fix(Link): export onNuxtReady from the Vue stubs

[IgorShevchik]

pnpm repl:build has been failing on main since #541, taking the Pages deploy down with it — three commits (89f81a79, dbd5e7dd, 6ae5680d), while ci stayed green the whole time.

[MISSING_EXPORT] "onNuxtReady" is not exported by "../../dist/runtime/vue/stubs/none.js"
   ╭─[ ../../src/runtime/components/Link.vue?vue&type=script&setup=true&lang.ts:9:46 ]

Link.vue imports onNuxtReady from #imports unconditionally and calls it only behind prefetchApi, which is undefined without NuxtLink — so no Vue build ever runs it. An import that is never called still has to resolve, and rolldown fails the whole bundle rather than warning.

The fix

onNuxtReady in src/runtime/vue/stubs/base.ts: runs the callback at the next idle moment, does nothing on the server. Nuxt's own waits for hydration and then for idle; a plain Vue app has no hydration to wait for, so only the idle half is left — and that is the half callers want, since this is how work is kept out of the mount path.

Why ci could not see it

#imports is a Nuxt alias. typecheck, test and build all resolve it against Nuxt's generated types, where every name is present. The stubs are consulted only by a bundle built without Nuxt, which in this repository is pnpm repl:build — a step in deploy.yml and in no other workflow. So a missing export leaves ci green, merges, and takes the deploy down on main.

test/utils/vue-stub-imports.spec.ts closes that gap where it is cheap: it reads every import { … } from '#imports' under src/runtime as text and asserts all three stub entry points (none, vue-router, inertia) cover the names. Read as text on purpose — importing #imports from a test resolves it through Nuxt, which is the resolution this check exists to bypass.

Mutation-checked:

mutation	result
un-export onNuxtReady from base.ts (the regression itself)	red, and names src/runtime/components/Link.vue as the caller
delete export * from './base' in none.ts	red on none only
rename one inherited export (useAppConfig)	red on all three

One of those mutations was invalid on the first attempt and worth recording: commenting the re-export out with // left the test green, because the helper's export \* from '\./base' regex matches inside a comment too. Deleting the line is the mutation that actually removes the behaviour.

Verified against the real criterion, not just the guard: pnpm repl:build fails on main and succeeds here.

Local gate green: lint · typecheck · build (3.87 MB) · test (345 files, 7855 passed, 6 skipped) · test:module · repl:build.

Not fixed here

ci still builds nothing against the stubs, so a repl breakage of a different shape would still reach main the same way. Adding repl:build to ci.yml is the general fix and a separate call — this PR covers the specific failure class cheaply.

🤖 Generated with Claude Code

https://claude.ai/code/session_01JS8ypVfQSFzYVZzkTHhURb

---

Generated by Claude Code