tatterdemalion @ tatterdemalion @programming.dev

Professional software engineer, musician, gamer, stoic, democratic socialis

---

Read more

Posts

6

Comments

524

Joined

2 yr. ago

Yea this looks productive.

Imagine how useless the LSP suggestions are.

I'm actually building a new work station right now.

I know for a fact that my company's build process is twice as slow on Windows/WSL than on vanilla Linux. We have benchmarks from many different user environments.

Windows managed to brick itself when I booted for the first time in a month. I only wanted it for the Karafun app, but I guess I can live without it.

This is dangerous misinformation.

Not my main rig, but my most unusual is 32-bit Yocto Linux on an Intel Edison that I got for free from a college professor that worked for Intel.

Yocto is awful. I mean it has a niche I guess, but there is basically no package manager. Somehow I managed to install a Rust toolchain on it, but it couldn't build the web server I wanted to run on it.

I'd much rather have a Pi running a sane distro.

It's an investment for the next time you install on a new dev machine. After install, I will literally run a single command to return to the exact state of my dev environment.

Oh good point I totally forgot it's a JVM lang.

Fair. I was mostly thinking about it from the perspective of what the writers had intended at the time.

I thought the Empire in Star Wars was supposed to be Third Reich Germany. They literally both have "stormtroopers."

Erlang really is the necromancer's language.

Why is Swift... like that?

I assume Scala is like a "gateway" (drug) to functional programming by way of also supporting imperative and object oriented paradigms.

You come off as incredibly patronizing. You should take that to heart.

K what do I know, it's just what I've done on code bases with ~5 million lines of code, at the point where no one understands even half of it.

Let's step back. We are trying to talk about code review, right? That's when you are getting a final check that your code looks roughly correct and does what you already planned for it to do with your colleagues, give or take a few details you couldn't plan for. So at this point, usually at least one other person knows what you're trying to do, is bought in, and should be able to review it.

I'll concede that if you are desperate to merge a PR and people are on vacation or something, then you might want to grab someone else to review. I'd send them the PR and say, "let me know if you have questions." Then if they do, you can have a quick chat or meet one on one.

That's code review.

You are talking about some other thing. I've been in internal release meetings with lots of people that needed to understand the basics of a feature in order to support it or use it. That happens way after the PR is merged. And I've been in meetings to figure out how we should solve a problem, which happens before the PR is reviewed. I have gotten PSA emails or slack messages about cool new tools that make my life easier or changed our coding standards in the organization.

I have never been asked to meet with multiple people about what a PR is about before reviewing it.

IME those meetings only happen if the PR is objectionable upon review and it's hard to negotiate what must be changed.

Do you not do demo meetings after introducing entirely new features?

Sure but it's rare that a single PR introduces anything but a small slice of a feature, and the demo is usually something that can just be recorded in under a minute and sent over chat. The PR reviewer is likely to try building and using the new functionality too.

And if it’s an internal tool/service for fellow devs to use to make their lives easier, yes, it likely deserves a meeting so the devs can have a chance to QnA about it.

Just document the tool or service. That must exist anyway if it's going to be used by many people. And we're not really talking about a PR review at this point, we're talking about an internal release ceremony, which may involve parties that are not going to look at the PR.

If you create a new tool like that abd then just hand it off with all the devs have to go on being “here’s the manual, figure it out” you know what happens?

Almost no one reads it, and pretty much no one uses it

You're still straying from the topic of conversation: PR review process. But regardless... If no one uses it then it's probably not useful. Why did you build something that didn't compel people to use it from a one-paragraph internal PSA?

It's surprising how Github is sorely lacking in git features and even encourages bad practices.

1. There is only a linear history view of commits without any child/parent relationships visualized. I have coworkers that tell me it's annoying how I use merge commits because they just see a big list of small commits in the history, and they'd rather see one commit per PR. Well... I sympathize, but also, fuck that! I'm not squashing my carefully crafted commits because Github has a shitty UI. Use git log --first-parent or something like lazygit. It's sad that many open source projects have adopted a "squash and merge" mantra for PRs. All of the merge methods have a valid use case, but people are so beholden to the damned Github history view.
2. PRs have no concept of what a patch is. If you rebase, the entire history disappears. Reviewers cannot see changes made to individual patches unless they are applied first in new commits and then manually squashed into old patches after the review is done. Phabricator does this better IMO; each patch has its own history. I've even heard devs at other companies tell me they never make PRs bigger than 100 LOC. That just seems like a huge waste of time to me, unless you have some special tools to support that bizarre workflow.
3. Merge queues only support a single merge method, one of merge commit, rebase, or squash. So we just don't use this desirable feature because of this limitation.

Or use a tool like StackedGit which makes the atomic commit workflow incredibly simple. Build atomic commits as you go instead of after you've written all of the code.

If you need a whole meeting to explain what's going on in your PR then it's already too complicated, IMO. That explanation should be done as a combination of PR description, commit messages, documentation, comments, and the code itself.

Everyone will forget a meeting. The stuff I described will pay dividends for future maintainers.

The meetings should happen at design and planning phase.

Even though the "open" option would only allow Threads accounts to post via mention, I am still concerned about bots. I imagine as soon as bot owners find out they can get access to more views via federation, they will spam as many instances as they can get away with.

So I am voting at best "limited" to curtail spam in the feed. I am less concerned about spam in the comments because it will hopefully get downvoted. But who knows, bot accounts may be upvote farming.

I am split between "limited" and "blocked" at the moment. I don't mind real Threads users being able to subscribe and comment, but bots make me worry.