Johannes Ewald

Oh ok, that’s a political reference, I didn’t get that. Nevermind 😃

I have a vague idea, but I always forget what is what 🥲

I thought esbuild produces a single file but maybe there‘s an option

What tool does Node use under the hood?

What's currently the fastest way to transpile TS to individual JS files without type checking? We can't just strip types because we use enums.

Is it? Did you check every digit? 😃

Yes, it was meant to be abstract, not about TypeScript.

Is this something experienced Rust developers would solve easily or is it something that is just really hard in Rust?

(I‘ve encountered this in my pet project and I gave up 😅)

Why "Choose the right tool for the right job" might be true, but unhelpful (and why choosing the right programming language is less important than people think):

That’s so cool to see it in a real-world project 😃

Thank you! 🙏 So it‘s not useless 😌

To be fair: I don’t maintain it anymore because I don’t use it. I gave edi9999 maintainer rights but they‘re doing a great job as far as I can tell :)

As far as I remember I needed a safe way to evaluate expressions without using eval(). I was using Angular.js in a different project back then so I decided to extract "the good parts" ;)

So nice to hear that it’s still of great use for you 😌. What‘s the topic of your talk?

Hat mich auch erwischt 😅🥵

Not sure what you mean. I don’t see any error in my playground…

I agree! Comments should mostly explain the Why, not the What 😃

This summarizes my experiences with AI coding assistants so far. After my initial reservations, I started to appreciate them for quick feedback or easy refactorings.

They’re impressive for the first 50% of the task, but I often hit the point where I think it would be faster now to do it by hand.

I don’t know but I wish they’d use Array<Type> as the canonical form :)

This hurts 🤦‍♂️

Nice!!

I hate that AIs are still bad at refactoring code. There‘s so much refactoring that‘s too sophisticated for string replace but too dumb for humans 😃

That‘s what I did :) I decide against rules that complain about code that works and isn’t flawed in some other way.
I asked @mattpocock.com the other day and he also turned it off

Regarding skipLibCheck: I agree with your sentiment but I often find myself turning it on it big projects. It’s mostly necessary because of conflicting types in node_modules (e.g. Cypress bringing their own type definitions for Mocha, etc.).

Regarding exactOptionalPropertyTypes: it sounded very promising, but it complained about lot of code that was just ok. Most of the time people don’t distinguish between non-existing properties and undefined and it’s kind of ok.

Regarding noUncheckedIndexedAccess: I‘d turn it on. I don’t think it‘s about holes, but rather indices out of bounds. I saw several instances where it caught actual bugs in sloppy code.

Very interesting, you covered everything 😀

JavaScript devs: when your backend is written in another language and it returns snake_case, do you transform them to camelCase or keep them as they are?

I gave it a shot:

So lassen sich mit CSS, JavaScript, SVG und Canvas bezaubernde Interaktionen und kleine magische Details gestalten.

Hier verrate ich meine besten Tricks!

@joshwcomeau.com I‘d go with this translation, it sounds very natural to me :)

Only change I‘d make: instead of "reizende Details schaffen kannst" I‘d use "reizende Details kreierst“, but that‘s just my personal taste.

Totally agree with this one :)

For the best result, you‘d need to translate it more freely. The other translations here are correct, but they still sound translated.