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         * @return Boolean success.
 245         */
 246         public function parse( $content, $allOfIt = true, $reset = false ) {
 247                 if ( $reset ) {
 248                         $this->resetXMLParser();
 249                 }
 250                 try {
 251
 252                         // detect encoding by looking for BOM which is supposed to be in processing instruction.
 253                         // see page 12 of http://www.adobe.com/devnet/xmp/pdfs/XMPSpecificationPart3.pdf
 254                         if ( !$this->charset ) {
 255                                 $bom = array();
 256                                 if ( preg_match( '/\xEF\xBB\xBF|\xFE\xFF|\x00\x00\xFE\xFF|\xFF\xFE\x00\x00|\xFF\xFE/',
 257                                          $content, $bom )
 258                                 ) {
 259                                         switch ( $bom[0] ) {
 260                                                 case "\xFE\xFF":
 261                                                         $this->charset = 'UTF-16BE';
 262                                                         break;
 263                                                 case "\xFF\xFE":
 264                                                         $this->charset = 'UTF-16LE';
 265                                                         break;
 266                                                 case "\x00\x00\xFE\xFF":
 267                                                         $this->charset = 'UTF-32BE';
 268                                                         break;
 269                                                 case "\xFF\xFE\x00\x00":
 270                                                         $this->charset = 'UTF-32LE';
 271                                                         break;
 272                                                 case "\xEF\xBB\xBF":
 273                                                         $this->charset = 'UTF-8';
 274                                                         break;
 275                                                 default:
 276                                                         //this should be impossible to get to
 277                                                         throw new MWException("Invalid BOM");
 278                                                         break;
 279
 280                                         }
 281
 282                                 } else {
 283                                         // standard specifically says, if no bom assume utf-8
 284                                         $this->charset = 'UTF-8';
 285                                 }
 286                         }
 287                         if ( $this->charset !== 'UTF-8' ) {
 288                                 //don't convert if already utf-8
 289                                 wfSuppressWarnings();
 290                                 $content = iconv( $this->charset, 'UTF-8//IGNORE', $content );
 291                                 wfRestoreWarnings();
 292                         }
 293
 294                         $ok = xml_parse( $this->xmlParser, $content, $allOfIt );
 295                         if ( !$ok ) {
 296                                 $error = xml_error_string( xml_get_error_code( $this->xmlParser ) );
 297                                 $where = 'line: ' . xml_get_current_line_number( $this->xmlParser )
 298                                         . ' column: ' . xml_get_current_column_number( $this->xmlParser )
 299                                         . ' byte offset: ' . xml_get_current_byte_index( $this->xmlParser );
 300
 301                                 wfDebugLog( 'XMP', "XMPReader::parse : Error reading XMP content: $error ($where)" );
 302                                 $this->results = array(); // blank if error.
 303                                 return false;
 304                         }
 305                 } catch ( MWException $e ) {
 306                         wfDebugLog( 'XMP', 'XMP parse error: ' . $e );
 307                         $this->results = array();
 308                         return false;
 309                 }
 310                 return true;
 311         }
 312
 313         /** Entry point for XMPExtended blocks in jpeg files
 314          *
 315          * @todo In serious need of testing
 316          * @see http://www.adobe.ge/devnet/xmp/pdfs/XMPSpecificationPart3.pdf XMP spec part 3 page 20
 317          * @param String $content XMPExtended block minus the namespace signature
 318          * @return Boolean If it succeeded.
 319          */
 320         public function parseExtended( $content ) {
 321                 // @todo FIXME: This is untested. Hard to find example files
 322                 // or programs that make such files..
 323                 $guid = substr( $content, 0, 32 );
 324                 if ( !isset( $this->results['xmp-special']['HasExtendedXMP'] )
 325                         || $this->results['xmp-special']['HasExtendedXMP'] !== $guid ) {
 326                         wfDebugLog('XMP', __METHOD__ . " Ignoring XMPExtended block due to wrong guid (guid= '$guid' )");
 327                         return false;
 328                 }
 329                 $len  = unpack( 'Nlength/Noffset', substr( $content, 32, 8 ) );
 330
 331                 if (!$len || $len['length'] < 4 || $len['offset'] < 0 || $len['offset'] > $len['length'] ) {
 332                         wfDebugLog('XMP', __METHOD__ . 'Error reading extended XMP block, invalid length or offset.');
 333                         return false;
 334                 }
 335
 336
 337                 // we're not very robust here. we should accept it in the wrong order. To quote
 338                 // the xmp standard:
 339                 // "A JPEG writer should write the ExtendedXMP marker segments in order, immediately following the
 340                 // StandardXMP. However, the JPEG standard does not require preservation of marker segment order. A
 341                 // robust JPEG reader should tolerate the marker segments in any order."
 342                 //
 343                 // otoh the probability that an image will have more than 128k of metadata is rather low...
 344                 // so the probability that it will have > 128k, and be in the wrong order is very low...
 345
 346                 if ( $len['offset'] !== $this->extendedXMPOffset ) {
 347                         wfDebugLog('XMP', __METHOD__ . 'Ignoring XMPExtended block due to wrong order. (Offset was '
 348                                 . $len['offset'] . ' but expected ' . $this->extendedXMPOffset . ')');
 349                         return false;
 350                 }
 351
 352                 if ( $len['offset'] === 0 ) {
 353                         // if we're starting the extended block, we've probably already
 354                         // done the XMPStandard block, so reset.
 355                         $this->resetXMLParser();
 356                 }
 357
 358                 $this->extendedXMPOffset += $len['length'];
 359
 360                 $actualContent = substr( $content, 40 );
 361
 362                 if ( $this->extendedXMPOffset === strlen( $actualContent ) ) {
 363                         $atEnd = true;
 364                 } else {
 365                         $atEnd = false;
 366                 }
 367
 368                 wfDebugLog('XMP', __METHOD__ . 'Parsing a XMPExtended block');
 369                 return $this->parse( $actualContent, $atEnd );
 370         }
 371
 372         /**
 373         * Character data handler
 374         * Called whenever character data is found in the xmp document.
 375         *
 376         * does nothing if we're in MODE_IGNORE or if the data is whitespace
 377         * throws an error if we're not in MODE_SIMPLE (as we're not allowed to have character
 378         * data in the other modes).
 379         *
 380         * As an example, this happens when we encounter XMP like:
 381         * <exif:DigitalZoomRatio>0/10</exif:DigitalZoomRatio>
 382         * and are processing the 0/10 bit.
 383         *
 384         * @param $parser XMLParser reference to the xml parser
 385         * @param $data String Character data
 386         * @throws MWException on invalid data
 387         */
 388         function char( $parser, $data ) {
 389
 390                 $data = trim( $data );
 391                 if ( trim( $data ) === "" ) {
 392                         return;
 393                 }
 394
 395                 if ( !isset( $this->mode[0] ) ) {
 396                         throw new MWException( 'Unexpected character data before first rdf:Description element' );
 397                 }
 398
 399                 if ( $this->mode[0] === self::MODE_IGNORE ) return;
 400
 401                 if ( $this->mode[0] !== self::MODE_SIMPLE
 402                         && $this->mode[0] !== self::MODE_QDESC
 403                 ) {
 404                         throw new MWException( 'character data where not expected. (mode ' . $this->mode[0] . ')' );
 405                 }
 406
 407                 // to check, how does this handle w.s.
 408                 if ( $this->charContent === false ) {
 409                         $this->charContent = $data;
 410                 } else {
 411                         $this->charContent .= $data;
 412                 }
 413
 414         }
 415
 416         /** When we hit a closing element in MODE_IGNORE
 417         * Check to see if this is the element we started to ignore,
 418         * in which case we get out of MODE_IGNORE
 419         *
 420         * @param $elm String Namespace of element followed by a space and then tag name of element.
 421         */
 422         private function endElementModeIgnore ( $elm ) {
 423
 424                 if ( $this->curItem[0] === $elm ) {
 425                         array_shift( $this->curItem );
 426                         array_shift( $this->mode );
 427                 }
 428                 return;
 429
 430         }
 431
 432         /**
 433         * Hit a closing element when in MODE_SIMPLE.
 434         * This generally means that we finished processing a
 435         * property value, and now have to save the result to the
 436         * results array
 437         *
 438         * For example, when processing:
 439         * <exif:DigitalZoomRatio>0/10</exif:DigitalZoomRatio>
 440         * this deals with when we hit </exif:DigitalZoomRatio>.
 441         *
 442         * Or it could be if we hit the end element of a property
 443         * of a compound data structure (like a member of an array).
 444         *
 445         * @param $elm String namespace, space, and tag name.
 446         */
 447         private function endElementModeSimple ( $elm ) {
 448                 if ( $this->charContent !== false ) {
 449                         if ( $this->processingArray ) {
 450                                 // if we're processing an array, use the original element
 451                                 // name instead of rdf:li.
 452                                 list( $ns, $tag ) = explode( ' ', $this->curItem[0], 2 );
 453                         } else {
 454                                 list( $ns, $tag ) = explode( ' ', $elm, 2 );
 455                         }
 456                         $this->saveValue( $ns, $tag, $this->charContent );
 457
 458                         $this->charContent = false; // reset
 459                 }
 460                 array_shift( $this->curItem );
 461                 array_shift( $this->mode );
 462
 463         }
 464
 465         /**
 466         * Hit a closing element in MODE_STRUCT, MODE_SEQ, MODE_BAG
 467         * generally means we've finished processing a nested structure.
 468         * resets some internal variables to indicate that.
 469         *
 470         * Note this means we hit the closing element not the "</rdf:Seq>".
 471         *
 472         * @par For example, when processing:
 473         * @code{,xml}
 474         * <exif:ISOSpeedRatings> <rdf:Seq> <rdf:li>64</rdf:li>
 475         *   </rdf:Seq> </exif:ISOSpeedRatings>
 476         * @endcode
 477         *
 478         * This method is called when we hit the "</exif:ISOSpeedRatings>" tag.
 479         *
 480         * @param $elm String namespace . space . tag name.
 481         */
 482         private function endElementNested( $elm ) {
 483
 484                 /* cur item must be the same as $elm, unless if in MODE_STRUCT
 485                    in which case it could also be rdf:Description */
 486                 if ( $this->curItem[0] !== $elm
 487                         && !( $elm === self::NS_RDF . ' Description'
 488                                 && $this->mode[0] === self::MODE_STRUCT )
 489                  ) {
 490                         throw new MWException( "nesting mismatch. got a </$elm> but expected a </" . $this->curItem[0] . '>' );
 491                 }
 492
 493                 // Validate structures.
 494                 list( $ns, $tag ) = explode( ' ', $elm, 2 );
 495                 if ( isset( $this->items[$ns][$tag]['validate'] ) ) {
 496
 497                         $info =& $this->items[$ns][$tag];
 498                         $finalName = isset( $info['map_name'] )
 499                                 ? $info['map_name'] : $tag;
 500
 501                         $validate = is_array( $info['validate'] ) ? $info['validate']
 502                                 : array( 'XMPValidate', $info['validate'] );
 503
 504                         if ( !isset( $this->results['xmp-' . $info['map_group']][$finalName] ) ) {
 505                                 // This can happen if all the members of the struct failed validation.
 506                                 wfDebugLog( 'XMP', __METHOD__ . " <$ns:$tag> has no valid members." );
 507
 508                         } elseif ( is_callable( $validate ) ) {
 509                                 $val =& $this->results['xmp-' . $info['map_group']][$finalName];
 510                                 call_user_func_array( $validate, array( $info, &$val, false ) );
 511                                 if ( is_null( $val ) ) {
 512                                         // the idea being the validation function will unset the variable if
 513                                         // its invalid.
 514                                         wfDebugLog( 'XMP', __METHOD__ . " <$ns:$tag> failed validation." );
 515                                         unset( $this->results['xmp-' . $info['map_group']][$finalName] );
 516                                 }
 517                         } else {
 518                                 wfDebugLog( 'XMP', __METHOD__ . " Validation function for $finalName ("
 519                                         . $validate[0] . '::' . $validate[1] . '()) is not callable.' );
 520                         }
 521                 }
 522
 523                 array_shift( $this->curItem );
 524                 array_shift( $this->mode );
 525                 $this->ancestorStruct = false;
 526                 $this->processingArray = false;
 527                 $this->itemLang = false;
 528         }
 529
 530         /**
 531         * Hit a closing element in MODE_LI (either rdf:Seq, or rdf:Bag )
 532         * Add information about what type of element this is.
 533         *
 534         * Note we still have to hit the outer "</property>"
 535         *
 536         * @par For example, when processing:
 537         * @code{,xml}
 538         * <exif:ISOSpeedRatings> <rdf:Seq> <rdf:li>64</rdf:li>
 539         *   </rdf:Seq> </exif:ISOSpeedRatings>
 540         * @endcode
 541         *
 542         * This method is called when we hit the "</rdf:Seq>".
 543         * (For comparison, we call endElementModeSimple when we
 544         * hit the "</rdf:li>")
 545         *
 546         * @param $elm String namespace . ' ' . element name
 547         */
 548         private function endElementModeLi( $elm ) {
 549
 550                 list( $ns, $tag ) = explode( ' ', $this->curItem[0], 2 );
 551                 $info = $this->items[$ns][$tag];
 552                 $finalName = isset( $info['map_name'] )
 553                         ? $info['map_name'] : $tag;
 554
 555                 array_shift( $this->mode );
 556
 557                 if ( !isset( $this->results['xmp-' . $info['map_group']][$finalName] ) ) {
 558                         wfDebugLog( 'XMP', __METHOD__ . " Empty compund element $finalName." );
 559                         return;
 560                 }
 561
 562                 if ( $elm === self::NS_RDF . ' Seq' ) {
 563                         $this->results['xmp-' . $info['map_group']][$finalName]['_type'] = 'ol';
 564                 } elseif ( $elm === self::NS_RDF . ' Bag' ) {
 565                         $this->results['xmp-' . $info['map_group']][$finalName]['_type'] = 'ul';
 566                 } elseif ( $elm === self::NS_RDF . ' Alt' ) {
 567                         // extra if needed as you could theoretically have a non-language alt.
 568                         if ( $info['mode'] === self::MODE_LANG ) {
 569                                 $this->results['xmp-' . $info['map_group']][$finalName]['_type'] = 'lang';
 570                         }
 571
 572                 } else {
 573                         throw new MWException( __METHOD__ . " expected </rdf:seq> or </rdf:bag> but instead got $elm." );
 574                 }
 575         }
 576
 577         /**
 578         * End element while in MODE_QDESC
 579         * mostly when ending an element when we have a simple value
 580         * that has qualifiers.
 581         *
 582         * Qualifiers aren't all that common, and we don't do anything
 583         * with them.
 584         *
 585         * @param $elm String namespace and element
 586         */
 587         private function endElementModeQDesc( $elm ) {
 588
 589                 if ( $elm === self::NS_RDF . ' value' ) {
 590                         list( $ns, $tag ) = explode( ' ', $this->curItem[0], 2 );
 591                         $this->saveValue( $ns, $tag, $this->charContent );
 592                         return;
 593                 } else {
 594                         array_shift( $this->mode );
 595                         array_shift( $this->curItem );
 596                 }
 597
 598
 599         }
 600
 601         /**
 602         * Handler for hitting a closing element.
 603         *
 604         * generally just calls a helper function depending on what
 605         * mode we're in.
 606         *
 607         * Ignores the outer wrapping elements that are optional in
 608         * xmp and have no meaning.
 609         *
 610         * @param $parser XMLParser
 611         * @param $elm String namespace . ' ' . element name
 612         */
 613         function endElement( $parser, $elm ) {
 614                 if ( $elm === ( self::NS_RDF . ' RDF' )
 615                         || $elm === 'adobe:ns:meta/ xmpmeta'
 616                         || $elm === 'adobe:ns:meta/ xapmeta' )
 617                 {
 618                         // ignore these.
 619                         return;
 620                 }
 621
 622                 if ( $elm === self::NS_RDF . ' type' ) {
 623                         // these aren't really supported properly yet.
 624                         // However, it appears they almost never used.
 625                         wfDebugLog( 'XMP', __METHOD__ . ' encountered <rdf:type>' );
 626                 }
 627
 628                 if ( strpos( $elm, ' ' ) === false ) {
 629                         // This probably shouldn't happen.
 630                         // However, there is a bug in an adobe product
 631                         // that forgets the namespace on some things.
 632                         // (Luckily they are unimportant things).
 633                         wfDebugLog( 'XMP', __METHOD__ . " Encountered </$elm> which has no namespace. Skipping." );
 634                         return;
 635                 }
 636
 637                 if ( count( $this->mode[0] ) === 0 ) {
 638                         // This should never ever happen and means
 639                         // there is a pretty major bug in this class.
 640                         throw new MWException( 'Encountered end element with no mode' );
 641                 }
 642
 643                 if ( count( $this->curItem ) == 0 && $this->mode[0] !== self::MODE_INITIAL ) {
 644                         // just to be paranoid. Should always have a curItem, except for initially
 645                         // (aka during MODE_INITAL).
 646                         throw new MWException( "Hit end element </$elm> but no curItem" );
 647                 }
 648
 649                 switch( $this->mode[0] ) {
 650                         case self::MODE_IGNORE:
 651                                 $this->endElementModeIgnore( $elm );
 652                                 break;
 653                         case self::MODE_SIMPLE:
 654                                 $this->endElementModeSimple( $elm );
 655                                 break;
 656                         case self::MODE_STRUCT:
 657                         case self::MODE_SEQ:
 658                         case self::MODE_BAG:
 659                         case self::MODE_LANG:
 660                         case self::MODE_BAGSTRUCT:
 661                                 $this->endElementNested( $elm );
 662                                 break;
 663                         case self::MODE_INITIAL:
 664                                 if ( $elm === self::NS_RDF . ' Description' ) {
 665                                         array_shift( $this->mode );
 666                                 } else {
 667                                         throw new MWException( 'Element ended unexpectedly while in MODE_INITIAL' );
 668                                 }
 669                                 break;
 670                         case self::MODE_LI:
 671                         case self::MODE_LI_LANG:
 672                                 $this->endElementModeLi( $elm );
 673                                 break;
 674                         case self::MODE_QDESC:
 675                                 $this->endElementModeQDesc( $elm );
 676                                 break;
 677                         default:
 678                                 wfDebugLog( 'XMP', __METHOD__ . " no mode (elm = $elm)" );
 679                                 break;
 680                 }
 681         }
 682
 683         /**
 684         * Hit an opening element while in MODE_IGNORE
 685         *
 686         * XMP is extensible, so ignore any tag we don't understand.
 687         *
 688         * Mostly ignores, unless we encounter the element that we are ignoring.
 689         * in which case we add it to the item stack, so we can ignore things
 690         * that are nested, correctly.
 691         *
 692         * @param $elm String namespace . ' ' . tag name
 693         */
 694         private function startElementModeIgnore( $elm ) {
 695                 if ( $elm === $this->curItem[0] ) {
 696                         array_unshift( $this->curItem, $elm );
 697                         array_unshift( $this->mode, self::MODE_IGNORE );
 698                 }
 699         }
 700
 701         /**
 702         *  Start element in MODE_BAG (unordered array)
 703         * this should always be <rdf:Bag>
 704         *
 705         * @param $elm String namespace . ' ' . tag
 706         * @throws MWException if we have an element that's not <rdf:Bag>
 707         */
 708         private function startElementModeBag( $elm ) {
 709                 if ( $elm === self::NS_RDF . ' Bag' ) {
 710                         array_unshift( $this->mode, self::MODE_LI );
 711                 } else {
 712                         throw new MWException( "Expected <rdf:Bag> but got $elm." );
 713                 }
 714
 715         }
 716
 717         /**
 718         * Start element in MODE_SEQ (ordered array)
 719         * this should always be <rdf:Seq>
 720         *
 721         * @param $elm String namespace . ' ' . tag
 722         * @throws MWException if we have an element that's not <rdf:Seq>
 723         */
 724         private function startElementModeSeq( $elm ) {
 725                 if ( $elm === self::NS_RDF . ' Seq' ) {
 726                         array_unshift( $this->mode, self::MODE_LI );
 727                 } elseif ( $elm === self::NS_RDF . ' Bag' ) {
 728                         # bug 27105
 729                         wfDebugLog( 'XMP', __METHOD__ . ' Expected an rdf:Seq, but got an rdf:Bag. Pretending'
 730                                 . ' it is a Seq, since some buggy software is known to screw this up.' );
 731                         array_unshift( $this->mode, self::MODE_LI );
 732                 } else {
 733                         throw new MWException( "Expected <rdf:Seq> but got $elm." );
 734                 }
 735
 736         }
 737
 738         /**
 739         * Start element in MODE_LANG (language alternative)
 740         * this should always be <rdf:Alt>
 741         *
 742         * This tag tends to be used for metadata like describe this
 743         * picture, which can be translated into multiple languages.
 744         *
 745         * XMP supports non-linguistic alternative selections,
 746         * which are really only used for thumbnails, which
 747         * we don't care about.
 748         *
 749         * @param $elm String namespace . ' ' . tag
 750         * @throws MWException if we have an element that's not <rdf:Alt>
 751         */
 752         private function startElementModeLang( $elm ) {
 753                 if ( $elm === self::NS_RDF . ' Alt' ) {
 754                         array_unshift( $this->mode, self::MODE_LI_LANG );
 755                 } else {
 756                         throw new MWException( "Expected <rdf:Seq> but got $elm." );
 757                 }
 758
 759         }
 760
 761         /**
 762         * Handle an opening element when in MODE_SIMPLE
 763         *
 764         * This should not happen often. This is for if a simple element
 765         * already opened has a child element. Could happen for a
 766         * qualified element.
 767         *
 768         * For example:
 769         * <exif:DigitalZoomRatio><rdf:Description><rdf:value>0/10</rdf:value>
 770         *   <foo:someQualifier>Bar</foo:someQualifier> </rdf:Description>
 771         *   </exif:DigitalZoomRatio>
 772         *
 773         * This method is called when processing the <rdf:Description> element
 774         *
 775         * @param $elm String namespace and tag names separated by space.
 776         * @param $attribs Array Attributes of the element.
 777         */
 778         private function startElementModeSimple( $elm, $attribs ) {
 779                 if ( $elm === self::NS_RDF . ' Description' ) {
 780                         // If this value has qualifiers
 781                         array_unshift( $this->mode, self::MODE_QDESC );
 782                         array_unshift( $this->curItem, $this->curItem[0] );
 783
 784                         if ( isset( $attribs[self::NS_RDF . ' value'] ) ) {
 785                                 list( $ns, $tag ) = explode( ' ', $this->curItem[0], 2 );
 786                                 $this->saveValue( $ns, $tag, $attribs[self::NS_RDF . ' value'] );
 787                         }
 788                 } elseif ( $elm === self::NS_RDF . ' value' ) {
 789                         // This should not be here.
 790                         throw new MWException( __METHOD__ . ' Encountered <rdf:value> where it was unexpected.' );
 791
 792                 } else {
 793                         // something else we don't recognize, like a qualifier maybe.
 794                         wfDebugLog( 'XMP', __METHOD__ . " Encountered element <$elm> where only expecting character data as value of " . $this->curItem[0] );
 795                         array_unshift( $this->mode, self::MODE_IGNORE );
 796                         array_unshift( $this->curItem, $elm );
 797
 798                 }
 799
 800         }
 801
 802         /**
 803         * Start an element when in MODE_QDESC.
 804         * This generally happens when a simple element has an inner
 805         * rdf:Description to hold qualifier elements.
 806         *
 807         * For example in:
 808         * <exif:DigitalZoomRatio><rdf:Description><rdf:value>0/10</rdf:value>
 809         *   <foo:someQualifier>Bar</foo:someQualifier> </rdf:Description>
 810         *   </exif:DigitalZoomRatio>
 811         * Called when processing the <rdf:value> or <foo:someQualifier>.
 812         *
 813         * @param $elm String namespace and tag name separated by a space.
 814         *
 815         */
 816         private function startElementModeQDesc( $elm ) {
 817                 if ( $elm === self::NS_RDF . ' value' ) {
 818                         return; // do nothing
 819                 } else {
 820                         // otherwise its a qualifier, which we ignore
 821                         array_unshift( $this->mode, self::MODE_IGNORE );
 822                         array_unshift( $this->curItem, $elm );
 823                 }
 824         }
 825
 826         /**
 827         * Starting an element when in MODE_INITIAL
 828         * This usually happens when we hit an element inside
 829         * the outer rdf:Description
 830         *
 831         * This is generally where most properties start.
 832         *
 833         * @param $ns String Namespace
 834         * @param $tag String tag name (without namespace prefix)
 835         * @param $attribs Array array of attributes
 836         */
 837         private function startElementModeInitial( $ns, $tag, $attribs ) {
 838                 if ( $ns !== self::NS_RDF ) {
 839
 840                         if ( isset( $this->items[$ns][$tag] ) ) {
 841                                 if ( isset( $this->items[$ns][$tag]['structPart'] ) ) {
 842                                         // If this element is supposed to appear only as
 843                                         // a child of a structure, but appears here (not as
 844                                         // a child of a struct), then something weird is
 845                                         // happening, so ignore this element and its children.
 846
 847                                         wfDebugLog( 'XMP', "Encountered <$ns:$tag> outside"
 848                                                 . " of its expected parent. Ignoring." );
 849
 850                                         array_unshift( $this->mode, self::MODE_IGNORE );
 851                                         array_unshift( $this->curItem, $ns . ' ' . $tag );
 852                                         return;
 853                                 }
 854                                 $mode = $this->items[$ns][$tag]['mode'];
 855                                 array_unshift( $this->mode, $mode );
 856                                 array_unshift( $this->curItem, $ns . ' ' . $tag );
 857                                 if ( $mode === self::MODE_STRUCT ) {
 858                                         $this->ancestorStruct = isset( $this->items[$ns][$tag]['map_name'] )
 859                                                 ? $this->items[$ns][$tag]['map_name'] : $tag;
 860                                 }
 861                                 if ( $this->charContent !== false ) {
 862                                         // Something weird.
 863                                         // Should not happen in valid XMP.
 864                                         throw new MWException( 'tag nested in non-whitespace characters.' );
 865                                 }
 866                         } else {
 867                                 // This element is not on our list of allowed elements so ignore.
 868                                 wfDebugLog( 'XMP', __METHOD__ . " Ignoring unrecognized element <$ns:$tag>." );
 869                                 array_unshift( $this->mode, self::MODE_IGNORE );
 870                                 array_unshift( $this->curItem, $ns . ' ' . $tag );
 871                                 return;
 872                         }
 873
 874                 }
 875                 // process attributes
 876                 $this->doAttribs( $attribs );
 877         }
 878
 879         /**
 880         * Hit an opening element when in a Struct (MODE_STRUCT)
 881         * This is generally for fields of a compound property.
 882         *
 883         * Example of a struct (abbreviated; flash has more properties):
 884         *
 885         * <exif:Flash> <rdf:Description> <exif:Fired>True</exif:Fired>
 886         *  <exif:Mode>1</exif:Mode></rdf:Description></exif:Flash>
 887         *
 888         * or:
 889         *
 890         * <exif:Flash rdf:parseType='Resource'> <exif:Fired>True</exif:Fired>
 891         *  <exif:Mode>1</exif:Mode></exif:Flash>
 892         *
 893         * @param $ns String namespace
 894         * @param $tag String tag name (no ns)
 895         * @param $attribs Array array of attribs w/ values.
 896         */
 897         private function startElementModeStruct( $ns, $tag, $attribs ) {
 898                 if ( $ns !== self::NS_RDF ) {
 899
 900                         if ( isset( $this->items[$ns][$tag] ) ) {
 901                                 if ( isset( $this->items[$ns][$this->ancestorStruct]['children'] )
 902                                         && !isset( $this->items[$ns][$this->ancestorStruct]['children'][$tag] ) )
 903                                 {
 904                                         // This assumes that we don't have inter-namespace nesting
 905                                         // which we don't in all the properties we're interested in.
 906                                         throw new MWException( " <$tag> appeared nested in <" . $this->ancestorStruct
 907                                                 . "> where it is not allowed." );
 908                                 }
 909                                 array_unshift( $this->mode, $this->items[$ns][$tag]['mode'] );
 910                                 array_unshift( $this->curItem, $ns . ' ' . $tag );
 911                                 if ( $this->charContent !== false ) {
 912                                         // Something weird.
 913                                         // Should not happen in valid XMP.
 914                                         throw new MWException( "tag <$tag> nested in non-whitespace characters (" . $this->charContent . ")." );
 915                                 }
 916                         } else {
 917                                 array_unshift( $this->mode, self::MODE_IGNORE );
 918                                 array_unshift( $this->curItem, $elm );
 919                                 return;
 920                         }
 921
 922                 }
 923
 924                 if ( $ns === self::NS_RDF && $tag === 'Description' ) {
 925                         $this->doAttribs( $attribs );
 926                         array_unshift( $this->mode, self::MODE_STRUCT );
 927                         array_unshift( $this->curItem, $this->curItem[0] );
 928                 }
 929         }
 930
 931         /**
 932         * opening element in MODE_LI
 933         * process elements of arrays.
 934         *
 935         * Example:
 936         * <exif:ISOSpeedRatings> <rdf:Seq> <rdf:li>64</rdf:li>
 937         *   </rdf:Seq> </exif:ISOSpeedRatings>
 938         * This method is called when we hit the <rdf:li> element.
 939         *
 940         * @param $elm String: namespace . ' ' . tagname
 941         * @param $attribs Array: Attributes. (needed for BAGSTRUCTS)
 942         * @throws MWException if gets a tag other than <rdf:li>
 943         */
 944         private function startElementModeLi( $elm, $attribs ) {
 945                 if ( ( $elm ) !== self::NS_RDF . ' li' ) {
 946                         throw new MWException( "<rdf:li> expected but got $elm." );
 947                 }
 948
 949                 if ( !isset( $this->mode[1] ) ) {
 950                         // This should never ever ever happen. Checking for it
 951                         // to be paranoid.
 952                         throw new MWException( 'In mode Li, but no 2xPrevious mode!' );
 953                 }
 954
 955                 if ( $this->mode[1] === self::MODE_BAGSTRUCT ) {
 956                         // This list item contains a compound (STRUCT) value.
 957                         array_unshift( $this->mode, self::MODE_STRUCT );
 958                         array_unshift( $this->curItem, $elm );
 959                         $this->processingArray = true;
 960
 961                         if ( !isset( $this->curItem[1] ) ) {
 962                                 // be paranoid.
 963                                 throw new MWException( 'Can not find parent of BAGSTRUCT.' );
 964                         }
 965                         list( $curNS, $curTag ) = explode( ' ', $this->curItem[1] );
 966                         $this->ancestorStruct = isset( $this->items[$curNS][$curTag]['map_name'] )
 967                                 ? $this->items[$curNS][$curTag]['map_name'] : $curTag;
 968
 969                         $this->doAttribs( $attribs );
 970
 971                 } else {
 972                         // Normal BAG or SEQ containing simple values.
 973                         array_unshift( $this->mode, self::MODE_SIMPLE );
 974                         // need to add curItem[0] on again since one is for the specific item
 975                         // and one is for the entire group.
 976                         array_unshift( $this->curItem, $this->curItem[0] );
 977                         $this->processingArray = true;
 978                 }
 979
 980         }
 981
 982         /**
 983         * Opening element in MODE_LI_LANG.
 984         * process elements of language alternatives
 985         *
 986         * Example:
 987         * <dc:title> <rdf:Alt> <rdf:li xml:lang="x-default">My house
 988         *  </rdf:li> </rdf:Alt> </dc:title>
 989         *
 990         * This method is called when we hit the <rdf:li> element.
 991         *
 992         * @param $elm String namespace . ' ' . tag
 993         * @param $attribs array array of elements (most importantly xml:lang)
 994         * @throws MWException if gets a tag other than <rdf:li> or if no xml:lang
 995         */
 996         private function startElementModeLiLang( $elm, $attribs ) {
 997                 if ( $elm !== self::NS_RDF . ' li' ) {
 998                         throw new MWException( __METHOD__ . " <rdf:li> expected but got $elm." );
 999                 }
1000                 if ( !isset( $attribs[ self::NS_XML . ' lang'] )
1001                         || !preg_match( '/^[-A-Za-z0-9]{2,}$/D', $attribs[ self::NS_XML . ' lang' ] ) )
1002                 {
1003                         throw new MWException( __METHOD__
1004                                 . " <rdf:li> did not contain, or has invalid xml:lang attribute in lang alternative" );
1005                 }
1006
1007                 // Lang is case-insensitive.
1008                 $this->itemLang = strtolower( $attribs[ self::NS_XML . ' lang' ] );
1009
1010                 // need to add curItem[0] on again since one is for the specific item
1011                 // and one is for the entire group.
1012                 array_unshift( $this->curItem, $this->curItem[0] );
1013                 array_unshift( $this->mode, self::MODE_SIMPLE );
1014                 $this->processingArray = true;
1015         }
1016
1017         /**
1018         * Hits an opening element.
1019         * Generally just calls a helper based on what MODE we're in.
1020         * Also does some initial set up for the wrapper element
1021         *
1022         * @param $parser XMLParser
1023         * @param $elm String namespace "<space>" element
1024         * @param $attribs Array attribute name => value
1025         */
1026         function startElement( $parser, $elm, $attribs ) {
1027
1028                 if ( $elm === self::NS_RDF . ' RDF'
1029                         || $elm === 'adobe:ns:meta/ xmpmeta'
1030                         || $elm === 'adobe:ns:meta/ xapmeta')
1031                 {
1032                         /* ignore. */
1033                         return;
1034                 } elseif ( $elm === self::NS_RDF . ' Description' ) {
1035                         if ( count( $this->mode ) === 0 ) {
1036                                 // outer rdf:desc
1037                                 array_unshift( $this->mode, self::MODE_INITIAL );
1038                         }
1039                 } elseif ( $elm === self::NS_RDF . ' type' ) {
1040                         // This doesn't support rdf:type properly.
1041                         // In practise I have yet to see a file that
1042                         // uses this element, however it is mentioned
1043                         // on page 25 of part 1 of the xmp standard.
1044                         //
1045                         // also it seems as if exiv2 and exiftool do not support
1046                         // this either (That or I misunderstand the standard)
1047                         wfDebugLog( 'XMP', __METHOD__ . ' Encountered <rdf:type> which isn\'t currently supported' );
1048                 }
1049
1050                 if ( strpos( $elm, ' ' ) === false ) {
1051                         // This probably shouldn't happen.
1052                         wfDebugLog( 'XMP', __METHOD__ . " Encountered <$elm> which has no namespace. Skipping." );
1053                         return;
1054                 }
1055
1056                 list( $ns, $tag ) = explode( ' ', $elm, 2 );
1057
1058                 if ( count( $this->mode ) === 0 ) {
1059                         // This should not happen.
1060                         throw new MWException('Error extracting XMP, '
1061                                 . "encountered <$elm> with no mode" );
1062                 }
1063
1064                 switch( $this->mode[0] ) {
1065                         case self::MODE_IGNORE:
1066                                 $this->startElementModeIgnore( $elm );
1067                                 break;
1068                         case self::MODE_SIMPLE:
1069                                 $this->startElementModeSimple( $elm, $attribs );
1070                                 break;
1071                         case self::MODE_INITIAL:
1072                                 $this->startElementModeInitial( $ns, $tag, $attribs );
1073                                 break;
1074                         case self::MODE_STRUCT:
1075                                 $this->startElementModeStruct( $ns, $tag, $attribs );
1076                                 break;
1077                         case self::MODE_BAG:
1078                         case self::MODE_BAGSTRUCT:
1079                                 $this->startElementModeBag( $elm );
1080                                 break;
1081                         case self::MODE_SEQ:
1082                                 $this->startElementModeSeq( $elm );
1083                                 break;
1084                         case self::MODE_LANG:
1085                                 $this->startElementModeLang( $elm );
1086                                 break;
1087                         case self::MODE_LI_LANG:
1088                                 $this->startElementModeLiLang( $elm, $attribs );
1089                                 break;
1090                         case self::MODE_LI:
1091                                 $this->startElementModeLi( $elm, $attribs );
1092                                 break;
1093                         case self::MODE_QDESC:
1094                                 $this->startElementModeQDesc( $elm );
1095                                 break;
1096                         default:
1097                                 throw new MWException( 'StartElement in unknown mode: ' . $this->mode[0] );
1098                                 break;
1099                 }
1100         }
1101
1102         /**
1103         * Process attributes.
1104         * Simple values can be stored as either a tag or attribute
1105         *
1106         * Often the initial "<rdf:Description>" tag just has all the simple
1107         * properties as attributes.
1108         *
1109         * @par Example:
1110         * @code
1111         * <rdf:Description rdf:about="" xmlns:exif="http://ns.adobe.com/exif/1.0/" exif:DigitalZoomRatio="0/10">
1112         * @endcode
1113         *
1114         * @param $attribs Array attribute=>value array.
1115         */
1116         private function doAttribs( $attribs ) {
1117
1118                 // first check for rdf:parseType attribute, as that can change
1119                 // how the attributes are interperted.
1120
1121                 if ( isset( $attribs[self::NS_RDF . ' parseType'] )
1122                         && $attribs[self::NS_RDF . ' parseType'] === 'Resource'
1123                         && $this->mode[0] === self::MODE_SIMPLE )
1124                 {
1125                         // this is equivalent to having an inner rdf:Description
1126                         $this->mode[0] = self::MODE_QDESC;
1127                 }
1128                 foreach ( $attribs as $name => $val ) {
1129
1130
1131                         if ( strpos( $name, ' ' ) === false ) {
1132                                 // This shouldn't happen, but so far some old software forgets namespace
1133                                 // on rdf:about.
1134                                 wfDebugLog( 'XMP', __METHOD__ . ' Encountered non-namespaced attribute: '
1135                                         . " $name=\"$val\". Skipping. " );
1136                                 continue;
1137                         }
1138                         list( $ns, $tag ) = explode( ' ', $name, 2 );
1139                         if ( $ns === self::NS_RDF ) {
1140                                 if ( $tag === 'value' || $tag === 'resource' ) {
1141                                         // resource is for url.
1142                                         // value attribute is a weird way of just putting the contents.
1143                                         $this->char( $this->xmlParser, $val );
1144                                 }
1145                         } elseif ( isset( $this->items[$ns][$tag] ) ) {
1146                                 if ( $this->mode[0] === self::MODE_SIMPLE ) {
1147                                         throw new MWException( __METHOD__
1148                                                 . " $ns:$tag found as attribute where not allowed" );
1149                                 }
1150                                 $this->saveValue( $ns, $tag, $val );
1151                         } else {
1152                                 wfDebugLog( 'XMP', __METHOD__ . " Ignoring unrecognized element <$ns:$tag>." );
1153                         }
1154                 }
1155         }
1156
1157         /**
1158         * Given an extracted value, save it to results array
1159         *
1160         * note also uses $this->ancestorStruct and
1161         * $this->processingArray to determine what name to
1162         * save the value under. (in addition to $tag).
1163         *
1164         * @param $ns String namespace of tag this is for
1165         * @param $tag String tag name
1166         * @param $val String value to save
1167         */
1168         private function saveValue( $ns, $tag, $val ) {
1169
1170                 $info =& $this->items[$ns][$tag];
1171                 $finalName = isset( $info['map_name'] )
1172                         ? $info['map_name'] : $tag;
1173                 if ( isset( $info['validate'] ) ) {
1174                         $validate = is_array( $info['validate'] ) ? $info['validate']
1175                                 : array( 'XMPValidate', $info['validate'] );
1176
1177                         if ( is_callable( $validate ) ) {
1178                                 call_user_func_array( $validate, array( $info, &$val, true ) );
1179                                 // the reasoning behind using &$val instead of using the return value
1180                                 // is to be consistent between here and validating structures.
1181                                 if ( is_null( $val ) ) {
1182                                         wfDebugLog( 'XMP', __METHOD__ . " <$ns:$tag> failed validation." );
1183                                         return;
1184                                 }
1185                         } else {
1186                                 wfDebugLog( 'XMP', __METHOD__ . " Validation function for $finalName ("
1187                                         . $validate[0] . '::' . $validate[1] . '()) is not callable.' );
1188                         }
1189                 }
1190
1191                 if ( $this->ancestorStruct && $this->processingArray ) {
1192                         // Aka both an array and a struct. ( self::MODE_BAGSTRUCT )
1193                         $this->results['xmp-' . $info['map_group']][$this->ancestorStruct][][$finalName] = $val;
1194                 } elseif ( $this->ancestorStruct ) {
1195                         $this->results['xmp-' . $info['map_group']][$this->ancestorStruct][$finalName] = $val;
1196                 } elseif ( $this->processingArray ) {
1197                         if ( $this->itemLang === false ) {
1198                                 // normal array
1199                                 $this->results['xmp-' . $info['map_group']][$finalName][] = $val;
1200                         } else {
1201                                 // lang array.
1202                                 $this->results['xmp-' . $info['map_group']][$finalName][$this->itemLang] = $val;
1203                         }
1204                 } else {
1205                         $this->results['xmp-' . $info['map_group']][$finalName] = $val;
1206                 }
1207         }
1208 }