| 1 | RewriteEngine On |
|
| 2 | ||
| 3 | RewriteCond %{HTTP_USER_AGENT} SemrushBot [NC] |
|
| 4 | RewriteRule .* - [F] |
|
| 5 | ||
| 6 | RewriteRule ^(git|pages|render)/ - [F] |
|
| 7 | ||
| 8 | RewriteCond %{REQUEST_URI} !^/index\.php$ |
|
| 9 | RewriteCond %{REQUEST_FILENAME} -f |
|
| 10 | RewriteRule \.php$ - [F] |
|
| 11 | ||
| 12 | RewriteCond %{REQUEST_FILENAME} !-f |
|
| 13 | RewriteRule ^(.*)$ index.php [QSA,L] |
|
| 14 | ||
| 15 | deny from 185.213.83.81 |
|
| 16 | ||
| 17 | 1 |
| 9 | 9 | ini_set( 'log_errors', 1 ); |
| 10 | 10 | ini_set( 'error_log', __DIR__ . self::LOG_FILE ); |
| 11 | ||
| 12 | set_time_limit( 0 ); |
|
| 11 | 13 | } |
| 12 | 14 |
| 1 | # Requirements |
|
| 2 | ||
| 3 | TreeTrek was built and tested using the following software components: |
|
| 4 | ||
| 5 | * Debian Linux 13 (kernel 6.12.69) |
|
| 6 | * nginx 1.26.3 |
|
| 7 | * PHP 8.4.16 |
|
| 8 | * Git 2.53.0 |
|
| 9 | ||
| 10 | # Installation |
|
| 11 | ||
| 12 | To install: |
|
| 13 | ||
| 14 | * Download code |
|
| 15 | * Configure settings |
|
| 16 | * Configure web server routing and security |
|
| 17 | * Define display order |
|
| 18 | ||
| 19 | Replace placeholder values with the configuration details. |
|
| 20 | ||
| 21 | 1. Set placeholders: |
|
| 22 | ``` bash |
|
| 23 | export REPO=repo.domain.com |
|
| 24 | export OWNER=username |
|
| 25 | export WEBDIR=/var/www |
|
| 26 | ``` |
|
| 27 | 1. Download code: |
|
| 28 | ```bash |
|
| 29 | mkdir -p /var/www/${REPO} |
|
| 30 | cd /var/www/${REPO} |
|
| 31 | git clone https://repo.autonoma.ca/repo/treetrek |
|
| 32 | sudo usermod -aG ${OWNER} www-data |
|
| 33 | ``` |
|
| 34 | 1. Edit `${WEBDIR}/${REPO}/Config.php`. |
|
| 35 | 1. Set `SITE_TITLE`. |
|
| 36 | 1. Set `REPOS_SUBDIR`. |
|
| 37 | 1. Save the file. |
|
| 38 | 1. Edit `${WEBDIR}/${REPO}/order.txt` |
|
| 39 | 1. Add repositories you want to come first, hide using a hyphen prefix. |
|
| 40 | 1. Edit nginx configuration file. |
|
| 41 | 1. Add routing and security rules: |
|
| 42 | ```nginx |
|
| 43 | location ^~ /images/ { |
|
| 44 | allow all; |
|
| 45 | } |
|
| 46 | ||
| 47 | location ~* ^/images/.*\.svg$ { |
|
| 48 | add_header Content-Type image/svg+xml; |
|
| 49 | } |
|
| 50 | ||
| 51 | location ~ ^/(git|pages|render)/ { |
|
| 52 | deny all; |
|
| 53 | return 403; |
|
| 54 | } |
|
| 55 | ||
| 56 | location ^~ /repo/ { |
|
| 57 | try_files $uri $uri/ /index.php?$query_string; |
|
| 58 | } |
|
| 59 | ||
| 60 | location / { |
|
| 61 | try_files $uri $uri/ /index.php?$query_string; |
|
| 62 | } |
|
| 63 | ||
| 64 | location = /index.php { |
|
| 65 | include snippets/fastcgi-php.conf; |
|
| 66 | fastcgi_pass unix:/run/php/php8.4-fpm.sock; |
|
| 67 | } |
|
| 68 | ||
| 69 | location ~ \.php$ { |
|
| 70 | deny all; |
|
| 71 | return 403; |
|
| 72 | } |
|
| 73 | ``` |
|
| 74 | 1. Apply changes: |
|
| 75 | ```bash |
|
| 76 | sudo nginx -t && sudo systemctl reload nginx |
|
| 77 | ``` |
|
| 78 | ||
| 79 | The application is installed. |
|
| 80 | ||
| 1 | 81 |
| 189 | 189 | |
| 190 | 190 | if( $real && strpos( $real, $repo ) === 0 ) { |
| 191 | $result = readfile( $path ) !== false; |
|
| 192 | } |
|
| 193 | } |
|
| 194 | } |
|
| 195 | ||
| 196 | return $result; |
|
| 197 | } |
|
| 198 | ||
| 199 | public function eachRef( callable $callback ): void { |
|
| 200 | $head = $this->resolve( 'HEAD' ); |
|
| 201 | ||
| 202 | if( $head !== '' ) { |
|
| 203 | $callback( 'HEAD', $head ); |
|
| 204 | } |
|
| 205 | ||
| 206 | $this->refs->scanRefs( 'refs/heads', function( $n, $s ) use ( $callback ) { |
|
| 207 | $callback( "refs/heads/$n", $s ); |
|
| 208 | } ); |
|
| 209 | ||
| 210 | $this->refs->scanRefs( 'refs/tags', function( $n, $s ) use ( $callback ) { |
|
| 211 | $callback( "refs/tags/$n", $s ); |
|
| 212 | } ); |
|
| 213 | } |
|
| 214 | ||
| 215 | public function generatePackfile( array $objs ): string { |
|
| 216 | $pData = "PACK" . pack( 'N', 2 ) . pack( 'N', 0 ); |
|
| 217 | ||
| 218 | if( !empty( $objs ) ) { |
|
| 219 | $data = ''; |
|
| 220 | ||
| 221 | foreach( $objs as $sha => $info ) { |
|
| 222 | $cont = $this->read( $sha ); |
|
| 223 | $size = strlen( $cont ); |
|
| 224 | $byte = $info['type'] << 4 | $size & 0x0f; |
|
| 225 | $size >>= 4; |
|
| 226 | ||
| 227 | while( $size > 0 ) { |
|
| 228 | $data .= chr( $byte | 0x80 ); |
|
| 229 | $byte = $size & 0x7f; |
|
| 230 | $size >>= 7; |
|
| 231 | } |
|
| 232 | ||
| 233 | $data .= chr( $byte ) . gzcompress( $cont ); |
|
| 234 | } |
|
| 235 | ||
| 236 | $pData = "PACK" . pack( 'N', 2 ) . pack( 'N', count( $objs ) ) . $data; |
|
| 237 | } |
|
| 238 | ||
| 239 | return $pData . hash( 'sha1', $pData, true ); |
|
| 240 | } |
|
| 241 | ||
| 242 | private function getTreeSha( string $commitOrTreeSha ): string { |
|
| 243 | $data = $this->read( $commitOrTreeSha ); |
|
| 244 | $sha = $commitOrTreeSha; |
|
| 245 | ||
| 246 | if( preg_match( '/^object ([0-9a-f]{40})/m', $data, $matches ) ) { |
|
| 247 | $sha = $this->getTreeSha( $matches[1] ); |
|
| 248 | } |
|
| 249 | ||
| 250 | if( $sha === $commitOrTreeSha && |
|
| 251 | preg_match( '/^tree ([0-9a-f]{40})/m', $data, $matches ) ) { |
|
| 252 | $sha = $matches[1]; |
|
| 253 | } |
|
| 254 | ||
| 255 | return $sha; |
|
| 256 | } |
|
| 257 | ||
| 258 | private function resolvePath( string $treeSha, string $path ): array { |
|
| 259 | $parts = explode( '/', trim( $path, '/' ) ); |
|
| 260 | $sha = $treeSha; |
|
| 261 | $mode = '40000'; |
|
| 262 | ||
| 263 | foreach( $parts as $part ) { |
|
| 264 | $entry = [ 'sha' => '', 'mode' => '' ]; |
|
| 265 | ||
| 266 | if( $part !== '' && $sha !== '' ) { |
|
| 267 | $entry = $this->findTreeEntry( $sha, $part ); |
|
| 268 | } |
|
| 269 | ||
| 270 | $sha = $entry['sha']; |
|
| 271 | $mode = $entry['mode']; |
|
| 272 | } |
|
| 273 | ||
| 274 | return [ |
|
| 275 | 'sha' => $sha, |
|
| 276 | 'mode' => $mode, |
|
| 277 | 'isDir' => $mode === '40000' || $mode === '040000' |
|
| 278 | ]; |
|
| 279 | } |
|
| 280 | ||
| 281 | private function findTreeEntry( string $treeSha, string $name ): array { |
|
| 282 | $data = $this->read( $treeSha ); |
|
| 283 | $pos = 0; |
|
| 284 | $len = strlen( $data ); |
|
| 285 | $entry = [ 'sha' => '', 'mode' => '' ]; |
|
| 286 | ||
| 287 | while( $pos < $len ) { |
|
| 288 | $space = strpos( $data, ' ', $pos ); |
|
| 289 | $eos = strpos( $data, "\0", $space ); |
|
| 290 | ||
| 291 | if( $space === false || $eos === false ) { |
|
| 292 | break; |
|
| 293 | } |
|
| 294 | ||
| 295 | if( substr( $data, $space + 1, $eos - $space - 1 ) === $name ) { |
|
| 296 | $entry = [ |
|
| 297 | 'sha' => bin2hex( substr( $data, $eos + 1, 20 ) ), |
|
| 298 | 'mode' => substr( $data, $pos, $space - $pos ) |
|
| 299 | ]; |
|
| 300 | break; |
|
| 301 | } |
|
| 302 | ||
| 303 | $pos = $eos + 21; |
|
| 304 | } |
|
| 305 | ||
| 306 | return $entry; |
|
| 307 | } |
|
| 308 | ||
| 309 | private function parseTagData( |
|
| 310 | string $name, |
|
| 311 | string $sha, |
|
| 312 | string $data |
|
| 313 | ): Tag { |
|
| 314 | $isAnn = strncmp( $data, 'object ', 7 ) === 0; |
|
| 315 | $pattern = $isAnn |
|
| 316 | ? '/^tagger (.*) <(.*)> (\d+) [+\-]\d{4}$/m' |
|
| 317 | : '/^author (.*) <(.*)> (\d+) [+\-]\d{4}$/m'; |
|
| 318 | $id = $this->parseIdentity( $data, $pattern ); |
|
| 319 | $target = $isAnn |
|
| 320 | ? $this->extractPattern( $data, '/^object (.*)$/m', 1, $sha ) |
|
| 321 | : $sha; |
|
| 322 | ||
| 323 | return new Tag( |
|
| 324 | $name, |
|
| 325 | $sha, |
|
| 326 | $target, |
|
| 327 | $id['timestamp'], |
|
| 328 | $this->extractMessage( $data ), |
|
| 329 | $id['name'] |
|
| 330 | ); |
|
| 331 | } |
|
| 332 | ||
| 333 | private function extractPattern( |
|
| 334 | string $data, |
|
| 335 | string $pattern, |
|
| 336 | int $group, |
|
| 337 | string $default = '' |
|
| 338 | ): string { |
|
| 339 | return preg_match( $pattern, $data, $matches ) |
|
| 340 | ? $matches[$group] |
|
| 341 | : $default; |
|
| 342 | } |
|
| 343 | ||
| 344 | private function parseIdentity( string $data, string $pattern ): array { |
|
| 345 | $found = preg_match( $pattern, $data, $matches ); |
|
| 346 | ||
| 347 | return [ |
|
| 348 | 'name' => $found ? trim( $matches[1] ) : 'Unknown', |
|
| 349 | 'email' => $found ? $matches[2] : '', |
|
| 350 | 'timestamp' => $found ? (int)$matches[3] : 0 |
|
| 351 | ]; |
|
| 352 | } |
|
| 353 | ||
| 354 | private function extractMessage( string $data ): string { |
|
| 355 | $pos = strpos( $data, "\n\n" ); |
|
| 356 | ||
| 357 | return $pos !== false ? trim( substr( $data, $pos + 2 ) ) : ''; |
|
| 358 | } |
|
| 359 | ||
| 360 | private function slurp( string $sha, callable $callback ): void { |
|
| 361 | $path = $this->getLoosePath( $sha ); |
|
| 362 | ||
| 363 | if( is_file( $path ) ) { |
|
| 364 | $this->slurpLooseObject( $path, $callback ); |
|
| 365 | } else { |
|
| 366 | $this->slurpPackedObject( $sha, $callback ); |
|
| 367 | } |
|
| 368 | } |
|
| 369 | ||
| 370 | private function slurpLooseObject( string $path, callable $callback ): void { |
|
| 371 | $this->iterateInflated( |
|
| 372 | $path, |
|
| 373 | function( $chunk ) use ( $callback ) { |
|
| 374 | if( $chunk !== '' ) { |
|
| 375 | $callback( $chunk ); |
|
| 376 | } |
|
| 377 | return true; |
|
| 378 | } |
|
| 379 | ); |
|
| 380 | } |
|
| 381 | ||
| 382 | private function slurpPackedObject( string $sha, callable $callback ): void { |
|
| 383 | $streamed = $this->packs->stream( $sha, $callback ); |
|
| 384 | ||
| 385 | if( !$streamed ) { |
|
| 386 | $data = $this->packs->read( $sha ); |
|
| 387 | ||
| 388 | if( $data !== '' ) { |
|
| 389 | $callback( $data ); |
|
| 390 | } |
|
| 391 | } |
|
| 392 | } |
|
| 393 | ||
| 394 | private function iterateInflated( |
|
| 395 | string $path, |
|
| 396 | callable $processor |
|
| 397 | ): void { |
|
| 398 | $handle = fopen( $path, 'rb' ); |
|
| 399 | $infl = $handle ? inflate_init( ZLIB_ENCODING_DEFLATE ) : null; |
|
| 400 | $found = false; |
|
| 401 | $buffer = ''; |
|
| 402 | ||
| 403 | if( $handle && $infl ) { |
|
| 404 | while( !feof( $handle ) ) { |
|
| 405 | $chunk = fread( $handle, 16384 ); |
|
| 406 | $inflated = inflate_add( $infl, $chunk ); |
|
| 407 | ||
| 408 | if( $inflated === false ) { |
|
| 409 | break; |
|
| 410 | } |
|
| 411 | ||
| 412 | if( !$found ) { |
|
| 413 | $buffer .= $inflated; |
|
| 414 | $eos = strpos( $buffer, "\0" ); |
|
| 415 | ||
| 416 | if( $eos !== false ) { |
|
| 417 | $found = true; |
|
| 418 | $body = substr( $buffer, $eos + 1 ); |
|
| 419 | $head = substr( $buffer, 0, $eos ); |
|
| 420 | ||
| 421 | if( $processor( $body, $head ) === false ) { |
|
| 422 | break; |
|
| 423 | } |
|
| 424 | } |
|
| 425 | } elseif( $processor( $inflated, null ) === false ) { |
|
| 426 | break; |
|
| 427 | } |
|
| 428 | } |
|
| 429 | ||
| 430 | fclose( $handle ); |
|
| 431 | } |
|
| 432 | } |
|
| 433 | ||
| 434 | private function peekLooseObject( string $sha, int $length ): string { |
|
| 435 | $path = $this->getLoosePath( $sha ); |
|
| 436 | $buf = ''; |
|
| 437 | ||
| 438 | if( is_file( $path ) ) { |
|
| 439 | $this->iterateInflated( |
|
| 440 | $path, |
|
| 441 | function( $chunk ) use ( $length, &$buf ) { |
|
| 442 | $buf .= $chunk; |
|
| 443 | return strlen( $buf ) < $length; |
|
| 444 | } |
|
| 445 | ); |
|
| 446 | } |
|
| 447 | ||
| 448 | return substr( $buf, 0, $length ); |
|
| 449 | } |
|
| 450 | ||
| 451 | private function parseCommit( string $sha ): object { |
|
| 452 | $data = $this->read( $sha ); |
|
| 453 | $result = (object)[ 'sha' => '' ]; |
|
| 454 | ||
| 455 | if( $data !== '' ) { |
|
| 456 | $id = $this->parseIdentity( |
|
| 457 | $data, |
|
| 458 | '/^author (.*) <(.*)> (\d+)/m' |
|
| 459 | ); |
|
| 460 | ||
| 461 | $result = (object)[ |
|
| 462 | 'sha' => $sha, |
|
| 463 | 'message' => $this->extractMessage( $data ), |
|
| 464 | 'author' => $id['name'], |
|
| 465 | 'email' => $id['email'], |
|
| 466 | 'date' => $id['timestamp'], |
|
| 467 | 'parentSha' => $this->extractPattern( $data, '/^parent (.*)$/m', 1 ) |
|
| 468 | ]; |
|
| 469 | } |
|
| 470 | ||
| 471 | return $result; |
|
| 472 | } |
|
| 473 | ||
| 474 | private function walkTree( string $sha, callable $callback ): void { |
|
| 475 | $data = $this->read( $sha ); |
|
| 476 | $tree = $data; |
|
| 477 | ||
| 478 | if( $data !== '' && preg_match( '/^tree (.*)$/m', $data, $m ) ) { |
|
| 479 | $tree = $this->read( $m[1] ); |
|
| 480 | } |
|
| 481 | ||
| 482 | if( $tree !== '' && $this->isTreeData( $tree ) ) { |
|
| 483 | $this->processTree( $tree, $callback ); |
|
| 484 | } |
|
| 485 | } |
|
| 486 | ||
| 487 | private function processTree( string $data, callable $callback ): void { |
|
| 488 | $pos = 0; |
|
| 489 | $len = strlen( $data ); |
|
| 490 | ||
| 491 | while( $pos < $len ) { |
|
| 492 | $space = strpos( $data, ' ', $pos ); |
|
| 493 | $eos = strpos( $data, "\0", $space ); |
|
| 494 | $entry = null; |
|
| 495 | ||
| 496 | if( $space !== false && $eos !== false && $eos + 21 <= $len ) { |
|
| 497 | $mode = substr( $data, $pos, $space - $pos ); |
|
| 498 | $sha = bin2hex( substr( $data, $eos + 1, 20 ) ); |
|
| 499 | $isD = $mode === '40000' || $mode === '040000'; |
|
| 500 | ||
| 501 | $entry = [ |
|
| 502 | 'file' => new File( |
|
| 503 | substr( $data, $space + 1, $eos - $space - 1 ), |
|
| 504 | $sha, |
|
| 505 | $mode, |
|
| 506 | 0, |
|
| 507 | $isD ? 0 : $this->getObjectSize( $sha ), |
|
| 508 | $isD ? '' : $this->peek( $sha ) |
|
| 509 | ), |
|
| 510 | 'nextPosition' => $eos + 21 |
|
| 511 | ]; |
|
| 512 | } |
|
| 513 | ||
| 514 | if( $entry === null ) { |
|
| 515 | break; |
|
| 516 | } |
|
| 517 | ||
| 518 | $callback( $entry['file'] ); |
|
| 519 | $pos = $entry['nextPosition']; |
|
| 520 | } |
|
| 521 | } |
|
| 522 | ||
| 523 | private function isTreeData( string $data ): bool { |
|
| 524 | $len = strlen( $data ); |
|
| 525 | $patt = '/^(40000|100644|100755|120000|160000) /'; |
|
| 526 | $match = $len >= 25 && preg_match( $patt, $data ); |
|
| 527 | $eos = $match ? strpos( $data, "\0" ) : false; |
|
| 528 | ||
| 529 | return $match && $eos !== false && $eos + 21 <= $len; |
|
| 530 | } |
|
| 531 | ||
| 532 | private function getLoosePath( string $sha ): string { |
|
| 533 | return "{$this->objPath}/" . substr( $sha, 0, 2 ) . "/" . |
|
| 534 | substr( $sha, 2 ); |
|
| 535 | } |
|
| 536 | ||
| 537 | private function getLooseObjectSize( string $sha ): int { |
|
| 538 | $path = $this->getLoosePath( $sha ); |
|
| 539 | $size = 0; |
|
| 540 | ||
| 541 | if( is_file( $path ) ) { |
|
| 542 | $this->iterateInflated( |
|
| 543 | $path, |
|
| 544 | function( $c, $head ) use ( &$size ) { |
|
| 545 | if( $head !== null ) { |
|
| 546 | $parts = explode( ' ', $head ); |
|
| 547 | $size = isset( $parts[1] ) ? (int)$parts[1] : 0; |
|
| 548 | } |
|
| 549 | return false; |
|
| 550 | } |
|
| 551 | ); |
|
| 552 | } |
|
| 553 | ||
| 554 | return $size; |
|
| 555 | } |
|
| 556 | ||
| 557 | public function collectObjects( array $wants, array $haves = [] ): array { |
|
| 558 | $objs = []; |
|
| 559 | $result = []; |
|
| 560 | ||
| 561 | foreach( $wants as $sha ) { |
|
| 562 | $objs = $this->collectObjectsRecursive( $sha, $objs, 0 ); |
|
| 563 | } |
|
| 564 | ||
| 565 | foreach( $haves as $sha ) { |
|
| 566 | if( isset( $objs[$sha] ) ) { |
|
| 567 | unset( $objs[$sha] ); |
|
| 568 | } |
|
| 569 | } |
|
| 570 | ||
| 571 | $result = $objs; |
|
| 572 | ||
| 573 | return $result; |
|
| 574 | } |
|
| 575 | ||
| 576 | private function collectObjectsRecursive( |
|
| 577 | string $sha, |
|
| 578 | array $objs, |
|
| 579 | int $expectedType = 0 |
|
| 580 | ): array { |
|
| 581 | $result = $objs; |
|
| 582 | ||
| 583 | if( !isset( $result[$sha] ) ) { |
|
| 584 | $data = $this->read( $sha ); |
|
| 585 | $type = $expectedType === 0 |
|
| 586 | ? $this->getObjectType( $data ) |
|
| 587 | : $expectedType; |
|
| 588 | ||
| 589 | $result[$sha] = [ |
|
| 590 | 'type' => $type, |
|
| 591 | 'size' => strlen( $data ) |
|
| 592 | ]; |
|
| 593 | ||
| 594 | if( $type === 1 ) { |
|
| 595 | $hasTree = preg_match( '/^tree ([0-9a-f]{40})/m', $data, $m ); |
|
| 596 | ||
| 597 | if( $hasTree ) { |
|
| 598 | $result = $this->collectObjectsRecursive( $m[1], $result, 2 ); |
|
| 599 | } |
|
| 600 | ||
| 601 | $hasParents = preg_match_all( |
|
| 602 | '/^parent ([0-9a-f]{40})/m', |
|
| 603 | $data, |
|
| 604 | $m |
|
| 605 | ); |
|
| 606 | ||
| 607 | if( $hasParents ) { |
|
| 608 | foreach( $m[1] as $parentSha ) { |
|
| 609 | $result = $this->collectObjectsRecursive( |
|
| 610 | $parentSha, |
|
| 611 | $result, |
|
| 612 | 1 |
|
| 613 | ); |
|
| 614 | } |
|
| 615 | } |
|
| 616 | } |
|
| 617 | ||
| 618 | if( $type === 2 ) { |
|
| 619 | $pos = 0; |
|
| 620 | $len = strlen( $data ); |
|
| 621 | ||
| 622 | while( $pos < $len ) { |
|
| 623 | $sp = strpos( $data, ' ', $pos ); |
|
| 624 | $eos = strpos( $data, "\0", $sp ); |
|
| 625 | ||
| 626 | if( $sp === false || $eos === false ) { |
|
| 627 | break; |
|
| 628 | } |
|
| 629 | ||
| 630 | $mode = substr( $data, $pos, $sp - $pos ); |
|
| 631 | $s = bin2hex( substr( $data, $eos + 1, 20 ) ); |
|
| 632 | $isDir = $mode === '40000' || $mode === '040000'; |
|
| 633 | $nextType = $isDir ? 2 : 3; |
|
| 634 | $pos = $eos + 21; |
|
| 635 | ||
| 636 | $result = $this->collectObjectsRecursive( |
|
| 637 | $s, |
|
| 638 | $result, |
|
| 639 | $nextType |
|
| 640 | ); |
|
| 641 | } |
|
| 642 | } |
|
| 643 | ||
| 644 | $isTagTarget = $type === 4 && |
|
| 645 | preg_match( '/^object ([0-9a-f]{40})/m', $data, $m ); |
|
| 646 | ||
| 647 | if( $isTagTarget ) { |
|
| 648 | $nextType = 1; |
|
| 649 | ||
| 650 | if( preg_match( '/^type (commit|tree|blob|tag)/m', $data, $t ) ) { |
|
| 651 | $map = [ |
|
| 652 | 'commit' => 1, |
|
| 653 | 'tree' => 2, |
|
| 654 | 'blob' => 3, |
|
| 655 | 'tag' => 4 |
|
| 656 | ]; |
|
| 657 | ||
| 658 | $nextType = $map[$t[1]] ?? 1; |
|
| 659 | } |
|
| 660 | ||
| 661 | $result = $this->collectObjectsRecursive( |
|
| 662 | $m[1], |
|
| 663 | $result, |
|
| 664 | $nextType |
|
| 665 | ); |
|
| 666 | } |
|
| 667 | } |
|
| 668 | ||
| 669 | return $result; |
|
| 191 | $result = $this->streamFileContent( $path ); |
|
| 192 | } |
|
| 193 | } |
|
| 194 | } |
|
| 195 | ||
| 196 | return $result; |
|
| 197 | } |
|
| 198 | ||
| 199 | private function streamFileContent( string $path ): bool { |
|
| 200 | $result = false; |
|
| 201 | ||
| 202 | if( $path !== '' ) { |
|
| 203 | header( 'X-Accel-Redirect: ' . $path ); |
|
| 204 | header( 'Content-Type: application/octet-stream' ); |
|
| 205 | ||
| 206 | $result = true; |
|
| 207 | } |
|
| 208 | ||
| 209 | return $result; |
|
| 210 | } |
|
| 211 | ||
| 212 | public function eachRef( callable $callback ): void { |
|
| 213 | $head = $this->resolve( 'HEAD' ); |
|
| 214 | ||
| 215 | if( $head !== '' ) { |
|
| 216 | $callback( 'HEAD', $head ); |
|
| 217 | } |
|
| 218 | ||
| 219 | $this->refs->scanRefs( 'refs/heads', function( $n, $s ) use ( $callback ) { |
|
| 220 | $callback( "refs/heads/$n", $s ); |
|
| 221 | } ); |
|
| 222 | ||
| 223 | $this->refs->scanRefs( 'refs/tags', function( $n, $s ) use ( $callback ) { |
|
| 224 | $callback( "refs/tags/$n", $s ); |
|
| 225 | } ); |
|
| 226 | } |
|
| 227 | ||
| 228 | public function generatePackfile( array $objs ): Generator { |
|
| 229 | $ctx = hash_init( 'sha1' ); |
|
| 230 | $head = "PACK" . pack( 'N', 2 ) . pack( 'N', count( $objs ) ); |
|
| 231 | ||
| 232 | hash_update( $ctx, $head ); |
|
| 233 | yield $head; |
|
| 234 | ||
| 235 | foreach( $objs as $sha => $type ) { |
|
| 236 | $size = $this->getObjectSize( $sha ); |
|
| 237 | $byte = $type << 4 | $size & 0x0f; |
|
| 238 | $sz = $size >> 4; |
|
| 239 | $hdr = ''; |
|
| 240 | ||
| 241 | while( $sz > 0 ) { |
|
| 242 | $hdr .= chr( $byte | 0x80 ); |
|
| 243 | $byte = $sz & 0x7f; |
|
| 244 | $sz >>= 7; |
|
| 245 | } |
|
| 246 | ||
| 247 | $hdr .= chr( $byte ); |
|
| 248 | hash_update( $ctx, $hdr ); |
|
| 249 | yield $hdr; |
|
| 250 | ||
| 251 | $deflate = deflate_init( ZLIB_ENCODING_DEFLATE ); |
|
| 252 | ||
| 253 | foreach( $this->slurpChunks( $sha ) as $raw ) { |
|
| 254 | $compressed = deflate_add( $deflate, $raw, ZLIB_NO_FLUSH ); |
|
| 255 | ||
| 256 | if( $compressed !== '' ) { |
|
| 257 | hash_update( $ctx, $compressed ); |
|
| 258 | yield $compressed; |
|
| 259 | } |
|
| 260 | } |
|
| 261 | ||
| 262 | $final = deflate_add( $deflate, '', ZLIB_FINISH ); |
|
| 263 | ||
| 264 | if( $final !== '' ) { |
|
| 265 | hash_update( $ctx, $final ); |
|
| 266 | yield $final; |
|
| 267 | } |
|
| 268 | } |
|
| 269 | ||
| 270 | yield hash_final( $ctx, true ); |
|
| 271 | } |
|
| 272 | ||
| 273 | private function slurpChunks( string $sha ): Generator { |
|
| 274 | $path = $this->getLoosePath( $sha ); |
|
| 275 | ||
| 276 | if( is_file( $path ) ) { |
|
| 277 | yield from $this->looseObjectChunks( $path ); |
|
| 278 | } else { |
|
| 279 | $any = false; |
|
| 280 | ||
| 281 | foreach( $this->packs->streamGenerator( $sha ) as $chunk ) { |
|
| 282 | $any = true; |
|
| 283 | yield $chunk; |
|
| 284 | } |
|
| 285 | ||
| 286 | if( !$any ) { |
|
| 287 | $data = $this->packs->read( $sha ); |
|
| 288 | ||
| 289 | if( $data !== '' ) { |
|
| 290 | yield $data; |
|
| 291 | } |
|
| 292 | } |
|
| 293 | } |
|
| 294 | } |
|
| 295 | ||
| 296 | private function looseObjectChunks( string $path ): Generator { |
|
| 297 | $handle = fopen( $path, 'rb' ); |
|
| 298 | $infl = $handle ? inflate_init( ZLIB_ENCODING_DEFLATE ) : null; |
|
| 299 | ||
| 300 | if( !$handle || !$infl ) { |
|
| 301 | return; |
|
| 302 | } |
|
| 303 | ||
| 304 | $found = false; |
|
| 305 | $buffer = ''; |
|
| 306 | ||
| 307 | while( !feof( $handle ) ) { |
|
| 308 | $chunk = fread( $handle, 16384 ); |
|
| 309 | $inflated = inflate_add( $infl, $chunk ); |
|
| 310 | ||
| 311 | if( $inflated === false ) { |
|
| 312 | break; |
|
| 313 | } |
|
| 314 | ||
| 315 | if( !$found ) { |
|
| 316 | $buffer .= $inflated; |
|
| 317 | $eos = strpos( $buffer, "\0" ); |
|
| 318 | ||
| 319 | if( $eos !== false ) { |
|
| 320 | $found = true; |
|
| 321 | $body = substr( $buffer, $eos + 1 ); |
|
| 322 | ||
| 323 | if( $body !== '' ) { |
|
| 324 | yield $body; |
|
| 325 | } |
|
| 326 | ||
| 327 | $buffer = ''; |
|
| 328 | } |
|
| 329 | } elseif( $inflated !== '' ) { |
|
| 330 | yield $inflated; |
|
| 331 | } |
|
| 332 | } |
|
| 333 | ||
| 334 | fclose( $handle ); |
|
| 335 | } |
|
| 336 | ||
| 337 | private function streamCompressedObject( string $sha, $ctx ): Generator { |
|
| 338 | $deflate = deflate_init( ZLIB_ENCODING_DEFLATE ); |
|
| 339 | $buffer = ''; |
|
| 340 | ||
| 341 | $this->slurp( $sha, function( $chunk ) use ( |
|
| 342 | $deflate, |
|
| 343 | $ctx, |
|
| 344 | &$buffer |
|
| 345 | ) { |
|
| 346 | $compressed = deflate_add( $deflate, $chunk, ZLIB_NO_FLUSH ); |
|
| 347 | ||
| 348 | if( $compressed !== '' ) { |
|
| 349 | hash_update( $ctx, $compressed ); |
|
| 350 | $buffer .= $compressed; |
|
| 351 | } |
|
| 352 | } ); |
|
| 353 | ||
| 354 | $final = deflate_add( $deflate, '', ZLIB_FINISH ); |
|
| 355 | ||
| 356 | if( $final !== '' ) { |
|
| 357 | hash_update( $ctx, $final ); |
|
| 358 | $buffer .= $final; |
|
| 359 | } |
|
| 360 | ||
| 361 | $pos = 0; |
|
| 362 | $len = strlen( $buffer ); |
|
| 363 | ||
| 364 | while( $pos < $len ) { |
|
| 365 | $chunk = substr( $buffer, $pos, 32768 ); |
|
| 366 | yield $chunk; |
|
| 367 | $pos += 32768; |
|
| 368 | } |
|
| 369 | } |
|
| 370 | ||
| 371 | private function getTreeSha( string $commitOrTreeSha ): string { |
|
| 372 | $data = $this->read( $commitOrTreeSha ); |
|
| 373 | $sha = $commitOrTreeSha; |
|
| 374 | ||
| 375 | if( preg_match( '/^object ([0-9a-f]{40})/m', $data, $matches ) ) { |
|
| 376 | $sha = $this->getTreeSha( $matches[1] ); |
|
| 377 | } |
|
| 378 | ||
| 379 | if( $sha === $commitOrTreeSha && |
|
| 380 | preg_match( '/^tree ([0-9a-f]{40})/m', $data, $matches ) ) { |
|
| 381 | $sha = $matches[1]; |
|
| 382 | } |
|
| 383 | ||
| 384 | return $sha; |
|
| 385 | } |
|
| 386 | ||
| 387 | private function resolvePath( string $treeSha, string $path ): array { |
|
| 388 | $parts = explode( '/', trim( $path, '/' ) ); |
|
| 389 | $sha = $treeSha; |
|
| 390 | $mode = '40000'; |
|
| 391 | ||
| 392 | foreach( $parts as $part ) { |
|
| 393 | $entry = [ 'sha' => '', 'mode' => '' ]; |
|
| 394 | ||
| 395 | if( $part !== '' && $sha !== '' ) { |
|
| 396 | $entry = $this->findTreeEntry( $sha, $part ); |
|
| 397 | } |
|
| 398 | ||
| 399 | $sha = $entry['sha']; |
|
| 400 | $mode = $entry['mode']; |
|
| 401 | } |
|
| 402 | ||
| 403 | return [ |
|
| 404 | 'sha' => $sha, |
|
| 405 | 'mode' => $mode, |
|
| 406 | 'isDir' => $mode === '40000' || $mode === '040000' |
|
| 407 | ]; |
|
| 408 | } |
|
| 409 | ||
| 410 | private function findTreeEntry( string $treeSha, string $name ): array { |
|
| 411 | $data = $this->read( $treeSha ); |
|
| 412 | $pos = 0; |
|
| 413 | $len = strlen( $data ); |
|
| 414 | $entry = [ 'sha' => '', 'mode' => '' ]; |
|
| 415 | ||
| 416 | while( $pos < $len ) { |
|
| 417 | $space = strpos( $data, ' ', $pos ); |
|
| 418 | $eos = strpos( $data, "\0", $space ); |
|
| 419 | ||
| 420 | if( $space === false || $eos === false ) { |
|
| 421 | break; |
|
| 422 | } |
|
| 423 | ||
| 424 | if( substr( $data, $space + 1, $eos - $space - 1 ) === $name ) { |
|
| 425 | $entry = [ |
|
| 426 | 'sha' => bin2hex( substr( $data, $eos + 1, 20 ) ), |
|
| 427 | 'mode' => substr( $data, $pos, $space - $pos ) |
|
| 428 | ]; |
|
| 429 | break; |
|
| 430 | } |
|
| 431 | ||
| 432 | $pos = $eos + 21; |
|
| 433 | } |
|
| 434 | ||
| 435 | return $entry; |
|
| 436 | } |
|
| 437 | ||
| 438 | private function parseTagData( |
|
| 439 | string $name, |
|
| 440 | string $sha, |
|
| 441 | string $data |
|
| 442 | ): Tag { |
|
| 443 | $isAnn = strncmp( $data, 'object ', 7 ) === 0; |
|
| 444 | $pattern = $isAnn |
|
| 445 | ? '/^tagger (.*) <(.*)> (\d+) [+\-]\d{4}$/m' |
|
| 446 | : '/^author (.*) <(.*)> (\d+) [+\-]\d{4}$/m'; |
|
| 447 | $id = $this->parseIdentity( $data, $pattern ); |
|
| 448 | $target = $isAnn |
|
| 449 | ? $this->extractPattern( $data, '/^object (.*)$/m', 1, $sha ) |
|
| 450 | : $sha; |
|
| 451 | ||
| 452 | return new Tag( |
|
| 453 | $name, |
|
| 454 | $sha, |
|
| 455 | $target, |
|
| 456 | $id['timestamp'], |
|
| 457 | $this->extractMessage( $data ), |
|
| 458 | $id['name'] |
|
| 459 | ); |
|
| 460 | } |
|
| 461 | ||
| 462 | private function extractPattern( |
|
| 463 | string $data, |
|
| 464 | string $pattern, |
|
| 465 | int $group, |
|
| 466 | string $default = '' |
|
| 467 | ): string { |
|
| 468 | return preg_match( $pattern, $data, $matches ) |
|
| 469 | ? $matches[$group] |
|
| 470 | : $default; |
|
| 471 | } |
|
| 472 | ||
| 473 | private function parseIdentity( string $data, string $pattern ): array { |
|
| 474 | $found = preg_match( $pattern, $data, $matches ); |
|
| 475 | ||
| 476 | return [ |
|
| 477 | 'name' => $found ? trim( $matches[1] ) : 'Unknown', |
|
| 478 | 'email' => $found ? $matches[2] : '', |
|
| 479 | 'timestamp' => $found ? (int)$matches[3] : 0 |
|
| 480 | ]; |
|
| 481 | } |
|
| 482 | ||
| 483 | private function extractMessage( string $data ): string { |
|
| 484 | $pos = strpos( $data, "\n\n" ); |
|
| 485 | ||
| 486 | return $pos !== false ? trim( substr( $data, $pos + 2 ) ) : ''; |
|
| 487 | } |
|
| 488 | ||
| 489 | private function slurp( string $sha, callable $callback ): void { |
|
| 490 | $path = $this->getLoosePath( $sha ); |
|
| 491 | ||
| 492 | if( is_file( $path ) ) { |
|
| 493 | $this->slurpLooseObject( $path, $callback ); |
|
| 494 | } else { |
|
| 495 | $this->slurpPackedObject( $sha, $callback ); |
|
| 496 | } |
|
| 497 | } |
|
| 498 | ||
| 499 | private function slurpLooseObject( string $path, callable $callback ): void { |
|
| 500 | $this->iterateInflated( |
|
| 501 | $path, |
|
| 502 | function( $chunk ) use ( $callback ) { |
|
| 503 | if( $chunk !== '' ) { |
|
| 504 | $callback( $chunk ); |
|
| 505 | } |
|
| 506 | return true; |
|
| 507 | } |
|
| 508 | ); |
|
| 509 | } |
|
| 510 | ||
| 511 | private function slurpPackedObject( string $sha, callable $callback ): void { |
|
| 512 | $streamed = $this->packs->stream( $sha, $callback ); |
|
| 513 | ||
| 514 | if( !$streamed ) { |
|
| 515 | $data = $this->packs->read( $sha ); |
|
| 516 | ||
| 517 | if( $data !== '' ) { |
|
| 518 | $callback( $data ); |
|
| 519 | } |
|
| 520 | } |
|
| 521 | } |
|
| 522 | ||
| 523 | private function iterateInflated( |
|
| 524 | string $path, |
|
| 525 | callable $processor |
|
| 526 | ): void { |
|
| 527 | $handle = fopen( $path, 'rb' ); |
|
| 528 | $infl = $handle ? inflate_init( ZLIB_ENCODING_DEFLATE ) : null; |
|
| 529 | $found = false; |
|
| 530 | $buffer = ''; |
|
| 531 | ||
| 532 | if( $handle && $infl ) { |
|
| 533 | while( !feof( $handle ) ) { |
|
| 534 | $chunk = fread( $handle, 16384 ); |
|
| 535 | $inflated = inflate_add( $infl, $chunk ); |
|
| 536 | ||
| 537 | if( $inflated === false ) { |
|
| 538 | break; |
|
| 539 | } |
|
| 540 | ||
| 541 | if( !$found ) { |
|
| 542 | $buffer .= $inflated; |
|
| 543 | $eos = strpos( $buffer, "\0" ); |
|
| 544 | ||
| 545 | if( $eos !== false ) { |
|
| 546 | $found = true; |
|
| 547 | $body = substr( $buffer, $eos + 1 ); |
|
| 548 | $head = substr( $buffer, 0, $eos ); |
|
| 549 | ||
| 550 | if( $processor( $body, $head ) === false ) { |
|
| 551 | break; |
|
| 552 | } |
|
| 553 | } |
|
| 554 | } elseif( $processor( $inflated, null ) === false ) { |
|
| 555 | break; |
|
| 556 | } |
|
| 557 | } |
|
| 558 | ||
| 559 | fclose( $handle ); |
|
| 560 | } |
|
| 561 | } |
|
| 562 | ||
| 563 | private function peekLooseObject( string $sha, int $length ): string { |
|
| 564 | $path = $this->getLoosePath( $sha ); |
|
| 565 | $buf = ''; |
|
| 566 | ||
| 567 | if( is_file( $path ) ) { |
|
| 568 | $this->iterateInflated( |
|
| 569 | $path, |
|
| 570 | function( $chunk ) use ( $length, &$buf ) { |
|
| 571 | $buf .= $chunk; |
|
| 572 | return strlen( $buf ) < $length; |
|
| 573 | } |
|
| 574 | ); |
|
| 575 | } |
|
| 576 | ||
| 577 | return substr( $buf, 0, $length ); |
|
| 578 | } |
|
| 579 | ||
| 580 | private function parseCommit( string $sha ): object { |
|
| 581 | $data = $this->read( $sha ); |
|
| 582 | $result = (object)[ 'sha' => '' ]; |
|
| 583 | ||
| 584 | if( $data !== '' ) { |
|
| 585 | $id = $this->parseIdentity( |
|
| 586 | $data, |
|
| 587 | '/^author (.*) <(.*)> (\d+)/m' |
|
| 588 | ); |
|
| 589 | ||
| 590 | $result = (object)[ |
|
| 591 | 'sha' => $sha, |
|
| 592 | 'message' => $this->extractMessage( $data ), |
|
| 593 | 'author' => $id['name'], |
|
| 594 | 'email' => $id['email'], |
|
| 595 | 'date' => $id['timestamp'], |
|
| 596 | 'parentSha' => $this->extractPattern( $data, '/^parent (.*)$/m', 1 ) |
|
| 597 | ]; |
|
| 598 | } |
|
| 599 | ||
| 600 | return $result; |
|
| 601 | } |
|
| 602 | ||
| 603 | private function walkTree( string $sha, callable $callback ): void { |
|
| 604 | $data = $this->read( $sha ); |
|
| 605 | $tree = $data; |
|
| 606 | ||
| 607 | if( $data !== '' && preg_match( '/^tree (.*)$/m', $data, $m ) ) { |
|
| 608 | $tree = $this->read( $m[1] ); |
|
| 609 | } |
|
| 610 | ||
| 611 | if( $tree !== '' && $this->isTreeData( $tree ) ) { |
|
| 612 | $this->processTree( $tree, $callback ); |
|
| 613 | } |
|
| 614 | } |
|
| 615 | ||
| 616 | private function processTree( string $data, callable $callback ): void { |
|
| 617 | $pos = 0; |
|
| 618 | $len = strlen( $data ); |
|
| 619 | ||
| 620 | while( $pos < $len ) { |
|
| 621 | $space = strpos( $data, ' ', $pos ); |
|
| 622 | $eos = strpos( $data, "\0", $space ); |
|
| 623 | $entry = null; |
|
| 624 | ||
| 625 | if( $space !== false && $eos !== false && $eos + 21 <= $len ) { |
|
| 626 | $mode = substr( $data, $pos, $space - $pos ); |
|
| 627 | $sha = bin2hex( substr( $data, $eos + 1, 20 ) ); |
|
| 628 | $dir = $mode === '40000' || $mode === '040000'; |
|
| 629 | $isSub = $mode === '160000'; |
|
| 630 | ||
| 631 | $entry = [ |
|
| 632 | 'file' => new File( |
|
| 633 | substr( $data, $space + 1, $eos - $space - 1 ), |
|
| 634 | $sha, |
|
| 635 | $mode, |
|
| 636 | 0, |
|
| 637 | $dir || $isSub ? 0 : $this->getObjectSize( $sha ), |
|
| 638 | $dir || $isSub ? '' : $this->peek( $sha ) |
|
| 639 | ), |
|
| 640 | 'nextPosition' => $eos + 21 |
|
| 641 | ]; |
|
| 642 | } |
|
| 643 | ||
| 644 | if( $entry === null ) { |
|
| 645 | break; |
|
| 646 | } |
|
| 647 | ||
| 648 | $callback( $entry['file'] ); |
|
| 649 | $pos = $entry['nextPosition']; |
|
| 650 | } |
|
| 651 | } |
|
| 652 | ||
| 653 | private function isTreeData( string $data ): bool { |
|
| 654 | $len = strlen( $data ); |
|
| 655 | $patt = '/^(40000|100644|100755|120000|160000) /'; |
|
| 656 | $match = $len >= 25 && preg_match( $patt, $data ); |
|
| 657 | $eos = $match ? strpos( $data, "\0" ) : false; |
|
| 658 | ||
| 659 | return $match && $eos !== false && $eos + 21 <= $len; |
|
| 660 | } |
|
| 661 | ||
| 662 | private function getLoosePath( string $sha ): string { |
|
| 663 | return "{$this->objPath}/" . substr( $sha, 0, 2 ) . "/" . |
|
| 664 | substr( $sha, 2 ); |
|
| 665 | } |
|
| 666 | ||
| 667 | private function getLooseObjectSize( string $sha ): int { |
|
| 668 | $path = $this->getLoosePath( $sha ); |
|
| 669 | $size = 0; |
|
| 670 | ||
| 671 | if( is_file( $path ) ) { |
|
| 672 | $this->iterateInflated( |
|
| 673 | $path, |
|
| 674 | function( $c, $head ) use ( &$size ) { |
|
| 675 | if( $head !== null ) { |
|
| 676 | $parts = explode( ' ', $head ); |
|
| 677 | $size = isset( $parts[1] ) ? (int)$parts[1] : 0; |
|
| 678 | } |
|
| 679 | return false; |
|
| 680 | } |
|
| 681 | ); |
|
| 682 | } |
|
| 683 | ||
| 684 | return $size; |
|
| 685 | } |
|
| 686 | ||
| 687 | public function collectObjects( array $wants, array $haves = [] ): array { |
|
| 688 | $objs = $this->traverseObjects( $wants ); |
|
| 689 | $result = []; |
|
| 690 | ||
| 691 | if( !empty( $haves ) ) { |
|
| 692 | $haveObjs = $this->traverseObjects( $haves ); |
|
| 693 | ||
| 694 | foreach( $haveObjs as $sha => $type ) { |
|
| 695 | if( isset( $objs[$sha] ) ) { |
|
| 696 | unset( $objs[$sha] ); |
|
| 697 | } |
|
| 698 | } |
|
| 699 | } |
|
| 700 | ||
| 701 | $result = $objs; |
|
| 702 | ||
| 703 | return $result; |
|
| 704 | } |
|
| 705 | ||
| 706 | private function traverseObjects( array $roots ): array { |
|
| 707 | $objs = []; |
|
| 708 | $queue = []; |
|
| 709 | ||
| 710 | foreach( $roots as $sha ) { |
|
| 711 | $queue[] = [ 'sha' => $sha, 'type' => 0 ]; |
|
| 712 | } |
|
| 713 | ||
| 714 | while( !empty( $queue ) ) { |
|
| 715 | $item = array_pop( $queue ); |
|
| 716 | $sha = $item['sha']; |
|
| 717 | $type = $item['type']; |
|
| 718 | ||
| 719 | if( isset( $objs[$sha] ) ) { |
|
| 720 | continue; |
|
| 721 | } |
|
| 722 | ||
| 723 | $data = ''; |
|
| 724 | ||
| 725 | if( $type !== 3 ) { |
|
| 726 | $data = $this->read( $sha ); |
|
| 727 | ||
| 728 | if( $type === 0 ) { |
|
| 729 | $type = $this->getObjectType( $data ); |
|
| 730 | } |
|
| 731 | } |
|
| 732 | ||
| 733 | $objs[$sha] = $type; |
|
| 734 | ||
| 735 | if( $type === 1 ) { |
|
| 736 | $hasTree = preg_match( '/^tree ([0-9a-f]{40})/m', $data, $m ); |
|
| 737 | ||
| 738 | if( $hasTree ) { |
|
| 739 | $queue[] = [ 'sha' => $m[1], 'type' => 2 ]; |
|
| 740 | } |
|
| 741 | ||
| 742 | $hasParents = preg_match_all( |
|
| 743 | '/^parent ([0-9a-f]{40})/m', |
|
| 744 | $data, |
|
| 745 | $m |
|
| 746 | ); |
|
| 747 | ||
| 748 | if( $hasParents ) { |
|
| 749 | foreach( $m[1] as $parentSha ) { |
|
| 750 | $queue[] = [ 'sha' => $parentSha, 'type' => 1 ]; |
|
| 751 | } |
|
| 752 | } |
|
| 753 | } elseif( $type === 2 ) { |
|
| 754 | $pos = 0; |
|
| 755 | $len = strlen( $data ); |
|
| 756 | ||
| 757 | while( $pos < $len ) { |
|
| 758 | $space = strpos( $data, ' ', $pos ); |
|
| 759 | $eos = strpos( $data, "\0", $space ); |
|
| 760 | ||
| 761 | if( $space === false || $eos === false ) { |
|
| 762 | break; |
|
| 763 | } |
|
| 764 | ||
| 765 | $mode = substr( $data, $pos, $space - $pos ); |
|
| 766 | $hash = bin2hex( substr( $data, $eos + 1, 20 ) ); |
|
| 767 | ||
| 768 | if( $mode !== '160000' ) { |
|
| 769 | $isDir = $mode === '40000' || $mode === '040000'; |
|
| 770 | $queue[] = [ 'sha' => $hash, 'type' => $isDir ? 2 : 3 ]; |
|
| 771 | } |
|
| 772 | ||
| 773 | $pos = $eos + 21; |
|
| 774 | } |
|
| 775 | } elseif( $type === 4 ) { |
|
| 776 | $isTagTgt = preg_match( '/^object ([0-9a-f]{40})/m', $data, $m ); |
|
| 777 | ||
| 778 | if( $isTagTgt ) { |
|
| 779 | $nextType = 1; |
|
| 780 | ||
| 781 | if( preg_match( '/^type (commit|tree|blob|tag)/m', $data, $t ) ) { |
|
| 782 | $map = [ |
|
| 783 | 'commit' => 1, |
|
| 784 | 'tree' => 2, |
|
| 785 | 'blob' => 3, |
|
| 786 | 'tag' => 4 |
|
| 787 | ]; |
|
| 788 | $nextType = $map[$t[1]] ?? 1; |
|
| 789 | } |
|
| 790 | ||
| 791 | $queue[] = [ 'sha' => $m[1], 'type' => $nextType ]; |
|
| 792 | } |
|
| 793 | } |
|
| 794 | } |
|
| 795 | ||
| 796 | return $objs; |
|
| 670 | 797 | } |
| 671 | 798 |
| 3 | 3 | private const MAX_READ = 1040576; |
| 4 | 4 | private const MAX_RAM = 1048576; |
| 5 | private const MAX_BASE_RAM = 524288; |
|
| 6 | private const MAX_DEPTH = 50; |
|
| 7 | ||
| 8 | private string $objectsPath; |
|
| 9 | private array $packFiles; |
|
| 10 | private string $lastPack = ''; |
|
| 11 | private array $fileHandles; |
|
| 12 | private array $fanoutCache; |
|
| 13 | private array $shaBucketCache; |
|
| 14 | private array $offsetBucketCache; |
|
| 15 | ||
| 16 | public function __construct( string $objectsPath ) { |
|
| 17 | $this->objectsPath = $objectsPath; |
|
| 18 | $this->packFiles = glob( "{$this->objectsPath}/pack/*.idx" ) ?: []; |
|
| 19 | $this->fileHandles = []; |
|
| 20 | $this->fanoutCache = []; |
|
| 21 | $this->shaBucketCache = []; |
|
| 22 | $this->offsetBucketCache = []; |
|
| 23 | } |
|
| 24 | ||
| 25 | public function __destruct() { |
|
| 26 | foreach( $this->fileHandles as $handle ) { |
|
| 27 | if( is_resource( $handle ) ) { |
|
| 28 | fclose( $handle ); |
|
| 29 | } |
|
| 30 | } |
|
| 31 | } |
|
| 32 | ||
| 33 | public function peek( string $sha, int $len = 12 ): string { |
|
| 34 | $info = $this->findPackInfo( $sha ); |
|
| 35 | $result = ''; |
|
| 36 | ||
| 37 | if( $info['offset'] !== 0 ) { |
|
| 38 | $handle = $this->getHandle( $info['file'] ); |
|
| 39 | ||
| 40 | if( $handle ) { |
|
| 41 | $result = $this->readPackEntry( |
|
| 42 | $handle, |
|
| 43 | $info['offset'], |
|
| 44 | $len, |
|
| 45 | $len |
|
| 46 | ); |
|
| 47 | } |
|
| 48 | } |
|
| 49 | ||
| 50 | return $result; |
|
| 51 | } |
|
| 52 | ||
| 53 | public function read( string $sha ): string { |
|
| 54 | $info = $this->findPackInfo( $sha ); |
|
| 55 | $result = ''; |
|
| 56 | ||
| 57 | if( $info['offset'] !== 0 ) { |
|
| 58 | $size = $this->extractPackedSize( $info['file'], $info['offset'] ); |
|
| 59 | ||
| 60 | if( $size <= self::MAX_RAM ) { |
|
| 61 | $handle = $this->getHandle( $info['file'] ); |
|
| 62 | ||
| 63 | if( $handle ) { |
|
| 64 | $result = $this->readPackEntry( |
|
| 65 | $handle, |
|
| 66 | $info['offset'], |
|
| 67 | $size |
|
| 68 | ); |
|
| 69 | } |
|
| 70 | } |
|
| 71 | } |
|
| 72 | ||
| 73 | return $result; |
|
| 74 | } |
|
| 75 | ||
| 76 | public function stream( string $sha, callable $callback ): bool { |
|
| 77 | return $this->streamInternal( $sha, $callback, 0 ); |
|
| 78 | } |
|
| 79 | ||
| 80 | private function streamInternal( |
|
| 81 | string $sha, |
|
| 82 | callable $callback, |
|
| 83 | int $depth |
|
| 84 | ): bool { |
|
| 85 | $info = $this->findPackInfo( $sha ); |
|
| 86 | $result = false; |
|
| 87 | ||
| 88 | if( $info['offset'] !== 0 ) { |
|
| 89 | $size = $this->extractPackedSize( $info['file'], $info['offset'] ); |
|
| 90 | $handle = $this->getHandle( $info['file'] ); |
|
| 91 | ||
| 92 | if( $handle ) { |
|
| 93 | $result = $this->streamPackEntry( |
|
| 94 | $handle, |
|
| 95 | $info['offset'], |
|
| 96 | $size, |
|
| 97 | $callback, |
|
| 98 | $depth |
|
| 99 | ); |
|
| 100 | } |
|
| 101 | } |
|
| 102 | ||
| 103 | return $result; |
|
| 104 | } |
|
| 105 | ||
| 106 | public function getSize( string $sha ): int { |
|
| 107 | $info = $this->findPackInfo( $sha ); |
|
| 108 | $result = 0; |
|
| 109 | ||
| 110 | if( $info['offset'] !== 0 ) { |
|
| 111 | $result = $this->extractPackedSize( $info['file'], $info['offset'] ); |
|
| 112 | } |
|
| 113 | ||
| 114 | return $result; |
|
| 115 | } |
|
| 116 | ||
| 117 | private function findPackInfo( string $sha ): array { |
|
| 118 | $result = [ 'offset' => 0, 'file' => '' ]; |
|
| 119 | ||
| 120 | if( strlen( $sha ) === 40 && ctype_xdigit( $sha ) ) { |
|
| 121 | $binarySha = hex2bin( $sha ); |
|
| 122 | if( $this->lastPack !== '' ) { |
|
| 123 | $offset = $this->findInIdx( $this->lastPack, $binarySha ); |
|
| 124 | ||
| 125 | if( $offset !== 0 ) { |
|
| 126 | $result = [ |
|
| 127 | 'file' => str_replace( '.idx', '.pack', $this->lastPack ), |
|
| 128 | 'offset' => $offset |
|
| 129 | ]; |
|
| 130 | } |
|
| 131 | } |
|
| 132 | ||
| 133 | if( $result['offset'] === 0 ) { |
|
| 134 | foreach( $this->packFiles as $indexFile ) { |
|
| 135 | if( $indexFile !== $this->lastPack ) { |
|
| 136 | $offset = $this->findInIdx( $indexFile, $binarySha ); |
|
| 137 | ||
| 138 | if( $offset !== 0 ) { |
|
| 139 | $this->lastPack = $indexFile; |
|
| 140 | $result = [ |
|
| 141 | 'file' => str_replace( '.idx', '.pack', $indexFile ), |
|
| 142 | 'offset' => $offset |
|
| 143 | ]; |
|
| 144 | break; |
|
| 145 | } |
|
| 146 | } |
|
| 147 | } |
|
| 148 | } |
|
| 149 | } |
|
| 150 | ||
| 151 | return $result; |
|
| 152 | } |
|
| 153 | ||
| 154 | private function findInIdx( string $indexFile, string $binarySha ): int { |
|
| 155 | $handle = $this->getHandle( $indexFile ); |
|
| 156 | $result = 0; |
|
| 157 | ||
| 158 | if( $handle ) { |
|
| 159 | if( !isset( $this->fanoutCache[$indexFile] ) ) { |
|
| 160 | fseek( $handle, 0 ); |
|
| 161 | $head = fread( $handle, 8 ); |
|
| 162 | ||
| 163 | if( $head === "\377tOc\0\0\0\2" ) { |
|
| 164 | $this->fanoutCache[$indexFile] = array_values( |
|
| 165 | unpack( 'N*', fread( $handle, 1024 ) ) |
|
| 166 | ); |
|
| 167 | } |
|
| 168 | } |
|
| 169 | ||
| 170 | if( isset( $this->fanoutCache[$indexFile] ) ) { |
|
| 171 | $fanout = $this->fanoutCache[$indexFile]; |
|
| 172 | $byte = ord( $binarySha[0] ); |
|
| 173 | $start = $byte === 0 ? 0 : $fanout[$byte - 1]; |
|
| 174 | $end = $fanout[$byte]; |
|
| 175 | ||
| 176 | if( $end > $start ) { |
|
| 177 | $result = $this->binarySearchIdx( |
|
| 178 | $indexFile, |
|
| 179 | $handle, |
|
| 180 | $start, |
|
| 181 | $end, |
|
| 182 | $binarySha, |
|
| 183 | $fanout[255] |
|
| 184 | ); |
|
| 185 | } |
|
| 186 | } |
|
| 187 | } |
|
| 188 | ||
| 189 | return $result; |
|
| 190 | } |
|
| 191 | ||
| 192 | private function binarySearchIdx( |
|
| 193 | string $indexFile, |
|
| 194 | $handle, |
|
| 195 | int $start, |
|
| 196 | int $end, |
|
| 197 | string $binarySha, |
|
| 198 | int $total |
|
| 199 | ): int { |
|
| 200 | $key = "$indexFile:$start"; |
|
| 201 | $count = $end - $start; |
|
| 202 | $result = 0; |
|
| 203 | ||
| 204 | if( !isset( $this->shaBucketCache[$key] ) ) { |
|
| 205 | fseek( $handle, 1032 + ($start * 20) ); |
|
| 206 | $this->shaBucketCache[$key] = fread( $handle, $count * 20 ); |
|
| 207 | ||
| 208 | fseek( $handle, 1032 + ($total * 24) + ($start * 4) ); |
|
| 209 | $this->offsetBucketCache[$key] = fread( $handle, $count * 4 ); |
|
| 210 | } |
|
| 211 | ||
| 212 | $shaBlock = $this->shaBucketCache[$key]; |
|
| 213 | $low = 0; |
|
| 214 | $high = $count - 1; |
|
| 215 | $found = -1; |
|
| 216 | ||
| 217 | while( $low <= $high ) { |
|
| 218 | $mid = ($low + $high) >> 1; |
|
| 219 | $cmp = substr( $shaBlock, $mid * 20, 20 ); |
|
| 220 | ||
| 221 | if( $cmp < $binarySha ) { |
|
| 222 | $low = $mid + 1; |
|
| 223 | } elseif( $cmp > $binarySha ) { |
|
| 224 | $high = $mid - 1; |
|
| 225 | } else { |
|
| 226 | $found = $mid; |
|
| 227 | break; |
|
| 228 | } |
|
| 229 | } |
|
| 230 | ||
| 231 | if( $found !== -1 ) { |
|
| 232 | $packed = substr( $this->offsetBucketCache[$key], $found * 4, 4 ); |
|
| 233 | $offset = unpack( 'N', $packed )[1]; |
|
| 234 | ||
| 235 | if( $offset & 0x80000000 ) { |
|
| 236 | $pos64 = 1032 + ($total * 28) + (($offset & 0x7FFFFFFF) * 8); |
|
| 237 | fseek( $handle, $pos64 ); |
|
| 238 | $offset = unpack( 'J', fread( $handle, 8 ) )[1]; |
|
| 239 | } |
|
| 240 | $result = (int)$offset; |
|
| 241 | } |
|
| 242 | ||
| 243 | return $result; |
|
| 244 | } |
|
| 245 | ||
| 246 | private function readPackEntry( |
|
| 247 | $handle, |
|
| 248 | int $offset, |
|
| 249 | int $size, |
|
| 250 | int $cap = 0 |
|
| 251 | ): string { |
|
| 252 | fseek( $handle, $offset ); |
|
| 253 | $header = $this->readVarInt( $handle ); |
|
| 254 | $type = ($header['byte'] >> 4) & 7; |
|
| 255 | ||
| 256 | return ($type === 6) |
|
| 257 | ? $this->handleOfsDelta( $handle, $offset, $size, $cap ) |
|
| 258 | : (($type === 7) |
|
| 259 | ? $this->handleRefDelta( $handle, $size, $cap ) |
|
| 260 | : $this->decompressToString( $handle, $cap )); |
|
| 261 | } |
|
| 262 | ||
| 263 | private function streamPackEntry( |
|
| 264 | $handle, |
|
| 265 | int $offset, |
|
| 266 | int $size, |
|
| 267 | callable $callback, |
|
| 268 | int $depth = 0 |
|
| 269 | ): bool { |
|
| 270 | fseek( $handle, $offset ); |
|
| 271 | $header = $this->readVarInt( $handle ); |
|
| 272 | $type = ($header['byte'] >> 4) & 7; |
|
| 273 | ||
| 274 | return ($type === 6 || $type === 7) |
|
| 275 | ? $this->streamDeltaObject( $handle, $offset, $type, $callback, $depth ) |
|
| 276 | : $this->streamDecompression( $handle, $callback ); |
|
| 277 | } |
|
| 278 | ||
| 279 | private function streamDeltaObject( |
|
| 280 | $handle, |
|
| 281 | int $offset, |
|
| 282 | int $type, |
|
| 283 | callable $callback, |
|
| 284 | int $depth = 0 |
|
| 285 | ): bool { |
|
| 286 | if( $depth >= self::MAX_DEPTH ) { |
|
| 287 | return false; |
|
| 288 | } |
|
| 289 | ||
| 290 | fseek( $handle, $offset ); |
|
| 291 | $this->readVarInt( $handle ); |
|
| 292 | $result = false; |
|
| 293 | ||
| 294 | if( $type === 6 ) { |
|
| 295 | $neg = $this->readOffsetDelta( $handle ); |
|
| 296 | $deltaPos = ftell( $handle ); |
|
| 297 | $base = ''; |
|
| 298 | ||
| 299 | $baseSize = $this->extractPackedSize( $handle, $offset - $neg ); |
|
| 300 | ||
| 301 | if( $baseSize > self::MAX_BASE_RAM ) { |
|
| 302 | return false; |
|
| 303 | } |
|
| 304 | ||
| 305 | $this->streamPackEntry( |
|
| 306 | $handle, |
|
| 307 | $offset - $neg, |
|
| 308 | 0, |
|
| 309 | function( $c ) use ( &$base ) { $base .= $c; }, |
|
| 310 | $depth + 1 |
|
| 311 | ); |
|
| 312 | ||
| 313 | fseek( $handle, $deltaPos ); |
|
| 314 | $result = $this->applyDeltaStream( $handle, $base, $callback ); |
|
| 315 | } else { |
|
| 316 | $baseSha = bin2hex( fread( $handle, 20 ) ); |
|
| 317 | $baseSize = $this->getSize( $baseSha ); |
|
| 318 | ||
| 319 | if( $baseSize > self::MAX_BASE_RAM ) { |
|
| 320 | return false; |
|
| 321 | } |
|
| 322 | ||
| 323 | $base = ''; |
|
| 324 | ||
| 325 | if( $this->streamInternal( $baseSha, function( $c ) use ( &$base ) { |
|
| 326 | $base .= $c; |
|
| 327 | }, $depth + 1 ) ) { |
|
| 328 | $result = $this->applyDeltaStream( $handle, $base, $callback ); |
|
| 329 | } |
|
| 330 | } |
|
| 331 | ||
| 332 | return $result; |
|
| 333 | } |
|
| 334 | ||
| 335 | private function applyDeltaStream( |
|
| 336 | $handle, |
|
| 337 | string $base, |
|
| 338 | callable $callback |
|
| 339 | ): bool { |
|
| 340 | $infl = inflate_init( ZLIB_ENCODING_DEFLATE ); |
|
| 341 | $ok = false; |
|
| 342 | ||
| 343 | if( $infl ) { |
|
| 344 | $state = 0; |
|
| 345 | $buffer = ''; |
|
| 346 | $ok = true; |
|
| 347 | ||
| 348 | while( !feof( $handle ) ) { |
|
| 349 | $chunk = fread( $handle, 8192 ); |
|
| 350 | ||
| 351 | if( $chunk === '' ) { |
|
| 352 | break; |
|
| 353 | } |
|
| 354 | ||
| 355 | $data = @inflate_add( $infl, $chunk ); |
|
| 356 | ||
| 357 | if( $data === false ) { |
|
| 358 | $ok = false; |
|
| 359 | break; |
|
| 360 | } |
|
| 361 | ||
| 362 | $buffer .= $data; |
|
| 363 | ||
| 364 | while( true ) { |
|
| 365 | $len = strlen( $buffer ); |
|
| 366 | ||
| 367 | if( $len === 0 ) { |
|
| 368 | break; |
|
| 369 | } |
|
| 370 | ||
| 371 | if( $state < 2 ) { |
|
| 372 | $pos = 0; |
|
| 373 | while( $pos < $len && (ord( $buffer[$pos] ) & 128) ) { $pos++; } |
|
| 374 | ||
| 375 | if( $pos === $len && (ord( $buffer[$pos - 1] ) & 128) ) { |
|
| 376 | break; |
|
| 377 | } |
|
| 378 | ||
| 379 | $buffer = substr( $buffer, $pos + 1 ); |
|
| 380 | $state++; |
|
| 381 | continue; |
|
| 382 | } |
|
| 383 | ||
| 384 | $op = ord( $buffer[0] ); |
|
| 385 | ||
| 386 | if( $op & 128 ) { |
|
| 387 | $need = $this->getCopyInstructionSize( $op ); |
|
| 388 | ||
| 389 | if( $len < 1 + $need ) { |
|
| 390 | break; |
|
| 391 | } |
|
| 392 | ||
| 393 | $info = $this->parseCopyInstruction( $op, $buffer, 1 ); |
|
| 394 | ||
| 395 | $callback( substr( $base, $info['off'], $info['len'] ) ); |
|
| 396 | $buffer = substr( $buffer, 1 + $need ); |
|
| 397 | } else { |
|
| 398 | $ln = $op & 127; |
|
| 399 | ||
| 400 | if( $len < 1 + $ln ) { |
|
| 401 | break; |
|
| 402 | } |
|
| 403 | ||
| 404 | $callback( substr( $buffer, 1, $ln ) ); |
|
| 405 | $buffer = substr( $buffer, 1 + $ln ); |
|
| 406 | } |
|
| 407 | } |
|
| 408 | ||
| 409 | if( inflate_get_status( $infl ) === ZLIB_STREAM_END ) { |
|
| 410 | break; |
|
| 411 | } |
|
| 412 | } |
|
| 413 | } |
|
| 414 | ||
| 415 | return $ok; |
|
| 5 | private const MAX_BASE_RAM = 2097152; |
|
| 6 | private const MAX_DEPTH = 200; |
|
| 7 | ||
| 8 | private string $objectsPath; |
|
| 9 | private array $packFiles; |
|
| 10 | private string $lastPack = ''; |
|
| 11 | private array $fileHandles; |
|
| 12 | private array $fanoutCache; |
|
| 13 | private array $shaBucketCache; |
|
| 14 | private array $offsetBucketCache; |
|
| 15 | ||
| 16 | public function __construct( string $objectsPath ) { |
|
| 17 | $this->objectsPath = $objectsPath; |
|
| 18 | $this->packFiles = glob( "{$this->objectsPath}/pack/*.idx" ) ?: []; |
|
| 19 | $this->fileHandles = []; |
|
| 20 | $this->fanoutCache = []; |
|
| 21 | $this->shaBucketCache = []; |
|
| 22 | $this->offsetBucketCache = []; |
|
| 23 | } |
|
| 24 | ||
| 25 | public function __destruct() { |
|
| 26 | foreach( $this->fileHandles as $handle ) { |
|
| 27 | if( is_resource( $handle ) ) { |
|
| 28 | fclose( $handle ); |
|
| 29 | } |
|
| 30 | } |
|
| 31 | } |
|
| 32 | ||
| 33 | public function peek( string $sha, int $len = 12 ): string { |
|
| 34 | $info = $this->findPackInfo( $sha ); |
|
| 35 | $result = ''; |
|
| 36 | ||
| 37 | if( $info['offset'] !== 0 ) { |
|
| 38 | $handle = $this->getHandle( $info['file'] ); |
|
| 39 | ||
| 40 | if( $handle ) { |
|
| 41 | $result = $this->readPackEntry( |
|
| 42 | $handle, |
|
| 43 | $info['offset'], |
|
| 44 | $len, |
|
| 45 | $len |
|
| 46 | ); |
|
| 47 | } |
|
| 48 | } |
|
| 49 | ||
| 50 | return $result; |
|
| 51 | } |
|
| 52 | ||
| 53 | public function read( string $sha ): string { |
|
| 54 | $info = $this->findPackInfo( $sha ); |
|
| 55 | $result = ''; |
|
| 56 | ||
| 57 | if( $info['offset'] !== 0 ) { |
|
| 58 | $size = $this->extractPackedSize( $info['file'], $info['offset'] ); |
|
| 59 | ||
| 60 | if( $size <= self::MAX_RAM ) { |
|
| 61 | $handle = $this->getHandle( $info['file'] ); |
|
| 62 | ||
| 63 | if( $handle ) { |
|
| 64 | $result = $this->readPackEntry( |
|
| 65 | $handle, |
|
| 66 | $info['offset'], |
|
| 67 | $size |
|
| 68 | ); |
|
| 69 | } |
|
| 70 | } |
|
| 71 | } |
|
| 72 | ||
| 73 | return $result; |
|
| 74 | } |
|
| 75 | ||
| 76 | public function stream( string $sha, callable $callback ): bool { |
|
| 77 | return $this->streamInternal( $sha, $callback, 0 ); |
|
| 78 | } |
|
| 79 | ||
| 80 | public function streamGenerator( string $sha ): Generator { |
|
| 81 | $info = $this->findPackInfo( $sha ); |
|
| 82 | ||
| 83 | if( $info['offset'] !== 0 ) { |
|
| 84 | $handle = $this->getHandle( $info['file'] ); |
|
| 85 | ||
| 86 | if( $handle ) { |
|
| 87 | yield from $this->streamPackEntryGenerator( |
|
| 88 | $handle, |
|
| 89 | $info['offset'], |
|
| 90 | 0 |
|
| 91 | ); |
|
| 92 | } |
|
| 93 | } |
|
| 94 | } |
|
| 95 | ||
| 96 | private function streamInternal( |
|
| 97 | string $sha, |
|
| 98 | callable $callback, |
|
| 99 | int $depth |
|
| 100 | ): bool { |
|
| 101 | $info = $this->findPackInfo( $sha ); |
|
| 102 | $result = false; |
|
| 103 | ||
| 104 | if( $info['offset'] !== 0 ) { |
|
| 105 | $size = $this->extractPackedSize( $info['file'], $info['offset'] ); |
|
| 106 | $handle = $this->getHandle( $info['file'] ); |
|
| 107 | ||
| 108 | if( $handle ) { |
|
| 109 | $result = $this->streamPackEntry( |
|
| 110 | $handle, |
|
| 111 | $info['offset'], |
|
| 112 | $size, |
|
| 113 | $callback, |
|
| 114 | $depth |
|
| 115 | ); |
|
| 116 | } |
|
| 117 | } |
|
| 118 | ||
| 119 | return $result; |
|
| 120 | } |
|
| 121 | ||
| 122 | public function getSize( string $sha ): int { |
|
| 123 | $info = $this->findPackInfo( $sha ); |
|
| 124 | $result = 0; |
|
| 125 | ||
| 126 | if( $info['offset'] !== 0 ) { |
|
| 127 | $result = $this->extractPackedSize( $info['file'], $info['offset'] ); |
|
| 128 | } |
|
| 129 | ||
| 130 | return $result; |
|
| 131 | } |
|
| 132 | ||
| 133 | private function findPackInfo( string $sha ): array { |
|
| 134 | $result = [ 'offset' => 0, 'file' => '' ]; |
|
| 135 | ||
| 136 | if( strlen( $sha ) === 40 && ctype_xdigit( $sha ) ) { |
|
| 137 | $binarySha = hex2bin( $sha ); |
|
| 138 | ||
| 139 | if( $this->lastPack !== '' ) { |
|
| 140 | $offset = $this->findInIdx( $this->lastPack, $binarySha ); |
|
| 141 | ||
| 142 | if( $offset !== 0 ) { |
|
| 143 | $result = [ |
|
| 144 | 'file' => str_replace( '.idx', '.pack', $this->lastPack ), |
|
| 145 | 'offset' => $offset |
|
| 146 | ]; |
|
| 147 | } |
|
| 148 | } |
|
| 149 | ||
| 150 | if( $result['offset'] === 0 ) { |
|
| 151 | foreach( $this->packFiles as $indexFile ) { |
|
| 152 | if( $indexFile !== $this->lastPack ) { |
|
| 153 | $offset = $this->findInIdx( $indexFile, $binarySha ); |
|
| 154 | ||
| 155 | if( $offset !== 0 ) { |
|
| 156 | $this->lastPack = $indexFile; |
|
| 157 | $result = [ |
|
| 158 | 'file' => str_replace( '.idx', '.pack', $indexFile ), |
|
| 159 | 'offset' => $offset |
|
| 160 | ]; |
|
| 161 | break; |
|
| 162 | } |
|
| 163 | } |
|
| 164 | } |
|
| 165 | } |
|
| 166 | } |
|
| 167 | ||
| 168 | return $result; |
|
| 169 | } |
|
| 170 | ||
| 171 | private function findInIdx( string $indexFile, string $binarySha ): int { |
|
| 172 | $handle = $this->getHandle( $indexFile ); |
|
| 173 | $result = 0; |
|
| 174 | ||
| 175 | if( $handle ) { |
|
| 176 | if( !isset( $this->fanoutCache[$indexFile] ) ) { |
|
| 177 | fseek( $handle, 0 ); |
|
| 178 | $head = fread( $handle, 8 ); |
|
| 179 | ||
| 180 | if( $head === "\377tOc\0\0\0\2" ) { |
|
| 181 | $this->fanoutCache[$indexFile] = array_values( |
|
| 182 | unpack( 'N*', fread( $handle, 1024 ) ) |
|
| 183 | ); |
|
| 184 | } |
|
| 185 | } |
|
| 186 | ||
| 187 | if( isset( $this->fanoutCache[$indexFile] ) ) { |
|
| 188 | $fanout = $this->fanoutCache[$indexFile]; |
|
| 189 | $byte = ord( $binarySha[0] ); |
|
| 190 | $start = $byte === 0 ? 0 : $fanout[$byte - 1]; |
|
| 191 | $end = $fanout[$byte]; |
|
| 192 | ||
| 193 | if( $end > $start ) { |
|
| 194 | $result = $this->binarySearchIdx( |
|
| 195 | $indexFile, |
|
| 196 | $handle, |
|
| 197 | $start, |
|
| 198 | $end, |
|
| 199 | $binarySha, |
|
| 200 | $fanout[255] |
|
| 201 | ); |
|
| 202 | } |
|
| 203 | } |
|
| 204 | } |
|
| 205 | ||
| 206 | return $result; |
|
| 207 | } |
|
| 208 | ||
| 209 | private function binarySearchIdx( |
|
| 210 | string $indexFile, |
|
| 211 | $handle, |
|
| 212 | int $start, |
|
| 213 | int $end, |
|
| 214 | string $binarySha, |
|
| 215 | int $total |
|
| 216 | ): int { |
|
| 217 | $key = "$indexFile:$start"; |
|
| 218 | $count = $end - $start; |
|
| 219 | $result = 0; |
|
| 220 | ||
| 221 | if( !isset( $this->shaBucketCache[$key] ) ) { |
|
| 222 | fseek( $handle, 1032 + ($start * 20) ); |
|
| 223 | $this->shaBucketCache[$key] = fread( $handle, $count * 20 ); |
|
| 224 | ||
| 225 | fseek( $handle, 1032 + ($total * 24) + ($start * 4) ); |
|
| 226 | $this->offsetBucketCache[$key] = fread( $handle, $count * 4 ); |
|
| 227 | } |
|
| 228 | ||
| 229 | $shaBlock = $this->shaBucketCache[$key]; |
|
| 230 | $low = 0; |
|
| 231 | $high = $count - 1; |
|
| 232 | $found = -1; |
|
| 233 | ||
| 234 | while( $low <= $high ) { |
|
| 235 | $mid = ($low + $high) >> 1; |
|
| 236 | $cmp = substr( $shaBlock, $mid * 20, 20 ); |
|
| 237 | ||
| 238 | if( $cmp < $binarySha ) { |
|
| 239 | $low = $mid + 1; |
|
| 240 | } elseif( $cmp > $binarySha ) { |
|
| 241 | $high = $mid - 1; |
|
| 242 | } else { |
|
| 243 | $found = $mid; |
|
| 244 | break; |
|
| 245 | } |
|
| 246 | } |
|
| 247 | ||
| 248 | if( $found !== -1 ) { |
|
| 249 | $packed = substr( $this->offsetBucketCache[$key], $found * 4, 4 ); |
|
| 250 | $offset = unpack( 'N', $packed )[1]; |
|
| 251 | ||
| 252 | if( $offset & 0x80000000 ) { |
|
| 253 | $pos64 = 1032 + ($total * 28) + (($offset & 0x7FFFFFFF) * 8); |
|
| 254 | fseek( $handle, $pos64 ); |
|
| 255 | $offset = unpack( 'J', fread( $handle, 8 ) )[1]; |
|
| 256 | } |
|
| 257 | $result = (int)$offset; |
|
| 258 | } |
|
| 259 | ||
| 260 | return $result; |
|
| 261 | } |
|
| 262 | ||
| 263 | private function readPackEntry( |
|
| 264 | $handle, |
|
| 265 | int $offset, |
|
| 266 | int $size, |
|
| 267 | int $cap = 0 |
|
| 268 | ): string { |
|
| 269 | fseek( $handle, $offset ); |
|
| 270 | $header = $this->readVarInt( $handle ); |
|
| 271 | $type = ($header['byte'] >> 4) & 7; |
|
| 272 | ||
| 273 | return ($type === 6) |
|
| 274 | ? $this->handleOfsDelta( $handle, $offset, $size, $cap ) |
|
| 275 | : (($type === 7) |
|
| 276 | ? $this->handleRefDelta( $handle, $size, $cap ) |
|
| 277 | : $this->decompressToString( $handle, $cap )); |
|
| 278 | } |
|
| 279 | ||
| 280 | private function streamPackEntry( |
|
| 281 | $handle, |
|
| 282 | int $offset, |
|
| 283 | int $size, |
|
| 284 | callable $callback, |
|
| 285 | int $depth = 0 |
|
| 286 | ): bool { |
|
| 287 | fseek( $handle, $offset ); |
|
| 288 | $header = $this->readVarInt( $handle ); |
|
| 289 | $type = ($header['byte'] >> 4) & 7; |
|
| 290 | ||
| 291 | return ($type === 6 || $type === 7) |
|
| 292 | ? $this->streamDeltaObject( $handle, $offset, $type, $callback, $depth ) |
|
| 293 | : $this->streamDecompression( $handle, $callback ); |
|
| 294 | } |
|
| 295 | ||
| 296 | private function streamDeltaObject( |
|
| 297 | $handle, |
|
| 298 | int $offset, |
|
| 299 | int $type, |
|
| 300 | callable $callback, |
|
| 301 | int $depth = 0 |
|
| 302 | ): bool { |
|
| 303 | if( $depth >= self::MAX_DEPTH ) { |
|
| 304 | error_log( "[GitPacks] delta depth limit exceeded at offset $offset" ); |
|
| 305 | return false; |
|
| 306 | } |
|
| 307 | ||
| 308 | fseek( $handle, $offset ); |
|
| 309 | $this->readVarInt( $handle ); |
|
| 310 | $result = false; |
|
| 311 | ||
| 312 | if( $type === 6 ) { |
|
| 313 | $neg = $this->readOffsetDelta( $handle ); |
|
| 314 | $deltaPos = ftell( $handle ); |
|
| 315 | $base = ''; |
|
| 316 | $baseSize = $this->extractPackedSize( $handle, $offset - $neg ); |
|
| 317 | ||
| 318 | if( $baseSize > self::MAX_BASE_RAM ) { |
|
| 319 | error_log( "[GitPacks] ofs-delta base too large for RAM path: $baseSize" ); |
|
| 320 | return false; |
|
| 321 | } |
|
| 322 | ||
| 323 | $this->streamPackEntry( |
|
| 324 | $handle, |
|
| 325 | $offset - $neg, |
|
| 326 | 0, |
|
| 327 | function( $c ) use ( &$base ) { $base .= $c; }, |
|
| 328 | $depth + 1 |
|
| 329 | ); |
|
| 330 | ||
| 331 | fseek( $handle, $deltaPos ); |
|
| 332 | $result = $this->applyDeltaStream( $handle, $base, $callback ); |
|
| 333 | } else { |
|
| 334 | $baseSha = bin2hex( fread( $handle, 20 ) ); |
|
| 335 | $baseSize = $this->getSize( $baseSha ); |
|
| 336 | ||
| 337 | if( $baseSize > self::MAX_BASE_RAM ) { |
|
| 338 | error_log( "[GitPacks] ref-delta base too large for RAM path: $baseSize (sha=$baseSha)" ); |
|
| 339 | return false; |
|
| 340 | } |
|
| 341 | ||
| 342 | $base = ''; |
|
| 343 | ||
| 344 | if( $this->streamInternal( $baseSha, function( $c ) use ( &$base ) { |
|
| 345 | $base .= $c; |
|
| 346 | }, $depth + 1 ) ) { |
|
| 347 | $result = $this->applyDeltaStream( $handle, $base, $callback ); |
|
| 348 | } |
|
| 349 | } |
|
| 350 | ||
| 351 | return $result; |
|
| 352 | } |
|
| 353 | ||
| 354 | private function applyDeltaStream( |
|
| 355 | $handle, |
|
| 356 | string $base, |
|
| 357 | callable $callback |
|
| 358 | ): bool { |
|
| 359 | $infl = inflate_init( ZLIB_ENCODING_DEFLATE ); |
|
| 360 | $ok = false; |
|
| 361 | ||
| 362 | if( $infl ) { |
|
| 363 | $state = 0; |
|
| 364 | $buffer = ''; |
|
| 365 | $ok = true; |
|
| 366 | ||
| 367 | while( !feof( $handle ) ) { |
|
| 368 | $chunk = fread( $handle, 8192 ); |
|
| 369 | ||
| 370 | if( $chunk === '' ) { |
|
| 371 | break; |
|
| 372 | } |
|
| 373 | ||
| 374 | $data = @inflate_add( $infl, $chunk ); |
|
| 375 | ||
| 376 | if( $data === false ) { |
|
| 377 | $ok = false; |
|
| 378 | break; |
|
| 379 | } |
|
| 380 | ||
| 381 | $buffer .= $data; |
|
| 382 | ||
| 383 | while( true ) { |
|
| 384 | $len = strlen( $buffer ); |
|
| 385 | ||
| 386 | if( $len === 0 ) { |
|
| 387 | break; |
|
| 388 | } |
|
| 389 | ||
| 390 | if( $state < 2 ) { |
|
| 391 | $pos = 0; |
|
| 392 | while( $pos < $len && (ord( $buffer[$pos] ) & 128) ) { $pos++; } |
|
| 393 | ||
| 394 | if( $pos === $len && (ord( $buffer[$pos - 1] ) & 128) ) { |
|
| 395 | break; |
|
| 396 | } |
|
| 397 | ||
| 398 | $buffer = substr( $buffer, $pos + 1 ); |
|
| 399 | $state++; |
|
| 400 | continue; |
|
| 401 | } |
|
| 402 | ||
| 403 | $op = ord( $buffer[0] ); |
|
| 404 | ||
| 405 | if( $op & 128 ) { |
|
| 406 | $need = $this->getCopyInstructionSize( $op ); |
|
| 407 | ||
| 408 | if( $len < 1 + $need ) { |
|
| 409 | break; |
|
| 410 | } |
|
| 411 | ||
| 412 | $info = $this->parseCopyInstruction( $op, $buffer, 1 ); |
|
| 413 | ||
| 414 | $callback( substr( $base, $info['off'], $info['len'] ) ); |
|
| 415 | $buffer = substr( $buffer, 1 + $need ); |
|
| 416 | } else { |
|
| 417 | $ln = $op & 127; |
|
| 418 | ||
| 419 | if( $len < 1 + $ln ) { |
|
| 420 | break; |
|
| 421 | } |
|
| 422 | ||
| 423 | $callback( substr( $buffer, 1, $ln ) ); |
|
| 424 | $buffer = substr( $buffer, 1 + $ln ); |
|
| 425 | } |
|
| 426 | } |
|
| 427 | ||
| 428 | if( inflate_get_status( $infl ) === ZLIB_STREAM_END ) { |
|
| 429 | break; |
|
| 430 | } |
|
| 431 | } |
|
| 432 | } |
|
| 433 | ||
| 434 | return $ok; |
|
| 435 | } |
|
| 436 | ||
| 437 | private function streamPackEntryGenerator( |
|
| 438 | $handle, |
|
| 439 | int $offset, |
|
| 440 | int $depth |
|
| 441 | ): Generator { |
|
| 442 | fseek( $handle, $offset ); |
|
| 443 | $header = $this->readVarInt( $handle ); |
|
| 444 | $type = ($header['byte'] >> 4) & 7; |
|
| 445 | ||
| 446 | if( $type === 6 || $type === 7 ) { |
|
| 447 | yield from $this->streamDeltaObjectGenerator( |
|
| 448 | $handle, |
|
| 449 | $offset, |
|
| 450 | $type, |
|
| 451 | $depth |
|
| 452 | ); |
|
| 453 | } else { |
|
| 454 | yield from $this->streamDecompressionGenerator( $handle ); |
|
| 455 | } |
|
| 456 | } |
|
| 457 | ||
| 458 | /** |
|
| 459 | * Decompresses the pack entry at $baseOffset into a temp file and returns |
|
| 460 | * the open handle rewound to byte 0, or null if tmpfile() fails. |
|
| 461 | * The caller is responsible for fclose()-ing the handle. |
|
| 462 | */ |
|
| 463 | private function resolveBaseToTempFile( |
|
| 464 | $packHandle, |
|
| 465 | int $baseOffset, |
|
| 466 | int $depth |
|
| 467 | ) { |
|
| 468 | $tmpHandle = tmpfile(); |
|
| 469 | ||
| 470 | if( !$tmpHandle ) { |
|
| 471 | error_log( |
|
| 472 | "[GitPacks] tmpfile() failed for ofs-delta base at offset $baseOffset" |
|
| 473 | ); |
|
| 474 | return null; |
|
| 475 | } |
|
| 476 | ||
| 477 | foreach( $this->streamPackEntryGenerator( $packHandle, $baseOffset, $depth + 1 ) as $chunk ) { |
|
| 478 | fwrite( $tmpHandle, $chunk ); |
|
| 479 | } |
|
| 480 | ||
| 481 | rewind( $tmpHandle ); |
|
| 482 | ||
| 483 | return $tmpHandle; |
|
| 484 | } |
|
| 485 | ||
| 486 | private function streamDeltaObjectGenerator( |
|
| 487 | $handle, |
|
| 488 | int $offset, |
|
| 489 | int $type, |
|
| 490 | int $depth |
|
| 491 | ): Generator { |
|
| 492 | if( $depth >= self::MAX_DEPTH ) { |
|
| 493 | error_log( "[GitPacks] delta depth limit exceeded at offset $offset" ); |
|
| 494 | return; |
|
| 495 | } |
|
| 496 | ||
| 497 | fseek( $handle, $offset ); |
|
| 498 | $this->readVarInt( $handle ); |
|
| 499 | ||
| 500 | if( $type === 6 ) { |
|
| 501 | $neg = $this->readOffsetDelta( $handle ); |
|
| 502 | $deltaPos = ftell( $handle ); |
|
| 503 | $baseSize = $this->extractPackedSize( $handle, $offset - $neg ); |
|
| 504 | ||
| 505 | if( $baseSize > self::MAX_BASE_RAM ) { |
|
| 506 | $tmpHandle = $this->resolveBaseToTempFile( |
|
| 507 | $handle, |
|
| 508 | $offset - $neg, |
|
| 509 | $depth |
|
| 510 | ); |
|
| 511 | ||
| 512 | if( $tmpHandle === null ) { |
|
| 513 | return; |
|
| 514 | } |
|
| 515 | ||
| 516 | fseek( $handle, $deltaPos ); |
|
| 517 | yield from $this->applyDeltaStreamFromFileGenerator( |
|
| 518 | $handle, |
|
| 519 | $tmpHandle |
|
| 520 | ); |
|
| 521 | fclose( $tmpHandle ); |
|
| 522 | } else { |
|
| 523 | $base = ''; |
|
| 524 | $this->streamPackEntry( |
|
| 525 | $handle, |
|
| 526 | $offset - $neg, |
|
| 527 | 0, |
|
| 528 | function( $c ) use ( &$base ) { $base .= $c; }, |
|
| 529 | $depth + 1 |
|
| 530 | ); |
|
| 531 | fseek( $handle, $deltaPos ); |
|
| 532 | yield from $this->applyDeltaStreamGenerator( $handle, $base ); |
|
| 533 | } |
|
| 534 | } else { |
|
| 535 | $baseSha = bin2hex( fread( $handle, 20 ) ); |
|
| 536 | $baseSize = $this->getSize( $baseSha ); |
|
| 537 | ||
| 538 | if( $baseSize > self::MAX_BASE_RAM ) { |
|
| 539 | $tmpHandle = tmpfile(); |
|
| 540 | ||
| 541 | if( !$tmpHandle ) { |
|
| 542 | error_log( |
|
| 543 | "[GitPacks] tmpfile() failed for ref-delta base (sha=$baseSha)" |
|
| 544 | ); |
|
| 545 | return; |
|
| 546 | } |
|
| 547 | ||
| 548 | $written = $this->streamInternal( |
|
| 549 | $baseSha, |
|
| 550 | function( $c ) use ( $tmpHandle ) { fwrite( $tmpHandle, $c ); }, |
|
| 551 | $depth + 1 |
|
| 552 | ); |
|
| 553 | ||
| 554 | if( $written ) { |
|
| 555 | rewind( $tmpHandle ); |
|
| 556 | yield from $this->applyDeltaStreamFromFileGenerator( |
|
| 557 | $handle, |
|
| 558 | $tmpHandle |
|
| 559 | ); |
|
| 560 | } |
|
| 561 | ||
| 562 | fclose( $tmpHandle ); |
|
| 563 | } else { |
|
| 564 | $base = ''; |
|
| 565 | ||
| 566 | if( $this->streamInternal( |
|
| 567 | $baseSha, |
|
| 568 | function( $c ) use ( &$base ) { $base .= $c; }, |
|
| 569 | $depth + 1 |
|
| 570 | ) ) { |
|
| 571 | yield from $this->applyDeltaStreamGenerator( $handle, $base ); |
|
| 572 | } |
|
| 573 | } |
|
| 574 | } |
|
| 575 | } |
|
| 576 | ||
| 577 | private function applyDeltaStreamGenerator( |
|
| 578 | $handle, |
|
| 579 | string $base |
|
| 580 | ): Generator { |
|
| 581 | $infl = inflate_init( ZLIB_ENCODING_DEFLATE ); |
|
| 582 | ||
| 583 | if( !$infl ) { |
|
| 584 | return; |
|
| 585 | } |
|
| 586 | ||
| 587 | $state = 0; |
|
| 588 | $buffer = ''; |
|
| 589 | ||
| 590 | while( !feof( $handle ) ) { |
|
| 591 | $chunk = fread( $handle, 8192 ); |
|
| 592 | ||
| 593 | if( $chunk === '' ) { |
|
| 594 | break; |
|
| 595 | } |
|
| 596 | ||
| 597 | $data = @inflate_add( $infl, $chunk ); |
|
| 598 | ||
| 599 | if( $data === false ) { |
|
| 600 | break; |
|
| 601 | } |
|
| 602 | ||
| 603 | $buffer .= $data; |
|
| 604 | ||
| 605 | while( true ) { |
|
| 606 | $len = strlen( $buffer ); |
|
| 607 | ||
| 608 | if( $len === 0 ) { |
|
| 609 | break; |
|
| 610 | } |
|
| 611 | ||
| 612 | if( $state < 2 ) { |
|
| 613 | $pos = 0; |
|
| 614 | while( $pos < $len && (ord( $buffer[$pos] ) & 128) ) { $pos++; } |
|
| 615 | ||
| 616 | if( $pos === $len && (ord( $buffer[$pos - 1] ) & 128) ) { |
|
| 617 | break; |
|
| 618 | } |
|
| 619 | ||
| 620 | $buffer = substr( $buffer, $pos + 1 ); |
|
| 621 | $state++; |
|
| 622 | continue; |
|
| 623 | } |
|
| 624 | ||
| 625 | $op = ord( $buffer[0] ); |
|
| 626 | ||
| 627 | if( $op & 128 ) { |
|
| 628 | $need = $this->getCopyInstructionSize( $op ); |
|
| 629 | ||
| 630 | if( $len < 1 + $need ) { |
|
| 631 | break; |
|
| 632 | } |
|
| 633 | ||
| 634 | $info = $this->parseCopyInstruction( $op, $buffer, 1 ); |
|
| 635 | yield substr( $base, $info['off'], $info['len'] ); |
|
| 636 | $buffer = substr( $buffer, 1 + $need ); |
|
| 637 | } else { |
|
| 638 | $ln = $op & 127; |
|
| 639 | ||
| 640 | if( $len < 1 + $ln ) { |
|
| 641 | break; |
|
| 642 | } |
|
| 643 | ||
| 644 | yield substr( $buffer, 1, $ln ); |
|
| 645 | $buffer = substr( $buffer, 1 + $ln ); |
|
| 646 | } |
|
| 647 | } |
|
| 648 | ||
| 649 | if( inflate_get_status( $infl ) === ZLIB_STREAM_END ) { |
|
| 650 | break; |
|
| 651 | } |
|
| 652 | } |
|
| 653 | } |
|
| 654 | ||
| 655 | private function applyDeltaStreamFromFileGenerator( |
|
| 656 | $deltaHandle, |
|
| 657 | $baseHandle |
|
| 658 | ): Generator { |
|
| 659 | $infl = inflate_init( ZLIB_ENCODING_DEFLATE ); |
|
| 660 | ||
| 661 | if( !$infl ) { |
|
| 662 | return; |
|
| 663 | } |
|
| 664 | ||
| 665 | $state = 0; |
|
| 666 | $buffer = ''; |
|
| 667 | ||
| 668 | while( !feof( $deltaHandle ) ) { |
|
| 669 | $chunk = fread( $deltaHandle, 8192 ); |
|
| 670 | ||
| 671 | if( $chunk === '' ) { |
|
| 672 | break; |
|
| 673 | } |
|
| 674 | ||
| 675 | $data = @inflate_add( $infl, $chunk ); |
|
| 676 | ||
| 677 | if( $data === false ) { |
|
| 678 | break; |
|
| 679 | } |
|
| 680 | ||
| 681 | $buffer .= $data; |
|
| 682 | ||
| 683 | while( true ) { |
|
| 684 | $len = strlen( $buffer ); |
|
| 685 | ||
| 686 | if( $len === 0 ) { |
|
| 687 | break; |
|
| 688 | } |
|
| 689 | ||
| 690 | if( $state < 2 ) { |
|
| 691 | $pos = 0; |
|
| 692 | while( $pos < $len && (ord( $buffer[$pos] ) & 128) ) { $pos++; } |
|
| 693 | ||
| 694 | if( $pos === $len && (ord( $buffer[$pos - 1] ) & 128) ) { |
|
| 695 | break; |
|
| 696 | } |
|
| 697 | ||
| 698 | $buffer = substr( $buffer, $pos + 1 ); |
|
| 699 | $state++; |
|
| 700 | continue; |
|
| 701 | } |
|
| 702 | ||
| 703 | $op = ord( $buffer[0] ); |
|
| 704 | ||
| 705 | if( $op & 128 ) { |
|
| 706 | $need = $this->getCopyInstructionSize( $op ); |
|
| 707 | ||
| 708 | if( $len < 1 + $need ) { |
|
| 709 | break; |
|
| 710 | } |
|
| 711 | ||
| 712 | $info = $this->parseCopyInstruction( $op, $buffer, 1 ); |
|
| 713 | fseek( $baseHandle, $info['off'] ); |
|
| 714 | $remaining = $info['len']; |
|
| 715 | ||
| 716 | while( $remaining > 0 ) { |
|
| 717 | $slice = fread( $baseHandle, min( 65536, $remaining ) ); |
|
| 718 | ||
| 719 | if( $slice === false || $slice === '' ) { |
|
| 720 | break; |
|
| 721 | } |
|
| 722 | ||
| 723 | yield $slice; |
|
| 724 | $remaining -= strlen( $slice ); |
|
| 725 | } |
|
| 726 | ||
| 727 | $buffer = substr( $buffer, 1 + $need ); |
|
| 728 | } else { |
|
| 729 | $ln = $op & 127; |
|
| 730 | ||
| 731 | if( $len < 1 + $ln ) { |
|
| 732 | break; |
|
| 733 | } |
|
| 734 | ||
| 735 | yield substr( $buffer, 1, $ln ); |
|
| 736 | $buffer = substr( $buffer, 1 + $ln ); |
|
| 737 | } |
|
| 738 | } |
|
| 739 | ||
| 740 | if( inflate_get_status( $infl ) === ZLIB_STREAM_END ) { |
|
| 741 | break; |
|
| 742 | } |
|
| 743 | } |
|
| 744 | } |
|
| 745 | ||
| 746 | private function streamDecompressionGenerator( $handle ): Generator { |
|
| 747 | $infl = inflate_init( ZLIB_ENCODING_DEFLATE ); |
|
| 748 | ||
| 749 | if( !$infl ) { |
|
| 750 | return; |
|
| 751 | } |
|
| 752 | ||
| 753 | while( !feof( $handle ) ) { |
|
| 754 | $chunk = fread( $handle, 8192 ); |
|
| 755 | ||
| 756 | if( $chunk === '' ) { |
|
| 757 | break; |
|
| 758 | } |
|
| 759 | ||
| 760 | $data = @inflate_add( $infl, $chunk ); |
|
| 761 | ||
| 762 | if( $data !== false && $data !== '' ) { |
|
| 763 | yield $data; |
|
| 764 | } |
|
| 765 | ||
| 766 | if( $data === false || |
|
| 767 | inflate_get_status( $infl ) === ZLIB_STREAM_END ) { |
|
| 768 | break; |
|
| 769 | } |
|
| 770 | } |
|
| 416 | 771 | } |
| 417 | 772 |
| 16 | 16 | |
| 17 | 17 | if( $input === 'HEAD' && file_exists( $headFile ) ) { |
| 18 | $head = trim( file_get_contents( $headFile ) ); |
|
| 18 | $head = trim( file_get_contents( $headFile ) ); |
|
| 19 | 19 | $result = strpos( $head, 'ref: ' ) === 0 |
| 20 | 20 | ? $this->resolve( substr( $head, 5 ) ) |
| ... | ||
| 47 | 47 | } |
| 48 | 48 | |
| 49 | if( empty( $found ) ) { |
|
| 50 | $key = array_key_first( $branches ); |
|
| 51 | $found = $key |
|
| 52 | ? [ 'name' => $key, 'hash' => $branches[$key] ] |
|
| 53 | : [ 'name' => '', 'hash' => '' ]; |
|
| 49 | if( empty( $found ) && !empty( $branches ) ) { |
|
| 50 | $key = array_key_first( $branches ); |
|
| 51 | $found = [ 'name' => $key, 'hash' => $branches[$key] ]; |
|
| 52 | } elseif( empty( $found ) ) { |
|
| 53 | $found = [ 'name' => '', 'hash' => '' ]; |
|
| 54 | 54 | } |
| 55 | 55 | |
| 56 | 56 | return $found; |
| 57 | 57 | } |
| 58 | 58 | |
| 59 | 59 | public function scanRefs( string $prefix, callable $callback ): void { |
| 60 | 60 | $dir = "{$this->repoPath}/$prefix"; |
| 61 | ||
| 62 | $this->traverseDirectory( $dir, $callback, '' ); |
|
| 63 | } |
|
| 61 | 64 | |
| 65 | private function traverseDirectory( |
|
| 66 | string $dir, |
|
| 67 | callable $callback, |
|
| 68 | string $subPath |
|
| 69 | ): void { |
|
| 62 | 70 | if( is_dir( $dir ) ) { |
| 63 | 71 | $files = array_diff( scandir( $dir ), ['.', '..'] ); |
| 72 | ||
| 64 | 73 | foreach( $files as $file ) { |
| 65 | $callback( $file, trim( file_get_contents( "$dir/$file" ) ) ); |
|
| 74 | $path = "$dir/$file"; |
|
| 75 | $name = $subPath === '' ? $file : "$subPath/$file"; |
|
| 76 | ||
| 77 | if( is_dir( $path ) ) { |
|
| 78 | $this->traverseDirectory( $path, $callback, $name ); |
|
| 79 | } elseif( is_file( $path ) ) { |
|
| 80 | $sha = trim( file_get_contents( $path ) ); |
|
| 81 | ||
| 82 | if( preg_match( '/^[0-9a-f]{40}$/', $sha ) ) { |
|
| 83 | $callback( $name, $sha ); |
|
| 84 | } |
|
| 85 | } |
|
| 66 | 86 | } |
| 67 | 87 | } |
| ... | ||
| 74 | 94 | foreach( $paths as $ref ) { |
| 75 | 95 | $path = "{$this->repoPath}/$ref"; |
| 96 | ||
| 76 | 97 | if( file_exists( $path ) ) { |
| 77 | 98 | $result = trim( file_get_contents( $path ) ); |
| 78 | 99 | break; |
| 79 | 100 | } |
| 80 | 101 | } |
| 81 | 102 | |
| 82 | 103 | if( $result === '' ) { |
| 83 | 104 | $packedPath = "{$this->repoPath}/packed-refs"; |
| 105 | ||
| 84 | 106 | if( file_exists( $packedPath ) ) { |
| 85 | 107 | $result = $this->findInPackedRefs( $packedPath, $input ); |
| 86 | 108 | } |
| 109 | } |
|
| 110 | ||
| 111 | if( !preg_match( '/^[0-9a-f]{40}$/', $result ) ) { |
|
| 112 | $result = ''; |
|
| 87 | 113 | } |
| 88 | 114 | |
| ... | ||
| 98 | 124 | if( $line[0] !== '#' && $line[0] !== '^' ) { |
| 99 | 125 | $parts = explode( ' ', trim( $line ) ); |
| 126 | ||
| 100 | 127 | if( count( $parts ) >= 2 && in_array( $parts[1], $targets ) ) { |
| 101 | 128 | $result = $parts[0]; |
| 87 | 87 | |
| 88 | 88 | private function handleUploadPack(): void { |
| 89 | set_time_limit( 0 ); |
|
| 90 | ||
| 89 | 91 | header( 'Content-Type: application/x-git-upload-pack-result' ); |
| 90 | 92 | header( 'Cache-Control: no-cache' ); |
| 91 | ||
| 92 | $input = file_get_contents( 'php://input' ); |
|
| 93 | $wants = []; |
|
| 94 | $haves = []; |
|
| 95 | $offset = 0; |
|
| 96 | $isGzip = isset( $_SERVER['HTTP_CONTENT_ENCODING'] ) && |
|
| 97 | $_SERVER['HTTP_CONTENT_ENCODING'] === 'gzip'; |
|
| 98 | 93 | |
| 99 | if( $isGzip ) { |
|
| 100 | $decoded = gzdecode( $input ); |
|
| 94 | $wants = []; |
|
| 95 | $haves = []; |
|
| 96 | $handle = fopen( 'php://input', 'rb' ); |
|
| 101 | 97 | |
| 102 | if( is_string( $decoded ) ) { |
|
| 103 | $input = $decoded; |
|
| 98 | if( $handle ) { |
|
| 99 | // If the input is gzipped, we wrap the stream |
|
| 100 | if( isset( $_SERVER['HTTP_CONTENT_ENCODING'] ) && |
|
| 101 | $_SERVER['HTTP_CONTENT_ENCODING'] === 'gzip' ) { |
|
| 102 | stream_filter_append( $handle, 'zlib.inflate', STREAM_FILTER_READ, [ |
|
| 103 | 'window' => 31 |
|
| 104 | ] ); |
|
| 104 | 105 | } |
| 105 | } |
|
| 106 | 106 | |
| 107 | while( $offset < strlen( $input ) ) { |
|
| 108 | $result = $this->readPacketLine( $input, $offset ); |
|
| 109 | $line = $result[0]; |
|
| 110 | $next = $result[1]; |
|
| 107 | while( !feof( $handle ) ) { |
|
| 108 | $lenHex = fread( $handle, 4 ); |
|
| 111 | 109 | |
| 112 | if( $next === $offset || $line === 'done' ) { |
|
| 113 | break; |
|
| 114 | } |
|
| 110 | if( strlen( $lenHex ) < 4 ) { |
|
| 111 | break; |
|
| 112 | } |
|
| 115 | 113 | |
| 116 | $offset = $next; |
|
| 114 | $len = hexdec( $lenHex ); |
|
| 117 | 115 | |
| 118 | if( $line === '' ) { |
|
| 119 | continue; |
|
| 120 | } |
|
| 116 | if( $len === 0 ) { // Flush packet |
|
| 117 | break; |
|
| 118 | } |
|
| 121 | 119 | |
| 122 | $trim = trim( $line ); |
|
| 120 | if( $len <= 4 ) { |
|
| 121 | continue; |
|
| 122 | } |
|
| 123 | 123 | |
| 124 | if( strpos( $trim, 'want ' ) === 0 ) { |
|
| 125 | $wants[] = explode( ' ', $trim )[1]; |
|
| 126 | } elseif( strpos( $trim, 'have ' ) === 0 ) { |
|
| 127 | $haves[] = explode( ' ', $trim )[1]; |
|
| 124 | $line = fread( $handle, $len - 4 ); |
|
| 125 | $trim = trim( $line ); |
|
| 126 | ||
| 127 | if( strpos( $trim, 'want ' ) === 0 ) { |
|
| 128 | $wants[] = explode( ' ', $trim )[1]; |
|
| 129 | } elseif( strpos( $trim, 'have ' ) === 0 ) { |
|
| 130 | $haves[] = explode( ' ', $trim )[1]; |
|
| 131 | } |
|
| 132 | ||
| 133 | if( $trim === 'done' ) { |
|
| 134 | break; |
|
| 135 | } |
|
| 128 | 136 | } |
| 137 | ||
| 138 | fclose( $handle ); |
|
| 129 | 139 | } |
| 130 | 140 | |
| 131 | if( $wants ) { |
|
| 141 | if( !empty( $wants ) ) { |
|
| 132 | 142 | $this->packetWrite( "NAK\n" ); |
| 133 | 143 | |
| 134 | $objects = $this->git->collectObjects( $wants, $haves ); |
|
| 135 | $pack = $this->git->generatePackfile( $objects ); |
|
| 144 | $objects = $this->git->collectObjects( $wants, $haves ); |
|
| 145 | $lastHeartbeat = time(); |
|
| 136 | 146 | |
| 137 | $this->sendSidebandData( 1, $pack ); |
|
| 147 | foreach( $this->git->generatePackfile( $objects ) as $chunk ) { |
|
| 148 | if( $chunk !== '' ) { |
|
| 149 | $this->sendSidebandData( 1, $chunk ); |
|
| 150 | } |
|
| 151 | ||
| 152 | $now = time(); |
|
| 153 | if( $now - $lastHeartbeat >= 5 ) { |
|
| 154 | $this->sendSidebandData( 2, "\r" ); |
|
| 155 | $lastHeartbeat = $now; |
|
| 156 | } |
|
| 157 | } |
|
| 138 | 158 | } |
| 139 | 159 | |
| 140 | 160 | $this->packetFlush(); |
| 141 | 161 | } |
| 142 | 162 | |
| 143 | 163 | private function sendSidebandData( int $band, string $data ): void { |
| 144 | $chunkSize = 65000; |
|
| 164 | $chunkSize = 65515; |
|
| 145 | 165 | $len = strlen( $data ); |
| 146 | 166 |
| 1 | user-agent: * |
|
| 2 | disallow: |
|
| 3 | crawl-delay: 60 |
|
| 1 | user-agent: * |
|
| 2 | disallow: |
|
| 3 | crawl-delay: 10080 |
|
| 4 | 4 | |
| 5 | User-agent: Googlebot |
|
| 6 | Disallow: /_mobile/ |
|
| 7 | Disallow: /*repo= |
|
| 8 | Disallow: /*action= |
|
| 5 | user-agent: Googlebot |
|
| 6 | disallow: /_mobile/ |
|
| 7 | disallow: /*repo= |
|
| 8 | disallow: /*action= |
|
| 9 | ||
| 9 | 10 |