Lesson 02 · The loose object format

The loose object format

Before an object can be hashed or stored, Git wraps its contents in a tiny header that records the object's type and size. Getting these bytes exactly right is what makes our hashes match real Git, so today you build that serialization.

The goal

Serialize a piece of content into Git's loose object format, a header followed by the content.

Start here - the target

TO DO

Scenario: Content is wrapped in a type-and-size header

Giventhe content is the 6 bytes hello followed by a newline

Whenit is serialized as a blob with Serialize(blob, content)

Thenthe result is the bytes of the word blob, a space, the digit 6, a NUL (zero) byte, then the content

Andin hex that is 62 6c 6f 62 20 36 00 68 65 6c 6c 6f 0a, and serializing empty content as a blob gives blob, space, 0, NUL

Background

Every object Git stores, whatever its type, is serialized the same way: a short header of "<type> <size>" then a single NUL byte (a zero byte) then the raw content. The type is a word like blob, tree, or commit; the size is the length of the content in bytes as decimal ASCII digits. For our file hello.txt holding hello\n (5 letters plus a newline, so 6 bytes) the header is blob 6 then a zero byte.

This wrapper is the whole reason our hashes will match real Git: the hash is taken over these exact bytes, header and all, not over the file content alone. The size and the NUL separator are what let a reader split the header back off later. Keep this a pure function that takes a type and content and returns the serialized bytes; nothing is written to disk yet.

Make it work

// header is "<type> <size>\0", then the raw content
func Serialize(typ string, content []byte) []byte {
  header := fmt.Sprintf("%s %d\x00", typ, len(content))
  return append([]byte(header), content...)
}

CheckpointDONE

You can wrap any content in Git's loose object header. Commit and stop here.