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 use MediaWiki\MediaWikiServices;
  29
  30 /**
  31  * Content object implementation for representing flat text.
  32  *
  33  * TextContent instances are immutable
  34  *
  35  * @ingroup Content
  36  */
  37 class TextContent extends AbstractContent {
  38
  39         /**
  40          * @var string
  41          */
  42         protected $mText;
  43
  44         /**
  45          * @param string $text
  46          * @param string $model_id
  47          * @throws MWException
  48          */
  49         public function __construct( $text, $model_id = CONTENT_MODEL_TEXT ) {
  50                 parent::__construct( $model_id );
  51
  52                 if ( $text === null || $text === false ) {
  53                         wfWarn( "TextContent constructed with \$text = " . var_export( $text, true ) . "! "
  54                                 . "This may indicate an error in the caller's scope.", 2 );
  55
  56                         $text = '';
  57                 }
  58
  59                 if ( !is_string( $text ) ) {
  60                         throw new MWException( "TextContent expects a string in the constructor." );
  61                 }
  62
  63                 $this->mText = $text;
  64         }
  65
  66         /**
  67          * @note Mutable subclasses MUST override this to return a copy!
  68          *
  69          * @return Content $this
  70          */
  71         public function copy() {
  72                 return $this; # NOTE: this is ok since TextContent are immutable.
  73         }
  74
  75         public function getTextForSummary( $maxlength = 250 ) {
  76                 $text = $this->getText();
  77
  78                 $truncatedtext = MediaWikiServices::getInstance()->getContentLanguage()->
  79                         truncateForDatabase( preg_replace( "/[\n\r]/", ' ', $text ), max( 0, $maxlength ) );
  80
  81                 return $truncatedtext;
  82         }
  83
  84         /**
  85          * Returns the text's size in bytes.
  86          *
  87          * @return int
  88          */
  89         public function getSize() {
  90                 $text = $this->getText();
  91
  92                 return strlen( $text );
  93         }
  94
  95         /**
  96          * Returns true if this content is not a redirect, and $wgArticleCountMethod
  97          * is "any".
  98          *
  99          * @param bool|null $hasLinks If it is known whether this content contains links,
 100          * provide this information here, to avoid redundant parsing to find out.
 101          *
 102          * @return bool
 103          */
 104         public function isCountable( $hasLinks = null ) {
 105                 global $wgArticleCountMethod;
 106
 107                 if ( $this->isRedirect() ) {
 108                         return false;
 109                 }
 110
 111                 if ( $wgArticleCountMethod === 'any' ) {
 112                         return true;
 113                 }
 114
 115                 return false;
 116         }
 117
 118         /**
 119          * Returns the text represented by this Content object, as a string.
 120          *
 121          * @deprecated since 1.33 use getText() instead.
 122          *
 123          * @return string The raw text. Subclasses may guarantee a specific syntax here.
 124          */
 125         public function getNativeData() {
 126                 return $this->getText();
 127         }
 128
 129         /**
 130          * Returns the text represented by this Content object, as a string.
 131          *
 132          * @since 1.33
 133          *
 134          * @return string The raw text.
 135          */
 136         public function getText() {
 137                 return $this->mText;
 138         }
 139
 140         /**
 141          * Returns the text represented by this Content object, as a string.
 142          *
 143          * @return string The raw text.
 144          */
 145         public function getTextForSearchIndex() {
 146                 return $this->getText();
 147         }
 148
 149         /**
 150          * Returns attempts to convert this content object to wikitext,
 151          * and then returns the text string. The conversion may be lossy.
 152          *
 153          * @note this allows any text-based content to be transcluded as if it was wikitext.
 154          *
 155          * @return string|bool The raw text, or false if the conversion failed.
 156          */
 157         public function getWikitextForTransclusion() {
 158                 /** @var WikitextContent $wikitext */
 159                 $wikitext = $this->convert( CONTENT_MODEL_WIKITEXT, 'lossy' );
 160                 '@phan-var WikitextContent $wikitext';
 161
 162                 if ( $wikitext ) {
 163                         return $wikitext->getText();
 164                 } else {
 165                         return false;
 166                 }
 167         }
 168
 169         /**
 170          * Do a "\r\n" -> "\n" and "\r" -> "\n" transformation
 171          * as well as trim trailing whitespace
 172          *
 173          * This was formerly part of Parser::preSaveTransform, but
 174          * for non-wikitext content models they probably still want
 175          * to normalize line endings without all of the other PST
 176          * changes.
 177          *
 178          * @since 1.28
 179          * @param string $text
 180          * @return string
 181          */
 182         public static function normalizeLineEndings( $text ) {
 183                 return str_replace( [ "\r\n", "\r" ], "\n", rtrim( $text ) );
 184         }
 185
 186         /**
 187          * Returns a Content object with pre-save transformations applied.
 188          *
 189          * At a minimum, subclasses should make sure to call TextContent::normalizeLineEndings()
 190          * either directly or part of Parser::preSaveTransform().
 191          *
 192          * @param Title $title
 193          * @param User $user
 194          * @param ParserOptions $popts
 195          *
 196          * @return Content
 197          */
 198         public function preSaveTransform( Title $title, User $user, ParserOptions $popts ) {
 199                 $text = $this->getText();
 200                 $pst = self::normalizeLineEndings( $text );
 201
 202                 return ( $text === $pst ) ? $this : new static( $pst, $this->getModel() );
 203         }
 204
 205         /**
 206          * Diff this content object with another content object.
 207          *
 208          * @since 1.21
 209          *
 210          * @param Content $that The other content object to compare this content object to.
 211          * @param Language|null $lang The language object to use for text segmentation.
 212          *    If not given, the content language is used.
 213          *
 214          * @return Diff A diff representing the changes that would have to be
 215          *    made to this content object to make it equal to $that.
 216          */
 217         public function diff( Content $that, Language $lang = null ) {
 218                 $this->checkModelID( $that->getModel() );
 219                 /** @var self $that */
 220                 '@phan-var self $that';
 221                 // @todo could implement this in DifferenceEngine and just delegate here?
 222
 223                 if ( !$lang ) {
 224                         $lang = MediaWikiServices::getInstance()->getContentLanguage();
 225                 }
 226
 227                 $otext = $this->getText();
 228                 $ntext = $that->getText();
 229
 230                 # Note: Use native PHP diff, external engines don't give us abstract output
 231                 $ota = explode( "\n", $lang->segmentForDiff( $otext ) );
 232                 $nta = explode( "\n", $lang->segmentForDiff( $ntext ) );
 233
 234                 $diff = new Diff( $ota, $nta );
 235
 236                 return $diff;
 237         }
 238
 239         /**
 240          * Fills the provided ParserOutput object with information derived from the content.
 241          * Unless $generateHtml was false, this includes an HTML representation of the content
 242          * provided by getHtml().
 243          *
 244          * For content models listed in $wgTextModelsToParse, this method will call the MediaWiki
 245          * wikitext parser on the text to extract any (wikitext) links, magic words, etc.
 246          *
 247          * Subclasses may override this to provide custom content processing.
 248          * For custom HTML generation alone, it is sufficient to override getHtml().
 249          *
 250          * @param Title $title Context title for parsing
 251          * @param int $revId Revision ID (for {{REVISIONID}})
 252          * @param ParserOptions $options
 253          * @param bool $generateHtml Whether or not to generate HTML
 254          * @param ParserOutput &$output The output object to fill (reference).
 255          */
 256         protected function fillParserOutput( Title $title, $revId,
 257                 ParserOptions $options, $generateHtml, ParserOutput &$output
 258         ) {
 259                 global $wgTextModelsToParse;
 260
 261                 if ( in_array( $this->getModel(), $wgTextModelsToParse ) ) {
 262                         // parse just to get links etc into the database, HTML is replaced below.
 263                         $output = MediaWikiServices::getInstance()->getParser()
 264                                 ->parse( $this->getText(), $title, $options, true, true, $revId );
 265                 }
 266
 267                 if ( $generateHtml ) {
 268                         $html = $this->getHtml();
 269                 } else {
 270                         $html = '';
 271                 }
 272
 273                 $output->clearWrapperDivClass();
 274                 $output->setText( $html );
 275         }
 276
 277         /**
 278          * Generates an HTML version of the content, for display. Used by
 279          * fillParserOutput() to provide HTML for the ParserOutput object.
 280          *
 281          * Subclasses may override this to provide a custom HTML rendering.
 282          * If further information is to be derived from the content (such as
 283          * categories), the fillParserOutput() method can be overridden instead.
 284          *
 285          * For backwards-compatibility, this default implementation just calls
 286          * getHighlightHtml().
 287          *
 288          * @return string An HTML representation of the content
 289          */
 290         protected function getHtml() {
 291                 return $this->getHighlightHtml();
 292         }
 293
 294         /**
 295          * Generates an HTML version of the content, for display.
 296          *
 297          * This default implementation returns an HTML-escaped version
 298          * of the raw text content.
 299          *
 300          * @note The functionality of this method should really be implemented
 301          * in getHtml(), and subclasses should override getHtml() if needed.
 302          * getHighlightHtml() is kept around for backward compatibility with
 303          * extensions that already override it.
 304          *
 305          * @deprecated since 1.24. Use getHtml() instead. In particular, subclasses overriding
 306          *     getHighlightHtml() should override getHtml() instead.
 307          *
 308          * @return string An HTML representation of the content
 309          */
 310         protected function getHighlightHtml() {
 311                 return htmlspecialchars( $this->getText() );
 312         }
 313
 314         /**
 315          * This implementation provides lossless conversion between content models based
 316          * on TextContent.
 317          *
 318          * @param string $toModel The desired content model, use the CONTENT_MODEL_XXX flags.
 319          * @param string $lossy Flag, set to "lossy" to allow lossy conversion. If lossy conversion is not
 320          *     allowed, full round-trip conversion is expected to work without losing information.
 321          *
 322          * @return Content|bool A content object with the content model $toModel, or false if that
 323          *     conversion is not supported.
 324          *
 325          * @see Content::convert()
 326          */
 327         public function convert( $toModel, $lossy = '' ) {
 328                 $converted = parent::convert( $toModel, $lossy );
 329
 330                 if ( $converted !== false ) {
 331                         return $converted;
 332                 }
 333
 334                 $toHandler = ContentHandler::getForModelID( $toModel );
 335
 336                 if ( $toHandler instanceof TextContentHandler ) {
 337                         // NOTE: ignore content serialization format - it's just text anyway.
 338                         $text = $this->getText();
 339                         $converted = $toHandler->unserializeContent( $text );
 340                 }
 341
 342                 return $converted;
 343         }
 344
 345 }