This repository has been archived on 2023-06-11. You can view files and clone it, but cannot push or open issues or pull requests.
blog.lazkani.io-20200902-hi.../posts/revision-control/git-branching-and-merging.org

230 lines
7.5 KiB
Org Mode

#+BEGIN_COMMENT
.. title: Git! Branching and Merging
.. date: 2019-08-01
.. slug: git-branching-and-merging
.. updated: 2019-08-01
.. status: published
.. tags: git, revision-control
.. category: revision-control
.. authors: Elia el Lazkani
.. description: Explaining branches, branching and merging strategies.
.. type: text
#+END_COMMENT
In the previous post about /git/, we had a look at what /git/ is and got our feet wet with a bit of it.
In this post, I will be moving forward with the topic, I will be talking about branches, how to work with them and finally what merging is and how it works.
{{{TEASER_END}}}
* Requirements
The same requirement we had from the last post, obviously /git/.
* Branching and Merging
** What is a branch?
/git/ [[https://git-scm.com/book/en/v1/Git-Branching-What-a-Branch-Is][documentation]] describes it as:
#+BEGIN_QUOTE
"A branch in Git is simply a lightweight movable pointer to one of the[se] commits."
#+END_QUOTE
Usually, people coming from /svn/ think of *branches* differently. In /git/, a branch is simply a pointer to a commit.
So let's verify that claim to see if it's true.
Remember our example repository from the last post ? We'll be using it here.
First let's create a new branch.
#+BEGIN_EXAMPLE
$ git checkout -b mybranch
Switched to a new branch 'mybranch'
#+END_EXAMPLE
That was simple, wasn't it ?
Alright, let's test our hypothesis.
#+BEGIN_EXAMPLE
$ git log
commit 643a353370d74c26d7cbf5c80a0d73988a75e09e (HEAD -> mybranch, master)
Author: John Doe <johndoe@example.com>
Date: Thu Aug 1 19:50:45 2019 +0200
Second commit
#+END_EXAMPLE
The commit is, of course, different because this is a different computer with a different repository from scratch. Anyway, it seems from the log message that both /mybranch/ and /master/ are pointing to same commit /SHA/. Technically they are pointing to *HEAD*.
Now let's continue and add a new commit.
#+BEGIN_EXAMPLE
$ echo "" >> README.md
$ git add README.md
$ git commit -m "Adding an empty line"
[mybranch b30f4e0] Adding an empty line
1 file changed, 1 insertion(+)
#+END_EXAMPLE
After this last commit, let's check the log
#+BEGIN_EXAMPLE
$ git log
commit b30f4e0fa8f3b5c9f041c9ad1be982b2fed80851 (HEAD -> mybranch)
Author: John Doe <johndoe@example.com>
Date: Thu Aug 1 20:28:05 2019 +0200
Adding an empty line
commit 643a353370d74c26d7cbf5c80a0d73988a75e09e (master)
Author: John Doe <johndoe@example.com>
Date: Thu Aug 1 19:50:45 2019 +0200
Second commit
#+END_EXAMPLE
From reading the output of log, we can see that the /master/ branch points to a different commit than /mybranch/.
To visualize this, let's look at it in a different way.
#+BEGIN_EXAMPLE
$ git log --graph --oneline --all
* b30f4e0 (HEAD -> mybranch) Adding an empty line
* 643a353 (master) Second commit
#+END_EXAMPLE
What the above suggests is that our two branches have different contents at this stage. In other words, if I switch back to the /master/ branch what do you think we will find in =README.md= ?
#+BEGIN_EXAMPLE
$ git checkout master
Switched to branch 'master'
$ cat README.md
# Example
This is an example repository.
This repository is trying to give you a hands on experience with git to complement the post.
$
#+END_EXAMPLE
And if we switch back to /mybranch/.
#+BEGIN_EXAMPLE
$ git checkout mybranch
Switched to branch 'mybranch'
$ cat README.md
# Example
This is an example repository.
This repository is trying to give you a hands on experience with git to complement the post.
$
#+END_EXAMPLE
Let's add another commit to make easier to see the changes than an empty line.
#+BEGIN_EXAMPLE
$ echo "Let's add a line to mybranch." >> README.md
$ git add README.md
$ git commit -m "Adding more commits to mybranch"
[mybranch f25dd5d] Adding more commits to mybranch
1 file changed, 1 insertion(+)
#+END_EXAMPLE
Now let's check the tree again.
#+BEGIN_EXAMPLE
$ git log --graph --oneline --all
* f25dd5d (HEAD -> mybranch) Adding more commits to mybranch
* b30f4e0 Adding an empty line
* 643a353 (master) Second commit
#+END_EXAMPLE
Let's also check the difference between our /master/ branch and /mybranch/.
#+BEGIN_EXAMPLE
$ git diff master mybranch
diff --git a/README.md b/README.md
index b4734ad..f07e71e 100644
--- a/README.md
+++ b/README.md
@@ -2,3 +2,5 @@
This is an example repository.
This repository is trying to give you a hands on experience with git to complement the post.
+
+Let's add a line to mybranch.
#+END_EXAMPLE
The =+= suggests an addition and =-= suggests a deletion of a line. As we can see from the =+= shown before the two lines added to the =README.md= file, /mybranch/ has these additions.
You can read more about /git/ branches in the /git/ [[https://git-scm.com/book/en/v1/Git-Branching-What-a-Branch-Is][documentation]] page.
** What is merging ?
That's all fine so far, but how do I get these changes from /mybranch/ to the /master/ branch ?
The answer to that is also as easy as all the steps taken so far. /git/ merges *from* a branch you specify *to* the branch you are currently on.
#+BEGIN_EXAMPLE
$ # Checking which branch we are on
$ git branch
master
* mybranch
$ # We are on mybranch and we need to put these changes into master
$ # First we need to move to our master branch
$ git checkout master
Switched to branch 'master'
$ # Now we can merge from mybranch
$ git merge mybranch
Updating 643a353..f25dd5d
Fast-forward
README.md | 2 ++
1 file changed, 2 insertions(+)
#+END_EXAMPLE
As we can see. The changes in /mybranch/ have been merged into the /master/ branch.
#+BEGIN_EXAMPLE
$ git log
commit f25dd5da3e6f91d117177782a5811d5086f66799 (HEAD -> master, mybranch)
Author: John Doe <johndoe@example.com>
Date: Thu Aug 1 20:43:57 2019 +0200
Adding more commits to mybranch
commit b30f4e0fa8f3b5c9f041c9ad1be982b2fed80851
Author: John Doe <johndoe@example.com>
Date: Thu Aug 1 20:28:05 2019 +0200
Adding an empty line
commit 643a353370d74c26d7cbf5c80a0d73988a75e09e
Author: John Doe <johndoe@example.com>
Date: Thu Aug 1 19:50:45 2019 +0200
Second commit
#+END_EXAMPLE
* Merging Strategies
I'll explain to you how I like to work and my personal merging strategy. I will keep out some details as they use concepts that are more advanced than what has been discussed so far.
** /master/ branch
To me, the /master/ branch stays always up to date with the *remote* /master/ branch. In other words, I do not make commits against the /master/ branch in the project I'm working on.
** branch
If I want to work on the project, I start by updating the /master/ branch and then branching it as we've seen before. The name of the branch is always indicative on what it holds, or what kind of work I am doing on it.
As long as I am working on my dev branch, I keep updating the /master/ branch and then porting the changes into my dev branch. This way, at the end the code is compatible and I am testing with the latest version of the code. This is very helpful and makes merging later a breeze.
** merging
After my work is done, I push my branch to the remote server and ask for the maintainer of the project to merge my changes into the /master/ branch after reviewing it, of course. To explain this in a very simple manner, all that mumbo jumpo talk previously simply means someone else did the merge into master.
* Conclusion
In this post, I talked about what are branches. We went ahead and worked a little bit with branches and then mentioned merging. At the end of the post I talked a bit about my merging strategy.
In the next post, I will be talking about remotes.