Understanding Yarn Version Management: Comparing yarn set version berry and corepack prepare yarn@stable --activate

[wojtekKrol]

Previously, I had Yarn v1.22.x installed. I completely removed it by running rm -rf ~/.yarn and removing it from nvm node versions. Then, I followed the installation guide from https://yarnpkg.com/getting-started/install. When I ran corepack prepare yarn@stable --activate, it installed Yarn v3.5.0 globally in $HOME/.nvm/versions/node/v18.16.0/bin/yarn.

By using yarn set version berry, Yarn will use the locally installed version of Yarn within the context of the current project. Whereas, corepack prepare yarn@stable --activate prepares Corepack to use the latest stable version of Yarn at the package manager level. Both of these commands achieve a similar result, but they work at different levels of the project.

To clarify, yarn set version berry sets the Yarn version to "berry" within the project's context. It uses the version specified in $PROJECT_DIR/.yarn/releases/yarn-3.5.0.cjs, as defined in package.json using the "packageManager": "yarn@3.5.0" configuration - I suppose?

Now, I am wondering if I need to run yarn set version berry in every new project if I have the latest version of Yarn installed globally?

Because I think that I don't have to.
If I have the latest version of Yarn installed globally, then I do not need to run yarn set version berry to switch to Yarn's Berry. Berry is simply a different version of Yarn with significant changes to the way packages are managed. I don't have any project with yarn at all so it won't affect anything.