@oljimenez.bsky.social
I need a way to make it work even on frameworks like Nextjs that does SPA like navigation.
May 31, 2025 at 1:32 AM
I need a way to make it work even on frameworks like Nextjs that does SPA like navigation.
One thing i always tried to learn when using React
How do i fire an action when a page load? Should i use "load" event from browser?
Somehow in every codebase i worked people do this on a useEffect with no dependencies, and i knew that's wrong, but i never knew a better way
What do you recommend?
How do i fire an action when a page load? Should i use "load" event from browser?
Somehow in every codebase i worked people do this on a useEffect with no dependencies, and i knew that's wrong, but i never knew a better way
What do you recommend?
May 31, 2025 at 1:29 AM
One thing i always tried to learn when using React
How do i fire an action when a page load? Should i use "load" event from browser?
Somehow in every codebase i worked people do this on a useEffect with no dependencies, and i knew that's wrong, but i never knew a better way
What do you recommend?
How do i fire an action when a page load? Should i use "load" event from browser?
Somehow in every codebase i worked people do this on a useEffect with no dependencies, and i knew that's wrong, but i never knew a better way
What do you recommend?
I'm talking about a tree of React code. For example
```tsx
<SomeComponent>
<OtherComponent/>
</SomeComponent>
```
Who's the Server Component?
```tsx
<SomeComponent>
<OtherComponent/>
</SomeComponent>
```
Who's the Server Component?
May 22, 2025 at 9:24 AM
I'm talking about a tree of React code. For example
```tsx
<SomeComponent>
<OtherComponent/>
</SomeComponent>
```
Who's the Server Component?
```tsx
<SomeComponent>
<OtherComponent/>
</SomeComponent>
```
Who's the Server Component?
I really think the directives approach is the best one in terms of composition and splitting bundle.
But there's a subject that i think everyone has been annoying about
By looking a React tree, how do you know what components are client or server or hybrid?
Naming?
But there's a subject that i think everyone has been annoying about
By looking a React tree, how do you know what components are client or server or hybrid?
Naming?
May 22, 2025 at 8:15 AM
I really think the directives approach is the best one in terms of composition and splitting bundle.
But there's a subject that i think everyone has been annoying about
By looking a React tree, how do you know what components are client or server or hybrid?
Naming?
But there's a subject that i think everyone has been annoying about
By looking a React tree, how do you know what components are client or server or hybrid?
Naming?
Imagine a LLM tool like v0 or something, that receives your custom Design System, but all components have already integrated the business logic.
From client to server, with React we have it all to solve the problem.
From client to server, with React we have it all to solve the problem.
May 21, 2025 at 9:03 AM
Imagine a LLM tool like v0 or something, that receives your custom Design System, but all components have already integrated the business logic.
From client to server, with React we have it all to solve the problem.
From client to server, with React we have it all to solve the problem.
So I'm imagining this for example.
We build components for the LLM not for ourselves, and then give those tools (LLM + React Components Design System) to the marketing team so they can build whatever they want.
It's exciting!!!
And with RSC, we can give them full stack components!!!
We build components for the LLM not for ourselves, and then give those tools (LLM + React Components Design System) to the marketing team so they can build whatever they want.
It's exciting!!!
And with RSC, we can give them full stack components!!!
May 21, 2025 at 9:01 AM
So I'm imagining this for example.
We build components for the LLM not for ourselves, and then give those tools (LLM + React Components Design System) to the marketing team so they can build whatever they want.
It's exciting!!!
And with RSC, we can give them full stack components!!!
We build components for the LLM not for ourselves, and then give those tools (LLM + React Components Design System) to the marketing team so they can build whatever they want.
It's exciting!!!
And with RSC, we can give them full stack components!!!
I don't think we're going to stop programming any time soon. AI is just not there, can't think at all, can't create new ideas.
But one thing that is possible may be how to build UI interfaces.
>
But one thing that is possible may be how to build UI interfaces.
>
May 21, 2025 at 8:58 AM
I don't think we're going to stop programming any time soon. AI is just not there, can't think at all, can't create new ideas.
But one thing that is possible may be how to build UI interfaces.
>
But one thing that is possible may be how to build UI interfaces.
>
I literally see a lot of hate around RSC without ever understanding them right.
It's a little sad, because they're really powerful.
I guess time will make them understand.
It's a little sad, because they're really powerful.
I guess time will make them understand.
April 13, 2025 at 2:23 AM
I literally see a lot of hate around RSC without ever understanding them right.
It's a little sad, because they're really powerful.
I guess time will make them understand.
It's a little sad, because they're really powerful.
I guess time will make them understand.
It's a cool type, but i don't see a use case for that type 😅
April 8, 2025 at 2:48 PM
It's a cool type, but i don't see a use case for that type 😅
One advantage of this approach it's that it's very simple to do.
Just a class with less than 200 lines of code.
You have fully infered errors and if you use it more in your code, you see a stack of all possible errors in a type safe way.
Yeah, we got typesafe server errors all the way to RQ hooks
Just a class with less than 200 lines of code.
You have fully infered errors and if you use it more in your code, you see a stack of all possible errors in a type safe way.
Yeah, we got typesafe server errors all the way to RQ hooks
March 20, 2025 at 2:02 PM
One advantage of this approach it's that it's very simple to do.
Just a class with less than 200 lines of code.
You have fully infered errors and if you use it more in your code, you see a stack of all possible errors in a type safe way.
Yeah, we got typesafe server errors all the way to RQ hooks
Just a class with less than 200 lines of code.
You have fully infered errors and if you use it more in your code, you see a stack of all possible errors in a type safe way.
Yeah, we got typesafe server errors all the way to RQ hooks
Don't get me wrong, i use JS, well TS.
And i know it's normal to throw errors and then catch them in another layer or not.
But i never meet someone who likes that strategy, that's all.
Usually for my projects i use a custom wrapper, similar to "never-throw" but even more simpler in my opinion.
And i know it's normal to throw errors and then catch them in another layer or not.
But i never meet someone who likes that strategy, that's all.
Usually for my projects i use a custom wrapper, similar to "never-throw" but even more simpler in my opinion.
March 20, 2025 at 2:00 PM
Don't get me wrong, i use JS, well TS.
And i know it's normal to throw errors and then catch them in another layer or not.
But i never meet someone who likes that strategy, that's all.
Usually for my projects i use a custom wrapper, similar to "never-throw" but even more simpler in my opinion.
And i know it's normal to throw errors and then catch them in another layer or not.
But i never meet someone who likes that strategy, that's all.
Usually for my projects i use a custom wrapper, similar to "never-throw" but even more simpler in my opinion.
Sorry, I haven't met anyone who liked the way JS handles errors.
March 20, 2025 at 12:42 PM
Sorry, I haven't met anyone who liked the way JS handles errors.
Is this sarcastic post?
March 20, 2025 at 12:20 PM
Is this sarcastic post?
It can't because there's not much public codebases that use Arktype, so no much data to train the models.
March 17, 2025 at 4:53 PM
It can't because there's not much public codebases that use Arktype, so no much data to train the models.
One question about useEffect cases? How about this? Can i remove the useEffect from this hook?
It's a good pattern to use useEffect in hooks and declare kinds of events so people avoid using it in components?
It's a good pattern to use useEffect in hooks and declare kinds of events so people avoid using it in components?
March 13, 2025 at 8:48 AM
One question about useEffect cases? How about this? Can i remove the useEffect from this hook?
It's a good pattern to use useEffect in hooks and declare kinds of events so people avoid using it in components?
It's a good pattern to use useEffect in hooks and declare kinds of events so people avoid using it in components?
Or maybe it's automatic if all the data inside of the <Suspense> still be in cache?
March 13, 2025 at 5:12 AM
Or maybe it's automatic if all the data inside of the <Suspense> still be in cache?
Interesting, so how i will decide when i want that <Suspense> part to be cached (static) or not (dynamic) in future requests after the first one?
March 13, 2025 at 5:10 AM
Interesting, so how i will decide when i want that <Suspense> part to be cached (static) or not (dynamic) in future requests after the first one?
That's goot to hear. So i've some hope.
Thanks for everything you do.
Thanks for everything you do.
March 13, 2025 at 5:08 AM
That's goot to hear. So i've some hope.
Thanks for everything you do.
Thanks for everything you do.
One question, is there any comparative of popular validation libraries TS performance?
March 13, 2025 at 4:33 AM
One question, is there any comparative of popular validation libraries TS performance?
So I'm trying to find a way to in ISR without statics params, that first render make of use of Streaming, but the following be Static page.
Is this possible? Maybe in the future?
Any idea about this?
Is this possible? Maybe in the future?
Any idea about this?
March 13, 2025 at 4:20 AM
So I'm trying to find a way to in ISR without statics params, that first render make of use of Streaming, but the following be Static page.
Is this possible? Maybe in the future?
Any idea about this?
Is this possible? Maybe in the future?
Any idea about this?
Then next users will receive the a Static page.
But what happenes when that first SSR render is too slow? Like 5 seconds?
If i tried to use instead <Suspense> to make the users move to a new page and see a squeleton or something, now i lost every static benefits.
/continue
But what happenes when that first SSR render is too slow? Like 5 seconds?
If i tried to use instead <Suspense> to make the users move to a new page and see a squeleton or something, now i lost every static benefits.
/continue
March 13, 2025 at 4:18 AM
Then next users will receive the a Static page.
But what happenes when that first SSR render is too slow? Like 5 seconds?
If i tried to use instead <Suspense> to make the users move to a new page and see a squeleton or something, now i lost every static benefits.
/continue
But what happenes when that first SSR render is too slow? Like 5 seconds?
If i tried to use instead <Suspense> to make the users move to a new page and see a squeleton or something, now i lost every static benefits.
/continue