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