Sebastian Silbermann
@sebbie.dev
nextjs @vercel.com, @react.dev at night, @testing-library.com core | testing + a11y first | he/him
sounds like a rewrite not migration
October 9, 2025 at 9:39 PM
sounds like a rewrite not migration
Do you have a PR to share?
October 9, 2025 at 9:29 PM
Do you have a PR to share?
Reposted by Sebastian Silbermann
It's the "I don't have any talks to give so I better make sure the stuff other people talk about actually works"
September 30, 2025 at 10:19 PM
It's the "I don't have any talks to give so I better make sure the stuff other people talk about actually works"
Is the Transition completing in a microtask? iirc that would be batched by `act()`. If it takes at least a macrotask, you'd just assert on the pending state immediately and then `waitFor()` the resolved state.
September 9, 2025 at 5:35 PM
Is the Transition completing in a microtask? iirc that would be batched by `act()`. If it takes at least a macrotask, you'd just assert on the pending state immediately and then `waitFor()` the resolved state.
I imagine you'd show key and params then
September 3, 2025 at 1:05 PM
I imagine you'd show key and params then
Probably i18n keys. Can be annoying to update tests every time copy changes. You need to trust your i18n implementation for that though.
September 3, 2025 at 11:09 AM
Probably i18n keys. Can be annoying to update tests every time copy changes. You need to trust your i18n implementation for that though.
Actually didn't know. That feels like a spec bug to me. Probably means JS heavy apps need to trigger two paints for live regions. One with no content and one with the initial content. What's the rationale for not announcing the initial content when a live region is injected?
September 3, 2025 at 9:32 AM
Actually didn't know. That feels like a spec bug to me. Probably means JS heavy apps need to trigger two paints for live regions. One with no content and one with the initial content. What's the rationale for not announcing the initial content when a live region is injected?
> it’s possible Svelte’s … “unique choices” have made LLM support more difficult
Shouldn't that be a red-flag for the current state of LLMs? Is the future really just doing what everybody else is doing?
Shouldn't that be a red-flag for the current state of LLMs? Is the future really just doing what everybody else is doing?
September 1, 2025 at 9:06 AM
> it’s possible Svelte’s … “unique choices” have made LLM support more difficult
Shouldn't that be a red-flag for the current state of LLMs? Is the future really just doing what everybody else is doing?
Shouldn't that be a red-flag for the current state of LLMs? Is the future really just doing what everybody else is doing?