Lightnews — Scholar-powered news

Reposted by Seb ⚛️ ThisWeekInReact.com

Seb ⚛️ ThisWeekInReact.com @sebastienlorber.com · 21h

This Week In React 253 - Special React Conf

🍿 Read/subscribe: thisweekinreact.com/newsletter/253

⚛️
- Compiler 1.0
- React Foundation
-
-
- useEffectEvent
-
📱
- RN 0.82
- Hermes V1
- DOM APIs
- Vega OS
- Keyboard Controller
- IAP
- Skia

5 29

Seb ⚛️ ThisWeekInReact.com @sebastienlorber.com · 21h

This Week In React 253 - Special React Conf

🍿 Read/subscribe: thisweekinreact.com/newsletter/253

⚛️
- Compiler 1.0
- React Foundation
-
-
- useEffectEvent
-
📱
- RN 0.82
- Hermes V1
- DOM APIs
- Vega OS
- Keyboard Controller
- IAP
- Skia

5 29

Seb ⚛️ ThisWeekInReact.com @sebastienlorber.com · 1d

Thanks, that makes sense

2

Seb ⚛️ ThisWeekInReact.com @sebastienlorber.com · 1d

But if you include it in the dependency array because yu don't use ESLint or whatever, it does re-run, right?

I mean there's no magic to taint/tag the function and special handling in useEffect?

1 2

Seb ⚛️ ThisWeekInReact.com @sebastienlorber.com · 1d

Or maybe I'm reading your post wrong

I read it this way:
"the Effect doesn't re-run, even if you include the Effect Event fn in the dependency array and it's unstable"

Maybe what I should have understood:
"the Effect doesn't re-run, because you do not include the Effect Event fn in deps"

1

Seb ⚛️ ThisWeekInReact.com @sebastienlorber.com · 1d

"doesn't cause an Effect that contains the function to re-run even though the function is "unstable""

Would be curious to double-check that, because it's not exactly what @danabra.mov said.

bsky.app/profile/dana...

dan @danabra.mov · 7d

not quite although that's close. it's actually "unstable" on purpose so that you don't rely on stable-ness. it *does* see the latest props/state, but the real trick is that the linter will ask you to NOT include it in the dependencies.

2

Seb ⚛️ ThisWeekInReact.com @sebastienlorber.com · 1d

Almost as bad as the iOS wheel 👌😆

2

Seb ⚛️ ThisWeekInReact.com @sebastienlorber.com · 1d

I don't know either, but I'm interested, so I hope you'll have the answer 😅

3

Seb ⚛️ ThisWeekInReact.com @sebastienlorber.com · 1d

Is there a good reason not to?

1 4

Reposted by Seb ⚛️ ThisWeekInReact.com

Seb ⚛️ ThisWeekInReact.com @sebastienlorber.com · 1d

Just announced at React Conf - New React API available in canary:
🔥 <ViewTransition>
🔥 <Fragment ref>

Particularly excited for Fragment Refs, exposing platform-specific APIs such as:
- addEventListener()
- observeUsing()
- focus()
- getClientRects()
- scrollIntoView()

3 5 35

Seb ⚛️ ThisWeekInReact.com @sebastienlorber.com · 1d

Also curious

The code above taken from official docs doesn't really try to avoid unobserve/reobserve (it could use useEffectEvent?) so I guess ref fn could lead to the same result?

1

Seb ⚛️ ThisWeekInReact.com @sebastienlorber.com · 1d

Yes
Yet it has progressed recently from PRs I've seen. I expected it to eventually become canary today, but maybe too soon

1

Seb ⚛️ ThisWeekInReact.com @sebastienlorber.com · 1d

The ref is not an actual Dom node so it doesn't expose platform APIs dirextly

It's probably registering the observer to all child nodes of the fragment.

1

Reposted by Seb ⚛️ ThisWeekInReact.com

react.dev @react.dev · 1d

Introducing the React Foundation

1 15 96

Seb ⚛️ ThisWeekInReact.com @sebastienlorber.com · 1d

This slide summarizes well all the new things in React

And what to look forward to!

Also, React Compiler is now 1.0!

1 5 15

Seb ⚛️ ThisWeekInReact.com @sebastienlorber.com · 1d

Just announced at React Conf - New React API available in canary:
🔥 <ViewTransition>
🔥 <Fragment ref>

Particularly excited for Fragment Refs, exposing platform-specific APIs such as:
- addEventListener()
- observeUsing()
- focus()
- getClientRects()
- scrollIntoView()