Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
| Total | |
95.75% |
586 / 612 |
|
61.90% |
13 / 21 |
CRAP | |
0.00% |
0 / 1 |
| CommentParser | |
95.75% |
586 / 612 |
|
61.90% |
13 / 21 |
201 | |
0.00% |
0 / 1 |
| __construct | |
100.00% |
11 / 11 |
|
100.00% |
1 / 1 |
1 | |||
| parse | |
100.00% |
6 / 6 |
|
100.00% |
1 / 1 |
1 | |||
| nextInterestingLeafNode | |
95.24% |
20 / 21 |
|
0.00% |
0 / 1 |
8 | |||
| regexpAlternateGroup | |
100.00% |
3 / 3 |
|
100.00% |
1 / 1 |
1 | |||
| getMessages | |
100.00% |
3 / 3 |
|
100.00% |
1 / 1 |
1 | |||
| getTimestampRegexp | |
88.54% |
85 / 96 |
|
0.00% |
0 / 1 |
32.45 | |||
| getTimestampParser | |
93.33% |
126 / 135 |
|
0.00% |
0 / 1 |
51.77 | |||
| getLocalTimestampRegexps | |
100.00% |
9 / 9 |
|
100.00% |
1 / 1 |
1 | |||
| getLocalTimestampParsers | |
100.00% |
10 / 10 |
|
100.00% |
1 / 1 |
1 | |||
| getUsernameFromLink | |
96.97% |
32 / 33 |
|
0.00% |
0 / 1 |
15 | |||
| findSignature | |
100.00% |
42 / 42 |
|
100.00% |
1 / 1 |
15 | |||
| acceptOnlyNodesAllowingComments | |
100.00% |
18 / 18 |
|
100.00% |
1 / 1 |
11 | |||
| getCodepointOffset | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| findTimestamp | |
100.00% |
45 / 45 |
|
100.00% |
1 / 1 |
11 | |||
| adjustSigRange | |
100.00% |
14 / 14 |
|
100.00% |
1 / 1 |
3 | |||
| buildThreadItems | |
98.95% |
94 / 95 |
|
0.00% |
0 / 1 |
21 | |||
| truncateForId | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| computeId | |
96.67% |
29 / 30 |
|
0.00% |
0 / 1 |
12 | |||
| computeName | |
91.67% |
11 / 12 |
|
0.00% |
0 / 1 |
4.01 | |||
| buildThreads | |
95.24% |
20 / 21 |
|
0.00% |
0 / 1 |
9 | |||
| computeIdsAndNames | |
100.00% |
6 / 6 |
|
100.00% |
1 / 1 |
2 | |||
| 1 | <?php |
| 2 | |
| 3 | namespace MediaWiki\Extension\DiscussionTools; |
| 4 | |
| 5 | use Config; |
| 6 | use DateInterval; |
| 7 | use DateTime; |
| 8 | use DateTimeImmutable; |
| 9 | use DateTimeZone; |
| 10 | use InvalidArgumentException; |
| 11 | use Language; |
| 12 | use LogicException; |
| 13 | use MalformedTitleException; |
| 14 | use MediaWiki\Extension\DiscussionTools\ThreadItem\ContentCommentItem; |
| 15 | use MediaWiki\Extension\DiscussionTools\ThreadItem\ContentHeadingItem; |
| 16 | use MediaWiki\Extension\DiscussionTools\ThreadItem\ContentThreadItem; |
| 17 | use MediaWiki\Languages\LanguageConverterFactory; |
| 18 | use RuntimeException; |
| 19 | use TitleParser; |
| 20 | use TitleValue; |
| 21 | use Wikimedia\Assert\Assert; |
| 22 | use Wikimedia\IPUtils; |
| 23 | use Wikimedia\Parsoid\DOM\Element; |
| 24 | use Wikimedia\Parsoid\DOM\Node; |
| 25 | use Wikimedia\Parsoid\DOM\Text; |
| 26 | use Wikimedia\Parsoid\Utils\DOMCompat; |
| 27 | |
| 28 | // TODO consider making timestamp parsing not a returned function |
| 29 | |
| 30 | class CommentParser { |
| 31 | |
| 32 | /** |
| 33 | * How far backwards we look for a signature associated with a timestamp before giving up. |
| 34 | * Note that this is not a hard limit on the length of signatures we detect. |
| 35 | */ |
| 36 | private const SIGNATURE_SCAN_LIMIT = 100; |
| 37 | |
| 38 | private Config $config; |
| 39 | private Language $language; |
| 40 | private LanguageConverterFactory $languageConverterFactory; |
| 41 | private TitleParser $titleParser; |
| 42 | |
| 43 | private $dateFormat; |
| 44 | private $digits; |
| 45 | /** @var string[][] */ |
| 46 | private $contLangMessages; |
| 47 | private $localTimezone; |
| 48 | private $timezones; |
| 49 | private $specialContributionsName; |
| 50 | |
| 51 | private Element $rootNode; |
| 52 | private TitleValue $title; |
| 53 | |
| 54 | /** |
| 55 | * @param Config $config |
| 56 | * @param Language $language Content language |
| 57 | * @param LanguageConverterFactory $languageConverterFactory |
| 58 | * @param LanguageData $languageData |
| 59 | * @param TitleParser $titleParser |
| 60 | */ |
| 61 | public function __construct( |
| 62 | Config $config, |
| 63 | Language $language, |
| 64 | LanguageConverterFactory $languageConverterFactory, |
| 65 | LanguageData $languageData, |
| 66 | TitleParser $titleParser |
| 67 | ) { |
| 68 | $this->config = $config; |
| 69 | $this->language = $language; |
| 70 | $this->languageConverterFactory = $languageConverterFactory; |
| 71 | $this->titleParser = $titleParser; |
| 72 | |
| 73 | $data = $languageData->getLocalData(); |
| 74 | $this->dateFormat = $data['dateFormat']; |
| 75 | $this->digits = $data['digits']; |
| 76 | $this->contLangMessages = $data['contLangMessages']; |
| 77 | $this->localTimezone = $data['localTimezone']; |
| 78 | $this->timezones = $data['timezones']; |
| 79 | $this->specialContributionsName = $data['specialContributionsName']; |
| 80 | } |
| 81 | |
| 82 | /** |
| 83 | * Parse a discussion page. |
| 84 | * |
| 85 | * @param Element $rootNode Root node of content to parse |
| 86 | * @param TitleValue $title Title of the page being parsed |
| 87 | * @return ContentThreadItemSet |
| 88 | */ |
| 89 | public function parse( Element $rootNode, TitleValue $title ): ContentThreadItemSet { |
| 90 | $this->rootNode = $rootNode; |
| 91 | $this->title = $title; |
| 92 | |
| 93 | $result = $this->buildThreadItems(); |
| 94 | $this->buildThreads( $result ); |
| 95 | $this->computeIdsAndNames( $result ); |
| 96 | |
| 97 | return $result; |
| 98 | } |
| 99 | |
| 100 | /** |
| 101 | * Return the next leaf node in the tree order that is likely a part of a discussion comment, |
| 102 | * rather than some boring "separator" element. |
| 103 | * |
| 104 | * Currently, this can return a Text node with content other than whitespace, or an Element node |
| 105 | * that is a "void element" or "text element", except some special cases that we treat as comment |
| 106 | * separators (isCommentSeparator()). |
| 107 | * |
| 108 | * @param Node $node Node to start searching at. This node's children are ignored. |
| 109 | * @return Node |
| 110 | */ |
| 111 | private function nextInterestingLeafNode( Node $node ): Node { |
| 112 | $rootNode = $this->rootNode; |
| 113 | $treeWalker = new TreeWalker( |
| 114 | $rootNode, |
| 115 | NodeFilter::SHOW_ELEMENT | NodeFilter::SHOW_TEXT, |
| 116 | static function ( $n ) use ( $node, $rootNode ) { |
| 117 | // Ignore this node and its descendants |
| 118 | // (unless it's the root node, this is a special case for "fakeHeading" handling) |
| 119 | if ( $node !== $rootNode && ( $n === $node || $n->parentNode === $node ) ) { |
| 120 | return NodeFilter::FILTER_REJECT; |
| 121 | } |
| 122 | // Ignore some elements usually used as separators or headers (and their descendants) |
| 123 | if ( CommentUtils::isCommentSeparator( $n ) ) { |
| 124 | return NodeFilter::FILTER_REJECT; |
| 125 | } |
| 126 | // Ignore nodes with no rendering that mess up our indentation detection |
| 127 | if ( CommentUtils::isRenderingTransparentNode( $n ) ) { |
| 128 | return NodeFilter::FILTER_REJECT; |
| 129 | } |
| 130 | if ( CommentUtils::isCommentContent( $n ) ) { |
| 131 | return NodeFilter::FILTER_ACCEPT; |
| 132 | } |
| 133 | return NodeFilter::FILTER_SKIP; |
| 134 | } |
| 135 | ); |
| 136 | $treeWalker->currentNode = $node; |
| 137 | $treeWalker->nextNode(); |
| 138 | if ( !$treeWalker->currentNode ) { |
| 139 | throw new RuntimeException( 'nextInterestingLeafNode not found' ); |
| 140 | } |
| 141 | return $treeWalker->currentNode; |
| 142 | } |
| 143 | |
| 144 | /** |
| 145 | * @param string[] $values Values to match |
| 146 | * @return string Regular expression |
| 147 | */ |
| 148 | private static function regexpAlternateGroup( array $values ): string { |
| 149 | return '(' . implode( '|', array_map( static function ( string $x ) { |
| 150 | return preg_quote( $x, '/' ); |
| 151 | }, $values ) ) . ')'; |
| 152 | } |
| 153 | |
| 154 | /** |
| 155 | * Get text of localisation messages in content language. |
| 156 | * |
| 157 | * @param string $contLangVariant Content language variant |
| 158 | * @param string[] $messages Message keys |
| 159 | * @return string[] Message values |
| 160 | */ |
| 161 | private function getMessages( string $contLangVariant, array $messages ): array { |
| 162 | return array_map( function ( string $key ) use ( $contLangVariant ) { |
| 163 | return $this->contLangMessages[$contLangVariant][$key]; |
| 164 | }, $messages ); |
| 165 | } |
| 166 | |
| 167 | /** |
| 168 | * Get a regexp that matches timestamps generated using the given date format. |
| 169 | * |
| 170 | * This only supports format characters that are used by the default date format in any of |
| 171 | * MediaWiki's languages, namely: D, d, F, G, H, i, j, l, M, n, Y, xg, xkY (and escape characters), |
| 172 | * and only dates when MediaWiki existed, let's say 2000 onwards (Thai dates before 1941 are |
| 173 | * complicated). |
| 174 | * |
| 175 | * @param string $contLangVariant Content language variant |
| 176 | * @param string $format Date format |
| 177 | * @param string $digitsRegexp Regular expression matching a single localised digit, e.g. '[0-9]' |
| 178 | * @param array $tzAbbrs Associative array mapping localised timezone abbreviations to |
| 179 | * IANA abbreviations, for the local timezone, e.g. [ 'EDT' => 'EDT', 'EST' => 'EST' ] |
| 180 | * @return string Regular expression |
| 181 | */ |
| 182 | private function getTimestampRegexp( |
| 183 | string $contLangVariant, string $format, string $digitsRegexp, array $tzAbbrs |
| 184 | ): string { |
| 185 | $formatLength = strlen( $format ); |
| 186 | $s = ''; |
| 187 | $raw = false; |
| 188 | // Adapted from Language::sprintfDate() |
| 189 | for ( $p = 0; $p < $formatLength; $p++ ) { |
| 190 | $num = false; |
| 191 | $code = $format[ $p ]; |
| 192 | if ( $code === 'x' && $p < $formatLength - 1 ) { |
| 193 | $code .= $format[++$p]; |
| 194 | } |
| 195 | if ( $code === 'xk' && $p < $formatLength - 1 ) { |
| 196 | $code .= $format[++$p]; |
| 197 | } |
| 198 | |
| 199 | switch ( $code ) { |
| 200 | case 'xx': |
| 201 | $s .= 'x'; |
| 202 | break; |
| 203 | case 'xg': |
| 204 | $s .= static::regexpAlternateGroup( |
| 205 | $this->getMessages( $contLangVariant, Language::MONTH_GENITIVE_MESSAGES ) |
| 206 | ); |
| 207 | break; |
| 208 | case 'xn': |
| 209 | $raw = true; |
| 210 | break; |
| 211 | case 'd': |
| 212 | $num = '2'; |
| 213 | break; |
| 214 | case 'D': |
| 215 | $s .= static::regexpAlternateGroup( |
| 216 | $this->getMessages( $contLangVariant, Language::WEEKDAY_ABBREVIATED_MESSAGES ) |
| 217 | ); |
| 218 | break; |
| 219 | case 'j': |
| 220 | $num = '1,2'; |
| 221 | break; |
| 222 | case 'l': |
| 223 | $s .= static::regexpAlternateGroup( |
| 224 | $this->getMessages( $contLangVariant, Language::WEEKDAY_MESSAGES ) |
| 225 | ); |
| 226 | break; |
| 227 | case 'F': |
| 228 | $s .= static::regexpAlternateGroup( |
| 229 | $this->getMessages( $contLangVariant, Language::MONTH_MESSAGES ) |
| 230 | ); |
| 231 | break; |
| 232 | case 'M': |
| 233 | $s .= static::regexpAlternateGroup( |
| 234 | $this->getMessages( $contLangVariant, Language::MONTH_ABBREVIATED_MESSAGES ) |
| 235 | ); |
| 236 | break; |
| 237 | case 'm': |
| 238 | $num = '2'; |
| 239 | break; |
| 240 | case 'n': |
| 241 | $num = '1,2'; |
| 242 | break; |
| 243 | case 'Y': |
| 244 | $num = '4'; |
| 245 | break; |
| 246 | case 'xkY': |
| 247 | $num = '4'; |
| 248 | break; |
| 249 | case 'G': |
| 250 | $num = '1,2'; |
| 251 | break; |
| 252 | case 'H': |
| 253 | $num = '2'; |
| 254 | break; |
| 255 | case 'i': |
| 256 | $num = '2'; |
| 257 | break; |
| 258 | case 's': |
| 259 | $num = '2'; |
| 260 | break; |
| 261 | case '\\': |
| 262 | // Backslash escaping |
| 263 | if ( $p < $formatLength - 1 ) { |
| 264 | $s .= preg_quote( $format[++$p], '/' ); |
| 265 | } else { |
| 266 | $s .= preg_quote( '\\', '/' ); |
| 267 | } |
| 268 | break; |
| 269 | case '"': |
| 270 | // Quoted literal |
| 271 | if ( $p < $formatLength - 1 ) { |
| 272 | $endQuote = strpos( $format, '"', $p + 1 ); |
| 273 | if ( $endQuote === false ) { |
| 274 | // No terminating quote, assume literal " |
| 275 | $s .= '"'; |
| 276 | } else { |
| 277 | $s .= preg_quote( substr( $format, $p + 1, $endQuote - $p - 1 ), '/' ); |
| 278 | $p = $endQuote; |
| 279 | } |
| 280 | } else { |
| 281 | // Quote at end of string, assume literal " |
| 282 | $s .= '"'; |
| 283 | } |
| 284 | break; |
| 285 | default: |
| 286 | // Copy whole characters together, instead of single bytes |
| 287 | $char = mb_substr( mb_strcut( $format, $p, 4 ), 0, 1 ); |
| 288 | $s .= preg_quote( $char, '/' ); |
| 289 | $p += strlen( $char ) - 1; |
| 290 | } |
| 291 | if ( $num !== false ) { |
| 292 | if ( $raw ) { |
| 293 | $s .= '([0-9]{' . $num . '})'; |
| 294 | $raw = false; |
| 295 | } else { |
| 296 | $s .= '(' . $digitsRegexp . '{' . $num . '})'; |
| 297 | } |
| 298 | } |
| 299 | // Ignore some invisible Unicode characters that often sneak into copy-pasted timestamps (T308448) |
| 300 | $s .= '[\\x{200E}\\x{200F}]?'; |
| 301 | } |
| 302 | |
| 303 | $tzRegexp = static::regexpAlternateGroup( array_keys( $tzAbbrs ) ); |
| 304 | |
| 305 | // Hard-coded parentheses and space like in Parser::pstPass2 |
| 306 | // Ignore some invisible Unicode characters that often sneak into copy-pasted timestamps (T245784) |
| 307 | // \uNNNN syntax can only be used from PHP 7.3 |
| 308 | return '/' . $s . ' [\\x{200E}\\x{200F}]?\\(' . $tzRegexp . '\\)/u'; |
| 309 | } |
| 310 | |
| 311 | /** |
| 312 | * Get a function that parses timestamps generated using the given date format, based on the result |
| 313 | * of matching the regexp returned by getTimestampRegexp() |
| 314 | * |
| 315 | * @param string $contLangVariant Content language variant |
| 316 | * @param string $format Date format, as used by MediaWiki |
| 317 | * @param string[]|null $digits Localised digits from 0 to 9, e.g. `[ '0', '1', ..., '9' ]` |
| 318 | * @param string $localTimezone Local timezone IANA name, e.g. `America/New_York` |
| 319 | * @param array $tzAbbrs Map of localised timezone abbreviations to IANA abbreviations |
| 320 | * for the local timezone, e.g. [ 'EDT' => 'EDT', 'EST' => 'EST' ] |
| 321 | * @return callable Parser function |
| 322 | */ |
| 323 | private function getTimestampParser( |
| 324 | string $contLangVariant, string $format, ?array $digits, string $localTimezone, array $tzAbbrs |
| 325 | ): callable { |
| 326 | $untransformDigits = static function ( string $text ) use ( $digits ) { |
| 327 | if ( !$digits ) { |
| 328 | return $text; |
| 329 | } |
| 330 | return preg_replace_callback( |
| 331 | '/[' . implode( '', $digits ) . ']/u', |
| 332 | static function ( array $m ) use ( $digits ) { |
| 333 | return (string)array_search( $m[0], $digits, true ); |
| 334 | }, |
| 335 | $text |
| 336 | ); |
| 337 | }; |
| 338 | |
| 339 | $formatLength = strlen( $format ); |
| 340 | $matchingGroups = []; |
| 341 | for ( $p = 0; $p < $formatLength; $p++ ) { |
| 342 | $code = $format[$p]; |
| 343 | if ( $code === 'x' && $p < $formatLength - 1 ) { |
| 344 | $code .= $format[++$p]; |
| 345 | } |
| 346 | if ( $code === 'xk' && $p < $formatLength - 1 ) { |
| 347 | $code .= $format[++$p]; |
| 348 | } |
| 349 | |
| 350 | switch ( $code ) { |
| 351 | case 'xx': |
| 352 | case 'xn': |
| 353 | break; |
| 354 | case 'xg': |
| 355 | case 'd': |
| 356 | case 'j': |
| 357 | case 'D': |
| 358 | case 'l': |
| 359 | case 'F': |
| 360 | case 'M': |
| 361 | case 'm': |
| 362 | case 'n': |
| 363 | case 'Y': |
| 364 | case 'xkY': |
| 365 | case 'G': |
| 366 | case 'H': |
| 367 | case 'i': |
| 368 | case 's': |
| 369 | $matchingGroups[] = $code; |
| 370 | break; |
| 371 | case '\\': |
| 372 | // Backslash escaping |
| 373 | if ( $p < $formatLength - 1 ) { |
| 374 | $p++; |
| 375 | } |
| 376 | break; |
| 377 | case '"': |
| 378 | // Quoted literal |
| 379 | if ( $p < $formatLength - 1 ) { |
| 380 | $endQuote = strpos( $format, '"', $p + 1 ); |
| 381 | if ( $endQuote !== false ) { |
| 382 | $p = $endQuote; |
| 383 | } |
| 384 | } |
| 385 | break; |
| 386 | default: |
| 387 | break; |
| 388 | } |
| 389 | } |
| 390 | |
| 391 | return function ( array $match ) use ( |
| 392 | $matchingGroups, $untransformDigits, $localTimezone, $tzAbbrs, $contLangVariant |
| 393 | ) { |