toderesa97/gitrepository

Practising with git

Practising with git

by @toderesa97

Before starting ...

What is git?

Git is a helpful tool to keep track of our files in a project. Instead of copying files to test if a new feature is going to work, we can branch, and that means creating our 'developing' path, so we don't push to the master branch without being sure what we've done. It's supposed that the master branch contains the last version (and obviously it must work) of our software.

This is a repository which will be used to practice the git methodology. I'll be using IntelliJ and the VCS (Version control Software), mostly known as GIT. Also I'll be using GitBash (provides a BASH emulation used to run Git from the command line)

Commands

$ git status      # Retrieve the current status (if a file has been modified, or nothing to commit...)
$ git log -n      # Retrieve the last n transactions(commits) we've done. Argument n is not necessary you can just
type git log.

Be careful with this last command. If you have done lots of 'transactions' and you execute this command in the terminal (especially on GitBash - for Windows), it's going to be a mess. Instead redirect the output to a file using the redirection operator, as follows...

$ git log > log_file.txt

If this command is run in an OS based on UNIX you should enter to VIM (VI IMproved) terminal text editor. To exit type :q or :wq if you entered on VI.

Continuing with the commands...

$ git add <files>           # Adding files to be committed
$ git commit <file(s)>      # Commit the added files, $ git commit -m "message" , to commit the added file(s) with a message.

You will be redirected to VI editor in order to add a message of the selected file(s). That's a good practice because it allows you to add specific information depending of the file. If you want to commit everything just type $ git commit -am "committing everything with this message"

$ git push        # Push to the repository

If you executed the last command you would get an error because your computer does not know who you are. In order to push (send to your remote repository) there is an useful command.

$ git remote -v, and the output should be something similar to:

origin  https://github.com/toderesa97/gitRepository.git (fetch)

origin  https://github.com/toderesa97/gitRepository.git (push)

The important thing here is the first column, which represents the name of your remote.

If you don't have any remote, you must create a repository and execute the following

$ echo "# <repo_name>" >> README.md    # concatenates the message surrounded by double quotes into 
                                       a new file that is going to be created implicitly (Expendable)
$ git init                             # initializes a git repository in the actual path
$ git add README.md                    # must be added if it was created to commit it.
$ git commit -m "first commit messg"
$ git remote add <remote_name> <reposory_url>.git
$ git push -u origin master

Another important thing we must take into account is the branches. You can execute $ git branch to see your branches (if it's a new repository you'll have just the master branch lead by an asterisk)

Note: asterisk (*) indicates you in what branch are you working on.

Now to push, execute:

$ git push <remote> <branch> in our case $ git push origin master. You will be asked for your github credentials.

$ git branch      # See the actual branches
$ git branch <n>  # Creates a branch with name n from the actual branch where the command is executed (creates a copy)

When working with branches the way we push a project to the repository is $ git push origin <branch_name>. Obviously what is being pushed to a branch will not appear in another unless you merge. For example, imagine you are requested to add a new functionality in the software, first thing you should do is to create your branch (your 'develop path') in order to avoid adding modifications to the master branch without testing that the functionality and what is in the master branch are not acting up. If everything is perfect you can execute:

$ git branch                    # see actual branches
$ git checkout master           # To select among different branches
$ git pull origin master
$ git merge  <branch>  
$ git push origin master

$ git clone <url>               # Clones a repository from _url_ into the directory where is being executed (grab from server the repository and *paste* it into your computer)
$ git pull origin <branch>      # 'Updates' your workspace with the content in the repository. 

NOTE please that origin is the default remote name, however it could be changed when initializing the repo (explained above)

Also it's very helpful to add a file called .gitignore. This file contains all the extensions we don't want to keep track of. Examples of these files could be: the .xml, .iml, .class ... files that are generated by your IDE. So instead of selecting the files you want to commit, which will make you waste time, you can add this file.

This file should be something similar to:

# Created by .ignore support plugin (hsz.mobi)
### Java template
# Compiled class file
*.class

# Log file
*.log

# BlueJ files
*.ctxt

# Mobile Tools for Java (J2ME)
.mtj.tmp/

# Package Files #
*.jar
*.war
*.ear
*.zip
*.tar.gz
*.rar

# virtual machine crash logs, see http://www.java.com/en/download/help/error_hotspot.xml
hs_err_pid*

# Extensions added by me
*.xml
*.iml

Note: *.extension is a regular expression that matches all files regardless their names with extension .extension.

By @toderesa97

Any correction and/or suggestion will be helpful.