Changelog

In this blogpost we’ll talk about the old Nginx-based traffic infrastructure, its pain points, and the benefits we gained by migrating to Envoy. We’ll compare Nginx to Envoy across many software engineering and operational dimensions. We’ll also briefly touch on the migration process, its current state, and some of the problems encountered on the way.

Envoy, for the uninitiated, is a proxy server “designed for cloud-native applications”. It was created by Lyft and used by a lot of big players in the cloud/services world.

Not only is this article interesting as a “switching” story, it’s also fascinating because of the scale of the migration:

When we moved most of Dropbox traffic to Envoy, we had to seamlessly migrate a system that already handles tens of millions of open connections, millions of requests per second, and terabits of bandwidth.

High stakes!

John D. Cook:

I don’t recall where I read this, but someone recommended that if you need a tool, buy the cheapest one you can find. If it’s inadequate, or breaks, or you use it a lot, then buy the best one you can afford.

If you follow this strategy, you’ll sometimes waste a little money by buying a cheap tool before buying a good one. But you won’t waste money buying expensive tools that you rarely use. And you won’t waste money by buying a sequence of incrementally better tools until you finally buy a good one.

What follows is an application of that idea to software tools.

Two days ago on this repo appeared on the top starred repositories first timers list on Changelog Nightly…

In this repository, you can find the official GitHub public product roadmap. Our product roadmap is where you can learn about what features we’re working on, what stage they’re in, and when we expect to bring them to you.

The roadmap repository is for communicating GitHub’s roadmap. Existing issues are currently read-only, and we are locking conversations, as we get started. Interaction limits are also in place to ensure issues originate from GitHub. We’re planning to iterate on the format of the roadmap itself, and we see potential to engage more in discussions about the future of GitHub products and features.

It’s easy to forget that there’s a human on the other side of that <textarea>. So we tend not to give people the benefit of the doubt on the internet. This post is a gentle reminder of that fact and how active awareness of it would go a long way toward making it a more enjoyable place for all of us.

A solid, brief Q&A with a couple of Redis’ core team members about what’s next for the project now that its BDFL is no more.

This the first majorly bearish case I’ve read on remote work:

… remote work makes you vulnerable to outsourcing, reduces your job to a metric, creates frustrating change-averse bureaucracies, and stifles your career growth. The lack of scrutiny our remote future faces is going to result in frustrated workers and ineffective companies.

Let’s tackle these issues one at a time.

Robert and Ian join us to talk about the latest updates on generics in Go. What type of feedback are they looking for as developers get their hands on tools designed to experiment with generics and Go? What was the deal with the featherweight Go paper that also discussed generics? Why can’t we use angle brackets for generics?

You can use this set of guidelines, fork them or make your own - the key here is that you pick a style and stick to it.

This is a great guide, but the 🔑 is definitely in the highlight above. Consistency is tantamount to readability.

Yes, you can do keybindings directly. But for a mere ~400 bytes you can do keybindings like this instead:

import tinykeys from "tinykeys"

tinykeys(window, {
  "Shift+D": () => {
    alert("The 'Shift' and 'd' keys were pressed at the same time")
  },
  "y e e t": () => {
    alert("The keys 'y', 'e', 'e', and 't' were pressed in order")
  },
  "$mod+KeyD": () => {
    alert("Either 'Control+d' or 'Meta+d' were pressed")
  },
})

First things first… does it actually work?!

Yes! Quite well, actually - on macOS, Windows, and Linux. Bear in mind that this is written entirely in JavaScript, so please adjust your expectations. The virtual machine is emulating a 1991 Macintosh Quadra 900 with a Motorola CPU, which Apple used before switching to IBM’s PowerPC architecture in the late 1990s.

Ok, cool. Does it run my favorite game?!

The short answer is “Yes”. In fact, you’ll find various games and demos preinstalled, thanks to an old MacWorld Demo CD from 1997. Namely, Oregon Trail, Duke Nukem 3D, Civilization II, Alley 19 Bowling, Damage Incorporated, and Dungeons & Dragons.

There are also various apps and trials preinstalled, including Photoshop 3, Premiere 4, Illustrator 5.5, StuffIt Expander, the Apple Web Page Construction Kit, and more.

A nice primer on JavaScript’s abstract syntax tree (AST), how to understand it, and what you can do with it. If this is a subject that interests you, also take a listen to this episode of The Changelog with Amal Hussein.

Marius Andra added TypeScript support to his React state management library with kea-typegen:

I tried a lot of different ways to add TS support to Kea. At the end of the day what was needed was just not possible without an extra type generation step. I’ll document the failed attempts in a blog post one day. The issue boils down to the fact that each block inside a logic (e.g. reducers) gets the logic itself as a parameter (reducers: logic => (...)). That kind of loopy stuff is just not possible with TS right now. At least not to the level that it has to be for it to work in Kea (think of selectors that depend on each other). Ultimately, all of the pure-TS attempts I tried would have yielded a partial solution.

Thus kea-typegen was born. Even if it’s just a step along the way (TS compiler plugins a’la babel someday?) and still missing many features, it’s already really useful. It uses the TypeScript compiler API to analyse your source files and print out complete types (interfaces) for them.

Marius also recorded a video of the process in action.

You’ve likely heard a lot about Google’s monorepo and how it impacts the org’s development productivity, but have you heard how it makes managing their open source efforts easier as well?

Leading off the updates for Git 2.28 is the highly sought after ability to configure init.defaultBranch so folks can move from master to main as their default branch name.

From Taylor Blau on the GitHub blog:

When you initialize a new Git repository from scratch with git init, Git has always created an initial first branch with the name master. In Git 2.28, a new configuration option, init.defaultBranch is being introduced to replace the hard-coded term. (For more background on this change, this statement from the Software Freedom Conservancy is an excellent place to look).

Starting in Git 2.28, git init will instead look to the value of init.defaultBranch when creating the first branch in a new repository. If that value is unset, init.defaultBranch defaults to master…

Also check out github/renaming to learn more about the complementary changes GitHub is making. GitLab and Bitbucket are making similar changes.

Michel Terhar:

In the search for a comfy and portable developer experience, I’ve made a lot of compromises in the past. The experience has gotten significantly better recently thanks to VS Code and Kubernetes. This workflow also does a good job for underpowered laptops or when working with lots of different and conflicting versions of python or ruby.

This is a solid, balanced piece that doesn’t overly sell the workflow and walks you through setting it up for yourself.

Edward Wible, Nubank’s CTO, shares his perspective on the future of Clojure and Datomic, and how the powerful ideas that guide these technologies helped shape Nubank’s culture and business.

Funny timing. Just the other day José Valim was telling us how Nubank is big on Clojure:

Come to think about it, they are one of the biggest cases of large companies using Clojure at a really large scale… So it’s actually interesting to hear about everything they are doing with Clojure, and the interesting cases, and how they are using the Clojure stack… And they write about it, they give talks, so it’s very interesting to check that out.

So in that sense, if they brought me, maybe they would be bringing Michael Jordan to play soccer… 😆 That’s not going to be a good fit, right?

If you’re wondering if José is audacious enough to compare himself to Michael Jordan… he didn’t. I did.

Sash Rush, of Cornell Tech and Hugging Face, catches us up on all the things happening with Hugging Face and transformers. Last time we had Clem from Hugging Face on the show (episode 35), their transformers library wasn’t even a thing yet. Oh how things have changed! This time Sasha tells us all about Hugging Face’s open source NLP work, gives us an intro to the key components of transformers, and shares his perspective on the future of AI research conferences.

I’m not sure what makes this lightweight (their word, not mine), but it does load pretty fast from where I’m accessing it. I definitely see what they mean by powerful, though, as wiki.js boasts many features: multiple editors, multiple auth schemes, search functions, comments, multiple locales, the list goes on…

The demo is worth a thousand words.

Jerod assembles a team of WebRTC experts (Suz, Feross, Mikeal) for a deep, deep dive on this practically-ubiquitous yet still-complicated web API.

We review its history, share really cool applications using the tech, provide an excellent primer on what you need to know about it, and details some production gotchas. ALSO we celebrate how Feross single-handedly “upgraded the internet”! 🙌

Addy Osmani:

Native lazy-loading for images landed in Chrome 76 via the loading attribute and later came to Firefox. We are happy to share that native lazy-loading for iframes is now standardized and is also supported in Chrome and Chromium-based browsers.

<iframe src="https://changelog.com" loading="lazy" width="600" height="400"></iframe>

YaS!! Lazy load all the things 💚