includes/content/TextContent.php

   1 <?php
   2
   3 /**
   4  * Content object implementation for representing flat text.
   5  *
   6  * TextContent instances are immutable
   7  *
   8  * This program is free software; you can redistribute it and/or modify
   9  * it under the terms of the GNU General Public License as published by
  10  * the Free Software Foundation; either version 2 of the License, or
  11  * (at your option) any later version.
  12  *
  13  * This program is distributed in the hope that it will be useful,
  14  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  15  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  16  * GNU General Public License for more details.
  17  *
  18  * You should have received a copy of the GNU General Public License along
  19  * with this program; if not, write to the Free Software Foundation, Inc.,
  20  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  21  * http://www.gnu.org/copyleft/gpl.html
  22  *
  23  * @since 1.21
  24  *
  25  * @file
  26  * @ingroup Content
  27  *
  28  * @author Daniel Kinzler
  29  */
  30 class TextContent extends AbstractContent {
  31
  32         public function __construct( $text, $model_id = CONTENT_MODEL_TEXT ) {
  33                 parent::__construct( $model_id );
  34
  35                 if ( $text === null || $text === false ) {
  36                         wfWarn( "TextContent constructed with \$text = " . var_export( $text, true ) . "! "
  37                                         . "This may indicate an error in the caller's scope." );
  38
  39                         $text = '';
  40                 }
  41
  42                 if ( !is_string( $text ) ) {
  43                         throw new MWException( "TextContent expects a string in the constructor." );
  44                 }
  45
  46                 $this->mText = $text;
  47         }
  48
  49         public function copy() {
  50                 return $this; # NOTE: this is ok since TextContent are immutable.
  51         }
  52
  53         public function getTextForSummary( $maxlength = 250 ) {
  54                 global $wgContLang;
  55
  56                 $text = $this->getNativeData();
  57
  58                 $truncatedtext = $wgContLang->truncate(
  59                         preg_replace( "/[\n\r]/", ' ', $text ),
  60                         max( 0, $maxlength ) );
  61
  62                 return $truncatedtext;
  63         }
  64
  65         /**
  66          * returns the text's size in bytes.
  67          *
  68          * @return int The size
  69          */
  70         public function getSize( ) {
  71                 $text = $this->getNativeData( );
  72                 return strlen( $text );
  73         }
  74
  75         /**
  76          * Returns true if this content is not a redirect, and $wgArticleCountMethod
  77          * is "any".
  78          *
  79          * @param $hasLinks Bool: if it is known whether this content contains links,
  80          * provide this information here, to avoid redundant parsing to find out.
  81          *
  82          * @return bool True if the content is countable
  83          */
  84         public function isCountable( $hasLinks = null ) {
  85                 global $wgArticleCountMethod;
  86
  87                 if ( $this->isRedirect( ) ) {
  88                         return false;
  89                 }
  90
  91                 if ( $wgArticleCountMethod === 'any' ) {
  92                         return true;
  93                 }
  94
  95                 return false;
  96         }
  97
  98         /**
  99          * Returns the text represented by this Content object, as a string.
 100          *
 101          * @return string: the raw text
 102          */
 103         public function getNativeData( ) {
 104                 $text = $this->mText;
 105                 return $text;
 106         }
 107
 108         /**
 109          * Returns the text represented by this Content object, as a string.
 110          *
 111          * @return string: the raw text
 112          */
 113         public function getTextForSearchIndex( ) {
 114                 return $this->getNativeData();
 115         }
 116
 117         /**
 118          * Returns attempts to convert this content object to wikitext,
 119          * and then returns the text string. The conversion may be lossy.
 120          *
 121          * @note: this allows any text-based content to be transcluded as if it was wikitext.
 122          *
 123          * @return string|false: the raw text, or null if the conversion failed
 124          */
 125         public function getWikitextForTransclusion( ) {
 126                 $wikitext = $this->convert( CONTENT_MODEL_WIKITEXT, 'lossy' );
 127
 128                 if ( $wikitext ) {
 129                         return $wikitext->getNativeData();
 130                 } else {
 131                         return false;
 132                 }
 133         }
 134
 135         /**
 136          * Returns a Content object with pre-save transformations applied.
 137          * This implementation just trims trailing whitespace.
 138          *
 139          * @param $title Title
 140          * @param $user User
 141          * @param $popts ParserOptions
 142          * @return Content
 143          */
 144         public function preSaveTransform( Title $title, User $user, ParserOptions $popts ) {
 145                 $text = $this->getNativeData();
 146                 $pst = rtrim( $text );
 147
 148                 return ( $text === $pst ) ? $this : new WikitextContent( $pst );
 149         }
 150
 151         /**
 152          * Diff this content object with another content object.
 153          *
 154          * @since 1.21diff
 155          *
 156          * @param $that Content: The other content object to compare this content
 157          * object to.
 158          * @param $lang Language: The language object to use for text segmentation.
 159          *    If not given, $wgContentLang is used.
 160          *
 161          * @return DiffResult: A diff representing the changes that would have to be
 162          *    made to this content object to make it equal to $that.
 163          */
 164         public function diff( Content $that, Language $lang = null ) {
 165                 global $wgContLang;
 166
 167                 $this->checkModelID( $that->getModel() );
 168
 169                 # @todo: could implement this in DifferenceEngine and just delegate here?
 170
 171                 if ( !$lang ) $lang = $wgContLang;
 172
 173                 $otext = $this->getNativeData();
 174                 $ntext = $this->getNativeData();
 175
 176                 # Note: Use native PHP diff, external engines don't give us abstract output
 177                 $ota = explode( "\n", $wgContLang->segmentForDiff( $otext ) );
 178                 $nta = explode( "\n", $wgContLang->segmentForDiff( $ntext ) );
 179
 180                 $diff = new Diff( $ota, $nta );
 181                 return $diff;
 182         }
 183
 184
 185         /**
 186          * Returns a generic ParserOutput object, wrapping the HTML returned by
 187          * getHtml().
 188          *
 189          * @param $title Title Context title for parsing
 190          * @param $revId int|null Revision ID (for {{REVISIONID}})
 191          * @param $options ParserOptions|null Parser options
 192          * @param $generateHtml bool Whether or not to generate HTML
 193          *
 194          * @return ParserOutput representing the HTML form of the text
 195          */
 196         public function getParserOutput( Title $title,
 197                 $revId = null,
 198                 ParserOptions $options = null, $generateHtml = true
 199         ) {
 200                 global $wgParser, $wgTextModelsToParse;
 201
 202                 if ( !$options ) {
 203                         //NOTE: use canonical options per default to produce cacheable output
 204                         $options = $this->getContentHandler()->makeParserOptions( 'canonical' );
 205                 }
 206
 207                 if ( in_array( $this->getModel(), $wgTextModelsToParse ) ) {
 208                         // parse just to get links etc into the database
 209                         $po = $wgParser->parse( $this->getNativeData(), $title, $options, true, true, $revId );
 210                 } else {
 211                         $po = new ParserOutput();
 212                 }
 213
 214                 if ( $generateHtml ) {
 215                         $html = $this->getHtml();
 216                 } else {
 217                         $html = '';
 218                 }
 219
 220                 $po->setText( $html );
 221                 return $po;
 222         }
 223
 224         /**
 225          * Generates an HTML version of the content, for display. Used by
 226          * getParserOutput() to construct a ParserOutput object.
 227          *
 228          * This default implementation just calls getHighlightHtml(). Content
 229          * models that have another mapping to HTML (as is the case for markup
 230          * languages like wikitext) should override this method to generate the
 231          * appropriate HTML.
 232          *
 233          * @return string An HTML representation of the content
 234          */
 235         protected function getHtml() {
 236                 return $this->getHighlightHtml();
 237         }
 238
 239         /**
 240          * Generates a syntax-highlighted version of the content, as HTML.
 241          * Used by the default implementation of getHtml().
 242          *
 243          * @return string an HTML representation of the content's markup
 244          */
 245         protected function getHighlightHtml( ) {
 246                 # TODO: make Highlighter interface, use highlighter here, if available
 247                 return htmlspecialchars( $this->getNativeData() );
 248         }
 249
 250         /**
 251          * @see Content::convert()
 252          *
 253          * This implementation provides lossless conversion between content models based
 254          * on TextContent.
 255          *
 256          * @param String  $toModel the desired content model, use the CONTENT_MODEL_XXX flags.
 257          * @param String  $lossy flag, set to "lossy" to allow lossy conversion. If lossy conversion is
 258          * not allowed, full round-trip conversion is expected to work without losing information.
 259          *
 260          * @return Content|bool A content object with the content model $toModel, or false if
 261          * that conversion is not supported.
 262          */
 263         public function convert( $toModel, $lossy = '' ) {
 264                 $converted = parent::convert( $toModel, $lossy );
 265
 266                 if ( $converted !== false ) {
 267                         return $converted;
 268                 }
 269
 270                 $toHandler = ContentHandler::getForModelID( $toModel );
 271
 272                 if ( $toHandler instanceof TextContentHandler ) {
 273                         //NOTE: ignore content serialization format - it's just text anyway.
 274                         $text = $this->getNativeData();
 275                         $converted = $toHandler->unserializeContent( $text );
 276                 }
 277
 278                 return $converted;
 279         }
 280 }