1. In VS Code this setting is called "Editor > Sticky Scroll: Enabled" and it is used so that you can see what the context is of the code you are viewing even if it is too long to fit on one screen.

  2. The key is to use the fewest dependencies possible and the most primitive ones, (not what is in favor this month) which will rarely go out of support. So using Webpack/Babel/Rollup/etc directly instead of via a higher level tool. Or using React directly instead of a Metaframework, or using CSS/SCSS instead of some CSS-in-JS library. The core technologies rarely change or stop being supported. So depend on those and depend on the core libraries directly which makes it so much easier to change later when necessary.

  3. I'm honestly surprised that fuzzing isn't used more in web dev. It'd be really cool to see a testing library that can exhaustively check every edge case by trial and error and infer based on the new UI if that was an expected result or not.

  4. The problem isn't how you are importing it, it's simply that getCode doesn't exist on that type as the error indicates. So you first need to find the correct type, then you can figure out the syntax to use it.

  5. Axios' Http Interceptors is one reason to use Axios rather than have to deal with overriding Fetch or pulling in another library.

  6. In most cases you can just create your own function which wraps fetch and get all of the benefits without any hacky magic or libraries.

  7. But why would you recreate the wheel?

  8. Because it's not reinventing the wheel. Odds are you wouldn't be reimplementing all of the features of the library, maybe one or two. The reduced complexity, dependencies, and lines of code is almost always an inherent benefit in the long run.

  9. I was really surprised with that claim too. So I ran some tests. (the slice method is actually worse)

  10. Your test should ideally run multiple times, then discard the result from the first run, as otherwise the result can be off depending on the behaviour of the JIT

  11. This sounds like it'd be far easier as a TamperMonkey UserScript. No need for all of the extra boilerplate and hassle of making a whole extension when all you want to do is add/change some stuff on the current page.

  12. Not the current page, but across a whole tool which is hosted as SAAS website. Do user scripts still make sense?

  13. Yeah that's what I usually use. They are super easy to create and share and they can do anything within a webpage that you might need. It's effectively the equivalent of running a script in the browser console right after a page loads. So anything you can do there, you can do in a UserScript.

  14. They don't have ublock for iOS? That sucks. :(

  15. You probably have circular type refs, which uses a CRAP ton of memory for typescript will finally be like "yeah, I quit" and throw a circular type ref error, or if not configured probably, just eat them.

  16. There isn't a plugin or eslint plugin I'm aware of that can detect it, because what we're talking about here is circular type references, not circular import references, but those are bad too. You can detect/prevent circular imoports by...

  17. Okay so TLDR as long as you run the typescript compiler at some point and check the output that'll tell you if there's a problem?

  18. Whenever you encounter a problem that you cannot solve with the first few attempts, your next step should always be to make a minimal reproduction of the issue. Not only will that be helpful to others when asking for help (if they cannot reproduce, they cannot help), but often you'll find the solution yourself through going through that process.

  19. I think that’s sometimes overlooked but the key with typescript is:

  20. Yup that's because it's super hard to implement it correctly. Most servers set cookies immediately, so it is long before the user has a chance to consent. It's a lot more difficult to not do anything until you get explicit permission.

  21. I know it's fun to hate on job requirements, but what is the problem here? They aren't asking you to know everything listed, those are just examples for each category.

  22. I created the themes myself, you can see them here:

  23. I'm always so surprised with how big the internet is when the creator of something finds a post about it so soon after it was created and replies. Kudos!

  24. Honestly the best way I found is to setup strict tsconfig and TypeScript ESLint rules, and then just start coding. Anything which is bad/wrong/suboptimal will have a squiggly line and then you can learn about each one as you go and how to fix it. I find that extremely helpful for learning if you are already experienced with JavaScript.

  25. I played this for about an hour or so (up until level 100). It's a cool concept. But it quickly loses excitement when the skin wears off. Every gun is functionally the same. There's no decisions to make since you always want to max everything. And even the fact that the different categories of guns unlock sequentially only once you've complete the previous category makes the categories feel hollow rather than distinct.

  26. In general, all of these new tools make the trade off of robustness for speed. Babel can do almost anything you can imagine, and likely there's already a well-supported library to enable it easily. And if you want to do something novel, there's an insane amount of docs and examples plus a robust API to enable your crazy custom ideas.

  27. Yeah, I read your template and I know the advantages of using Babel myself. I'm here looking for the opinions of those experts who sincerely tried both in the past and know about the trade-offs. Did you try both?

  28. Are you talking about supporting ES5? Or are you asking why we should still support old versions at all? Firefox ESR is a great example of an enterprise browser version which updates far less frequently than the consumer browser so if you only support the last few versions of Firefox, you won't be supporting Firefox ESR users who might be a dozen versions behind latest.

  29. While I agree with you on many of your points against Jest, I don't think there's anything that fundamentally fixes any of them. There's vitest but that's basically a rewrite of Jest with most of the same paradigms.

  30. AWS is extremely cheap, at your scale it will be nearly free since you'll be at the free tier for almost everything you are doing. The trade off is, to get that absolute cheapest price, you need to build your app to take advantage of the specific services that AWS offers. So if you have an app that you've already built and aren't willing to rewrite any of it, then it's likely going to be much more expensive. Even still, you can get the smallest

  31. How do you keep your data on aws? I know a RDB instance gets expensive fast.

  32. It depends. But your code should just import something from "package-b" and it'll find it regardless of where npm decides to install it.

Leave a Reply

Your email address will not be published. Required fields are marked *

News Reporter