hash-object command

The hash-object command is used to compute the ID value "aka. name" of an object, and sometimes we use it to write the object, although it is not a frequently used command and it is considered as a low-level command still it is for sure an important one and in our implementation version, we'll use this function to do hashing functionality in upcoming commands.

What are objects?

Object:

• The unit of storage in Git. It is uniquely identified by the SHA-1 hash of its contents. Consequently, an object cannot be changed.

• Every object has an:

  • Object name (id)
    • The unique identifier of an object. The object name is usually represented by a 40-character hexadecimal string. Also colloquially called SHA-1.
    • That means object names are not just regular file names, it is a unique name calculated depending on the content of the file. This makes sure that all the files are unique from each other.
    • We conclude that we do not modify files in Git even if you change just one character in a file the name of the file changes too and is stored in a different location "Don't get overridden".
  • Object type
    • One of the identifiers "commit", "tree", "tag" or "blob" describes the type of an object.

Where objects are being stored?

The path where git stores its objects is created by calculating the SHA-1 hash of its contents. Git splits the hash into two parts: the first two characters of the hash, and the rest. Then it uses the first part as the directory name, and the rest as the file name.

example

./.mygit/objects/d4/5a25929cd70b0264caea33b209c9cfa845f6bc

How objects are stored?

First, we need to understand the storage format. An object has a header that specifies its type ("commit", "tree", "tag" or "blob"), the size of object contents in bytes, null byte, then the object compressed content, Git uses Zlib for compression.

example

blob <size>\0<content>

For example, if the file content is this is file the blob object file would look like this after decompression.

example

blob 12\0this is file

Object types:

1. Blob:

   • Stores the content of a file.
   • Contains no metadata like filenames or permissions.
   • Identified by a unique SHA-1 hash.

2. Tree:

   • Represents a directory and its structure.
   • Points to blobs (files) and other trees (subdirectories).
   • Stores filenames and permissions.

3. Commit:

   • Represents a snapshot of the repository.
   • Points to a tree object and parent commits.
   • Includes metadata (author, timestamp, message).

4. Tag:

   • Marks a specific commit, often for releases.
   • Can be lightweight (simple pointer) or annotated (with metadata).

For now, that's all the knowledge we need to start implementing our function.

Command Implementation

First, we define function command_hash_object($args), the $args is an array of arguments that get passed to our function from the user_call_func that we used earlier to read user input.

1. Parse arguments:

./index.php

$filePath = $args[0];
$writeEnable = $args[1] ?? false;

• Extract file path from arguments.
• Check if the second argument is provided to decide if the object should be written to the repo or not.

2. Check if the file exists:

./index.php

if (!file_exists($filePath)) {
  echo "Error: File not found.\n";
  return;
}

• Check if the file path provided exists and if not we display an error message and return.

3. Read the file:

./index.php

$fileContents = file_get_contents($filePath);

• Read the file contents and store them in $fileContents.

4. Create the object header:

./index.php

$header = "blob " . strlen($fileContents) . "\0";

• Create the object header by concatenating the following:
  • blob: which is the object type.
  • strlen($fileContents): which is the length of the contents inside the file.
  • \0: which is the null byte (used to separate the header from the file contents).

5. Concatenate header and contents:

./index.php

$fullContent = $header . $fileContents;

• concatenate the header with the file contents to get the final storage format
  which is blob <size>\0<content>.

6. Compute the SHA-1 hash:

./index.php

$hash = sha1($fullContent);

• calculate the SHA-1 hash of the object by using sha1() php built-in function.

7. Write object to repo if writeEnable is true:

./index.php

if ($writeEnable) {
  $compressed = gzcompress($fullContent);

  $dir = '.mygit/objects/' . substr($hash, 0, 2);
  if (!is_dir($dir)) {
    mkdir($dir, 0777, true);
  }

  $filePath = $dir . '/' . substr($hash, 2);
  file_put_contents($filePath, $compressed);
}

• If the $writeEnable value was true:
  • We compress the file content using gzcompress() built-in function which is the compress function for Zlib.
  • We create our directory path $dir by concatenating '.mygit/objects/' with the first two characters in $hash.
  • If the path $dir didn't exist we make a new directory using mkdir() built-in function.
  • We create our file path $filePath by concatenating $dir path with / and the rest of the file hash. Then we save the compressed content into our file.

8. Output and return the hash:

./index.php

echo $hash . PHP_EOL;
return $hash;

• At the end, we output the calculated hash and return it (because we will use this function later in upcoming commands).

run example

php ./index.php hash_object test.txt true #if we want it to write

output example

d0ab265094c565b783e6880d73ff0dbf315d8812

Full code for this section:

index.php

function command_hash_object($args)
{
  $filePath = $args[0];
  $writeEnable = $args[1] ?? false;

  if (!file_exists($filePath)) {
    echo "Error: File not found.\n";
    return;
  }

  $fileContents = file_get_contents($filePath);

  $header = "blob " . strlen($fileContents) . "\0";

  $fullContent = $header . $fileContents;

  $hash = sha1($fullContent);

  if ($writeEnable) {
    $compressed = gzcompress($fullContent);

    $dir = '.mygit/objects/' . substr($hash, 0, 2);
    if (!is_dir($dir)) {
      mkdir($dir, 0777, true);
    }

    $filePath = $dir . '/' . substr($hash, 2);
    file_put_contents($filePath, $compressed);
  }
  echo $hash . PHP_EOL;
  return $hash;
}