It’s cute. Maybe my favorite use of ai I’ve seen in a while.

I wish it looked at contributions instead of just the profile page. Much more accurate roasting.

Mine looks a little like that. It’s my job though. Everything’s on GitHub.

I think the technologies are pretty bubble based. We are 80/15/5 Mac/Linux/Windows and it’s been 15 years since I worked on a software team that’s thats mostly windows. But I talk to them from time to time. But if anything Mac feels underrepresented compared to my bubble.

I admit I’m probably biased in favor of believing the survey is representative. I work on one of the databases.

Speaking of databases, I don’t work on SQL Server but can see the appeal. It implements a huge array of features and it’s documentation is pretty good. Folks have told me it’s a lovely database to use.

I think all those are a little true. But I’m mostly guessing. I’m happy to change my mind if anyone knows better.

Either way, these folks are my hero.

Usually I use glob patterns for test selection.

But I did use reges yesterday to find something else. A java security file definition.

Thanks. I remember one of these had people being excited about it and I felt bad that I couldn’t try it. But Linux is hard and we are all so grumpy. I get it.

Is that the Mac only one?

The point of the license combination they use is to allow the enterprise version to be open and live in the same repo as everything else. Dunno if that’s what they do, but that’s why the elastic license exists.

The only surefire way is to read it all. And understand it all. That ain’t happening though. So you decide how much to do.

You should figure out how many people are landing patches and get a rough sense of why. Same for folks filing issues or talking about the project in general. Maybe you trust one of the contributors for some reason. Either way, you want to know how alive the project is.

You could land a patch.

You could spot check parts of the code.

You could run vulnerability scanners on it.

I dunno. It’s hard.

I’m not sure I’d attach any meaning to real names online. There’s a whole group of us whose online names are just things they thought were neat when they were 12. And they’ve just stuck forever. There’s lot of reasons.

But otherwise, yeah. I’ll spend ten minutes looking up someone’s online profile. Mostly for GitHub if I can find it. If someone’s commenting on public prs and seems nice that’s a big signal.

I agree. Light touch until you have a bunch of changes landed.

I was a professional open source contributor for a while. Still have the same job, but the license changed. Culture still quite similar though.

We squash. I’m not really interesting in your local journey to land the change. It’s sometimes useful during review, but after that it’s mostly the state of the main branch I care about. It’s what I need to bisect anyway.

I don’t like commits that are just references to issues. Copy the issue into the commit message so git blame tells you something useful. Unless it’s just closing a simple big. Then the title and issue reference are plenty.

Depends on the project I imagine.

I wonder what my last commit at each job was. I’ll bet it was boring. About 10% of my commit messages are genuinely interesting.

I review a ton of code and have a bunch reviewed in turn. I don’t remember that last time I’ve had this come up. Either direction really. I guess I’m lucky. We just split naturally in similar places.

I think it’s a bad analogy because it’ll distract some people.

I believe they were referring to this: https://youtu.be/9eyFDBPk4Yw?si=sb_v_EPhTM9C6bZZ

I imagine a delta encoding scheme similar to what the time series DBs use would work well for the geo points. Maybe even a delta-of-delta encoding for things like ships which move very consistently.

It’s probably not worth it given how small they’ve already go their data. But it is fun.

I don’t believe Prometheus supports geospatial data. Two minutes of googling though, so I could be wrong.

I think the last new instruction the JVM added was invokedynamic like 10 years ago. I believe they did it so lambdas could be called efficiently. Polymorphic incline cache and stuff.

But the JVM has grown more complex in other ways. The way to force simd instructions is pretty wild, for example.

I don’t know enough to call it a mess or not. It works though.

It’s not just Google. If you enjoy that sort of thing there are industries where it’s more important. Not every day. Not every team. And you’d have support like you say.

But you can go a lifetime without using it beyond rules of thumb.