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