• $t_unique_name = file_generate_unique_name( $t_file_hash . '-' . $t_file_name, $t_file_path );
• $t_disk_file_name = $t_file_path . $t_unique_name;

• $c_unique_name = db_prepare_string( $t_unique_name );

  $t_file_size = filesize( $t_tmp_file );
  if( 0 == $t_file_size ) {
  @@ -624,7 +713,7 @@ function file_add( $p_bug_id, $p_file, $p_table = 'bug', $p_title = '', $p_desc
  $query = "INSERT INTO $t_file_table
  (" . $p_table . "_id, title, description, diskfile, filename, folder, filesize, file_type, date_added, content)
  VALUES

• ($c_id, '$c_title', '$c_desc', '$c_disk_file_name', '$c_new_file_name', '$c_file_path', $c_file_size, '$c_file_type', '" . db_now() . "', $c_content)";
• ($c_id, '$c_title', '$c_desc', '$c_unique_name',
  +'$c_new_file_name', '$c_file_path', $c_file_size, '$c_file_type', '" .
  +db_now() . "', $c_content)";

This to me looks like we still store path information in the database - given we put all attachments in 1 folder and a project configuration specifies the location of that folder, we should only store the filename itself in the database, and no absolute/relative path information.

Paul

At the moment we store a folder field in the database that I believe is not currently used. I was planning to review that and remove the field as part of a separate issue. For some reason in the past we use to store the full path to the file, then the folder path. Now we only store the file name and the folder path.