype, $origname);
// Normalize and make the original filename more URL friendly.
2014-06-22 16:03:27 +01:00
$origname = basename($origname, ".$ext");
if (class_exists('Normalizer')) {
// http://php.net/manual/en/class.normalizer.php
// http://www.unicode.org/reports/tr15/
$origname = Normalizer::normalize($origname, Normalizer::FORM_KC);
}
$origname = preg_replace('/[^A-Za-z0-9\.\_]/', '_', $origname);
2014-06-22 16:03:27 +01:00
$nickname = $profile->getNickname();
$datestamp = strftime('%Y%m%d', time());
do {
// generate new random strings until we don't run into a filename collision.
$random = strtolower(common_confirmation_code(16));
$filename = "$nickname-$datestamp-$origname-$random.$ext";
} while (file_exists(self::path($filename)));
return $filename;
}
/**
* @param string $filename
* @return string|bool Value from the 'extblacklist' array, in the config
*/
public static function getSafeExtension(string $filename) {
if (preg_match('/^.+?\.([A-Za-z0-9]+)$/', $filename, $matches)) {
// we matched on a file extension, so let's see if it means something.
$ext = mb_strtolower($matches[1]);
$blacklist = common_config('attachments', 'extblacklist');
// If we got an extension from $filename we want to check if it's in a blacklist
// so we avoid people uploading restricted files
if (array_key_exists($ext, $blacklist)) {
if (!is_string($blacklist[$ext])) {
return false;
}
// return a safe replacement extension ('php' => 'phps' for example)
return $blacklist[$ext];
}
// the attachment extension based on its filename was not blacklisted so it's ok to use it
return $ext;
} else {
return false;
}
}
/**
* @param $mimetype string The mimetype we've discovered for this file.
* @param $filename string An optional filename which we can use on failure.
* @return mixed|string
* @throws ClientException
*/
public static function guessMimeExtension($mimetype, $filename=null)
{
try {
// first see if we know the extension for our mimetype
$ext = common_supported_mime_to_ext($mimetype);
// we do, so use it!
return $ext;
2016-07-06 07:59:16 +01:00
} catch (UnknownMimeExtensionException $e) {
// We don't know the extension for this mimetype, but let's guess.
2016-07-06 07:59:16 +01:00
// If we can't recognize the extension from the MIME, we try
// to guess based on filename, if one was supplied.
if (!is_null($filename)) {
$ext = getSafeExtension($filename);
if ($ext === false) {
// we don't have a safe replacement extension
throw new ClientException(_('Blacklisted file extension.'));
} else {
return $ext;
}
}
2016-07-06 07:59:16 +01:00
} catch (Exception $e) {
common_log(LOG_INFO, 'Problem when figuring out extension for mimetype: '._ve($e));
}
// If nothing else has given us a result, try to extract it from
// the mimetype value (this turns .jpg to .jpeg for example...)
$matches = array();
// Will get jpeg from image/jpeg as well as json from application/jrd+json
if (!preg_match('/[\/+-\.]([a-z0-9]+)/', mb_strtolower($mimetype), $matches)) {
throw new Exception("Malformed mimetype: {$mimetype}");
}
return mb_strtolower($matches[1]);
}
/**
* Validation for as-saved base filenames
* @param $filename
* @return false|int
*/
public static function validFilename($filename)
{
return preg_match('/^[A-Za-z0-9._-]+$/', $filename);
}
public static function tryFilename($filename)
{
if (!self::validFilename($filename)) {
throw new InvalidFilenameException($filename);
}
// if successful, return the filename for easy if-statementing
return $filename;
}
/**
* @param $filename
* @return string
* @throws InvalidFilenameException
*/
public static function path($filename)
{
self::tryFilename($filename);
$dir = common_config('attachments', 'dir');
2016-03-07 19:13:07 +00:00
if (!in_array($dir[mb_strlen($dir)-1], ['/', '\\'])) {
$dir .= DIRECTORY_SEPARATOR;
}
return $dir . $filename;
}
public static function url($filename)
{
self::tryFilename($filename);
if (common_config('site', 'private')) {
return common_local_url(
'getfile',
array('filename' => $filename)
);
}
if (GNUsocial::useHTTPS()) {
$sslserver = common_config('attachments', 'sslserver');
if (empty($sslserver)) {
// XXX: this assumes that background dir == site dir + /file/
// not true if there's another server
if (is_string(common_config('site', 'sslserver')) &&
mb_strlen(common_config('site', 'sslserver')) > 0) {
$server = common_config('site', 'sslserver');
} elseif (common_config('site', 'server')) {
$server = common_config('site', 'server');
}
$path = common_config('site', 'path') . '/file/';
} else {
$server = $sslserver;
$path = common_config('attachments', 'sslpath');
if (empty($path)) {
$path = common_config('attachments', 'path');
}
}
$protocol = 'https';
} else {
$path = common_config('attachments', 'path');
$server = common_config('attachments', 'server');
if (empty($server)) {
$server = common_config('site', 'server');
}
2010-02-11 22:06:57 +00:00
$ssl = common_config('attachments', 'ssl');
2010-02-11 22:06:57 +00:00
$protocol = ($ssl) ? 'https' : 'http';
}
2010-02-11 22:06:57 +00:00
if ($path[strlen($path)-1] != '/') {
$path .= '/';
}
if ($path[0] != '/') {
$path = '/'.$path;
}
return $protocol.'://'.$server.$path.$filename;
}
public static $_enclosures = array();
public function getEnclosure()
{
if (isset(self::$_enclosures[$this->getID()])) {
return self::$_enclosures[$this->getID()];
}
$enclosure = (object) array();
foreach (array('title', 'url', 'date', 'modified', 'size', 'mimetype', 'width', 'height') as $key) {
if ($this->$key !== '') {
$enclosure->$key = $this->$key;
}
}
$needMoreMetadataMimetypes = array(null, 'application/xhtml+xml', 'text/html');
if (!isset($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 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');
}
self::$_enclosures[$this->getID()] = $enclosure;
return $enclosure;
}
public function hasThumbnail()
{
try {
$this->getThumbnail();
} catch (Exception $e) {
return false;
}
return true;
}
/**
* Get the attachment's thumbnail record, if any.
* Make sure you supply proper 'int' typed variables (or null).
*
* @param $width int Max width of thumbnail in pixels. (if null, use common_config values)
* @param $height int Max height of thumbnail in pixels. (if null, square-crop to $width)
* @param $crop bool Crop to the max-values' aspect ratio
* @param $force_still bool Don't allow fallback to showing original (such as animated GIF)
* @param $upscale mixed Whether or not to scale smaller images up to larger thumbnail sizes. (null = site default)
*
* @return File_thumbnail
*
* @throws UseFileAsThumbnailException if the file is considered an image itself and should be itself as thumbnail
* @throws UnsupportedMediaException if, despite trying, we can't understand how to make a thumbnail for this format
* @throws ServerException on various other errors
*/
2016-02-10 03:40:54 +00:00
public function getThumbnail($width=null, $height=null, $crop=false, $force_still=true, $upscale=null)
{
// 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) {
try {
// remote files with animated GIFs as thumbnails will match this
return File_thumbnail::byFile($this);
} catch (NoResultException $e) {
// and if it's not a remote file, it'll be safe to use the locally stored File
throw new UseFileAsThumbnailException($this);
}
}
}
return $image->getFileThumbnail(
$width,
$height,
$crop,
!is_null($upscale) ? $upscale : common_config('thumbnail', 'upscale')
);
}
public function getPath()
{
$filepath = self::path($this->filename);
if (!file_exists($filepath)) {
throw new FileNotFoundException($filepath);
}
return $filepath;
}
public function getAttachmentUrl()
{
return common_local_url('attachment', array('attachment'=>$this->getID()));
}
public function getAttachmentDownloadUrl()
{
return common_local_url('attachment_download', array('attachment'=>$this->getID()));
}
public function getAttachmentViewUrl()
{
return common_local_url('attachment_view', array('attachment'=>$this->getID()));
}
/**
* @param mixed $use_local true means require local, null means prefer local, false means use whatever is stored
* @return string
* @throws FileNotStoredLocallyException
*/
public function getUrl($use_local=null)
{
if ($use_local !== false) {
if (is_string($this->filename) || !empty($this->filename)) {
// A locally stored file, so let's generate a URL for our instance.
return $this->getAttachmentViewUrl();
}
if ($use_local) {
// if the file wasn't stored locally (has filename) and we require a local URL
throw new FileNotStoredLocallyException($this);
}
}
// No local filename available, return the URL we have stored
return $this->url;
}
public static function getByUrl($url)
{
$file = new File();
$file->urlhash = self::hashurl($url);
if (!$file->find(true)) {
throw new NoResultException($file);
}
return $file;
}
/**
* @param string $hashstr String of (preferrably lower case) hexadecimal characters, same as result of 'hash_file(...)'
* @return File
* @throws NoResultException
*/
public static function getByHash($hashstr)
{
$file = new File();
$file->filehash = strtolower($hashstr);
if (!$file->find(true)) {
throw new NoResultException($file);
}
return $file;
}
public function updateUrl($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;
}
2010-12-28 20:57:31 +00:00
/**
* Blow the cache of notices that link to this URL
*
* @param boolean $last Whether to blow the "last" cache too
*
* @return void
*/
public function blowCache($last=false)
2010-12-28 19:58:55 +00:00
{
self::blow('file:notice-ids:%s', $this->id);
2010-12-28 19:58:55 +00:00
if ($last) {
self::blow('file:notice-ids:%s;last', $this->id);
2010-12-28 19:58:55 +00:00
}
self::blow('file:notice-count:%d', $this->id);
2010-12-28 19:58:55 +00:00
}
/**
* Stream of notices linking to this URL
*
* @param integer $offset Offset to show; default is 0
* @param integer $limit Limit of notices to show
* @param integer $since_id Since this notice
* @param integer $max_id Before this notice
*
* @return array ids of notices that link to this file
*/
public function stream($offset=0, $limit=NOTICES_PER_PAGE, $since_id=0, $max_id=0)
2010-12-28 19:58:55 +00:00
{
// FIXME: Try to get the Profile::current() here in some other way to avoid mixing
// the current session user with possibly background/queue processing.
$stream = new FileNoticeStream($this, Profile::current());
return $stream->getNotices($offset, $limit, $since_id, $max_id);
2010-12-28 19:58:55 +00:00
}
public function noticeCount()
{
$cacheKey = sprintf('file:notice-count:%d', $this->id);
$count = self::cacheGet($cacheKey);
if ($count === false) {
$f2p = new File_to_post();
$f2p->file_id = $this->id;
$count = $f2p->count();
self::cacheSet($cacheKey, $count);
}
return $count;
}
public function isLocal()
{
return !empty($this->filename);
}
public function delete($useWhere=false)
{
// Delete the file, if it exists locally
if (!empty($this->filename) && file_exists(self::path($this->filename))) {
$deleted = @unlink(self::path($this->filename));
if (!$deleted) {
common_log(LOG_ERR, sprintf('Could not unlink existing file: "%s"', self::path($this->filename)));
}
}
// Clear out related things in the database and filesystem, such as thumbnails
if (Event::handle('FileDeleteRelated', array($this))) {
$thumbs = new File_thumbnail();
$thumbs->file_id = $this->id;
if ($thumbs->find()) {
while ($thumbs->fetch()) {
$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 = $this->title ?: MediaFile::getDisplayName($this);
return $title ?: null;
}
2016-07-23 20:00:57 +01:00
public function setTitle($title)
{
$orig = clone($this);
$this->title = mb_strlen($title) > 0 ? $title : null;
return $this->update($orig);
}
public static function hashurl($url)
{
if (empty($url)) {
throw new Exception('No URL provided to hash algorithm.');
}
return hash(self::URLHASH_ALG, $url);
}
public static 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
2015-02-19 18:36:59 +00:00
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', $dupfile->_quote($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);
$origurl = $dupfile->url; // save for logging purposes
$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 ($dupfile->fetch()) {
common_log(LOG_INFO, sprintf('Deleting duplicate File entry of %1$d: %2$d (original URL: %3$s collides with these first 191 characters: %4$s', $dupfile->id, $file->id, $origurl, $file->shortenedurl));
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...";
}
2016-02-02 23:22:18 +00:00
}