@rangojs/router 0.0.0-experimental.35 → 0.0.0-experimental.358c376d

package/skills/links/SKILL.md

@@ -1,16 +1,25 @@
 ---
 name: links
-description: URL generation with ctx.reverse (server), href (client), useHref (mounted), useMount, and scopedReverse
-argument-hint: [href|useHref|useMount|scopedReverse]
+description: URL generation with ctx.reverse (server default), href (client), useHref (mounted), useMount, useReverse, and scopedReverse. Use when generating a link to a route by name instead of hardcoding a path, or a link breaks after routes move or get mounted elsewhere.
+argument-hint: [ctx.reverse|href|useHref|useMount|useReverse|scopedReverse]
 ---
 
 # Links & URL Generation
 
 @rangojs/router provides different href APIs for server and client contexts.
 
+**Default server API: `ctx.reverse()`.** Generate URLs from the handler context — it's typed, auto-fills mount params, and resolves local (`.name`) and absolute (`name.sub`) names.
+
+**On the client, two patterns:**
+
+1. **Receive URLs as props / loader data / action return.** The default. The server has the full route manifest and handler context — generate URLs there and hand strings to client components.
+2. **`useReverse(routes)`.** Import a generated `routes` map from a `urls()` module's `.gen.ts` and call `reverse("name", params?)` (the leading dot is optional). Mount-aware via `useMount()`, auto-fills params from `useParams()`, fully typed from the imported map. Use this when a client component needs to generate URLs into a known module without round-tripping through the server.
+
+`ctx.reverse()` itself is **server-only** — it depends on the full route manifest and handler context. Client components never import or call it.
+
 ## Server: ctx.reverse()
 
-Available in route handlers via HandlerContext. Resolves named routes using the full route map.
+Available in route handlers via HandlerContext. Resolves named routes using the full route map. This is the default way to generate URLs on the server.
 
 ```typescript
 import { urls, scopedReverse } from "@rangojs/router";
@@ -29,6 +38,18 @@ export const shopPatterns = urls(({ path, layout }) => [
 - **`.name`** — local route, resolved within the current `include()` scope
 - **`name`** — global route, from the named-routes definition
 
+The include site decides which global names exist:
+
+| Include site                                       | App-wide names                                     |
+| -------------------------------------------------- | -------------------------------------------------- |
+| `include("/shop", shopPatterns)`                   | none from the child; use `.name` inside the module |
+| `include("/shop", shopPatterns, { name: "shop" })` | `shop.index`, `shop.product`, ...                  |
+| `include("/shop", shopPatterns, { name: "" })`     | `index`, `product`, ... at the parent scope        |
+
+Flatten only when those names are intentionally unique across the whole parent
+map. Private includes are useful for reusable modules and still support
+mount-aware dot-local reversal.
+
 ```typescript
 // Inside a handler within shopPatterns (mounted at /shop)
 path("/product/:slug", (ctx) => {
@@ -103,7 +124,7 @@ path("/search", (ctx) => {
 
 ### scopedReverse() - type-safe ctx.reverse
 
-Wraps `ctx.reverse` with local route type information for autocomplete and validation:
+Wraps `ctx.reverse` with local route type information for autocomplete and validation. Runtime behavior is identical to `ctx.reverse` — `scopedReverse` is a type-only cast. The same dot-prefix rule applies: local names use `.name`, global names use `name.sub`.
 
 ```typescript
 import { scopedReverse } from "@rangojs/router";
@@ -111,18 +132,83 @@ import { scopedReverse } from "@rangojs/router";
 path("/product/:slug", (ctx) => {
   const reverse = scopedReverse<typeof shopPatterns>(ctx.reverse);
 
-  reverse("cart");                        // Type-safe local name
-  reverse("product", { slug: "widget" }); // Type-safe with params
-  reverse("blog.post");                   // Absolute names (dot notation) always allowed
-  reverse("/about");                      // Path-based always allowed
+  reverse(".cart");                        // Local name (dot-prefixed) — resolves in include scope
+  reverse(".product", { slug: "widget" }); // Local name with params
+  reverse("blog.post", { slug: "hi" });    // Global name (dotted) — full route map
 
   return <ProductPage slug={ctx.params.slug} />;
 }, { name: "product" })
 ```
 
+`reverse()` does not accept raw path strings (`"/about"`). For static paths in client components, use `href("/about")`; on the server, look up the route by name.
+
+## Client components: receive URLs as props
+
+`ctx.reverse()` is not available inside `"use client"` modules — there is no handler context in the browser bundle. For in-module names, prefer `useReverse(routes)` (see below) and import the relevant `urls/*.gen.js`. For cross-module URLs or one-off names, generate the URL on the server and hand it to the client component using one of these three patterns:
+
+1. Pass as a prop from a server component:
+
+```tsx
+// server
+function BlogPostPage(ctx: HandlerContext) {
+  return <ShareButton url={ctx.reverse(".post", { slug: ctx.params.slug })} />;
+}
+```
+
+```tsx
+"use client";
+
+export function ShareButton({ url }: { url: string }) {
+  return (
+    <button onClick={() => navigator.clipboard.writeText(url)}>Share</button>
+  );
+}
+```
+
+2. Return from a loader (attached to the route via the DSL):
+
+```tsx
+// server — loaders/nav.ts
+export const NavLoader = createLoader((ctx) => ({
+  home: ctx.reverse("home"),
+  blog: ctx.reverse("blog.index"),
+}));
+
+// server — urls.tsx: attach the loader so useLoader has data in context
+const urlpatterns = urls(({ path, loader }) => [
+  path("/", HomePage, { name: "home" }, () => [loader(NavLoader)]),
+]);
+```
+
+```tsx
+"use client";
+
+function Nav() {
+  const { data } = useLoader(NavLoader);
+  return <Link to={data.home}>Home</Link>;
+}
+```
+
+`useLoader()` requires the loader to be attached to an active route. If you need on-demand fetching instead, use `useFetchLoader()`.
+
+3. Return from a server action:
+
+```tsx
+"use server";
+
+export async function getProductUrl(slug: string) {
+  const ctx = getRequestContext();
+  return ctx.reverse("product", { slug });
+}
+```
+
+See `/server-actions` for the full action surface (`getRequestContext()` is the same context middleware and handlers use).
+
+For static path strings (not named routes), client components can use `href()` — see below.
+
 ## Client: href()
 
-Plain function for absolute path-based URLs. No hook needed - works anywhere.
+Plain function for absolute path-based URLs. No hook needed - works anywhere in client components. `href()` validates paths at compile time, but does **not** resolve named routes — for named routes, use one of the patterns above.
 
 ```typescript
 "use client";
@@ -139,7 +225,19 @@ function GlobalNav() {
 }
 ```
 
-`href()` is an identity function at runtime but provides compile-time validation via `ValidPaths` type. Paths are validated against registered route patterns using `PatternToPath`.
+`href()` provides compile-time validation via the `Rango.Path` type. Paths are validated against registered route patterns using `PatternToPath`.
+
+When wrapping `href()`, type the wrapper's path parameter as `Rango.Path` so it
+keeps the same generated-route validation. `Rango.Path` is ambient — no import,
+just like `Rango.Env` / `Rango.Vars`:
+
+```typescript
+import { href } from "@rangojs/router/client";
+
+export const appHref = (path: Rango.Path): string => href(path);
+```
+
+`href()` is a raw path helper — it is **not** basename-aware. It returns the path as-is (or with the include mount prefix via `useHref()`). For basename-aware navigation, use `Link`, `useRouter().push()`, or `reverse()`, which auto-prefix root-relative paths with the router's basename.
 
 ## Client: useHref()
 
@@ -183,15 +281,185 @@ function MountInfo() {
 
 `useMount()` reads from `MountContext`, which is automatically set by `include()` in the segment tree.
 
+## Client: useReverse(routes)
+
+Hook that returns a typed local reverse function for a `routes` map imported from a generated `.gen.ts` next to a `urls()` module. The route map is the **exposure boundary** — `useReverse` only knows about names in that map, never the full app manifest.
+
+> **Which map?** `useReverse` accepts any routes map. Prefer the per-module `routes` (e.g. `urls/blog.gen.ts`): it gives **mount-aware** local `.name` reverse (auto-prefixes the `include()` mount) and only that module's names enter the client bundle. You _can_ instead pass `router.named-routes.gen.ts` (`NamedRoutes`) for **global** names (`blog.post`; the leading dot is optional) — it is a plain importable map and works on the client (it is **not** server-only) — but its paths are **absolute** while `useReverse` mount-prefixes, so it is correct only at the root mount (under a non-root mount it double-prefixes), and importing it pulls every route name and pattern in the app into the client bundle (a small names-to-paths map — not components or loaders), versus the per-module map which exposes only one module's names. So the per-module map is preferred for in-module links; the named-routes map is the escape hatch for global names.
+
+```tsx
+"use client";
+import { Link, useReverse } from "@rangojs/router/client";
+import { routes as blogRoutes } from "../urls/blog.gen.js";
+
+export function BlogNav() {
+  const reverse = useReverse(blogRoutes);
+
+  return (
+    <nav>
+      <Link to={reverse("index")}>Blog</Link>
+      <Link to={reverse("post", { postId: "hello" })}>Post</Link>
+    </nav>
+  );
+}
+```
+
+### How it resolves
+
+1. Strips an optional leading `.` and looks up the name in the imported `routes` map.
+2. Joins the local pattern with the surrounding `useMount()` value — the include's URL pattern.
+3. Substitutes params: explicit params from the call, then auto-filled from `useParams()` for anything still unresolved (mount params like `:tenantId` flow in this way).
+4. Appends a query string if a search object is passed and the route has a `search` schema.
+
+### Mount-relativity
+
+Patterns in the generated `routes` map are **mount-relative** — they're the patterns as defined inside the `urls()` module, _not_ the full app paths. Mount-joining happens at runtime via `useMount()`, so the same component works under any include:
+
+```typescript
+// urls/blog.tsx
+export const blogPatterns = urls(({ path }) => [
+  path("/", BlogIndex, { name: "index" }),
+  path("/:postId", BlogPost, { name: "post" }),
+]);
+
+// Generated urls/blog.gen.ts
+// export const routes = { index: "/", post: "/:postId" } as const;
+
+// urls.tsx — same module mounted twice
+include("/news", blogPatterns, { name: "news" }),    // <BlogNav> renders /news, /news/hello
+include("/journal", blogPatterns, { name: "diary" }), // <BlogNav> renders /journal, /journal/hello
+```
+
+The `/` pattern under a non-root mount collapses cleanly: under `/news`, `reverse(".index")` returns `/news` (no trailing slash), matching `ctx.reverse(".index")` on the server.
+
+### Auto-filled params (mount params)
+
+When the include itself carries `:params`, those are auto-filled from `useParams()` so the caller doesn't have to thread them through:
+
+```typescript
+// urls.tsx
+include("/tenant/:tenantId", clientReversePatterns, { name: "tenant" });
+```
+
+```tsx
+// At /tenant/acme/posts/p1, useParams() = { tenantId: "acme", postId: "p1" }
+const reverse = useReverse(clientReverseRoutes);
+
+reverse(".index"); // "/tenant/acme"
+reverse(".post", { postId: "p2" }); // "/tenant/acme/posts/p2"   (tenantId auto-filled)
+reverse(".post", { tenantId: "other", postId: "p2" }); // "/tenant/other/posts/p2" (explicit override)
+```
+
+Auto-fill follows soft navigation — when the matched route changes, `useReverse` re-renders with the new params.
+
+### Search schemas
+
+Routes declared with a `search` schema accept a typed search object as the third argument:
+
+```typescript
+// urls/blog.tsx
+path("/search", SearchPage, {
+  name: "search",
+  search: { q: "string", page: "number?" },
+}),
+
+// Generated as: search: { path: "/search", search: { q: "string", page: "number?" } }
+```
+
+```tsx
+const reverse = useReverse(blogRoutes);
+reverse(".search", {}, { q: "hello world", page: 2 });
+// "/news/search?q=hello%20world&page=2"
+```
+
+### Errors
+
+- Unknown name: throws `Unknown route: ".not-a-route"`.
+- Missing required param: throws `Missing param "postId" for route ".detail"`.
+
+Both happen synchronously during `reverse()` — wrap calls in try/catch (or an ErrorBoundary if the throw happens during render) when you need to surface them as UI.
+
+### The leading dot is optional
+
+`reverse("post")` and `reverse(".post")` resolve **identically** — the leading dot is cosmetic. The map you import IS the scope, so there is no separate global namespace to disambiguate and the dot carries no meaning; it exists only as a readability convention and for parity with `ctx.reverse(".name")` on the server. To link into a different module, import that module's `routes`:
+
+```tsx
+import { routes as blogRoutes } from "../urls/blog.gen.js";
+import { routes as shopRoutes } from "../urls/shop.gen.js";
+
+function CrossNav() {
+  const blog = useReverse(blogRoutes);
+  const shop = useReverse(shopRoutes);
+  return (
+    <nav>
+      <Link to={blog("index")}>Blog</Link>
+      <Link to={shop("cart")}>Cart</Link>
+    </nav>
+  );
+}
+```
+
+### Codegen
+
+Each `urls()` module gets a sibling `.gen.ts` with the local route names and patterns, produced by `rango generate`:
+
+```bash
+pnpm exec rango generate src/urls/blog.tsx
+# or generate everything under a directory:
+pnpm exec rango generate src/urls --static
+```
+
+Don't edit the file by hand — re-run codegen when patterns change.
+
+**Today the Vite plugin only regenerates the router-level `*.named-routes.gen.ts`.** Per-module `urls/*.gen.ts` files are emitted only by the CLI (or `writePerModuleRouteTypesForFile` programmatically). Commit the generated files and re-run `rango generate` whenever a `urls()` module's `path()`/`include()` shape changes. A common workflow is to wire it into a `predev` script:
+
+```jsonc
+// package.json
+{
+  "scripts": {
+    "predev": "rango generate src",
+    "dev": "vite",
+  },
+}
+```
+
+## Prefetch boundaries
+
+`Link` follows its `prefetch` prop or the router's `defaultPrefetch`; eligible
+plain anchors follow the router default. When a whole DOM section must never
+speculate, mark its container instead of repeating per-link opt-outs:
+
+```tsx
+<section data-prefetch-scope="none">
+  <Link to="/activity" prefetch="viewport">
+    Activity
+  </Link>
+  <a href="/export">Export</a>
+</section>
+```
+
+The scope is a hard opt-out for every descendant Link and plain anchor; `"false"`
+and `"none"` are equivalent. An explicit `prefetch` prop or
+`data-prefetch="true"` cannot override it, and navigation still works normally.
+Use `data-prefetch="false"` or `"none"` directly on a plain anchor when only that
+anchor is unsafe. Dynamic scope changes re-evaluate only descendant anchors and
+mounted viewport/render Links. Links share one document-level scope observer,
+not one observer per Link or container. Adding a scope cannot recall work already
+queued or in flight; removing it re-arms both Link types.
+
 ## When to use what
 
-| Context          | API                             | Resolves                        | Use for                             |
-| ---------------- | ------------------------------- | ------------------------------- | ----------------------------------- |
-| Server handler   | `ctx.reverse("name")`           | Named routes (local + absolute) | Server-side URL generation          |
-| Server handler   | `scopedReverse<T>(ctx.reverse)` | Same, with type safety          | Type-safe server URLs               |
-| Client component | `href("/path")`                 | Absolute paths                  | Global navigation                   |
-| Client component | `useHref()`                     | Mount-prefixed paths            | Local navigation inside `include()` |
-| Client component | `useMount()`                    | Raw mount path                  | Custom mount-aware logic            |
+| Context          | API                                                | Resolves                                  | Use for                                                          |
+| ---------------- | -------------------------------------------------- | ----------------------------------------- | ---------------------------------------------------------------- |
+| Server handler   | `ctx.reverse("name")`                              | Named routes (local + absolute)           | **Default** server-side URL generation                           |
+| Server handler   | `scopedReverse<T>(ctx.reverse)`                    | Same, with type safety                    | Type-safe server URLs                                            |
+| Client component | `useReverse(routes)`                               | Local names from an imported `routes` map | Typed in-module URL generation without round-tripping the server |
+| Client component | (URL passed as prop / loader data / action return) | Named routes                              | Cross-module URLs or one-off names you don't want to import      |
+| Client component | `href("/path")`                                    | Absolute paths (static strings)           | Static navigation where no named-route lookup is needed          |
+| Client component | `useHref()`                                        | Mount-prefixed paths                      | Local navigation inside `include()`                              |
+| Client component | `useMount()`                                       | Raw mount path                            | Custom mount-aware logic                                         |
+
+> `ctx.reverse()` is server-only. On the client, either generate URLs on the server and pass them in, or import the `routes` map and use `useReverse(routes)` for in-module names.
 
 ## Complete example: mounted module