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          * @param string $text
  39          * @param string $model_id
  40          */
  41         public function __construct( $text, $model_id = CONTENT_MODEL_TEXT ) {
  42                 parent::__construct( $model_id );
  43
  44                 if ( $text === null || $text === false ) {
  45                         wfWarn( "TextContent constructed with \$text = " . var_export( $text, true ) . "! "
  46                                 . "This may indicate an error in the caller's scope.", 2 );
  47
  48                         $text = '';
  49                 }
  50
  51                 if ( !is_string( $text ) ) {
  52                         throw new MWException( "TextContent expects a string in the constructor." );
  53                 }
  54
  55                 $this->mText = $text;
  56         }
  57
  58         /**
  59          * @note Mutable subclasses MUST override this to return a copy!
  60          *
  61          * @return Content $this
  62          */
  63         public function copy() {
  64                 return $this; # NOTE: this is ok since TextContent are immutable.
  65         }
  66
  67         public function getTextForSummary( $maxlength = 250 ) {
  68                 global $wgContLang;
  69
  70                 $text = $this->getNativeData();
  71
  72                 $truncatedtext = $wgContLang->truncate(
  73                         preg_replace( "/[\n\r]/", ' ', $text ),
  74                         max( 0, $maxlength ) );
  75
  76                 return $truncatedtext;
  77         }
  78
  79         /**
  80          * Returns the text's size in bytes.
  81          *
  82          * @return int
  83          */
  84         public function getSize() {
  85                 $text = $this->getNativeData();
  86
  87                 return strlen( $text );
  88         }
  89
  90         /**
  91          * Returns true if this content is not a redirect, and $wgArticleCountMethod
  92          * is "any".
  93          *
  94          * @param bool $hasLinks If it is known whether this content contains links,
  95          * provide this information here, to avoid redundant parsing to find out.
  96          *
  97          * @return bool
  98          */
  99         public function isCountable( $hasLinks = null ) {
 100                 global $wgArticleCountMethod;
 101
 102                 if ( $this->isRedirect() ) {
 103                         return false;
 104                 }
 105
 106                 if ( $wgArticleCountMethod === 'any' ) {
 107                         return true;
 108                 }
 109
 110                 return false;
 111         }
 112
 113         /**
 114          * Returns the text represented by this Content object, as a string.
 115          *
 116          * @return string The raw text.
 117          */
 118         public function getNativeData() {
 119                 $text = $this->mText;
 120
 121                 return $text;
 122         }
 123
 124         /**
 125          * Returns the text represented by this Content object, as a string.
 126          *
 127          * @return string The raw text.
 128          */
 129         public function getTextForSearchIndex() {
 130                 return $this->getNativeData();
 131         }
 132
 133         /**
 134          * Returns attempts to convert this content object to wikitext,
 135          * and then returns the text string. The conversion may be lossy.
 136          *
 137          * @note: this allows any text-based content to be transcluded as if it was wikitext.
 138          *
 139          * @return string|bool The raw text, or false if the conversion failed.
 140          */
 141         public function getWikitextForTransclusion() {
 142                 $wikitext = $this->convert( CONTENT_MODEL_WIKITEXT, 'lossy' );
 143
 144                 if ( $wikitext ) {
 145                         return $wikitext->getNativeData();
 146                 } else {
 147                         return false;
 148                 }
 149         }
 150
 151         /**
 152          * Returns a Content object with pre-save transformations applied.
 153          * This implementation just trims trailing whitespace.
 154          *
 155          * @param Title $title
 156          * @param User $user
 157          * @param ParserOptions $popts
 158          *
 159          * @return Content
 160          */
 161         public function preSaveTransform( Title $title, User $user, ParserOptions $popts ) {
 162                 $text = $this->getNativeData();
 163                 $pst = rtrim( $text );
 164
 165                 return ( $text === $pst ) ? $this : new static( $pst );
 166         }
 167
 168         /**
 169          * Diff this content object with another content object.
 170          *
 171          * @since 1.21
 172          *
 173          * @param Content $that The other content object to compare this content object to.
 174          * @param Language $lang The language object to use for text segmentation.
 175          *    If not given, $wgContentLang is used.
 176          *
 177          * @return Diff A diff representing the changes that would have to be
 178          *    made to this content object to make it equal to $that.
 179          */
 180         public function diff( Content $that, Language $lang = null ) {
 181                 global $wgContLang;
 182
 183                 $this->checkModelID( $that->getModel() );
 184
 185                 // @todo could implement this in DifferenceEngine and just delegate here?
 186
 187                 if ( !$lang ) {
 188                         $lang = $wgContLang;
 189                 }
 190
 191                 $otext = $this->getNativeData();
 192                 $ntext = $that->getNativeData();
 193
 194                 # Note: Use native PHP diff, external engines don't give us abstract output
 195                 $ota = explode( "\n", $lang->segmentForDiff( $otext ) );
 196                 $nta = explode( "\n", $lang->segmentForDiff( $ntext ) );
 197
 198                 $diff = new Diff( $ota, $nta );
 199
 200                 return $diff;
 201         }
 202
 203         /**
 204          * Returns a generic ParserOutput object, wrapping the HTML returned by
 205          * getHtml().
 206          *
 207          * @param Title $title Context title for parsing
 208          * @param int $revId Revision ID (for {{REVISIONID}})
 209          * @param ParserOptions $options Parser options
 210          * @param bool $generateHtml Whether or not to generate HTML
 211          *
 212          * @return ParserOutput Representing the HTML form of the text.
 213          */
 214         public function getParserOutput( Title $title, $revId = null,
 215                 ParserOptions $options = null, $generateHtml = true ) {
 216                 global $wgParser, $wgTextModelsToParse;
 217
 218                 if ( !$options ) {
 219                         //NOTE: use canonical options per default to produce cacheable output
 220                         $options = $this->getContentHandler()->makeParserOptions( 'canonical' );
 221                 }
 222
 223                 if ( in_array( $this->getModel(), $wgTextModelsToParse ) ) {
 224                         // parse just to get links etc into the database
 225                         $po = $wgParser->parse( $this->getNativeData(), $title, $options, true, true, $revId );
 226                 } else {
 227                         $po = new ParserOutput();
 228                 }
 229
 230                 if ( $generateHtml ) {
 231                         $html = $this->getHtml();
 232                 } else {
 233                         $html = '';
 234                 }
 235
 236                 $po->setText( $html );
 237
 238                 return $po;
 239         }
 240
 241         /**
 242          * Generates an HTML version of the content, for display. Used by
 243          * getParserOutput() to construct a ParserOutput object.
 244          *
 245          * Subclasses may override this to provide a custom HTML rendering.
 246          * If further information is to be derived from the content (such as
 247          * categories), the getParserOutput() method can be overridden instead.
 248          *
 249          * For backwards-compatibility, this default implementation just calls
 250          * getHighlightHtml().
 251          *
 252          * @return string An HTML representation of the content
 253          */
 254         protected function getHtml() {
 255                 return $this->getHighlightHtml();
 256         }
 257
 258         /**
 259          * Generates an HTML version of the content, for display.
 260          *
 261          * This default implementation returns an HTML-escaped version
 262          * of the raw text content.
 263          *
 264          * @note: The functionality of this method should really be implemented
 265          * in getHtml(), and subclasses should override getHtml() if needed.
 266          * getHighlightHtml() is kept around for backward compatibility with
 267          * extensions that already override it.
 268          *
 269          * @deprecated since 1.24. Use getHtml() instead. In particular, subclasses overriding
 270          *     getHighlightHtml() should override getHtml() instead.
 271          *
 272          * @return string An HTML representation of the content
 273          */
 274         protected function getHighlightHtml() {
 275                 return htmlspecialchars( $this->getNativeData() );
 276         }
 277
 278         /**
 279          * This implementation provides lossless conversion between content models based
 280          * on TextContent.
 281          *
 282          * @param string $toModel The desired content model, use the CONTENT_MODEL_XXX flags.
 283          * @param string $lossy Flag, set to "lossy" to allow lossy conversion. If lossy conversion is not
 284          *     allowed, full round-trip conversion is expected to work without losing information.
 285          *
 286          * @return Content|bool A content object with the content model $toModel, or false if that
 287          *     conversion is not supported.
 288          *
 289          * @see Content::convert()
 290          */
 291         public function convert( $toModel, $lossy = '' ) {
 292                 $converted = parent::convert( $toModel, $lossy );
 293
 294                 if ( $converted !== false ) {
 295                         return $converted;
 296                 }
 297
 298                 $toHandler = ContentHandler::getForModelID( $toModel );
 299
 300                 if ( $toHandler instanceof TextContentHandler ) {
 301                         // NOTE: ignore content serialization format - it's just text anyway.
 302                         $text = $this->getNativeData();
 303                         $converted = $toHandler->unserializeContent( $text );
 304                 }
 305
 306                 return $converted;
 307         }
 308
 309 }