Cloudflare Tunnel | React Compiler | refs during render

Ryan talks to Sam about reproducing iOS's new image background treatment for his Open Graph Preview tool, opengraph.ing. They talk about different approaches for generating gradients from images, including finding the vibrant color of an image, luminosity-weighted averages, k-means clustering, and more.Timestamps:0:00 - Intro3:07 - Apple's new OG image gradient10:06 - Luminosity-weighted average14:22 - Finding the vibrant color of an image21:41 - Contrast ratios on favicons32:21 - K-means clustering41:25 - Refactoring UI tip about rotating the hueLinks:Open Graph PreviewRefactoring UI

Sam and Ryan talk about React 19's useActionState hook. They discuss how adding async functions to a plain React app introduces lots of in-between states that developers must grapple with, and how useActionState allows React to collapse and eliminate these states, bringing the simplicity of React's sync mental model to our async code.Timestamps:0:00 - Intro1:51 - How React normally eliminates state in synchronous apps8:20 - How useActionState lets React eliminate state in asynchronous apps18:17 - Why you shouldn't just pass server actions into useActionState23:00 - TCP/IP and UDP analogy26:39 - Thinking of useActionState like enqueue34:55 - Why the term "reducer" is too loaded for best understanding useActionState51:07 - How useActionState helps you build a Todo app that stays responsive during pending actions

Sam and Ryan talk about using Cloudflare Tunnel for local development, the new React Compiler beta release, and why reading or writing refs during render violates the rules of React.Timestamps:0:00 - Intro1:42 - Cloudflare Tunnel7:06 - React Compiler14:21 - Reading or writing refs during renderLinks:Cloudflare TunnelReact Compiler Beta releaseReact Docs Pitfall on refs

Sam and Ryan talk about building a useAnimatedText hook that can animate streaming text. They also discuss how React code that uses state changes to approximate events can be simplified, and the benefits of having escape hatches when building UI with Catalyst.Timestamps:0:00 - Intro1:22 - Catalyst and escape hatches16:03 - Building a useAnimatedText hook (aka useBufferedText)54:10 - Avoiding using state changes to approximate eventsLinks:CatalystSam's Tweet about useBufferedText and AnimatedScrolleruseAnimatedText recipeIntl.Segmenter on MDNDavid K's Tweet on eventsRicky's Tweet on Intersection ObserverBuild UI's new course

Sam and Ryan talk about how frameworks and infrastructure evolve with each other, using Next.js as a representative example. They discuss how hosting providers like Heroku have always imposed certain constraints on apps, what features those constraints enable hosting providers to support, how burdensome those constraints are across different frameworks, and how frameworks that add infra-specific APIs can best communicate the costs of those APIs and benefits they enable.Timestamps:0:00 - Intro3:03 - Heroku and the Twelve-Factor App7:39 - GitHub Pages and static sites13:57 - Serverless and JAMstack17:30 - Vercel and CDNs, self-hosting, and Next.js19:00 - How framework APIs can nudge an app towards a particular hosting solution23:09 - What constraints does Next.js impose on your app (e.g. middleware doesn't run node), and what benefits do those constraints give you?36:13 - How Next.js APIs are motivated by wanting to tease apart static and dynamic code, in an attempt to support the needs of any web app with a single stack40:33 - What is the relationship between frameworks and infra?47:37 - How can frameworks that add infra-specific APIs best communicate the costs of those APIs and the benefits they enable?Links:The Twelve-Factor App

Tom Occhino, Chief Product Officer at Vercel and former Engineering Director at Facebook, joins Sam to talk about the pivotal moments in React's history. He talks about how React popularized the ideas of declarative rendering and unidirectional data flow, how GraphQL furthered React's goal of co-locating all the concerns of a particular piece of UI, the problems that GraphQL led to at Facebook and how Relay solved them, and how Suspense, Server Components, and PPR are the generalized spiritual successors to the stack used at Facebook.Timestamps:0:00 - Intro2:53 - Declarative rendering as React's legacy8:12 - How GraphQL enabled complex components to be fully self-contained20:12 - How React's goal has always been to co-locate all the concerns of a particular piece of UI22:58 - The problem with co-locating GraphQL with components, and how Relay solved it26:28 - How RSC is the generalized spiritual successor to BigPipe and GraphQL34:46 - What PPR is, and how it and Suspense fit into this story55:55 - The general paradigm shift of getting static code to the device as soon as possibleLinks:Tom Occhino with Ben DunphyReact: The DocumentaryReact Roundtable with Andrew Clark and Sebastian MarkbågeTom Occhino on Twitter

Sam and Ryan talk about render props in React. They discuss where they came from, how Hooks superseded them for sharing stateful logic, how data attributes compare to them for customizing styling, and how for certain complex components like forms they're still a great solution for accessing slices of internal state.Timestamps:0:00 - Intro3:40 - Where did render props come from?6:01 - How hooks replaced many use cases for render props8:14 - Using render props for custom styling10:32 - Data attributes vs. render props for custom styling16:43 - Using render props to peek into an uncontrolled component's internal state21:05 - Forms example and using render props for a smaller public API24:10 - React docs mention of render props25:48 - Where render props shineLinks:Headless UI MenuRadix FormsReact Aria FormsReact docs - Render propsOur upcoming course on React Component PatternsEmail SamEmail Ryan

Sam and Ryan discuss controlled and uncontrolled components in React. They talk about how uncontrolled components can be thought of as components that manage their own internal state, why you should model your complex React components after the simpler APIs of native HTML elements like inputs, why you shouldn't try to make components that are both controlled and uncontrolled, and why making a new component boundary is sometimes all you need to make your custom components behave more predictably.Timestamps:0:00 - Intro1:41 - What are controlled and uncontrolled components?6:11 - How to change a component from uncontrolled to controlled8:48 - How do you decide when to use a controlled or uncontrolled component?12:00 - Sortable table example and a single source of truth15:27 - Is it always either controlled or uncontrolled?21:09 - Color picker example and not exposing internal state28:46 - Sortable list example with Framer Motion39:45 - Component boundaries and wearing two hats: the library author vs. library consumer41:43 - How do you know if you are using the wrong approach?Links:New course: Advanced React Component PatternsReact docs: Controlled and uncontrolled componentsReact docs: Storing state from previous rendersEmail SamEmail RyanFrontend First on Twitter

Sam and Ryan talk about the pattern of building unstyled components with React. They discuss why unstyled components were created, how they improve upon composition patterns from UI libraries like Bootstrap, how they can be used to share behavior and logic without prescribing any styling opinions, and how they fit into a larger collection of React patterns that can be used to build more powerful components that avoid premature abstractions.Timestamps:0:00 - Intro1:36 - What are unstyled components?2:24 - How do unstyled components improve upon earlier patterns?6:44 - Why would you want to use an unstyled component?9:58 - How can you compose an unstyled component with a styled component?16:41 - How to decide which pattern is best suited for the code you want to share19:36 - Using patterns that preserve React's locality of behavior24:49 - How do you know if an abstraction is good?32:54 - Build UI's upcoming course on Advanced React Component PatternsLinks:Build UI's newest course: Advanced React Component PatternsSam's YouTube video on Recursive ComponentsEmail SamEmail Ryan

Sam and Ryan talk about what sorts of capabilities a tool should have to be considered a web framework. They discuss how frameworks tackle the complexity of getting different systems to communicate with each other, how good frameworks embrace the strengths and patterns of the language they're written in, and why frameworks and services are not in opposition to each other.Timestamps:0:00 - Intro3:58 - Adapter pattern and cohesive boundaries9:43 - Rails is Omakase13:47 - Configurable, but still cohesive17:04 - Frontend frameworks try to “work with everything”21:42 - Does composition mean a React framework will look different than Rails?29:29 - Why taste still matters34:20 - A framework is a shell of adapters and a brain that coordinates them35:16 - When using services, complexity still exists in the in-between37:59 - A fullstack dev is someone who acknowledges and understands how all the parts come together44:06 - Tweets about the hard problems that Laravel tackles, and the deep design it took to get there49:15 - Frameworks should embrace the strengths and patterns of their language and ecosystem50:35 - Why RSCs and Server Actions mean the “Rails for JS” may end up looking nothing like Rails52:11 - Why users of a “fullstack framework” shouldn’t even care about where the code is running55:31 - Why libraries or services that are easy to install and set up are not a replacement for frameworksLinks:Sam's BigSkyDevCon talkRails is omakasePovilas' Laravel tweetMary's Laravel tweetBuild UI's upcoming React Component Patterns courseEmail SamEmail Ryan