perf(ci): parallelize verification via a vp run.tasks graph

Replace ci.yml's five serial steps with a single `vp run ci` whose graph lives
in vite.config.ts (run.tasks). The vp task runner fans out independent branches
concurrently: fmt and lint start immediately while check:type and the test
suites wait on build (their consumers resolve @vue-tui/runtime from the built
dist/*.d.mts). The wall-clock critical path becomes build -> test:pty instead
of the sum of every check.

Measured cold (no task cache, no prebuilt dist — the real CI condition): the
graph completes in ~41s vs ~60s serial, ~32% faster, with build correctly
fanned out before the type and test branches. `vp run --last-details` (and the
run summary) still pinpoints which sub-task failed.

Also set run.cache=false so neither local nor CI verification depends on any
task-cache replay. The serial `ready` script in package.json is kept for simple
local use.

.github/workflows/ci.yml

@@ -39,22 +39,13 @@ jobs:
           node-version: "22.12.0"
           cache: true
 
-      # Steps mirror `vp run ready` as individual, labeled concerns so the run
-      # shows exactly which one failed. Build MUST precede check:type and test:
-      # @vue-tui/runtime exposes no "types" export, so testing / runtime-tests
-      # resolve its types and runtime from the built dist/*.d.mts. Steps in one
-      # job share the filesystem, so the build output persists to later steps.
-      - name: Format
-        run: vp run check:fmt
-
-      - name: Lint
-        run: vp run check:lint
-
-      - name: Build
-        run: vp run build
-
-      - name: Type check
-        run: vp run check:type
-
-      - name: Test
-        run: vp run test
+      # `ci` is the parallel verification graph defined in vite.config.ts
+      # (run.tasks). The vp task runner fans out independent branches
+      # concurrently — fmt and lint start immediately, while check:type and the
+      # test suites wait on build (their consumers resolve @vue-tui/runtime from
+      # the built dist) — so the wall-clock critical path is build -> test:pty
+      # rather than the sum of every check. On a cache miss this finishes in
+      # ~40s vs ~60s run serially. `vp run --last-details` (or the run summary)
+      # shows which sub-task failed when one does.
+      - name: Verify (vp run ci)
+        run: vp run ci

vite.config.ts

@@ -7,6 +7,28 @@ export default defineConfig({
   fmt: {},
   lint: { options: { typeAware: true, typeCheck: false } },
   run: {
-    cache: true,
+    cache: false,
+    // `ci` is the parallel verification graph used by .github/workflows/ci.yml.
+    // The task runner fans out independent branches concurrently; the wall-clock
+    // critical path is build -> test:pty. fmt and lint have no build dependency,
+    // so they run immediately alongside build. check:type and the test suites
+    // need the built dist (@vue-tui/runtime exposes no "types" export — its
+    // consumers resolve types and runtime from dist/*.d.mts), so they depend on
+    // build. The serial `ready` script in package.json stays for simple local use.
+    tasks: {
+      "ci:build": { command: "vp run build" },
+      "ci:fmt": { command: "vp run check:fmt" },
+      "ci:lint": { command: "vp run check:lint" },
+      "ci:type": { command: "vp run check:type", dependsOn: ["ci:build"] },
+      "ci:test:integration": {
+        command: "vp run -r test:integration",
+        dependsOn: ["ci:build"],
+      },
+      "ci:test:pty": { command: "vp run -r test:pty", dependsOn: ["ci:build"] },
+      ci: {
+        command: "echo ci ok",
+        dependsOn: ["ci:fmt", "ci:lint", "ci:type", "ci:test:integration", "ci:test:pty"],
+      },
+    },
   },
 });