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