Written from your under the hood note, the untitled one about the build and runtime worlds, the start-up sequence and Nitro. The matching study section is Under the hood.
Nothing in this note contradicts the documentation. Two points worth adding to it:
"I only have access to what the modules before set up" is right, and it is why cross-module work goes into hooks that run later, such as modules:done or components:extend.
Because h3 matches server/ routes before the Vue renderer's catch-all, a server route and a page at the same path never both run: the server route answers.
setup() and in a modules:done hook how many components are registered, with one more module installed after yours.server/routes/about.get.ts next to a pages/about.vue and see which one answers."It loads the config with c12: resolves layers from extends, loads .env, applies environment overrides and merges everything with defu, project first. The result is resolved against Nuxt's schema. Then the hookable Nuxt instance is created and modules are installed in order, internal ones first; each can change options and register hooks, but only sees what earlier modules did, so cross-module work goes into modules:done. Then the app is resolved (plugins, layouts, middleware, pages), templates are generated into .nuxt, Vite builds the client and server bundles, and Nitro bundles the server for its preset."
"h3 is a small HTTP framework built on web standards: event handlers, routing and helpers for requests and responses, so the same code runs on Node, edge workers and serverless. Nitro is the server toolkit around it: it scans server/ for routes, middleware and plugins, adds storage, caching and tasks, and bundles everything for a deployment preset. In Nuxt, a request first goes through the server routes, and anything unmatched falls through to the Vue SSR renderer."
Nothing yet. Notes filed here show up on this page and the section overview.
Best practices
Questions and flashcards from your notes on the Vue docs' Best Practices chapter — shallow reactivity for large data, v-once, v-memo, lazy components and virtual scrolling, focus after route changes, heading levels, landmarks and their roles, aria-hidden, labels and descriptions, the aria-labelledby spelling, URL injection and safe rich text — with a fact check of the notes.
Conrad caching
Preparation for the caching part of the Conrad interview, written from your Conrad Caching note — HTTP cache headers and validation, Cloudflare's cache, Nitro's cache with Redis, and the cache layers of a large e-shop.