includes/media/Exif.php

   1 <?php
   2 /**
   3  * This program is free software; you can redistribute it and/or modify
   4  * it under the terms of the GNU General Public License as published by
   5  * the Free Software Foundation; either version 2 of the License, or
   6  * (at your option) any later version.
   7  *
   8  * This program is distributed in the hope that it will be useful,
   9  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  10  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  11  * GNU General Public License for more details.
  12  *
  13  * You should have received a copy of the GNU General Public License along
  14  * with this program; if not, write to the Free Software Foundation, Inc.,
  15  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  16  * http://www.gnu.org/copyleft/gpl.html
  17  *
  18  * @ingroup Media
  19  * @author Ævar Arnfjörð Bjarmason <avarab@gmail.com>
  20  * @copyright Copyright © 2005, Ævar Arnfjörð Bjarmason, 2009 Brent Garber
  21  * @license http://www.gnu.org/copyleft/gpl.html GNU General Public License
  22  * @see http://exif.org/Exif2-2.PDF The Exif 2.2 specification
  23  * @file
  24  */
  25
  26 /**
  27  * Class to extract and validate Exif data from jpeg (and possibly tiff) files.
  28  * @ingroup Media
  29  */
  30 class Exif {
  31
  32         const BYTE      = 1;    //!< An 8-bit (1-byte) unsigned integer.
  33         const ASCII     = 2;    //!< An 8-bit byte containing one 7-bit ASCII code. The final byte is terminated with NULL.
  34         const SHORT     = 3;    //!< A 16-bit (2-byte) unsigned integer.
  35         const LONG      = 4;    //!< A 32-bit (4-byte) unsigned integer.
  36         const RATIONAL  = 5;    //!< Two LONGs. The first LONG is the numerator and the second LONG expresses the denominator
  37         const UNDEFINED = 7;    //!< An 8-bit byte that can take any value depending on the field definition
  38         const SLONG     = 9;    //!< A 32-bit (4-byte) signed integer (2's complement notation),
  39         const SRATIONAL = 10;   //!< Two SLONGs. The first SLONG is the numerator and the second SLONG is the denominator.
  40         const IGNORE    = -1;   // A fake value for things we don't want or don't support.
  41
  42         //@{
  43         /* @var array
  44          * @private
  45          */
  46
  47         /**
  48          * Exif tags grouped by category, the tagname itself is the key and the type
  49          * is the value, in the case of more than one possible value type they are
  50          * separated by commas.
  51          */
  52         var $mExifTags;
  53
  54         /**
  55          * The raw Exif data returned by exif_read_data()
  56          */
  57         var $mRawExifData;
  58
  59         /**
  60          * A Filtered version of $mRawExifData that has been pruned of invalid
  61          * tags and tags that contain content they shouldn't contain according
  62          * to the Exif specification
  63          */
  64         var $mFilteredExifData;
  65
  66         /**
  67          * Filtered and formatted Exif data, see FormatMetadata::getFormattedData()
  68          */
  69         var $mFormattedExifData;
  70
  71         //@}
  72
  73         //@{
  74         /* @var string
  75          * @private
  76          */
  77
  78         /**
  79          * The file being processed
  80          */
  81         var $file;
  82
  83         /**
  84          * The basename of the file being processed
  85          */
  86         var $basename;
  87
  88         /**
  89          * The private log to log to, e.g. 'exif'
  90          */
  91         var $log = false;
  92
  93         /**
  94          * The byte order of the file. Needed because php's
  95          * extension doesn't fully process some obscure props.
  96          */
  97         private $byteOrder;
  98         //@}
  99
 100         /**
 101          * Constructor
 102          *
 103          * @param $file String: filename.
 104          * @todo FIXME: The following are broke:
 105          * SubjectArea. Need to test the more obscure tags.
 106          *
 107          * DigitalZoomRatio = 0/0 is rejected. need to determine if that's valid.
 108          * possibly should treat 0/0 = 0. need to read exif spec on that.
 109          */
 110         function __construct( $file, $byteOrder = '' ) {
 111                 /**
 112                  * Page numbers here refer to pages in the EXIF 2.2 standard
 113                  *
 114                  * Note, Exif::UNDEFINED is treated as a string, not as an array of bytes
 115                  * so don't put a count parameter for any UNDEFINED values.
 116                  *
 117                  * @link http://exif.org/Exif2-2.PDF The Exif 2.2 specification
 118                  */
 119                 $this->mExifTags = array(
 120                         # TIFF Rev. 6.0 Attribute Information (p22)
 121                         'IFD0' => array(
 122                                 # Tags relating to image structure
 123                                 'ImageWidth' => Exif::SHORT.','.Exif::LONG,             # Image width
 124                                 'ImageLength' => Exif::SHORT.','.Exif::LONG,            # Image height
 125                                 'BitsPerSample' => array( Exif::SHORT, 3 ),             # Number of bits per component
 126                                 # "When a primary image is JPEG compressed, this designation is not"
 127                                 # "necessary and is omitted." (p23)
 128                                 'Compression' => Exif::SHORT,                           # Compression scheme #p23
 129                                 'PhotometricInterpretation' => Exif::SHORT,             # Pixel composition #p23
 130                                 'Orientation' => Exif::SHORT,                           # Orientation of image #p24
 131                                 'SamplesPerPixel' => Exif::SHORT,                       # Number of components
 132                                 'PlanarConfiguration' => Exif::SHORT,                   # Image data arrangement #p24
 133                                 'YCbCrSubSampling' => array( Exif::SHORT, 2),           # Subsampling ratio of Y to C #p24
 134                                 'YCbCrPositioning' => Exif::SHORT,                      # Y and C positioning #p24-25
 135                                 'XResolution' => Exif::RATIONAL,                        # Image resolution in width direction
 136                                 'YResolution' => Exif::RATIONAL,                        # Image resolution in height direction
 137                                 'ResolutionUnit' => Exif::SHORT,                        # Unit of X and Y resolution #(p26)
 138
 139                                 # Tags relating to recording offset
 140                                 'StripOffsets' => Exif::SHORT.','.Exif::LONG,                   # Image data location
 141                                 'RowsPerStrip' => Exif::SHORT.','.Exif::LONG,                   # Number of rows per strip
 142                                 'StripByteCounts' => Exif::SHORT.','.Exif::LONG,                # Bytes per compressed strip
 143                                 'JPEGInterchangeFormat' => Exif::SHORT.','.Exif::LONG,          # Offset to JPEG SOI
 144                                 'JPEGInterchangeFormatLength' => Exif::SHORT.','.Exif::LONG,    # Bytes of JPEG data
 145
 146                                 # Tags relating to image data characteristics
 147                                 'TransferFunction' => Exif::IGNORE,                     # Transfer function
 148                                 'WhitePoint' => array( Exif::RATIONAL, 2),              # White point chromaticity
 149                                 'PrimaryChromaticities' => array( Exif::RATIONAL, 6),   # Chromaticities of primarities
 150                                 'YCbCrCoefficients' => array( Exif::RATIONAL, 3),       # Color space transformation matrix coefficients #p27
 151                                 'ReferenceBlackWhite' => array( Exif::RATIONAL, 6),     # Pair of black and white reference values
 152
 153                                 # Other tags
 154                                 'DateTime' => Exif::ASCII,                              # File change date and time
 155                                 'ImageDescription' => Exif::ASCII,                      # Image title
 156                                 'Make' => Exif::ASCII,                                  # Image input equipment manufacturer
 157                                 'Model' => Exif::ASCII,                                 # Image input equipment model
 158                                 'Software' => Exif::ASCII,                              # Software used
 159                                 'Artist' => Exif::ASCII,                                # Person who created the image
 160                                 'Copyright' => Exif::ASCII,                             # Copyright holder
 161                         ),
 162
 163                         # Exif IFD Attribute Information (p30-31)
 164                         'EXIF' => array(
 165                                 # TODO: NOTE: Nonexistence of this field is taken to mean nonconformance
 166                                 # to the EXIF 2.1 AND 2.2 standards
 167                                 'ExifVersion' =>  Exif::UNDEFINED,                      # Exif version
 168                                 'FlashPixVersion' => Exif::UNDEFINED,                   # Supported Flashpix version #p32
 169
 170                                 # Tags relating to Image Data Characteristics
 171                                 'ColorSpace' => Exif::SHORT,                            # Color space information #p32
 172
 173                                 # Tags relating to image configuration
 174                                 'ComponentsConfiguration' => Exif::UNDEFINED,                   # Meaning of each component #p33
 175                                 'CompressedBitsPerPixel' => Exif::RATIONAL,                     # Image compression mode
 176                                 'PixelYDimension' => Exif::SHORT.','.Exif::LONG,                # Valid image width
 177                                 'PixelXDimension' => Exif::SHORT.','.Exif::LONG,                # Valid image height
 178
 179                                 # Tags relating to related user information
 180                                 'MakerNote' => Exif::IGNORE,                            # Manufacturer notes
 181                                 'UserComment' => Exif::UNDEFINED,                       # User comments #p34
 182
 183                                 # Tags relating to related file information
 184                                 'RelatedSoundFile' => Exif::ASCII,                      # Related audio file
 185
 186                                 # Tags relating to date and time
 187                                 'DateTimeOriginal' => Exif::ASCII,                      # Date and time of original data generation #p36
 188                                 'DateTimeDigitized' => Exif::ASCII,                     # Date and time of original data generation
 189                                 'SubSecTime' => Exif::ASCII,                            # DateTime subseconds
 190                                 'SubSecTimeOriginal' => Exif::ASCII,                    # DateTimeOriginal subseconds
 191                                 'SubSecTimeDigitized' => Exif::ASCII,                   # DateTimeDigitized subseconds
 192
 193                                 # Tags relating to picture-taking conditions (p31)
 194                                 'ExposureTime' => Exif::RATIONAL,                       # Exposure time
 195                                 'FNumber' => Exif::RATIONAL,                            # F Number
 196                                 'ExposureProgram' => Exif::SHORT,                       # Exposure Program #p38
 197                                 'SpectralSensitivity' => Exif::ASCII,                   # Spectral sensitivity
 198                                 'ISOSpeedRatings' => Exif::SHORT,                       # ISO speed rating
 199                                 'OECF' => Exif::IGNORE,
 200                                 # Optoelectronic conversion factor. Note: We don't have support for this atm.
 201                                 'ShutterSpeedValue' => Exif::SRATIONAL,                 # Shutter speed
 202                                 'ApertureValue' => Exif::RATIONAL,                      # Aperture
 203                                 'BrightnessValue' => Exif::SRATIONAL,                   # Brightness
 204                                 'ExposureBiasValue' => Exif::SRATIONAL,                 # Exposure bias
 205                                 'MaxApertureValue' => Exif::RATIONAL,                   # Maximum land aperture
 206                                 'SubjectDistance' => Exif::RATIONAL,                    # Subject distance
 207                                 'MeteringMode' => Exif::SHORT,                          # Metering mode #p40
 208                                 'LightSource' => Exif::SHORT,                           # Light source #p40-41
 209                                 'Flash' => Exif::SHORT,                                 # Flash #p41-42
 210                                 'FocalLength' => Exif::RATIONAL,                        # Lens focal length
 211                                 'SubjectArea' => array( Exif::SHORT, 4 ),               # Subject area
 212                                 'FlashEnergy' => Exif::RATIONAL,                        # Flash energy
 213                                 'SpatialFrequencyResponse' => Exif::IGNORE,             # Spatial frequency response. Not supported atm.
 214                                 'FocalPlaneXResolution' => Exif::RATIONAL,              # Focal plane X resolution
 215                                 'FocalPlaneYResolution' => Exif::RATIONAL,              # Focal plane Y resolution
 216                                 'FocalPlaneResolutionUnit' => Exif::SHORT,              # Focal plane resolution unit #p46
 217                                 'SubjectLocation' => array( Exif::SHORT, 2),            # Subject location
 218                                 'ExposureIndex' => Exif::RATIONAL,                      # Exposure index
 219                                 'SensingMethod' => Exif::SHORT,                         # Sensing method #p46
 220                                 'FileSource' => Exif::UNDEFINED,                        # File source #p47
 221                                 'SceneType' => Exif::UNDEFINED,                         # Scene type #p47
 222                                 'CFAPattern' => Exif::IGNORE,                           # CFA pattern. not supported atm.
 223                                 'CustomRendered' => Exif::SHORT,                        # Custom image processing #p48
 224                                 'ExposureMode' => Exif::SHORT,                          # Exposure mode #p48
 225                                 'WhiteBalance' => Exif::SHORT,                          # White Balance #p49
 226                                 'DigitalZoomRatio' => Exif::RATIONAL,                   # Digital zoom ration
 227                                 'FocalLengthIn35mmFilm' => Exif::SHORT,                 # Focal length in 35 mm film
 228                                 'SceneCaptureType' => Exif::SHORT,                      # Scene capture type #p49
 229                                 'GainControl' => Exif::SHORT,                           # Scene control #p49-50
 230                                 'Contrast' => Exif::SHORT,                              # Contrast #p50
 231                                 'Saturation' => Exif::SHORT,                            # Saturation #p50
 232                                 'Sharpness' => Exif::SHORT,                             # Sharpness #p50
 233                                 'DeviceSettingDescription' => Exif::IGNORE,
 234                                 # Device settings description. This could maybe be supported. Need to find an
 235                                 # example file that uses this to see if it has stuff of interest in it.
 236                                 'SubjectDistanceRange' => Exif::SHORT,                  # Subject distance range #p51
 237
 238                                 'ImageUniqueID' => Exif::ASCII,                         # Unique image ID
 239                         ),
 240
 241                         # GPS Attribute Information (p52)
 242                         'GPS' => array(
 243                                 'GPSVersion' => Exif::UNDEFINED,
 244                                 # Should be an array of 4 Exif::BYTE's. However php treats it as an undefined
 245                                 # Note exif standard calls this GPSVersionID, but php doesn't like the id suffix
 246                                 'GPSLatitudeRef' => Exif::ASCII,                        # North or South Latitude #p52-53
 247                                 'GPSLatitude' => array( Exif::RATIONAL, 3 ),            # Latitude
 248                                 'GPSLongitudeRef' => Exif::ASCII,                       # East or West Longitude #p53
 249                                 'GPSLongitude' => array( Exif::RATIONAL, 3),            # Longitude
 250                                 'GPSAltitudeRef' => Exif::UNDEFINED,
 251                                 # Altitude reference. Note, the exif standard says this should be an EXIF::Byte,
 252                                 # but php seems to disagree.
 253                                 'GPSAltitude' => Exif::RATIONAL,                        # Altitude
 254                                 'GPSTimeStamp' => array( Exif::RATIONAL, 3),            # GPS time (atomic clock)
 255                                 'GPSSatellites' => Exif::ASCII,                         # Satellites used for measurement
 256                                 'GPSStatus' => Exif::ASCII,                             # Receiver status #p54
 257                                 'GPSMeasureMode' => Exif::ASCII,                        # Measurement mode #p54-55
 258                                 'GPSDOP' => Exif::RATIONAL,                             # Measurement precision
 259                                 'GPSSpeedRef' => Exif::ASCII,                           # Speed unit #p55
 260                                 'GPSSpeed' => Exif::RATIONAL,                           # Speed of GPS receiver
 261                                 'GPSTrackRef' => Exif::ASCII,                           # Reference for direction of movement #p55
 262                                 'GPSTrack' => Exif::RATIONAL,                           # Direction of movement
 263                                 'GPSImgDirectionRef' => Exif::ASCII,                    # Reference for direction of image #p56
 264                                 'GPSImgDirection' => Exif::RATIONAL,                    # Direction of image
 265                                 'GPSMapDatum' => Exif::ASCII,                           # Geodetic survey data used
 266                                 'GPSDestLatitudeRef' => Exif::ASCII,                    # Reference for latitude of destination #p56
 267                                 'GPSDestLatitude' => array( Exif::RATIONAL, 3 ),        # Latitude destination
 268                                 'GPSDestLongitudeRef' => Exif::ASCII,                   # Reference for longitude of destination #p57
 269                                 'GPSDestLongitude' => array( Exif::RATIONAL, 3 ),       # Longitude of destination
 270                                 'GPSDestBearingRef' => Exif::ASCII,                     # Reference for bearing of destination #p57
 271                                 'GPSDestBearing' => Exif::RATIONAL,                     # Bearing of destination
 272                                 'GPSDestDistanceRef' => Exif::ASCII,                    # Reference for distance to destination #p57-58
 273                                 'GPSDestDistance' => Exif::RATIONAL,                    # Distance to destination
 274                                 'GPSProcessingMethod' => Exif::UNDEFINED,               # Name of GPS processing method
 275                                 'GPSAreaInformation' => Exif::UNDEFINED,                # Name of GPS area
 276                                 'GPSDateStamp' => Exif::ASCII,                          # GPS date
 277                                 'GPSDifferential' => Exif::SHORT,                       # GPS differential correction
 278                         ),
 279                 );
 280
 281                 $this->file = $file;
 282                 $this->basename = wfBaseName( $this->file );
 283                 if ( $byteOrder === 'BE' || $byteOrder === 'LE' ) {
 284                         $this->byteOrder = $byteOrder;
 285                 } else {
 286                         // Only give a warning for b/c, since originally we didn't
 287                         // require this. The number of things affected by this is
 288                         // rather small.
 289                         wfWarn( 'Exif class did not have byte order specified. '
 290                          . 'Some properties may be decoded incorrectly.' );
 291                         $this->byteOrder = 'BE'; // BE seems about twice as popular as LE in jpg's.
 292                 }
 293
 294                 $this->debugFile( $this->basename, __FUNCTION__, true );
 295                 if( function_exists( 'exif_read_data' ) ) {
 296                         wfSuppressWarnings();
 297                         $data = exif_read_data( $this->file, 0, true );
 298                         wfRestoreWarnings();
 299                 } else {
 300                         throw new MWException( "Internal error: exif_read_data not present. \$wgShowEXIF may be incorrectly set or not checked by an extension." );
 301                 }
 302                 /**
 303                  * exif_read_data() will return false on invalid input, such as
 304                  * when somebody uploads a file called something.jpeg
 305                  * containing random gibberish.
 306                  */
 307                 $this->mRawExifData = $data ? $data : array();
 308                 $this->makeFilteredData();
 309                 $this->collapseData();
 310                 $this->debugFile( __FUNCTION__, false );
 311         }
 312
 313         /**
 314          * Make $this->mFilteredExifData
 315          */
 316         function makeFilteredData() {
 317                 $this->mFilteredExifData = Array();
 318
 319                 foreach ( array_keys( $this->mRawExifData ) as $section ) {
 320                         if ( !in_array( $section, array_keys( $this->mExifTags ) ) ) {
 321                                 $this->debug( $section , __FUNCTION__, "'$section' is not a valid Exif section" );
 322                                 continue;
 323                         }
 324
 325                         foreach ( array_keys( $this->mRawExifData[$section] ) as $tag ) {
 326                                 if ( !in_array( $tag, array_keys( $this->mExifTags[$section] ) ) ) {
 327                                         $this->debug( $tag, __FUNCTION__, "'$tag' is not a valid tag in '$section'" );
 328                                         continue;
 329                                 }
 330
 331                                 $this->mFilteredExifData[$tag] = $this->mRawExifData[$section][$tag];
 332                                 // This is ok, as the tags in the different sections do not conflict.
 333                                 // except in computed and thumbnail section, which we don't use.
 334
 335                                 $value = $this->mRawExifData[$section][$tag];
 336                                 if ( !$this->validate( $section, $tag, $value ) ) {
 337                                         $this->debug( $value, __FUNCTION__, "'$tag' contained invalid data" );
 338                                         unset( $this->mFilteredExifData[$tag] );
 339                                 }
 340                         }
 341                 }
 342         }
 343
 344         /**
 345         * Collapse some fields together.
 346         * This converts some fields from exif form, to a more friendly form.
 347         * For example GPS latitude to a single number.
 348         *
 349         * The rationale behind this is that we're storing data, not presenting to the user
 350         * For example a longitude is a single number describing how far away you are from
 351         * the prime meridian. Well it might be nice to split it up into minutes and seconds
 352         * for the user, it doesn't really make sense to split a single number into 4 parts
 353         * for storage. (degrees, minutes, second, direction vs single floating point number).
 354         *
 355         * Other things this might do (not really sure if they make sense or not):
 356         * Dates -> mediawiki date format.
 357         * convert values that can be in different units to be in one standardized unit.
 358         *
 359         * As an alternative approach, some of this could be done in the validate phase
 360         * if we make up our own types like Exif::DATE.
 361         */
 362         function collapseData( ) {
 363
 364                 $this->exifGPStoNumber( 'GPSLatitude' );
 365                 $this->exifGPStoNumber( 'GPSDestLatitude' );
 366                 $this->exifGPStoNumber( 'GPSLongitude' );
 367                 $this->exifGPStoNumber( 'GPSDestLongitude' );
 368
 369                 if ( isset( $this->mFilteredExifData['GPSAltitude'] ) && isset( $this->mFilteredExifData['GPSAltitudeRef'] ) ) {
 370                         if ( $this->mFilteredExifData['GPSAltitudeRef'] === "\1" ) {
 371                                 $this->mFilteredExifData['GPSAltitude'] *= - 1;
 372                         }
 373                         unset( $this->mFilteredExifData['GPSAltitudeRef'] );
 374                 }
 375
 376                 $this->exifPropToOrd( 'FileSource' );
 377                 $this->exifPropToOrd( 'SceneType' );
 378
 379                 $this->charCodeString( 'UserComment' );
 380                 $this->charCodeString( 'GPSProcessingMethod');
 381                 $this->charCodeString( 'GPSAreaInformation' );
 382
 383                 //ComponentsConfiguration should really be an array instead of a string...
 384                 //This turns a string of binary numbers into an array of numbers.
 385
 386                 if ( isset ( $this->mFilteredExifData['ComponentsConfiguration'] ) ) {
 387                         $val = $this->mFilteredExifData['ComponentsConfiguration'];
 388                         $ccVals = array();
 389                         for ($i = 0; $i < strlen($val); $i++) {
 390                                 $ccVals[$i] = ord( substr($val, $i, 1) );
 391                         }
 392                         $ccVals['_type'] = 'ol'; //this is for formatting later.
 393                         $this->mFilteredExifData['ComponentsConfiguration'] = $ccVals;
 394                 }
 395
 396                 //GPSVersion(ID) is treated as the wrong type by php exif support.
 397                 //Go through each byte turning it into a version string.
 398                 //For example: "\x02\x02\x00\x00" -> "2.2.0.0"
 399
 400                 //Also change exif tag name from GPSVersion (what php exif thinks it is)
 401                 //to GPSVersionID (what the exif standard thinks it is).
 402
 403                 if ( isset ( $this->mFilteredExifData['GPSVersion'] ) ) {
 404                         $val = $this->mFilteredExifData['GPSVersion'];
 405                         $newVal = '';
 406                         for ($i = 0; $i < strlen($val); $i++) {
 407                                 if ( $i !== 0 ) {
 408                                         $newVal .= '.';
 409                                 }
 410                                 $newVal .= ord( substr($val, $i, 1) );
 411                         }
 412                         if ( $this->byteOrder === 'LE' ) {
 413                                 // Need to reverse the string
 414                                 $newVal2 = '';
 415                                 for ( $i = strlen( $newVal ) - 1; $i >= 0; $i-- ) {
 416                                         $newVal2 .= substr( $newVal, $i, 1 );
 417                                 }
 418                                 $this->mFilteredExifData['GPSVersionID'] = $newVal2;
 419                         } else {
 420                                 $this->mFilteredExifData['GPSVersionID'] = $newVal;
 421                         }
 422                         unset( $this->mFilteredExifData['GPSVersion'] );
 423                 }
 424
 425         }
 426         /**
 427         * Do userComment tags and similar. See pg. 34 of exif standard.
 428         * basically first 8 bytes is charset, rest is value.
 429         * This has not been tested on any shift-JIS strings.
 430         * @param $prop String prop name.
 431         */
 432         private function charCodeString ( $prop ) {
 433                 if ( isset( $this->mFilteredExifData[$prop] ) ) {
 434
 435                         if ( strlen($this->mFilteredExifData[$prop]) <= 8 ) {
 436                                 //invalid. Must be at least 9 bytes long.
 437
 438                                 $this->debug( $this->mFilteredExifData[$prop] , __FUNCTION__, false );
 439                                 unset($this->mFilteredExifData[$prop]);
 440                                 return;
 441                         }
 442                         $charCode = substr( $this->mFilteredExifData[$prop], 0, 8);
 443                         $val = substr( $this->mFilteredExifData[$prop], 8);
 444
 445
 446                         switch ($charCode) {
 447                                 case "\x4A\x49\x53\x00\x00\x00\x00\x00":
 448                                         //JIS
 449                                         $charset = "Shift-JIS";
 450                                         break;
 451                                 case "UNICODE\x00":
 452                                         $charset = "UTF-16" . $this->byteOrder;
 453                                         break;
 454                                 default: //ascii or undefined.
 455                                         $charset = "";
 456                                         break;
 457                         }
 458                         // This could possibly check to see if iconv is really installed
 459                         // or if we're using the compatibility wrapper in globalFunctions.php
 460                         if ($charset) {
 461                                 wfSuppressWarnings();
 462                                 $val = iconv($charset, 'UTF-8//IGNORE', $val);
 463                                 wfRestoreWarnings();
 464                         } else {
 465                                 // if valid utf-8, assume that, otherwise assume windows-1252
 466                                 $valCopy = $val;
 467                                 UtfNormal::quickIsNFCVerify( $valCopy ); //validates $valCopy.
 468                                 if ( $valCopy !== $val ) {
 469                                         wfSuppressWarnings();
 470                                         $val = iconv('Windows-1252', 'UTF-8//IGNORE', $val);
 471                                         wfRestoreWarnings();
 472                                 }
 473                         }
 474
 475                         //trim and check to make sure not only whitespace.
 476                         $val = trim($val);
 477                         if ( strlen( $val ) === 0 ) {
 478                                 //only whitespace.
 479                                 $this->debug( $this->mFilteredExifData[$prop] , __FUNCTION__, "$prop: Is only whitespace" );
 480                                 unset($this->mFilteredExifData[$prop]);
 481                                 return;
 482                         }
 483
 484                         //all's good.
 485                         $this->mFilteredExifData[$prop] = $val;
 486                 }
 487         }
 488         /**
 489         * Convert an Exif::UNDEFINED from a raw binary string
 490         * to its value. This is sometimes needed depending on
 491         * the type of UNDEFINED field
 492         * @param $prop String name of property
 493         */
 494         private function exifPropToOrd ( $prop ) {
 495                 if ( isset( $this->mFilteredExifData[$prop] ) ) {
 496                         $this->mFilteredExifData[$prop] = ord( $this->mFilteredExifData[$prop] );
 497                 }
 498         }
 499         /**
 500         * Convert gps in exif form to a single floating point number
 501         * for example 10 degress 20`40`` S -> -10.34444
 502         * @param String $prop a gps coordinate exif tag name (like GPSLongitude)
 503         */
 504         private function exifGPStoNumber ( $prop ) {
 505                 $loc =& $this->mFilteredExifData[$prop];
 506                 $dir =& $this->mFilteredExifData[$prop . 'Ref'];
 507                 $res = false;
 508
 509                 if ( isset( $loc ) && isset( $dir ) && ( $dir === 'N' || $dir === 'S' || $dir === 'E' || $dir === 'W' ) ) {
 510                         list( $num, $denom ) = explode( '/', $loc[0] );
 511                         $res = $num / $denom;
 512                         list( $num, $denom ) = explode( '/', $loc[1] );
 513                         $res += ( $num / $denom ) * ( 1 / 60 );
 514                         list( $num, $denom ) = explode( '/', $loc[2] );
 515                         $res += ( $num / $denom ) * ( 1 / 3600 );
 516
 517                         if ( $dir === 'S' || $dir === 'W' ) {
 518                                 $res *= - 1; // make negative
 519                         }
 520                 }
 521
 522                 // update the exif records.
 523
 524                 if ( $res !== false ) { // using !== as $res could potentially be 0
 525                         $this->mFilteredExifData[$prop] = $res;
 526                         unset( $this->mFilteredExifData[$prop . 'Ref'] );
 527                 } else { // if invalid
 528                         unset( $this->mFilteredExifData[$prop] );
 529                         unset( $this->mFilteredExifData[$prop . 'Ref'] );
 530                 }
 531         }
 532
 533         /**
 534          * Use FormatMetadata to create formatted values for display to user
 535          * (is this ever used?)
 536          *
 537          * @deprecated since 1.18
 538          */
 539         function makeFormattedData( ) {
 540                 wfDeprecated( __METHOD__ );
 541                 $this->mFormattedExifData = FormatMetadata::getFormattedData(
 542                         $this->mFilteredExifData );
 543         }
 544         /**#@-*/
 545
 546         /**#@+
 547          * @return array
 548          */
 549         /**
 550          * Get $this->mRawExifData
 551          */
 552         function getData() {
 553                 return $this->mRawExifData;
 554         }
 555
 556         /**
 557          * Get $this->mFilteredExifData
 558          */
 559         function getFilteredData() {
 560                 return $this->mFilteredExifData;
 561         }
 562
 563         /**
 564          * Get $this->mFormattedExifData
 565          *
 566          * This returns the data for display to user.
 567          * Its unclear if this is ever used.
 568          *
 569          * @deprecated since 1.18
 570          */
 571         function getFormattedData() {
 572                 wfDeprecated( __METHOD__ );
 573                 if (!$this->mFormattedExifData) {
 574                         $this->makeFormattedData();
 575                 }
 576                 return $this->mFormattedExifData;
 577         }
 578         /**#@-*/
 579
 580         /**
 581          * The version of the output format
 582          *
 583          * Before the actual metadata information is saved in the database we
 584          * strip some of it since we don't want to save things like thumbnails
 585          * which usually accompany Exif data. This value gets saved in the
 586          * database along with the actual Exif data, and if the version in the
 587          * database doesn't equal the value returned by this function the Exif
 588          * data is regenerated.
 589          *
 590          * @return int
 591          */
 592         public static function version() {
 593                 return 2; // We don't need no bloddy constants!
 594         }
 595
 596         /**#@+
 597          * Validates if a tag value is of the type it should be according to the Exif spec
 598          *
 599          * @private
 600          *
 601          * @param $in Mixed: the input value to check
 602          * @return bool
 603          */
 604         private function isByte( $in ) {
 605                 if ( !is_array( $in ) && sprintf('%d', $in) == $in && $in >= 0 && $in <= 255 ) {
 606                         $this->debug( $in, __FUNCTION__, true );
 607                         return true;
 608                 } else {
 609                         $this->debug( $in, __FUNCTION__, false );
 610                         return false;
 611                 }
 612         }
 613
 614         /**
 615          * @param $in
 616          * @return bool
 617          */
 618         private function isASCII( $in ) {
 619                 if ( is_array( $in ) ) {
 620                         return false;
 621                 }
 622
 623                 if ( preg_match( "/[^\x0a\x20-\x7e]/", $in ) ) {
 624                         $this->debug( $in, __FUNCTION__, 'found a character not in our whitelist' );
 625                         return false;
 626                 }
 627
 628                 if ( preg_match( '/^\s*$/', $in ) ) {
 629                         $this->debug( $in, __FUNCTION__, 'input consisted solely of whitespace' );
 630                         return false;
 631                 }
 632
 633                 return true;
 634         }
 635
 636         /**
 637          * @param $in
 638          * @return bool
 639          */
 640         private function isShort( $in ) {
 641                 if ( !is_array( $in ) && sprintf('%d', $in) == $in && $in >= 0 && $in <= 65536 ) {
 642                         $this->debug( $in, __FUNCTION__, true );
 643                         return true;
 644                 } else {
 645                         $this->debug( $in, __FUNCTION__, false );
 646                         return false;
 647                 }
 648         }
 649
 650         /**
 651          * @param $in
 652          * @return bool
 653          */
 654         private function isLong( $in ) {
 655                 if ( !is_array( $in ) && sprintf('%d', $in) == $in && $in >= 0 && $in <= 4294967296 ) {
 656                         $this->debug( $in, __FUNCTION__, true );
 657                         return true;
 658                 } else {
 659                         $this->debug( $in, __FUNCTION__, false );
 660                         return false;
 661                 }
 662         }
 663
 664         /**
 665          * @param $in
 666          * @return bool
 667          */
 668         private function isRational( $in ) {
 669                 $m = array();
 670                 if ( !is_array( $in ) && @preg_match( '/^(\d+)\/(\d+[1-9]|[1-9]\d*)$/', $in, $m ) ) { # Avoid division by zero
 671                         return $this->isLong( $m[1] ) && $this->isLong( $m[2] );
 672                 } else {
 673                         $this->debug( $in, __FUNCTION__, 'fed a non-fraction value' );
 674                         return false;
 675                 }
 676         }
 677
 678         /**
 679          * @param $in
 680          * @return bool
 681          */
 682         private function isUndefined( $in ) {
 683                 $this->debug( $in, __FUNCTION__, true );
 684                 return true;
 685         }
 686
 687         /**
 688          * @param $in
 689          * @return bool
 690          */
 691         private function isSlong( $in ) {
 692                 if ( $this->isLong( abs( $in ) ) ) {
 693                         $this->debug( $in, __FUNCTION__, true );
 694                         return true;
 695                 } else {
 696                         $this->debug( $in, __FUNCTION__, false );
 697                         return false;
 698                 }
 699         }
 700
 701         /**
 702          * @param $in
 703          * @return bool
 704          */
 705         private function isSrational( $in ) {
 706                 $m = array();
 707                 if ( !is_array( $in ) && preg_match( '/^(-?\d+)\/(\d+[1-9]|[1-9]\d*)$/', $in, $m ) ) { # Avoid division by zero
 708                         return $this->isSlong( $m[0] ) && $this->isSlong( $m[1] );
 709                 } else {
 710                         $this->debug( $in, __FUNCTION__, 'fed a non-fraction value' );
 711                         return false;
 712                 }
 713         }
 714         /**#@-*/
 715
 716         /**
 717          * Validates if a tag has a legal value according to the Exif spec
 718          *
 719          * @private
 720          * @param $section String: section where tag is located.
 721          * @param $tag String: the tag to check.
 722          * @param $val Mixed: the value of the tag.
 723          * @param $recursive Boolean: true if called recursively for array types.
 724          * @return bool
 725          */
 726         private function validate( $section, $tag, $val, $recursive = false ) {
 727                 $debug = "tag is '$tag'";
 728                 $etype = $this->mExifTags[$section][$tag];
 729                 $ecount = 1;
 730                 if( is_array( $etype ) ) {
 731                         list( $etype, $ecount ) = $etype;
 732                         if ( $recursive )
 733                                 $ecount = 1; // checking individual elements
 734                 }
 735                 $count = count( $val );
 736                 if( $ecount != $count ) {
 737                         $this->debug( $val, __FUNCTION__, "Expected $ecount elements for $tag but got $count" );
 738                         return false;
 739                 }
 740                 if( $count > 1 ) {
 741                         foreach( $val as $v ) {
 742                                 if( !$this->validate( $section, $tag, $v, true ) ) {
 743                                         return false;
 744                                 }
 745                         }
 746                         return true;
 747                 }
 748                 // Does not work if not typecast
 749                 switch( (string)$etype ) {
 750                         case (string)Exif::BYTE:
 751                                 $this->debug( $val, __FUNCTION__, $debug );
 752                                 return $this->isByte( $val );
 753                         case (string)Exif::ASCII:
 754                                 $this->debug( $val, __FUNCTION__, $debug );
 755                                 return $this->isASCII( $val );
 756                         case (string)Exif::SHORT:
 757                                 $this->debug( $val, __FUNCTION__, $debug );
 758                                 return $this->isShort( $val );
 759                         case (string)Exif::LONG:
 760                                 $this->debug( $val, __FUNCTION__, $debug );
 761                                 return $this->isLong( $val );
 762                         case (string)Exif::RATIONAL:
 763                                 $this->debug( $val, __FUNCTION__, $debug );
 764                                 return $this->isRational( $val );
 765                         case (string)Exif::UNDEFINED:
 766                                 $this->debug( $val, __FUNCTION__, $debug );
 767                                 return $this->isUndefined( $val );
 768                         case (string)Exif::SLONG:
 769                                 $this->debug( $val, __FUNCTION__, $debug );
 770                                 return $this->isSlong( $val );
 771                         case (string)Exif::SRATIONAL:
 772                                 $this->debug( $val, __FUNCTION__, $debug );
 773                                 return $this->isSrational( $val );
 774                         case (string)Exif::SHORT.','.Exif::LONG:
 775                                 $this->debug( $val, __FUNCTION__, $debug );
 776                                 return $this->isShort( $val ) || $this->isLong( $val );
 777                         case (string)Exif::IGNORE:
 778                                 $this->debug( $val, __FUNCTION__, $debug );
 779                                 return false;
 780                         default:
 781                                 $this->debug( $val, __FUNCTION__, "The tag '$tag' is unknown" );
 782                                 return false;
 783                 }
 784         }
 785
 786         /**
 787          * Convenience function for debugging output
 788          *
 789          * @private
 790          *
 791          * @param $in Mixed:
 792          * @param $fname String:
 793          * @param $action Mixed: , default NULL.
 794          */
 795         private function debug( $in, $fname, $action = null ) {
 796                 if ( !$this->log ) {
 797                         return;
 798                 }
 799                 $type = gettype( $in );
 800                 $class = ucfirst( __CLASS__ );
 801                 if ( $type === 'array' ) {
 802                         $in = print_r( $in, true );
 803                 }
 804
 805                 if ( $action === true ) {
 806                         wfDebugLog( $this->log, "$class::$fname: accepted: '$in' (type: $type)\n");
 807                 } elseif ( $action === false ) {
 808                         wfDebugLog( $this->log, "$class::$fname: rejected: '$in' (type: $type)\n");
 809                 } elseif ( $action === null ) {
 810                         wfDebugLog( $this->log, "$class::$fname: input was: '$in' (type: $type)\n");
 811                 } else {
 812                         wfDebugLog( $this->log, "$class::$fname: $action (type: $type; content: '$in')\n");
 813                 }
 814         }
 815
 816         /**
 817          * Convenience function for debugging output
 818          *
 819          * @private
 820          *
 821          * @param $fname String: the name of the function calling this function
 822          * @param $io Boolean: Specify whether we're beginning or ending
 823          */
 824         private function debugFile( $fname, $io ) {
 825                 if ( !$this->log ) {
 826                         return;
 827                 }
 828                 $class = ucfirst( __CLASS__ );
 829                 if ( $io ) {
 830                         wfDebugLog( $this->log, "$class::$fname: begin processing: '{$this->basename}'\n" );
 831                 } else {
 832                         wfDebugLog( $this->log, "$class::$fname: end processing: '{$this->basename}'\n" );
 833                 }
 834         }
 835 }
 836