maintenance/checkSyntax.php

   1 <?php
   2 /**
   3  * Check syntax of all PHP files in MediaWiki
   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 Maintenance
  22  */
  23
  24 require_once __DIR__ . '/Maintenance.php';
  25
  26 /**
  27  * Maintenance script to check syntax of all PHP files in MediaWiki.
  28  *
  29  * @ingroup Maintenance
  30  */
  31 class CheckSyntax extends Maintenance {
  32
  33         // List of files we're going to check
  34         private $mFiles = array(), $mFailures = array(), $mWarnings = array();
  35         private $mIgnorePaths = array(), $mNoStyleCheckPaths = array();
  36
  37         public function __construct() {
  38                 parent::__construct();
  39                 $this->addDescription( 'Check syntax for all PHP files in MediaWiki' );
  40                 $this->addOption( 'with-extensions', 'Also recurse the extensions folder' );
  41                 $this->addOption(
  42                         'path',
  43                         'Specific path (file or directory) to check, either with absolute path or '
  44                                 . 'relative to the root of this MediaWiki installation',
  45                         false,
  46                         true
  47                 );
  48                 $this->addOption(
  49                         'list-file',
  50                         'Text file containing list of files or directories to check',
  51                         false,
  52                         true
  53                 );
  54                 $this->addOption(
  55                         'modified',
  56                         'Check only files that were modified (requires Git command-line client)'
  57                 );
  58                 $this->addOption( 'syntax-only', 'Check for syntax validity only, skip code style warnings' );
  59         }
  60
  61         public function getDbType() {
  62                 return Maintenance::DB_NONE;
  63         }
  64
  65         public function execute() {
  66                 $this->buildFileList();
  67
  68                 // ParseKit is broken on PHP 5.3+, disabled until this is fixed
  69                 $useParseKit = function_exists( 'parsekit_compile_file' )
  70                         && version_compare( PHP_VERSION, '5.3', '<' );
  71
  72                 $str = 'Checking syntax (using ' . ( $useParseKit ?
  73                         'parsekit' : ' php -l, this can take a long time' ) . ")\n";
  74                 $this->output( $str );
  75                 foreach ( $this->mFiles as $f ) {
  76                         if ( $useParseKit ) {
  77                                 $this->checkFileWithParsekit( $f );
  78                         } else {
  79                                 $this->checkFileWithCli( $f );
  80                         }
  81                         if ( !$this->hasOption( 'syntax-only' ) ) {
  82                                 $this->checkForMistakes( $f );
  83                         }
  84                 }
  85                 $this->output( "\nDone! " . count( $this->mFiles ) . " files checked, " .
  86                         count( $this->mFailures ) . " failures and " . count( $this->mWarnings ) .
  87                         " warnings found\n" );
  88         }
  89
  90         /**
  91          * Build the list of files we'll check for syntax errors
  92          */
  93         private function buildFileList() {
  94                 global $IP;
  95
  96                 $this->mIgnorePaths = array(
  97                         // Compat stuff, explodes on PHP 5.3
  98                         "includes/NamespaceCompat.php$",
  99                 );
 100
 101                 $this->mNoStyleCheckPaths = array(
 102                         // Third-party code we don't care about
 103                         "/activemq_stomp/",
 104                         "EmailPage/PHPMailer",
 105                         "FCKeditor/fckeditor/",
 106                         '\bphplot-',
 107                         "/svggraph/",
 108                         "\bjsmin.php$",
 109                         "PEAR/File_Ogg/",
 110                         "QPoll/Excel/",
 111                         "/geshi/",
 112                         "/smarty/",
 113                 );
 114
 115                 if ( $this->hasOption( 'path' ) ) {
 116                         $path = $this->getOption( 'path' );
 117                         if ( !$this->addPath( $path ) ) {
 118                                 $this->error( "Error: can't find file or directory $path\n", true );
 119                         }
 120
 121                         return; // process only this path
 122                 } elseif ( $this->hasOption( 'list-file' ) ) {
 123                         $file = $this->getOption( 'list-file' );
 124                         MediaWiki\suppressWarnings();
 125                         $f = fopen( $file, 'r' );
 126                         MediaWiki\restoreWarnings();
 127                         if ( !$f ) {
 128                                 $this->error( "Can't open file $file\n", true );
 129                         }
 130                         $path = trim( fgets( $f ) );
 131                         while ( $path ) {
 132                                 $this->addPath( $path );
 133                         }
 134                         fclose( $f );
 135
 136                         return;
 137                 } elseif ( $this->hasOption( 'modified' ) ) {
 138                         $this->output( "Retrieving list from Git... " );
 139                         $files = $this->getGitModifiedFiles( $IP );
 140                         $this->output( "done\n" );
 141                         foreach ( $files as $file ) {
 142                                 if ( $this->isSuitableFile( $file ) && !is_dir( $file ) ) {
 143                                         $this->mFiles[] = $file;
 144                                 }
 145                         }
 146
 147                         return;
 148                 }
 149
 150                 $this->output( 'Building file list...', 'listfiles' );
 151
 152                 // Only check files in these directories.
 153                 // Don't just put $IP, because the recursive dir thingie goes into all subdirs
 154                 $dirs = array(
 155                         $IP . '/includes',
 156                         $IP . '/mw-config',
 157                         $IP . '/languages',
 158                         $IP . '/maintenance',
 159                         $IP . '/skins',
 160                 );
 161                 if ( $this->hasOption( 'with-extensions' ) ) {
 162                         $dirs[] = $IP . '/extensions';
 163                 }
 164
 165                 foreach ( $dirs as $d ) {
 166                         $this->addDirectoryContent( $d );
 167                 }
 168
 169                 // Manually add two user-editable files that are usually sources of problems
 170                 if ( file_exists( "$IP/LocalSettings.php" ) ) {
 171                         $this->mFiles[] = "$IP/LocalSettings.php";
 172                 }
 173
 174                 $this->output( 'done.', 'listfiles' );
 175         }
 176
 177         /**
 178          * Returns a list of tracked files in a Git work tree differing from the master branch.
 179          * @param string $path Path to the repository
 180          * @return array Resulting list of changed files
 181          */
 182         private function getGitModifiedFiles( $path ) {
 183
 184                 global $wgMaxShellMemory;
 185
 186                 if ( !is_dir( "$path/.git" ) ) {
 187                         $this->error( "Error: Not a Git repository!\n", true );
 188                 }
 189
 190                 // git diff eats memory.
 191                 $oldMaxShellMemory = $wgMaxShellMemory;
 192                 if ( $wgMaxShellMemory < 1024000 ) {
 193                         $wgMaxShellMemory = 1024000;
 194                 }
 195
 196                 $ePath = wfEscapeShellArg( $path );
 197
 198                 // Find an ancestor in common with master (rather than just using its HEAD)
 199                 // to prevent files only modified there from showing up in the list.
 200                 $cmd = "cd $ePath && git merge-base master HEAD";
 201                 $retval = 0;
 202                 $output = wfShellExec( $cmd, $retval );
 203                 if ( $retval !== 0 ) {
 204                         $this->error( "Error retrieving base SHA1 from Git!\n", true );
 205                 }
 206
 207                 // Find files in the working tree that changed since then.
 208                 $eBase = wfEscapeShellArg( rtrim( $output, "\n" ) );
 209                 $cmd = "cd $ePath && git diff --name-only --diff-filter AM $eBase";
 210                 $retval = 0;
 211                 $output = wfShellExec( $cmd, $retval );
 212                 if ( $retval !== 0 ) {
 213                         $this->error( "Error retrieving list from Git!\n", true );
 214                 }
 215
 216                 $wgMaxShellMemory = $oldMaxShellMemory;
 217
 218                 $arr = array();
 219                 $filename = strtok( $output, "\n" );
 220                 while ( $filename !== false ) {
 221                         if ( $filename !== '' ) {
 222                                 $arr[] = "$path/$filename";
 223                         }
 224                         $filename = strtok( "\n" );
 225                 }
 226
 227                 return $arr;
 228         }
 229
 230         /**
 231          * Returns true if $file is of a type we can check
 232          * @param string $file
 233          * @return bool
 234          */
 235         private function isSuitableFile( $file ) {
 236                 $file = str_replace( '\\', '/', $file );
 237                 $ext = pathinfo( $file, PATHINFO_EXTENSION );
 238                 if ( $ext != 'php' && $ext != 'inc' && $ext != 'php5' ) {
 239                         return false;
 240                 }
 241                 foreach ( $this->mIgnorePaths as $regex ) {
 242                         $m = array();
 243                         if ( preg_match( "~{$regex}~", $file, $m ) ) {
 244                                 return false;
 245                         }
 246                 }
 247
 248                 return true;
 249         }
 250
 251         /**
 252          * Add given path to file list, searching it in include path if needed
 253          * @param string $path
 254          * @return bool
 255          */
 256         private function addPath( $path ) {
 257                 global $IP;
 258
 259                 return $this->addFileOrDir( $path ) || $this->addFileOrDir( "$IP/$path" );
 260         }
 261
 262         /**
 263          * Add given file to file list, or, if it's a directory, add its content
 264          * @param string $path
 265          * @return bool
 266          */
 267         private function addFileOrDir( $path ) {
 268                 if ( is_dir( $path ) ) {
 269                         $this->addDirectoryContent( $path );
 270                 } elseif ( file_exists( $path ) ) {
 271                         $this->mFiles[] = $path;
 272                 } else {
 273                         return false;
 274                 }
 275
 276                 return true;
 277         }
 278
 279         /**
 280          * Add all suitable files in given directory or its subdirectories to the file list
 281          *
 282          * @param string $dir Directory to process
 283          */
 284         private function addDirectoryContent( $dir ) {
 285                 $iterator = new RecursiveIteratorIterator(
 286                         new RecursiveDirectoryIterator( $dir ),
 287                         RecursiveIteratorIterator::SELF_FIRST
 288                 );
 289                 foreach ( $iterator as $file ) {
 290                         if ( $this->isSuitableFile( $file->getRealPath() ) ) {
 291                                 $this->mFiles[] = $file->getRealPath();
 292                         }
 293                 }
 294         }
 295
 296         /**
 297          * Check a file for syntax errors using Parsekit. Shamelessly stolen
 298          * from tools/lint.php by TimStarling
 299          * @param string $file Path to a file to check for syntax errors
 300          * @return bool
 301          */
 302         private function checkFileWithParsekit( $file ) {
 303                 static $okErrors = array(
 304                         'Redefining already defined constructor',
 305                         'Assigning the return value of new by reference is deprecated',
 306                 );
 307                 $errors = array();
 308                 parsekit_compile_file( $file, $errors, PARSEKIT_SIMPLE );
 309                 $ret = true;
 310                 if ( $errors ) {
 311                         foreach ( $errors as $error ) {
 312                                 foreach ( $okErrors as $okError ) {
 313                                         if ( substr( $error['errstr'], 0, strlen( $okError ) ) == $okError ) {
 314                                                 continue 2;
 315                                         }
 316                                 }
 317                                 $ret = false;
 318                                 $this->output( "Error in $file line {$error['lineno']}: {$error['errstr']}\n" );
 319                                 $this->mFailures[$file] = $errors;
 320                         }
 321                 }
 322
 323                 return $ret;
 324         }
 325
 326         /**
 327          * Check a file for syntax errors using php -l
 328          * @param string $file Path to a file to check for syntax errors
 329          * @return bool
 330          */
 331         private function checkFileWithCli( $file ) {
 332                 $res = exec( 'php -l ' . wfEscapeShellArg( $file ) );
 333                 if ( strpos( $res, 'No syntax errors detected' ) === false ) {
 334                         $this->mFailures[$file] = $res;
 335                         $this->output( $res . "\n" );
 336
 337                         return false;
 338                 }
 339
 340                 return true;
 341         }
 342
 343         /**
 344          * Check a file for non-fatal coding errors, such as byte-order marks in the beginning
 345          * or pointless ?> closing tags at the end.
 346          *
 347          * @param string $file String Path to a file to check for errors
 348          */
 349         private function checkForMistakes( $file ) {
 350                 foreach ( $this->mNoStyleCheckPaths as $regex ) {
 351                         $m = array();
 352                         if ( preg_match( "~{$regex}~", $file, $m ) ) {
 353                                 return;
 354                         }
 355                 }
 356
 357                 $text = file_get_contents( $file );
 358                 $tokens = token_get_all( $text );
 359
 360                 $this->checkEvilToken( $file, $tokens, '@', 'Error supression operator (@)' );
 361                 $this->checkRegex( $file, $text, '/^[\s\r\n]+<\?/', 'leading whitespace' );
 362                 $this->checkRegex( $file, $text, '/\?>[\s\r\n]*$/', 'trailing ?>' );
 363                 $this->checkRegex( $file, $text, '/^[\xFF\xFE\xEF]/', 'byte-order mark' );
 364         }
 365
 366         private function checkRegex( $file, $text, $regex, $desc ) {
 367                 if ( !preg_match( $regex, $text ) ) {
 368                         return;
 369                 }
 370
 371                 if ( !isset( $this->mWarnings[$file] ) ) {
 372                         $this->mWarnings[$file] = array();
 373                 }
 374                 $this->mWarnings[$file][] = $desc;
 375                 $this->output( "Warning in file $file: $desc found.\n" );
 376         }
 377
 378         private function checkEvilToken( $file, $tokens, $evilToken, $desc ) {
 379                 if ( !in_array( $evilToken, $tokens ) ) {
 380                         return;
 381                 }
 382
 383                 if ( !isset( $this->mWarnings[$file] ) ) {
 384                         $this->mWarnings[$file] = array();
 385                 }
 386                 $this->mWarnings[$file][] = $desc;
 387                 $this->output( "Warning in file $file: $desc found.\n" );
 388         }
 389 }
 390
 391 $maintClass = "CheckSyntax";
 392 require_once RUN_MAINTENANCE_IF_MAIN;