Git

Your website looks good on your computer? Good, now you want to deploy it (= get it available online)! But first, you need to copy your code to your git server (here: Gitlab).

git is a powerful software with numerous commands and options but, I can safely assume you are the only developper of your website, so you just have to know 5 git commands. There are numerous tutorials for Git and cheat sheet. Here we just see the basic command that you should / will have to use.

Basic commands

Command Short description
git clone source destination Copy a whole repository source into destination. This is done once.
git add file 2 usages: first, specifies git to track file. Second, add that file to the next commit.
git commit -m "message of the commit" Do a commit, that is like a screenshot or a save, of every added files.
git push Send repository (and last commits) from the computer to the remote server (Gitlab). See next section before using it!
git pull Receive the repository (and missing commits) from the remote server (Gitlab) to the computer. You might use this command if you work on different computer and/or multiple developpers on same project (kind of synchronisation)

More “advanced” (but not so much)

The following commands are also useful but not in your “every day life”:

Command Short description
git init Initialize a git repository. This must be used if the folder do not comes from a git clone for example.
git status Display the status of the repository: files that has been changed but not commited, files ready to be commited, …
git log Display the history of the last commits.
git checkout file Cancel the modification of the file and reset it to the last commit. Useful if you are unhappy with your changes and want to come back
git remote -v Display the remote server of your git repository.

Reset to commit

Huge warning: you will lose everything that has not be committed! This command is destructive!

If you completely messed up your working directory and want to come back to the last commit:

git reset --hard HEAD

or if it’s not working:

  1. First get the SHA of the last commit (identifier) (copy/past on shell is possible using shift + ctrl + c)

    git log
    > commit 69c453f2d0314af860ee790809ed98861d27ff9b (HEAD -> master)
    
  2. Revert (either the full SHA or the first 6 characters)

git reset --hard 69c453

`.gitignore``

You may already have a .gitignore file at the root of your folder. This file tells git which file it should not tracked. This is very useful as many file should not be tracked and stored by git. You probably do not have to modify it, though.

(Very) short example

The git commit command has an option -a that forces to commit every changes of every tracked file:

git commit -m "message of commit" -a

This is quite usefull as, by doing so, you do not need to git add every files you want to commit (-a stands for -all).

Let says that I changed the content/home/about.md file. Doing

git status
> Changes not staged for commit:
> 	modified:   content/home/about.md (modified content)

I know add the changes to the next commit and commit the changes

git add content/home/about.md
git commit -m "Update the about page"
Précédent
Suivant