Ari Perkkiö

Started to rollback local `vite` commit by commit. Test works after reverting https://github.com/vitejs/vite/pull/11595 (:wave: @bluwy). It seems that entrypoint for Svelte is resolved to `svelte/ssr.mjs` instead of `svelte/index.mjs`. In `[email protected]`...

> I think Svelte team wants Vitest to use browser condition (resolve.conditions in config) here. Using `resolve.conditions` seems to work. Maybe Vitest should expose `conditions` in configuration API, so that...

I tried the reproduction case on Github Codespaces but wasn't able to reproduce the issue. > make some minor changes in the test files Could you write down more specific...

> this happens on my machine only when node is version 20 or 21. I'm unable to reproduce this locally using Node v21 with MacOS.

I can reproduce this issue using the provided `Dockerfile`. Is `fastify` or some of your dependencies using Node's native `fetch`? It seems that as soon as I switch Vitest to...

> I am not using in any way `fetch` from `node`. It's possible that the root cause of https://github.com/nodejs/undici/issues/2026 that makes `node:worker_threads` stuck is not directly `fetch` - it could...

Using just `fastify` seems to be enough to make `node:worker_threads` stuck. Having contents below in two different test files and running them parallel in the docker reproduces the issue. Next...

Did Rollup v4 switch to SWC's parser? - https://github.com/swc-project/swc/issues/8462 - https://github.com/swc-project/swc/blob/9ac304be9e3948e2d723afb80a8ff6cc9811d1a9/crates/swc_ecma_parser/src/parser/expr.rs#L343

I'm planning to focus on this next. Starting probably this or next week. Before starting we'll need a smaller **minimal reproduction setup** with clear goals of what should be seen...

There's now a minimal reproduction setup at https://github.com/AriPerkkio/reproductions, branch `vitest-changed-coverage`. Make changes to source files and run `pnpm run coverage --changed=HEAD` to see the issue. There's also branch `jest-changed-coverage` which...