Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
| Total | |
63.06% |
99 / 157 |
|
0.00% |
0 / 5 |
CRAP | |
0.00% |
0 / 1 |
| GIFMetadataExtractor | |
63.06% |
99 / 157 |
|
0.00% |
0 / 5 |
170.05 | |
0.00% |
0 / 1 |
| getMetadata | |
79.20% |
99 / 125 |
|
0.00% |
0 / 1 |
47.66 | |||
| readGCT | |
0.00% |
0 / 3 |
|
0.00% |
0 / 1 |
6 | |||
| decodeBPP | |
0.00% |
0 / 7 |
|
0.00% |
0 / 1 |
6 | |||
| skipBlock | |
0.00% |
0 / 8 |
|
0.00% |
0 / 1 |
20 | |||
| readBlock | |
0.00% |
0 / 14 |
|
0.00% |
0 / 1 |
30 | |||
| 1 | <?php |
| 2 | /** |
| 3 | * GIF frame counter. |
| 4 | * |
| 5 | * Originally written in Perl by Steve Sanbeg. |
| 6 | * Ported to PHP by Andrew Garrett |
| 7 | * Deliberately not using MWExceptions to avoid external dependencies, encouraging |
| 8 | * redistribution. |
| 9 | * |
| 10 | * This program is free software; you can redistribute it and/or modify |
| 11 | * it under the terms of the GNU General Public License as published by |
| 12 | * the Free Software Foundation; either version 2 of the License, or |
| 13 | * (at your option) any later version. |
| 14 | * |
| 15 | * This program is distributed in the hope that it will be useful, |
| 16 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 17 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 18 | * GNU General Public License for more details. |
| 19 | * |
| 20 | * You should have received a copy of the GNU General Public License along |
| 21 | * with this program; if not, write to the Free Software Foundation, Inc., |
| 22 | * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. |
| 23 | * http://www.gnu.org/copyleft/gpl.html |
| 24 | * |
| 25 | * @file |
| 26 | * @ingroup Media |
| 27 | */ |
| 28 | |
| 29 | use Wikimedia\AtEase\AtEase; |
| 30 | |
| 31 | /** |
| 32 | * GIF frame counter. |
| 33 | * |
| 34 | * @ingroup Media |
| 35 | */ |
| 36 | class GIFMetadataExtractor { |
| 37 | /** @var string */ |
| 38 | private static $gifFrameSep; |
| 39 | |
| 40 | /** @var string */ |
| 41 | private static $gifExtensionSep; |
| 42 | |
| 43 | /** @var string */ |
| 44 | private static $gifTerm; |
| 45 | |
| 46 | public const VERSION = 1; |
| 47 | |
| 48 | // Each sub-block is less than or equal to 255 bytes. |
| 49 | // Most of the time its 255 bytes, except for in XMP |
| 50 | // blocks, where it's usually between 32-127 bytes each. |
| 51 | private const MAX_SUBBLOCKS = 262144; // 5 MiB divided by 20. |
| 52 | |
| 53 | /** |
| 54 | * @throws Exception |
| 55 | * @param string $filename |
| 56 | * @return array |
| 57 | */ |
| 58 | public static function getMetadata( $filename ) { |
| 59 | self::$gifFrameSep = pack( "C", ord( "," ) ); // 2C |
| 60 | self::$gifExtensionSep = pack( "C", ord( "!" ) ); // 21 |
| 61 | self::$gifTerm = pack( "C", ord( ";" ) ); // 3B |
| 62 | |
| 63 | $frameCount = 0; |
| 64 | $duration = 0.0; |
| 65 | $isLooped = false; |
| 66 | $xmp = ""; |
| 67 | $comment = []; |
| 68 | |
| 69 | if ( !$filename ) { |
| 70 | throw new InvalidArgumentException( 'No file name specified' ); |
| 71 | } |
| 72 | if ( !file_exists( $filename ) || is_dir( $filename ) ) { |
| 73 | throw new InvalidArgumentException( "File $filename does not exist" ); |
| 74 | } |
| 75 | |
| 76 | $fh = fopen( $filename, 'rb' ); |
| 77 | |
| 78 | if ( !$fh ) { |
| 79 | throw new InvalidArgumentException( "Unable to open file $filename" ); |
| 80 | } |
| 81 | |
| 82 | // Check for the GIF header |
| 83 | $buf = fread( $fh, 6 ); |
| 84 | if ( !( $buf === 'GIF87a' || $buf === 'GIF89a' ) ) { |
| 85 | throw new InvalidArgumentException( "Not a valid GIF file; header: $buf" ); |
| 86 | } |
| 87 | |
| 88 | // Read width and height. |
| 89 | $buf = fread( $fh, 2 ); |
| 90 | if ( strlen( $buf ) < 2 ) { |
| 91 | throw new InvalidArgumentException( "Not a valid GIF file; Unable to read width." ); |
| 92 | } |
| 93 | $width = unpack( 'v', $buf )[1]; |
| 94 | $buf = fread( $fh, 2 ); |
| 95 | if ( strlen( $buf ) < 2 ) { |
| 96 | throw new InvalidArgumentException( "Not a valid GIF file; Unable to read height." ); |
| 97 | } |
| 98 | $height = unpack( 'v', $buf )[1]; |
| 99 | |
| 100 | // Read BPP |
| 101 | $buf = fread( $fh, 1 ); |
| 102 | [ $bpp, $have_map ] = self::decodeBPP( $buf ); |
| 103 | |
| 104 | // Skip over background and aspect ratio |
| 105 | // @phan-suppress-next-line PhanPluginUseReturnValueInternalKnown |
| 106 | fread( $fh, 2 ); |
| 107 | |
| 108 | // Skip over the GCT |
| 109 | if ( $have_map ) { |
| 110 | self::readGCT( $fh, $bpp ); |
| 111 | } |
| 112 | |
| 113 | while ( !feof( $fh ) ) { |
| 114 | $buf = fread( $fh, 1 ); |
| 115 | |
| 116 | if ( $buf === self::$gifFrameSep ) { |
| 117 | // Found a frame |
| 118 | $frameCount++; |
| 119 | |
| 120 | # # Skip bounding box |
| 121 | // @phan-suppress-next-line PhanPluginUseReturnValueInternalKnown |
| 122 | fread( $fh, 8 ); |
| 123 | |
| 124 | # # Read BPP |
| 125 | $buf = fread( $fh, 1 ); |
| 126 | [ $bpp, $have_map ] = self::decodeBPP( $buf ); |
| 127 | |
| 128 | # # Read GCT |
| 129 | if ( $have_map ) { |
| 130 | self::readGCT( $fh, $bpp ); |
| 131 | } |
| 132 | // @phan-suppress-next-line PhanPluginUseReturnValueInternalKnown |
| 133 | fread( $fh, 1 ); |
| 134 | self::skipBlock( $fh ); |
| 135 | } elseif ( $buf === self::$gifExtensionSep ) { |
| 136 | $buf = fread( $fh, 1 ); |
| 137 | if ( strlen( $buf ) < 1 ) { |
| 138 | throw new InvalidArgumentException( |
| 139 | "Not a valid GIF file; Unable to read graphics control extension." |
| 140 | ); |
| 141 | } |
| 142 | $extension_code = unpack( 'C', $buf )[1]; |
| 143 | |
| 144 | if ( $extension_code === 0xF9 ) { |
| 145 | // Graphics Control Extension. |
| 146 | // @phan-suppress-next-line PhanPluginUseReturnValueInternalKnown |
| 147 | fread( $fh, 1 ); // Block size |
| 148 | |
| 149 | // @phan-suppress-next-next-line PhanPluginUseReturnValueInternalKnown |
| 150 | // @phan-suppress-next-line PhanPluginDuplicateAdjacentStatement |
| 151 | fread( $fh, 1 ); // Transparency, disposal method, user input |
| 152 | |
| 153 | $buf = fread( $fh, 2 ); // Delay, in hundredths of seconds. |
| 154 | if ( strlen( $buf ) < 2 ) { |
| 155 | throw new InvalidArgumentException( "Not a valid GIF file; Unable to read delay" ); |
| 156 | } |
| 157 | $delay = unpack( 'v', $buf )[1]; |
| 158 | $duration += $delay * 0.01; |
| 159 | |
| 160 | // @phan-suppress-next-line PhanPluginUseReturnValueInternalKnown |
| 161 | fread( $fh, 1 ); // Transparent colour index |
| 162 | |
| 163 | $term = fread( $fh, 1 ); // Should be a terminator |
| 164 | if ( strlen( $term ) < 1 ) { |
| 165 | throw new InvalidArgumentException( "Not a valid GIF file; Unable to read terminator byte" ); |
| 166 | } |
| 167 | $term = unpack( 'C', $term )[1]; |
| 168 | if ( $term != 0 ) { |
| 169 | throw new InvalidArgumentException( "Malformed Graphics Control Extension block" ); |
| 170 | } |
| 171 | } elseif ( $extension_code === 0xFE ) { |
| 172 | // Comment block(s). |
| 173 | $data = self::readBlock( $fh ); |
| 174 | if ( $data === "" ) { |
| 175 | throw new InvalidArgumentException( 'Read error, zero-length comment block' ); |
| 176 | } |
| 177 | |
| 178 | // The standard says this should be ASCII, however its unclear if |
| 179 | // thats true in practise. Check to see if its valid utf-8, if so |
| 180 | // assume its that, otherwise assume its windows-1252 (iso-8859-1) |
| 181 | $dataCopy = $data; |
| 182 | // quickIsNFCVerify has the side effect of replacing any invalid characters |
| 183 | UtfNormal\Validator::quickIsNFCVerify( $dataCopy ); |
| 184 | |
| 185 | if ( $dataCopy !== $data ) { |
| 186 | AtEase::suppressWarnings(); |
| 187 | $data = iconv( 'windows-1252', 'UTF-8', $data ); |
| 188 | AtEase::restoreWarnings(); |
| 189 | } |
| 190 | |
| 191 | $commentCount = count( $comment ); |
| 192 | if ( $commentCount === 0 |
| 193 | // @phan-suppress-next-line PhanTypeInvalidDimOffset |
| 194 | || $comment[$commentCount - 1] !== $data |
| 195 | ) { |
| 196 | // Some applications repeat the same comment on each |
| 197 | // frame of an animated GIF image, so if this comment |
| 198 | // is identical to the last, only extract once. |
| 199 | $comment[] = $data; |
| 200 | } |
| 201 | } elseif ( $extension_code === 0xFF ) { |
| 202 | // Application extension (Netscape info about the animated gif) |
| 203 | // or XMP (or theoretically any other type of extension block) |
| 204 | $blockLength = fread( $fh, 1 ); |
| 205 | if ( strlen( $blockLength ) < 1 ) { |
| 206 | throw new InvalidArgumentException( "Not a valid GIF file; Unable to read block length" ); |
| 207 | } |
| 208 | $blockLength = unpack( 'C', $blockLength )[1]; |
| 209 | $data = fread( $fh, $blockLength ); |
| 210 | |
| 211 | if ( $blockLength !== 11 ) { |
| 212 | wfDebug( __METHOD__ . " GIF application block with wrong length" ); |
| 213 | fseek( $fh, -( $blockLength + 1 ), SEEK_CUR ); |
| 214 | self::skipBlock( $fh ); |
| 215 | continue; |
| 216 | } |
| 217 | |
| 218 | // NETSCAPE2.0 (application name for animated gif) |
| 219 | if ( $data === 'NETSCAPE2.0' ) { |
| 220 | $data = fread( $fh, 2 ); // Block length and introduction, should be 03 01 |
| 221 | |
| 222 | if ( $data !== "\x03\x01" ) { |
| 223 | throw new InvalidArgumentException( "Expected \x03\x01, got $data" ); |
| 224 | } |
| 225 | |
| 226 | // Unsigned little-endian integer, loop count or zero for "forever" |
| 227 | $loopData = fread( $fh, 2 ); |
| 228 | if ( strlen( $loopData ) < 2 ) { |
| 229 | throw new InvalidArgumentException( "Not a valid GIF file; Unable to read loop count" ); |
| 230 | } |
| 231 | $loopCount = unpack( 'v', $loopData )[1]; |
| 232 | |
| 233 | if ( $loopCount !== 1 ) { |
| 234 | $isLooped = true; |
| 235 | } |
| 236 | |
| 237 | // Read out terminator byte |
| 238 | // @phan-suppress-next-line PhanPluginUseReturnValueInternalKnown |
| 239 | fread( $fh, 1 ); |
| 240 | } elseif ( $data === 'XMP DataXMP' ) { |
| 241 | // application name for XMP data. |
| 242 | // see pg 18 of XMP spec part 3. |
| 243 | |
| 244 | $xmp = self::readBlock( $fh, true ); |
| 245 | |
| 246 | if ( substr( $xmp, -257, 3 ) !== "\x01\xFF\xFE" |
| 247 | || substr( $xmp, -4 ) !== "\x03\x02\x01\x00" |
| 248 | ) { |
| 249 | throw new InvalidArgumentException( "XMP does not have magic trailer!" ); |
| 250 | } |
| 251 | |
| 252 | // strip out trailer. |
| 253 | $xmp = substr( $xmp, 0, -257 ); |
| 254 | } else { |
| 255 | // unrecognized extension block |
| 256 | fseek( $fh, -( $blockLength + 1 ), SEEK_CUR ); |
| 257 | self::skipBlock( $fh ); |
| 258 | } |
| 259 | } else { |
| 260 | self::skipBlock( $fh ); |
| 261 | } |
| 262 | } elseif ( $buf === self::$gifTerm ) { |
| 263 | break; |
| 264 | } else { |
| 265 | if ( strlen( $buf ) < 1 ) { |
| 266 | throw new InvalidArgumentException( "Not a valid GIF file; Unable to read unknown byte." ); |
| 267 | } |
| 268 | $byte = unpack( 'C', $buf )[1]; |
| 269 | throw new InvalidArgumentException( "At position: " . ftell( $fh ) . ", Unknown byte " . $byte ); |
| 270 | } |
| 271 | } |
| 272 | |
| 273 | return [ |
| 274 | 'frameCount' => $frameCount, |
| 275 | 'looped' => $isLooped, |
| 276 | 'duration' => $duration, |
| 277 | 'xmp' => $xmp, |
| 278 | 'comment' => $comment, |
| 279 | 'width' => $width, |
| 280 | 'height' => $height, |
| 281 | 'bits' => $bpp, |
| 282 | ]; |
| 283 | } |
| 284 | |
| 285 | /** |
| 286 | * @param resource $fh |
| 287 | * @param int $bpp |
| 288 | * @return void |
| 289 | */ |
| 290 | private static function readGCT( $fh, $bpp ) { |
| 291 | $max = 2 ** $bpp; |
| 292 | for ( $i = 1; $i <= $max; ++$i ) { |
| 293 | // @phan-suppress-next-line PhanPluginUseReturnValueInternalKnown |
| 294 | fread( $fh, 3 ); |
| 295 | } |
| 296 | } |
| 297 | |
| 298 | /** |
| 299 | * @param string $data |
| 300 | * @throws Exception |
| 301 | * @return array [ int bits per channel, bool have GCT ] |
| 302 | */ |
| 303 | private static function decodeBPP( $data ) { |
| 304 | if ( strlen( $data ) < 1 ) { |
| 305 | throw new InvalidArgumentException( "Not a valid GIF file; Unable to read bits per channel." ); |
| 306 | } |
| 307 | $buf = unpack( 'C', $data )[1]; |
| 308 | $bpp = ( $buf & 7 ) + 1; |
| 309 | $buf >>= 7; |
| 310 | |
| 311 | $have_map = $buf & 1; |
| 312 | |
| 313 | return [ $bpp, $have_map ]; |
| 314 | } |
| 315 | |
| 316 | /** |
| 317 | * @param resource $fh |
| 318 | * @throws Exception |
| 319 | */ |
| 320 | private static function skipBlock( $fh ) { |
| 321 | while ( !feof( $fh ) ) { |
| 322 | $buf = fread( $fh, 1 ); |
| 323 | if ( strlen( $buf ) < 1 ) { |
| 324 | throw new InvalidArgumentException( "Not a valid GIF file; Unable to read block length." ); |
| 325 | } |
| 326 | $block_len = unpack( 'C', $buf )[1]; |
| 327 | if ( $block_len == 0 ) { |
| 328 | return; |
| 329 | } |
| 330 | // @phan-suppress-next-line PhanPluginUseReturnValueInternalKnown |
| 331 | fread( $fh, $block_len ); |
| 332 | } |
| 333 | } |
| 334 | |
| 335 | /** |
| 336 | * Read a block. In the GIF format, a block is made up of |
| 337 | * several sub-blocks. Each sub block starts with one byte |
| 338 | * saying how long the sub-block is, followed by the sub-block. |
| 339 | * The entire block is terminated by a sub-block of length |
| 340 | * 0. |
| 341 | * @param resource $fh File handle |
| 342 | * @param bool $includeLengths Include the length bytes of the |
| 343 | * sub-blocks in the returned value. Normally this is false, |
| 344 | * except XMP is weird and does a hack where you need to keep |
| 345 | * these length bytes. |
| 346 | * @throws Exception |
| 347 | * @return string The data. |
| 348 | */ |
| 349 | private static function readBlock( $fh, $includeLengths = false ) { |
| 350 | $data = ''; |
| 351 | $subLength = fread( $fh, 1 ); |
| 352 | $blocks = 0; |
| 353 | |
| 354 | while ( $subLength !== "\0" ) { |
| 355 | $blocks++; |
| 356 | if ( $blocks > self::MAX_SUBBLOCKS ) { |
| 357 | throw new InvalidArgumentException( "MAX_SUBBLOCKS exceeded (over $blocks sub-blocks)" ); |
| 358 | } |
| 359 | if ( feof( $fh ) ) { |
| 360 | throw new InvalidArgumentException( "Read error: Unexpected EOF." ); |
| 361 | } |
| 362 | if ( $includeLengths ) { |
| 363 | $data .= $subLength; |
| 364 | } |
| 365 | |
| 366 | $data .= fread( $fh, ord( $subLength ) ); |
| 367 | $subLength = fread( $fh, 1 ); |
| 368 | } |
| 369 | |
| 370 | return $data; |
| 371 | } |
| 372 | } |