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