Bifalgorithm post

"Dude, GitHub just released stacked PRs. What the heck is a stacked PR? Exactly what it sounds like. Basically, you're now able to stack PRs on top of each other so you don't have to wait for one to merge before you keep working on it. Okay, and what exactly is the point of this? This is how most developers work. You get assigned a fairly large task and you create a PR to get the first part of it done. You submit that for review, but because there's other changes waiting to get reviewed, it takes your teammates a couple hours to a couple days to review that change. And during that time, you don't want to start work on another PR because what if this one has to change? So you end up getting blocked on your second PR. for these guys to review your code. Well, with this, instead of waiting, you can just create another PR on top of that initial PR, and each PR only contains specific changes. So you could start off with a large PR, covering a database change, then you could do an API level change, and a front-end change, and people could review each part individually. That's kind of nice, but I could already achieve this with branches. You could, but it was kind of a pain to manage. Now this is a first-class feature, and there are dedicated commands to do it. I mean, I guess that's nice. Is there any reason why they decided to do this now? One of the biggest reasons is because with AI, people are writing a lot more code. As a developer, you could just patch an agent that might create a 4000 line PR in 20 minutes. But with stacked PRs, you can have your agent break that up into multiple smaller PRs. So if you want to review the code, it becomes a lot easier to do so. I'm still not convinced. What do you guys think?"

"Dude, GitHub just released stacked PRs. What the heck is a stacked PR? Exactly what it sounds like. Basically, you're now able to stack PRs on top of each oth…