This lesson is under construction. Learn from it at your own risk. If you have any feedback, please fill out our General Feedback Survey.
VCS is how one tracks changes, modifications, and updates to source files over time. Creating a history of changes for a project over time.
- Used for:
- Other Names Include:
- Source Control Management (SCM)
- Version Control Software
- Revision Control Software
- Version control solves a lot of problems:
- I have changes I want to integrate (merge) into the main project.
- I want to track the state of this project over time.
- I want to make some changes without possibly breaking what I have.
- ... and much more.
Most VCS software does this by storing metadata about a project, specifically differences (diffs) between between points in history.
VCS is a broad and complicated topic. It might not come easy to you, but experience will help.
Below is a list of concepts / vocabulary you’ll encounter when working with VCS. Some systems use slightly different verbiage to describe these ideas but you can get by with these words.
“I created a new repository (repo) for my school project so we can collaborate more easily.”
The delta (additions and deletions) between two states of a project.
“The diff between draft one and two was very long thanks to the help of the skilled editor.”
A snapshot of your project’s state at a point in history. Records the difference between two points in history with a diff.
“Your last commit modified two methods and introduced a bug somewhere.”
Modifications to a project (main branch = trunk) made in parallel with the a main branch, but not affecting the main branch.
“My branch introduces some changes which might break production so I’m not going to merge it until it’s well tested.”
Introducing changes from one branch into another.
“I merged three commits into the Master branch so we can have those features in the next release.”
- Downloading a local copy of a project.
- Your own version of somebody else’s project where you take the original code-base and make modifications. May include many changes or just a few bug-fixes. Sometimes you end up merging those changes back upstream.
There are two main types of VCS: distributed and centralized. They each have their pros and cons.
- Centralized VCS
“I’m going to work at the designated work-bench. Nobody else can work on this until I’m done.”
Centralized Version Control Systems require a persistent connection with a centralized server.
- Examples include:
- Distributed VCS
“I’m going to work over here for a while and tell you about what I did later.”
Distributed version control systems are those in which each person working on a project downloads a local copy of the project, makes their changes, and manually fetches changes from other copies.
- Examples include:
Git is a Free and Open Source distributed version control system designed to handle everything from small to very large projects with speed and efficiency. ( https://git-scm.com )
Git has become the de facto VCS tools used by new Open Source projects. It is supported on platforms including Github, Bitbucket, and Gitlab and is used by projects like the Linux Kernel, the Go language, and of course Git itself.
It is not the only VCS out there. Your future job may require you to use a different tool for VCS. Thankfully the principles of SCM are very similar everywhere you go, so pay attention! There’s a lot to learn here and it’s not all about the commands you should type.
Before using Git we need to make sure we have it configured. We will do this by telling Git what our name, email, and preferred text editor is.
$ git config --global user.name "My Name" $ git config --global user.email "firstname.lastname@example.org" $ git config --global core.editor "nano"
First we will create a repo. In Git a repo is just a directory that we have run git init inside of.
Create a project with Git:
$ mkdir my-project $ cd my-project # Always run `git init` inside of a project folder! $ git init # Never inside of your home directory.
Next we are going to add a file to our project. This is as easy as creating it, staging it (git add <filename>) and committing it.
Add and commit a file to your project with Git:
$ touch newfile.txt $ git add newfile.txt $ git commit # Edit message in Nano, save the file, exit to commit.
To see which files are staged, unstaged, or untracked:
$ git status
To look through your repository history:
$ git log
To create and checkout a branch:
#Note the `*` which indicates the current branch $ git checkout -b "new-branch" $ git branch master * new-branch
Checkout a new feature branch on your repository.
$ git checkout -b "add-awesome-feature"
Create/Edit files on the new branch.
$ echo "Some awesome text" > awesomefile.txt $ git status # On branch add-awesome-feature # Untracked files: ... # awesomefile.txt ... $ git add awesomefile.txt $ git commit -m "Short awesome commit message"
View the diff between the two.
$ git diff master diff --git a/awesomefile.txt b/awesomefile.txt new file mode 100644 index 0000000..08cec7f --- /dev/null +++ b/awesomefile.txt @@ -0,0 +1 @@ +Some awesome text
Locally merge the changes from your new branch into Master.
$ git checkout master $ git merge add-awesome-feature Updating 459de26..5c4ca48 Fast-forward awesomefile.txt | 1 + 1 file changed, 1 insertion(+) create mode 100644 awesomefile.txt
- Do not delete the .git directory.
- The .git directory is the file which contains your entire Git repository. All history and metadata about your repo is kept in that folder, deleting it deletes your project except for the code you have right now.
- Do not change history.
You will eventually learn some commands to modify history (change commit messages or changes in a commit). You can do this but you should only do it with caution and only when you fully understand what you’re doing. It has the potential to ruin other people’s day.
Well – don’t change history with the exception of the next two points...
- Do not make redundant commits or revert commits.
- After making a commit, use git commit --amend if you want to make changes to avoid cluttering your history with Oops, changing that messages.
- Don’t wait too long between commits.
- Commits are cheap and you can never make too many. If you wait too long you might as well not use it! Once you’re done with a feature you can squash many commits into one.
- Avoid redundant copies of the same work.
- If you want to make changes to something, but also want to keep the original, make a branch! There are very few cases where making a copy of something in the same repository is the best thing to do.
- Don’t commit secrets. They never go away.
- Ars Technica wrote about the topic, check it out: http://arstechnica.com/security/2013/01/psa-dont-upload-your-important-passwords-to-github/
Everybody uses VCS differently. Choose the workflow that works best for everybody involved.
Unfortunately we can’t really prescribe a best Git workflow. There are definitely better workflows out there, but you should do what works for you, and when you join a project explicitly ask what their workflow is so you can stay on the same page as your peers.
While Git is distributed it can be made more centralized by pushing and pulling with a common remote repository. Tools for centralizing Git include:
- Open Source, free to run, feature rich.
- Very popular. Not Open Source but free for Open Source projects.
- Also popular, similar to Github, unlimited free private and public repositories.
- Bare-bones. Fewer features than the previous three. Open Source, useful for learning the nitty-gritty on how Git really works.
Each of these services is either free to use, or Open Source (so you can setup and run it yourself).
To contribute to someone else’s repository you first need to clone the repo.
$ cd /path/to/my/projects $ git clone <some git url> $ cd <new repo directory> $ ls
Once you clone a repository you can make as many local changes as you want without affecting the original (central) copy. You can experiment and work without the original owner even knowing what you’re doing!
In this exercise we will clone an application and run it in our remote GNU/Linux box.
First, clone the tinsy-flask-app repository from the DevOpsBootcamp organization on Github.
$ cd ~ $ git clone https://github.com/DevOpsBootcamp/tinsy-flask-app.git $ cd tinsy-flask-app
See http://git.io/vcVmB for more details about the tinsy-flask-app repository.
After cloning we need to setup the project. This is similar to building an application from source because we downloaded the source code and want to run it!
Most repositories include a file called README which includes instructions on how to setup and run the code included in the repo. Those instructions tell us to run the following commands.
If you are using our docker image, you won’t be able to access the webserver without adding a port forwarding rule for TCP port 8080 on the container, similar to how SSH is forwarded on the container. See ‘Setting up Docker’ on the left for info on how to add port mappings to a docker container.
#Setup python virtual environment $ virtualenv venv $ source venv/bin/activate (venv) $ pip install -r requirements.txt #Run server (venv) $ python script.py #When finished, deactivate virtual environment (venv) $ deactivate $
You may need to install the python, virtualenv, and pip packages. How would you do that?
Now if you go to <your ip address>:<http port> in your web-browser to see a live version of the app!
- The Online Git Docs
- This is a portal to all of the official docs on git-scm.com. It includes everything from Getting Started to Git Internals. Check it out!
- Git workflow tutorial
- This is the tutorial provided on https://git-scm.com/about/distributed. It is a good high-level overview of some common git workflows.
- A successful Git branching model
- This blogpost describes a git workflow (git-flow) that the Open Source Lab bases their workflow on.