GIT AND GITHUB

GIT

 Git is a distributed version control system (collaborated with people and work on the same project)
 Can track changes in files in any folder (repository)
 Git preserves history of all changes in the files
 Git could be used locally without internet and locally you will have full access to the files change
history
 Git stores all changes locally and does not rely on internet connection

HOW DOES IT DO THAT?

 Git has its own file structure (special file system)

 Every file is stored in the separate document and has unique hash
 Git is simply a persistent hash map that stores key-value pairs
 Key is a hash of the file and value is content of the file

GITHUB

 Github is a repository hosting service

 Github is a backup for local repositories
 Github is used primarily by distributed teams that work on the same projects
 Capable of website hosting and desktop use

SHELL

 Mkdir – make directory

 Ls – list files and directories
 Cd – change directory
 . – alias of the current directory
 .. – alias of the parent directory
 nano – edit file
 clear – clear terminal
 tab – autocomplete command
 echo – print to terminal
 man – help on specific command
 pwd – print working directory
 touch – create new files
 > - write to the file
 >> - append to the file
 Cat – list contents of the file
  Rm – remove files and directories
 Rm -rf - deletes directories

HOW GIT WORKS?

 Git init – Initialises a new git repository (with default branch as master)
 Git repository will be initially empty even if directory where you initialize git is not empty and
contains files and other directories
 Never manually change .git folder, let git manage itself

GIT OBJECT TYPES

 Blob – stores any files with extensions (.txt, .mp3, .mp4, etc)
o Have no file names
o Size and type is stored in each blob
 Tree – stores information about directories / folders
 Commit – storing different versions of the project
 Annotated tag – persistent text pointer to specific commits

GIT LOW LEVEL COMMANDS

 Git hash – object – able to create new object and write it in git repository
o Hash = folder name + file name
o For e.g. echo “HELLO” | Git hash-object –stdin -w
 Git cat -file – read information about git object
 Git mktree – make tree
JSON VS GIT

 Javascript Object Notation

o Format for data exchange between different servers (retrieve data from API from server to
web browser or send data to database to another server etc)
 Within JSON, keys should be unique, but it is possible to create the same keys but not
recommended
 However, keys in git is equal to hash of each object
o Key in git is generated based on value
o Git stores files based on their hash

HASH

 Takes any variable length input and creates a fixed length hash
 Hash is generated based on the contents of the file
 Hash are one-way functions (create based on input to create hash)
 Same input will generate same hash
 Hash types
o MD5 (128 bit)
o SHA1 (160 bits)
o SHA256 (256 bits)
o SHA384 (384 bits)
o SHA512 (512 bits)
 Git uses SHA1 hash function (which is hexadecimal format)

SHA1 HASH FUNCTION

 160 bits means (git can store 2^160 files in the same repository)
o Therefore the probability of a hash collision is 1/ (2^320)
 40 hexadecimal (base 16  0 – 9 and A - F ) characters
 Slight change to input data can lead to completely different hashes
 Commands such as shasum can be used in git bash to generate a hash
 Each time you save it creates a unique ID which is the SHA1 Hash

GIT CAT-FILE OPTIONS

 Git cat-file -p <hash (can enter first 4 characters atleast) > - contents of the object
 Git cat-file -s <hash> - size of the object (in bytes)
 Git cat-file -t <hash> - type of the object
 Git cat-file command doesn’t have options for retrieving the filename from the blob because blobs
don’t store filenames
GIT HASH-OBJECT

 Echo “hello, Git” | git hash-object --stdin -w (stdin  standard input and -w  create git object)
o Git hash-object <filename> -w
 Git repository stores files independently in its own file system in the objects directory
 Git generates SHA1 hash based on input+type+size

GIT OBJECT

 Do not store file names of the source file

 Contains contents, type, size and delimiters (\0)
 Name of each blob is based on a SHA1 hash of each file

TREES

 Representation of repositories
 Structure is the same of a blob (type, length and hash)
 Contents include permissions, type (tree), SHA1 hash and file name
PERMISSIONS

 040000 – directory (common)

 100644 – regular non-executable file (common)
 100664 – Regular non-executable group-writeable file
 100755 – Regular executable file
 120000 – Symbolic link
 160000 – Gitlink
 Find files
o Find .git/objects -type f

STEPS FOR MAKING TREE

 Nano temp-tree.txt
 <permission type> (space) <object type> (space) <sha1 hash of the object> (tab) <filename>

WORKING DIRECTORY, STAGING AREA AND GIT REPO

 Working directory – local directory

 Staging area (index) – responsible for repairing files to be inserted into git repository or prepare
files taken from git repository to be put into working directory
o Git ls-files -s : lists files in the staging area
o Git read-tree <hash> (gets files from git repo and puts in staging area)
o Git ls-files (lists files in git in staging areas)
o Git checkout-index -a (puts from staging area into working directory)
 Git repository – storage of files and folders using SHA1 hash

HOW MANY FOLDERS CAN GIT CREATE?

 16 (due to 16 different hexadecimal characters) * 16 (due to the folders having 2 characters) which
means 16^2 = 256 in git objects folder
COMMITS

 Commit contains same structure as blob and tree (object type, length and delimiter)
 Commit has information such as author name, email, commit description and parent
 Each commit has its own SHA1 hash (due to author name and email)
 Commits allow us to store in git database for different versions of the project
 Able to very easily and fast move to any version of the project by checking out specific commit
 Commit is a wrapper for the tree object and contains pointer to a specific tree

 By moving to different commits (checking out – taking files out of git repository and putting it into
your directory) you are able to “travel” between different “versions” of the project

SET GIT AUTHOR NAME AND EMAIL

 Git config - -global user.name <Name>

 Git config - -global user.email <Email>
 Git config - -list

CREATING COMMIT

 Git commit -m “HELLO”

o -m : Allows you to add a description
GIT BASIC COMMANDS

 Git status – current state of the git repository

 Git add – add files to staging area
o Git add . – stage all files
o Git add <filename> - stage specific file
 Git commit – write changes to git repository (creates new blob entries such as files in git repository)
 Git log – history of changes (commits)
 Git checkout – checkout commit or branch

GIT FILES LIFECYCLE

UNSTAGE FILES USING GIT RM

 Git rm –cached file3.txt

o Unstages files but does not remove them from the repository

PARENT COMMIT

 Check commit contents

o Git cat-file -p 44e073
 Parent is a pointer to the root commit (initial commit)

GIT CHECKOUT

 Git checkout – jump into certain version of the project

WHAT IS A BRANCH

 Branch is just a text reference to the commit (separate history of your project)
 Default branch is master
 Multiple branches can exist in the same repository
 Pointers for all branches are located in .git/refs/heads folder
 Current branch tracks new commits and creates new commit objects
o For example, you are in master branch and made several commits and checkout to a new-
branch and have made couple of commits. However, when you checkout to master branch,
you will not see commits from new-branch as the current branch tracks new commits and
those commits are not added to other branches but there are other ways to marriage
branches or rebase them.
 Branch pointer moves automatically after every new commit
 Change branch git checkout <branch>

WHAT IS HEAD IN GIT ?

 HEAD is reference to the currently checked-out branch or commit

 HEAD is locally significant
o When head points to specific commit instead of branch it is called detached HEAD
 Pointer is located in the cat .git/HEAD file
 Default pointer is ref: refs/heads/master
 Change reference to specific branch git checkout <branch>
 Change reference to specific commit git checkout <sha1>

GIT BRANCHES MANAGEMENT

 Git branch – list all local branches

 Git branch <name> - create new branch
o New branch will have contents from the current commit of the branch
 Git checkout <name> - checkout specific branch
 Git checkout -b <branch name> - shortcut for creating a branch with checkout
 Git branch -d <name> - delete specific branch (will delete only merged branch but if you want to
delete non-merged branch use “-D” option)
 Git branch -m <old> <new> - rename specific branch
GIT DIFF COMMAND

 Tracks different changes that are made between files.