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                 return $this->mText;
 120         }
 121
 122         /**
 123          * Returns the text represented by this Content object, as a string.
 124          *
 125          * @return string The raw text.
 126          */
 127         public function getTextForSearchIndex() {
 128                 return $this->getNativeData();
 129         }
 130
 131         /**
 132          * Returns attempts to convert this content object to wikitext,
 133          * and then returns the text string. The conversion may be lossy.
 134          *
 135          * @note this allows any text-based content to be transcluded as if it was wikitext.
 136          *
 137          * @return string|bool The raw text, or false if the conversion failed.
 138          */
 139         public function getWikitextForTransclusion() {
 140                 $wikitext = $this->convert( CONTENT_MODEL_WIKITEXT, 'lossy' );
 141
 142                 if ( $wikitext ) {
 143                         return $wikitext->getNativeData();
 144                 } else {
 145                         return false;
 146                 }
 147         }
 148
 149         /**
 150          * Returns a Content object with pre-save transformations applied.
 151          * This implementation just trims trailing whitespace.
 152          *
 153          * @param Title $title
 154          * @param User $user
 155          * @param ParserOptions $popts
 156          *
 157          * @return Content
 158          */
 159         public function preSaveTransform( Title $title, User $user, ParserOptions $popts ) {
 160                 $text = $this->getNativeData();
 161                 $pst = rtrim( $text );
 162
 163                 return ( $text === $pst ) ? $this : new static( $pst );
 164         }
 165
 166         /**
 167          * Diff this content object with another content object.
 168          *
 169          * @since 1.21
 170          *
 171          * @param Content $that The other content object to compare this content object to.
 172          * @param Language $lang The language object to use for text segmentation.
 173          *    If not given, $wgContentLang is used.
 174          *
 175          * @return Diff A diff representing the changes that would have to be
 176          *    made to this content object to make it equal to $that.
 177          */
 178         public function diff( Content $that, Language $lang = null ) {
 179                 global $wgContLang;
 180
 181                 $this->checkModelID( $that->getModel() );
 182
 183                 // @todo could implement this in DifferenceEngine and just delegate here?
 184
 185                 if ( !$lang ) {
 186                         $lang = $wgContLang;
 187                 }
 188
 189                 $otext = $this->getNativeData();
 190                 $ntext = $that->getNativeData();
 191
 192                 # Note: Use native PHP diff, external engines don't give us abstract output
 193                 $ota = explode( "\n", $lang->segmentForDiff( $otext ) );
 194                 $nta = explode( "\n", $lang->segmentForDiff( $ntext ) );
 195
 196                 $diff = new Diff( $ota, $nta );
 197
 198                 return $diff;
 199         }
 200
 201         /**
 202          * Fills the provided ParserOutput object with information derived from the content.
 203          * Unless $generateHtml was false, this includes an HTML representation of the content
 204          * provided by getHtml().
 205          *
 206          * For content models listed in $wgTextModelsToParse, this method will call the MediaWiki
 207          * wikitext parser on the text to extract any (wikitext) links, magic words, etc.
 208          *
 209          * Subclasses may override this to provide custom content processing.
 210          * For custom HTML generation alone, it is sufficient to override getHtml().
 211          *
 212          * @param Title $title Context title for parsing
 213          * @param int $revId Revision ID (for {{REVISIONID}})
 214          * @param ParserOptions $options Parser options
 215          * @param bool $generateHtml Whether or not to generate HTML
 216          * @param ParserOutput $output The output object to fill (reference).
 217          */
 218         protected function fillParserOutput( Title $title, $revId,
 219                 ParserOptions $options, $generateHtml, ParserOutput &$output
 220         ) {
 221                 global $wgParser, $wgTextModelsToParse;
 222
 223                 if ( in_array( $this->getModel(), $wgTextModelsToParse ) ) {
 224                         // parse just to get links etc into the database, HTML is replaced below.
 225                         $output = $wgParser->parse( $this->getNativeData(), $title, $options, true, true, $revId );
 226                 }
 227
 228                 if ( $generateHtml ) {
 229                         $html = $this->getHtml();
 230                 } else {
 231                         $html = '';
 232                 }
 233
 234                 $output->setText( $html );
 235         }
 236
 237         /**
 238          * Generates an HTML version of the content, for display. Used by
 239          * fillParserOutput() to provide HTML for the ParserOutput object.
 240          *
 241          * Subclasses may override this to provide a custom HTML rendering.
 242          * If further information is to be derived from the content (such as
 243          * categories), the fillParserOutput() method can be overridden instead.
 244          *
 245          * For backwards-compatibility, this default implementation just calls
 246          * getHighlightHtml().
 247          *
 248          * @return string An HTML representation of the content
 249          */
 250         protected function getHtml() {
 251                 return $this->getHighlightHtml();
 252         }
 253
 254         /**
 255          * Generates an HTML version of the content, for display.
 256          *
 257          * This default implementation returns an HTML-escaped version
 258          * of the raw text content.
 259          *
 260          * @note The functionality of this method should really be implemented
 261          * in getHtml(), and subclasses should override getHtml() if needed.
 262          * getHighlightHtml() is kept around for backward compatibility with
 263          * extensions that already override it.
 264          *
 265          * @deprecated since 1.24. Use getHtml() instead. In particular, subclasses overriding
 266          *     getHighlightHtml() should override getHtml() instead.
 267          *
 268          * @return string An HTML representation of the content
 269          */
 270         protected function getHighlightHtml() {
 271                 return htmlspecialchars( $this->getNativeData() );
 272         }
 273
 274         /**
 275          * This implementation provides lossless conversion between content models based
 276          * on TextContent.
 277          *
 278          * @param string $toModel The desired content model, use the CONTENT_MODEL_XXX flags.
 279          * @param string $lossy Flag, set to "lossy" to allow lossy conversion. If lossy conversion is not
 280          *     allowed, full round-trip conversion is expected to work without losing information.
 281          *
 282          * @return Content|bool A content object with the content model $toModel, or false if that
 283          *     conversion is not supported.
 284          *
 285          * @see Content::convert()
 286          */
 287         public function convert( $toModel, $lossy = '' ) {
 288                 $converted = parent::convert( $toModel, $lossy );
 289
 290                 if ( $converted !== false ) {
 291                         return $converted;
 292                 }
 293
 294                 $toHandler = ContentHandler::getForModelID( $toModel );
 295
 296                 if ( $toHandler instanceof TextContentHandler ) {
 297                         // NOTE: ignore content serialization format - it's just text anyway.
 298                         $text = $this->getNativeData();
 299                         $converted = $toHandler->unserializeContent( $text );
 300                 }
 301
 302                 return $converted;
 303         }
 304
 305 }