So creating a new repo on GitHub, you get a set of getting started steps. They changed the default branchname to "main" from "master" due to its connotations with slavery.
When I create a new repo now, the initial getting started steps recommend creating a branch named "master" as opposed to "main" as it was a while ago.
It's especially weird since the line git branch -M master is completely unnecessary, since git init still sets you up with a "master" branch.
Disclaimer: I have a bunch of private repos, and my default branchnames are pretty much all "master".
Is this a recent change?
Edit: Mystery solved, my default branchname is "master". Thanks [email protected] !
Again, you’re conflating your own stubbornness with correctness and that just ain’t how it goes. Branch names are frivolous. So much so that changing the strategy or retargeting a branch one time shouldn’t be such a nightmare for your pipelines that you have to pretend like you’re the big dick on campus spouting accomplishments when someone mildly suggests there’s a mistake in your thinking. Look inward if you’re so upset by this that you have to make up irrelevant insults in a vague attempt to protect your own ego, then go fix your pipelines to make it easier to do for the next person after you’re gone.
@[email protected]
@[email protected] @[email protected] @[email protected]
The posts you are replying to ha e been deleted. I'm really currious what they said because we have one vendor who claims to be/is locked into usung "master". This either requires us to write CI that merges main -> master and mirrors master back to main or use master. This can confuse junior devs once or twice, but it is really not an issue. The ONLY time I felt compelled to use master because of this vendor was when working with a group using GitLab. GitLab has a feature called Pull Mirroring that is MUCH more reliable than a pull/mirror action in GitHub that does the same thing, but to use that the branch names had to be the same.
I see both sides of this argument. The master/slave relationship in tech is NOT like masterworks or mastering a craft. It is based on one "owning" the other, but I don't think that allowing technology to work that way is violating its rights. Obviously changing the name doesn't change the behavior and isn't it really only when that behavior is applied to people that we have a problem with it?
I never fully supported the effort required to change, but I've also never written anything in a way it would be difficult to change. I recognize that it could be considered a micro aggression, but it's not like we are going to stop ants or bees from treating other classes as forced labor. Slavery exists. It is bad when applied to people. It accurately describes tech. Changing the name of the master db or branch did NOT free the slaves.
The stupidest problem I've seen arising from this debate was that with one employer they had a legally required retention policy, and instead of implementing it in their GitLab server software, they did it directly by coming in between GitLab and git. The result was that they had no idea which to use, so they protected both.
On one repo, we mistakenly made both branches, and there was no way to get rid of either, so it kinda just stayed there. It confused the hell out of new people.
I use master for my own stuff because of muscle memory, on the job IDC I use whatever, usually main these years.
Wait, this is a thread about branch names in git. The "master" in question would be more akin to a "master recording" from music, not master/slave software or system architecture.
While it may be true that the master branch is more akin to a master record, not everyone knows the nuance and quite frankly it doesn’t matter, if it makes people uncomfortable then it shouldn’t be a problem to accommodate a simple change, most of the tech world has already done so. Computers used to have a literal slave/master relationship with hardware components and control systems and we moved past that just fine despite still having controllers and actors everywhere.
It was someone ranting about the many hours and days of lost productivity and cost of manually switching over 70+ legacy build pipelines all because of a branch name change. Also lots of condescending and insulting language from someone who thought their stubbornness and “standards” meant they were better than everyone else. Honestly I just probably set them off in my first message and they wouldn’t let it go, leading to increased levels of ranting and insults from them attempting to spout accomplishments while detailing their failings in the same breath. Admittedly that above description is a bit belittling from my end, I’m just annoyed they couldn’t keep their messages up for all to see.
I still stand by the opinion that changing branch strategies, names, or targets should not be a multi-hour multi-resource process and if it is, that’s a failure of the systems engineers / ops who put together such a plan. It’s possible to have CI/CD pipelines that run for years on end attached to critical infrastructure while being flexible enough for such simple config changes and maintained by one engineer.