includes/GitInfo.php

   1 <?php
   2 /**
   3  * A class to help return information about a git repo MediaWiki may be inside
   4  * This is used by Special:Version and is also useful for the LocalSettings.php
   5  * of anyone working on large branches in git to setup config that show up only
   6  * when specific branches are currently checked out.
   7  *
   8  * This program is free software; you can redistribute it and/or modify
   9  * it under the terms of the GNU General Public License as published by
  10  * the Free Software Foundation; either version 2 of the License, or
  11  * (at your option) any later version.
  12  *
  13  * This program is distributed in the hope that it will be useful,
  14  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  15  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  16  * GNU General Public License for more details.
  17  *
  18  * You should have received a copy of the GNU General Public License along
  19  * with this program; if not, write to the Free Software Foundation, Inc.,
  20  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  21  * http://www.gnu.org/copyleft/gpl.html
  22  *
  23  * @file
  24  */
  25
  26 use MediaWiki\Shell\Shell;
  27
  28 class GitInfo {
  29
  30         /**
  31          * Singleton for the repo at $IP
  32          */
  33         protected static $repo = null;
  34
  35         /**
  36          * Location of the .git directory
  37          */
  38         protected $basedir;
  39
  40         /**
  41          * Path to JSON cache file for pre-computed git information.
  42          */
  43         protected $cacheFile;
  44
  45         /**
  46          * Cached git information.
  47          */
  48         protected $cache = [];
  49
  50         /**
  51          * @var array|false Map of repo URLs to viewer URLs. Access via static method getViewers().
  52          */
  53         private static $viewers = false;
  54
  55         /**
  56          * @param string $repoDir The root directory of the repo where .git can be found
  57          * @param bool $usePrecomputed Use precomputed information if available
  58          * @see precomputeValues
  59          */
  60         public function __construct( $repoDir, $usePrecomputed = true ) {
  61                 $this->cacheFile = self::getCacheFilePath( $repoDir );
  62                 wfDebugLog( 'gitinfo',
  63                         "Computed cacheFile={$this->cacheFile} for {$repoDir}"
  64                 );
  65                 if ( $usePrecomputed &&
  66                         $this->cacheFile !== null &&
  67                         is_readable( $this->cacheFile )
  68                 ) {
  69                         $this->cache = FormatJson::decode(
  70                                 file_get_contents( $this->cacheFile ),
  71                                 true
  72                         );
  73                         wfDebugLog( 'gitinfo', "Loaded git data from cache for {$repoDir}" );
  74                 }
  75
  76                 if ( !$this->cacheIsComplete() ) {
  77                         wfDebugLog( 'gitinfo', "Cache incomplete for {$repoDir}" );
  78                         $this->basedir = $repoDir . DIRECTORY_SEPARATOR . '.git';
  79                         if ( is_readable( $this->basedir ) && !is_dir( $this->basedir ) ) {
  80                                 $GITfile = file_get_contents( $this->basedir );
  81                                 if ( strlen( $GITfile ) > 8 &&
  82                                         substr( $GITfile, 0, 8 ) === 'gitdir: '
  83                                 ) {
  84                                         $path = rtrim( substr( $GITfile, 8 ), "\r\n" );
  85                                         if ( $path[0] === '/' || substr( $path, 1, 1 ) === ':' ) {
  86                                                 // Path from GITfile is absolute
  87                                                 $this->basedir = $path;
  88                                         } else {
  89                                                 $this->basedir = $repoDir . DIRECTORY_SEPARATOR . $path;
  90                                         }
  91                                 }
  92                         }
  93                 }
  94         }
  95
  96         /**
  97          * Compute the path to the cache file for a given directory.
  98          *
  99          * @param string $repoDir The root directory of the repo where .git can be found
 100          * @return string Path to GitInfo cache file in $wgGitInfoCacheDirectory or
 101          * fallback in the extension directory itself
 102          * @since 1.24
 103          */
 104         protected static function getCacheFilePath( $repoDir ) {
 105                 global $IP, $wgGitInfoCacheDirectory;
 106
 107                 if ( $wgGitInfoCacheDirectory ) {
 108                         // Convert both $IP and $repoDir to canonical paths to protect against
 109                         // $IP having changed between the settings files and runtime.
 110                         $realIP = realpath( $IP );
 111                         $repoName = realpath( $repoDir );
 112                         if ( $repoName === false ) {
 113                                 // Unit tests use fake path names
 114                                 $repoName = $repoDir;
 115                         }
 116                         if ( strpos( $repoName, $realIP ) === 0 ) {
 117                                 // Strip $IP from path
 118                                 $repoName = substr( $repoName, strlen( $realIP ) );
 119                         }
 120                         // Transform path to git repo to something we can safely embed in
 121                         // a filename
 122                         $repoName = strtr( $repoName, DIRECTORY_SEPARATOR, '-' );
 123                         $fileName = 'info' . $repoName . '.json';
 124                         $cachePath = "{$wgGitInfoCacheDirectory}/{$fileName}";
 125                         if ( is_readable( $cachePath ) ) {
 126                                 return $cachePath;
 127                         }
 128                 }
 129
 130                 return "$repoDir/gitinfo.json";
 131         }
 132
 133         /**
 134          * Get the singleton for the repo at $IP
 135          *
 136          * @return GitInfo
 137          */
 138         public static function repo() {
 139                 if ( is_null( self::$repo ) ) {
 140                         global $IP;
 141                         self::$repo = new self( $IP );
 142                 }
 143                 return self::$repo;
 144         }
 145
 146         /**
 147          * Check if a string looks like a hex encoded SHA1 hash
 148          *
 149          * @param string $str The string to check
 150          * @return bool Whether or not the string looks like a SHA1
 151          */
 152         public static function isSHA1( $str ) {
 153                 return !!preg_match( '/^[0-9A-F]{40}$/i', $str );
 154         }
 155
 156         /**
 157          * Get the HEAD of the repo (without any opening "ref: ")
 158          *
 159          * @return string|bool The HEAD (git reference or SHA1) or false
 160          */
 161         public function getHead() {
 162                 if ( !isset( $this->cache['head'] ) ) {
 163                         $headFile = "{$this->basedir}/HEAD";
 164                         $head = false;
 165
 166                         if ( is_readable( $headFile ) ) {
 167                                 $head = file_get_contents( $headFile );
 168
 169                                 if ( preg_match( "/ref: (.*)/", $head, $m ) ) {
 170                                         $head = rtrim( $m[1] );
 171                                 } else {
 172                                         $head = rtrim( $head );
 173                                 }
 174                         }
 175                         $this->cache['head'] = $head;
 176                 }
 177                 return $this->cache['head'];
 178         }
 179
 180         /**
 181          * Get the SHA1 for the current HEAD of the repo
 182          *
 183          * @return string|bool A SHA1 or false
 184          */
 185         public function getHeadSHA1() {
 186                 if ( !isset( $this->cache['headSHA1'] ) ) {
 187                         $head = $this->getHead();
 188                         $sha1 = false;
 189
 190                         // If detached HEAD may be a SHA1
 191                         if ( self::isSHA1( $head ) ) {
 192                                 $sha1 = $head;
 193                         } else {
 194                                 // If not a SHA1 it may be a ref:
 195                                 $refFile = "{$this->basedir}/{$head}";
 196                                 $packedRefs = "{$this->basedir}/packed-refs";
 197                                 $headRegex = preg_quote( $head, '/' );
 198                                 if ( is_readable( $refFile ) ) {
 199                                         $sha1 = rtrim( file_get_contents( $refFile ) );
 200                                 } elseif ( is_readable( $packedRefs ) &&
 201                                         preg_match( "/^([0-9A-Fa-f]{40}) $headRegex$/m", file_get_contents( $packedRefs ), $matches )
 202                                 ) {
 203                                         $sha1 = $matches[1];
 204                                 }
 205                         }
 206                         $this->cache['headSHA1'] = $sha1;
 207                 }
 208                 return $this->cache['headSHA1'];
 209         }
 210
 211         /**
 212          * Get the commit date of HEAD entry of the git code repository
 213          *
 214          * @since 1.22
 215          * @return int|bool Commit date (UNIX timestamp) or false
 216          */
 217         public function getHeadCommitDate() {
 218                 global $wgGitBin;
 219
 220                 if ( !isset( $this->cache['headCommitDate'] ) ) {
 221                         $date = false;
 222                         if ( is_file( $wgGitBin ) &&
 223                                 is_executable( $wgGitBin ) &&
 224                                 $this->getHead() !== false
 225                         ) {
 226                                 $cmd = [
 227                                         $wgGitBin,
 228                                         'show',
 229                                         '-s',
 230                                         '--format=format:%ct',
 231                                         'HEAD',
 232                                 ];
 233                                 $result = Shell::command( $cmd )
 234                                         ->environment( [ 'GIT_DIR' => $this->basedir ] )
 235                                         ->restrict( Shell::RESTRICT_DEFAULT | Shell::NO_NETWORK )
 236                                         ->whitelistPaths( [ $this->basedir ] )
 237                                         ->execute();
 238
 239                                 if ( $result->getExitCode() === 0 ) {
 240                                         $date = (int)$result->getStdout();
 241                                 }
 242                         }
 243                         $this->cache['headCommitDate'] = $date;
 244                 }
 245                 return $this->cache['headCommitDate'];
 246         }
 247
 248         /**
 249          * Get the name of the current branch, or HEAD if not found
 250          *
 251          * @return string|bool The branch name, HEAD, or false
 252          */
 253         public function getCurrentBranch() {
 254                 if ( !isset( $this->cache['branch'] ) ) {
 255                         $branch = $this->getHead();
 256                         if ( $branch &&
 257                                 preg_match( "#^refs/heads/(.*)$#", $branch, $m )
 258                         ) {
 259                                 $branch = $m[1];
 260                         }
 261                         $this->cache['branch'] = $branch;
 262                 }
 263                 return $this->cache['branch'];
 264         }
 265
 266         /**
 267          * Get an URL to a web viewer link to the HEAD revision.
 268          *
 269          * @return string|bool String if a URL is available or false otherwise
 270          */
 271         public function getHeadViewUrl() {
 272                 $url = $this->getRemoteUrl();
 273                 if ( $url === false ) {
 274                         return false;
 275                 }
 276                 foreach ( self::getViewers() as $repo => $viewer ) {
 277                         $pattern = '#^' . $repo . '$#';
 278                         if ( preg_match( $pattern, $url, $matches ) ) {
 279                                 $viewerUrl = preg_replace( $pattern, $viewer, $url );
 280                                 $headSHA1 = $this->getHeadSHA1();
 281                                 $replacements = [
 282                                         '%h' => substr( $headSHA1, 0, 7 ),
 283                                         '%H' => $headSHA1,
 284                                         '%r' => urlencode( $matches[1] ),
 285                                         '%R' => $matches[1],
 286                                 ];
 287                                 return strtr( $viewerUrl, $replacements );
 288                         }
 289                 }
 290                 return false;
 291         }
 292
 293         /**
 294          * Get the URL of the remote origin.
 295          * @return string|bool String if a URL is available or false otherwise.
 296          */
 297         protected function getRemoteUrl() {
 298                 if ( !isset( $this->cache['remoteURL'] ) ) {
 299                         $config = "{$this->basedir}/config";
 300                         $url = false;
 301                         if ( is_readable( $config ) ) {
 302                                 MediaWiki\suppressWarnings();
 303                                 $configArray = parse_ini_file( $config, true );
 304                                 MediaWiki\restoreWarnings();
 305                                 $remote = false;
 306
 307                                 // Use the "origin" remote repo if available or any other repo if not.
 308                                 if ( isset( $configArray['remote origin'] ) ) {
 309                                         $remote = $configArray['remote origin'];
 310                                 } elseif ( is_array( $configArray ) ) {
 311                                         foreach ( $configArray as $sectionName => $sectionConf ) {
 312                                                 if ( substr( $sectionName, 0, 6 ) == 'remote' ) {
 313                                                         $remote = $sectionConf;
 314                                                 }
 315                                         }
 316                                 }
 317
 318                                 if ( $remote !== false && isset( $remote['url'] ) ) {
 319                                         $url = $remote['url'];
 320                                 }
 321                         }
 322                         $this->cache['remoteURL'] = $url;
 323                 }
 324                 return $this->cache['remoteURL'];
 325         }
 326
 327         /**
 328          * Check to see if the current cache is fully populated.
 329          *
 330          * Note: This method is public only to make unit testing easier. There's
 331          * really no strong reason that anything other than a test should want to
 332          * call this method.
 333          *
 334          * @return bool True if all expected cache keys exist, false otherwise
 335          */
 336         public function cacheIsComplete() {
 337                 return isset( $this->cache['head'] ) &&
 338                         isset( $this->cache['headSHA1'] ) &&
 339                         isset( $this->cache['headCommitDate'] ) &&
 340                         isset( $this->cache['branch'] ) &&
 341                         isset( $this->cache['remoteURL'] );
 342         }
 343
 344         /**
 345          * Precompute and cache git information.
 346          *
 347          * Creates a JSON file in the cache directory associated with this
 348          * GitInfo instance. This cache file will be used by subsequent GitInfo objects referencing
 349          * the same directory to avoid needing to examine the .git directory again.
 350          *
 351          * @since 1.24
 352          */
 353         public function precomputeValues() {
 354                 if ( $this->cacheFile !== null ) {
 355                         // Try to completely populate the cache
 356                         $this->getHead();
 357                         $this->getHeadSHA1();
 358                         $this->getHeadCommitDate();
 359                         $this->getCurrentBranch();
 360                         $this->getRemoteUrl();
 361
 362                         if ( !$this->cacheIsComplete() ) {
 363                                 wfDebugLog( 'gitinfo',
 364                                         "Failed to compute GitInfo for \"{$this->basedir}\""
 365                                 );
 366                                 return;
 367                         }
 368
 369                         $cacheDir = dirname( $this->cacheFile );
 370                         if ( !file_exists( $cacheDir ) &&
 371                                 !wfMkdirParents( $cacheDir, null, __METHOD__ )
 372                         ) {
 373                                 throw new MWException( "Unable to create GitInfo cache \"{$cacheDir}\"" );
 374                         }
 375
 376                         file_put_contents( $this->cacheFile, FormatJson::encode( $this->cache ) );
 377                 }
 378         }
 379
 380         /**
 381          * @see self::getHeadSHA1
 382          * @return string
 383          */
 384         public static function headSHA1() {
 385                 return self::repo()->getHeadSHA1();
 386         }
 387
 388         /**
 389          * @see self::getCurrentBranch
 390          * @return string
 391          */
 392         public static function currentBranch() {
 393                 return self::repo()->getCurrentBranch();
 394         }
 395
 396         /**
 397          * @see self::getHeadViewUrl()
 398          * @return bool|string
 399          */
 400         public static function headViewUrl() {
 401                 return self::repo()->getHeadViewUrl();
 402         }
 403
 404         /**
 405          * Gets the list of repository viewers
 406          * @return array
 407          */
 408         protected static function getViewers() {
 409                 global $wgGitRepositoryViewers;
 410
 411                 if ( self::$viewers === false ) {
 412                         self::$viewers = $wgGitRepositoryViewers;
 413                         Hooks::run( 'GitViewers', [ &self::$viewers ] );
 414                 }
 415
 416                 return self::$viewers;
 417         }
 418 }