MB

[MB @essenmitsosse]

Finally decided to put my responsive pixel illustrations on GitHub. The original code was actually written without git. Crazy me back in 2014.

There is still some unpublished stuff that I will add later.

https://t.co/H9bvvrUdrj

[MB @essenmitsosse]

@aylacroft @t3dotgg Had exactly the same vibe and was surprised not more people noticed it for how obvious it was.

[MB @essenmitsosse]

Zengarden was a mistake.

[Theo - t3.gg @theo]

Might have went a bit too hard on this one

[MB @essenmitsosse]

@siteinspire Why am I getting advertising from @designdeck_io via the mail I used for SiteInspire? Not cool.

[MB @essenmitsosse]

@AwayCaludio @dhh Could potentially work, if the function itself doesn’t rely on „this“. Would still consider this bad practice though.

[MB @essenmitsosse]

So far, the biggest speed improvement bun has brought me is saving three characters because I can just do "bun lint" instead of "npm run lint".

The actual command execution isn't measurably quicker.

[MB @essenmitsosse]

@WebReflection @anatudor Just to be clear: never do that. Order is not guaranteed and might change for any kind of reason. This will lead to horribly hard to debug bugs.

[MB @essenmitsosse]

@mattpocockuk I was about to argue that it is not technically wrong by TypeScript standards (even though TypeScript is stricter about tuples than you might expect). But one thing made it obvious this must be a bug: you can also call pop — which definitely ruins everything.

[MB @essenmitsosse]

XML was a mistake.

[James Ross @CherryJimbo]

React was a mistake

[MB @essenmitsosse]

@shoeboxdnb I don’t think @Netlify has spending limits either. I am getting really suspicious of the way these services make their money.

[MB @essenmitsosse]

Why would you have a pricing model that so heavily disincentives good practices by charging you more if you split your assertions into separate tests?

[MB @essenmitsosse]

@Cypress_io is going crazy with their prices recently. They charge 6$/1000 test runs. That means if your suite has 50 tests, they charge you 30 cents. Per commit. For using their interface. That doesn't include the machine that runs the tests.

[MB @essenmitsosse]

Yarn are you alright? Need a hug?

I am getting serious „gave up on life“-vibes. Wouldn’t be surprised if in some section of the Yarn docs there would be an emo poem about jQuery.

[MB @essenmitsosse]

@marceloabsousa @housecor I think that idea is not realistic. PRs and commits are different concepts and enforcing a 1to1 relationship between the two will weaken one of them. Either your PRs become less useful or your commits to broad.

[MB @essenmitsosse]

@mattpocockuk This is great. I used a similar pattern before but wondered where the name „branded“ comes from. Is there some history to it?

What I would be careful with is selling this as a security feature. People might take that the wrong way and rely only on types for security.

[MB @essenmitsosse]

@philipvollet @BramVanroy That's a weird looking MacBook 😆

All the best for the new chapter!

[MB @essenmitsosse]

Liking CSS is caused by Stockholm syndrom. I am going to die on that hill.

[Ben Holmes @BHolmesDev]

I genuinely don’t understand why so many people hate CSS

[MB @essenmitsosse]

@mattpocockuk @oezguerisbert I was seriously worried for a second.

[MB @essenmitsosse]

@MafVosburgh @codinghorror Then he could have just figured that out in court if it’s such a no-brainer. Surely the lawyers would have been the cheaper alternative.

[MB @essenmitsosse]

@thethinkingbat @chinmay @shl @Tesla That aged well

[MB @essenmitsosse]

@housecor 2. when you do Type-driven development. Type the function so it can already be consumed by others without error, then implement it.