maintenance/findHooks.php

   1 <?php
   2 /**
   3  * Simple script that try to find documented hook and hooks actually
   4  * in the code and show what's missing.
   5  *
   6  * This script assumes that:
   7  * - hooks names in hooks.txt are at the beginning of a line and single quoted.
   8  * - hooks names in code are the first parameter of wfRunHooks.
   9  *
  10  * if --online option is passed, the script will compare the hooks in the code
  11  * with the ones at http://www.mediawiki.org/wiki/Manual:Hooks
  12  *
  13  * Any instance of wfRunHooks that doesn't meet these parameters will be noted.
  14  *
  15  * Copyright © Antoine Musso
  16  *
  17  * This program is free software; you can redistribute it and/or modify
  18  * it under the terms of the GNU General Public License as published by
  19  * the Free Software Foundation; either version 2 of the License, or
  20  * (at your option) any later version.
  21  *
  22  * This program is distributed in the hope that it will be useful,
  23  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  24  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  25  * GNU General Public License for more details.
  26  *
  27  * You should have received a copy of the GNU General Public License along
  28  * with this program; if not, write to the Free Software Foundation, Inc.,
  29  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  30  * http://www.gnu.org/copyleft/gpl.html
  31  *
  32  * @file
  33  * @ingroup Maintenance
  34  * @author Antoine Musso <hashar at free dot fr>
  35  */
  36
  37 require_once __DIR__ . '/Maintenance.php';
  38
  39 /**
  40  * Maintenance script that compares documented and actually present mismatches.
  41  *
  42  * @ingroup Maintenance
  43  */
  44 class FindHooks extends Maintenance {
  45         /*
  46          * Hooks that are ignored
  47          */
  48         protected static $ignore = array( 'testRunLegacyHooks' );
  49
  50         public function __construct() {
  51                 parent::__construct();
  52                 $this->addDescription( 'Find hooks that are undocumented, missing, or just plain wrong' );
  53                 $this->addOption( 'online', 'Check against MediaWiki.org hook documentation' );
  54         }
  55
  56         public function getDbType() {
  57                 return Maintenance::DB_NONE;
  58         }
  59
  60         public function execute() {
  61                 global $IP;
  62
  63                 $documentedHooks = $this->getHooksFromDoc( $IP . '/docs/hooks.txt' );
  64                 $potentialHooks = array();
  65                 $bad = array();
  66
  67                 // TODO: Don't hardcode the list of directories
  68                 $pathinc = array(
  69                         $IP . '/',
  70                         $IP . '/includes/',
  71                         $IP . '/includes/actions/',
  72                         $IP . '/includes/api/',
  73                         $IP . '/includes/cache/',
  74                         $IP . '/includes/changes/',
  75                         $IP . '/includes/changetags/',
  76                         $IP . '/includes/clientpool/',
  77                         $IP . '/includes/content/',
  78                         $IP . '/includes/context/',
  79                         $IP . '/includes/dao/',
  80                         $IP . '/includes/db/',
  81                         $IP . '/includes/debug/',
  82                         $IP . '/includes/deferred/',
  83                         $IP . '/includes/diff/',
  84                         $IP . '/includes/exception/',
  85                         $IP . '/includes/externalstore/',
  86                         $IP . '/includes/filebackend/',
  87                         $IP . '/includes/filerepo/',
  88                         $IP . '/includes/filerepo/file/',
  89                         $IP . '/includes/gallery/',
  90                         $IP . '/includes/htmlform/',
  91                         $IP . '/includes/installer/',
  92                         $IP . '/includes/interwiki/',
  93                         $IP . '/includes/jobqueue/',
  94                         $IP . '/includes/json/',
  95                         $IP . '/includes/logging/',
  96                         $IP . '/includes/mail/',
  97                         $IP . '/includes/media/',
  98                         $IP . '/includes/page/',
  99                         $IP . '/includes/parser/',
 100                         $IP . '/includes/password/',
 101                         $IP . '/includes/rcfeed/',
 102                         $IP . '/includes/resourceloader/',
 103                         $IP . '/includes/revisiondelete/',
 104                         $IP . '/includes/search/',
 105                         $IP . '/includes/site/',
 106                         $IP . '/includes/skins/',
 107                         $IP . '/includes/specialpage/',
 108                         $IP . '/includes/specials/',
 109                         $IP . '/includes/upload/',
 110                         $IP . '/includes/utils/',
 111                         $IP . '/languages/',
 112                         $IP . '/maintenance/',
 113                         $IP . '/maintenance/language/',
 114                         $IP . '/tests/',
 115                         $IP . '/tests/parser/',
 116                         $IP . '/tests/phpunit/suites/',
 117                 );
 118
 119                 foreach ( $pathinc as $dir ) {
 120                         $potentialHooks = array_merge( $potentialHooks, $this->getHooksFromPath( $dir ) );
 121                         $bad = array_merge( $bad, $this->getBadHooksFromPath( $dir ) );
 122                 }
 123
 124                 $documented = array_keys( $documentedHooks );
 125                 $potential = array_keys( $potentialHooks );
 126                 $potential = array_unique( $potential );
 127                 $bad = array_diff( array_unique( $bad ), self::$ignore );
 128                 $todo = array_diff( $potential, $documented, self::$ignore );
 129                 $deprecated = array_diff( $documented, $potential, self::$ignore );
 130
 131                 // Check parameter count and references
 132                 $badParameterCount = $badParameterReference = array();
 133                 foreach ( $potentialHooks as $hook => $args ) {
 134                         if ( !isset( $documentedHooks[$hook] ) ) {
 135                                 // Not documented, but that will also be in $todo
 136                                 continue;
 137                         }
 138                         $argsDoc = $documentedHooks[$hook];
 139                         if ( $args === 'unknown' || $argsDoc === 'unknown' ) {
 140                                 // Could not get parameter information
 141                                 continue;
 142                         }
 143                         if ( count( $argsDoc ) !== count( $args ) ) {
 144                                 $badParameterCount[] = $hook . ': Doc: ' . count( $argsDoc ) . ' vs. Code: ' . count( $args );
 145                         } else {
 146                                 // Check if & is equal
 147                                 foreach ( $argsDoc as $index => $argDoc ) {
 148                                         $arg = $args[$index];
 149                                         if ( ( $arg[0] === '&' ) !== ( $argDoc[0] === '&' ) ) {
 150                                                 $badParameterReference[] = $hook . ': References different: Doc: ' . $argDoc .
 151                                                         ' vs. Code: ' . $arg;
 152                                         }
 153                                 }
 154                         }
 155                 }
 156
 157                 // let's show the results:
 158                 $this->printArray( 'Undocumented', $todo );
 159                 $this->printArray( 'Documented and not found', $deprecated );
 160                 $this->printArray( 'Unclear hook calls', $bad );
 161                 $this->printArray( 'Different parameter count', $badParameterCount );
 162                 $this->printArray( 'Different parameter reference', $badParameterReference );
 163
 164                 if ( count( $todo ) == 0 && count( $deprecated ) == 0 && count( $bad ) == 0
 165                         && count( $badParameterCount ) == 0 && count( $badParameterReference ) == 0
 166                 ) {
 167                         $this->output( "Looks good!\n" );
 168                 } else {
 169                         $this->error( 'The script finished with errors.', 1 );
 170                 }
 171         }
 172
 173         /**
 174          * Get the hook documentation, either locally or from MediaWiki.org
 175          * @param string $doc
 176          * @return array Array: key => hook name; value => array of arguments or string 'unknown'
 177          */
 178         private function getHooksFromDoc( $doc ) {
 179                 if ( $this->hasOption( 'online' ) ) {
 180                         return $this->getHooksFromOnlineDoc();
 181                 } else {
 182                         return $this->getHooksFromLocalDoc( $doc );
 183                 }
 184         }
 185
 186         /**
 187          * Get hooks from a local file (for example docs/hooks.txt)
 188          * @param string $doc Filename to look in
 189          * @return array Array: key => hook name; value => array of arguments or string 'unknown'
 190          */
 191         private function getHooksFromLocalDoc( $doc ) {
 192                 $m = array();
 193                 $content = file_get_contents( $doc );
 194                 preg_match_all(
 195                         "/\n'(.*?)':.*((?:\n.+)*)/",
 196                         $content,
 197                         $m,
 198                         PREG_SET_ORDER
 199                 );
 200
 201                 // Extract the documented parameter
 202                 $hooks = array();
 203                 foreach ( $m as $match ) {
 204                         $args = array();
 205                         if ( isset( $match[2] ) ) {
 206                                 $n = array();
 207                                 if ( preg_match_all( "/\n(&?\\$\w+):.+/", $match[2], $n ) ) {
 208                                         $args = $n[1];
 209                                 }
 210                         }
 211                         $hooks[$match[1]] = $args;
 212                 }
 213                 return $hooks;
 214         }
 215
 216         /**
 217          * Get hooks from www.mediawiki.org using the API
 218          * @return array Array: key => hook name; value => string 'unknown'
 219          */
 220         private function getHooksFromOnlineDoc() {
 221                 $allhooks = $this->getHooksFromOnlineDocCategory( 'MediaWiki_hooks' );
 222                 $removed = $this->getHooksFromOnlineDocCategory( 'Removed_hooks' );
 223                 return array_diff_key( $allhooks, $removed );
 224         }
 225
 226         /**
 227          * @param string $title
 228          * @return array
 229          */
 230         private function getHooksFromOnlineDocCategory( $title ) {
 231                 $params = array(
 232                         'action' => 'query',
 233                         'list' => 'categorymembers',
 234                         'cmtitle' => "Category:$title",
 235                         'cmlimit' => 500,
 236                         'format' => 'json',
 237                         'continue' => '',
 238                 );
 239
 240                 $retval = array();
 241                 while ( true ) {
 242                         $json = Http::get(
 243                                 wfAppendQuery( 'http://www.mediawiki.org/w/api.php', $params ),
 244                                 array(),
 245                                 __METHOD__
 246                         );
 247                         $data = FormatJson::decode( $json, true );
 248                         foreach ( $data['query']['categorymembers'] as $page ) {
 249                                 if ( preg_match( '/Manual\:Hooks\/([a-zA-Z0-9- :]+)/', $page['title'], $m ) ) {
 250                                         // parameters are unknown, because that needs parsing of wikitext
 251                                         $retval[str_replace( ' ', '_', $m[1] )] = 'unknown';
 252                                 }
 253                         }
 254                         if ( !isset( $data['continue'] ) ) {
 255                                 return $retval;
 256                         }
 257                         $params = array_replace( $params, $data['continue'] );
 258                 }
 259         }
 260
 261         /**
 262          * Get hooks from a PHP file
 263          * @param string $file Full filename to the PHP file.
 264          * @return array Array: key => hook name; value => array of arguments or string 'unknown'
 265          */
 266         private function getHooksFromFile( $file ) {
 267                 $content = file_get_contents( $file );
 268                 $m = array();
 269                 preg_match_all(
 270                         // All functions which runs hooks
 271                         '/(?:wfRunHooks|Hooks\:\:run|ContentHandler\:\:runLegacyHooks)\s*\(\s*' .
 272                                 // First argument is the hook name as string
 273                                 '([\'"])(.*?)\1' .
 274                                 // Comma for second argument
 275                                 '(?:\s*(,))?' .
 276                                 // Second argument must start with array to be processed
 277                                 '(?:\s*array\s*\(' .
 278                                 // Matching inside array - allows one deep of brackets
 279                                 '((?:[^\(\)]|\([^\(\)]*\))*)' .
 280                                 // End
 281                                 '\))?/',
 282                         $content,
 283                         $m,
 284                         PREG_SET_ORDER
 285                 );
 286
 287                 // Extract parameter
 288                 $hooks = array();
 289                 foreach ( $m as $match ) {
 290                         $args = array();
 291                         if ( isset( $match[4] ) ) {
 292                                 $n = array();
 293                                 if ( preg_match_all( '/((?:[^,\(\)]|\([^\(\)]*\))+)/', $match[4], $n ) ) {
 294                                         $args = array_map( 'trim', $n[1] );
 295                                 }
 296                         } elseif ( isset( $match[3] ) ) {
 297                                 // Found a parameter for Hooks::run,
 298                                 // but could not extract the hooks argument,
 299                                 // because there are given by a variable
 300                                 $args = 'unknown';
 301                         }
 302                         $hooks[$match[2]] = $args;
 303                 }
 304
 305                 return $hooks;
 306         }
 307
 308         /**
 309          * Get hooks from the source code.
 310          * @param string $path Directory where the include files can be found
 311          * @return array Array: key => hook name; value => array of arguments or string 'unknown'
 312          */
 313         private function getHooksFromPath( $path ) {
 314                 $hooks = array();
 315                 $dh = opendir( $path );
 316                 if ( $dh ) {
 317                         while ( ( $file = readdir( $dh ) ) !== false ) {
 318                                 if ( filetype( $path . $file ) == 'file' ) {
 319                                         $hooks = array_merge( $hooks, $this->getHooksFromFile( $path . $file ) );
 320                                 }
 321                         }
 322                         closedir( $dh );
 323                 }
 324
 325                 return $hooks;
 326         }
 327
 328         /**
 329          * Get bad hooks (where the hook name could not be determined) from a PHP file
 330          * @param string $file Full filename to the PHP file.
 331          * @return array Array of bad wfRunHooks() lines
 332          */
 333         private function getBadHooksFromFile( $file ) {
 334                 $content = file_get_contents( $file );
 335                 $m = array();
 336                 # We want to skip the "function wfRunHooks()" one.  :)
 337                 preg_match_all( '/(?<!function )wfRunHooks\(\s*[^\s\'"].*/', $content, $m );
 338                 $list = array();
 339                 foreach ( $m[0] as $match ) {
 340                         $list[] = $match . "(" . $file . ")";
 341                 }
 342
 343                 return $list;
 344         }
 345
 346         /**
 347          * Get bad hooks from the source code.
 348          * @param string $path Directory where the include files can be found
 349          * @return array Array of bad wfRunHooks() lines
 350          */
 351         private function getBadHooksFromPath( $path ) {
 352                 $hooks = array();
 353                 $dh = opendir( $path );
 354                 if ( $dh ) {
 355                         while ( ( $file = readdir( $dh ) ) !== false ) {
 356                                 # We don't want to read this file as it contains bad calls to wfRunHooks()
 357                                 if ( filetype( $path . $file ) == 'file' && !$path . $file == __FILE__ ) {
 358                                         $hooks = array_merge( $hooks, $this->getBadHooksFromFile( $path . $file ) );
 359                                 }
 360                         }
 361                         closedir( $dh );
 362                 }
 363
 364                 return $hooks;
 365         }
 366
 367         /**
 368          * Nicely output the array
 369          * @param string $msg A message to show before the value
 370          * @param array $arr
 371          * @param bool $sort Whether to sort the array (Default: true)
 372          */
 373         private function printArray( $msg, $arr, $sort = true ) {
 374                 if ( $sort ) {
 375                         asort( $arr );
 376                 }
 377
 378                 foreach ( $arr as $v ) {
 379                         $this->output( "$msg: $v\n" );
 380                 }
 381         }
 382 }
 383
 384 $maintClass = 'FindHooks';
 385 require_once RUN_MAINTENANCE_IF_MAIN;