Documentation/technical/commit-graph-format.txt

   1 Git commit graph format
   2 =======================
   3
   4 The Git commit graph stores a list of commit OIDs and some associated
   5 metadata, including:
   6
   7 - The generation number of the commit. Commits with no parents have
   8   generation number 1; commits with parents have generation number
   9   one more than the maximum generation number of its parents. We
  10   reserve zero as special, and can be used to mark a generation
  11   number invalid or as "not computed".
  12
  13 - The root tree OID.
  14
  15 - The commit date.
  16
  17 - The parents of the commit, stored using positional references within
  18   the graph file.
  19
  20 These positional references are stored as unsigned 32-bit integers
  21 corresponding to the array position within the list of commit OIDs. Due
  22 to some special constants we use to track parents, we can store at most
  23 (1 << 30) + (1 << 29) + (1 << 28) - 1 (around 1.8 billion) commits.
  24
  25 == Commit graph files have the following format:
  26
  27 In order to allow extensions that add extra data to the graph, we organize
  28 the body into "chunks" and provide a binary lookup table at the beginning
  29 of the body. The header includes certain values, such as number of chunks
  30 and hash type.
  31
  32 All 4-byte numbers are in network order.
  33
  34 HEADER:
  35
  36   4-byte signature:
  37       The signature is: {'C', 'G', 'P', 'H'}
  38
  39   1-byte version number:
  40       Currently, the only valid version is 1.
  41
  42   1-byte Hash Version (1 = SHA-1)
  43       We infer the hash length (H) from this value.
  44
  45   1-byte number (C) of "chunks"
  46
  47   1-byte (reserved for later use)
  48      Current clients should ignore this value.
  49
  50 CHUNK LOOKUP:
  51
  52   (C + 1) * 12 bytes listing the table of contents for the chunks:
  53       First 4 bytes describe the chunk id. Value 0 is a terminating label.
  54       Other 8 bytes provide the byte-offset in current file for chunk to
  55       start. (Chunks are ordered contiguously in the file, so you can infer
  56       the length using the next chunk position if necessary.) Each chunk
  57       ID appears at most once.
  58
  59   The remaining data in the body is described one chunk at a time, and
  60   these chunks may be given in any order. Chunks are required unless
  61   otherwise specified.
  62
  63 CHUNK DATA:
  64
  65   OID Fanout (ID: {'O', 'I', 'D', 'F'}) (256 * 4 bytes)
  66       The ith entry, F[i], stores the number of OIDs with first
  67       byte at most i. Thus F[255] stores the total
  68       number of commits (N).
  69
  70   OID Lookup (ID: {'O', 'I', 'D', 'L'}) (N * H bytes)
  71       The OIDs for all commits in the graph, sorted in ascending order.
  72
  73   Commit Data (ID: {'C', 'D', 'A', 'T' }) (N * (H + 16) bytes)
  74     * The first H bytes are for the OID of the root tree.
  75     * The next 8 bytes are for the positions of the first two parents
  76       of the ith commit. Stores value 0x7000000 if no parent in that
  77       position. If there are more than two parents, the second value
  78       has its most-significant bit on and the other bits store an array
  79       position into the Extra Edge List chunk.
  80     * The next 8 bytes store the generation number of the commit and
  81       the commit time in seconds since EPOCH. The generation number
  82       uses the higher 30 bits of the first 4 bytes, while the commit
  83       time uses the 32 bits of the second 4 bytes, along with the lowest
  84       2 bits of the lowest byte, storing the 33rd and 34th bit of the
  85       commit time.
  86
  87   Extra Edge List (ID: {'E', 'D', 'G', 'E'}) [Optional]
  88       This list of 4-byte values store the second through nth parents for
  89       all octopus merges. The second parent value in the commit data stores
  90       an array position within this list along with the most-significant bit
  91       on. Starting at that array position, iterate through this list of commit
  92       positions for the parents until reaching a value with the most-significant
  93       bit on. The other bits correspond to the position of the last parent.
  94
  95 TRAILER:
  96
  97         H-byte HASH-checksum of all of the above.