includes/content/TextContent.php

   1 <?php
   2 /**
   3  * Content object implementation for representing flat text.
   4  *
   5  * This program is free software; you can redistribute it and/or modify
   6  * it under the terms of the GNU General Public License as published by
   7  * the Free Software Foundation; either version 2 of the License, or
   8  * (at your option) any later version.
   9  *
  10  * This program is distributed in the hope that it will be useful,
  11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  13  * GNU General Public License for more details.
  14  *
  15  * You should have received a copy of the GNU General Public License along
  16  * with this program; if not, write to the Free Software Foundation, Inc.,
  17  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  18  * http://www.gnu.org/copyleft/gpl.html
  19  *
  20  * @since 1.21
  21  *
  22  * @file
  23  * @ingroup Content
  24  *
  25  * @author Daniel Kinzler
  26  */
  27
  28 /**
  29  * Content object implementation for representing flat text.
  30  *
  31  * TextContent instances are immutable
  32  *
  33  * @ingroup Content
  34  */
  35 class TextContent extends AbstractContent {
  36
  37         /**
  38          * @var string
  39          */
  40         protected $mText;
  41
  42         /**
  43          * @param string $text
  44          * @param string $model_id
  45          * @throws MWException
  46          */
  47         public function __construct( $text, $model_id = CONTENT_MODEL_TEXT ) {
  48                 parent::__construct( $model_id );
  49
  50                 if ( $text === null || $text === false ) {
  51                         wfWarn( "TextContent constructed with \$text = " . var_export( $text, true ) . "! "
  52                                 . "This may indicate an error in the caller's scope.", 2 );
  53
  54                         $text = '';
  55                 }
  56
  57                 if ( !is_string( $text ) ) {
  58                         throw new MWException( "TextContent expects a string in the constructor." );
  59                 }
  60
  61                 $this->mText = $text;
  62         }
  63
  64         /**
  65          * @note Mutable subclasses MUST override this to return a copy!
  66          *
  67          * @return Content $this
  68          */
  69         public function copy() {
  70                 return $this; # NOTE: this is ok since TextContent are immutable.
  71         }
  72
  73         public function getTextForSummary( $maxlength = 250 ) {
  74                 global $wgContLang;
  75
  76                 $text = $this->getNativeData();
  77
  78                 $truncatedtext = $wgContLang->truncate(
  79                         preg_replace( "/[\n\r]/", ' ', $text ),
  80                         max( 0, $maxlength ) );
  81
  82                 return $truncatedtext;
  83         }
  84
  85         /**
  86          * Returns the text's size in bytes.
  87          *
  88          * @return int
  89          */
  90         public function getSize() {
  91                 $text = $this->getNativeData();
  92
  93                 return strlen( $text );
  94         }
  95
  96         /**
  97          * Returns true if this content is not a redirect, and $wgArticleCountMethod
  98          * is "any".
  99          *
 100          * @param bool|null $hasLinks If it is known whether this content contains links,
 101          * provide this information here, to avoid redundant parsing to find out.
 102          *
 103          * @return bool
 104          */
 105         public function isCountable( $hasLinks = null ) {
 106                 global $wgArticleCountMethod;
 107
 108                 if ( $this->isRedirect() ) {
 109                         return false;
 110                 }
 111
 112                 if ( $wgArticleCountMethod === 'any' ) {
 113                         return true;
 114                 }
 115
 116                 return false;
 117         }
 118
 119         /**
 120          * Returns the text represented by this Content object, as a string.
 121          *
 122          * @return string The raw text.
 123          */
 124         public function getNativeData() {
 125                 return $this->mText;
 126         }
 127
 128         /**
 129          * Returns the text represented by this Content object, as a string.
 130          *
 131          * @return string The raw text.
 132          */
 133         public function getTextForSearchIndex() {
 134                 return $this->getNativeData();
 135         }
 136
 137         /**
 138          * Returns attempts to convert this content object to wikitext,
 139          * and then returns the text string. The conversion may be lossy.
 140          *
 141          * @note this allows any text-based content to be transcluded as if it was wikitext.
 142          *
 143          * @return string|bool The raw text, or false if the conversion failed.
 144          */
 145         public function getWikitextForTransclusion() {
 146                 $wikitext = $this->convert( CONTENT_MODEL_WIKITEXT, 'lossy' );
 147
 148                 if ( $wikitext ) {
 149                         return $wikitext->getNativeData();
 150                 } else {
 151                         return false;
 152                 }
 153         }
 154
 155         /**
 156          * Do a "\r\n" -> "\n" and "\r" -> "\n" transformation
 157          * as well as trim trailing whitespace
 158          *
 159          * This was formerly part of Parser::preSaveTransform, but
 160          * for non-wikitext content models they probably still want
 161          * to normalize line endings without all of the other PST
 162          * changes.
 163          *
 164          * @since 1.28
 165          * @param string $text
 166          * @return string
 167          */
 168         public static function normalizeLineEndings( $text ) {
 169                 return str_replace( [ "\r\n", "\r" ], "\n", rtrim( $text ) );
 170         }
 171
 172         /**
 173          * Returns a Content object with pre-save transformations applied.
 174          *
 175          * At a minimum, subclasses should make sure to call TextContent::normalizeLineEndings()
 176          * either directly or part of Parser::preSaveTransform().
 177          *
 178          * @param Title $title
 179          * @param User $user
 180          * @param ParserOptions $popts
 181          *
 182          * @return Content
 183          */
 184         public function preSaveTransform( Title $title, User $user, ParserOptions $popts ) {
 185                 $text = $this->getNativeData();
 186                 $pst = self::normalizeLineEndings( $text );
 187
 188                 return ( $text === $pst ) ? $this : new static( $pst, $this->getModel() );
 189         }
 190
 191         /**
 192          * Diff this content object with another content object.
 193          *
 194          * @since 1.21
 195          *
 196          * @param Content $that The other content object to compare this content object to.
 197          * @param Language $lang The language object to use for text segmentation.
 198          *    If not given, $wgContentLang is used.
 199          *
 200          * @return Diff A diff representing the changes that would have to be
 201          *    made to this content object to make it equal to $that.
 202          */
 203         public function diff( Content $that, Language $lang = null ) {
 204                 global $wgContLang;
 205
 206                 $this->checkModelID( $that->getModel() );
 207
 208                 // @todo could implement this in DifferenceEngine and just delegate here?
 209
 210                 if ( !$lang ) {
 211                         $lang = $wgContLang;
 212                 }
 213
 214                 $otext = $this->getNativeData();
 215                 $ntext = $that->getNativeData();
 216
 217                 # Note: Use native PHP diff, external engines don't give us abstract output
 218                 $ota = explode( "\n", $lang->segmentForDiff( $otext ) );
 219                 $nta = explode( "\n", $lang->segmentForDiff( $ntext ) );
 220
 221                 $diff = new Diff( $ota, $nta );
 222
 223                 return $diff;
 224         }
 225
 226         /**
 227          * Fills the provided ParserOutput object with information derived from the content.
 228          * Unless $generateHtml was false, this includes an HTML representation of the content
 229          * provided by getHtml().
 230          *
 231          * For content models listed in $wgTextModelsToParse, this method will call the MediaWiki
 232          * wikitext parser on the text to extract any (wikitext) links, magic words, etc.
 233          *
 234          * Subclasses may override this to provide custom content processing.
 235          * For custom HTML generation alone, it is sufficient to override getHtml().
 236          *
 237          * @param Title $title Context title for parsing
 238          * @param int $revId Revision ID (for {{REVISIONID}})
 239          * @param ParserOptions $options Parser options
 240          * @param bool $generateHtml Whether or not to generate HTML
 241          * @param ParserOutput &$output The output object to fill (reference).
 242          */
 243         protected function fillParserOutput( Title $title, $revId,
 244                 ParserOptions $options, $generateHtml, ParserOutput &$output
 245         ) {
 246                 global $wgParser, $wgTextModelsToParse;
 247
 248                 if ( in_array( $this->getModel(), $wgTextModelsToParse ) ) {
 249                         // parse just to get links etc into the database, HTML is replaced below.
 250                         $output = $wgParser->parse( $this->getNativeData(), $title, $options, true, true, $revId );
 251                 }
 252
 253                 if ( $generateHtml ) {
 254                         $html = $this->getHtml();
 255                 } else {
 256                         $html = '';
 257                 }
 258
 259                 $output->setText( $html );
 260         }
 261
 262         /**
 263          * Generates an HTML version of the content, for display. Used by
 264          * fillParserOutput() to provide HTML for the ParserOutput object.
 265          *
 266          * Subclasses may override this to provide a custom HTML rendering.
 267          * If further information is to be derived from the content (such as
 268          * categories), the fillParserOutput() method can be overridden instead.
 269          *
 270          * For backwards-compatibility, this default implementation just calls
 271          * getHighlightHtml().
 272          *
 273          * @return string An HTML representation of the content
 274          */
 275         protected function getHtml() {
 276                 return $this->getHighlightHtml();
 277         }
 278
 279         /**
 280          * Generates an HTML version of the content, for display.
 281          *
 282          * This default implementation returns an HTML-escaped version
 283          * of the raw text content.
 284          *
 285          * @note The functionality of this method should really be implemented
 286          * in getHtml(), and subclasses should override getHtml() if needed.
 287          * getHighlightHtml() is kept around for backward compatibility with
 288          * extensions that already override it.
 289          *
 290          * @deprecated since 1.24. Use getHtml() instead. In particular, subclasses overriding
 291          *     getHighlightHtml() should override getHtml() instead.
 292          *
 293          * @return string An HTML representation of the content
 294          */
 295         protected function getHighlightHtml() {
 296                 return htmlspecialchars( $this->getNativeData() );
 297         }
 298
 299         /**
 300          * This implementation provides lossless conversion between content models based
 301          * on TextContent.
 302          *
 303          * @param string $toModel The desired content model, use the CONTENT_MODEL_XXX flags.
 304          * @param string $lossy Flag, set to "lossy" to allow lossy conversion. If lossy conversion is not
 305          *     allowed, full round-trip conversion is expected to work without losing information.
 306          *
 307          * @return Content|bool A content object with the content model $toModel, or false if that
 308          *     conversion is not supported.
 309          *
 310          * @see Content::convert()
 311          */
 312         public function convert( $toModel, $lossy = '' ) {
 313                 $converted = parent::convert( $toModel, $lossy );
 314
 315                 if ( $converted !== false ) {
 316                         return $converted;
 317                 }
 318
 319                 $toHandler = ContentHandler::getForModelID( $toModel );
 320
 321                 if ( $toHandler instanceof TextContentHandler ) {
 322                         // NOTE: ignore content serialization format - it's just text anyway.
 323                         $text = $this->getNativeData();
 324                         $converted = $toHandler->unserializeContent( $text );
 325                 }
 326
 327                 return $converted;
 328         }
 329
 330 }