| 1 | <?php |
| 2 | |
| 3 | /** |
| 4 | * mime.php |
| 5 | * |
| 6 | * Copyright (c) 1999-2002 The SquirrelMail Project Team |
| 7 | * Licensed under the GNU GPL. For full terms see the file COPYING. |
| 8 | * |
| 9 | * This contains the functions necessary to detect and decode MIME |
| 10 | * messages. |
| 11 | * |
| 12 | * $Id$ |
| 13 | */ |
| 14 | |
| 15 | require_once('../functions/imap.php'); |
| 16 | require_once('../functions/attachment_common.php'); |
| 17 | |
| 18 | /** Setting up the objects that have the structure for the message **/ |
| 19 | class msg_header { |
| 20 | /** msg_header contains generic variables for values that **/ |
| 21 | /** could be in a header. **/ |
| 22 | |
| 23 | var $type0 = '', $type1 = '', $boundary = '', $charset = '', |
| 24 | $encoding = '', $size = 0, $to = array(), $from = '', $date = '', |
| 25 | $cc = array(), $bcc = array(), $reply_to = '', $subject = '', |
| 26 | $id = 0, $mailbox = '', $description = '', $filename = '', |
| 27 | $entity_id = 0, $message_id = 0, $name = '', $priority = 3; |
| 28 | } |
| 29 | |
| 30 | class message { |
| 31 | /** message is the object that contains messages. It is a recursive |
| 32 | object in that through the $entities variable, it can contain |
| 33 | more objects of type message. See documentation in mime.txt for |
| 34 | a better description of how this works. |
| 35 | **/ |
| 36 | var $header = ''; |
| 37 | var $entities = array(); |
| 38 | |
| 39 | function addEntity ($msg) { |
| 40 | $this->entities[] = $msg; |
| 41 | } |
| 42 | } |
| 43 | |
| 44 | /* --------------------------------------------------------------------------------- */ |
| 45 | /* MIME DECODING */ |
| 46 | /* --------------------------------------------------------------------------------- */ |
| 47 | |
| 48 | /* This function gets the structure of a message and stores it in the "message" class. |
| 49 | * It will return this object for use with all relevant header information and |
| 50 | * fully parsed into the standard "message" object format. |
| 51 | */ |
| 52 | function mime_structure ($imap_stream, $header) { |
| 53 | |
| 54 | sqimap_messages_flag ($imap_stream, $header->id, $header->id, 'Seen'); |
| 55 | $ssid = sqimap_session_id(); |
| 56 | $lsid = strlen( $ssid ); |
| 57 | $id = $header->id; |
| 58 | fputs ($imap_stream, "$ssid FETCH $id BODYSTRUCTURE\r\n"); |
| 59 | // |
| 60 | // This should use sqimap_read_data instead of reading it itself |
| 61 | // |
| 62 | $read = fgets ($imap_stream, 10000); |
| 63 | $bodystructure = ''; |
| 64 | while ( substr($read, 0, $lsid) <> $ssid && |
| 65 | !feof( $imap_stream ) ) { |
| 66 | $bodystructure .= $read; |
| 67 | $read = fgets ($imap_stream, 10000); |
| 68 | } |
| 69 | $read = $bodystructure; |
| 70 | |
| 71 | // isolate the body structure and remove beginning and end parenthesis |
| 72 | $read = trim(substr ($read, strpos(strtolower($read), 'bodystructure') + 13)); |
| 73 | $read = trim(substr ($read, 0, -1)); |
| 74 | $end = mime_match_parenthesis(0, $read); |
| 75 | while ($end == strlen($read)-1) { |
| 76 | $read = trim(substr ($read, 0, -1)); |
| 77 | $read = trim(substr ($read, 1)); |
| 78 | $end = mime_match_parenthesis(0, $read); |
| 79 | } |
| 80 | |
| 81 | $msg = mime_parse_structure ($read, 0); |
| 82 | $msg->header = $header; |
| 83 | |
| 84 | return( $msg ); |
| 85 | } |
| 86 | |
| 87 | /* this starts the parsing of a particular structure. It is called recursively, |
| 88 | * so it can be passed different structures. It returns an object of type |
| 89 | * $message. |
| 90 | * First, it checks to see if it is a multipart message. If it is, then it |
| 91 | * handles that as it sees is necessary. If it is just a regular entity, |
| 92 | * then it parses it and adds the necessary header information (by calling out |
| 93 | * to mime_get_elements() |
| 94 | */ |
| 95 | function mime_parse_structure ($structure, $ent_id) { |
| 96 | |
| 97 | $msg = new message(); |
| 98 | if ($structure{0} == '(') { |
| 99 | $ent_id = mime_new_element_level($ent_id); |
| 100 | $start = $end = -1; |
| 101 | do { |
| 102 | $start = $end+1; |
| 103 | $end = mime_match_parenthesis ($start, $structure); |
| 104 | |
| 105 | $element = substr($structure, $start+1, ($end - $start)-1); |
| 106 | $ent_id = mime_increment_id ($ent_id); |
| 107 | $newmsg = mime_parse_structure ($element, $ent_id); |
| 108 | $msg->addEntity ($newmsg); |
| 109 | } while ($structure{$end+1} == '('); |
| 110 | } else { |
| 111 | // parse the elements |
| 112 | $msg = mime_get_element ($structure, $msg, $ent_id); |
| 113 | } |
| 114 | return $msg; |
| 115 | } |
| 116 | |
| 117 | /* Increments the element ID. An element id can look like any of |
| 118 | * the following: 1, 1.2, 4.3.2.4.1, etc. This function increments |
| 119 | * the last number of the element id, changing 1.2 to 1.3. |
| 120 | */ |
| 121 | function mime_increment_id ($id) { |
| 122 | |
| 123 | if (strpos($id, ".")) { |
| 124 | $first = substr($id, 0, strrpos($id, ".")); |
| 125 | $last = substr($id, strrpos($id, ".")+1); |
| 126 | $last++; |
| 127 | $new = $first . "." .$last; |
| 128 | } else { |
| 129 | $new = $id + 1; |
| 130 | } |
| 131 | |
| 132 | return $new; |
| 133 | } |
| 134 | |
| 135 | /* |
| 136 | * See comment for mime_increment_id(). |
| 137 | * This adds another level on to the entity_id changing 1.3 to 1.3.0 |
| 138 | * NOTE: 1.3.0 is not a valid element ID. It MUST be incremented |
| 139 | * before it can be used. I left it this way so as not to have |
| 140 | * to make a special case if it is the first entity_id. It |
| 141 | * always increments it, and that works fine. |
| 142 | */ |
| 143 | function mime_new_element_level ($id) { |
| 144 | |
| 145 | if (!$id) { |
| 146 | $id = 0; |
| 147 | } else { |
| 148 | $id = $id . '.0'; |
| 149 | } |
| 150 | |
| 151 | return( $id ); |
| 152 | } |
| 153 | |
| 154 | function mime_get_element (&$structure, $msg, $ent_id) { |
| 155 | |
| 156 | $elem_num = 1; |
| 157 | $msg->header = new msg_header(); |
| 158 | $msg->header->entity_id = $ent_id; |
| 159 | $properties = array(); |
| 160 | |
| 161 | while (strlen($structure) > 0) { |
| 162 | $structure = trim($structure); |
| 163 | $char = $structure{0}; |
| 164 | |
| 165 | if (strtolower(substr($structure, 0, 3)) == 'nil') { |
| 166 | $text = ''; |
| 167 | $structure = substr($structure, 3); |
| 168 | } else if ($char == '"') { |
| 169 | // loop through until we find the matching quote, and return that as a string |
| 170 | $pos = 1; |
| 171 | $text = ''; |
| 172 | while ( ($char = $structure{$pos} ) <> '"' && $pos < strlen($structure)) { |
| 173 | $text .= $char; |
| 174 | $pos++; |
| 175 | } |
| 176 | $structure = substr($structure, strlen($text) + 2); |
| 177 | } else if ($char == '(') { |
| 178 | // comment me |
| 179 | $end = mime_match_parenthesis (0, $structure); |
| 180 | $sub = substr($structure, 1, $end-1); |
| 181 | $properties = mime_get_props($properties, $sub); |
| 182 | $structure = substr($structure, strlen($sub) + 2); |
| 183 | } else { |
| 184 | // loop through until we find a space or an end parenthesis |
| 185 | $pos = 0; |
| 186 | $char = $structure{$pos}; |
| 187 | $text = ''; |
| 188 | while ($char != ' ' && $char != ')' && $pos < strlen($structure)) { |
| 189 | $text .= $char; |
| 190 | $pos++; |
| 191 | $char = $structure{$pos}; |
| 192 | } |
| 193 | $structure = substr($structure, strlen($text)); |
| 194 | } |
| 195 | |
| 196 | // This is where all the text parts get put into the header |
| 197 | switch ($elem_num) { |
| 198 | case 1: |
| 199 | $msg->header->type0 = strtolower($text); |
| 200 | break; |
| 201 | case 2: |
| 202 | $msg->header->type1 = strtolower($text); |
| 203 | break; |
| 204 | case 4: // Id |
| 205 | // Invisimail enclose images with <> |
| 206 | $msg->header->id = str_replace( '<', '', str_replace( '>', '', $text ) ); |
| 207 | break; |
| 208 | case 5: |
| 209 | $msg->header->description = $text; |
| 210 | break; |
| 211 | case 6: |
| 212 | $msg->header->encoding = strtolower($text); |
| 213 | break; |
| 214 | case 7: |
| 215 | $msg->header->size = $text; |
| 216 | break; |
| 217 | default: |
| 218 | if ($msg->header->type0 == 'text' && $elem_num == 8) { |
| 219 | // This is a plain text message, so lets get the number of lines |
| 220 | // that it contains. |
| 221 | $msg->header->num_lines = $text; |
| 222 | |
| 223 | } else if ($msg->header->type0 == 'message' && $msg->header->type1 == 'rfc822' && $elem_num == 8) { |
| 224 | // This is an encapsulated message, so lets start all over again and |
| 225 | // parse this message adding it on to the existing one. |
| 226 | $structure = trim($structure); |
| 227 | if ( $structure{0} == '(' ) { |
| 228 | $e = mime_match_parenthesis (0, $structure); |
| 229 | $structure = substr($structure, 0, $e); |
| 230 | $structure = substr($structure, 1); |
| 231 | $m = mime_parse_structure($structure, $msg->header->entity_id); |
| 232 | |
| 233 | // the following conditional is there to correct a bug that wasn't |
| 234 | // incrementing the entity IDs correctly because of the special case |
| 235 | // that message/rfc822 is. This fixes it fine. |
| 236 | if (substr($structure, 1, 1) != '(') |
| 237 | $m->header->entity_id = mime_increment_id(mime_new_element_level($ent_id)); |
| 238 | |
| 239 | // Now we'll go through and reformat the results. |
| 240 | if ($m->entities) { |
| 241 | for ($i=0; $i < count($m->entities); $i++) { |
| 242 | $msg->addEntity($m->entities[$i]); |
| 243 | } |
| 244 | } else { |
| 245 | $msg->addEntity($m); |
| 246 | } |
| 247 | $structure = ""; |
| 248 | } |
| 249 | } |
| 250 | break; |
| 251 | } |
| 252 | $elem_num++; |
| 253 | $text = ""; |
| 254 | } |
| 255 | // loop through the additional properties and put those in the various headers |
| 256 | if ($msg->header->type0 != 'message') { |
| 257 | for ($i=0; $i < count($properties); $i++) { |
| 258 | $msg->header->{$properties[$i]['name']} = $properties[$i]['value']; |
| 259 | } |
| 260 | } |
| 261 | |
| 262 | return $msg; |
| 263 | } |
| 264 | |
| 265 | /* |
| 266 | * I did most of the MIME stuff yesterday (June 20, 2000), but I couldn't |
| 267 | * figure out how to do this part, so I decided to go to bed. I woke up |
| 268 | * in the morning and had a flash of insight. I went to the white-board |
| 269 | * and scribbled it out, then spent a bit programming it, and this is the |
| 270 | * result. Nothing complicated, but I think my brain was fried yesterday. |
| 271 | * Funny how that happens some times. |
| 272 | * |
| 273 | * This gets properties in a nested parenthesisized list. For example, |
| 274 | * this would get passed something like: ("attachment" ("filename" "luke.tar.gz")) |
| 275 | * This returns an array called $props with all paired up properties. |
| 276 | * It ignores the "attachment" for now, maybe that should change later |
| 277 | * down the road. In this case, what is returned is: |
| 278 | * $props[0]["name"] = "filename"; |
| 279 | * $props[0]["value"] = "luke.tar.gz"; |
| 280 | */ |
| 281 | function mime_get_props ($props, $structure) { |
| 282 | |
| 283 | while (strlen($structure) > 0) { |
| 284 | $structure = trim($structure); |
| 285 | $char = $structure{0}; |
| 286 | |
| 287 | if ($char == '"') { |
| 288 | $pos = 1; |
| 289 | $tmp = ''; |
| 290 | while ( ( $char = $structure{$pos} ) != '"' && |
| 291 | $pos < strlen($structure)) { |
| 292 | $tmp .= $char; |
| 293 | $pos++; |
| 294 | } |
| 295 | $structure = trim(substr($structure, strlen($tmp) + 2)); |
| 296 | $char = $structure{0}; |
| 297 | |
| 298 | if ($char == '"') { |
| 299 | $pos = 1; |
| 300 | $value = ''; |
| 301 | while ( ( $char = $structure{$pos} ) != '"' && |
| 302 | $pos < strlen($structure) ) { |
| 303 | $value .= $char; |
| 304 | $pos++; |
| 305 | } |
| 306 | $structure = trim(substr($structure, strlen($tmp) + 2)); |
| 307 | |
| 308 | $k = count($props); |
| 309 | $props[$k]['name'] = strtolower($tmp); |
| 310 | $props[$k]['value'] = $value; |
| 311 | } else if ($char == '(') { |
| 312 | $end = mime_match_parenthesis (0, $structure); |
| 313 | $sub = substr($structure, 1, $end-1); |
| 314 | if (! isset($props)) |
| 315 | $props = array(); |
| 316 | $props = mime_get_props($props, $sub); |
| 317 | $structure = substr($structure, strlen($sub) + 2); |
| 318 | } |
| 319 | return $props; |
| 320 | } else if ($char == '(') { |
| 321 | $end = mime_match_parenthesis (0, $structure); |
| 322 | $sub = substr($structure, 1, $end-1); |
| 323 | $props = mime_get_props($props, $sub); |
| 324 | $structure = substr($structure, strlen($sub) + 2); |
| 325 | return $props; |
| 326 | } else { |
| 327 | return $props; |
| 328 | } |
| 329 | } |
| 330 | } |
| 331 | |
| 332 | /* |
| 333 | * Matches parenthesis. It will return the position of the matching |
| 334 | * parenthesis in $structure. For instance, if $structure was: |
| 335 | * ("text" "plain" ("val1name", "1") nil ... ) |
| 336 | * x x |
| 337 | * then this would return 42 to match up those two. |
| 338 | */ |
| 339 | function mime_match_parenthesis ($pos, $structure) { |
| 340 | |
| 341 | $j = strlen( $structure ); |
| 342 | |
| 343 | // ignore all extra characters |
| 344 | // If inside of a string, skip string -- Boundary IDs and other |
| 345 | // things can have ) in them. |
| 346 | if ( $structure{$pos} != '(' ) { |
| 347 | return( $j ); |
| 348 | } |
| 349 | |
| 350 | while ( $pos < $j ) { |
| 351 | $pos++; |
| 352 | if ($structure{$pos} == ')') { |
| 353 | return $pos; |
| 354 | } elseif ($structure{$pos} == '"') { |
| 355 | $pos++; |
| 356 | while ( $structure{$pos} != '"' && |
| 357 | $pos < $j ) { |
| 358 | if (substr($structure, $pos, 2) == '\\"') { |
| 359 | $pos++; |
| 360 | } elseif (substr($structure, $pos, 2) == '\\\\') { |
| 361 | $pos++; |
| 362 | } |
| 363 | $pos++; |
| 364 | } |
| 365 | } elseif ( $structure{$pos} == '(' ) { |
| 366 | $pos = mime_match_parenthesis ($pos, $structure); |
| 367 | } |
| 368 | } |
| 369 | echo _("Error decoding mime structure. Report this as a bug!") . '<br>'; |
| 370 | return( $pos ); |
| 371 | } |
| 372 | |
| 373 | function mime_fetch_body($imap_stream, $id, $ent_id ) { |
| 374 | |
| 375 | /* |
| 376 | * do a bit of error correction. If we couldn't find the entity id, just guess |
| 377 | * that it is the first one. That is usually the case anyway. |
| 378 | */ |
| 379 | if (!$ent_id) { |
| 380 | $ent_id = 1; |
| 381 | } |
| 382 | |
| 383 | $cmd = "FETCH $id BODY[$ent_id]"; |
| 384 | $data = sqimap_run_command ($imap_stream, $cmd, true, $response, $message); |
| 385 | $topline = array_shift($data); |
| 386 | while (! ereg('\\* [0-9]+ FETCH ', $topline) && $data) { |
| 387 | $topline = array_shift($data); |
| 388 | } |
| 389 | $wholemessage = implode('', $data); |
| 390 | if (ereg('\\{([^\\}]*)\\}', $topline, $regs)) { |
| 391 | $ret = substr( $wholemessage, 0, $regs[1] ); |
| 392 | /* |
| 393 | There is some information in the content info header that could be important |
| 394 | in order to parse html messages. Let's get them here. |
| 395 | */ |
| 396 | if ( $ret{0} == '<' ) { |
| 397 | $data = sqimap_run_command ($imap_stream, "FETCH $id BODY[$ent_id.MIME]", true, $response, $message); |
| 398 | $base = ''; |
| 399 | $k = 10; |
| 400 | foreach( $data as $d ) { |
| 401 | if ( substr( $d, 0, 13 ) == 'Content-Base:' ) { |
| 402 | $j = strlen( $d ); |
| 403 | $i = 13; |
| 404 | $base = ''; |
| 405 | while ( $i < $j && |
| 406 | ( !isNoSep( $d{$i} ) || $d{$i} == '"' ) ) |
| 407 | $i++; |
| 408 | while ( $i < $j ) { |
| 409 | if ( isNoSep( $d{$i} ) ) |
| 410 | $base .= $d{$i}; |
| 411 | $i++; |
| 412 | } |
| 413 | $k = 0; |
| 414 | } elseif ( $k == 1 && !isnosep( $d{0} ) ) { |
| 415 | $base .= substr( $d, 1 ); |
| 416 | } |
| 417 | $k++; |
| 418 | } |
| 419 | if ( $base <> '' ) { |
| 420 | $ret = "<base href=\"$base\">" . $ret; |
| 421 | } |
| 422 | } |
| 423 | } else if (ereg('"([^"]*)"', $topline, $regs)) { |
| 424 | $ret = $regs[1]; |
| 425 | } else { |
| 426 | global $where, $what, $mailbox, $passed_id, $startMessage; |
| 427 | $par = 'mailbox=' . urlencode($mailbox) . "&passed_id=$passed_id"; |
| 428 | if (isset($where) && isset($what)) { |
| 429 | $par .= '&where='. urlencode($where) . "&what=" . urlencode($what); |
| 430 | } else { |
| 431 | $par .= "&startMessage=$startMessage&show_more=0"; |
| 432 | } |
| 433 | $par .= '&response=' . urlencode($response) . |
| 434 | '&message=' . urlencode($message). |
| 435 | '&topline=' . urlencode($topline); |
| 436 | |
| 437 | echo '<tt><br>' . |
| 438 | '<table width="80%"><tr>' . |
| 439 | '<tr><td colspan=2>' . |
| 440 | _("Body retrieval error. The reason for this is most probably that the message is malformed. Please help us making future versions better by submitting this message to the developers knowledgebase!") . |
| 441 | " <A HREF=\"../src/retrievalerror.php?$par\"><br>" . |
| 442 | _("Submit message") . '</A><BR> ' . |
| 443 | '</td></tr>' . |
| 444 | '<td><b>' . _("Command:") . "</td><td>$cmd</td></tr>" . |
| 445 | '<td><b>' . _("Response:") . "</td><td>$response</td></tr>" . |
| 446 | '<td><b>' . _("Message:") . "</td><td>$message</td></tr>" . |
| 447 | '<td><b>' . _("FETCH line:") . "</td><td>$topline</td></tr>" . |
| 448 | "</table><BR></tt></font><hr>"; |
| 449 | |
| 450 | $data = sqimap_run_command ($imap_stream, "FETCH $passed_id BODY[]", true, $response, $message); |
| 451 | array_shift($data); |
| 452 | $wholemessage = implode('', $data); |
| 453 | |
| 454 | $ret = $wholemessage; |
| 455 | } |
| 456 | return( $ret ); |
| 457 | } |
| 458 | |
| 459 | function mime_print_body_lines ($imap_stream, $id, $ent_id, $encoding) { |
| 460 | // do a bit of error correction. If we couldn't find the entity id, just guess |
| 461 | // that it is the first one. That is usually the case anyway. |
| 462 | if (!$ent_id) { |
| 463 | $ent_id = 1; |
| 464 | } |
| 465 | $sid = sqimap_session_id(); |
| 466 | // Don't kill the connection if the browser is over a dialup |
| 467 | // and it would take over 30 seconds to download it. |
| 468 | |
| 469 |