includes/parser/BlockLevelPass.php

   1 <?php
   2
   3 /**
   4  * This is the part of the wikitext parser which handles automatic paragraphs
   5  * and conversion of start-of-line prefixes to HTML lists.
   6  *
   7  * This program is free software; you can redistribute it and/or modify
   8  * it under the terms of the GNU General Public License as published by
   9  * the Free Software Foundation; either version 2 of the License, or
  10  * (at your option) any later version.
  11  *
  12  * This program is distributed in the hope that it will be useful,
  13  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  14  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  15  * GNU General Public License for more details.
  16  *
  17  * You should have received a copy of the GNU General Public License along
  18  * with this program; if not, write to the Free Software Foundation, Inc.,
  19  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  20  * http://www.gnu.org/copyleft/gpl.html
  21  *
  22  * @file
  23  * @ingroup Parser
  24  */
  25 class BlockLevelPass {
  26         private $DTopen = false;
  27         private $inPre = false;
  28         private $lastSection = '';
  29         private $linestart;
  30         private $text;
  31
  32         # State constants for the definition list colon extraction
  33         const COLON_STATE_TEXT = 0;
  34         const COLON_STATE_TAG = 1;
  35         const COLON_STATE_TAGSTART = 2;
  36         const COLON_STATE_CLOSETAG = 3;
  37         const COLON_STATE_TAGSLASH = 4;
  38         const COLON_STATE_COMMENT = 5;
  39         const COLON_STATE_COMMENTDASH = 6;
  40         const COLON_STATE_COMMENTDASHDASH = 7;
  41         const COLON_STATE_LC = 8;
  42
  43         /**
  44          * Make lists from lines starting with ':', '*', '#', etc.
  45          *
  46          * @param string $text
  47          * @param bool $lineStart Whether or not this is at the start of a line.
  48          * @return string The lists rendered as HTML
  49          */
  50         public static function doBlockLevels( $text, $lineStart ) {
  51                 $pass = new self( $text, $lineStart );
  52                 return $pass->execute();
  53         }
  54
  55         /**
  56          * Private constructor
  57          */
  58         private function __construct( $text, $lineStart ) {
  59                 $this->text = $text;
  60                 $this->lineStart = $lineStart;
  61         }
  62
  63         /**
  64          * If a pre or p is open, return the corresponding close tag and update
  65          * the state. If no tag is open, return an empty string.
  66          * @return string
  67          */
  68         private function closeParagraph() {
  69                 $result = '';
  70                 if ( $this->lastSection !== '' ) {
  71                         $result = '</' . $this->lastSection . ">\n";
  72                 }
  73                 $this->inPre = false;
  74                 $this->lastSection = '';
  75                 return $result;
  76         }
  77
  78         /**
  79          * getCommon() returns the length of the longest common substring
  80          * of both arguments, starting at the beginning of both.
  81          *
  82          * @param string $st1
  83          * @param string $st2
  84          *
  85          * @return int
  86          */
  87         private function getCommon( $st1, $st2 ) {
  88                 $shorter = min( strlen( $st1 ), strlen( $st2 ) );
  89
  90                 for ( $i = 0; $i < $shorter; ++$i ) {
  91                         if ( $st1[$i] !== $st2[$i] ) {
  92                                 break;
  93                         }
  94                 }
  95                 return $i;
  96         }
  97
  98         /**
  99          * Open the list item element identified by the prefix character.
 100          *
 101          * @param string $char
 102          *
 103          * @return string
 104          */
 105         private function openList( $char ) {
 106                 $result = $this->closeParagraph();
 107
 108                 if ( '*' === $char ) {
 109                         $result .= "<ul><li>";
 110                 } elseif ( '#' === $char ) {
 111                         $result .= "<ol><li>";
 112                 } elseif ( ':' === $char ) {
 113                         $result .= "<dl><dd>";
 114                 } elseif ( ';' === $char ) {
 115                         $result .= "<dl><dt>";
 116                         $this->DTopen = true;
 117                 } else {
 118                         $result = '<!-- ERR 1 -->';
 119                 }
 120
 121                 return $result;
 122         }
 123
 124         /**
 125          * Close the current list item and open the next one.
 126          * @param string $char
 127          *
 128          * @return string
 129          */
 130         private function nextItem( $char ) {
 131                 if ( '*' === $char || '#' === $char ) {
 132                         return "</li>\n<li>";
 133                 } elseif ( ':' === $char || ';' === $char ) {
 134                         $close = "</dd>\n";
 135                         if ( $this->DTopen ) {
 136                                 $close = "</dt>\n";
 137                         }
 138                         if ( ';' === $char ) {
 139                                 $this->DTopen = true;
 140                                 return $close . '<dt>';
 141                         } else {
 142                                 $this->DTopen = false;
 143                                 return $close . '<dd>';
 144                         }
 145                 }
 146                 return '<!-- ERR 2 -->';
 147         }
 148
 149         /**
 150          * Close the current list item identified by the prefix character.
 151          * @param string $char
 152          *
 153          * @return string
 154          */
 155         private function closeList( $char ) {
 156                 if ( '*' === $char ) {
 157                         $text = "</li></ul>";
 158                 } elseif ( '#' === $char ) {
 159                         $text = "</li></ol>";
 160                 } elseif ( ':' === $char ) {
 161                         if ( $this->DTopen ) {
 162                                 $this->DTopen = false;
 163                                 $text = "</dt></dl>";
 164                         } else {
 165                                 $text = "</dd></dl>";
 166                         }
 167                 } else {
 168                         return '<!-- ERR 3 -->';
 169                 }
 170                 return $text;
 171         }
 172
 173         /**
 174          * Execute the pass.
 175          * @return string
 176          */
 177         private function execute() {
 178                 $text = $this->text;
 179                 # Parsing through the text line by line.  The main thing
 180                 # happening here is handling of block-level elements p, pre,
 181                 # and making lists from lines starting with * # : etc.
 182                 $textLines = StringUtils::explode( "\n", $text );
 183
 184                 $lastPrefix = $output = '';
 185                 $this->DTopen = $inBlockElem = false;
 186                 $prefixLength = 0;
 187                 $pendingPTag = false;
 188                 $inBlockquote = false;
 189
 190                 foreach ( $textLines as $inputLine ) {
 191                         # Fix up $lineStart
 192                         if ( !$this->lineStart ) {
 193                                 $output .= $inputLine;
 194                                 $this->lineStart = true;
 195                                 continue;
 196                         }
 197                         # * = ul
 198                         # # = ol
 199                         # ; = dt
 200                         # : = dd
 201
 202                         $lastPrefixLength = strlen( $lastPrefix );
 203                         $preCloseMatch = preg_match( '/<\\/pre/i', $inputLine );
 204                         $preOpenMatch = preg_match( '/<pre/i', $inputLine );
 205                         # If not in a <pre> element, scan for and figure out what prefixes are there.
 206                         if ( !$this->inPre ) {
 207                                 # Multiple prefixes may abut each other for nested lists.
 208                                 $prefixLength = strspn( $inputLine, '*#:;' );
 209                                 $prefix = substr( $inputLine, 0, $prefixLength );
 210
 211                                 # eh?
 212                                 # ; and : are both from definition-lists, so they're equivalent
 213                                 #  for the purposes of determining whether or not we need to open/close
 214                                 #  elements.
 215                                 $prefix2 = str_replace( ';', ':', $prefix );
 216                                 $t = substr( $inputLine, $prefixLength );
 217                                 $this->inPre = (bool)$preOpenMatch;
 218                         } else {
 219                                 # Don't interpret any other prefixes in preformatted text
 220                                 $prefixLength = 0;
 221                                 $prefix = $prefix2 = '';
 222                                 $t = $inputLine;
 223                         }
 224
 225                         # List generation
 226                         if ( $prefixLength && $lastPrefix === $prefix2 ) {
 227                                 # Same as the last item, so no need to deal with nesting or opening stuff
 228                                 $output .= $this->nextItem( substr( $prefix, -1 ) );
 229                                 $pendingPTag = false;
 230
 231                                 if ( substr( $prefix, -1 ) === ';' ) {
 232                                         # The one nasty exception: definition lists work like this:
 233                                         # ; title : definition text
 234                                         # So we check for : in the remainder text to split up the
 235                                         # title and definition, without b0rking links.
 236                                         $term = $t2 = '';
 237                                         if ( $this->findColonNoLinks( $t, $term, $t2 ) !== false ) {
 238                                                 $t = $t2;
 239                                                 $output .= $term . $this->nextItem( ':' );
 240                                         }
 241                                 }
 242                         } elseif ( $prefixLength || $lastPrefixLength ) {
 243                                 # We need to open or close prefixes, or both.
 244
 245                                 # Either open or close a level...
 246                                 $commonPrefixLength = $this->getCommon( $prefix, $lastPrefix );
 247                                 $pendingPTag = false;
 248
 249                                 # Close all the prefixes which aren't shared.
 250                                 while ( $commonPrefixLength < $lastPrefixLength ) {
 251                                         $output .= $this->closeList( $lastPrefix[$lastPrefixLength - 1] );
 252                                         --$lastPrefixLength;
 253                                 }
 254
 255                                 # Continue the current prefix if appropriate.
 256                                 if ( $prefixLength <= $commonPrefixLength && $commonPrefixLength > 0 ) {
 257                                         $output .= $this->nextItem( $prefix[$commonPrefixLength - 1] );
 258                                 }
 259
 260                                 # Open prefixes where appropriate.
 261                                 if ( $lastPrefix && $prefixLength > $commonPrefixLength ) {
 262                                         $output .= "\n";
 263                                 }
 264                                 while ( $prefixLength > $commonPrefixLength ) {
 265                                         $char = substr( $prefix, $commonPrefixLength, 1 );
 266                                         $output .= $this->openList( $char );
 267
 268                                         if ( ';' === $char ) {
 269                                                 # @todo FIXME: This is dupe of code above
 270                                                 if ( $this->findColonNoLinks( $t, $term, $t2 ) !== false ) {
 271                                                         $t = $t2;
 272                                                         $output .= $term . $this->nextItem( ':' );
 273                                                 }
 274                                         }
 275                                         ++$commonPrefixLength;
 276                                 }
 277                                 if ( !$prefixLength && $lastPrefix ) {
 278                                         $output .= "\n";
 279                                 }
 280                                 $lastPrefix = $prefix2;
 281                         }
 282
 283                         # If we have no prefixes, go to paragraph mode.
 284                         if ( 0 == $prefixLength ) {
 285                                 # No prefix (not in list)--go to paragraph mode
 286                                 # @todo consider using a stack for nestable elements like span, table and div
 287                                 $openMatch = preg_match(
 288                                         '/(?:<table|<h1|<h2|<h3|<h4|<h5|<h6|<pre|<tr|'
 289                                                 . '<p|<ul|<ol|<dl|<li|<\\/tr|<\\/td|<\\/th)\\b/iS',
 290                                         $t
 291                                 );
 292                                 $closeMatch = preg_match(
 293                                         '/(?:<\\/table|<\\/h1|<\\/h2|<\\/h3|<\\/h4|<\\/h5|<\\/h6|'
 294                                                 . '<td|<th|<\\/?blockquote|<\\/?div|<hr|<\\/pre|<\\/p|<\\/mw:|'
 295                                                 . Parser::MARKER_PREFIX
 296                                                 . '-pre|<\\/li|<\\/ul|<\\/ol|<\\/dl|<\\/?center)\\b/iS',
 297                                         $t
 298                                 );
 299
 300                                 if ( $openMatch || $closeMatch ) {
 301                                         $pendingPTag = false;
 302                                         # @todo T7718: paragraph closed
 303                                         $output .= $this->closeParagraph();
 304                                         if ( $preOpenMatch && !$preCloseMatch ) {
 305                                                 $this->inPre = true;
 306                                         }
 307                                         $bqOffset = 0;
 308                                         while ( preg_match( '/<(\\/?)blockquote[\s>]/i', $t,
 309                                                 $bqMatch, PREG_OFFSET_CAPTURE, $bqOffset )
 310                                         ) {
 311                                                 $inBlockquote = !$bqMatch[1][0]; // is this a close tag?
 312                                                 $bqOffset = $bqMatch[0][1] + strlen( $bqMatch[0][0] );
 313                                         }
 314                                         $inBlockElem = !$closeMatch;
 315                                 } elseif ( !$inBlockElem && !$this->inPre ) {
 316                                         if ( ' ' == substr( $t, 0, 1 )
 317                                                 && ( $this->lastSection === 'pre' || trim( $t ) != '' )
 318                                                 && !$inBlockquote
 319                                         ) {
 320                                                 # pre
 321                                                 if ( $this->lastSection !== 'pre' ) {
 322                                                         $pendingPTag = false;
 323                                                         $output .= $this->closeParagraph() . '<pre>';
 324                                                         $this->lastSection = 'pre';
 325                                                 }
 326                                                 $t = substr( $t, 1 );
 327                                         } else {
 328                                                 # paragraph
 329                                                 if ( trim( $t ) === '' ) {
 330                                                         if ( $pendingPTag ) {
 331                                                                 $output .= $pendingPTag . '<br />';
 332                                                                 $pendingPTag = false;
 333                                                                 $this->lastSection = 'p';
 334                                                         } else {
 335                                                                 if ( $this->lastSection !== 'p' ) {
 336                                                                         $output .= $this->closeParagraph();
 337                                                                         $this->lastSection = '';
 338                                                                         $pendingPTag = '<p>';
 339                                                                 } else {
 340                                                                         $pendingPTag = '</p><p>';
 341                                                                 }
 342                                                         }
 343                                                 } else {
 344                                                         if ( $pendingPTag ) {
 345                                                                 $output .= $pendingPTag;
 346                                                                 $pendingPTag = false;
 347                                                                 $this->lastSection = 'p';
 348                                                         } elseif ( $this->lastSection !== 'p' ) {
 349                                                                 $output .= $this->closeParagraph() . '<p>';
 350                                                                 $this->lastSection = 'p';
 351                                                         }
 352                                                 }
 353                                         }
 354                                 }
 355                         }
 356                         # somewhere above we forget to get out of pre block (T2785)
 357                         if ( $preCloseMatch && $this->inPre ) {
 358                                 $this->inPre = false;
 359                         }
 360                         if ( $pendingPTag === false ) {
 361                                 $output .= $t;
 362                                 if ( $prefixLength === 0 ) {
 363                                         $output .= "\n";
 364                                 }
 365                         }
 366                 }
 367                 while ( $prefixLength ) {
 368                         $output .= $this->closeList( $prefix2[$prefixLength - 1] );
 369                         --$prefixLength;
 370                         if ( !$prefixLength ) {
 371                                 $output .= "\n";
 372                         }
 373                 }
 374                 if ( $this->lastSection !== '' ) {
 375                         $output .= '</' . $this->lastSection . '>';
 376                         $this->lastSection = '';
 377                 }
 378
 379                 return $output;
 380         }
 381
 382         /**
 383          * Split up a string on ':', ignoring any occurrences inside tags
 384          * to prevent illegal overlapping.
 385          *
 386          * @param string $str The string to split
 387          * @param string &$before Set to everything before the ':'
 388          * @param string &$after Set to everything after the ':'
 389          * @throws MWException
 390          * @return string The position of the ':', or false if none found
 391          */
 392         private function findColonNoLinks( $str, &$before, &$after ) {
 393                 if ( !preg_match( '/:|<|-\{/', $str, $m, PREG_OFFSET_CAPTURE ) ) {
 394                         # Nothing to find!
 395                         return false;
 396                 }
 397
 398                 if ( $m[0][0] === ':' ) {
 399                         # Easy; no tag nesting to worry about
 400                         $colonPos = $m[0][1];
 401                         $before = substr( $str, 0, $colonPos );
 402                         $after = substr( $str, $colonPos + 1 );
 403                         return $colonPos;
 404                 }
 405
 406                 # Ugly state machine to walk through avoiding tags.
 407                 $state = self::COLON_STATE_TEXT;
 408                 $ltLevel = 0;
 409                 $lcLevel = 0;
 410                 $len = strlen( $str );
 411                 for ( $i = $m[0][1]; $i < $len; $i++ ) {
 412                         $c = $str[$i];
 413
 414                         switch ( $state ) {
 415                         case self::COLON_STATE_TEXT:
 416                                 switch ( $c ) {
 417                                 case "<":
 418                                         # Could be either a <start> tag or an </end> tag
 419                                         $state = self::COLON_STATE_TAGSTART;
 420                                         break;
 421                                 case ":":
 422                                         if ( $ltLevel === 0 ) {
 423                                                 # We found it!
 424                                                 $before = substr( $str, 0, $i );
 425                                                 $after = substr( $str, $i + 1 );
 426                                                 return $i;
 427                                         }
 428                                         # Embedded in a tag; don't break it.
 429                                         break;
 430                                 default:
 431                                         # Skip ahead looking for something interesting
 432                                         if ( !preg_match( '/:|<|-\{/', $str, $m, PREG_OFFSET_CAPTURE, $i ) ) {
 433                                                 # Nothing else interesting
 434                                                 return false;
 435                                         }
 436                                         if ( $m[0][0] === '-{' ) {
 437                                                 $state = self::COLON_STATE_LC;
 438                                                 $lcLevel++;
 439                                                 $i = $m[0][1] + 1;
 440                                         } else {
 441                                                 # Skip ahead to next interesting character.
 442                                                 $i = $m[0][1] - 1;
 443                                         }
 444                                         break;
 445                                 }
 446                                 break;
 447                         case self::COLON_STATE_LC:
 448                                 # In language converter markup -{ ... }-
 449                                 if ( !preg_match( '/-\{|\}-/', $str, $m, PREG_OFFSET_CAPTURE, $i ) ) {
 450                                         # Nothing else interesting to find; abort!
 451                                         # We're nested in language converter markup, but there
 452                                         # are no close tags left.  Abort!
 453                                         break 2;
 454                                 } elseif ( $m[0][0] === '-{' ) {
 455                                         $i = $m[0][1] + 1;
 456                                         $lcLevel++;
 457                                 } elseif ( $m[0][0] === '}-' ) {
 458                                         $i = $m[0][1] + 1;
 459                                         $lcLevel--;
 460                                         if ( $lcLevel === 0 ) {
 461                                                 $state = self::COLON_STATE_TEXT;
 462                                         }
 463                                 }
 464                                 break;
 465                         case self::COLON_STATE_TAG:
 466                                 # In a <tag>
 467                                 switch ( $c ) {
 468                                 case ">":
 469                                         $ltLevel++;
 470                                         $state = self::COLON_STATE_TEXT;
 471                                         break;
 472                                 case "/":
 473                                         # Slash may be followed by >?
 474                                         $state = self::COLON_STATE_TAGSLASH;
 475                                         break;
 476                                 default:
 477                                         # ignore
 478                                 }
 479                                 break;
 480                         case self::COLON_STATE_TAGSTART:
 481                                 switch ( $c ) {
 482                                 case "/":
 483                                         $state = self::COLON_STATE_CLOSETAG;
 484                                         break;
 485                                 case "!":
 486                                         $state = self::COLON_STATE_COMMENT;
 487                                         break;
 488                                 case ">":
 489                                         # Illegal early close? This shouldn't happen D:
 490                                         $state = self::COLON_STATE_TEXT;
 491                                         break;
 492                                 default:
 493                                         $state = self::COLON_STATE_TAG;
 494                                 }
 495                                 break;
 496                         case self::COLON_STATE_CLOSETAG:
 497                                 # In a </tag>
 498                                 if ( $c === ">" ) {
 499                                         if ( $ltLevel > 0 ) {
 500                                                 $ltLevel--;
 501                                         } else {
 502                                                 # ignore the excess close tag, but keep looking for
 503                                                 # colons. (This matches Parsoid behavior.)
 504                                                 wfDebug( __METHOD__ . ": Invalid input; too many close tags\n" );
 505                                         }
 506                                         $state = self::COLON_STATE_TEXT;
 507                                 }
 508                                 break;
 509                         case self::COLON_STATE_TAGSLASH:
 510                                 if ( $c === ">" ) {
 511                                         # Yes, a self-closed tag <blah/>
 512                                         $state = self::COLON_STATE_TEXT;
 513                                 } else {
 514                                         # Probably we're jumping the gun, and this is an attribute
 515                                         $state = self::COLON_STATE_TAG;
 516                                 }
 517                                 break;
 518                         case self::COLON_STATE_COMMENT:
 519                                 if ( $c === "-" ) {
 520                                         $state = self::COLON_STATE_COMMENTDASH;
 521                                 }
 522                                 break;
 523                         case self::COLON_STATE_COMMENTDASH:
 524                                 if ( $c === "-" ) {
 525                                         $state = self::COLON_STATE_COMMENTDASHDASH;
 526                                 } else {
 527                                         $state = self::COLON_STATE_COMMENT;
 528                                 }
 529                                 break;
 530                         case self::COLON_STATE_COMMENTDASHDASH:
 531                                 if ( $c === ">" ) {
 532                                         $state = self::COLON_STATE_TEXT;
 533                                 } else {
 534                                         $state = self::COLON_STATE_COMMENT;
 535                                 }
 536                                 break;
 537                         default:
 538                                 throw new MWException( "State machine error in " . __METHOD__ );
 539                         }
 540                 }
 541                 if ( $ltLevel > 0 || $lcLevel > 0 ) {
 542                         wfDebug(
 543                                 __METHOD__ . ": Invalid input; not enough close tags " .
 544                                 "(level $ltLevel/$lcLevel, state $state)\n"
 545                         );
 546                         return false;
 547                 }
 548                 return false;
 549         }
 550 }