includes/json/FormatJson.php

   1 <?php
   2 /**
   3  * Wrapper for json_encode and json_decode.
   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  */
  22
  23 /**
  24  * JSON formatter wrapper class
  25  */
  26 class FormatJson {
  27         /**
  28          * Skip escaping most characters above U+007F for readability and compactness.
  29          * This encoding option saves 3 to 8 bytes (uncompressed) for each such character;
  30          * however, it could break compatibility with systems that incorrectly handle UTF-8.
  31          *
  32          * @since 1.22
  33          */
  34         const UTF8_OK = 1;
  35
  36         /**
  37          * Skip escaping the characters '<', '>', and '&', which have special meanings in
  38          * HTML and XML.
  39          *
  40          * @warning Do not use this option for JSON that could end up in inline scripts.
  41          * - HTML5, §4.3.1.2 Restrictions for contents of script elements
  42          * - XML 1.0 (5th Ed.), §2.4 Character Data and Markup
  43          *
  44          * @since 1.22
  45          */
  46         const XMLMETA_OK = 2;
  47
  48         /**
  49          * Skip escaping as many characters as reasonably possible.
  50          *
  51          * @warning When generating inline script blocks, use FormatJson::UTF8_OK instead.
  52          *
  53          * @since 1.22
  54          */
  55         const ALL_OK = 3;
  56
  57         /**
  58          * If set, treat json objects '{...}' as associative arrays. Without this option,
  59          * json objects will be converted to stdClass.
  60          * The value is set to 1 to be backward compatible with 'true' that was used before.
  61          *
  62          * @since 1.24
  63          */
  64         const FORCE_ASSOC = 0x100;
  65
  66         /**
  67          * If set, attempts to fix invalid json.
  68          *
  69          * @since 1.24
  70          */
  71         const TRY_FIXING = 0x200;
  72
  73         /**
  74          * If set, strip comments from input before parsing as JSON.
  75          *
  76          * @since 1.25
  77          */
  78         const STRIP_COMMENTS = 0x400;
  79
  80         /**
  81          * Regex that matches whitespace inside empty arrays and objects.
  82          *
  83          * This doesn't affect regular strings inside the JSON because those can't
  84          * have a real line break (\n) in them, at this point they are already escaped
  85          * as the string "\n" which this doesn't match.
  86          *
  87          * @private
  88          */
  89         const WS_CLEANUP_REGEX = '/(?<=[\[{])\n\s*+(?=[\]}])/';
  90
  91         /**
  92          * Characters problematic in JavaScript.
  93          *
  94          * @note These are listed in ECMA-262 (5.1 Ed.), §7.3 Line Terminators along with U+000A (LF)
  95          *       and U+000D (CR). However, PHP already escapes LF and CR according to RFC 4627.
  96          */
  97         private static $badChars = array(
  98                 "\xe2\x80\xa8", // U+2028 LINE SEPARATOR
  99                 "\xe2\x80\xa9", // U+2029 PARAGRAPH SEPARATOR
 100         );
 101
 102         /**
 103          * Escape sequences for characters listed in FormatJson::$badChars.
 104          */
 105         private static $badCharsEscaped = array(
 106                 '\u2028', // U+2028 LINE SEPARATOR
 107                 '\u2029', // U+2029 PARAGRAPH SEPARATOR
 108         );
 109
 110         /**
 111          * Returns the JSON representation of a value.
 112          *
 113          * @note Empty arrays are encoded as numeric arrays, not as objects, so cast any associative
 114          *       array that might be empty to an object before encoding it.
 115          *
 116          * @note In pre-1.22 versions of MediaWiki, using this function for generating inline script
 117          *       blocks may result in an XSS vulnerability, and quite likely will in XML documents
 118          *       (cf. FormatJson::XMLMETA_OK). Use Xml::encodeJsVar() instead in such cases.
 119          *
 120          * @param mixed $value The value to encode. Can be any type except a resource.
 121          * @param string|bool $pretty If a string, add non-significant whitespace to improve
 122          *   readability, using that string for indentation. If true, use the default indent
 123          *   string (four spaces).
 124          * @param int $escaping Bitfield consisting of _OK class constants
 125          * @return string|false String if successful; false upon failure
 126          */
 127         public static function encode( $value, $pretty = false, $escaping = 0 ) {
 128                 if ( !is_string( $pretty ) ) {
 129                         $pretty = $pretty ? '    ' : false;
 130                 }
 131
 132                 if ( defined( 'JSON_UNESCAPED_UNICODE' ) ) {
 133                         return self::encode54( $value, $pretty, $escaping );
 134                 }
 135
 136                 return self::encode53( $value, $pretty, $escaping );
 137         }
 138
 139         /**
 140          * Decodes a JSON string. It is recommended to use FormatJson::parse(),
 141          * which returns more comprehensive result in case of an error, and has
 142          * more parsing options.
 143          *
 144          * @param string $value The JSON string being decoded
 145          * @param bool $assoc When true, returned objects will be converted into associative arrays.
 146          *
 147          * @return mixed The value encoded in JSON in appropriate PHP type.
 148          * `null` is returned if $value represented `null`, if $value could not be decoded,
 149          * or if the encoded data was deeper than the recursion limit.
 150          * Use FormatJson::parse() to distinguish between types of `null` and to get proper error code.
 151          */
 152         public static function decode( $value, $assoc = false ) {
 153                 return json_decode( $value, $assoc );
 154         }
 155
 156         /**
 157          * Decodes a JSON string.
 158          * Unlike FormatJson::decode(), if $value represents null value, it will be
 159          * properly decoded as valid.
 160          *
 161          * @param string $value The JSON string being decoded
 162          * @param int $options A bit field that allows FORCE_ASSOC, TRY_FIXING,
 163          * STRIP_COMMENTS
 164          * @return Status If valid JSON, the value is available in $result->getValue()
 165          */
 166         public static function parse( $value, $options = 0 ) {
 167                 if ( $options & self::STRIP_COMMENTS ) {
 168                         $value = self::stripComments( $value );
 169                 }
 170                 $assoc = ( $options & self::FORCE_ASSOC ) !== 0;
 171                 $result = json_decode( $value, $assoc );
 172                 $code = json_last_error();
 173
 174                 if ( $code === JSON_ERROR_SYNTAX && ( $options & self::TRY_FIXING ) !== 0 ) {
 175                         // The most common error is the trailing comma in a list or an object.
 176                         // We cannot simply replace /,\s*[}\]]/ because it could be inside a string value.
 177                         // But we could use the fact that JSON does not allow multi-line string values,
 178                         // And remove trailing commas if they are et the end of a line.
 179                         // JSON only allows 4 control characters: [ \t\r\n].  So we must not use '\s' for matching.
 180                         // Regex match   ,]<any non-quote chars>\n   or   ,\n]   with optional spaces/tabs.
 181                         $count = 0;
 182                         $value =
 183                                 preg_replace( '/,([ \t]*[}\]][^"\r\n]*([\r\n]|$)|[ \t]*[\r\n][ \t\r\n]*[}\]])/', '$1',
 184                                         $value, - 1, $count );
 185                         if ( $count > 0 ) {
 186                                 $result = json_decode( $value, $assoc );
 187                                 if ( JSON_ERROR_NONE === json_last_error() ) {
 188                                         // Report warning
 189                                         $st = Status::newGood( $result );
 190                                         $st->warning( wfMessage( 'json-warn-trailing-comma' )->numParams( $count ) );
 191                                         return $st;
 192                                 }
 193                         }
 194                 }
 195
 196                 switch ( $code ) {
 197                         case JSON_ERROR_NONE:
 198                                 return Status::newGood( $result );
 199                         default:
 200                                 return Status::newFatal( wfMessage( 'json-error-unknown' )->numParams( $code ) );
 201                         case JSON_ERROR_DEPTH:
 202                                 $msg = 'json-error-depth';
 203                                 break;
 204                         case JSON_ERROR_STATE_MISMATCH:
 205                                 $msg = 'json-error-state-mismatch';
 206                                 break;
 207                         case JSON_ERROR_CTRL_CHAR:
 208                                 $msg = 'json-error-ctrl-char';
 209                                 break;
 210                         case JSON_ERROR_SYNTAX:
 211                                 $msg = 'json-error-syntax';
 212                                 break;
 213                         case JSON_ERROR_UTF8:
 214                                 $msg = 'json-error-utf8';
 215                                 break;
 216                         case JSON_ERROR_RECURSION:
 217                                 $msg = 'json-error-recursion';
 218                                 break;
 219                         case JSON_ERROR_INF_OR_NAN:
 220                                 $msg = 'json-error-inf-or-nan';
 221                                 break;
 222                         case JSON_ERROR_UNSUPPORTED_TYPE:
 223                                 $msg = 'json-error-unsupported-type';
 224                                 break;
 225                 }
 226                 return Status::newFatal( $msg );
 227         }
 228
 229         /**
 230          * JSON encoder wrapper for PHP >= 5.4, which supports useful encoding options.
 231          *
 232          * @param mixed $value
 233          * @param string|bool $pretty
 234          * @param int $escaping
 235          * @return string|false
 236          */
 237         private static function encode54( $value, $pretty, $escaping ) {
 238                 static $bug66021;
 239                 if ( $pretty !== false && $bug66021 === null ) {
 240                         $bug66021 = json_encode( array(), JSON_PRETTY_PRINT ) !== '[]';
 241                 }
 242
 243                 // PHP escapes '/' to prevent breaking out of inline script blocks using '</script>',
 244                 // which is hardly useful when '<' and '>' are escaped (and inadequate), and such
 245                 // escaping negatively impacts the human readability of URLs and similar strings.
 246                 $options = JSON_UNESCAPED_SLASHES;
 247                 $options |= $pretty !== false ? JSON_PRETTY_PRINT : 0;
 248                 $options |= ( $escaping & self::UTF8_OK ) ? JSON_UNESCAPED_UNICODE : 0;
 249                 $options |= ( $escaping & self::XMLMETA_OK ) ? 0 : ( JSON_HEX_TAG | JSON_HEX_AMP );
 250                 $json = json_encode( $value, $options );
 251                 if ( $json === false ) {
 252                         return false;
 253                 }
 254
 255                 if ( $pretty !== false ) {
 256                         // Workaround for <https://bugs.php.net/bug.php?id=66021>
 257                         if ( $bug66021 ) {
 258                                 $json = preg_replace( self::WS_CLEANUP_REGEX, '', $json );
 259                         }
 260                         if ( $pretty !== '    ' ) {
 261                                 // Change the four-space indent to a tab indent
 262                                 $json = str_replace( "\n    ", "\n\t", $json );
 263                                 while ( strpos( $json, "\t    " ) !== false ) {
 264                                         $json = str_replace( "\t    ", "\t\t", $json );
 265                                 }
 266
 267                                 if ( $pretty !== "\t" ) {
 268                                         // Change the tab indent to the provided indent
 269                                         $json = str_replace( "\t", $pretty, $json );
 270                                 }
 271                         }
 272                 }
 273                 if ( $escaping & self::UTF8_OK ) {
 274                         $json = str_replace( self::$badChars, self::$badCharsEscaped, $json );
 275                 }
 276
 277                 return $json;
 278         }
 279
 280         /**
 281          * JSON encoder wrapper for PHP 5.3, which lacks native support for some encoding options.
 282          * Therefore, the missing options are implemented here purely in PHP code.
 283          *
 284          * @param mixed $value
 285          * @param string|bool $pretty
 286          * @param int $escaping
 287          * @return string|false
 288          */
 289         private static function encode53( $value, $pretty, $escaping ) {
 290                 $options = ( $escaping & self::XMLMETA_OK ) ? 0 : ( JSON_HEX_TAG | JSON_HEX_AMP );
 291                 $json = json_encode( $value, $options );
 292                 if ( $json === false ) {
 293                         return false;
 294                 }
 295
 296                 // Emulate JSON_UNESCAPED_SLASHES. Because the JSON contains no unescaped slashes
 297                 // (only escaped slashes), a simple string replacement works fine.
 298                 $json = str_replace( '\/', '/', $json );
 299
 300                 if ( $escaping & self::UTF8_OK ) {
 301                         // JSON hex escape sequences follow the format \uDDDD, where DDDD is four hex digits
 302                         // indicating the equivalent UTF-16 code unit's value. To most efficiently unescape
 303                         // them, we exploit the JSON extension's built-in decoder.
 304                         // * We escape the input a second time, so any such sequence becomes \\uDDDD.
 305                         // * To avoid interpreting escape sequences that were in the original input,
 306                         //   each double-escaped backslash (\\\\) is replaced with \\\u005c.
 307                         // * We strip one of the backslashes from each of the escape sequences to unescape.
 308                         // * Then the JSON decoder can perform the actual unescaping.
 309                         $json = str_replace( "\\\\\\\\", "\\\\\\u005c", addcslashes( $json, '\"' ) );
 310                         $json = json_decode( preg_replace( "/\\\\\\\\u(?!00[0-7])/", "\\\\u", "\"$json\"" ) );
 311                         $json = str_replace( self::$badChars, self::$badCharsEscaped, $json );
 312                 }
 313
 314                 if ( $pretty !== false ) {
 315                         return self::prettyPrint( $json, $pretty );
 316                 }
 317
 318                 return $json;
 319         }
 320
 321         /**
 322          * Adds non-significant whitespace to an existing JSON representation of an object.
 323          * Only needed for PHP < 5.4, which lacks the JSON_PRETTY_PRINT option.
 324          *
 325          * @param string $json
 326          * @param string $indentString
 327          * @return string
 328          */
 329         private static function prettyPrint( $json, $indentString ) {
 330                 $buf = '';
 331                 $indent = 0;
 332                 $json = strtr( $json, array( '\\\\' => '\\\\', '\"' => "\x01" ) );
 333                 for ( $i = 0, $n = strlen( $json ); $i < $n; $i += $skip ) {
 334                         $skip = 1;
 335                         switch ( $json[$i] ) {
 336                                 case ':':
 337                                         $buf .= ': ';
 338                                         break;
 339                                 case '[':
 340                                 case '{':
 341                                         ++$indent;
 342                                         // falls through
 343                                 case ',':
 344                                         $buf .= $json[$i] . "\n" . str_repeat( $indentString, $indent );
 345                                         break;
 346                                 case ']':
 347                                 case '}':
 348                                         $buf .= "\n" . str_repeat( $indentString, --$indent ) . $json[$i];
 349                                         break;
 350                                 case '"':
 351                                         $skip = strcspn( $json, '"', $i + 1 ) + 2;
 352                                         $buf .= substr( $json, $i, $skip );
 353                                         break;
 354                                 default:
 355                                         $skip = strcspn( $json, ',]}"', $i + 1 ) + 1;
 356                                         $buf .= substr( $json, $i, $skip );
 357                         }
 358                 }
 359                 $buf = preg_replace( self::WS_CLEANUP_REGEX, '', $buf );
 360
 361                 return str_replace( "\x01", '\"', $buf );
 362         }
 363
 364         /**
 365          * Remove multiline and single line comments from an otherwise valid JSON
 366          * input string. This can be used as a preprocessor for to allow JSON
 367          * formatted configuration files to contain comments.
 368          *
 369          * @param string $json
 370          * @return string JSON with comments removed
 371          */
 372         public static function stripComments( $json ) {
 373                 // Ensure we have a string
 374                 $str = (string) $json;
 375                 $buffer = '';
 376                 $maxLen = strlen( $str );
 377                 $mark = 0;
 378
 379                 $inString = false;
 380                 $inComment = false;
 381                 $multiline = false;
 382
 383                 for ($idx = 0; $idx < $maxLen; $idx++) {
 384                         switch ( $str[$idx] ) {
 385                                 case '"':
 386                                         $lookBehind = ( $idx - 1 >= 0 ) ? $str[$idx - 1] : '';
 387                                         if ( !$inComment && $lookBehind !== '\\' ) {
 388                                                 // Either started or ended a string
 389                                                 $inString = !$inString;
 390                                         }
 391                                         break;
 392
 393                                 case '/':
 394                                         $lookAhead = ( $idx + 1 < $maxLen ) ? $str[$idx + 1] : '';
 395                                         $lookBehind = ( $idx - 1 >= 0 ) ? $str[$idx - 1] : '';
 396                                         if ( $inString ) {
 397                                                 continue;
 398
 399                                         } elseif ( !$inComment &&
 400                                                 ( $lookAhead === '/' || $lookAhead === '*' )
 401                                         ) {
 402                                                 // Transition into a comment
 403                                                 // Add characters seen to buffer
 404                                                 $buffer .= substr( $str, $mark, $idx - $mark );
 405                                                 // Consume the look ahead character
 406                                                 $idx++;
 407                                                 // Track state
 408                                                 $inComment = true;
 409                                                 $multiline = $lookAhead === '*';
 410
 411                                         } elseif ( $multiline && $lookBehind === '*' ) {
 412                                                 // Found the end of the current comment
 413                                                 $mark = $idx + 1;
 414                                                 $inComment = false;
 415                                                 $multiline = false;
 416                                         }
 417                                         break;
 418
 419                                 case "\n":
 420                                         if ( $inComment && !$multiline ) {
 421                                                 // Found the end of the current comment
 422                                                 $mark = $idx + 1;
 423                                                 $inComment = false;
 424                                         }
 425                                         break;
 426                         }
 427                 }
 428                 if ( $inComment ) {
 429                         // Comment ends with input
 430                         // Technically we should check to ensure that we aren't in
 431                         // a multiline comment that hasn't been properly ended, but this
 432                         // is a strip filter, not a validating parser.
 433                         $mark = $maxLen;
 434                 }
 435                 // Add final chunk to buffer before returning
 436                 return $buffer . substr( $str, $mark, $maxLen - $mark );
 437         }
 438 }