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->mDescription = "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                         return; // process only this path
 121                 } elseif ( $this->hasOption( 'list-file' ) ) {
 122                         $file = $this->getOption( 'list-file' );
 123                         wfSuppressWarnings();
 124                         $f = fopen( $file, 'r' );
 125                         wfRestoreWarnings();
 126                         if ( !$f ) {
 127                                 $this->error( "Can't open file $file\n", true );
 128                         }
 129                         $path = trim( fgets( $f ) );
 130                         while ( $path ) {
 131                                 $this->addPath( $path );
 132                         }
 133                         fclose( $f );
 134                         return;
 135                 } elseif ( $this->hasOption( 'modified' ) ) {
 136                         $this->output( "Retrieving list from Git... " );
 137                         $files = $this->getGitModifiedFiles( $IP );
 138                         $this->output( "done\n" );
 139                         foreach ( $files as $file ) {
 140                                 if ( $this->isSuitableFile( $file ) && !is_dir( $file ) ) {
 141                                         $this->mFiles[] = $file;
 142                                 }
 143                         }
 144                         return;
 145                 }
 146
 147                 $this->output( 'Building file list...', 'listfiles' );
 148
 149                 // Only check files in these directories.
 150                 // Don't just put $IP, because the recursive dir thingie goes into all subdirs
 151                 $dirs = array(
 152                         $IP . '/includes',
 153                         $IP . '/mw-config',
 154                         $IP . '/languages',
 155                         $IP . '/maintenance',
 156                         $IP . '/skins',
 157                 );
 158                 if ( $this->hasOption( 'with-extensions' ) ) {
 159                         $dirs[] = $IP . '/extensions';
 160                 }
 161
 162                 foreach ( $dirs as $d ) {
 163                         $this->addDirectoryContent( $d );
 164                 }
 165
 166                 // Manually add two user-editable files that are usually sources of problems
 167                 if ( file_exists( "$IP/LocalSettings.php" ) ) {
 168                         $this->mFiles[] = "$IP/LocalSettings.php";
 169                 }
 170
 171                 $this->output( 'done.', 'listfiles' );
 172         }
 173
 174         /**
 175          * Returns a list of tracked files in a Git work tree differing from the master branch.
 176          * @param string $path Path to the repository
 177          * @return array Resulting list of changed files
 178          */
 179         private function getGitModifiedFiles( $path ) {
 180
 181                 global $wgMaxShellMemory;
 182
 183                 if ( !is_dir( "$path/.git" ) ) {
 184                         $this->error( "Error: Not a Git repository!\n", true );
 185                 }
 186
 187                 // git diff eats memory.
 188                 $oldMaxShellMemory = $wgMaxShellMemory;
 189                 if ( $wgMaxShellMemory < 1024000 ) {
 190                         $wgMaxShellMemory = 1024000;
 191                 }
 192
 193                 $ePath = wfEscapeShellArg( $path );
 194
 195                 // Find an ancestor in common with master (rather than just using its HEAD)
 196                 // to prevent files only modified there from showing up in the list.
 197                 $cmd = "cd $ePath && git merge-base master HEAD";
 198                 $retval = 0;
 199                 $output = wfShellExec( $cmd, $retval );
 200                 if ( $retval !== 0 ) {
 201                         $this->error( "Error retrieving base SHA1 from Git!\n", true );
 202                 }
 203
 204                 // Find files in the working tree that changed since then.
 205                 $eBase = wfEscapeShellArg( rtrim( $output, "\n" ) );
 206                 $cmd = "cd $ePath && git diff --name-only --diff-filter AM $eBase";
 207                 $retval = 0;
 208                 $output = wfShellExec( $cmd, $retval );
 209                 if ( $retval !== 0 ) {
 210                         $this->error( "Error retrieving list from Git!\n", true );
 211                 }
 212
 213                 $wgMaxShellMemory = $oldMaxShellMemory;
 214
 215                 $arr = array();
 216                 $filename = strtok( $output, "\n" );
 217                 while ( $filename !== false ) {
 218                         if ( $filename !== '' ) {
 219                                 $arr[] = "$path/$filename";
 220                         }
 221                         $filename = strtok( "\n" );
 222                 }
 223
 224                 return $arr;
 225         }
 226
 227         /**
 228          * Returns true if $file is of a type we can check
 229          * @param string $file
 230          * @return bool
 231          */
 232         private function isSuitableFile( $file ) {
 233                 $file = str_replace( '\\', '/', $file );
 234                 $ext = pathinfo( $file, PATHINFO_EXTENSION );
 235                 if ( $ext != 'php' && $ext != 'inc' && $ext != 'php5' ) {
 236                         return false;
 237                 }
 238                 foreach ( $this->mIgnorePaths as $regex ) {
 239                         $m = array();
 240                         if ( preg_match( "~{$regex}~", $file, $m ) ) {
 241                                 return false;
 242                         }
 243                 }
 244                 return true;
 245         }
 246
 247         /**
 248          * Add given path to file list, searching it in include path if needed
 249          * @param string $path
 250          * @return bool
 251          */
 252         private function addPath( $path ) {
 253                 global $IP;
 254                 return $this->addFileOrDir( $path ) || $this->addFileOrDir( "$IP/$path" );
 255         }
 256
 257         /**
 258          * Add given file to file list, or, if it's a directory, add its content
 259          * @param string $path
 260          * @return bool
 261          */
 262         private function addFileOrDir( $path ) {
 263                 if ( is_dir( $path ) ) {
 264                         $this->addDirectoryContent( $path );
 265                 } elseif ( file_exists( $path ) ) {
 266                         $this->mFiles[] = $path;
 267                 } else {
 268                         return false;
 269                 }
 270                 return true;
 271         }
 272
 273         /**
 274          * Add all suitable files in given directory or its subdirectories to the file list
 275          *
 276          * @param string $dir Directory to process
 277          */
 278         private function addDirectoryContent( $dir ) {
 279                 $iterator = new RecursiveIteratorIterator(
 280                         new RecursiveDirectoryIterator( $dir ),
 281                         RecursiveIteratorIterator::SELF_FIRST
 282                 );
 283                 foreach ( $iterator as $file ) {
 284                         if ( $this->isSuitableFile( $file->getRealPath() ) ) {
 285                                 $this->mFiles[] = $file->getRealPath();
 286                         }
 287                 }
 288         }
 289
 290         /**
 291          * Check a file for syntax errors using Parsekit. Shamelessly stolen
 292          * from tools/lint.php by TimStarling
 293          * @param string $file Path to a file to check for syntax errors
 294          * @return bool
 295          */
 296         private function checkFileWithParsekit( $file ) {
 297                 static $okErrors = array(
 298                         'Redefining already defined constructor',
 299                         'Assigning the return value of new by reference is deprecated',
 300                 );
 301                 $errors = array();
 302                 parsekit_compile_file( $file, $errors, PARSEKIT_SIMPLE );
 303                 $ret = true;
 304                 if ( $errors ) {
 305                         foreach ( $errors as $error ) {
 306                                 foreach ( $okErrors as $okError ) {
 307                                         if ( substr( $error['errstr'], 0, strlen( $okError ) ) == $okError ) {
 308                                                 continue 2;
 309                                         }
 310                                 }
 311                                 $ret = false;
 312                                 $this->output( "Error in $file line {$error['lineno']}: {$error['errstr']}\n" );
 313                                 $this->mFailures[$file] = $errors;
 314                         }
 315                 }
 316                 return $ret;
 317         }
 318
 319         /**
 320          * Check a file for syntax errors using php -l
 321          * @param string $file Path to a file to check for syntax errors
 322          * @return bool
 323          */
 324         private function checkFileWithCli( $file ) {
 325                 $res = exec( 'php -l ' . wfEscapeShellArg( $file ) );
 326                 if ( strpos( $res, 'No syntax errors detected' ) === false ) {
 327                         $this->mFailures[$file] = $res;
 328                         $this->output( $res . "\n" );
 329                         return false;
 330                 }
 331                 return true;
 332         }
 333
 334         /**
 335          * Check a file for non-fatal coding errors, such as byte-order marks in the beginning
 336          * or pointless ?> closing tags at the end.
 337          *
 338          * @param string $file String Path to a file to check for errors
 339          * @return bool
 340          */
 341         private function checkForMistakes( $file ) {
 342                 foreach ( $this->mNoStyleCheckPaths as $regex ) {
 343                         $m = array();
 344                         if ( preg_match( "~{$regex}~", $file, $m ) ) {
 345                                 return;
 346                         }
 347                 }
 348
 349                 $text = file_get_contents( $file );
 350                 $tokens = token_get_all( $text );
 351
 352                 $this->checkEvilToken( $file, $tokens, '@', 'Error supression operator (@)' );
 353                 $this->checkRegex( $file, $text, '/^[\s\r\n]+<\?/', 'leading whitespace' );
 354                 $this->checkRegex( $file, $text, '/\?>[\s\r\n]*$/', 'trailing ?>' );
 355                 $this->checkRegex( $file, $text, '/^[\xFF\xFE\xEF]/', 'byte-order mark' );
 356         }
 357
 358         private function checkRegex( $file, $text, $regex, $desc ) {
 359                 if ( !preg_match( $regex, $text ) ) {
 360                         return;
 361                 }
 362
 363                 if ( !isset( $this->mWarnings[$file] ) ) {
 364                         $this->mWarnings[$file] = array();
 365                 }
 366                 $this->mWarnings[$file][] = $desc;
 367                 $this->output( "Warning in file $file: $desc found.\n" );
 368         }
 369
 370         private function checkEvilToken( $file, $tokens, $evilToken, $desc ) {
 371                 if ( !in_array( $evilToken, $tokens ) ) {
 372                         return;
 373                 }
 374
 375                 if ( !isset( $this->mWarnings[$file] ) ) {
 376                         $this->mWarnings[$file] = array();
 377                 }
 378                 $this->mWarnings[$file][] = $desc;
 379                 $this->output( "Warning in file $file: $desc found.\n" );
 380         }
 381 }
 382
 383 $maintClass = "CheckSyntax";
 384 require_once RUN_MAINTENANCE_IF_MAIN;