includes/media/XMP.php

   1 <?php
   2 /**
   3  * Reader for XMP data containing properties relevant to images.
   4  *
   5  * This program is free software; you can redistribute it and/or modify
   6  * it under the terms of the GNU General Public License as published by
   7  * the Free Software Foundation; either version 2 of the License, or
   8  * (at your option) any later version.
   9  *
  10  * This program is distributed in the hope that it will be useful,
  11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  13  * GNU General Public License for more details.
  14  *
  15  * You should have received a copy of the GNU General Public License along
  16  * with this program; if not, write to the Free Software Foundation, Inc.,
  17  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  18  * http://www.gnu.org/copyleft/gpl.html
  19  *
  20  * @file
  21  * @ingroup Media
  22  */
  23
  24 /**
  25 * Class for reading xmp data containing properties relevant to
  26 * images, and spitting out an array that FormatExif accepts.
  27 *
  28 * Note, this is not meant to recognize every possible thing you can
  29 * encode in XMP. It should recognize all the properties we want.
  30 * For example it doesn't have support for structures with multiple
  31 * nesting levels, as none of the properties we're supporting use that
  32 * feature. If it comes across properties it doesn't recognize, it should
  33 * ignore them.
  34 *
  35 * The public methods one would call in this class are
  36 * - parse( $content )
  37 *       Reads in xmp content.
  38 *       Can potentially be called multiple times with partial data each time.
  39 * - parseExtended( $content )
  40 *       Reads XMPExtended blocks (jpeg files only).
  41 * - getResults
  42 *       Outputs a results array.
  43 *
  44 * Note XMP kind of looks like rdf. They are not the same thing - XMP is
  45 * encoded as a specific subset of rdf. This class can read XMP. It cannot
  46 * read rdf.
  47 *
  48 */
  49 class XMPReader {
  50
  51         private $curItem = array();        // array to hold the current element (and previous element, and so on)
  52         private $ancestorStruct = false;   // the structure name when processing nested structures.
  53         private $charContent = false;      // temporary holder for character data that appears in xmp doc.
  54         private $mode = array();           // stores the state the xmpreader is in (see MODE_FOO constants)
  55         private $results = array();        // array to hold results
  56         private $processingArray = false;  // if we're doing a seq or bag.
  57         private $itemLang = false;         // used for lang alts only
  58
  59         private $xmlParser;
  60         private $charset = false;
  61         private $extendedXMPOffset = 0;
  62
  63         protected $items;
  64
  65         /**
  66         * These are various mode constants.
  67         * they are used to figure out what to do
  68         * with an element when its encountered.
  69         *
  70         * For example, MODE_IGNORE is used when processing
  71         * a property we're not interested in. So if a new
  72         * element pops up when we're in that mode, we ignore it.
  73         */
  74         const MODE_INITIAL = 0;
  75         const MODE_IGNORE  = 1;
  76         const MODE_LI      = 2;
  77         const MODE_LI_LANG = 3;
  78         const MODE_QDESC   = 4;
  79
  80         // The following MODE constants are also used in the
  81         // $items array to denote what type of property the item is.
  82         const MODE_SIMPLE    = 10;
  83         const MODE_STRUCT    = 11; // structure (associative array)
  84         const MODE_SEQ       = 12; // ordered list
  85         const MODE_BAG       = 13; // unordered list
  86         const MODE_LANG      = 14;
  87         const MODE_ALT       = 15; // non-language alt. Currently not implemented, and not needed atm.
  88         const MODE_BAGSTRUCT = 16; // A BAG of Structs.
  89
  90         const NS_RDF = 'http://www.w3.org/1999/02/22-rdf-syntax-ns#';
  91         const NS_XML = 'http://www.w3.org/XML/1998/namespace';
  92
  93
  94         /**
  95         * Constructor.
  96         *
  97         * Primary job is to initialize the XMLParser
  98         */
  99         function __construct() {
 100
 101                 if ( !function_exists( 'xml_parser_create_ns' ) ) {
 102                         // this should already be checked by this point
 103                         throw new MWException( 'XMP support requires XML Parser' );
 104                 }
 105
 106                 $this->items = XMPInfo::getItems();
 107
 108                 $this->resetXMLParser();
 109
 110         }
 111         /**
 112         * Main use is if a single item has multiple xmp documents describing it.
 113         * For example in jpeg's with extendedXMP
 114         */
 115         private function resetXMLParser() {
 116
 117                 if ($this->xmlParser) {
 118                         //is this needed?
 119                         xml_parser_free( $this->xmlParser );
 120                 }
 121
 122                 $this->xmlParser = xml_parser_create_ns( 'UTF-8', ' ' );
 123                 xml_parser_set_option( $this->xmlParser, XML_OPTION_CASE_FOLDING, 0 );
 124                 xml_parser_set_option( $this->xmlParser, XML_OPTION_SKIP_WHITE, 1 );
 125
 126                 xml_set_element_handler( $this->xmlParser,
 127                         array( $this, 'startElement' ),
 128                         array( $this, 'endElement' ) );
 129
 130                 xml_set_character_data_handler( $this->xmlParser, array( $this, 'char' ) );
 131         }
 132
 133         /** Destroy the xml parser
 134         *
 135         * Not sure if this is actually needed.
 136         */
 137         function __destruct() {
 138                 // not sure if this is needed.
 139                 xml_parser_free( $this->xmlParser );
 140         }
 141
 142         /** Get the result array. Do some post-processing before returning
 143         * the array, and transform any metadata that is special-cased.
 144         *
 145         * @return Array array of results as an array of arrays suitable for
 146         *       FormatMetadata::getFormattedData().
 147         */
 148         public function getResults() {
 149                 // xmp-special is for metadata that affects how stuff
 150                 // is extracted. For example xmpNote:HasExtendedXMP.
 151
 152                 // It is also used to handle photoshop:AuthorsPosition
 153                 // which is weird and really part of another property,
 154                 // see 2:85 in IPTC. See also pg 21 of IPTC4XMP standard.
 155                 // The location fields also use it.
 156
 157                 $data = $this->results;
 158
 159                 wfRunHooks('XMPGetResults', Array(&$data));
 160
 161                 if ( isset( $data['xmp-special']['AuthorsPosition'] )
 162                         && is_string( $data['xmp-special']['AuthorsPosition'] )
 163                         && isset( $data['xmp-general']['Artist'][0] )
 164                 ) {
 165                         // Note, if there is more than one creator,
 166                         // this only applies to first. This also will
 167                         // only apply to the dc:Creator prop, not the
 168                         // exif:Artist prop.
 169
 170                         $data['xmp-general']['Artist'][0] =
 171                                 $data['xmp-special']['AuthorsPosition'] . ', '
 172                                 . $data['xmp-general']['Artist'][0];
 173                 }
 174
 175                 // Go through the LocationShown and LocationCreated
 176                 // changing it to the non-hierarchal form used by
 177                 // the other location fields.
 178
 179                 if ( isset( $data['xmp-special']['LocationShown'][0] )
 180                         && is_array( $data['xmp-special']['LocationShown'][0] )
 181                 ) {
 182                         // the is_array is just paranoia. It should always
 183                         // be an array.
 184                         foreach( $data['xmp-special']['LocationShown'] as $loc ) {
 185                                 if ( !is_array( $loc ) ) {
 186                                         // To avoid copying over the _type meta-fields.
 187                                         continue;
 188                                 }
 189                                 foreach( $loc as $field => $val ) {
 190                                         $data['xmp-general'][$field . 'Dest'][] = $val;
 191                                 }
 192                         }
 193                 }
 194                 if ( isset( $data['xmp-special']['LocationCreated'][0] )
 195                         && is_array( $data['xmp-special']['LocationCreated'][0] )
 196                 ) {
 197                         // the is_array is just paranoia. It should always
 198                         // be an array.
 199                         foreach( $data['xmp-special']['LocationCreated'] as $loc ) {
 200                                 if ( !is_array( $loc ) ) {
 201                                         // To avoid copying over the _type meta-fields.
 202                                         continue;
 203                                 }
 204                                 foreach(  $loc as $field => $val ) {
 205                                         $data['xmp-general'][$field . 'Created'][] = $val;
 206                                 }
 207                         }
 208                 }
 209
 210
 211                 // We don't want to return the special values, since they're
 212                 // special and not info to be stored about the file.
 213                 unset( $data['xmp-special'] );
 214
 215                 // Convert GPSAltitude to negative if below sea level.
 216                 if ( isset( $data['xmp-exif']['GPSAltitudeRef'] )
 217                         && isset( $data['xmp-exif']['GPSAltitude'] )
 218                 ) {
 219
 220                         // Must convert to a real before multiplying by -1
 221                         // XMPValidate guarantees there will always be a '/' in this value.
 222                         list( $nom, $denom ) = explode( '/', $data['xmp-exif']['GPSAltitude'] );
 223                         $data['xmp-exif']['GPSAltitude'] = $nom / $denom;
 224
 225                         if ( $data['xmp-exif']['GPSAltitudeRef'] == '1' ) {
 226                                 $data['xmp-exif']['GPSAltitude'] *= -1;
 227                         }
 228                         unset( $data['xmp-exif']['GPSAltitudeRef'] );
 229                 }
 230
 231                 return $data;
 232         }
 233
 234         /**
 235          * Main function to call to parse XMP. Use getResults to
 236          * get results.
 237          *
 238          * Also catches any errors during processing, writes them to
 239          * debug log, blanks result array and returns false.
 240          *
 241          * @param $content String: XMP data
 242          * @param $allOfIt Boolean: If this is all the data (true) or if its split up (false). Default true
 243          * @param $reset Boolean: does xml parser need to be reset. Default false
 244          * @throws MWException
 245          * @return Boolean success.
 246          */
 247         public function parse( $content, $allOfIt = true, $reset = false ) {
 248                 if ( $reset ) {
 249                         $this->resetXMLParser();
 250                 }
 251                 try {
 252
 253                         // detect encoding by looking for BOM which is supposed to be in processing instruction.
 254                         // see page 12 of http://www.adobe.com/devnet/xmp/pdfs/XMPSpecificationPart3.pdf
 255                         if ( !$this->charset ) {
 256                                 $bom = array();
 257                                 if ( preg_match( '/\xEF\xBB\xBF|\xFE\xFF|\x00\x00\xFE\xFF|\xFF\xFE\x00\x00|\xFF\xFE/',
 258                                          $content, $bom )
 259                                 ) {
 260                                         switch ( $bom[0] ) {
 261                                                 case "\xFE\xFF":
 262                                                         $this->charset = 'UTF-16BE';
 263                                                         break;
 264                                                 case "\xFF\xFE":
 265                                                         $this->charset = 'UTF-16LE';
 266                                                         break;
 267                                                 case "\x00\x00\xFE\xFF":
 268                                                         $this->charset = 'UTF-32BE';
 269                                                         break;
 270                                                 case "\xFF\xFE\x00\x00":
 271                                                         $this->charset = 'UTF-32LE';
 272                                                         break;
 273                                                 case "\xEF\xBB\xBF":
 274                                                         $this->charset = 'UTF-8';
 275                                                         break;
 276                                                 default:
 277                                                         //this should be impossible to get to
 278                                                         throw new MWException("Invalid BOM");
 279                                                         break;
 280
 281                                         }
 282
 283                                 } else {
 284                                         // standard specifically says, if no bom assume utf-8
 285                                         $this->charset = 'UTF-8';
 286                                 }
 287                         }
 288                         if ( $this->charset !== 'UTF-8' ) {
 289                                 //don't convert if already utf-8
 290                                 wfSuppressWarnings();
 291                                 $content = iconv( $this->charset, 'UTF-8//IGNORE', $content );
 292                                 wfRestoreWarnings();
 293                         }
 294
 295                         $ok = xml_parse( $this->xmlParser, $content, $allOfIt );
 296                         if ( !$ok ) {
 297                                 $error = xml_error_string( xml_get_error_code( $this->xmlParser ) );
 298                                 $where = 'line: ' . xml_get_current_line_number( $this->xmlParser )
 299                                         . ' column: ' . xml_get_current_column_number( $this->xmlParser )
 300                                         . ' byte offset: ' . xml_get_current_byte_index( $this->xmlParser );
 301
 302                                 wfDebugLog( 'XMP', "XMPReader::parse : Error reading XMP content: $error ($where)" );
 303                                 $this->results = array(); // blank if error.
 304                                 return false;
 305                         }
 306                 } catch ( MWException $e ) {
 307                         wfDebugLog( 'XMP', 'XMP parse error: ' . $e );
 308                         $this->results = array();
 309                         return false;
 310                 }
 311                 return true;
 312         }
 313
 314         /** Entry point for XMPExtended blocks in jpeg files
 315          *
 316          * @todo In serious need of testing
 317          * @see http://www.adobe.ge/devnet/xmp/pdfs/XMPSpecificationPart3.pdf XMP spec part 3 page 20
 318          * @param String $content XMPExtended block minus the namespace signature
 319          * @return Boolean If it succeeded.
 320          */
 321         public function parseExtended( $content ) {
 322                 // @todo FIXME: This is untested. Hard to find example files
 323                 // or programs that make such files..
 324                 $guid = substr( $content, 0, 32 );
 325                 if ( !isset( $this->results['xmp-special']['HasExtendedXMP'] )
 326                         || $this->results['xmp-special']['HasExtendedXMP'] !== $guid ) {
 327                         wfDebugLog('XMP', __METHOD__ . " Ignoring XMPExtended block due to wrong guid (guid= '$guid' )");
 328                         return false;
 329                 }
 330                 $len  = unpack( 'Nlength/Noffset', substr( $content, 32, 8 ) );
 331
 332                 if (!$len || $len['length'] < 4 || $len['offset'] < 0 || $len['offset'] > $len['length'] ) {
 333                         wfDebugLog('XMP', __METHOD__ . 'Error reading extended XMP block, invalid length or offset.');
 334                         return false;
 335                 }
 336
 337
 338                 // we're not very robust here. we should accept it in the wrong order. To quote
 339                 // the xmp standard:
 340                 // "A JPEG writer should write the ExtendedXMP marker segments in order, immediately following the
 341                 // StandardXMP. However, the JPEG standard does not require preservation of marker segment order. A
 342                 // robust JPEG reader should tolerate the marker segments in any order."
 343                 //
 344                 // otoh the probability that an image will have more than 128k of metadata is rather low...
 345                 // so the probability that it will have > 128k, and be in the wrong order is very low...
 346
 347                 if ( $len['offset'] !== $this->extendedXMPOffset ) {
 348                         wfDebugLog('XMP', __METHOD__ . 'Ignoring XMPExtended block due to wrong order. (Offset was '
 349                                 . $len['offset'] . ' but expected ' . $this->extendedXMPOffset . ')');
 350                         return false;
 351                 }
 352
 353                 if ( $len['offset'] === 0 ) {
 354                         // if we're starting the extended block, we've probably already
 355                         // done the XMPStandard block, so reset.
 356                         $this->resetXMLParser();
 357                 }
 358
 359                 $this->extendedXMPOffset += $len['length'];
 360
 361                 $actualContent = substr( $content, 40 );
 362
 363                 if ( $this->extendedXMPOffset === strlen( $actualContent ) ) {
 364                         $atEnd = true;
 365                 } else {
 366                         $atEnd = false;
 367                 }
 368
 369                 wfDebugLog('XMP', __METHOD__ . 'Parsing a XMPExtended block');
 370                 return $this->parse( $actualContent, $atEnd );
 371         }
 372
 373         /**
 374         * Character data handler
 375         * Called whenever character data is found in the xmp document.
 376         *
 377         * does nothing if we're in MODE_IGNORE or if the data is whitespace
 378         * throws an error if we're not in MODE_SIMPLE (as we're not allowed to have character
 379         * data in the other modes).
 380         *
 381         * As an example, this happens when we encounter XMP like:
 382         * <exif:DigitalZoomRatio>0/10</exif:DigitalZoomRatio>
 383         * and are processing the 0/10 bit.
 384         *
 385         * @param $parser XMLParser reference to the xml parser
 386         * @param $data String Character data
 387         * @throws MWException on invalid data
 388         */
 389         function char( $parser, $data ) {
 390
 391                 $data = trim( $data );
 392                 if ( trim( $data ) === "" ) {
 393                         return;
 394                 }
 395
 396                 if ( !isset( $this->mode[0] ) ) {
 397                         throw new MWException( 'Unexpected character data before first rdf:Description element' );
 398                 }
 399
 400                 if ( $this->mode[0] === self::MODE_IGNORE ) return;
 401
 402                 if ( $this->mode[0] !== self::MODE_SIMPLE
 403                         && $this->mode[0] !== self::MODE_QDESC
 404                 ) {
 405                         throw new MWException( 'character data where not expected. (mode ' . $this->mode[0] . ')' );
 406                 }
 407
 408                 // to check, how does this handle w.s.
 409                 if ( $this->charContent === false ) {
 410                         $this->charContent = $data;
 411                 } else {
 412                         $this->charContent .= $data;
 413                 }
 414
 415         }
 416
 417         /** When we hit a closing element in MODE_IGNORE
 418         * Check to see if this is the element we started to ignore,
 419         * in which case we get out of MODE_IGNORE
 420         *
 421         * @param $elm String Namespace of element followed by a space and then tag name of element.
 422         */
 423         private function endElementModeIgnore ( $elm ) {
 424
 425                 if ( $this->curItem[0] === $elm ) {
 426                         array_shift( $this->curItem );
 427                         array_shift( $this->mode );
 428                 }
 429                 return;
 430
 431         }
 432
 433         /**
 434         * Hit a closing element when in MODE_SIMPLE.
 435         * This generally means that we finished processing a
 436         * property value, and now have to save the result to the
 437         * results array
 438         *
 439         * For example, when processing:
 440         * <exif:DigitalZoomRatio>0/10</exif:DigitalZoomRatio>
 441         * this deals with when we hit </exif:DigitalZoomRatio>.
 442         *
 443         * Or it could be if we hit the end element of a property
 444         * of a compound data structure (like a member of an array).
 445         *
 446         * @param $elm String namespace, space, and tag name.
 447         */
 448         private function endElementModeSimple ( $elm ) {
 449                 if ( $this->charContent !== false ) {
 450                         if ( $this->processingArray ) {
 451                                 // if we're processing an array, use the original element
 452                                 // name instead of rdf:li.
 453                                 list( $ns, $tag ) = explode( ' ', $this->curItem[0], 2 );
 454                         } else {
 455                                 list( $ns, $tag ) = explode( ' ', $elm, 2 );
 456                         }
 457                         $this->saveValue( $ns, $tag, $this->charContent );
 458
 459                         $this->charContent = false; // reset
 460                 }
 461                 array_shift( $this->curItem );
 462                 array_shift( $this->mode );
 463
 464         }
 465
 466         /**
 467          * Hit a closing element in MODE_STRUCT, MODE_SEQ, MODE_BAG
 468          * generally means we've finished processing a nested structure.
 469          * resets some internal variables to indicate that.
 470          *
 471          * Note this means we hit the closing element not the "</rdf:Seq>".
 472          *
 473          * @par For example, when processing:
 474          * @code{,xml}
 475          * <exif:ISOSpeedRatings> <rdf:Seq> <rdf:li>64</rdf:li>
 476          *   </rdf:Seq> </exif:ISOSpeedRatings>
 477          * @endcode
 478          *
 479          * This method is called when we hit the "</exif:ISOSpeedRatings>" tag.
 480          *
 481          * @param $elm String namespace . space . tag name.
 482          * @throws MWException
 483          */
 484         private function endElementNested( $elm ) {
 485
 486                 /* cur item must be the same as $elm, unless if in MODE_STRUCT
 487                    in which case it could also be rdf:Description */
 488                 if ( $this->curItem[0] !== $elm
 489                         && !( $elm === self::NS_RDF . ' Description'
 490                                 && $this->mode[0] === self::MODE_STRUCT )
 491                  ) {
 492                         throw new MWException( "nesting mismatch. got a </$elm> but expected a </" . $this->curItem[0] . '>' );
 493                 }
 494
 495                 // Validate structures.
 496                 list( $ns, $tag ) = explode( ' ', $elm, 2 );
 497                 if ( isset( $this->items[$ns][$tag]['validate'] ) ) {
 498
 499                         $info =& $this->items[$ns][$tag];
 500                         $finalName = isset( $info['map_name'] )
 501                                 ? $info['map_name'] : $tag;
 502
 503                         $validate = is_array( $info['validate'] ) ? $info['validate']
 504                                 : array( 'XMPValidate', $info['validate'] );
 505
 506                         if ( !isset( $this->results['xmp-' . $info['map_group']][$finalName] ) ) {
 507                                 // This can happen if all the members of the struct failed validation.
 508                                 wfDebugLog( 'XMP', __METHOD__ . " <$ns:$tag> has no valid members." );
 509
 510                         } elseif ( is_callable( $validate ) ) {
 511                                 $val =& $this->results['xmp-' . $info['map_group']][$finalName];
 512                                 call_user_func_array( $validate, array( $info, &$val, false ) );
 513                                 if ( is_null( $val ) ) {
 514                                         // the idea being the validation function will unset the variable if
 515                                         // its invalid.
 516                                         wfDebugLog( 'XMP', __METHOD__ . " <$ns:$tag> failed validation." );
 517                                         unset( $this->results['xmp-' . $info['map_group']][$finalName] );
 518                                 }
 519                         } else {
 520                                 wfDebugLog( 'XMP', __METHOD__ . " Validation function for $finalName ("
 521                                         . $validate[0] . '::' . $validate[1] . '()) is not callable.' );
 522                         }
 523                 }
 524
 525                 array_shift( $this->curItem );
 526                 array_shift( $this->mode );
 527                 $this->ancestorStruct = false;
 528                 $this->processingArray = false;
 529                 $this->itemLang = false;
 530         }
 531
 532         /**
 533          * Hit a closing element in MODE_LI (either rdf:Seq, or rdf:Bag )
 534          * Add information about what type of element this is.
 535          *
 536          * Note we still have to hit the outer "</property>"
 537          *
 538          * @par For example, when processing:
 539          * @code{,xml}
 540          * <exif:ISOSpeedRatings> <rdf:Seq> <rdf:li>64</rdf:li>
 541          *   </rdf:Seq> </exif:ISOSpeedRatings>
 542          * @endcode
 543          *
 544          * This method is called when we hit the "</rdf:Seq>".
 545          * (For comparison, we call endElementModeSimple when we
 546          * hit the "</rdf:li>")
 547          *
 548          * @param $elm String namespace . ' ' . element name
 549          * @throws MWException
 550          */
 551         private function endElementModeLi( $elm ) {
 552
 553                 list( $ns, $tag ) = explode( ' ', $this->curItem[0], 2 );
 554                 $info = $this->items[$ns][$tag];
 555                 $finalName = isset( $info['map_name'] )
 556                         ? $info['map_name'] : $tag;
 557
 558                 array_shift( $this->mode );
 559
 560                 if ( !isset( $this->results['xmp-' . $info['map_group']][$finalName] ) ) {
 561                         wfDebugLog( 'XMP', __METHOD__ . " Empty compund element $finalName." );
 562                         return;
 563                 }
 564
 565                 if ( $elm === self::NS_RDF . ' Seq' ) {
 566                         $this->results['xmp-' . $info['map_group']][$finalName]['_type'] = 'ol';
 567                 } elseif ( $elm === self::NS_RDF . ' Bag' ) {
 568                         $this->results['xmp-' . $info['map_group']][$finalName]['_type'] = 'ul';
 569                 } elseif ( $elm === self::NS_RDF . ' Alt' ) {
 570                         // extra if needed as you could theoretically have a non-language alt.
 571                         if ( $info['mode'] === self::MODE_LANG ) {
 572                                 $this->results['xmp-' . $info['map_group']][$finalName]['_type'] = 'lang';
 573                         }
 574
 575                 } else {
 576                         throw new MWException( __METHOD__ . " expected </rdf:seq> or </rdf:bag> but instead got $elm." );
 577                 }
 578         }
 579
 580         /**
 581         * End element while in MODE_QDESC
 582         * mostly when ending an element when we have a simple value
 583         * that has qualifiers.
 584         *
 585         * Qualifiers aren't all that common, and we don't do anything
 586         * with them.
 587         *
 588         * @param $elm String namespace and element
 589         */
 590         private function endElementModeQDesc( $elm ) {
 591
 592                 if ( $elm === self::NS_RDF . ' value' ) {
 593                         list( $ns, $tag ) = explode( ' ', $this->curItem[0], 2 );
 594                         $this->saveValue( $ns, $tag, $this->charContent );
 595                         return;
 596                 } else {
 597                         array_shift( $this->mode );
 598                         array_shift( $this->curItem );
 599                 }
 600
 601
 602         }
 603
 604         /**
 605          * Handler for hitting a closing element.
 606          *
 607          * generally just calls a helper function depending on what
 608          * mode we're in.
 609          *
 610          * Ignores the outer wrapping elements that are optional in
 611          * xmp and have no meaning.
 612          *
 613          * @param $parser XMLParser
 614          * @param $elm String namespace . ' ' . element name
 615          * @throws MWException
 616          */
 617         function endElement( $parser, $elm ) {
 618                 if ( $elm === ( self::NS_RDF . ' RDF' )
 619                         || $elm === 'adobe:ns:meta/ xmpmeta'
 620                         || $elm === 'adobe:ns:meta/ xapmeta' )
 621                 {
 622                         // ignore these.
 623                         return;
 624                 }
 625
 626                 if ( $elm === self::NS_RDF . ' type' ) {
 627                         // these aren't really supported properly yet.
 628                         // However, it appears they almost never used.
 629                         wfDebugLog( 'XMP', __METHOD__ . ' encountered <rdf:type>' );
 630                 }
 631
 632                 if ( strpos( $elm, ' ' ) === false ) {
 633                         // This probably shouldn't happen.
 634                         // However, there is a bug in an adobe product
 635                         // that forgets the namespace on some things.
 636                         // (Luckily they are unimportant things).
 637                         wfDebugLog( 'XMP', __METHOD__ . " Encountered </$elm> which has no namespace. Skipping." );
 638                         return;
 639                 }
 640
 641                 if ( count( $this->mode[0] ) === 0 ) {
 642                         // This should never ever happen and means
 643                         // there is a pretty major bug in this class.
 644                         throw new MWException( 'Encountered end element with no mode' );
 645                 }
 646
 647                 if ( count( $this->curItem ) == 0 && $this->mode[0] !== self::MODE_INITIAL ) {
 648                         // just to be paranoid. Should always have a curItem, except for initially
 649                         // (aka during MODE_INITAL).
 650                         throw new MWException( "Hit end element </$elm> but no curItem" );
 651                 }
 652
 653                 switch( $this->mode[0] ) {
 654                         case self::MODE_IGNORE:
 655                                 $this->endElementModeIgnore( $elm );
 656                                 break;
 657                         case self::MODE_SIMPLE:
 658                                 $this->endElementModeSimple( $elm );
 659                                 break;
 660                         case self::MODE_STRUCT:
 661                         case self::MODE_SEQ:
 662                         case self::MODE_BAG:
 663                         case self::MODE_LANG:
 664                         case self::MODE_BAGSTRUCT:
 665                                 $this->endElementNested( $elm );
 666                                 break;
 667                         case self::MODE_INITIAL:
 668                                 if ( $elm === self::NS_RDF . ' Description' ) {
 669                                         array_shift( $this->mode );
 670                                 } else {
 671                                         throw new MWException( 'Element ended unexpectedly while in MODE_INITIAL' );
 672                                 }
 673                                 break;
 674                         case self::MODE_LI:
 675                         case self::MODE_LI_LANG:
 676                                 $this->endElementModeLi( $elm );
 677                                 break;
 678                         case self::MODE_QDESC:
 679                                 $this->endElementModeQDesc( $elm );
 680                                 break;
 681                         default:
 682                                 wfDebugLog( 'XMP', __METHOD__ . " no mode (elm = $elm)" );
 683                                 break;
 684                 }
 685         }
 686
 687         /**
 688         * Hit an opening element while in MODE_IGNORE
 689         *
 690         * XMP is extensible, so ignore any tag we don't understand.
 691         *
 692         * Mostly ignores, unless we encounter the element that we are ignoring.
 693         * in which case we add it to the item stack, so we can ignore things
 694         * that are nested, correctly.
 695         *
 696         * @param $elm String namespace . ' ' . tag name
 697         */
 698         private function startElementModeIgnore( $elm ) {
 699                 if ( $elm === $this->curItem[0] ) {
 700                         array_unshift( $this->curItem, $elm );
 701                         array_unshift( $this->mode, self::MODE_IGNORE );
 702                 }
 703         }
 704
 705         /**
 706         *  Start element in MODE_BAG (unordered array)
 707         * this should always be <rdf:Bag>
 708         *
 709         * @param $elm String namespace . ' ' . tag
 710         * @throws MWException if we have an element that's not <rdf:Bag>
 711         */
 712         private function startElementModeBag( $elm ) {
 713                 if ( $elm === self::NS_RDF . ' Bag' ) {
 714                         array_unshift( $this->mode, self::MODE_LI );
 715                 } else {
 716                         throw new MWException( "Expected <rdf:Bag> but got $elm." );
 717                 }
 718
 719         }
 720
 721         /**
 722         * Start element in MODE_SEQ (ordered array)
 723         * this should always be <rdf:Seq>
 724         *
 725         * @param $elm String namespace . ' ' . tag
 726         * @throws MWException if we have an element that's not <rdf:Seq>
 727         */
 728         private function startElementModeSeq( $elm ) {
 729                 if ( $elm === self::NS_RDF . ' Seq' ) {
 730                         array_unshift( $this->mode, self::MODE_LI );
 731                 } elseif ( $elm === self::NS_RDF . ' Bag' ) {
 732                         # bug 27105
 733                         wfDebugLog( 'XMP', __METHOD__ . ' Expected an rdf:Seq, but got an rdf:Bag. Pretending'
 734                                 . ' it is a Seq, since some buggy software is known to screw this up.' );
 735                         array_unshift( $this->mode, self::MODE_LI );
 736                 } else {
 737                         throw new MWException( "Expected <rdf:Seq> but got $elm." );
 738                 }
 739
 740         }
 741
 742         /**
 743         * Start element in MODE_LANG (language alternative)
 744         * this should always be <rdf:Alt>
 745         *
 746         * This tag tends to be used for metadata like describe this
 747         * picture, which can be translated into multiple languages.
 748         *
 749         * XMP supports non-linguistic alternative selections,
 750         * which are really only used for thumbnails, which
 751         * we don't care about.
 752         *
 753         * @param $elm String namespace . ' ' . tag
 754         * @throws MWException if we have an element that's not <rdf:Alt>
 755         */
 756         private function startElementModeLang( $elm ) {
 757                 if ( $elm === self::NS_RDF . ' Alt' ) {
 758                         array_unshift( $this->mode, self::MODE_LI_LANG );
 759                 } else {
 760                         throw new MWException( "Expected <rdf:Seq> but got $elm." );
 761                 }
 762
 763         }
 764
 765         /**
 766          * Handle an opening element when in MODE_SIMPLE
 767          *
 768          * This should not happen often. This is for if a simple element
 769          * already opened has a child element. Could happen for a
 770          * qualified element.
 771          *
 772          * For example:
 773          * <exif:DigitalZoomRatio><rdf:Description><rdf:value>0/10</rdf:value>
 774          *   <foo:someQualifier>Bar</foo:someQualifier> </rdf:Description>
 775          *   </exif:DigitalZoomRatio>
 776          *
 777          * This method is called when processing the <rdf:Description> element
 778          *
 779          * @param $elm String namespace and tag names separated by space.
 780          * @param $attribs Array Attributes of the element.
 781          * @throws MWException
 782          */
 783         private function startElementModeSimple( $elm, $attribs ) {
 784                 if ( $elm === self::NS_RDF . ' Description' ) {
 785                         // If this value has qualifiers
 786                         array_unshift( $this->mode, self::MODE_QDESC );
 787                         array_unshift( $this->curItem, $this->curItem[0] );
 788
 789                         if ( isset( $attribs[self::NS_RDF . ' value'] ) ) {
 790                                 list( $ns, $tag ) = explode( ' ', $this->curItem[0], 2 );
 791                                 $this->saveValue( $ns, $tag, $attribs[self::NS_RDF . ' value'] );
 792                         }
 793                 } elseif ( $elm === self::NS_RDF . ' value' ) {
 794                         // This should not be here.
 795                         throw new MWException( __METHOD__ . ' Encountered <rdf:value> where it was unexpected.' );
 796
 797                 } else {
 798                         // something else we don't recognize, like a qualifier maybe.
 799                         wfDebugLog( 'XMP', __METHOD__ . " Encountered element <$elm> where only expecting character data as value of " . $this->curItem[0] );
 800                         array_unshift( $this->mode, self::MODE_IGNORE );
 801                         array_unshift( $this->curItem, $elm );
 802
 803                 }
 804
 805         }
 806
 807         /**
 808         * Start an element when in MODE_QDESC.
 809         * This generally happens when a simple element has an inner
 810         * rdf:Description to hold qualifier elements.
 811         *
 812         * For example in:
 813         * <exif:DigitalZoomRatio><rdf:Description><rdf:value>0/10</rdf:value>
 814         *   <foo:someQualifier>Bar</foo:someQualifier> </rdf:Description>
 815         *   </exif:DigitalZoomRatio>
 816         * Called when processing the <rdf:value> or <foo:someQualifier>.
 817         *
 818         * @param $elm String namespace and tag name separated by a space.
 819         *
 820         */
 821         private function startElementModeQDesc( $elm ) {
 822                 if ( $elm === self::NS_RDF . ' value' ) {
 823                         return; // do nothing
 824                 } else {
 825                         // otherwise its a qualifier, which we ignore
 826                         array_unshift( $this->mode, self::MODE_IGNORE );
 827                         array_unshift( $this->curItem, $elm );
 828                 }
 829         }
 830
 831         /**
 832          * Starting an element when in MODE_INITIAL
 833          * This usually happens when we hit an element inside
 834          * the outer rdf:Description
 835          *
 836          * This is generally where most properties start.
 837          *
 838          * @param $ns String Namespace
 839          * @param $tag String tag name (without namespace prefix)
 840          * @param $attribs Array array of attributes
 841          * @throws MWException
 842          */
 843         private function startElementModeInitial( $ns, $tag, $attribs ) {
 844                 if ( $ns !== self::NS_RDF ) {
 845
 846                         if ( isset( $this->items[$ns][$tag] ) ) {
 847                                 if ( isset( $this->items[$ns][$tag]['structPart'] ) ) {
 848                                         // If this element is supposed to appear only as
 849                                         // a child of a structure, but appears here (not as
 850                                         // a child of a struct), then something weird is
 851                                         // happening, so ignore this element and its children.
 852
 853                                         wfDebugLog( 'XMP', "Encountered <$ns:$tag> outside"
 854                                                 . " of its expected parent. Ignoring." );
 855
 856                                         array_unshift( $this->mode, self::MODE_IGNORE );
 857                                         array_unshift( $this->curItem, $ns . ' ' . $tag );
 858                                         return;
 859                                 }
 860                                 $mode = $this->items[$ns][$tag]['mode'];
 861                                 array_unshift( $this->mode, $mode );
 862                                 array_unshift( $this->curItem, $ns . ' ' . $tag );
 863                                 if ( $mode === self::MODE_STRUCT ) {
 864                                         $this->ancestorStruct = isset( $this->items[$ns][$tag]['map_name'] )
 865                                                 ? $this->items[$ns][$tag]['map_name'] : $tag;
 866                                 }
 867                                 if ( $this->charContent !== false ) {
 868                                         // Something weird.
 869                                         // Should not happen in valid XMP.
 870                                         throw new MWException( 'tag nested in non-whitespace characters.' );
 871                                 }
 872                         } else {
 873                                 // This element is not on our list of allowed elements so ignore.
 874                                 wfDebugLog( 'XMP', __METHOD__ . " Ignoring unrecognized element <$ns:$tag>." );
 875                                 array_unshift( $this->mode, self::MODE_IGNORE );
 876                                 array_unshift( $this->curItem, $ns . ' ' . $tag );
 877                                 return;
 878                         }
 879
 880                 }
 881                 // process attributes
 882                 $this->doAttribs( $attribs );
 883         }
 884
 885         /**
 886          * Hit an opening element when in a Struct (MODE_STRUCT)
 887          * This is generally for fields of a compound property.
 888          *
 889          * Example of a struct (abbreviated; flash has more properties):
 890          *
 891          * <exif:Flash> <rdf:Description> <exif:Fired>True</exif:Fired>
 892          *  <exif:Mode>1</exif:Mode></rdf:Description></exif:Flash>
 893          *
 894          * or:
 895          *
 896          * <exif:Flash rdf:parseType='Resource'> <exif:Fired>True</exif:Fired>
 897          *  <exif:Mode>1</exif:Mode></exif:Flash>
 898          *
 899          * @param $ns String namespace
 900          * @param $tag String tag name (no ns)
 901          * @param $attribs Array array of attribs w/ values.
 902          * @throws MWException
 903          */
 904         private function startElementModeStruct( $ns, $tag, $attribs ) {
 905                 if ( $ns !== self::NS_RDF ) {
 906
 907                         if ( isset( $this->items[$ns][$tag] ) ) {
 908                                 if ( isset( $this->items[$ns][$this->ancestorStruct]['children'] )
 909                                         && !isset( $this->items[$ns][$this->ancestorStruct]['children'][$tag] ) )
 910                                 {
 911                                         // This assumes that we don't have inter-namespace nesting
 912                                         // which we don't in all the properties we're interested in.
 913                                         throw new MWException( " <$tag> appeared nested in <" . $this->ancestorStruct
 914                                                 . "> where it is not allowed." );
 915                                 }
 916                                 array_unshift( $this->mode, $this->items[$ns][$tag]['mode'] );
 917                                 array_unshift( $this->curItem, $ns . ' ' . $tag );
 918                                 if ( $this->charContent !== false ) {
 919                                         // Something weird.
 920                                         // Should not happen in valid XMP.
 921                                         throw new MWException( "tag <$tag> nested in non-whitespace characters (" . $this->charContent . ")." );
 922                                 }
 923                         } else {
 924                                 array_unshift( $this->mode, self::MODE_IGNORE );
 925                                 array_unshift( $this->curItem, $elm );
 926                                 return;
 927                         }
 928
 929                 }
 930
 931                 if ( $ns === self::NS_RDF && $tag === 'Description' ) {
 932                         $this->doAttribs( $attribs );
 933                         array_unshift( $this->mode, self::MODE_STRUCT );
 934                         array_unshift( $this->curItem, $this->curItem[0] );
 935                 }
 936         }
 937
 938         /**
 939         * opening element in MODE_LI
 940         * process elements of arrays.
 941         *
 942         * Example:
 943         * <exif:ISOSpeedRatings> <rdf:Seq> <rdf:li>64</rdf:li>
 944         *   </rdf:Seq> </exif:ISOSpeedRatings>
 945         * This method is called when we hit the <rdf:li> element.
 946         *
 947         * @param $elm String: namespace . ' ' . tagname
 948         * @param $attribs Array: Attributes. (needed for BAGSTRUCTS)
 949         * @throws MWException if gets a tag other than <rdf:li>
 950         */
 951         private function startElementModeLi( $elm, $attribs ) {
 952                 if ( ( $elm ) !== self::NS_RDF . ' li' ) {
 953                         throw new MWException( "<rdf:li> expected but got $elm." );
 954                 }
 955
 956                 if ( !isset( $this->mode[1] ) ) {
 957                         // This should never ever ever happen. Checking for it
 958                         // to be paranoid.
 959                         throw new MWException( 'In mode Li, but no 2xPrevious mode!' );
 960                 }
 961
 962                 if ( $this->mode[1] === self::MODE_BAGSTRUCT ) {
 963                         // This list item contains a compound (STRUCT) value.
 964                         array_unshift( $this->mode, self::MODE_STRUCT );
 965                         array_unshift( $this->curItem, $elm );
 966                         $this->processingArray = true;
 967
 968                         if ( !isset( $this->curItem[1] ) ) {
 969                                 // be paranoid.
 970                                 throw new MWException( 'Can not find parent of BAGSTRUCT.' );
 971                         }
 972                         list( $curNS, $curTag ) = explode( ' ', $this->curItem[1] );
 973                         $this->ancestorStruct = isset( $this->items[$curNS][$curTag]['map_name'] )
 974                                 ? $this->items[$curNS][$curTag]['map_name'] : $curTag;
 975
 976                         $this->doAttribs( $attribs );
 977
 978                 } else {
 979                         // Normal BAG or SEQ containing simple values.
 980                         array_unshift( $this->mode, self::MODE_SIMPLE );
 981                         // need to add curItem[0] on again since one is for the specific item
 982                         // and one is for the entire group.
 983                         array_unshift( $this->curItem, $this->curItem[0] );
 984                         $this->processingArray = true;
 985                 }
 986
 987         }
 988
 989         /**
 990         * Opening element in MODE_LI_LANG.
 991         * process elements of language alternatives
 992         *
 993         * Example:
 994         * <dc:title> <rdf:Alt> <rdf:li xml:lang="x-default">My house
 995         *  </rdf:li> </rdf:Alt> </dc:title>
 996         *
 997         * This method is called when we hit the <rdf:li> element.
 998         *
 999         * @param $elm String namespace . ' ' . tag
1000         * @param $attribs array array of elements (most importantly xml:lang)
1001         * @throws MWException if gets a tag other than <rdf:li> or if no xml:lang
1002         */
1003         private function startElementModeLiLang( $elm, $attribs ) {
1004                 if ( $elm !== self::NS_RDF . ' li' ) {
1005                         throw new MWException( __METHOD__ . " <rdf:li> expected but got $elm." );
1006                 }
1007                 if ( !isset( $attribs[ self::NS_XML . ' lang'] )
1008                         || !preg_match( '/^[-A-Za-z0-9]{2,}$/D', $attribs[ self::NS_XML . ' lang' ] ) )
1009                 {
1010                         throw new MWException( __METHOD__
1011                                 . " <rdf:li> did not contain, or has invalid xml:lang attribute in lang alternative" );
1012                 }
1013
1014                 // Lang is case-insensitive.
1015                 $this->itemLang = strtolower( $attribs[ self::NS_XML . ' lang' ] );
1016
1017                 // need to add curItem[0] on again since one is for the specific item
1018                 // and one is for the entire group.
1019                 array_unshift( $this->curItem, $this->curItem[0] );
1020                 array_unshift( $this->mode, self::MODE_SIMPLE );
1021                 $this->processingArray = true;
1022         }
1023
1024         /**
1025          * Hits an opening element.
1026          * Generally just calls a helper based on what MODE we're in.
1027          * Also does some initial set up for the wrapper element
1028          *
1029          * @param $parser XMLParser
1030          * @param $elm String namespace "<space>" element
1031          * @param $attribs Array attribute name => value
1032          * @throws MWException
1033          */
1034         function startElement( $parser, $elm, $attribs ) {
1035
1036                 if ( $elm === self::NS_RDF . ' RDF'
1037                         || $elm === 'adobe:ns:meta/ xmpmeta'
1038                         || $elm === 'adobe:ns:meta/ xapmeta')
1039                 {
1040                         /* ignore. */
1041                         return;
1042                 } elseif ( $elm === self::NS_RDF . ' Description' ) {
1043                         if ( count( $this->mode ) === 0 ) {
1044                                 // outer rdf:desc
1045                                 array_unshift( $this->mode, self::MODE_INITIAL );
1046                         }
1047                 } elseif ( $elm === self::NS_RDF . ' type' ) {
1048                         // This doesn't support rdf:type properly.
1049                         // In practise I have yet to see a file that
1050                         // uses this element, however it is mentioned
1051                         // on page 25 of part 1 of the xmp standard.
1052                         //
1053                         // also it seems as if exiv2 and exiftool do not support
1054                         // this either (That or I misunderstand the standard)
1055                         wfDebugLog( 'XMP', __METHOD__ . ' Encountered <rdf:type> which isn\'t currently supported' );
1056                 }
1057
1058                 if ( strpos( $elm, ' ' ) === false ) {
1059                         // This probably shouldn't happen.
1060                         wfDebugLog( 'XMP', __METHOD__ . " Encountered <$elm> which has no namespace. Skipping." );
1061                         return;
1062                 }
1063
1064                 list( $ns, $tag ) = explode( ' ', $elm, 2 );
1065
1066                 if ( count( $this->mode ) === 0 ) {
1067                         // This should not happen.
1068                         throw new MWException('Error extracting XMP, '
1069                                 . "encountered <$elm> with no mode" );
1070                 }
1071
1072                 switch( $this->mode[0] ) {
1073                         case self::MODE_IGNORE:
1074                                 $this->startElementModeIgnore( $elm );
1075                                 break;
1076                         case self::MODE_SIMPLE:
1077                                 $this->startElementModeSimple( $elm, $attribs );
1078                                 break;
1079                         case self::MODE_INITIAL:
1080                                 $this->startElementModeInitial( $ns, $tag, $attribs );
1081                                 break;
1082                         case self::MODE_STRUCT:
1083                                 $this->startElementModeStruct( $ns, $tag, $attribs );
1084                                 break;
1085                         case self::MODE_BAG:
1086                         case self::MODE_BAGSTRUCT:
1087                                 $this->startElementModeBag( $elm );
1088                                 break;
1089                         case self::MODE_SEQ:
1090                                 $this->startElementModeSeq( $elm );
1091                                 break;
1092                         case self::MODE_LANG:
1093                                 $this->startElementModeLang( $elm );
1094                                 break;
1095                         case self::MODE_LI_LANG:
1096                                 $this->startElementModeLiLang( $elm, $attribs );
1097                                 break;
1098                         case self::MODE_LI:
1099                                 $this->startElementModeLi( $elm, $attribs );
1100                                 break;
1101                         case self::MODE_QDESC:
1102                                 $this->startElementModeQDesc( $elm );
1103                                 break;
1104                         default:
1105                                 throw new MWException( 'StartElement in unknown mode: ' . $this->mode[0] );
1106                                 break;
1107                 }
1108         }
1109
1110         /**
1111          * Process attributes.
1112          * Simple values can be stored as either a tag or attribute
1113          *
1114          * Often the initial "<rdf:Description>" tag just has all the simple
1115          * properties as attributes.
1116          *
1117          * @par Example:
1118          * @code
1119          * <rdf:Description rdf:about="" xmlns:exif="http://ns.adobe.com/exif/1.0/" exif:DigitalZoomRatio="0/10">
1120          * @endcode
1121          *
1122          * @param $attribs Array attribute=>value array.
1123          * @throws MWException
1124          */
1125         private function doAttribs( $attribs ) {
1126
1127                 // first check for rdf:parseType attribute, as that can change
1128                 // how the attributes are interperted.
1129
1130                 if ( isset( $attribs[self::NS_RDF . ' parseType'] )
1131                         && $attribs[self::NS_RDF . ' parseType'] === 'Resource'
1132                         && $this->mode[0] === self::MODE_SIMPLE )
1133                 {
1134                         // this is equivalent to having an inner rdf:Description
1135                         $this->mode[0] = self::MODE_QDESC;
1136                 }
1137                 foreach ( $attribs as $name => $val ) {
1138
1139
1140                         if ( strpos( $name, ' ' ) === false ) {
1141                                 // This shouldn't happen, but so far some old software forgets namespace
1142                                 // on rdf:about.
1143                                 wfDebugLog( 'XMP', __METHOD__ . ' Encountered non-namespaced attribute: '
1144                                         . " $name=\"$val\". Skipping. " );
1145                                 continue;
1146                         }
1147                         list( $ns, $tag ) = explode( ' ', $name, 2 );
1148                         if ( $ns === self::NS_RDF ) {
1149                                 if ( $tag === 'value' || $tag === 'resource' ) {
1150                                         // resource is for url.
1151                                         // value attribute is a weird way of just putting the contents.
1152                                         $this->char( $this->xmlParser, $val );
1153                                 }
1154                         } elseif ( isset( $this->items[$ns][$tag] ) ) {
1155                                 if ( $this->mode[0] === self::MODE_SIMPLE ) {
1156                                         throw new MWException( __METHOD__
1157                                                 . " $ns:$tag found as attribute where not allowed" );
1158                                 }
1159                                 $this->saveValue( $ns, $tag, $val );
1160                         } else {
1161                                 wfDebugLog( 'XMP', __METHOD__ . " Ignoring unrecognized element <$ns:$tag>." );
1162                         }
1163                 }
1164         }
1165
1166         /**
1167         * Given an extracted value, save it to results array
1168         *
1169         * note also uses $this->ancestorStruct and
1170         * $this->processingArray to determine what name to
1171         * save the value under. (in addition to $tag).
1172         *
1173         * @param $ns String namespace of tag this is for
1174         * @param $tag String tag name
1175         * @param $val String value to save
1176         */
1177         private function saveValue( $ns, $tag, $val ) {
1178
1179                 $info =& $this->items[$ns][$tag];
1180                 $finalName = isset( $info['map_name'] )
1181                         ? $info['map_name'] : $tag;
1182                 if ( isset( $info['validate'] ) ) {
1183                         $validate = is_array( $info['validate'] ) ? $info['validate']
1184                                 : array( 'XMPValidate', $info['validate'] );
1185
1186                         if ( is_callable( $validate ) ) {
1187                                 call_user_func_array( $validate, array( $info, &$val, true ) );
1188                                 // the reasoning behind using &$val instead of using the return value
1189                                 // is to be consistent between here and validating structures.
1190                                 if ( is_null( $val ) ) {
1191                                         wfDebugLog( 'XMP', __METHOD__ . " <$ns:$tag> failed validation." );
1192                                         return;
1193                                 }
1194                         } else {
1195                                 wfDebugLog( 'XMP', __METHOD__ . " Validation function for $finalName ("
1196                                         . $validate[0] . '::' . $validate[1] . '()) is not callable.' );
1197                         }
1198                 }
1199
1200                 if ( $this->ancestorStruct && $this->processingArray ) {
1201                         // Aka both an array and a struct. ( self::MODE_BAGSTRUCT )
1202                         $this->results['xmp-' . $info['map_group']][$this->ancestorStruct][][$finalName] = $val;
1203                 } elseif ( $this->ancestorStruct ) {
1204                         $this->results['xmp-' . $info['map_group']][$this->ancestorStruct][$finalName] = $val;
1205                 } elseif ( $this->processingArray ) {
1206                         if ( $this->itemLang === false ) {
1207                                 // normal array
1208                                 $this->results['xmp-' . $info['map_group']][$finalName][] = $val;
1209                         } else {
1210                                 // lang array.
1211                                 $this->results['xmp-' . $info['map_group']][$finalName][$this->itemLang] = $val;
1212                         }
1213                 } else {
1214                         $this->results['xmp-' . $info['map_group']][$finalName] = $val;
1215                 }
1216         }
1217 }