🪚 Use pnpm instead of yarn (LayerZero-Labs#150)

Author: janjakubnanista

.github/workflows/reusable-publish.yaml

@@ -24,21 +24,27 @@ jobs:
     steps:
       - uses: actions/checkout@v3
 
+      - uses: pnpm/action-setup@v2
+        name: Install pnpm
+        with:
+          version: 8
+          run_install: false
+
       - name: Setup Node.js
         uses: actions/setup-node@v3
         with:
           node-version-file: ".nvmrc"
-          cache: "yarn"
+          cache: "pnpm"
 
-      # Install the dependencies without running the post-install scripts
-      - name: Install Dependencies
-        run: yarn install --frozen-lockfile --prefer-offline --ignore-scripts
+      # Fetch the dependencies without running the post-install scripts
+      - name: Fetch Dependencies
+        run: pnpm fetch --frozen-lockfile --prefer-offline --ignore-scripts
         env:
           NPM_TOKEN: ${{ secrets.NPM_TOKEN }}
 
-      # Run the post-install scripts
-      - name: Build Dependencies
-        run: yarn install --frozen-lockfile --offline
+      # Install the dependencies and run the post-install scripts
+      - name: Install Dependencies
+        run: pnpm install --frozen-lockfile --offline
 
       # Cache build artifacts from turbo
       #
@@ -64,7 +70,7 @@ jobs:
             ${{ runner.os }}-hardhat-
 
       - name: Build
-        run: yarn build
+        run: pnpm build
         env:
           NODE_ENV: production
 
@@ -82,8 +88,8 @@ jobs:
       - name: Publish packages / create version bump PRs
         uses: changesets/action@v1
         with:
-          version: yarn changeset version
-          publish: yarn changeset publish
+          version: pnpm changeset version
+          publish: pnpm changeset publish
           title: "🚀 Version packages"
 
         env:

.github/workflows/reusable-test.yaml

@@ -24,21 +24,27 @@ jobs:
     steps:
       - uses: actions/checkout@v3
 
+      - uses: pnpm/action-setup@v2
+        name: Install pnpm
+        with:
+          version: 8
+          run_install: false
+
       - name: Setup Node.js
         uses: actions/setup-node@v3
         with:
           node-version-file: ".nvmrc"
-          cache: "yarn"
+          cache: "pnpm"
 
-      # Install the dependencies without running the post-install scripts
-      - name: Install Dependencies
-        run: yarn install --frozen-lockfile --prefer-offline --ignore-scripts
+      # Fetch the dependencies without running the post-install scripts
+      - name: Fetch Dependencies
+        run: pnpm fetch --frozen-lockfile --prefer-offline --ignore-scripts
         env:
           NPM_TOKEN: ${{ secrets.NPM_TOKEN }}
 
-      # Run the post-install scripts
-      - name: Build Dependencies
-        run: yarn install --frozen-lockfile --offline
+      # Install the dependencies and run the post-install scripts
+      - name: Install Dependencies
+        run: pnpm install --frozen-lockfile --offline
 
       # Cache build artifacts from turbo
       #
@@ -64,13 +70,13 @@ jobs:
             ${{ runner.os }}-hardhat-
 
       - name: Lint
-        run: yarn lint
+        run: pnpm lint
 
       - name: Build
-        run: yarn build
+        run: pnpm build
 
       - name: Test
-        run: yarn test
+        run: pnpm test
         # Since we're running the E2E tests in docker, we'll need to reinstall
         # the node modules. This is a temporary solution to bridge us between
         # non-containerized and containerized development

.gitignore

@@ -1,8 +1,8 @@
 # Logs
 logs
 *.log
-yarn-debug.log*
-yarn-error.log*
+pnpm-debug.log*
+pnpm-error.log*
 
 # Dependency directories
 node_modules/

.husky/pre-commit

@@ -1,4 +1,4 @@
 #!/usr/bin/env sh
 . "$(dirname -- "$0")/_/husky.sh"
 
-npx lint-staged
+pnpm lint-staged

.npmrc

@@ -1 +1,19 @@
 //registry.npmjs.org/:_authToken=${NPM_TOKEN}
+
+# pnpm does not by default execute pre & post scripts
+# 
+# since we kinda want to keep them, we need to enable this option
+# see https://pnpm.io/cli/run#differences-with-npm-run for more info
+enable-pre-post-scripts=true
+
+# pnpm will by default use the workspace protocol when installing
+# local dependencies
+# 
+# in order not to modify the publishing step just yet, we'll disable this behavior
+# 
+# TODO reenable this behavior and use solutions provided in the linked PRs
+# in this github issue for changesets https://github.com/changesets/changesets/issues/432
+# 
+# see more here https://pnpm.io/npmrc#save-workspace-protocol
+save-workspace-protocol=false
+prefer-workspace-packages=true

.prettierignore

@@ -1,6 +1,6 @@
-yarn.lock
 *.log
 .eslintignore
 .prettierignore
 .netrc
 *ignore
+pnpm-*.yaml

DEVELOPMENT.md

@@ -24,26 +24,30 @@ The code is arranged into:
 # or nvm install if nvm use fails
 nvm use
 
-yarn
+# If pnpm is not installed on your local machine, you can install it using corepack
+corepack enable
+
+# Install project dependencies
+pnpm install
 
 # Build the entire project
-yarn build
+pnpm build
 
 # Lints the entire project
-yarn lint
+pnpm lint
 
 # Tests the entire project
-yarn test
+pnpm test
 
 # Runs the project in development mode
-yarn dev
+pnpm dev
 ```
 
 This project is built using `turborepo`. The above commands are just aliases to `turbo` CLI and as such support all the `turbo` options:
 
 ```bash
 # To start the development mode for create-lz-oapp and its dependencies
-yarn dev --filter=create-lz-oapp...
+pnpm dev --filter=create-lz-oapp...
 ```
 
 ### Running tests
@@ -56,7 +60,7 @@ The tests are by default executed in a containerized environment that sets up tw
 You can run the whole test suite within this environment by running:
 
 ```bash
-yarn test
+pnpm test
 ```
 
 #### Refining tested packages
@@ -65,21 +69,21 @@ To only run a specific test suite, you can define `DOCKER_COMPOSE_RUN_TESTS_TURB
 
 ```bash
 # To only run tests for @layerzerolabs/ua-devtools-evm-hardhat-test package
-DOCKER_COMPOSE_RUN_TESTS_TURBO_ARGS=--filter=ua-devtools-evm-hardhat-test yarn test
+DOCKER_COMPOSE_RUN_TESTS_TURBO_ARGS=--filter=ua-devtools-evm-hardhat-test pnpm test
 ```
 
 #### Rebuilding containers
 
 `docker compose` will by default reuse images for containers it has already built. If by any chance you are seeing code changes not being reflected in your test runs, you can force docker to rebuild the images by defining `DOCKER_COMPOSE_RUN_TESTS_ARGS` environment variable. This variable will be passed to the underlying `docker compose run` command and can contain any arguments that this command understands, for example:
 
 ```bash
-DOCKER_COMPOSE_RUN_TESTS_ARGS=--build yarn test
+DOCKER_COMPOSE_RUN_TESTS_ARGS=--build pnpm test
 ```
 
 You also combine the environment variables:
 
 ```bash
-DOCKER_COMPOSE_RUN_TESTS_TURBO_ARGS=--filter=ua-devtools-evm-hardhat-test DOCKER_COMPOSE_RUN_TESTS_ARGS=--build yarn test
+DOCKER_COMPOSE_RUN_TESTS_TURBO_ARGS=--filter=ua-devtools-evm-hardhat-test DOCKER_COMPOSE_RUN_TESTS_ARGS=--build pnpm test
 ```
 
 #### Container logs
@@ -97,7 +101,7 @@ This allows you to monitor logs coming from e.g. the `hardhat` nodes
 It is possible to expose the test networks defined in `docker-compose.yaml` on your host machine. To do this, you can run:
 
 ```bash
-yarn start
+pnpm start
 ```
 
 Once the networks are running, you can go to the `ua-devtools-evm-hardhat-test` package:
@@ -109,25 +113,25 @@ cd packages/ua-devtools-hardhat-test
 Setup the default `EndpointV2` and `DefaultOApp`:
 
 ```bash
-npx hardhat lz:test:oapp:deploy
+pnpm hardhat lz:test:oapp:deploy
 ```
 
 And execute `hardhat` tasks as usual:
 
 ```bash
-npx hardhat lz:oapp:getDefaultConfig
+pnpm hardhat lz:oapp:getDefaultConfig
 ```
 
 If you are developing tasks, it's useful to build the code when it changes. To do this, run the following from the project root:
 
 ```bash
-yarn dev
+pnpm dev
 ```
 
 To stop the network containers, just run:
 
 ```bash
-yarn stop
+pnpm stop
 ```
 
 **Don't forget that the state of the local networks disappears after they are stopped and any deployment files created in one session will be invalid in the next one.**
@@ -144,21 +148,9 @@ export NVM_DIR="$HOME/.nvm"
 [ -s "$NVM_DIR/nvm.sh" ] && \. "$NVM_DIR/nvm.sh"
 ```
 
-#### Problems with package updating
-
-To update external `@layerzerolabs` packages, you can use the builtin `yarn` utility:
-
-```bash
-yarn upgrade-interactive --scope @layerzerolabs --latest
-```
-
-However, this utility has an issue with packages that are listed both at the workspace root and in the individual packages, e.g. `@layerzerolabs/prettier-config-next` - it errors out saying that a workspace package could not be found.
-
-To work around this (since this version of yarn is outdated and a fix for this problem will not be provided), you can remove the entries from the root `package.json` before running the command, then add them back (just don't forget to update their versions).
-
 #### Problems using the `dev` script
 
-`turbo` might complain about concurrency issues when running `yarn dev`:
+`turbo` might complain about concurrency issues when running `pnpm dev`:
 
 ```diff
 - error preparing engine: Invalid persistent task configuration:
@@ -168,7 +160,7 @@ To work around this (since this version of yarn is outdated and a fix for this p
 If you see this error, just follow turbo's lead and use:
 
 ```bash
-yarn dev --concurrency 19
+pnpm dev --concurrency 19
 ```
 
 #### Problems with snapshots
@@ -183,5 +175,5 @@ For some packages the snapshot output depends on environment variables and other
 If you encounter errors when running these tests, just set the environment variable before running the test:
 
 ```bash
-CI=1 yarn test
+CI=1 pnpm test
 ```