Lesson 03 · The object id is the SHA-1 of the object

The object id is the SHA-1 of the object

An object's identity in Git is the SHA-1 hash of its serialized bytes, written as 40 lowercase hex characters. This is the content-addressable idea at the heart of Git, and today your ids start matching real Git exactly.

The goal

Compute an object id as the SHA-1 of its loose object bytes.

Start here - the target

TO DO

Scenario: The id is the SHA-1 of the serialized object

Giventhe serialized loose object bytes for some content

WhenHashObject(blob, content) takes the SHA-1 of those bytes as 40 lowercase hex characters

ThenHashObject(blob, hello and a newline) is ce013625030ba8dba906f756967f9e9ca394464a

AndHashObject(blob, empty content) is e69de29bb2d1d6434b8b29ae775ad8c2e48c5391, the well-known empty blob id

Background

This is the moment the whole design clicks: an object’s id is its content, run through SHA-1. Hash the serialized bytes (header, NUL, and content together) and render the 20-byte digest as 40 lowercase hex characters. Two files with identical content get the same id automatically, because the id is a pure function of the bytes. That is what “content-addressable” means.

The values are not ours to choose; they are whatever SHA-1 produces, and they are identical to real Git because we hash the exact same wrapped bytes Git does. The empty blob e69de29b... is a constant you will see in real repositories, and hello\n always hashes to ce013625.... We lean on the standard library for SHA-1 here - earlier projects built hashing from scratch; this one is about the object model, not the hash function.

Make it work

// hash the serialized bytes, not the raw content
func HashObject(typ string, content []byte) string {
  sum := sha1.Sum(Serialize(typ, content))
  return hex.EncodeToString(sum[:])
}