Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
| Total | |
38.36% |
89 / 232 |
|
10.53% |
2 / 19 |
CRAP | |
0.00% |
0 / 1 |
| SyntaxHighlight | |
38.36% |
89 / 232 |
|
10.53% |
2 / 19 |
1502.73 | |
0.00% |
0 / 1 |
| getMaxLines | |
0.00% |
0 / 2 |
|
0.00% |
0 / 1 |
2 | |||
| getMaxBytes | |
100.00% |
2 / 2 |
|
100.00% |
1 / 1 |
1 | |||
| getLexer | |
46.15% |
6 / 13 |
|
0.00% |
0 / 1 |
8.90 | |||
| onParserFirstCallInit | |
0.00% |
0 / 2 |
|
0.00% |
0 / 1 |
2 | |||
| parserHookSource | |
0.00% |
0 / 2 |
|
0.00% |
0 / 1 |
2 | |||
| getModuleStyles | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| processContent | |
0.00% |
0 / 12 |
|
0.00% |
0 / 1 |
20 | |||
| parserHook | |
0.00% |
0 / 8 |
|
0.00% |
0 / 1 |
12 | |||
| sourceToDom | |
0.00% |
0 / 3 |
|
0.00% |
0 / 1 |
2 | |||
| unwrap | |
83.33% |
5 / 6 |
|
0.00% |
0 / 1 |
3.04 | |||
| plainCodeWrap | |
85.71% |
6 / 7 |
|
0.00% |
0 / 1 |
2.01 | |||
| highlightInner | |
64.06% |
41 / 64 |
|
0.00% |
0 / 1 |
44.46 | |||
| highlight | |
56.25% |
27 / 48 |
|
0.00% |
0 / 1 |
18.37 | |||
| makeCacheKeyHash | |
100.00% |
2 / 2 |
|
100.00% |
1 / 1 |
1 | |||
| parseHighlightLines | |
0.00% |
0 / 14 |
|
0.00% |
0 / 1 |
56 | |||
| validHighlightRange | |
0.00% |
0 / 3 |
|
0.00% |
0 / 1 |
12 | |||
| onContentGetParserOutput | |
0.00% |
0 / 25 |
|
0.00% |
0 / 1 |
42 | |||
| onApiFormatHighlight | |
0.00% |
0 / 16 |
|
0.00% |
0 / 1 |
20 | |||
| onSoftwareInfo | |
0.00% |
0 / 2 |
|
0.00% |
0 / 1 |
6 | |||
| 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 | |
| 19 | namespace MediaWiki\SyntaxHighlight; |
| 20 | |
| 21 | use Content; |
| 22 | use ExtensionRegistry; |
| 23 | use FormatJson; |
| 24 | use IContextSource; |
| 25 | use MediaWiki\Api\Hook\ApiFormatHighlightHook; |
| 26 | use MediaWiki\Content\Hook\ContentGetParserOutputHook; |
| 27 | use MediaWiki\Hook\ParserFirstCallInitHook; |
| 28 | use MediaWiki\Hook\SoftwareInfoHook; |
| 29 | use MediaWiki\Html\Html; |
| 30 | use MediaWiki\MediaWikiServices; |
| 31 | use MediaWiki\Parser\ParserOutput; |
| 32 | use MediaWiki\Parser\Sanitizer; |
| 33 | use MediaWiki\Status\Status; |
| 34 | use MediaWiki\Title\Title; |
| 35 | use Parser; |
| 36 | use ParserOptions; |
| 37 | use RuntimeException; |
| 38 | use TextContent; |
| 39 | use WANObjectCache; |
| 40 | use Wikimedia\Parsoid\DOM\DocumentFragment; |
| 41 | use Wikimedia\Parsoid\Ext\ExtensionTagHandler; |
| 42 | use Wikimedia\Parsoid\Ext\ParsoidExtensionAPI; |
| 43 | |
| 44 | class SyntaxHighlight extends ExtensionTagHandler implements |
| 45 | ParserFirstCallInitHook, |
| 46 | ContentGetParserOutputHook, |
| 47 | ApiFormatHighlightHook, |
| 48 | SoftwareInfoHook |
| 49 | { |
| 50 | |
| 51 | /** @var string CSS class for syntax-highlighted code. Public as used by the updateCSS maintenance script. */ |
| 52 | public const HIGHLIGHT_CSS_CLASS = 'mw-highlight'; |
| 53 | |
| 54 | /** @var int Cache version. Increment whenever the HTML changes. */ |
| 55 | private const CACHE_VERSION = 2; |
| 56 | |
| 57 | /** @var array<string,string> Mapping of MIME-types to lexer names. */ |
| 58 | private static $mimeLexers = [ |
| 59 | 'text/javascript' => 'javascript', |
| 60 | 'application/json' => 'javascript', |
| 61 | 'text/xml' => 'xml', |
| 62 | ]; |
| 63 | |
| 64 | /** |
| 65 | * Returns the maximum number of lines that may be selected for highlighting |
| 66 | * |
| 67 | * @return int |
| 68 | */ |
| 69 | private static function getMaxLines(): int { |
| 70 | $config = MediaWikiServices::getInstance()->getMainConfig(); |
| 71 | return $config->get( 'SyntaxHighlightMaxLines' ); |
| 72 | } |
| 73 | |
| 74 | /** |
| 75 | * Returns the maximum input size for the highlighter |
| 76 | * |
| 77 | * @return int |
| 78 | */ |
| 79 | private static function getMaxBytes(): int { |
| 80 | $config = MediaWikiServices::getInstance()->getMainConfig(); |
| 81 | return $config->get( 'SyntaxHighlightMaxBytes' ); |
| 82 | } |
| 83 | |
| 84 | /** |
| 85 | * Get the Pygments lexer name for a particular language. |
| 86 | * |
| 87 | * @param string|null $lang Language name. |
| 88 | * @return string|null Lexer name, or null if no matching lexer. |
| 89 | */ |
| 90 | private static function getLexer( $lang ) { |
| 91 | static $lexers = null; |
| 92 | |
| 93 | if ( $lang === null ) { |
| 94 | return null; |
| 95 | } |
| 96 | |
| 97 | $lexers ??= Pygmentize::getLexers(); |
| 98 | |
| 99 | $lexer = strtolower( $lang ); |
| 100 | |
| 101 | if ( isset( $lexers[$lexer] ) ) { |
| 102 | return $lexer; |
| 103 | } |
| 104 | |
| 105 | $geshi2pygments = SyntaxHighlightGeSHiCompat::getGeSHiToPygmentsMap(); |
| 106 | |
| 107 | // Check if this is a GeSHi lexer name for which there exists |
| 108 | // a compatible Pygments lexer with a different name. |
| 109 | if ( isset( $geshi2pygments[$lexer] ) ) { |
| 110 | $lexer = $geshi2pygments[$lexer]; |
| 111 | if ( isset( $lexers[$lexer] ) ) { |
| 112 | return $lexer; |
| 113 | } |
| 114 | } |
| 115 | |
| 116 | return null; |
| 117 | } |
| 118 | |
| 119 | /** |
| 120 | * Register parser hook |
| 121 | * |
| 122 | * @param Parser $parser |
| 123 | */ |
| 124 | public function onParserFirstCallInit( $parser ) { |
| 125 | $parser->setHook( 'source', [ self::class, 'parserHookSource' ] ); |
| 126 | $parser->setHook( 'syntaxhighlight', [ self::class, 'parserHook' ] ); |
| 127 | } |
| 128 | |
| 129 | /** |
| 130 | * Parser hook for <source> to add deprecated tracking category |
| 131 | * |
| 132 | * @param ?string $text |
| 133 | * @param array $args |
| 134 | * @param Parser $parser |
| 135 | * @return string |
| 136 | */ |
| 137 | public static function parserHookSource( $text, $args, $parser ) { |
| 138 | $parser->addTrackingCategory( 'syntaxhighlight-source-category' ); |
| 139 | return self::parserHook( $text, $args, $parser ); |
| 140 | } |
| 141 | |
| 142 | /** |
| 143 | * @return string[] |
| 144 | */ |
| 145 | private static function getModuleStyles(): array { |
| 146 | return [ 'ext.pygments' ]; |
| 147 | } |
| 148 | |
| 149 | /** |
| 150 | * |
| 151 | * @param string $text |
| 152 | * @param array $args |
| 153 | * @param ?Parser $parser |
| 154 | * @return array |
| 155 | */ |
| 156 | private static function processContent( string $text, array $args, ?Parser $parser = null ): array { |
| 157 | // Don't trim leading spaces away, just the linefeeds |
| 158 | $out = rtrim( trim( $text, "\n" ) ); |
| 159 | $trackingCats = []; |
| 160 | |
| 161 | // Convert deprecated attributes |
| 162 | if ( isset( $args['enclose'] ) ) { |
| 163 | if ( $args['enclose'] === 'none' ) { |
| 164 | $args['inline'] = true; |
| 165 | } |
| 166 | unset( $args['enclose'] ); |
| 167 | $trackingCats[] = 'syntaxhighlight-enclose-category'; |
| 168 | } |
| 169 | |
| 170 | $lexer = $args['lang'] ?? ''; |
| 171 | |
| 172 | $result = self::highlight( $out, $lexer, $args, $parser ); |
| 173 | if ( !$result->isGood() ) { |
| 174 | $trackingCats[] = 'syntaxhighlight-error-category'; |
| 175 | } |
| 176 | |
| 177 | return [ 'html' => $result->getValue(), 'cats' => $trackingCats ]; |
| 178 | } |
| 179 | |
| 180 | /** |
| 181 | * Parser hook for both <source> and <syntaxhighlight> logic |
| 182 | * |
| 183 | * @param ?string $text |
| 184 | * @param array $args |
| 185 | * @param Parser $parser |
| 186 | * @return string |
| 187 | */ |
| 188 | public static function parserHook( $text, $args, $parser ) { |
| 189 | // Replace strip markers (For e.g. {{#tag:syntaxhighlight|<nowiki>...}}) |
| 190 | $out = $parser->getStripState()->unstripNoWiki( $text ?? '' ); |
| 191 | |
| 192 | $result = self::processContent( $out, $args, $parser ); |
| 193 | foreach ( $result['cats'] as $cat ) { |
| 194 | $parser->addTrackingCategory( $cat ); |
| 195 | } |
| 196 | |
| 197 | // Register CSS |
| 198 | $parser->getOutput()->addModuleStyles( self::getModuleStyles() ); |
| 199 | if ( !empty( $args['linelinks'] ) ) { |
| 200 | $parser->getOutput()->addModules( [ 'ext.pygments.linenumbers' ] ); |
| 201 | } |
| 202 | |
| 203 | return $result['html']; |
| 204 | } |
| 205 | |
| 206 | /** @inheritDoc */ |
| 207 | public function sourceToDom( |
| 208 | ParsoidExtensionAPI $extApi, string $text, array $extArgs |
| 209 | ): ?DocumentFragment { |
| 210 | $result = self::processContent( $text, $extApi->extArgsToArray( $extArgs ) ); |
| 211 | |
| 212 | // FIXME: There is no API method in Parsoid to add tracking categories |
| 213 | // So, $result['cats'] is being ignored |
| 214 | |
| 215 | // Register CSS |
| 216 | $extApi->addModuleStyles( self::getModuleStyles() ); |
| 217 | |
| 218 | return $extApi->htmlToDom( $result['html'] ); |
| 219 | } |
| 220 | |
| 221 | /** |
| 222 | * Unwrap the <div> wrapper of the Pygments output |
| 223 | * |
| 224 | * @param string $out Output |
| 225 | * @return string Unwrapped output |
| 226 | */ |
| 227 | private static function unwrap( string $out ): string { |
| 228 | if ( $out !== '' ) { |
| 229 | $m = []; |
| 230 | if ( preg_match( '/^<div class="?mw-highlight"?>(.*)<\/div>$/s', trim( $out ), $m ) ) { |
| 231 | $out = trim( $m[1] ); |
| 232 | } else { |
| 233 | throw new RuntimeException( 'Unexpected output from Pygments encountered' ); |
| 234 | } |
| 235 | } |
| 236 | return $out; |
| 237 | } |
| 238 | |
| 239 | /** |
| 240 | * @param string $code |
| 241 | * @param bool $isInline |
| 242 | * @return string HTML |
| 243 | */ |
| 244 | private static function plainCodeWrap( $code, $isInline ) { |
| 245 | if ( $isInline ) { |
| 246 | return htmlspecialchars( $code, ENT_NOQUOTES ); |
| 247 | } |
| 248 | |
| 249 | return Html::rawElement( |
| 250 | 'div', |
| 251 | [ 'class' => self::HIGHLIGHT_CSS_CLASS ], |
| 252 | Html::element( 'pre', [], $code ) |
| 253 | ); |
| 254 | } |
| 255 | |
| 256 | /** |
| 257 | * @param string $code |
| 258 | * @param string|null $lang |
| 259 | * @param array $args |
| 260 | * @param Parser|null $parser Parser, if generating content to be parsed. |
| 261 | * @return Status |
| 262 | */ |
| 263 | private static function highlightInner( $code, $lang = null, $args = [], ?Parser $parser = null ) { |
| 264 | $status = new Status; |
| 265 | |
| 266 | $lexer = self::getLexer( $lang ); |
| 267 | if ( $lexer === null && $lang !== null ) { |
| 268 | $status->warning( 'syntaxhighlight-error-unknown-language', $lang ); |
| 269 | } |
| 270 | |
| 271 | // For empty tag, output nothing instead of empty <pre>. |
| 272 | if ( $code === '' ) { |
| 273 | $status->value = ''; |
| 274 | return $status; |
| 275 | } |
| 276 | |
| 277 | $length = strlen( $code ); |
| 278 | if ( strlen( $code ) > self::getMaxBytes() ) { |
| 279 | // Disable syntax highlighting |
| 280 | $lexer = null; |
| 281 | $status->warning( |
| 282 | 'syntaxhighlight-error-exceeds-size-limit', |
| 283 | $length, |
| 284 | self::getMaxBytes() |
| 285 | ); |
| 286 | } |
| 287 | |
| 288 | $isInline = isset( $args['inline'] ); |
| 289 | $showLines = isset( $args['line'] ); |
| 290 | |
| 291 | if ( $isInline ) { |
| 292 | $code = trim( $code ); |
| 293 | } |
| 294 | |
| 295 | if ( $lexer === null ) { |
| 296 | // When syntax highlighting is disabled.. |
| 297 | $status->value = self::plainCodeWrap( $code, $isInline ); |
| 298 | return $status; |
| 299 | } |
| 300 | |
| 301 | if ( $parser && !$parser->incrementExpensiveFunctionCount() ) { |
| 302 | // Highlighting is expensive, return unstyled |
| 303 | $status->value = self::plainCodeWrap( $code, $isInline ); |
| 304 | return $status; |
| 305 | } |
| 306 | |
| 307 | $options = [ |
| 308 | 'cssclass' => self::HIGHLIGHT_CSS_CLASS, |
| 309 | 'encoding' => 'utf-8', |
| 310 | ]; |
| 311 | |
| 312 | // Line numbers |
| 313 | if ( $showLines ) { |
| 314 | $options['linenos'] = 'inline'; |
| 315 | } |
| 316 | |
| 317 | if ( $lexer === 'php' && strpos( $code, '<?php' ) === false ) { |
| 318 | $options['startinline'] = 1; |
| 319 | } |
| 320 | |
| 321 | // Highlight specific lines |
| 322 | if ( isset( $args['highlight'] ) ) { |
| 323 | $lines = self::parseHighlightLines( $args['highlight'] ); |
| 324 | if ( count( $lines ) ) { |
| 325 | $options['hl_lines'] = implode( ' ', $lines ); |
| 326 | } |
| 327 | } |
| 328 | |
| 329 | // Starting line number |
| 330 | if ( isset( $args['start'] ) && ctype_digit( $args['start'] ) ) { |
| 331 | $options['linenostart'] = (int)$args['start']; |
| 332 | } |
| 333 | |
| 334 | if ( !empty( $args['linelinks'] ) ) { |
| 335 | $options['linespans'] = $args['linelinks']; |
| 336 | } |
| 337 | |
| 338 | if ( $isInline ) { |
| 339 | $options['nowrap'] = 1; |
| 340 | } |
| 341 | |
| 342 | $cache = MediaWikiServices::getInstance()->getMainWANObjectCache(); |
| 343 | $error = null; |
| 344 | $output = $cache->getWithSetCallback( |
| 345 | $cache->makeGlobalKey( 'highlight', self::makeCacheKeyHash( $code, $lexer, $options ) ), |
| 346 | $cache::TTL_MONTH, |
| 347 | static function ( $oldValue, &$ttl ) use ( $code, $lexer, $options, &$error ) { |
| 348 | try { |
| 349 | return Pygmentize::highlight( $lexer, $code, $options ); |
| 350 | } catch ( PygmentsException $e ) { |
| 351 | $ttl = WANObjectCache::TTL_UNCACHEABLE; |
| 352 | $error = $e->getMessage(); |
| 353 | return null; |
| 354 | } |
| 355 | } |
| 356 | ); |
| 357 | |
| 358 | if ( $error !== null || $output === null ) { |
| 359 | $status->warning( 'syntaxhighlight-error-pygments-invocation-failure' ); |
| 360 | if ( $error !== null ) { |
| 361 | wfWarn( 'Failed to invoke Pygments: ' . $error ); |
| 362 | } else { |
| 363 | wfWarn( 'Invoking Pygments returned blank output with no error response' ); |
| 364 | } |
| 365 | |
| 366 | // Fall back to preformatted code without syntax highlighting |
| 367 | $output = self::plainCodeWrap( $code, $isInline ); |
| 368 | } |
| 369 | |
| 370 | $status->value = $output; |
| 371 | |
| 372 | return $status; |
| 373 | } |
| 374 | |
| 375 | /** |
| 376 | * Highlight a code-block using a particular lexer. |
| 377 | * |
| 378 | * This produces raw HTML (wrapped by Status), the caller is responsible |
| 379 | * for making sure the "ext.pygments" module is loaded in the output. |
| 380 | * |
| 381 | * @param string $code Code to highlight. |
| 382 | * @param string|null $lang Language name, or null to use plain markup. |
| 383 | * @param array $args Associative array of additional arguments. |
| 384 | * If it contains a 'line' key, the output will include line numbers. |
| 385 | * If it includes a 'highlight' key, the value will be parsed as a |
| 386 | * comma-separated list of lines and line-ranges to highlight. |
| 387 | * If it contains a 'start' key, the value will be used as the line at which to |
| 388 | * start highlighting. |
| 389 | * If it contains a 'inline' key, the output will not be wrapped in `<div><pre/></div>`. |
| 390 | * If it contains a 'linelinks' key, lines will have links and anchors with a prefix |
| 391 | * of the value. Similar to the lineanchors+linespans features in Pygments. |
| 392 | * @param Parser|null $parser Parser, if generating content to be parsed. |
| 393 | * @return Status Status object, with HTML representing the highlighted |
| 394 | * code as its value. |
| 395 | */ |
| 396 | public static function highlight( $code, $lang = null, $args = [], ?Parser $parser = null ) { |
| 397 | $status = self::highlightInner( $code, $lang, $args, $parser ); |
| 398 | $output = $status->getValue(); |
| 399 | |
| 400 | $isInline = isset( $args['inline'] ); |
| 401 | $showLines = isset( $args['line'] ); |
| 402 | $lexer = self::getLexer( $lang ); |
| 403 | |
| 404 | // Post-Pygment HTML transformations. |
| 405 | |
| 406 | if ( $showLines ) { |
| 407 | $lineReplace = Html::element( 'span', [ 'class' => 'linenos', 'data-line' => '$1' ] ); |
| 408 | if ( !empty( $args['linelinks'] ) ) { |
| 409 | $lineReplace = Html::rawElement( |
| 410 | 'a', |
| 411 | [ 'href' => '#' . $args['linelinks'] . '-$1' ], |
| 412 | $lineReplace |
| 413 | ); |
| 414 | } |
| 415 | // Convert line numbers to data attributes so they |
| 416 | // can be displayed as CSS generated content and be |
| 417 | // unselectable in all browsers. |
| 418 | $output = preg_replace( |
| 419 | '`<span class="linenos">\s*([^<]*)\s*</span>`', |
| 420 | $lineReplace, |
| 421 | $output |
| 422 | ); |
| 423 | } |
| 424 | |
| 425 | // Allow certain HTML attributes |
| 426 | $htmlAttribs = Sanitizer::validateAttributes( |
| 427 | $args, array_flip( [ 'style', 'class', 'id' ] ) |
| 428 | ); |
| 429 | |
| 430 | $dir = ( isset( $args['dir'] ) && $args['dir'] === 'rtl' ) ? 'rtl' : 'ltr'; |
| 431 | |
| 432 | // Build class list |
| 433 | $classList = []; |
| 434 | if ( isset( $htmlAttribs['class'] ) ) { |
| 435 | $classList[] = $htmlAttribs['class']; |
| 436 | } |
| 437 | $classList[] = self::HIGHLIGHT_CSS_CLASS; |
| 438 | if ( $lexer !== null ) { |
| 439 | $classList[] = self::HIGHLIGHT_CSS_CLASS . '-lang-' . $lexer; |
| 440 | } |
| 441 | $classList[] = 'mw-content-' . $dir; |
| 442 | if ( $showLines ) { |
| 443 | $classList[] = self::HIGHLIGHT_CSS_CLASS . '-lines'; |
| 444 | } |
| 445 | $htmlAttribs['class'] = implode( ' ', $classList ); |
| 446 | $htmlAttribs['dir'] = $dir; |
| 447 | '@phan-var array{class:string,dir:string} $htmlAttribs'; |
| 448 | |
| 449 | if ( $isInline ) { |
| 450 | // We've already trimmed the input $code before highlighting, |
| 451 | // but pygment's standard out adds a line break afterwards, |
| 452 | // which would then be preserved in the paragraph that wraps this, |
| 453 | // and become visible as a space. Avoid that. |
| 454 | $output = trim( $output ); |
| 455 | |
| 456 | // Enforce inlineness. Stray newlines may result in unexpected list and paragraph processing |
| 457 | // (also known as doBlockLevels()). |
| 458 | $output = str_replace( "\n", ' ', $output ); |
| 459 | $output = Html::rawElement( 'code', $htmlAttribs, $output ); |
| 460 | } else { |
| 461 | $output = self::unwrap( $output ); |
| 462 | |
| 463 | if ( $parser ) { |
| 464 | // Use 'nowiki' strip marker to prevent list processing (also known as doBlockLevels()). |
| 465 | // However, leave the wrapping <div/> outside to prevent <p/>-wrapping. |
| 466 | $marker = $parser::MARKER_PREFIX . '-syntaxhighlightinner-' . |
| 467 | sprintf( '%08X', $parser->mMarkerIndex++ ) . $parser::MARKER_SUFFIX; |
| 468 | $parser->getStripState()->addNoWiki( $marker, $output ); |
| 469 | $output = $marker; |
| 470 | } |
| 471 | |
| 472 | $output = Html::openElement( 'div', $htmlAttribs ) . |
| 473 | $output . |
| 474 | Html::closeElement( 'div' ); |
| 475 | } |
| 476 | |
| 477 | $status->value = $output; |
| 478 | return $status; |
| 479 | } |
| 480 | |
| 481 | /** |
| 482 | * Construct a cache key for the results of a Pygments invocation. |
| 483 | * |
| 484 | * @param string $code Code to be highlighted. |
| 485 | * @param string $lexer Lexer name. |
| 486 | * @param array $options Options array. |
| 487 | * @return string Cache key. |
| 488 | */ |
| 489 | private static function makeCacheKeyHash( $code, $lexer, $options ) { |
| 490 | $optionString = FormatJson::encode( $options, false, FormatJson::ALL_OK ); |
| 491 | return md5( "{$code}|{$lexer}|{$optionString}|" . self::CACHE_VERSION ); |
| 492 | } |
| 493 | |
| 494 | /** |
| 495 | * Take an input specifying a list of lines to highlight, returning |
| 496 | * a raw list of matching line numbers. |
| 497 | * |
| 498 | * Input is comma-separated list of lines or line ranges. |
| 499 | * |
| 500 | * @param string $lineSpec |
| 501 | * @return int[] Line numbers. |
| 502 | */ |
| 503 | protected static function parseHighlightLines( $lineSpec ) { |
| 504 | $lines = []; |
| 505 | $values = array_map( 'trim', explode( ',', $lineSpec ) ); |
| 506 | foreach ( $values as $value ) { |
| 507 | if ( ctype_digit( $value ) ) { |
| 508 | $lines[] = (int)$value; |
| 509 | } elseif ( strpos( $value, '-' ) !== false ) { |
| 510 | [ $start, $end ] = array_map( 'intval', explode( '-', $value ) ); |
| 511 | if ( self::validHighlightRange( $start, $end ) ) { |
| 512 | for ( $i = $start; $i <= $end; $i++ ) { |
| 513 | $lines[] = $i; |
| 514 | } |
| 515 | } |
| 516 | } |
| 517 | if ( count( $lines ) > self::getMaxLines() ) { |
| 518 | $lines = array_slice( $lines, 0, self::getMaxLines() ); |
| 519 | break; |
| 520 | } |
| 521 | } |
| 522 | return $lines; |
| 523 | } |
| 524 | |
| 525 | /** |
| 526 | * Validate a provided input range |
| 527 | * @param int $start |
| 528 | * @param int $end |
| 529 | * @return bool |
| 530 | */ |
| 531 | protected static function validHighlightRange( $start, $end ) { |
| 532 | // Since we're taking this tiny range and producing a an |
| 533 | // array of every integer between them, it would be trivial |
| 534 | // to DoS the system by asking for a huge range. |
| 535 | // Impose an arbitrary limit on the number of lines in a |
| 536 | // given range to reduce the impact. |
| 537 | return $start > 0 && |
| 538 | $start < $end && |
| 539 | $end - $start < self::getMaxLines(); |
| 540 | } |
| 541 | |
| 542 | /** |
| 543 | * Hook into Content::getParserOutput to provide syntax highlighting for |
| 544 | * script content. |
| 545 | * |
| 546 | * @param Content $content |
| 547 | * @param Title $title |
| 548 | * @param int $revId |
| 549 | * @param ParserOptions $options |
| 550 | * @param bool $generateHtml |
| 551 | * @param ParserOutput &$parserOutput |
| 552 | * @return bool |
| 553 | * @since MW 1.21 |
| 554 | */ |
| 555 | public function onContentGetParserOutput( $content, $title, |
| 556 | $revId, $options, $generateHtml, &$parserOutput |
| 557 | ) { |
| 558 | global $wgTextModelsToParse; |
| 559 | |
| 560 | // Hope that the "SyntaxHighlightModels" attribute does not contain silly types. |
| 561 | if ( !( $content instanceof TextContent ) ) { |
| 562 | // Oops! Non-text content? Let MediaWiki handle this. |
| 563 | return true; |
| 564 | } |
| 565 | |
| 566 | if ( !$generateHtml ) { |
| 567 | // Nothing special for us to do, let MediaWiki handle this. |
| 568 | return true; |
| 569 | } |
| 570 | |
| 571 | // Determine the SyntaxHighlight language from the page's |
| 572 | // content model. Extensions can extend the default CSS/JS |
| 573 | // mapping by setting the SyntaxHighlightModels attribute. |
| 574 | $extension = ExtensionRegistry::getInstance(); |
| 575 | $models = $extension->getAttribute( 'SyntaxHighlightModels' ) + [ |
| 576 | CONTENT_MODEL_CSS => 'css', |
| 577 | CONTENT_MODEL_JAVASCRIPT => 'javascript', |
| 578 | ]; |
| 579 | $model = $content->getModel(); |
| 580 | if ( !isset( $models[$model] ) ) { |
| 581 | // We don't care about this model, carry on. |
| 582 | return true; |
| 583 | } |
| 584 | $lexer = $models[$model]; |
| 585 | $text = $content->getText(); |
| 586 | |
| 587 | // Parse using the standard parser to get links etc. into the database, HTML is replaced below. |
| 588 | // We could do this using $content->fillParserOutput(), but alas it is 'protected'. |
| 589 | if ( in_array( $model, $wgTextModelsToParse, true ) ) { |
| 590 | $parserOutput = MediaWikiServices::getInstance()->getParser() |
| 591 | ->parse( $text, $title, $options, true, true, $revId ); |
| 592 | } |
| 593 | |
| 594 | $status = self::highlight( $text, $lexer, [ 'line' => true, 'linelinks' => 'L' ] ); |
| 595 | if ( !$status->isOK() ) { |
| 596 | return true; |
| 597 | } |
| 598 | $out = $status->getValue(); |
| 599 | |
| 600 | $parserOutput->addModuleStyles( self::getModuleStyles() ); |
| 601 | $parserOutput->addModules( [ 'ext.pygments.linenumbers' ] ); |
| 602 | $parserOutput->setText( $out ); |
| 603 | |
| 604 | // Inform MediaWiki that we have parsed this page and it shouldn't mess with it. |
| 605 | return false; |
| 606 | } |
| 607 | |
| 608 | /** |
| 609 | * Hook to provide syntax highlighting for API pretty-printed output |
| 610 | * |
| 611 | * @param IContextSource $context |
| 612 | * @param string $text |
| 613 | * @param string $mime |
| 614 | * @param string $format |
| 615 | * @since MW 1.24 |
| 616 | * @return bool |
| 617 | */ |
| 618 | public function onApiFormatHighlight( $context, $text, $mime, $format ) { |
| 619 | if ( !isset( self::$mimeLexers[$mime] ) ) { |
| 620 | return true; |
| 621 | } |
| 622 | |
| 623 | $lexer = self::$mimeLexers[$mime]; |
| 624 | $status = self::highlight( $text, $lexer ); |
| 625 | if ( !$status->isOK() ) { |
| 626 | return true; |
| 627 | } |
| 628 | |
| 629 | $out = $status->getValue(); |
| 630 | if ( preg_match( '/^<pre([^>]*)>/i', $out, $m ) ) { |
| 631 | $attrs = Sanitizer::decodeTagAttributes( $m[1] ); |
| 632 | $attrs['class'] .= ' api-pretty-content'; |
| 633 | $encodedAttrs = Sanitizer::safeEncodeTagAttributes( $attrs ); |
| 634 | $out = '<pre' . $encodedAttrs . '>' . substr( $out, strlen( $m[0] ) ); |
| 635 | } |
| 636 | $output = $context->getOutput(); |
| 637 | $output->addModuleStyles( self::getModuleStyles() ); |
| 638 | $output->addHTML( '<div dir="ltr">' . $out . '</div>' ); |
| 639 | |
| 640 | // Inform MediaWiki that we have parsed this page and it shouldn't mess with it. |
| 641 | return false; |
| 642 | } |
| 643 | |
| 644 | /** |
| 645 | * Hook to add Pygments version to Special:Version |
| 646 | * |
| 647 | * @see https://www.mediawiki.org/wiki/Manual:Hooks/SoftwareInfo |
| 648 | * @param array &$software |
| 649 | */ |
| 650 | public function onSoftwareInfo( &$software ) { |
| 651 | try { |
| 652 | $software['[https://pygments.org/ Pygments]'] = Pygmentize::getVersion(); |
| 653 | } catch ( PygmentsException $e ) { |
| 654 | // pass |
| 655 | } |
| 656 | } |
| 657 | } |