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         public function __construct( $text, $model_id = CONTENT_MODEL_TEXT ) {
  37                 parent::__construct( $model_id );
  38
  39                 if ( $text === null || $text === false ) {
  40                         wfWarn( "TextContent constructed with \$text = " . var_export( $text, true ) . "! "
  41                                 . "This may indicate an error in the caller's scope." );
  42
  43                         $text = '';
  44                 }
  45
  46                 if ( !is_string( $text ) ) {
  47                         throw new MWException( "TextContent expects a string in the constructor." );
  48                 }
  49
  50                 $this->mText = $text;
  51         }
  52
  53         public function copy() {
  54                 return $this; # NOTE: this is ok since TextContent are immutable.
  55         }
  56
  57         public function getTextForSummary( $maxlength = 250 ) {
  58                 global $wgContLang;
  59
  60                 $text = $this->getNativeData();
  61
  62                 $truncatedtext = $wgContLang->truncate(
  63                         preg_replace( "/[\n\r]/", ' ', $text ),
  64                         max( 0, $maxlength ) );
  65
  66                 return $truncatedtext;
  67         }
  68
  69         /**
  70          * returns the text's size in bytes.
  71          *
  72          * @return int The size
  73          */
  74         public function getSize() {
  75                 $text = $this->getNativeData();
  76
  77                 return strlen( $text );
  78         }
  79
  80         /**
  81          * Returns true if this content is not a redirect, and $wgArticleCountMethod
  82          * is "any".
  83          *
  84          * @param bool $hasLinks if it is known whether this content contains links,
  85          * provide this information here, to avoid redundant parsing to find out.
  86          *
  87          * @return bool True if the content is countable
  88          */
  89         public function isCountable( $hasLinks = null ) {
  90                 global $wgArticleCountMethod;
  91
  92                 if ( $this->isRedirect() ) {
  93                         return false;
  94                 }
  95
  96                 if ( $wgArticleCountMethod === 'any' ) {
  97                         return true;
  98                 }
  99
 100                 return false;
 101         }
 102
 103         /**
 104          * Returns the text represented by this Content object, as a string.
 105          *
 106          * @return string: the raw text
 107          */
 108         public function getNativeData() {
 109                 $text = $this->mText;
 110
 111                 return $text;
 112         }
 113
 114         /**
 115          * Returns the text represented by this Content object, as a string.
 116          *
 117          * @return string: the raw text
 118          */
 119         public function getTextForSearchIndex() {
 120                 return $this->getNativeData();
 121         }
 122
 123         /**
 124          * Returns attempts to convert this content object to wikitext,
 125          * and then returns the text string. The conversion may be lossy.
 126          *
 127          * @note: this allows any text-based content to be transcluded as if it was wikitext.
 128          *
 129          * @return string|false: the raw text, or null if the conversion failed
 130          */
 131         public function getWikitextForTransclusion() {
 132                 $wikitext = $this->convert( CONTENT_MODEL_WIKITEXT, 'lossy' );
 133
 134                 if ( $wikitext ) {
 135                         return $wikitext->getNativeData();
 136                 } else {
 137                         return false;
 138                 }
 139         }
 140
 141         /**
 142          * Returns a Content object with pre-save transformations applied.
 143          * This implementation just trims trailing whitespace.
 144          *
 145          * @param $title Title
 146          * @param $user User
 147          * @param $popts ParserOptions
 148          * @return Content
 149          */
 150         public function preSaveTransform( Title $title, User $user, ParserOptions $popts ) {
 151                 $text = $this->getNativeData();
 152                 $pst = rtrim( $text );
 153
 154                 return ( $text === $pst ) ? $this : new static( $pst );
 155         }
 156
 157         /**
 158          * Diff this content object with another content object.
 159          *
 160          * @since 1.21
 161          *
 162          * @param $that Content: The other content object to compare this content
 163          * object to.
 164          * @param $lang Language: The language object to use for text segmentation.
 165          *    If not given, $wgContentLang is used.
 166          *
 167          * @return Diff A diff representing the changes that would have to be
 168          *    made to this content object to make it equal to $that.
 169          */
 170         public function diff( Content $that, Language $lang = null ) {
 171                 global $wgContLang;
 172
 173                 $this->checkModelID( $that->getModel() );
 174
 175                 // @todo could implement this in DifferenceEngine and just delegate here?
 176
 177                 if ( !$lang ) {
 178                         $lang = $wgContLang;
 179                 }
 180
 181                 $otext = $this->getNativeData();
 182                 $ntext = $that->getNativeData();
 183
 184                 # Note: Use native PHP diff, external engines don't give us abstract output
 185                 $ota = explode( "\n", $lang->segmentForDiff( $otext ) );
 186                 $nta = explode( "\n", $lang->segmentForDiff( $ntext ) );
 187
 188                 $diff = new Diff( $ota, $nta );
 189
 190                 return $diff;
 191         }
 192
 193         /**
 194          * Returns a generic ParserOutput object, wrapping the HTML returned by
 195          * getHtml().
 196          *
 197          * @param $title Title Context title for parsing
 198          * @param int|null $revId Revision ID (for {{REVISIONID}})
 199          * @param $options ParserOptions|null Parser options
 200          * @param bool $generateHtml Whether or not to generate HTML
 201          *
 202          * @return ParserOutput representing the HTML form of the text
 203          */
 204         public function getParserOutput( Title $title,
 205                 $revId = null,
 206                 ParserOptions $options = null, $generateHtml = true
 207         ) {
 208                 global $wgParser, $wgTextModelsToParse;
 209
 210                 if ( !$options ) {
 211                         //NOTE: use canonical options per default to produce cacheable output
 212                         $options = $this->getContentHandler()->makeParserOptions( 'canonical' );
 213                 }
 214
 215                 if ( in_array( $this->getModel(), $wgTextModelsToParse ) ) {
 216                         // parse just to get links etc into the database
 217                         $po = $wgParser->parse( $this->getNativeData(), $title, $options, true, true, $revId );
 218                 } else {
 219                         $po = new ParserOutput();
 220                 }
 221
 222                 if ( $generateHtml ) {
 223                         $html = $this->getHtml();
 224                 } else {
 225                         $html = '';
 226                 }
 227
 228                 $po->setText( $html );
 229
 230                 return $po;
 231         }
 232
 233         /**
 234          * Generates an HTML version of the content, for display. Used by
 235          * getParserOutput() to construct a ParserOutput object.
 236          *
 237          * This default implementation just calls getHighlightHtml(). Content
 238          * models that have another mapping to HTML (as is the case for markup
 239          * languages like wikitext) should override this method to generate the
 240          * appropriate HTML.
 241          *
 242          * @return string An HTML representation of the content
 243          */
 244         protected function getHtml() {
 245                 return $this->getHighlightHtml();
 246         }
 247
 248         /**
 249          * Generates a syntax-highlighted version of the content, as HTML.
 250          * Used by the default implementation of getHtml().
 251          *
 252          * @return string an HTML representation of the content's markup
 253          */
 254         protected function getHighlightHtml() {
 255                 # TODO: make Highlighter interface, use highlighter here, if available
 256                 return htmlspecialchars( $this->getNativeData() );
 257         }
 258
 259         /**
 260          * @see Content::convert()
 261          *
 262          * This implementation provides lossless conversion between content models based
 263          * on TextContent.
 264          *
 265          * @param string $toModel the desired content model, use the CONTENT_MODEL_XXX flags.
 266          * @param string $lossy flag, set to "lossy" to allow lossy conversion. If lossy conversion is
 267          * not allowed, full round-trip conversion is expected to work without losing information.
 268          *
 269          * @return Content|bool A content object with the content model $toModel, or false if
 270          * that conversion is not supported.
 271          */
 272         public function convert( $toModel, $lossy = '' ) {
 273                 $converted = parent::convert( $toModel, $lossy );
 274
 275                 if ( $converted !== false ) {
 276                         return $converted;
 277                 }
 278
 279                 $toHandler = ContentHandler::getForModelID( $toModel );
 280
 281                 if ( $toHandler instanceof TextContentHandler ) {
 282                         //NOTE: ignore content serialization format - it's just text anyway.
 283                         $text = $this->getNativeData();
 284                         $converted = $toHandler->unserializeContent( $text );
 285                 }
 286
 287                 return $converted;
 288         }
 289 }