Repository navigation
Define precise disableCache trigger points for uncacheable operations #482
Description
Problem
disableCache() currently has an imprecise contract in practice. Tool integrations can call it while configuring a server because the configuration may later lead to an uncacheable operation. That is too early.
The observed failure mode is Vite/Vitest: a task can create a Vite server for transforms or test execution without actually listening on a port or watching the filesystem. If disableCache() fires during server/config setup, cacheable tasks such as vitest run are reported as Not cached: the task opted out of caching even though no uncacheable operation happened.
Principle
disableCache() should be called immediately before the concrete operation that makes the task uncacheable, not when reading configuration that might eventually lead to such an operation.
Concrete examples:
- Listening on a port should disable caching immediately before the listen/bind operation.
- Observing the filesystem should disable caching immediately before the watcher starts observing paths.
- Creating/configuring an object that might later listen or watch should not disable caching by itself.
Temporary workaround
PR #481 makes vite_task_client::Client::disable_cache() a no-op so false opt-outs stop affecting downstream task caching while the real fix is designed and shipped.
Real solution
Move the Vite/Vitest integration points to operation boundaries:
- In Vite, call
disableCache()from the HTTP server start/listen path immediately before binding a port. - In Vite, call
disableCache()from the watcher start path immediately before filesystem observation begins. - Do not call
disableCache()from generic server creation/config resolution paths. - Verify Vitest
runwithout watch/API does not listen or watch and therefore remains cacheable. - Verify Vitest API/server mode and watch mode still opt out because they perform uncacheable operations.
After those semantics are implemented and covered, revert the vite-task client no-op workaround and restore disableCache() as an effective client request.
Activity
fix: ignore disableCache in client
jasonkuhrt commentedon Oct 24, 2026
A use case for the redesign, from a tool that knows the exact operation that makes its run uncacheable.
We run Playwright suites as Vite Task tasks with a ratchet: a reporter skips the cases earlier runs on the branch already passed, so a re-push runs only what is still owed. A run that skipped cases exits 0 but is not a pass of the whole suite, so it must never be stored: a later cache hit would report every case passed on code where most of them never ran. A run that ran every case is an ordinary pass and should be stored and replayed.
The reporter knows this at one point, in Playwright's
onBegin, where it callstestRun.excludeon the settled cases. That fits the principle here: disable caching immediately before the operation that makes the run uncacheable.With
disable_cachea no-op since #481, there is no way to say it, so these tasks arecache: falseeverywhere and can't replay a full pass either. Could the redesign keep a way for a process inside the task (here a reporter two processes below the task's command) to mark the current run as not to be stored, leaving lookup as it is? Happy to test a branch.