Clement Yan

Just ran into a similar problem. I have done some investigation and determined both issues have the same root cause. A little background: `typeRoots` are effectively implemented by implicitly adding...

> at the moment, browserslist: ['last 1 firefox versions'] has less polyfilling than browserslist: false -- so I was hoping to get to the bottom of why. 🙃 Probably duplicate...

Looks like there is an additional change needed so the `ecmaVersion` and `sourceType` are read from `langaugeOptions` if they are not given in `languageOptions.parserOptions`, [to mimic ESLint's behavior](https://github.com/eslint/eslint/blob/60f6a06f521c514e3834dd9f82821b10c69a5f00/lib/linter/linter.js#L750-L763)

As far as I understand, the playground was essentially just an API (+ Next.js app?) hosted on CodeSandbox? What makes it difficult to maintain? What eventually completely broke it? I...

UX is definitely important for both sherlock and the playground, but I want to get the playground back up first to at least have something. I can get a PoC...

Turns out, the package info page uses some of docusaurus's internal components and is broken after the migration. Marking the PR as draft for now.

In this case, yarn is unable to automatically determine the ESM loader is needed. You need to force it by setting [`pnpEnableEsmLoader`](https://yarnpkg.com/configuration/yarnrc#pnpEnableEsmLoader).

I believe this is fixed via #5600

This happens whenever importing the config crosses from ESM into CJS. [Minimal reproduction](https://stackblitz.com/edit/stackblitz-starters-2phvoo?file=eslint.config.mjs) Looks like a limitation of `bundle-require`

First of all, `YARN_CACHE_FOLDER` has no effect here because the `enableGlobalCache` defaults to `true` since 4.0.0. The main issue here is `RUN --mount` mounts a **build cache**, which means it...