Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
| Total | |
54.33% |
270 / 497 |
|
31.11% |
14 / 45 |
CRAP | |
0.00% |
0 / 1 |
| LanguageConverter | |
54.44% |
270 / 496 |
|
31.11% |
14 / 45 |
3280.11 | |
0.00% |
0 / 1 |
| __construct | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| getMainCode | n/a |
0 / 0 |
n/a |
0 / 0 |
0 | |||||
| getStaticDefaultVariant | |
100.00% |
2 / 2 |
|
100.00% |
1 / 1 |
1 | |||
| getLanguageVariants | n/a |
0 / 0 |
n/a |
0 / 0 |
0 | |||||
| getVariantsFallbacks | n/a |
0 / 0 |
n/a |
0 / 0 |
0 | |||||
| getFlags | |
0.00% |
0 / 13 |
|
0.00% |
0 / 1 |
6 | |||
| getAdditionalFlags | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| getManualLevel | |
0.00% |
0 / 7 |
|
0.00% |
0 / 1 |
12 | |||
| getAdditionalManualLevel | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| getDescCodeSeparator | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| getDescVarSeparator | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| getVariantNames | |
0.00% |
0 / 3 |
|
0.00% |
0 / 1 |
2 | |||
| getVariants | |
100.00% |
3 / 3 |
|
100.00% |
1 / 1 |
1 | |||
| getVariantFallbacks | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| getConvRuleTitle | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| getPreferredVariant | |
100.00% |
14 / 14 |
|
100.00% |
1 / 1 |
6 | |||
| getDefaultVariant | |
0.00% |
0 / 6 |
|
0.00% |
0 / 1 |
12 | |||
| validateVariant | |
100.00% |
9 / 9 |
|
100.00% |
1 / 1 |
5 | |||
| getURLVariant | |
87.50% |
7 / 8 |
|
0.00% |
0 / 1 |
3.02 | |||
| getUserVariant | |
84.62% |
11 / 13 |
|
0.00% |
0 / 1 |
5.09 | |||
| getHeaderVariant | |
64.00% |
16 / 25 |
|
0.00% |
0 / 1 |
16.65 | |||
| autoConvert | |
70.27% |
52 / 74 |
|
0.00% |
0 / 1 |
28.49 | |||
| translate | |
100.00% |
4 / 4 |
|
100.00% |
1 / 1 |
2 | |||
| translateWithoutRomanNumbers | |
70.59% |
12 / 17 |
|
0.00% |
0 / 1 |
3.23 | |||
| autoConvertToAllVariants | |
0.00% |
0 / 5 |
|
0.00% |
0 / 1 |
6 | |||
| applyManualConv | |
35.71% |
5 / 14 |
|
0.00% |
0 / 1 |
20.02 | |||
| convertSplitTitle | |
100.00% |
6 / 6 |
|
100.00% |
1 / 1 |
1 | |||
| convertTitle | |
100.00% |
4 / 4 |
|
100.00% |
1 / 1 |
2 | |||
| convertNamespace | |
100.00% |
12 / 12 |
|
100.00% |
1 / 1 |
2 | |||
| computeNsVariantText | |
85.71% |
12 / 14 |
|
0.00% |
0 / 1 |
5.07 | |||
| convert | |
0.00% |
0 / 2 |
|
0.00% |
0 / 1 |
2 | |||
| convertTo | |
83.33% |
5 / 6 |
|
0.00% |
0 / 1 |
3.04 | |||
| recursiveConvertTopLevel | |
100.00% |
26 / 26 |
|
100.00% |
1 / 1 |
6 | |||
| recursiveConvertRule | |
51.28% |
20 / 39 |
|
0.00% |
0 / 1 |
24.99 | |||
| findVariantLink | |
0.00% |
0 / 43 |
|
0.00% |
0 / 1 |
420 | |||
| getExtraHashOptions | |
0.00% |
0 / 2 |
|
0.00% |
0 / 1 |
2 | |||
| guessVariant | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| loadDefaultTables | n/a |
0 / 0 |
n/a |
0 / 0 |
0 | |||||
| loadTables | |
95.24% |
20 / 21 |
|
0.00% |
0 / 1 |
4 | |||
| postLoadTables | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| reloadTables | |
0.00% |
0 / 4 |
|
0.00% |
0 / 1 |
6 | |||
| parseCachedTable | |
17.19% |
11 / 64 |
|
0.00% |
0 / 1 |
296.87 | |||
| markNoConversion | |
0.00% |
0 / 3 |
|
0.00% |
0 / 1 |
12 | |||
| convertCategoryKey | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| updateConversionTable | |
0.00% |
0 / 5 |
|
0.00% |
0 / 1 |
30 | |||
| getVarSeparatorPattern | |
93.33% |
14 / 15 |
|
0.00% |
0 / 1 |
5.01 | |||
| hasVariants | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| hasVariant | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
2 | |||
| convertHtml | |
0.00% |
0 / 1 |
|
0.00% |
0 / 1 |
2 | |||
| 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 | * @file |
| 19 | * @author Zhengzhu Feng <zhengzhu@gmail.com> |
| 20 | * @author fdcn <fdcn64@gmail.com> |
| 21 | * @author shinjiman <shinjiman@gmail.com> |
| 22 | * @author PhiLiP <philip.npc@gmail.com> |
| 23 | */ |
| 24 | |
| 25 | namespace MediaWiki\Language; |
| 26 | |
| 27 | use InvalidArgumentException; |
| 28 | use MediaWiki\Context\RequestContext; |
| 29 | use MediaWiki\Debug\DeprecationHelper; |
| 30 | use MediaWiki\HookContainer\HookRunner; |
| 31 | use MediaWiki\Html\Html; |
| 32 | use MediaWiki\Linker\LinkTarget; |
| 33 | use MediaWiki\Logger\LoggerFactory; |
| 34 | use MediaWiki\MainConfigNames; |
| 35 | use MediaWiki\MediaWikiServices; |
| 36 | use MediaWiki\Page\PageIdentity; |
| 37 | use MediaWiki\Parser\Parser; |
| 38 | use MediaWiki\Parser\Sanitizer; |
| 39 | use MediaWiki\Revision\RevisionRecord; |
| 40 | use MediaWiki\Revision\SlotRecord; |
| 41 | use MediaWiki\StubObject\StubUserLang; |
| 42 | use MediaWiki\Title\Title; |
| 43 | use MediaWiki\User\User; |
| 44 | use RuntimeException; |
| 45 | use StringUtils; |
| 46 | use UnexpectedValueException; |
| 47 | use Wikimedia\ObjectCache\BagOStuff; |
| 48 | |
| 49 | /** |
| 50 | * Base class for multi-variant language conversion. |
| 51 | * |
| 52 | * @ingroup Language |
| 53 | */ |
| 54 | abstract class LanguageConverter implements ILanguageConverter { |
| 55 | use DeprecationHelper; |
| 56 | |
| 57 | /** |
| 58 | * languages supporting variants |
| 59 | * @since 1.20 |
| 60 | * @var string[] |
| 61 | */ |
| 62 | public static $languagesWithVariants = [ |
| 63 | 'ban', |
| 64 | 'en', |
| 65 | 'crh', |
| 66 | 'gan', |
| 67 | 'iu', |
| 68 | 'ku', |
| 69 | 'mni', |
| 70 | 'sh', |
| 71 | 'shi', |
| 72 | 'sr', |
| 73 | 'tg', |
| 74 | 'tly', |
| 75 | 'uz', |
| 76 | 'wuu', |
| 77 | 'zgh', |
| 78 | 'zh', |
| 79 | ]; |
| 80 | |
| 81 | /** |
| 82 | * static default variant of languages supporting variants |
| 83 | * for use with DefaultOptionsLookup.php |
| 84 | * @since 1.40 |
| 85 | * @var array<string,string> |
| 86 | */ |
| 87 | public static $languagesWithStaticDefaultVariant = [ |
| 88 | 'ban' => 'ban', |
| 89 | 'en' => 'en', |
| 90 | 'crh' => 'crh', |
| 91 | 'gan' => 'gan', |
| 92 | 'iu' => 'iu', |
| 93 | 'ku' => 'ku', |
| 94 | 'mni' => 'mni', |
| 95 | 'sh' => 'sh-latn', |
| 96 | 'shi' => 'shi', |
| 97 | 'sr' => 'sr', |
| 98 | 'tg' => 'tg', |
| 99 | 'tly' => 'tly', |
| 100 | 'uz' => 'uz', |
| 101 | 'wuu' => 'wuu', |
| 102 | 'zgh' => 'zgh', |
| 103 | 'zh' => 'zh', |
| 104 | ]; |
| 105 | |
| 106 | /** @var bool */ |
| 107 | private $mTablesLoaded = false; |
| 108 | /** @var ReplacementArray[] */ |
| 109 | protected $mTables = []; |
| 110 | /** @var Language|StubUserLang */ |
| 111 | private $mLangObj; |
| 112 | /** @var string|false */ |
| 113 | private $mConvRuleTitle = false; |
| 114 | /** @var string|null */ |
| 115 | private $mURLVariant; |
| 116 | /** @var string|null */ |
| 117 | private $mUserVariant; |
| 118 | /** @var string|null */ |
| 119 | private $mHeaderVariant; |
| 120 | /** @var int */ |
| 121 | private $mMaxDepth = 10; |
| 122 | /** @var string|null */ |
| 123 | private $mVarSeparatorPattern; |
| 124 | |
| 125 | private const CACHE_VERSION_KEY = 'VERSION 7'; |
| 126 | |
| 127 | /** |
| 128 | * @param Language|StubUserLang $langobj |
| 129 | */ |
| 130 | public function __construct( $langobj ) { |
| 131 | $this->mLangObj = $langobj; |
| 132 | } |
| 133 | |
| 134 | /** |
| 135 | * Get the language code with converter (the "main" language code). |
| 136 | * Page language code would be the same of the language code with converter. |
| 137 | * Note that this code might not be included as one of the variant languages. |
| 138 | * @since 1.36 |
| 139 | * |
| 140 | * @return string |
| 141 | */ |
| 142 | abstract public function getMainCode(): string; |
| 143 | |
| 144 | /** |
| 145 | * Get static default variant. |
| 146 | * For use of specify the default variant form when it different from the |
| 147 | * default "unconverted/mixed-variant form". |
| 148 | * @since 1.40 |
| 149 | * |
| 150 | * @return string |
| 151 | */ |
| 152 | protected function getStaticDefaultVariant(): string { |
| 153 | $code = $this->getMainCode(); |
| 154 | return self::$languagesWithStaticDefaultVariant[$code] ?? $code; |
| 155 | } |
| 156 | |
| 157 | /** |
| 158 | * Get supported variants of the language. |
| 159 | * @since 1.36 |
| 160 | * |
| 161 | * @return array |
| 162 | */ |
| 163 | abstract protected function getLanguageVariants(): array; |
| 164 | |
| 165 | /** |
| 166 | * Get language variants fallbacks. |
| 167 | * @since 1.36 |
| 168 | * |
| 169 | * @return array |
| 170 | */ |
| 171 | abstract public function getVariantsFallbacks(): array; |
| 172 | |
| 173 | /** |
| 174 | * Get the strings that map to the flags. |
| 175 | * @since 1.36 |
| 176 | * |
| 177 | * @return array |
| 178 | */ |
| 179 | final public function getFlags(): array { |
| 180 | $defaultflags = [ |
| 181 | // 'S' show the converted text |
| 182 | // '+' add rules for alltext |
| 183 | // 'E' the flags have an error |
| 184 | // these flags above are reserved for program |
| 185 | 'A' => 'A', // add rule for convert code (all text converted) |
| 186 | 'T' => 'T', // title convert |
| 187 | 'R' => 'R', // raw content |
| 188 | 'D' => 'D', // convert description (subclass implement) |
| 189 | '-' => '-', // remove convert (not implement) |
| 190 | 'H' => 'H', // add rule for convert code (but no display in placed code) |
| 191 | 'N' => 'N', // current variant name |
| 192 | ]; |
| 193 | $flags = array_merge( $defaultflags, $this->getAdditionalFlags() ); |
| 194 | foreach ( $this->getVariants() as $v ) { |
| 195 | $flags[$v] = $v; |
| 196 | } |
| 197 | return $flags; |
| 198 | } |
| 199 | |
| 200 | /** |
| 201 | * Provides additional flags for converter. By default, it returns empty array and |
| 202 | * typically should be overridden by implementation of converter. |
| 203 | * |
| 204 | * @return array |
| 205 | */ |
| 206 | protected function getAdditionalFlags(): array { |
| 207 | return []; |
| 208 | } |
| 209 | |
| 210 | /** |
| 211 | * Get manual level limit for supported variants. |
| 212 | * @since 1.36 |
| 213 | * |
| 214 | * @return array |
| 215 | */ |
| 216 | final public function getManualLevel() { |
| 217 | $manualLevel = $this->getAdditionalManualLevel(); |
| 218 | $result = []; |
| 219 | foreach ( $this->getVariants() as $v ) { |
| 220 | if ( array_key_exists( $v, $manualLevel ) ) { |
| 221 | $result[$v] = $manualLevel[$v]; |
| 222 | } else { |
| 223 | $result[$v] = 'bidirectional'; |
| 224 | } |
| 225 | } |
| 226 | return $result; |
| 227 | } |
| 228 | |
| 229 | /** |
| 230 | * Provides additional flags for converter. By default, this function returns an empty array and |
| 231 | * typically should be overridden by the implementation of converter. |
| 232 | * @since 1.36 |
| 233 | * |
| 234 | * @return array |
| 235 | */ |
| 236 | protected function getAdditionalManualLevel(): array { |
| 237 | return []; |
| 238 | } |
| 239 | |
| 240 | /** |
| 241 | * Get desc code separator. By default returns ":", can be overridden by |
| 242 | * implementation of converter. |
| 243 | * @since 1.36 |
| 244 | * |
| 245 | * @return string |
| 246 | */ |
| 247 | public function getDescCodeSeparator(): string { |
| 248 | return ':'; |
| 249 | } |
| 250 | |
| 251 | /** |
| 252 | * Get desc var separator. By default returns ";", can be overridden by |
| 253 | * implementation of converter. |
| 254 | * @since 1.36 |
| 255 | * |
| 256 | * @return string |
| 257 | */ |
| 258 | public function getDescVarSeparator(): string { |
| 259 | return ';'; |
| 260 | } |
| 261 | |
| 262 | /** |
| 263 | * Get variant names. |
| 264 | * |
| 265 | * @return array |
| 266 | */ |
| 267 | public function getVariantNames(): array { |
| 268 | return MediaWikiServices::getInstance() |
| 269 | ->getLanguageNameUtils() |
| 270 | ->getLanguageNames(); |
| 271 | } |
| 272 | |
| 273 | final public function getVariants() { |
| 274 | $disabledVariants = MediaWikiServices::getInstance()->getMainConfig()->get( |
| 275 | MainConfigNames::DisabledVariants ); |
| 276 | return array_diff( $this->getLanguageVariants(), $disabledVariants ); |
| 277 | } |
| 278 | |
| 279 | public function getVariantFallbacks( $variant ) { |
| 280 | return $this->getVariantsFallbacks()[$variant] ?? $this->getStaticDefaultVariant(); |
| 281 | } |
| 282 | |
| 283 | public function getConvRuleTitle() { |
| 284 | return $this->mConvRuleTitle; |
| 285 | } |
| 286 | |
| 287 | public function getPreferredVariant() { |
| 288 | $req = $this->getURLVariant(); |
| 289 | |
| 290 | $services = MediaWikiServices::getInstance(); |
| 291 | ( new HookRunner( $services->getHookContainer() ) )->onGetLangPreferredVariant( $req ); |
| 292 | |
| 293 | if ( !$req ) { |
| 294 | $user = RequestContext::getMain()->getUser(); |
| 295 | // NOTE: For some calls there may not be a context user or session that is safe |
| 296 | // to use, see (T235360) |
| 297 | // Use case: During user autocreation, UserNameUtils::isUsable is called which uses interface |
| 298 | // messages for reserved usernames. |
| 299 | if ( $user->isSafeToLoad() && $user->isRegistered() ) { |
| 300 | $req = $this->getUserVariant( $user ); |
| 301 | } else { |
| 302 | $req = $this->getHeaderVariant(); |
| 303 | } |
| 304 | } |
| 305 | |
| 306 | $defaultLanguageVariant = $services->getMainConfig() |
| 307 | ->get( MainConfigNames::DefaultLanguageVariant ); |
| 308 | if ( !$req && $defaultLanguageVariant ) { |
| 309 | $req = $this->validateVariant( $defaultLanguageVariant ); |
| 310 | } |
| 311 | |
| 312 | $req = $this->validateVariant( $req ); |
| 313 | |
| 314 | // This function, unlike the other get*Variant functions, is |
| 315 | // not memoized (i.e., there return value is not cached) since |
| 316 | // new information might appear during processing after this |
| 317 | // is first called. |
| 318 | return $req ?? $this->getStaticDefaultVariant(); |
| 319 | } |
| 320 | |
| 321 | public function getDefaultVariant() { |
| 322 | $defaultLanguageVariant = MediaWikiServices::getInstance()->getMainConfig()->get( |
| 323 | MainConfigNames::DefaultLanguageVariant ); |
| 324 | |
| 325 | $req = $this->getURLVariant() ?? $this->getHeaderVariant(); |
| 326 | |
| 327 | if ( !$req && $defaultLanguageVariant ) { |
| 328 | $req = $this->validateVariant( $defaultLanguageVariant ); |
| 329 | } |
| 330 | |
| 331 | return $req ?? $this->getStaticDefaultVariant(); |
| 332 | } |
| 333 | |
| 334 | public function validateVariant( $variant = null ) { |
| 335 | if ( $variant === null ) { |
| 336 | return null; |
| 337 | } |
| 338 | // Our internal variants are always lower-case; the variant we |
| 339 | // are validating may have mixed cases. |
| 340 | $variant = LanguageCode::replaceDeprecatedCodes( strtolower( $variant ) ); |
| 341 | if ( in_array( $variant, $this->getVariants() ) ) { |
| 342 | return $variant; |
| 343 | } |
| 344 | // Browsers are supposed to use BCP 47 standard in the |
| 345 | // Accept-Language header, but not all of our internal |
| 346 | // mediawiki variant codes are BCP 47. Map BCP 47 code |
| 347 | // to our internal code. |
| 348 | foreach ( $this->getVariants() as $v ) { |
| 349 | // Case-insensitive match (BCP 47 is mixed-case) |
| 350 | if ( strtolower( LanguageCode::bcp47( $v ) ) === $variant ) { |
| 351 | return $v; |
| 352 | } |
| 353 | } |
| 354 | return null; |
| 355 | } |
| 356 | |
| 357 | public function getURLVariant() { |
| 358 | if ( $this->mURLVariant ) { |
| 359 | return $this->mURLVariant; |
| 360 | } |
| 361 | |
| 362 | $request = RequestContext::getMain()->getRequest(); |
| 363 | // see if the preference is set in the request |
| 364 | $ret = $request->getText( 'variant' ); |
| 365 | |
| 366 | if ( !$ret ) { |
| 367 | $ret = $request->getVal( 'uselang' ); |
| 368 | } |
| 369 | |
| 370 | $this->mURLVariant = $this->validateVariant( $ret ); |
| 371 | return $this->mURLVariant; |
| 372 | } |
| 373 | |
| 374 | /** |
| 375 | * Determine if the user has a variant set. |
| 376 | * |
| 377 | * @param User $user |
| 378 | * @return string|null Variant if one found, null otherwise |
| 379 | */ |
| 380 | protected function getUserVariant( User $user ) { |
| 381 | // This should only be called within the class after the user is known to be |
| 382 | // safe to load and logged in, but check just in case. |
| 383 | if ( !$user->isSafeToLoad() ) { |
| 384 | return null; |
| 385 | } |
| 386 | |
| 387 | if ( !$this->mUserVariant ) { |
| 388 | $services = MediaWikiServices::getInstance(); |
| 389 | if ( $user->isRegistered() ) { |
| 390 | // Get language variant preference from logged in users |
| 391 | if ( |
| 392 | $this->getMainCode() === |
| 393 | $services->getContentLanguage()->getCode() |
| 394 | ) { |
| 395 | $optionName = 'variant'; |
| 396 | } else { |
| 397 | $optionName = 'variant-' . $this->getMainCode(); |
| 398 | } |
| 399 | } else { |
| 400 | // figure out user lang without constructing wgLang to avoid |
| 401 | // infinite recursion |
| 402 | $optionName = 'language'; |
| 403 | } |
| 404 | $ret = $services->getUserOptionsLookup()->getOption( $user, $optionName ); |
| 405 | |
| 406 | $this->mUserVariant = $this->validateVariant( $ret ); |
| 407 | } |
| 408 | |
| 409 | return $this->mUserVariant; |
| 410 | } |
| 411 | |
| 412 | /** |
| 413 | * Determine the language variant from the Accept-Language header. |
| 414 | * |
| 415 | * @return string|null Variant if one found, null otherwise |
| 416 | */ |
| 417 | protected function getHeaderVariant() { |
| 418 | if ( $this->mHeaderVariant ) { |
| 419 | return $this->mHeaderVariant; |
| 420 | } |
| 421 | |
| 422 | $request = RequestContext::getMain()->getRequest(); |
| 423 | // See if some supported language variant is set in the |
| 424 | // HTTP header. |
| 425 | $languages = array_keys( $request->getAcceptLang() ); |
| 426 | if ( !$languages ) { |
| 427 | return null; |
| 428 | } |
| 429 | |
| 430 | $fallbackLanguages = []; |
| 431 | foreach ( $languages as $language ) { |
| 432 | $this->mHeaderVariant = $this->validateVariant( $language ); |
| 433 | if ( $this->mHeaderVariant ) { |
| 434 | break; |
| 435 | } |
| 436 | |
| 437 | // To see if there are fallbacks of current language. |
| 438 | // We record these fallback variants, and process |
| 439 | // them later. |
| 440 | $fallbacks = $this->getVariantFallbacks( $language ); |
| 441 | if ( |
| 442 | is_string( $fallbacks ) && |
| 443 | $fallbacks !== $this->getStaticDefaultVariant() |
| 444 | ) { |
| 445 | $fallbackLanguages[] = $fallbacks; |
| 446 | } elseif ( is_array( $fallbacks ) ) { |
| 447 | $fallbackLanguages = |
| 448 | array_merge( $fallbackLanguages, $fallbacks ); |
| 449 | } |
| 450 | } |
| 451 | |
| 452 | if ( !$this->mHeaderVariant ) { |
| 453 | // process fallback languages now |
| 454 | $fallback_languages = array_unique( $fallbackLanguages ); |
| 455 | foreach ( $fallback_languages as $language ) { |
| 456 | $this->mHeaderVariant = $this->validateVariant( $language ); |
| 457 | if ( $this->mHeaderVariant ) { |
| 458 | break; |
| 459 | } |
| 460 | } |
| 461 | } |
| 462 | |
| 463 | return $this->mHeaderVariant; |
| 464 | } |
| 465 | |
| 466 | public function autoConvert( $text, $toVariant = false ) { |
| 467 | $this->loadTables(); |
| 468 | |
| 469 | if ( !$toVariant ) { |
| 470 | $toVariant = $this->getPreferredVariant(); |
| 471 | if ( !$toVariant ) { |
| 472 | return $text; |
| 473 | } |
| 474 | } |
| 475 | |
| 476 | if ( $this->guessVariant( $text, $toVariant ) ) { |
| 477 | return $text; |
| 478 | } |
| 479 | /** |
| 480 | * We convert everything except: |
| 481 | * 1. HTML markups (anything between < and >) |
| 482 | * 2. HTML entities |
| 483 | * 3. placeholders created by the parser |
| 484 | * IMPORTANT: Beware of failure from pcre.backtrack_limit (T124404). |
| 485 | * Minimize the use of backtracking where possible. |
| 486 | */ |
| 487 | static $reg; |
| 488 | if ( $reg === null ) { |
| 489 | $marker = '|' . Parser::MARKER_PREFIX . '[^\x7f]++\x7f'; |
| 490 | |
| 491 | // this one is needed when the text is inside an HTML markup |
| 492 | $htmlfix = '|<[^>\004]++(?=\004$)|^[^<>]*+>'; |
| 493 | |
| 494 | // Optimize for the common case where these tags have |
| 495 | // few or no children. Thus try and possessively get as much as |
| 496 | // possible, and only engage in backtracking when we hit a '<'. |
| 497 | |
| 498 | // disable convert to variants between <code> tags |
| 499 | $codefix = '<code>[^<]*+(?:(?:(?!<\/code>).)[^<]*+)*+<\/code>|'; |
| 500 | // disable conversion of <script> tags |
| 501 | $scriptfix = '<script[^>]*+>[^<]*+(?:(?:(?!<\/script>).)[^<]*+)*+<\/script>|'; |
| 502 | // disable conversion of <pre> tags |
| 503 | $prefix = '<pre[^>]*+>[^<]*+(?:(?:(?!<\/pre>).)[^<]*+)*+<\/pre>|'; |
| 504 | // The "|.*+)" at the end, is in case we missed some part of html syntax, |
| 505 | // we will fail securely (hopefully) by matching the rest of the string. |
| 506 | $htmlFullTag = '<(?:[^>=]*+(?>[^>=]*+=\s*+(?:"[^"]*"|\'[^\']*\'|[^\'">\s]*+))*+[^>=]*+>|.*+)|'; |
| 507 | |
| 508 | $reg = '/' . $codefix . $scriptfix . $prefix . $htmlFullTag . |
| 509 | '&[a-zA-Z#][a-z0-9]++;' . $marker . $htmlfix . '|\004$/s'; |
| 510 | } |
| 511 | $startPos = 0; |
| 512 | $sourceBlob = ''; |
| 513 | $literalBlob = ''; |
| 514 | |
| 515 | // Guard against delimiter nulls in the input |
| 516 | // (should never happen: see T159174) |
| 517 | $text = str_replace( "\000", '', $text ); |
| 518 | $text = str_replace( "\004", '', $text ); |
| 519 | |
| 520 | $markupMatches = null; |
| 521 | $elementMatches = null; |
| 522 | |
| 523 | // We add a marker (\004) at the end of text, to ensure we always match the |
| 524 | // entire text (Otherwise, pcre.backtrack_limit might cause silent failure) |
| 525 | $textWithMarker = $text . "\004"; |
| 526 | while ( $startPos < strlen( $text ) ) { |
| 527 | if ( preg_match( $reg, $textWithMarker, $markupMatches, PREG_OFFSET_CAPTURE, $startPos ) ) { |
| 528 | $elementPos = $markupMatches[0][1]; |
| 529 | $element = $markupMatches[0][0]; |
| 530 | if ( $element === "\004" ) { |
| 531 | // We hit the end. |
| 532 | $elementPos = strlen( $text ); |
| 533 | $element = ''; |
| 534 | } elseif ( substr( $element, -1 ) === "\004" ) { |
| 535 | // This can sometimes happen if we have |
| 536 | // unclosed html tags. For example, |
| 537 | // when converting a title attribute |
| 538 | // during a recursive call that contains |
| 539 | // a < e.g. <div title="<">. |
| 540 | $element = substr( $element, 0, -1 ); |
| 541 | } |
| 542 | } else { |
| 543 | // If we hit here, then Language Converter could be tricked |
| 544 | // into doing an XSS, so we refuse to translate. |
| 545 | // If expected input manages to reach this code path, |
| 546 | // we should consider it a bug. |
| 547 | $log = LoggerFactory::getInstance( 'languageconverter' ); |
| 548 | $log->error( "Hit pcre.backtrack_limit in " . __METHOD__ |
| 549 | . ". Disabling language conversion for this page.", |
| 550 | [ |
| 551 | "method" => __METHOD__, |
| 552 | "variant" => $toVariant, |
| 553 | "startOfText" => substr( $text, 0, 500 ) |
| 554 | ] |
| 555 | ); |
| 556 | return $text; |
| 557 | } |
| 558 | // Queue the part before the markup for translation in a batch |
| 559 | $sourceBlob .= substr( $text, $startPos, $elementPos - $startPos ) . "\000"; |
| 560 | |
| 561 | // Advance to the next position |
| 562 | $startPos = $elementPos + strlen( $element ); |
| 563 | |
| 564 | // Translate any alt or title attributes inside the matched element |
| 565 | if ( $element !== '' |
| 566 | && preg_match( '/^(<[^>\s]*+)\s([^>]*+)(.*+)$/', $element, $elementMatches ) |
| 567 | ) { |
| 568 | // FIXME, this decodes entities, so if you have something |
| 569 | // like <div title="foo<bar"> the bar won't get |
| 570 | // translated since after entity decoding it looks like |
| 571 | // unclosed html and we call this method recursively |
| 572 | // on attributes. |
| 573 | $attrs = Sanitizer::decodeTagAttributes( $elementMatches[2] ); |
| 574 | // Ensure self-closing tags stay self-closing. |
| 575 | $close = substr( $elementMatches[2], -1 ) === '/' ? ' /' : ''; |
| 576 | $changed = false; |
| 577 | foreach ( [ 'title', 'alt' ] as $attrName ) { |
| 578 | if ( !isset( $attrs[$attrName] ) ) { |
| 579 | continue; |
| 580 | } |
| 581 | $attr = $attrs[$attrName]; |
| 582 | // Don't convert URLs |
| 583 | if ( !str_contains( $attr, '://' ) ) { |
| 584 | $attr = $this->recursiveConvertTopLevel( $attr, $toVariant ); |
| 585 | } |
| 586 | |
| 587 | if ( $attr !== $attrs[$attrName] ) { |
| 588 | $attrs[$attrName] = $attr; |
| 589 | $changed = true; |
| 590 | } |
| 591 | } |
| 592 | if ( $changed ) { |
| 593 | // @phan-suppress-next-line SecurityCheck-DoubleEscaped Explained above with decodeTagAttributes |
| 594 | $element = $elementMatches[1] . Html::expandAttributes( $attrs ) . |
| 595 | $close . $elementMatches[3]; |
| 596 | } |
| 597 | } |
| 598 | $literalBlob .= $element . "\000"; |
| 599 | } |
| 600 | |
| 601 | // Do the main translation batch |
| 602 | $translatedBlob = $this->translate( $sourceBlob, $toVariant ); |
| 603 | |
| 604 | // Put the output back together |
| 605 | $translatedIter = StringUtils::explode( "\000", $translatedBlob ); |
| 606 | $literalIter = StringUtils::explode( "\000", $literalBlob ); |
| 607 | $output = ''; |
| 608 | while ( $translatedIter->valid() && $literalIter->valid() ) { |
| 609 | $output .= $translatedIter->current(); |
| 610 | $output .= $literalIter->current(); |
| 611 | $translatedIter->next(); |
| 612 | $literalIter->next(); |
| 613 | } |
| 614 | |
| 615 | return $output; |
| 616 | } |
| 617 | |
| 618 | public function translate( $text, $variant ) { |
| 619 | // If $text is empty or only includes spaces, do nothing |
| 620 | // Otherwise translate it |
| 621 | if ( trim( $text ) ) { |
| 622 | $this->loadTables(); |
| 623 | $text = $this->mTables[$variant]->replace( $text ); |
| 624 | } |
| 625 | return $text; |
| 626 | } |
| 627 | |
| 628 | /** |
| 629 | * @param string $text Text to convert |
| 630 | * @param string $variant Variant language code |
| 631 | * @return string Translated text |
| 632 | */ |
| 633 | protected function translateWithoutRomanNumbers( $text, $variant ) { |
| 634 | $breaks = '[^\w\x80-\xff]'; |
| 635 | |
| 636 | // regexp for roman numbers |
| 637 | // Lookahead assertion ensures $roman doesn't match the empty string |
| 638 | $roman = '(?=[MDCLXVI])M{0,4}(C[DM]|D?C{0,3})(X[LC]|L?X{0,3})(I[VX]|V?I{0,3})'; |
| 639 | |
| 640 | $reg = '/^' . $roman . '$|^' . $roman . $breaks . '|' . $breaks |
| 641 | . $roman . '$|' . $breaks . $roman . $breaks . '/'; |
| 642 | |
| 643 | $matches = preg_split( $reg, $text, -1, PREG_SPLIT_OFFSET_CAPTURE ); |
| 644 | |
| 645 | $m = array_shift( $matches ); |
| 646 | $this->loadTables(); |
| 647 | if ( !isset( $this->mTables[$variant] ) ) { |
| 648 | throw new RuntimeException( "Broken variant table: " |
| 649 | . implode( ',', array_keys( $this->mTables ) ) ); |
| 650 | } |
| 651 | $ret = $this->mTables[$variant]->replace( $m[0] ); |
| 652 | $mstart = (int)$m[1] + strlen( $m[0] ); |
| 653 | foreach ( $matches as $m ) { |
| 654 | $ret .= substr( $text, $mstart, (int)$m[1] - $mstart ); |
| 655 | $ret .= $this->translate( $m[0], $variant ); |
| 656 | $mstart = (int)$m[1] + strlen( $m[0] ); |
| 657 | } |
| 658 | |
| 659 | return $ret; |
| 660 | } |
| 661 | |
| 662 | public function autoConvertToAllVariants( $text ) { |
| 663 | $this->loadTables(); |
| 664 | |
| 665 | $ret = []; |
| 666 | foreach ( $this->getVariants() as $variant ) { |
| 667 | $ret[$variant] = $this->translate( $text, $variant ); |
| 668 | } |
| 669 | |
| 670 | return $ret; |
| 671 | } |
| 672 | |
| 673 | /** |
| 674 | &n |