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