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