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