Fri, 10 Apr 2026 19:28:26 +0000 https://bugs.horde.org/ticket/6060 Incorrect return type from cache miss I have Horde FW_3 from CVS as of about 2 hours ago. My installation is configured to use the MySQL cache backend. While calling $perms->exists() I see that a cache miss results in an erroneous $perms result. Fortunately it is a "fail-closed" so no extra permissions are granted, but the unsuspecting user will not see the application to which he has been granted. From framework/Perms/Perms/datatree.php line 139: $perm = $this->_cache->get('perm_' . $name, $GLOBALS['conf']['cache']['default_lifetime']); if ($perm === false) { $this->_cache->set('perm_' . $name, serialize($perm), $GLOBALS['conf']['cache']['default_lifetime']); $permsCache[$name] = $perm; } This code expects a cache miss to return a false. Inspection of the code seems to take some care to ensure this happens (for instance on a DB error). However the return type for DB::common::getOne() is NULL if the row does not exist. This causes entries which have not yet been cached to return a type of NULL. From line 150 of framework/Cache/Cache/sql.php: $result = $this->_db->getOne($query, $values); if (is_a($result, 'PEAR_Error')) { Horde::logMessage($result, __FILE__, __LINE__, PEAR_LOG_ERR); return false; } else { if ($this->_mc) { $this->_mc->set($key, $result); } return $result; } I *think* the correct action here would be to check for a return type of NULL from getOne() and then return false, indicating a cache miss. However, I'm not very coherent on the bigger picture for the Cache system yet, so I'll leave it for discussion here. Did I miss something? Practical example: In my case I have a module called "site" which has granted "show/read" perms to Guests. When a guest goes into that app they are prompted with a login screen. Inserting debug statements into the code shows that while I have indeed defined a "site" permission, it is never checked directly because the Perms/datatree code receives a NULL from the Cache driver. Wed, 02 Jan 2008 03:43:56 +0000 https://bugs.horde.org/ticket/6060#t40616 Yup, needed an is_null check. Thanks. Wed, 02 Jan 2008 19:38:32 +0000 https://bugs.horde.org/ticket/6060#t40626