this post was submitted on 20 Feb 2025
59 points (100.0% liked)

ProgrammingCircleJerk

61 readers
9 users here now

Programming CircleJerkCommunity at request of a user.

founded 2 weeks ago
MODERATORS
all 24 comments
sorted by: hot top controversial new old
[–] [email protected] 18 points 2 days ago* (last edited 2 days ago) (1 children)

From the comments:

I can't wrap my mind around 10 people pushing changes to main all the time. It feels like a merge-conflict-rollback-preventing nightmare to me.

That's ... literally what one of my colleagues is unironically advocating for over a decade. Despite us trying to explain to him that this might work for team of 2-3 people but is really a bad idea for team of 10+ developers, 5 or so testers, POs needing guarantee that they can deploy at any time and tools like Gitlab/Github. Thankfully we overruled him few years ago

[–] [email protected] 14 points 2 days ago (3 children)

This is obviously crazy talk, but so is keeping feature branches alive over extended periods of time. Depending on the development tempo of your project and the number of people involved, you should figure out a way to land things into main (in pieces if needed) within a few days.

[–] [email protected] 3 points 2 days ago

Or never killing branches... I've been on projects with over 100 branches.

[–] [email protected] 3 points 2 days ago

Feature flipping

[–] ulterno 1 points 2 days ago

I had created a feature branch before leaving my previous company, completed the feature and uploaded a build for testing, after which it was supposed to be merged.

Months passed and it has still been neither merged nor rejected.

[–] [email protected] 6 points 2 days ago

That hurts my soul.

[–] onlinepersona 3 points 2 days ago* (last edited 2 days ago) (2 children)

Any CVS and SVN lovers in chat? Those were the days, eh? Got and mercurial were a mistake. Bring back the good old days 😭😭😭

/s

Anti Commercial-AI license

[–] [email protected] 5 points 1 day ago* (last edited 1 day ago)

Back during the master/main debate I proposed going back to “trunk”, but there were too few of us old farts to make it stick.

That said, I didn’t miss subversion and CVS even a bit. They were rough compared to modern source control. I did like both hg and bazaar better than git though.

[–] [email protected] 3 points 2 days ago (1 children)

Real men use classic TFS. Real men then pursue a career in a different field so they don't have to use TFS anymore.

[–] NostraDavid 1 points 2 days ago

Sneakernet and floppies, just the way Carmack intended!

[–] FizzyOrange 1 points 2 days ago (2 children)

What is even the motivation for this? How do you make PRs?

[–] [email protected] 8 points 2 days ago (1 children)

That's the neat part: you don't

[–] FizzyOrange 3 points 2 days ago

A joke presumably? Kind of hard to tell!

[–] onlinepersona 6 points 2 days ago* (last edited 2 days ago) (1 children)

It's called "trunk based development". Look it up and watch videos about it. I swear they are stuck in the 90s yearning for CVS and SVN.

Anti Commercial-AI license

[–] brian 2 points 1 day ago (1 children)

trunk based development generally still merges branches into a main one, see this. idk if only using main even has a name

[–] onlinepersona 2 points 1 day ago

Trust me, people still call it trunk based development. Just like "agile" has many forms, trunk based development does too.

Managers will often read about some process, tech, or strategy and tout it as the solution to their problems without understanding context nor wanting to invest in what's actually required for a successful implementation. For example trunk based development should have automated testing, a deployment and a fallback strategy. And if you're working on something with hefty features that you can't slice, feature flags or something similar are probably required. Not giving the team time to develop a strategy to implement that is a great way to mess up a product.

Anti Commercial-AI license

[–] [email protected] -4 points 2 days ago