GIT 3 Branching Strategy | Github and GitLab Branching Strategy | Must know for all engineers

Summary

Highlights

• Understand the purpose of each branch: develop, release, and master – streamline your workflow 🧠.
• Handle hotfixes efficiently on the master branch, without impacting ongoing developments 🔥.
• Prevent regression test delays by separating hotfixes and ongoing updates, keeping projects on track 🕒.
• Ensure all branches are up-to-date with back merges to avoid losing critical fixes 🔗.
• Ideal for MNCs thanks to its organized structure, though less frequent in startups 🏢.

Key Takeaways

• Three major branches in Git: develop, release, and master – each serving a distinct purpose in code management 🚀.
• Hotfixes are handled smoothly with this strategy, ensuring live production isn't delayed or disrupted 🛠️.
• Efficient regression testing keeps the code quality intact without derailing the development flow ✔️.
• Back merges are key to synchronizing updates across branches, preventing version conflicts 🔄.
• This strategy prevents disruptions and maintains continuous integration and deployment, ideal for large-scale environments 🌐.

Overview

Chapters

• 00:00 - 00:30: Introduction to Git 3 Branching Strategy The chapter discusses the 'Git 3 Branching Strategy,' which is commonly used in multinational corporations (MNCs). The strategy may not be prevalent in startups, but it is quite popular in MNCs. The narrator, Shreyansh, explains that this strategy involves having three main branches, typically observed in platforms like GitHub or GitLab, where one of the branches is 'develop.'
• 00:30 - 01:30: Branch Structure and Initial Setup This chapter discusses the initial setup and branch structure in a version-controlled environment. It mentions different branches such as release and master, and describes a scenario where the current code version is V1. In this setup, the V1 version is running in production, and engineers are beginning to raise pull requests for further code developments.
• 01:30 - 02:30: Release Process and Changes The chapter discusses the process involved in transitioning changes from the development phase to the release phase. It starts by merging changes onto the developer's segment. Upon initiating the release process, the primary task is to transfer changes from the development to the release phase. The details of the procedures and modifications involved in this transition are highlighted.
• 02:30 - 03:30: Regression and Deployment The chapter 'Regression and Deployment' discusses the process of preparing and deploying new versions of software. It begins with the scenario where the development branch, called 'develop', has several changes merged, resulting in version V2. Meanwhile, the 'release' and 'master' branches, as well as the production environment, are still on version V1. The chapter emphasizes that to initiate the release process, the first step is to synchronize changes from the 'develop' branch to the 'release' branch, ensuring the latest updates are included for deployment.
• 03:30 - 05:30: Importance of Master Branch and Hotfix Handling The chapter discusses the process of updating the release version by integrating changes from the develop branch. This transition marks the release version as V2, indicating that all developments are now included in the release branch. Additionally, the chapter highlights the importance of performing regression testing on the release branch. This involves running functional test cases to ensure stability and functionality before finalizing the release.
• 05:30 - 07:30: Managing Versions and Back Merging The chapter titled 'Managing Versions and Back Merging' discusses the process of handling numerous functional test cases, especially in large MNCs. It highlights the importance of running regression tests on build versions (specifically V2) to ensure that no existing functionalities are broken. The process aims to validate stability before the final step of pushing the updates to the live environment.
• 07:30 - 09:00: Conclusion on Git 3 Branching Strategy The chapter discusses the process of pushing updates to a live environment in the context of a Git 3 branching strategy. Initially, the production (live) version is at V1 and the development version is at V2. After completing regression testing on V2, the next step is to deploy this version to the live environment. The chapter concludes by mentioning the subsequent step, which involves switching the live build to V2 and monitoring the changes.

GIT 3 Branching Strategy | Github and GitLab Branching Strategy | Must know for all engineers Transcription

• 00:00 - 00:30 hey guys welcome to concept encoding this is shreyansh and today I'm going to tell you about the get three branching strategy this is mostly used in the MNC so in mncs mostly you will find this three branching strategy only uh in startup I doubt that they might be using it but at least you will see in it is very popular in the MNC at least right so let's say what it is uh three branching strategies so in the if you are using GitHub or gitlab you will see that generally there will have three branches one is develop
• 00:30 - 01:00 second one is release third one is master now let's say the version of the code currently it has V1 V1 V1 and let's say in the production the build which is generated out of this V1 code is currently running okay now let's say all the engineers starting to raise their pull request
• 01:00 - 01:30 on get there and get it merge onto the developer itself right so now let's say uh when we when we start the release process what we will do is first take the changes from develop to release take the changes from develop to release okay so what we will do as a part of this change so now let's say your
• 01:30 - 02:00 develop release master so develop let's say after merging the couple of PRS got merged and after that develop version is let's say V2 release is still on V1 Master is still on V1 and in production the code is which is running on the V1 build that's it but develop has certain more changes and let's say its version is V2 as of today and now we wanted to start the release process so what is the first step take the changes from develop to release so
• 02:00 - 02:30 what we will do is we will take the changes from develop to release so release version now will become to V2 what does V2 means all the changes which are present into the develop now it's also presented to the release okay now what is the Second Step you start the regression on release branch okay so now what will happen is uh we do regression means running the Ft so you have my let's say functional test cases
• 02:30 - 03:00 100 functional discusses 200 functional distances and if you join certain big mnc's companies you will find that there might be a thousand of functional test cases also very very common right so you run your regression you run all the functional test cases on this build which is generated out of V2 so we are validating that we are not breaking anything in life right and let's say if regression is done properly what is the third step is push to life
• 03:00 - 03:30 push to live so now develop is you have V2 releases having V2 Master is having V1 in live build was running on V1 build so now as a part of third step after regression is completed on this V2 what we have to do is push to live what we will do is push to life so now we will change it to V2 build okay so what's the four step monitor
• 03:30 - 04:00 so now we will monitor the live that everything is working fine on the live traffic nothing is breaking it's all going good right so once we very validate that yes everything is fine this build is proper now what we will do is fifth step back much right so what we will do is we will do back merge two of uh update the master also and update the developer also so developers already has V2 itself no
• 04:00 - 04:30 let's say no changes has merged through the develop so no changes will come to the develop but Master will change it to V2 so now everything is V2 V2 V2 and and production also V2 build is running so now you can ask me a question that hey what is the use of Master then so we did a regression on release and from release itself we pushed to the live then what is the use of Master I'll tell you what is the use of Master then let's
• 04:30 - 05:00 say you have a developed release Master production the same let's say it has V1 V1 V1 and it is running on V1 build okay so now let's say couple of PR merge into the develop and the day which you started so developers has V2 release and master bar having V1 V1 and in production it was V1 okay so the day you started that yes I wanted to do a release there are so many
• 05:00 - 05:30 PR's much to develop develop I have certain V2 version of the code so we bring the changes to the release first step now in the second step regression started regression started let's say that regression might take one day regression takes one day okay so you are doing a regression now let's say while your regression is running ah hot fix you have to take what does
• 05:30 - 06:00 hotfix means hotfix means there is an issue in the live which you have to fix immediately right you can't wait like you can't just fix it into the develop then you start the release process then do regression and then do it has to be merged immediately fix it immediately if you have to take hot fix now release has V2 version because we have taken the changes from develop to release V2
• 06:00 - 06:30 version and regression is going on regression takes regression is very very effort taking step and it generally sometimes take one day also right because it might have thousand test cases and all so if you have to take immediately the hotfix how you will take it then they are the master Branch comes into the picture so Master is always equivalent to the production right so what we will do is we'll take
• 06:30 - 07:00 the hotfix here so let's say we'll take version 1.1 so certain fix on top of this generate a bill and push it so live would have 1.1 build so with the hotfix right so now let's say what happen is let's say you have taken a hot fix then what we have to do we have to take this changes to the release right otherwise otherwise if you take don't take the changes to release what will happen is
• 07:00 - 07:30 when this release regression is done if it puts to the production this hot fix changes will get loose right so we should not it this should not happen so what happen is once you have taken a hot fix you have uh put a fix over this master branch which was equivalent to the production you have made your changes push it to the live now customer impact is solved now what you have to do is you have to bring the master hotfix to release
• 07:30 - 08:00 okay so this hotfix changes bring to the release now you see that only you have to generate a build again so you have to generate a new build but your regression will won't impact whatever the time you have taken let's say out of 1000 test cases there were 50 failures and you are analyzing these 50 failures you don't have to start the regression again because you know that everything it has only one hotfix is coming so you you know that it's not
• 08:00 - 08:30 bringing certain additional changes so that you don't have to you have to start the regression again right otherwise let's say if you do this just assume with this one you have let's say V2 V1 V1 and let's say this is your production okay now you want to start the release process you change it to V2 and you started the regression
• 08:30 - 09:00 you started the regression regression let's say the version only take one day and your there are 100 failures and you are analyzing it now is certain hotfix come and let's say after this V2 there are certain PRS merge to the develop PR Marshall the develop and current version of the develop is currently let's say V3 okay so now let's say the hotfix come if you don't have the master branch what you have to do you have to revert
• 09:00 - 09:30 the changes in first of all release you have to make it so what you have to do is first again so develop is your V3 from R was your V2 now so you have to revert it back to the V1 so you have to revert the pr which you have bring the changes from develop to release right develop to release from V1 you become V2 so you have to revert that PR you have to again make it to the version one now you do a hotfix changes so you will make 1.1 generate a build and put into the
• 09:30 - 10:00 production okay if there is no master and then after this is done then you have to again bring this changes back from develop to release but now it will be now a kind of V3 and now you see that whatever regression you did now you have to start it over again because earlier whatever the changes you have done over the V2 after that certain more
• 10:00 - 10:30 changes has been done right so you have to validate those changes also right so you have to start the regression again and I can tell you this regression is very very very time consuming process there might be 2000 test cases certain time 200 failures you have to analyze each one why they are failing right so this is very very critical step so that's why in three branching system if any hot fix comes it is very easy
• 10:30 - 11:00 then it do not impact our regression process release process so if release is going in parallel then hotfix comes your release process doesn't get impact it will continue it will continue to do a regression and all if hotfix comes in parallel they are also not blocking that I have to revert it and then do this thing then release process will stop no so release guy will continue the regression and all hot fix guy who else has to take they will do it on Master
• 11:00 - 11:30 done it and bring these changes to the release only right so the release guy don't have to do the regression again because it know that what changes are coming and it's already in life so it doesn't have to new these are those are not new change which has to validate got it so you know right why develop release Master comes and one thing also is that you know right why we are doing this back much so back merge is also sometimes necessary
• 11:30 - 12:00 now let's say I told you that let's say it is V1 V1 V1 right it's a production let's say develop got few PR's merge it change to B2 now you wanted to start a release process it goes to the release it becomes V2 right and then let's say it everything is fine it goes to this live so it is on V2 build now currently it's M1 and it's V1 so now
• 12:00 - 12:30 everything is live now let's say while this is happen and develop certain more PR God merge and currently developed version is going V3 okay so now let's say this is live now V2 build V2 build is live but we are waiting we are currently in monitoring phase so that's why we haven't updated in master yet because currently it's monitoring whatever the build we have pushed to the live we are monitoring hey
• 12:30 - 13:00 okay monitoring can be one day we are just checking everything is going good or not right so now it is monitoring well now let's say we found certain bug in this build if we found certain bug in this build right and let's say we have to fix it immediately we can't take it from Master now because master has V1 so what we will do is we'll put a hotfix here in the V2 release so in the release so your develop is V3
• 13:00 - 13:30 release was V2 master was V1 in live it was running on V2 build now if with this V2 build during monitoring you find that hey it's an issue and we have to fix it so what we'll do is you will put a hotfix here in here in release so let's say you put a hotfix and you got certain changes in this merge now you uh take a build just do a regression and put it again here
• 13:30 - 14:00 so now live has v 2.1 build okay now you are doing monitoring now everything goes fine yes this build is perfect one day no no error nothing everything is good now you have to do is back merge why back merge is important because here if you see that foreign doesn't have this hot fixed changes
• 14:00 - 14:30 right if you don't do back merge and then when the next person start the release process it will merge from develop to release and when it merge from develop to release due to hot fix version which you have much to release will get loose right it get lost so what you have to do is after everything is fine you have to do backwards so that developed also get this hotfix changes into this one kind of V 3.1 you can say right not V 3.1 but develop will get
• 14:30 - 15:00 this hot fixed changes and similarly we will update the master that yes live build is up to date working properly now we can also have this version 2.1 this is version 2.1 and all version any hotfix also will also goes there so everything would be in sync now I hope you got it right why this three Bill uh three branch system is required right because hotfix can come
• 15:00 - 15:30 even in the live while regression is happening while regression is happening code has not been pushed but in the live build it comes in the Life code so hot fix can come then Master can help you if you have pushed the code from release right with the new code and any hot and you have to take hot fix then with the release you can do and then you have to do back merge okay and update the master
• 15:30 - 16:00 I hope you got it guys so this three branching system is very very popular and I think maybe in your companies you might be using it if not and I'm pretty much sure that maybe in your next company and soon you will see that this is how most of the companies are using it okay okay guys thank you bye