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                                         }
 280                                 } else {
 281                                         // standard specifically says, if no bom assume utf-8
 282                                         $this->charset = 'UTF-8';
 283                                 }
 284                         }
 285                         if ( $this->charset !== 'UTF-8' ) {
 286                                 //don't convert if already utf-8
 287                                 wfSuppressWarnings();
 288                                 $content = iconv( $this->charset, 'UTF-8//IGNORE', $content );
 289                                 wfRestoreWarnings();
 290                         }
 291
 292                         $ok = xml_parse( $this->xmlParser, $content, $allOfIt );
 293                         if ( !$ok ) {
 294                                 $error = xml_error_string( xml_get_error_code( $this->xmlParser ) );
 295                                 $where = 'line: ' . xml_get_current_line_number( $this->xmlParser )
 296                                         . ' column: ' . xml_get_current_column_number( $this->xmlParser )
 297                                         . ' byte offset: ' . xml_get_current_byte_index( $this->xmlParser );
 298
 299                                 wfDebugLog( 'XMP', "XMPReader::parse : Error reading XMP content: $error ($where)" );
 300                                 $this->results = array(); // blank if error.
 301                                 return false;
 302                         }
 303                 } catch ( MWException $e ) {
 304                         wfDebugLog( 'XMP', 'XMP parse error: ' . $e );
 305                         $this->results = array();
 306                         return false;
 307                 }
 308                 return true;
 309         }
 310
 311         /** Entry point for XMPExtended blocks in jpeg files
 312          *
 313          * @todo In serious need of testing
 314          * @see http://www.adobe.ge/devnet/xmp/pdfs/XMPSpecificationPart3.pdf XMP spec part 3 page 20
 315          * @param String $content XMPExtended block minus the namespace signature
 316          * @return Boolean If it succeeded.
 317          */
 318         public function parseExtended( $content ) {
 319                 // @todo FIXME: This is untested. Hard to find example files
 320                 // or programs that make such files..
 321                 $guid = substr( $content, 0, 32 );
 322                 if ( !isset( $this->results['xmp-special']['HasExtendedXMP'] )
 323                         || $this->results['xmp-special']['HasExtendedXMP'] !== $guid ) {
 324                         wfDebugLog( 'XMP', __METHOD__ . " Ignoring XMPExtended block due to wrong guid (guid= '$guid')" );
 325                         return false;
 326                 }
 327                 $len = unpack( 'Nlength/Noffset', substr( $content, 32, 8 ) );
 328
 329                 if ( !$len || $len['length'] < 4 || $len['offset'] < 0 || $len['offset'] > $len['length'] ) {
 330                         wfDebugLog( 'XMP', __METHOD__ . 'Error reading extended XMP block, invalid length or offset.' );
 331                         return false;
 332                 }
 333
 334
 335                 // we're not very robust here. we should accept it in the wrong order. To quote
 336                 // the xmp standard:
 337                 // "A JPEG writer should write the ExtendedXMP marker segments in order, immediately following the
 338                 // StandardXMP. However, the JPEG standard does not require preservation of marker segment order. A
 339                 // robust JPEG reader should tolerate the marker segments in any order."
 340                 //
 341                 // otoh the probability that an image will have more than 128k of metadata is rather low...
 342                 // so the probability that it will have > 128k, and be in the wrong order is very low...
 343
 344                 if ( $len['offset'] !== $this->extendedXMPOffset ) {
 345                         wfDebugLog( 'XMP', __METHOD__ . 'Ignoring XMPExtended block due to wrong order. (Offset was '
 346                                 . $len['offset'] . ' but expected ' . $this->extendedXMPOffset . ')' );
 347                         return false;
 348                 }
 349
 350                 if ( $len['offset'] === 0 ) {
 351                         // if we're starting the extended block, we've probably already
 352                         // done the XMPStandard block, so reset.
 353                         $this->resetXMLParser();
 354                 }
 355
 356                 $this->extendedXMPOffset += $len['length'];
 357
 358                 $actualContent = substr( $content, 40 );
 359
 360                 if ( $this->extendedXMPOffset === strlen( $actualContent ) ) {
 361                         $atEnd = true;
 362                 } else {
 363                         $atEnd = false;
 364                 }
 365
 366                 wfDebugLog( 'XMP', __METHOD__ . 'Parsing a XMPExtended block' );
 367                 return $this->parse( $actualContent, $atEnd );
 368         }
 369
 370         /**
 371         * Character data handler
 372         * Called whenever character data is found in the xmp document.
 373         *
 374         * does nothing if we're in MODE_IGNORE or if the data is whitespace
 375         * throws an error if we're not in MODE_SIMPLE (as we're not allowed to have character
 376         * data in the other modes).
 377         *
 378         * As an example, this happens when we encounter XMP like:
 379         * <exif:DigitalZoomRatio>0/10</exif:DigitalZoomRatio>
 380         * and are processing the 0/10 bit.
 381         *
 382         * @param $parser XMLParser reference to the xml parser
 383         * @param $data String Character data
 384         * @throws MWException on invalid data
 385         */
 386         function char( $parser, $data ) {
 387
 388                 $data = trim( $data );
 389                 if ( trim( $data ) === "" ) {
 390                         return;
 391                 }
 392
 393                 if ( !isset( $this->mode[0] ) ) {
 394                         throw new MWException( 'Unexpected character data before first rdf:Description element' );
 395                 }
 396
 397                 if ( $this->mode[0] === self::MODE_IGNORE ) return;
 398
 399                 if ( $this->mode[0] !== self::MODE_SIMPLE
 400                         && $this->mode[0] !== self::MODE_QDESC
 401                 ) {
 402                         throw new MWException( 'character data where not expected. (mode ' . $this->mode[0] . ')' );
 403                 }
 404
 405                 // to check, how does this handle w.s.
 406                 if ( $this->charContent === false ) {
 407                         $this->charContent = $data;
 408                 } else {
 409                         $this->charContent .= $data;
 410                 }
 411
 412         }
 413
 414         /** When we hit a closing element in MODE_IGNORE
 415         * Check to see if this is the element we started to ignore,
 416         * in which case we get out of MODE_IGNORE
 417         *
 418         * @param $elm String Namespace of element followed by a space and then tag name of element.
 419         */
 420         private function endElementModeIgnore ( $elm ) {
 421                 if ( $this->curItem[0] === $elm ) {
 422                         array_shift( $this->curItem );
 423                         array_shift( $this->mode );
 424                 }
 425         }
 426
 427         /**
 428         * Hit a closing element when in MODE_SIMPLE.
 429         * This generally means that we finished processing a
 430         * property value, and now have to save the result to the
 431         * results array
 432         *
 433         * For example, when processing:
 434         * <exif:DigitalZoomRatio>0/10</exif:DigitalZoomRatio>
 435         * this deals with when we hit </exif:DigitalZoomRatio>.
 436         *
 437         * Or it could be if we hit the end element of a property
 438         * of a compound data structure (like a member of an array).
 439         *
 440         * @param $elm String namespace, space, and tag name.
 441         */
 442         private function endElementModeSimple ( $elm ) {
 443                 if ( $this->charContent !== false ) {
 444                         if ( $this->processingArray ) {
 445                                 // if we're processing an array, use the original element
 446                                 // name instead of rdf:li.
 447                                 list( $ns, $tag ) = explode( ' ', $this->curItem[0], 2 );
 448                         } else {
 449                                 list( $ns, $tag ) = explode( ' ', $elm, 2 );
 450                         }
 451                         $this->saveValue( $ns, $tag, $this->charContent );
 452
 453                         $this->charContent = false; // reset
 454                 }
 455                 array_shift( $this->curItem );
 456                 array_shift( $this->mode );
 457
 458         }
 459
 460         /**
 461          * Hit a closing element in MODE_STRUCT, MODE_SEQ, MODE_BAG
 462          * generally means we've finished processing a nested structure.
 463          * resets some internal variables to indicate that.
 464          *
 465          * Note this means we hit the closing element not the "</rdf:Seq>".
 466          *
 467          * @par For example, when processing:
 468          * @code{,xml}
 469          * <exif:ISOSpeedRatings> <rdf:Seq> <rdf:li>64</rdf:li>
 470          *   </rdf:Seq> </exif:ISOSpeedRatings>
 471          * @endcode
 472          *
 473          * This method is called when we hit the "</exif:ISOSpeedRatings>" tag.
 474          *
 475          * @param $elm String namespace . space . tag name.
 476          * @throws MWException
 477          */
 478         private function endElementNested( $elm ) {
 479
 480                 /* cur item must be the same as $elm, unless if in MODE_STRUCT
 481                    in which case it could also be rdf:Description */
 482                 if ( $this->curItem[0] !== $elm
 483                         && !( $elm === self::NS_RDF . ' Description'
 484                                 && $this->mode[0] === self::MODE_STRUCT )
 485                 ) {
 486                         throw new MWException( "nesting mismatch. got a </$elm> but expected a </" . $this->curItem[0] . '>' );
 487                 }
 488
 489                 // Validate structures.
 490                 list( $ns, $tag ) = explode( ' ', $elm, 2 );
 491                 if ( isset( $this->items[$ns][$tag]['validate'] ) ) {
 492
 493                         $info =& $this->items[$ns][$tag];
 494                         $finalName = isset( $info['map_name'] )
 495                                 ? $info['map_name'] : $tag;
 496
 497                         $validate = is_array( $info['validate'] ) ? $info['validate']
 498                                 : array( 'XMPValidate', $info['validate'] );
 499
 500                         if ( !isset( $this->results['xmp-' . $info['map_group']][$finalName] ) ) {
 501                                 // This can happen if all the members of the struct failed validation.
 502                                 wfDebugLog( 'XMP', __METHOD__ . " <$ns:$tag> has no valid members." );
 503
 504                         } elseif ( is_callable( $validate ) ) {
 505                                 $val =& $this->results['xmp-' . $info['map_group']][$finalName];
 506                                 call_user_func_array( $validate, array( $info, &$val, false ) );
 507                                 if ( is_null( $val ) ) {
 508                                         // the idea being the validation function will unset the variable if
 509                                         // its invalid.
 510                                         wfDebugLog( 'XMP', __METHOD__ . " <$ns:$tag> failed validation." );
 511                                         unset( $this->results['xmp-' . $info['map_group']][$finalName] );
 512                                 }
 513                         } else {
 514                                 wfDebugLog( 'XMP', __METHOD__ . " Validation function for $finalName ("
 515                                         . $validate[0] . '::' . $validate[1] . '()) is not callable.' );
 516                         }
 517                 }
 518
 519                 array_shift( $this->curItem );
 520                 array_shift( $this->mode );
 521                 $this->ancestorStruct = false;
 522                 $this->processingArray = false;
 523                 $this->itemLang = false;
 524         }
 525
 526         /**
 527          * Hit a closing element in MODE_LI (either rdf:Seq, or rdf:Bag )
 528          * Add information about what type of element this is.
 529          *
 530          * Note we still have to hit the outer "</property>"
 531          *
 532          * @par For example, when processing:
 533          * @code{,xml}
 534          * <exif:ISOSpeedRatings> <rdf:Seq> <rdf:li>64</rdf:li>
 535          *   </rdf:Seq> </exif:ISOSpeedRatings>
 536          * @endcode
 537          *
 538          * This method is called when we hit the "</rdf:Seq>".
 539          * (For comparison, we call endElementModeSimple when we
 540          * hit the "</rdf:li>")
 541          *
 542          * @param $elm String namespace . ' ' . element name
 543          * @throws MWException
 544          */
 545         private function endElementModeLi( $elm ) {
 546
 547                 list( $ns, $tag ) = explode( ' ', $this->curItem[0], 2 );
 548                 $info = $this->items[$ns][$tag];
 549                 $finalName = isset( $info['map_name'] )
 550                         ? $info['map_name'] : $tag;
 551
 552                 array_shift( $this->mode );
 553
 554                 if ( !isset( $this->results['xmp-' . $info['map_group']][$finalName] ) ) {
 555                         wfDebugLog( 'XMP', __METHOD__ . " Empty compund element $finalName." );
 556                         return;
 557                 }
 558
 559                 if ( $elm === self::NS_RDF . ' Seq' ) {
 560                         $this->results['xmp-' . $info['map_group']][$finalName]['_type'] = 'ol';
 561                 } elseif ( $elm === self::NS_RDF . ' Bag' ) {
 562                         $this->results['xmp-' . $info['map_group']][$finalName]['_type'] = 'ul';
 563                 } elseif ( $elm === self::NS_RDF . ' Alt' ) {
 564                         // extra if needed as you could theoretically have a non-language alt.
 565                         if ( $info['mode'] === self::MODE_LANG ) {
 566                                 $this->results['xmp-' . $info['map_group']][$finalName]['_type'] = 'lang';
 567                         }
 568
 569                 } else {
 570                         throw new MWException( __METHOD__ . " expected </rdf:seq> or </rdf:bag> but instead got $elm." );
 571                 }
 572         }
 573
 574         /**
 575         * End element while in MODE_QDESC
 576         * mostly when ending an element when we have a simple value
 577         * that has qualifiers.
 578         *
 579         * Qualifiers aren't all that common, and we don't do anything
 580         * with them.
 581         *
 582         * @param $elm String namespace and element
 583         */
 584         private function endElementModeQDesc( $elm ) {
 585
 586                 if ( $elm === self::NS_RDF . ' value' ) {
 587                         list( $ns, $tag ) = explode( ' ', $this->curItem[0], 2 );
 588                         $this->saveValue( $ns, $tag, $this->charContent );
 589                         return;
 590                 } else {
 591                         array_shift( $this->mode );
 592                         array_shift( $this->curItem );
 593                 }
 594
 595
 596         }
 597
 598         /**
 599          * Handler for hitting a closing element.
 600          *
 601          * generally just calls a helper function depending on what
 602          * mode we're in.
 603          *
 604          * Ignores the outer wrapping elements that are optional in
 605          * xmp and have no meaning.
 606          *
 607          * @param $parser XMLParser
 608          * @param $elm String namespace . ' ' . element name
 609          * @throws MWException
 610          */
 611         function endElement( $parser, $elm ) {
 612                 if ( $elm === ( self::NS_RDF . ' RDF' )
 613                         || $elm === 'adobe:ns:meta/ xmpmeta'
 614                         || $elm === 'adobe:ns:meta/ xapmeta' )
 615                 {
 616                         // ignore these.
 617                         return;
 618                 }
 619
 620                 if ( $elm === self::NS_RDF . ' type' ) {
 621                         // these aren't really supported properly yet.
 622                         // However, it appears they almost never used.
 623                         wfDebugLog( 'XMP', __METHOD__ . ' encountered <rdf:type>' );
 624                 }
 625
 626                 if ( strpos( $elm, ' ' ) === false ) {
 627                         // This probably shouldn't happen.
 628                         // However, there is a bug in an adobe product
 629                         // that forgets the namespace on some things.
 630                         // (Luckily they are unimportant things).
 631                         wfDebugLog( 'XMP', __METHOD__ . " Encountered </$elm> which has no namespace. Skipping." );
 632                         return;
 633                 }
 634
 635                 if ( count( $this->mode[0] ) === 0 ) {
 636                         // This should never ever happen and means
 637                         // there is a pretty major bug in this class.
 638                         throw new MWException( 'Encountered end element with no mode' );
 639                 }
 640
 641                 if ( count( $this->curItem ) == 0 && $this->mode[0] !== self::MODE_INITIAL ) {
 642                         // just to be paranoid. Should always have a curItem, except for initially
 643                         // (aka during MODE_INITAL).
 644                         throw new MWException( "Hit end element </$elm> but no curItem" );
 645                 }
 646
 647                 switch( $this->mode[0] ) {
 648                         case self::MODE_IGNORE:
 649                                 $this->endElementModeIgnore( $elm );
 650                                 break;
 651                         case self::MODE_SIMPLE:
 652                                 $this->endElementModeSimple( $elm );
 653                                 break;
 654                         case self::MODE_STRUCT:
 655                         case self::MODE_SEQ:
 656                         case self::MODE_BAG:
 657                         case self::MODE_LANG:
 658                         case self::MODE_BAGSTRUCT:
 659                                 $this->endElementNested( $elm );
 660                                 break;
 661                         case self::MODE_INITIAL:
 662                                 if ( $elm === self::NS_RDF . ' Description' ) {
 663                                         array_shift( $this->mode );
 664                                 } else {
 665                                         throw new MWException( 'Element ended unexpectedly while in MODE_INITIAL' );
 666                                 }
 667                                 break;
 668                         case self::MODE_LI:
 669                         case self::MODE_LI_LANG:
 670                                 $this->endElementModeLi( $elm );
 671                                 break;
 672                         case self::MODE_QDESC:
 673                                 $this->endElementModeQDesc( $elm );
 674                                 break;
 675                         default:
 676                                 wfDebugLog( 'XMP', __METHOD__ . " no mode (elm = $elm)" );
 677                                 break;
 678                 }
 679         }
 680
 681         /**
 682         * Hit an opening element while in MODE_IGNORE
 683         *
 684         * XMP is extensible, so ignore any tag we don't understand.
 685         *
 686         * Mostly ignores, unless we encounter the element that we are ignoring.
 687         * in which case we add it to the item stack, so we can ignore things
 688         * that are nested, correctly.
 689         *
 690         * @param $elm String namespace . ' ' . tag name
 691         */
 692         private function startElementModeIgnore( $elm ) {
 693                 if ( $elm === $this->curItem[0] ) {
 694                         array_unshift( $this->curItem, $elm );
 695                         array_unshift( $this->mode, self::MODE_IGNORE );
 696                 }
 697         }
 698
 699         /**
 700         *  Start element in MODE_BAG (unordered array)
 701         * this should always be <rdf:Bag>
 702         *
 703         * @param $elm String namespace . ' ' . tag
 704         * @throws MWException if we have an element that's not <rdf:Bag>
 705         */
 706         private function startElementModeBag( $elm ) {
 707                 if ( $elm === self::NS_RDF . ' Bag' ) {
 708                         array_unshift( $this->mode, self::MODE_LI );
 709                 } else {
 710                         throw new MWException( "Expected <rdf:Bag> but got $elm." );
 711                 }
 712
 713         }
 714
 715         /**
 716         * Start element in MODE_SEQ (ordered array)
 717         * this should always be <rdf:Seq>
 718         *
 719         * @param $elm String namespace . ' ' . tag
 720         * @throws MWException if we have an element that's not <rdf:Seq>
 721         */
 722         private function startElementModeSeq( $elm ) {
 723                 if ( $elm === self::NS_RDF . ' Seq' ) {
 724                         array_unshift( $this->mode, self::MODE_LI );
 725                 } elseif ( $elm === self::NS_RDF . ' Bag' ) {
 726                         # bug 27105
 727                         wfDebugLog( 'XMP', __METHOD__ . ' Expected an rdf:Seq, but got an rdf:Bag. Pretending'
 728                                 . ' it is a Seq, since some buggy software is known to screw this up.' );
 729                         array_unshift( $this->mode, self::MODE_LI );
 730                 } else {
 731                         throw new MWException( "Expected <rdf:Seq> but got $elm." );
 732                 }
 733
 734         }
 735
 736         /**
 737         * Start element in MODE_LANG (language alternative)
 738         * this should always be <rdf:Alt>
 739         *
 740         * This tag tends to be used for metadata like describe this
 741         * picture, which can be translated into multiple languages.
 742         *
 743         * XMP supports non-linguistic alternative selections,
 744         * which are really only used for thumbnails, which
 745         * we don't care about.
 746         *
 747         * @param $elm String namespace . ' ' . tag
 748         * @throws MWException if we have an element that's not <rdf:Alt>
 749         */
 750         private function startElementModeLang( $elm ) {
 751                 if ( $elm === self::NS_RDF . ' Alt' ) {
 752                         array_unshift( $this->mode, self::MODE_LI_LANG );
 753                 } else {
 754                         throw new MWException( "Expected <rdf:Seq> but got $elm." );
 755                 }
 756
 757         }
 758
 759         /**
 760          * Handle an opening element when in MODE_SIMPLE
 761          *
 762          * This should not happen often. This is for if a simple element
 763          * already opened has a child element. Could happen for a
 764          * qualified element.
 765          *
 766          * For example:
 767          * <exif:DigitalZoomRatio><rdf:Description><rdf:value>0/10</rdf:value>
 768          *   <foo:someQualifier>Bar</foo:someQualifier> </rdf:Description>
 769          *   </exif:DigitalZoomRatio>
 770          *
 771          * This method is called when processing the <rdf:Description> element
 772          *
 773          * @param $elm String namespace and tag names separated by space.
 774          * @param $attribs Array Attributes of the element.
 775          * @throws MWException
 776          */
 777         private function startElementModeSimple( $elm, $attribs ) {
 778                 if ( $elm === self::NS_RDF . ' Description' ) {
 779                         // If this value has qualifiers
 780                         array_unshift( $this->mode, self::MODE_QDESC );
 781                         array_unshift( $this->curItem, $this->curItem[0] );
 782
 783                         if ( isset( $attribs[self::NS_RDF . ' value'] ) ) {
 784                                 list( $ns, $tag ) = explode( ' ', $this->curItem[0], 2 );
 785                                 $this->saveValue( $ns, $tag, $attribs[self::NS_RDF . ' value'] );
 786                         }
 787                 } elseif ( $elm === self::NS_RDF . ' value' ) {
 788                         // This should not be here.
 789                         throw new MWException( __METHOD__ . ' Encountered <rdf:value> where it was unexpected.' );
 790
 791                 } else {
 792                         // something else we don't recognize, like a qualifier maybe.
 793                         wfDebugLog( 'XMP', __METHOD__ . " Encountered element <$elm> where only expecting character data as value of " . $this->curItem[0] );
 794                         array_unshift( $this->mode, self::MODE_IGNORE );
 795                         array_unshift( $this->curItem, $elm );
 796
 797                 }
 798
 799         }
 800
 801         /**
 802         * Start an element when in MODE_QDESC.
 803         * This generally happens when a simple element has an inner
 804         * rdf:Description to hold qualifier elements.
 805         *
 806         * For example in:
 807         * <exif:DigitalZoomRatio><rdf:Description><rdf:value>0/10</rdf:value>
 808         *   <foo:someQualifier>Bar</foo:someQualifier> </rdf:Description>
 809         *   </exif:DigitalZoomRatio>
 810         * Called when processing the <rdf:value> or <foo:someQualifier>.
 811         *
 812         * @param $elm String namespace and tag name separated by a space.
 813         *
 814         */
 815         private function startElementModeQDesc( $elm ) {
 816                 if ( $elm === self::NS_RDF . ' value' ) {
 817                         return; // do nothing
 818                 } else {
 819                         // otherwise its a qualifier, which we ignore
 820                         array_unshift( $this->mode, self::MODE_IGNORE );
 821                         array_unshift( $this->curItem, $elm );
 822                 }
 823         }
 824
 825         /**
 826          * Starting an element when in MODE_INITIAL
 827          * This usually happens when we hit an element inside
 828          * the outer rdf:Description
 829          *
 830          * This is generally where most properties start.
 831          *
 832          * @param $ns String Namespace
 833          * @param $tag String tag name (without namespace prefix)
 834          * @param $attribs Array array of attributes
 835          * @throws MWException
 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          * @throws MWException
 897          */
 898         private function startElementModeStruct( $ns, $tag, $attribs ) {
 899                 if ( $ns !== self::NS_RDF ) {
 900
 901                         if ( isset( $this->items[$ns][$tag] ) ) {
 902                                 if ( isset( $this->items[$ns][$this->ancestorStruct]['children'] )
 903                                         && !isset( $this->items[$ns][$this->ancestorStruct]['children'][$tag] ) )
 904                                 {
 905                                         // This assumes that we don't have inter-namespace nesting
 906                                         // which we don't in all the properties we're interested in.
 907                                         throw new MWException( " <$tag> appeared nested in <" . $this->ancestorStruct
 908                                                 . "> where it is not allowed." );
 909                                 }
 910                                 array_unshift( $this->mode, $this->items[$ns][$tag]['mode'] );
 911                                 array_unshift( $this->curItem, $ns . ' ' . $tag );
 912                                 if ( $this->charContent !== false ) {
 913                                         // Something weird.
 914                                         // Should not happen in valid XMP.
 915                                         throw new MWException( "tag <$tag> nested in non-whitespace characters (" . $this->charContent . ")." );
 916                                 }
 917                         } else {
 918                                 array_unshift( $this->mode, self::MODE_IGNORE );
 919                                 array_unshift( $this->curItem, $elm );
 920                                 return;
 921                         }
 922
 923                 }
 924
 925                 if ( $ns === self::NS_RDF && $tag === 'Description' ) {
 926                         $this->doAttribs( $attribs );
 927                         array_unshift( $this->mode, self::MODE_STRUCT );
 928                         array_unshift( $this->curItem, $this->curItem[0] );
 929                 }
 930         }
 931
 932         /**
 933         * opening element in MODE_LI
 934         * process elements of arrays.
 935         *
 936         * Example:
 937         * <exif:ISOSpeedRatings> <rdf:Seq> <rdf:li>64</rdf:li>
 938         *   </rdf:Seq> </exif:ISOSpeedRatings>
 939         * This method is called when we hit the <rdf:li> element.
 940         *
 941         * @param $elm String: namespace . ' ' . tagname
 942         * @param $attribs Array: Attributes. (needed for BAGSTRUCTS)
 943         * @throws MWException if gets a tag other than <rdf:li>
 944         */
 945         private function startElementModeLi( $elm, $attribs ) {
 946                 if ( ( $elm ) !== self::NS_RDF . ' li' ) {
 947                         throw new MWException( "<rdf:li> expected but got $elm." );
 948                 }
 949
 950                 if ( !isset( $this->mode[1] ) ) {
 951                         // This should never ever ever happen. Checking for it
 952                         // to be paranoid.
 953                         throw new MWException( 'In mode Li, but no 2xPrevious mode!' );
 954                 }
 955
 956                 if ( $this->mode[1] === self::MODE_BAGSTRUCT ) {
 957                         // This list item contains a compound (STRUCT) value.
 958                         array_unshift( $this->mode, self::MODE_STRUCT );
 959                         array_unshift( $this->curItem, $elm );
 960                         $this->processingArray = true;
 961
 962                         if ( !isset( $this->curItem[1] ) ) {
 963                                 // be paranoid.
 964                                 throw new MWException( 'Can not find parent of BAGSTRUCT.' );
 965                         }
 966                         list( $curNS, $curTag ) = explode( ' ', $this->curItem[1] );
 967                         $this->ancestorStruct = isset( $this->items[$curNS][$curTag]['map_name'] )
 968                                 ? $this->items[$curNS][$curTag]['map_name'] : $curTag;
 969
 970                         $this->doAttribs( $attribs );
 971
 972                 } else {
 973                         // Normal BAG or SEQ containing simple values.
 974                         array_unshift( $this->mode, self::MODE_SIMPLE );
 975                         // need to add curItem[0] on again since one is for the specific item
 976                         // and one is for the entire group.
 977                         array_unshift( $this->curItem, $this->curItem[0] );
 978                         $this->processingArray = true;
 979                 }
 980
 981         }
 982
 983         /**
 984         * Opening element in MODE_LI_LANG.
 985         * process elements of language alternatives
 986         *
 987         * Example:
 988         * <dc:title> <rdf:Alt> <rdf:li xml:lang="x-default">My house
 989         *  </rdf:li> </rdf:Alt> </dc:title>
 990         *
 991         * This method is called when we hit the <rdf:li> element.
 992         *
 993         * @param $elm String namespace . ' ' . tag
 994         * @param $attribs array array of elements (most importantly xml:lang)
 995         * @throws MWException if gets a tag other than <rdf:li> or if no xml:lang
 996         */
 997         private function startElementModeLiLang( $elm, $attribs ) {
 998                 if ( $elm !== self::NS_RDF . ' li' ) {
 999                         throw new MWException( __METHOD__ . " <rdf:li> expected but got $elm." );
1000                 }
1001                 if ( !isset( $attribs[ self::NS_XML . ' lang'] )
1002                         || !preg_match( '/^[-A-Za-z0-9]{2,}$/D', $attribs[ self::NS_XML . ' lang' ] ) )
1003                 {
1004                         throw new MWException( __METHOD__
1005                                 . " <rdf:li> did not contain, or has invalid xml:lang attribute in lang alternative" );
1006                 }
1007
1008                 // Lang is case-insensitive.
1009                 $this->itemLang = strtolower( $attribs[ self::NS_XML . ' lang' ] );
1010
1011                 // need to add curItem[0] on again since one is for the specific item
1012                 // and one is for the entire group.
1013                 array_unshift( $this->curItem, $this->curItem[0] );
1014                 array_unshift( $this->mode, self::MODE_SIMPLE );
1015                 $this->processingArray = true;
1016         }
1017
1018         /**
1019          * Hits an opening element.
1020          * Generally just calls a helper based on what MODE we're in.
1021          * Also does some initial set up for the wrapper element
1022          *
1023          * @param $parser XMLParser
1024          * @param $elm String namespace "<space>" element
1025          * @param $attribs Array attribute name => value
1026          * @throws MWException
1027          */
1028         function startElement( $parser, $elm, $attribs ) {
1029
1030                 if ( $elm === self::NS_RDF . ' RDF'
1031                         || $elm === 'adobe:ns:meta/ xmpmeta'
1032                         || $elm === 'adobe:ns:meta/ xapmeta' )
1033                 {
1034                         /* ignore. */
1035                         return;
1036                 } elseif ( $elm === self::NS_RDF . ' Description' ) {
1037                         if ( count( $this->mode ) === 0 ) {
1038                                 // outer rdf:desc
1039                                 array_unshift( $this->mode, self::MODE_INITIAL );
1040                         }
1041                 } elseif ( $elm === self::NS_RDF . ' type' ) {
1042                         // This doesn't support rdf:type properly.
1043                         // In practise I have yet to see a file that
1044                         // uses this element, however it is mentioned
1045                         // on page 25 of part 1 of the xmp standard.
1046                         //
1047                         // also it seems as if exiv2 and exiftool do not support
1048                         // this either (That or I misunderstand the standard)
1049                         wfDebugLog( 'XMP', __METHOD__ . ' Encountered <rdf:type> which isn\'t currently supported' );
1050                 }
1051
1052                 if ( strpos( $elm, ' ' ) === false ) {
1053                         // This probably shouldn't happen.
1054                         wfDebugLog( 'XMP', __METHOD__ . " Encountered <$elm> which has no namespace. Skipping." );
1055                         return;
1056                 }
1057
1058                 list( $ns, $tag ) = explode( ' ', $elm, 2 );
1059
1060                 if ( count( $this->mode ) === 0 ) {
1061                         // This should not happen.
1062                         throw new MWException( 'Error extracting XMP, '
1063                                 . "encountered <$elm> with no mode" );
1064                 }
1065
1066                 switch( $this->mode[0] ) {
1067                         case self::MODE_IGNORE:
1068                                 $this->startElementModeIgnore( $elm );
1069                                 break;
1070                         case self::MODE_SIMPLE:
1071                                 $this->startElementModeSimple( $elm, $attribs );
1072                                 break;
1073                         case self::MODE_INITIAL:
1074                                 $this->startElementModeInitial( $ns, $tag, $attribs );
1075                                 break;
1076                         case self::MODE_STRUCT:
1077                                 $this->startElementModeStruct( $ns, $tag, $attribs );
1078                                 break;
1079                         case self::MODE_BAG:
1080                         case self::MODE_BAGSTRUCT:
1081                                 $this->startElementModeBag( $elm );
1082                                 break;
1083                         case self::MODE_SEQ:
1084                                 $this->startElementModeSeq( $elm );
1085                                 break;
1086                         case self::MODE_LANG:
1087                                 $this->startElementModeLang( $elm );
1088                                 break;
1089                         case self::MODE_LI_LANG:
1090                                 $this->startElementModeLiLang( $elm, $attribs );
1091                                 break;
1092                         case self::MODE_LI:
1093                                 $this->startElementModeLi( $elm, $attribs );
1094                                 break;
1095                         case self::MODE_QDESC:
1096                                 $this->startElementModeQDesc( $elm );
1097                                 break;
1098                         default:
1099                                 throw new MWException( 'StartElement in unknown mode: ' . $this->mode[0] );
1100                 }
1101         }
1102
1103         /**
1104          * Process attributes.
1105          * Simple values can be stored as either a tag or attribute
1106          *
1107          * Often the initial "<rdf:Description>" tag just has all the simple
1108          * properties as attributes.
1109          *
1110          * @par Example:
1111          * @code
1112          * <rdf:Description rdf:about="" xmlns:exif="http://ns.adobe.com/exif/1.0/" exif:DigitalZoomRatio="0/10">
1113          * @endcode
1114          *
1115          * @param $attribs Array attribute=>value array.
1116          * @throws MWException
1117          */
1118         private function doAttribs( $attribs ) {
1119
1120                 // first check for rdf:parseType attribute, as that can change
1121                 // how the attributes are interperted.
1122
1123                 if ( isset( $attribs[self::NS_RDF . ' parseType'] )
1124                         && $attribs[self::NS_RDF . ' parseType'] === 'Resource'
1125                         && $this->mode[0] === self::MODE_SIMPLE )
1126                 {
1127                         // this is equivalent to having an inner rdf:Description
1128                         $this->mode[0] = self::MODE_QDESC;
1129                 }
1130                 foreach ( $attribs as $name => $val ) {
1131
1132
1133                         if ( strpos( $name, ' ' ) === false ) {
1134                                 // This shouldn't happen, but so far some old software forgets namespace
1135                                 // on rdf:about.
1136                                 wfDebugLog( 'XMP', __METHOD__ . ' Encountered non-namespaced attribute: '
1137                                         . " $name=\"$val\". Skipping. " );
1138                                 continue;
1139                         }
1140                         list( $ns, $tag ) = explode( ' ', $name, 2 );
1141                         if ( $ns === self::NS_RDF ) {
1142                                 if ( $tag === 'value' || $tag === 'resource' ) {
1143                                         // resource is for url.
1144                                         // value attribute is a weird way of just putting the contents.
1145                                         $this->char( $this->xmlParser, $val );
1146                                 }
1147                         } elseif ( isset( $this->items[$ns][$tag] ) ) {
1148                                 if ( $this->mode[0] === self::MODE_SIMPLE ) {
1149                                         throw new MWException( __METHOD__
1150                                                 . " $ns:$tag found as attribute where not allowed" );
1151                                 }
1152                                 $this->saveValue( $ns, $tag, $val );
1153                         } else {
1154                                 wfDebugLog( 'XMP', __METHOD__ . " Ignoring unrecognized element <$ns:$tag>." );
1155                         }
1156                 }
1157         }
1158
1159         /**
1160         * Given an extracted value, save it to results array
1161         *
1162         * note also uses $this->ancestorStruct and
1163         * $this->processingArray to determine what name to
1164         * save the value under. (in addition to $tag).
1165         *
1166         * @param $ns String namespace of tag this is for
1167         * @param $tag String tag name
1168         * @param $val String value to save
1169         */
1170         private function saveValue( $ns, $tag, $val ) {
1171
1172                 $info =& $this->items[$ns][$tag];
1173                 $finalName = isset( $info['map_name'] )
1174                         ? $info['map_name'] : $tag;
1175                 if ( isset( $info['validate'] ) ) {
1176                         $validate = is_array( $info['validate'] ) ? $info['validate']
1177                                 : array( 'XMPValidate', $info['validate'] );
1178
1179                         if ( is_callable( $validate ) ) {
1180                                 call_user_func_array( $validate, array( $info, &$val, true ) );
1181                                 // the reasoning behind using &$val instead of using the return value
1182                                 // is to be consistent between here and validating structures.
1183                                 if ( is_null( $val ) ) {
1184                                         wfDebugLog( 'XMP', __METHOD__ . " <$ns:$tag> failed validation." );
1185                                         return;
1186                                 }
1187                         } else {
1188                                 wfDebugLog( 'XMP', __METHOD__ . " Validation function for $finalName ("
1189                                         . $validate[0] . '::' . $validate[1] . '()) is not callable.' );
1190                         }
1191                 }
1192
1193                 if ( $this->ancestorStruct && $this->processingArray ) {
1194                         // Aka both an array and a struct. ( self::MODE_BAGSTRUCT )
1195                         $this->results['xmp-' . $info['map_group']][$this->ancestorStruct][][$finalName] = $val;
1196                 } elseif ( $this->ancestorStruct ) {
1197                         $this->results['xmp-' . $info['map_group']][$this->ancestorStruct][$finalName] = $val;
1198                 } elseif ( $this->processingArray ) {
1199                         if ( $this->itemLang === false ) {
1200                                 // normal array
1201                                 $this->results['xmp-' . $info['map_group']][$finalName][] = $val;
1202                         } else {
1203                                 // lang array.
1204                                 $this->results['xmp-' . $info['map_group']][$finalName][$this->itemLang] = $val;
1205                         }
1206                 } else {
1207                         $this->results['xmp-' . $info['map_group']][$finalName] = $val;
1208                 }
1209         }
1210 }