includes/parser/Parser_LinkHooks.php

   1 <?php
   2 /**
   3  * Modified version of the PHP parser with hooks for wiki links; experimental
   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  * @file
  21  * @ingroup Parser
  22  */
  23
  24 /**
  25  * Parser with LinkHooks experiment
  26  * @ingroup Parser
  27  */
  28 class Parser_LinkHooks extends Parser {
  29         /**
  30          * Update this version number when the ParserOutput format
  31          * changes in an incompatible way, so the parser cache
  32          * can automatically discard old data.
  33          */
  34         const VERSION = '1.6.4';
  35
  36         # Flags for Parser::setLinkHook
  37         # Also available as global constants from Defines.php
  38         const SLH_PATTERN = 1;
  39
  40         # Constants needed for external link processing
  41         # Everything except bracket, space, or control characters
  42         const EXT_LINK_URL_CLASS = '[^][<>"\\x00-\\x20\\x7F]';
  43         const EXT_IMAGE_REGEX = '/^(http:\/\/|https:\/\/)([^][<>"\\x00-\\x20\\x7F]+)
  44                 \\/([A-Za-z0-9_.,~%\\-+&;#*?!=()@\\x80-\\xFF]+)\\.((?i)gif|png|jpg|jpeg)$/Sx';
  45
  46         /**#@+
  47          * @private
  48          */
  49         # Persistent:
  50         var $mLinkHooks;
  51
  52         /**#@-*/
  53
  54         /**
  55          * Constructor
  56          */
  57         public function __construct( $conf = array() ) {
  58                 parent::__construct( $conf );
  59                 $this->mLinkHooks = array();
  60         }
  61
  62         /**
  63          * Do various kinds of initialisation on the first call of the parser
  64          */
  65         function firstCallInit() {
  66                 parent::__construct();
  67                 if ( !$this->mFirstCall ) {
  68                         return;
  69                 }
  70                 $this->mFirstCall = false;
  71
  72                 wfProfileIn( __METHOD__ );
  73
  74                 $this->setHook( 'pre', array( $this, 'renderPreTag' ) );
  75                 CoreParserFunctions::register( $this );
  76                 CoreLinkFunctions::register( $this );
  77                 $this->initialiseVariables();
  78
  79                 wfRunHooks( 'ParserFirstCallInit', array( &$this ) );
  80                 wfProfileOut( __METHOD__ );
  81         }
  82
  83         /**
  84          * Create a link hook, e.g. [[Namepsace:...|display}}
  85          * The callback function should have the form:
  86          *    function myLinkCallback( $parser, $holders, $markers,
  87          *      Title $title, $titleText, &$sortText = null, &$leadingColon = false ) { ... }
  88          *
  89          * Or with SLH_PATTERN:
  90          *    function myLinkCallback( $parser, $holders, $markers, )
  91          *      &$titleText, &$sortText = null, &$leadingColon = false ) { ... }
  92          *
  93          * The callback may either return a number of different possible values:
  94          * String) Text result of the link
  95          * True) (Treat as link) Parse the link according to normal link rules
  96          * False) (Bad link) Just output the raw wikitext (You may modify the text first)
  97          *
  98          * @param $ns Integer or String: the Namespace ID or regex pattern if SLH_PATTERN is set
  99          * @param $callback Mixed: the callback function (and object) to use
 100          * @param $flags Integer: a combination of the following flags:
 101          *     SLH_PATTERN   Use a regex link pattern rather than a namespace
 102          *
 103          * @throws MWException
 104          * @return callback|null The old callback function for this name, if any
 105          */
 106         public function setLinkHook( $ns, $callback, $flags = 0 ) {
 107                 if( $flags & SLH_PATTERN && !is_string($ns) )
 108                         throw new MWException( __METHOD__ . '() expecting a regex string pattern.' );
 109                 elseif( $flags | ~SLH_PATTERN && !is_int( $ns ) )
 110                         throw new MWException( __METHOD__ . '() expecting a namespace index.' );
 111                 $oldVal = isset( $this->mLinkHooks[$ns] ) ? $this->mLinkHooks[$ns][0] : null;
 112                 $this->mLinkHooks[$ns] = array( $callback, $flags );
 113                 return $oldVal;
 114         }
 115
 116         /**
 117          * Get all registered link hook identifiers
 118          *
 119          * @return array
 120          */
 121         function getLinkHooks() {
 122                 return array_keys( $this->mLinkHooks );
 123         }
 124
 125         /**
 126          * Process [[ ]] wikilinks
 127          * @param $s
 128          * @throws MWException
 129          * @return LinkHolderArray
 130          *
 131          * @private
 132          */
 133         function replaceInternalLinks2( &$s ) {
 134                 wfProfileIn( __METHOD__ );
 135
 136                 wfProfileIn( __METHOD__ . '-setup' );
 137                 static $tc = false, $titleRegex; //$e1, $e1_img;
 138                 if( !$tc ) {
 139                         # the % is needed to support urlencoded titles as well
 140                         $tc = Title::legalChars() . '#%';
 141                         # Match a link having the form [[namespace:link|alternate]]trail
 142                         //$e1 = "/^([{$tc}]+)(?:\\|(.+?))?]](.*)\$/sD";
 143                         # Match cases where there is no "]]", which might still be images
 144                         //$e1_img = "/^([{$tc}]+)\\|(.*)\$/sD";
 145                         # Match a valid plain title
 146                         $titleRegex = "/^([{$tc}]+)$/sD";
 147                 }
 148
 149                 $holders = new LinkHolderArray( $this );
 150
 151                 if( is_null( $this->mTitle ) ) {
 152                         wfProfileOut( __METHOD__ . '-setup' );
 153                         wfProfileOut( __METHOD__ );
 154                         throw new MWException( __METHOD__ . ": \$this->mTitle is null\n" );
 155                 }
 156
 157                 wfProfileOut( __METHOD__ . '-setup' );
 158
 159                 $offset = 0;
 160                 $offsetStack = array();
 161                 $markers = new LinkMarkerReplacer( $this, $holders, array( &$this, 'replaceInternalLinksCallback' ) );
 162                 while( true ) {
 163                         $startBracketOffset = strpos( $s, '[[', $offset );
 164                         $endBracketOffset   = strpos( $s, ']]', $offset );
 165                         # Finish when there are no more brackets
 166                         if( $startBracketOffset === false && $endBracketOffset === false ) break;
 167                         # Determine if the bracket is a starting or ending bracket
 168                         # When we find both, use the first one
 169                         elseif( $startBracketOffset !== false && $endBracketOffset !== false )
 170                                 $isStart = $startBracketOffset <= $endBracketOffset;
 171                         # When we only found one, check which it is
 172                         else $isStart = $startBracketOffset !== false;
 173                         $bracketOffset = $isStart ? $startBracketOffset : $endBracketOffset;
 174                         if( $isStart ) {
 175                                 /** Opening bracket **/
 176                                 # Just push our current offset in the string onto the stack
 177                                 $offsetStack[] = $startBracketOffset;
 178                         } else {
 179                                 /** Closing bracket **/
 180                                 # Pop the start pos for our current link zone off the stack
 181                                 $startBracketOffset = array_pop( $offsetStack );
 182                                 # Just to clean up the code, lets place offsets on the outer ends
 183                                 $endBracketOffset += 2;
 184
 185                                 # Only do logic if we actually have a opening bracket for this
 186                                 if( isset( $startBracketOffset ) ) {
 187                                         # Extract text inside the link
 188                                         @list( $titleText, $paramText ) = explode( '|',
 189                                                 substr( $s, $startBracketOffset + 2, $endBracketOffset - $startBracketOffset - 4 ), 2 );
 190                                         # Create markers only for valid links
 191                                         if( preg_match( $titleRegex, $titleText ) ) {
 192                                                 # Store the text for the marker
 193                                                 $marker = $markers->addMarker( $titleText, $paramText );
 194                                                 # Replace the current link with the marker
 195                                                 $s = substr( $s, 0, $startBracketOffset ) .
 196                                                         $marker .
 197                                                         substr( $s, $endBracketOffset );
 198                                                 # We have modified $s, because of this we need to set the
 199                                                 # offset manually since the end position is different now
 200                                                 $offset = $startBracketOffset+strlen( $marker );
 201                                                 continue;
 202                                         }
 203                                         # ToDo: Some LinkHooks may allow recursive links inside of
 204                                         # the link text, create a regex that also matches our
 205                                         # <!-- LINKMARKER ### --> sequence in titles
 206                                         # ToDO: Some LinkHooks use patterns rather than namespaces
 207                                         # these need to be tested at this point here
 208                                 }
 209                         }
 210                         # Bump our offset to after our current bracket
 211                         $offset = $bracketOffset+2;
 212                 }
 213
 214                 # Now expand our tree
 215                 wfProfileIn( __METHOD__ . '-expand' );
 216                 $s = $markers->expand( $s );
 217                 wfProfileOut( __METHOD__ . '-expand' );
 218
 219                 wfProfileOut( __METHOD__ );
 220                 return $holders;
 221         }
 222
 223         function replaceInternalLinksCallback( $parser, $holders, $markers, $titleText, $paramText ) {
 224                 wfProfileIn( __METHOD__ );
 225                 $wt = isset( $paramText ) ? "[[$titleText|$paramText]]" : "[[$titleText]]";
 226                 wfProfileIn( __METHOD__ . "-misc" );
 227
 228                 # Don't allow internal links to pages containing
 229                 # PROTO: where PROTO is a valid URL protocol; these
 230                 # should be external links.
 231                 if( preg_match( '/^\b(?i:' . wfUrlProtocols() . ')/', $titleText ) ) {
 232                         wfProfileOut( __METHOD__ . "-misc" );
 233                         wfProfileOut( __METHOD__ );
 234                         return $wt;
 235                 }
 236
 237                 # Make subpage if necessary
 238                 if( $this->areSubpagesAllowed() ) {
 239                         $titleText = $this->maybeDoSubpageLink( $titleText, $paramText );
 240                 }
 241
 242                 # Check for a leading colon and strip it if it is there
 243                 $leadingColon = $titleText[0] == ':';
 244                 if( $leadingColon ) $titleText = substr( $titleText, 1 );
 245
 246                 wfProfileOut( __METHOD__ . "-misc" );
 247                 # Make title object
 248                 wfProfileIn( __METHOD__ . "-title" );
 249                 $title = Title::newFromText( $this->mStripState->unstripNoWiki( $titleText ) );
 250                 if( !$title ) {
 251                         wfProfileOut( __METHOD__ . "-title" );
 252                         wfProfileOut( __METHOD__ );
 253                         return $wt;
 254                 }
 255                 $ns = $title->getNamespace();
 256                 wfProfileOut( __METHOD__ . "-title" );
 257
 258                 # Default for Namespaces is a default link
 259                 # ToDo: Default for patterns is plain wikitext
 260                 $return = true;
 261                 if( isset( $this->mLinkHooks[$ns] ) ) {
 262                         list( $callback, $flags ) = $this->mLinkHooks[$ns];
 263                         if( $flags & SLH_PATTERN ) {
 264                                 $args = array( $parser, $holders, $markers, $titleText, &$paramText, &$leadingColon );
 265                         } else {
 266                                 $args = array( $parser, $holders, $markers, $title, $titleText, &$paramText, &$leadingColon );
 267                         }
 268                         # Workaround for PHP bug 35229 and similar
 269                         if ( !is_callable( $callback ) ) {
 270                                 throw new MWException( "Tag hook for namespace $ns is not callable\n" );
 271                         }
 272                         $return = call_user_func_array( $callback, $args );
 273                 }
 274                 if( $return === true ) {
 275                         # True (treat as plain link) was returned, call the defaultLinkHook
 276                         $return = CoreLinkFunctions::defaultLinkHook( $parser, $holders, $markers, $title,
 277                                 $titleText, $paramText, $leadingColon );
 278                 }
 279                 if( $return === false ) {
 280                         # False (no link) was returned, output plain wikitext
 281                         # Build it again as the hook is allowed to modify $paramText
 282                         $return = isset( $paramText ) ? "[[$titleText|$paramText]]" : "[[$titleText]]";
 283                 }
 284                 # Content was returned, return it
 285                 wfProfileOut( __METHOD__ );
 286                 return $return;
 287         }
 288
 289 }
 290
 291 class LinkMarkerReplacer {
 292
 293         protected $markers, $nextId, $parser, $holders, $callback;
 294
 295         function __construct( $parser, $holders, $callback ) {
 296                 $this->nextId   = 0;
 297                 $this->markers  = array();
 298                 $this->parser   = $parser;
 299                 $this->holders  = $holders;
 300                 $this->callback = $callback;
 301         }
 302
 303         function addMarker( $titleText, $paramText ) {
 304                 $id = $this->nextId++;
 305                 $this->markers[$id] = array( $titleText, $paramText );
 306                 return "<!-- LINKMARKER $id -->";
 307         }
 308
 309         function findMarker( $string ) {
 310                 return (bool) preg_match( '/<!-- LINKMARKER [0-9]+ -->/', $string );
 311         }
 312
 313         function expand( $string ) {
 314                 return StringUtils::delimiterReplaceCallback( "<!-- LINKMARKER ", " -->", array( &$this, 'callback' ), $string );
 315         }
 316
 317         function callback( $m ) {
 318                 $id = intval( $m[1] );
 319                 if( !array_key_exists( $id, $this->markers ) ) return $m[0];
 320                 $args = $this->markers[$id];
 321                 array_unshift( $args, $this );
 322                 array_unshift( $args, $this->holders );
 323                 array_unshift( $args, $this->parser );
 324                 return call_user_func_array( $this->callback, $args );
 325         }
 326 }