<?php
/*************************************************************
# 076 License

Copyright (c) テクニカル諏訪子

Permission is hereby granted to any person obtaining a copy of the software
Little Beast (the "Software") to use, modify, merge, copy, publish, distribute,
sublicense, and/or sell copies of the Software, subject to the following conditions:

    1. **Origin Attribution**:
       - You must not misrepresent the origin of the Software; you must not claim
         you created the original Software.
       - If the Software is used in a product, you must either:
         a. Provide clear attribution in the product's documentation, user interface,
            or other visible areas, **OR**
         b. Pay the original developers a fee they specify in writing.
    2. **Usage Restriction**:
       - The Software, or any derivative works, dependencies, or libraries
         incorporating it, must not be used for censorship or to suppress freedom of
         speech, expression, or creativity. Prohibited uses include, but are not
         limited to:
         - Censorship of so-called "hate speech", visuals, non-mainstream opinions,
           ideas, or objective reality.
         - Tools or systems designed to restrict access to information or
           artistic works.
    3. **Notice Preservation**:
       - This license and the above copyright notice must remain intact in all copies
         of the source code.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED,
INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A
PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT
HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF
CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE
OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 */
namespace Std\Lib\Git;

/**
 * パックファイル（objects/pack/pack-*.pack）とインデックス（.idx v2、v1）
 *
 * オブジェクトの種類：1 commit、2 tree、3 blob、4 tag、6 OFS_DELTA、7 REF_DELTA。
 */
final class Pack {
  public const COMMIT = 1;
  public const TREE = 2;
  public const BLOB = 3;
  public const TAG = 4;
  public const OFS_DELTA = 6;
  public const REF_DELTA = 7;
  public const TYPES = [1 => 'commit', 2 => 'tree', 3 => 'blob', 4 => 'tag'];

  /** @var resource|null */
  private $fh = null;
  private string $idx = '';
  private int $version = 2;
  private int $count = 0;
  private int $namesAt = 0;
  private int $crcAt = 0;
  private int $offsetsAt = 0;
  private int $largeAt = 0;
  private ?string $rev = null;
  public readonly int $size;

  /**
   * @param string $path  .packのパス
   * @param int $hashLen  20（SHA-1）又は32（SHA-256）
   * @param bool $withIndex  .idxを読む（受け取り中のパックはfalse）
   */
  public function __construct(public readonly string $path, private int $hashLen, bool $withIndex = true) {
    $fh = @fopen($path, 'rb');
    if ($fh === false) throw new GitException("パックを開けません：{$path}");
    $this->fh = $fh;
    $this->size = (int)fstat($fh)['size'];
    $head = (string)fread($fh, 12);
    if (strlen($head) < 12 || substr($head, 0, 4) !== 'PACK' || !in_array(unpack('N', $head, 4)[1], [2, 3], true)) throw new GitException("パックではありません：{$path}");
    if ($withIndex) $this->loadIndex(substr($path, 0, -5).'.idx');
  }

  public function __destruct() {
    if ($this->fh) fclose($this->fh);
  }

  private function loadIndex(string $file): void {
    $idx = @file_get_contents($file);
    if ($idx === false) throw new GitException("インデックスを読めません：{$file}");
    $this->idx = $idx;
    $h = $this->hashLen;
    if (substr($idx, 0, 4) === "\xfftOc") {
      $this->version = unpack('N', $idx, 4)[1];
      if ($this->version !== 2) throw new GitException("未対応のインデックスのバージョン：{$this->version}");
      $this->count = unpack('N', $idx, 8 + 255 * 4)[1];
      $this->namesAt = 8 + 256 * 4;
      $this->crcAt = $this->namesAt + $this->count * $h;
      $this->offsetsAt = $this->crcAt + $this->count * 4;
      $this->largeAt = $this->offsetsAt + $this->count * 4;
    } else {
      // v1：fanout、[オフセット(4) 名前]の配列
      $this->version = 1;
      $this->count = unpack('N', $idx, 255 * 4)[1];
      $this->namesAt = 256 * 4;
    }
    if (strlen($idx) < $this->namesAt + $this->count * $h) throw new GitException("壊れたインデックス：{$file}");
  }

  public function count(): int {
    return $this->count;
  }

  /** n番目（名前の順）のID（バイナリ） */
  public function idAt(int $n): string {
    return $this->version === 2 ? substr($this->idx, $this->namesAt + $n * $this->hashLen, $this->hashLen)
                                : substr($this->idx, $this->namesAt + $n * ($this->hashLen + 4) + 4, $this->hashLen);
  }

  /** n番目のオフセット */
  public function offsetAt(int $n): int {
    if ($this->version === 1) return unpack('N', $this->idx, $this->namesAt + $n * ($this->hashLen + 4))[1];
    $o = unpack('N', $this->idx, $this->offsetsAt + $n * 4)[1];
    if ($o & 0x80000000) $o = unpack('J', $this->idx, $this->largeAt + ($o & 0x7fffffff) * 8)[1];
    return $o;
  }

  public function crcAt(int $n): ?int {
    return $this->version === 2 ? unpack('N', $this->idx, $this->crcAt + $n * 4)[1] : null;
  }

  /**
   * IDの位置（名前の順の番号）
   *
   * @param string $id  バイナリのID
   */
  public function position(string $id): ?int {
    $first = ord($id[0]);
    $fan = $this->version === 2 ? 8 : 0;
    $lo = $first === 0 ? 0 : unpack('N', $this->idx, $fan + ($first - 1) * 4)[1];
    $hi = unpack('N', $this->idx, $fan + $first * 4)[1] - 1;
    while ($lo <= $hi) {
      $mid = ($lo + $hi) >> 1;
      $c = strcmp($this->idAt($mid), $id);
      if ($c === 0) return $mid;
      if ($c < 0) $lo = $mid + 1;
      else $hi = $mid - 1;
    }
    return null;
  }

  /** IDのオフセット（無ければnull） */
  public function find(string $id): ?int {
    $n = $this->position($id);
    return $n === null ? null : $this->offsetAt($n);
  }

  /**
   * 始まりが一致するID（短いIDの解決用）
   *
   * @param string $prefix  16進数（偶数でなくても良い）
   * @return string[]  16進数（最大2件）
   */
  public function findPrefix(string $prefix): array {
    $out = [];
    $first = hexdec(substr($prefix . '0', 0, 2));
    $fan = $this->version === 2 ? 8 : 0;
    $lo = $first === 0 ? 0 : unpack('N', $this->idx, $fan + ($first - 1) * 4)[1];
    $hi = unpack('N', $this->idx, $fan + (strlen($prefix) >= 2 ? $first : min(255, $first + 15)) * 4)[1];
    for ($n = $lo; $n < $hi && count($out) < 2; $n++) {
      $hex = bin2hex($this->idAt($n));
      if (str_starts_with($hex, $prefix)) $out[] = $hex;
    }
    return $out;
  }

  /** 全部のID（16進数） */
  public function ids(): \Generator {
    for ($n = 0; $n < $this->count; $n++) yield bin2hex($this->idAt($n));
  }

  /**
   * エントリーのヘッダー
   *
   * @return array  [種類, 大きさ, データの位置, 元のオフセット（OFS）, 元のID（REF、バイナリ）]
   */
  public function header(int $offset): array {
    fseek($this->fh, $offset);
    $buf = (string)fread($this->fh, 32 + $this->hashLen);
    $p = 0;
    $b = ord($buf[$p++]);
    $type = ($b >> 4) & 7;
    $size = $b & 15;
    $shift = 4;
    while ($b & 0x80) {
      $b = ord($buf[$p++]);
      $size |= ($b & 0x7f) << $shift;
      $shift += 7;
    }
    $baseOffset = null;
    $baseId = null;
    if ($type === self::OFS_DELTA) {
      $b = ord($buf[$p++]);
      $n = $b & 0x7f;
      while ($b & 0x80) {
        $b = ord($buf[$p++]);
        $n = (($n + 1) << 7) | ($b & 0x7f);
      }
      $baseOffset = $offset - $n;
      if ($baseOffset < 12 || $baseOffset >= $offset) throw new GitException("壊れたパック（OFS_DELTA）：{$this->path}");
    } else if ($type === self::REF_DELTA) {
      $baseId = substr($buf, $p, $this->hashLen);
      $p += $this->hashLen;
    } else if (!isset(self::TYPES[$type])) {
      throw new GitException("壊れたパック（種類{$type}、{$offset}）：{$this->path}");
    }
    return [$type, $size, $offset + $p, $baseOffset, $baseId];
  }

  /**
   * zlibのデータを展開する
   *
   * @param int $pos  データの位置
   * @param int $size  展開後の大きさ（分かっていれば読む量の目安）
   * @param int|null $consumed  圧縮されたデータの長さ
   */
  public function inflate(int $pos, int $size, ?int &$consumed = null): string {
    fseek($this->fh, $pos);
    $ctx = inflate_init(ZLIB_ENCODING_DEFLATE);
    $out = '';
    $chunk = min(max(256, $size + 64), 1048576);
    while (true) {
      $in = fread($this->fh, $chunk);
      if ($in === false || $in === '') throw new GitException("壊れたパック（途中で終わる）：{$this->path}");
      $res = @inflate_add($ctx, $in, ZLIB_SYNC_FLUSH);
      if ($res === false) throw new GitException("壊れたパック（zlib）：{$this->path}");
      $out .= $res;
      if (inflate_get_status($ctx) === ZLIB_STREAM_END) break;
      $chunk = 65536;
    }
    $consumed = inflate_get_read_len($ctx);
    if (strlen($out) !== $size) throw new GitException("壊れたパック（大きさ）：{$this->path}");
    return $out;
  }

  /** 位置のオフセットに対する次のエントリーの始まり（.revか、無ければ作る） */
  public function entryEnd(int $offset): int {
    $this->loadRev();
    // オフセットの順のn番目
    $lo = 0;
    $hi = $this->count - 1;
    while ($lo <= $hi) {
      $mid = ($lo + $hi) >> 1;
      $o = $this->offsetAt(unpack('N', $this->rev, $mid * 4)[1]);
      if ($o === $offset) return $mid + 1 < $this->count ? $this->offsetAt(unpack('N', $this->rev, ($mid + 1) * 4)[1]) : $this->size - $this->hashLen;
      if ($o < $offset) $lo = $mid + 1;
      else $hi = $mid - 1;
    }
    throw new GitException("パックにエントリーが有りません（{$offset}）：{$this->path}");
  }

  /** オフセットのID（バイナリ） */
  public function idAtOffset(int $offset): ?string {
    $this->loadRev();
    $lo = 0;
    $hi = $this->count - 1;
    while ($lo <= $hi) {
      $mid = ($lo + $hi) >> 1;
      $n = unpack('N', $this->rev, $mid * 4)[1];
      $o = $this->offsetAt($n);
      if ($o === $offset) return $this->idAt($n);
      if ($o < $offset) $lo = $mid + 1;
      else $hi = $mid - 1;
    }
    return null;
  }

  /** 逆インデックス（オフセットの順の名前の番号）。.revが有れば使う */
  private function loadRev(): void {
    if ($this->rev !== null) return;
    $file = substr($this->path, 0, -5).'.rev';
    $r = is_file($file) ? (string)file_get_contents($file) : '';
    if (strlen($r) >= 12 + $this->count * 4 && substr($r, 0, 4) === 'RIDX' && unpack('N', $r, 4)[1] === 1) {
      $this->rev = substr($r, 12, $this->count * 4);
      return;
    }
    $offs = [];
    for ($n = 0; $n < $this->count; $n++) $offs[$n] = $this->offsetAt($n);
    asort($offs);
    $this->rev = pack('N*', ...array_keys($offs));
  }

  /** パックの中身の生のバイト列（ヘッダーと圧縮されたデータ） */
  public function raw(int $start, int $end): string {
    fseek($this->fh, $start);
    $s = (string)fread($this->fh, $end - $start);
    if (strlen($s) !== $end - $start) throw new GitException("パックを読めません：{$this->path}");
    return $s;
  }
}