Merge pull request #17505 from github/parallelize-link-check-workflows

Parallelize link check workflows

.github/workflows/link-check-dotcom.yml

@@ -0,0 +1,57 @@
+name: 'Link Checker: Dotcom'
+
+on:
+  workflow_dispatch:
+  push:
+
+jobs:
+  see_if_should_skip:
+    continue-on-error: true
+    runs-on: ubuntu-latest
+    # Map a step output to a job output
+    outputs:
+      should_skip: ${{ steps.skip_check.outputs.should_skip }}
+    steps:
+      - id: skip_check
+        uses: fkirc/skip-duplicate-actions@36feb0d8d062137530c2e00bd278d138fe191289
+        with:
+          cancel_others: 'false'
+          github_token: ${{ github.token }}
+          paths: '[".github/workflows/link-check-dotcom.yml", "assets/**", "content/**", "data/**", "includes/**", "javascripts/**", "jest-puppeteer.config.js", "jest.config.js", "layouts/**", "lib/**", "middleware/**", "package-lock.json", "package.json", "server.js", "translations/**", "webpack.config.js"]'
+  build:
+    needs: see_if_should_skip
+    runs-on: ${{ fromJSON('["ubuntu-latest", "self-hosted"]')[github.repository == 'github/docs-internal'] }}
+    steps:
+      # Each of these ifs needs to be repeated at each step to make sure the required check still runs
+      # Even if if doesn't do anything
+      - if: ${{ needs.see_if_should_skip.outputs.should_skip != 'true' }}
+        name: Checkout
+        uses: actions/checkout@5a4ac9002d0be2fb38bd78e4b4dbde5606d7042f
+
+      - if: ${{ needs.see_if_should_skip.outputs.should_skip != 'true' }}
+        name: Setup node
+        uses: actions/setup-node@c46424eee26de4078d34105d3de3cc4992202b1e
+        with:
+          node-version: 14.x
+
+      - if: ${{ needs.see_if_should_skip.outputs.should_skip != 'true' }}
+        name: Install
+        run: npm ci
+
+      ## TODO
+      # - if: ${{ github.repository == 'github/docs-internal' &&  needs.see_if_should_skip.outputs.should_skip != 'true' }}
+      #   name: Clone early access
+      #   run: npm run heroku-postbuild
+      #   env:
+      #     DOCUBOT_REPO_PAT: ${{ secrets.DOCUBOT_REPO_PAT }}
+      #     GIT_BRANCH: ${{ github.ref }}
+
+      - if: ${{ needs.see_if_should_skip.outputs.should_skip != 'true' }}
+        name: Build
+        run: npm run build
+
+      - if: ${{ needs.see_if_should_skip.outputs.should_skip != 'true' }}
+        name: 'Link check: Dotcom'
+        env:
+          DOCS_VERSION: 'dotcom'
+        run: npm run link-check

.github/workflows/link-check-ghae.yml

@@ -1,4 +1,4 @@
-name: Link Checker
+name: 'Link Checker: GitHub AE'
 
 on:
   workflow_dispatch:
@@ -17,7 +17,7 @@ jobs:
         with:
           cancel_others: 'false'
           github_token: ${{ github.token }}
-          paths: '[".github/workflows/link-check-test.yml", "assets/**", "content/**", "data/**", "includes/**", "javascripts/**", "jest-puppeteer.config.js", "jest.config.js", "layouts/**", "lib/**", "middleware/**", "package-lock.json", "package.json", "server.js", "translations/**", "webpack.config.js"]'
+          paths: '[".github/workflows/link-check-ghae.yml", "assets/**", "content/**", "data/**", "includes/**", "javascripts/**", "jest-puppeteer.config.js", "jest.config.js", "layouts/**", "lib/**", "middleware/**", "package-lock.json", "package.json", "server.js", "translations/**", "webpack.config.js"]'
   build:
     needs: see_if_should_skip
     runs-on: ${{ fromJSON('["ubuntu-latest", "self-hosted"]')[github.repository == 'github/docs-internal'] }}
@@ -51,5 +51,7 @@ jobs:
         run: npm run build
 
       - if: ${{ needs.see_if_should_skip.outputs.should_skip != 'true' }}
-        name: Link check
+        name: 'Link check: GitHub AE'
+        env:
+          DOCS_VERSION: 'github-ae'
         run: npm run link-check

.github/workflows/link-check-ghes.yml

@@ -0,0 +1,57 @@
+name: 'Link Checker: Enterprise Server'
+
+on:
+  workflow_dispatch:
+  push:
+
+jobs:
+  see_if_should_skip:
+    continue-on-error: true
+    runs-on: ubuntu-latest
+    # Map a step output to a job output
+    outputs:
+      should_skip: ${{ steps.skip_check.outputs.should_skip }}
+    steps:
+      - id: skip_check
+        uses: fkirc/skip-duplicate-actions@36feb0d8d062137530c2e00bd278d138fe191289
+        with:
+          cancel_others: 'false'
+          github_token: ${{ github.token }}
+          paths: '[".github/workflows/link-check-ghes.yml", "assets/**", "content/**", "data/**", "includes/**", "javascripts/**", "jest-puppeteer.config.js", "jest.config.js", "layouts/**", "lib/**", "middleware/**", "package-lock.json", "package.json", "server.js", "translations/**", "webpack.config.js"]'
+  build:
+    needs: see_if_should_skip
+    runs-on: ${{ fromJSON('["ubuntu-latest", "self-hosted"]')[github.repository == 'github/docs-internal'] }}
+    steps:
+      # Each of these ifs needs to be repeated at each step to make sure the required check still runs
+      # Even if if doesn't do anything
+      - if: ${{ needs.see_if_should_skip.outputs.should_skip != 'true' }}
+        name: Checkout
+        uses: actions/checkout@5a4ac9002d0be2fb38bd78e4b4dbde5606d7042f
+
+      - if: ${{ needs.see_if_should_skip.outputs.should_skip != 'true' }}
+        name: Setup node
+        uses: actions/setup-node@c46424eee26de4078d34105d3de3cc4992202b1e
+        with:
+          node-version: 14.x
+
+      - if: ${{ needs.see_if_should_skip.outputs.should_skip != 'true' }}
+        name: Install
+        run: npm ci
+
+      ## TODO
+      # - if: ${{ github.repository == 'github/docs-internal' &&  needs.see_if_should_skip.outputs.should_skip != 'true' }}
+      #   name: Clone early access
+      #   run: npm run heroku-postbuild
+      #   env:
+      #     DOCUBOT_REPO_PAT: ${{ secrets.DOCUBOT_REPO_PAT }}
+      #     GIT_BRANCH: ${{ github.ref }}
+
+      - if: ${{ needs.see_if_should_skip.outputs.should_skip != 'true' }}
+        name: Build
+        run: npm run build
+
+      - if: ${{ needs.see_if_should_skip.outputs.should_skip != 'true' }}
+        name: 'Link check: Enterprise Server'
+        env:
+          DOCS_VERSION: 'enterprise-server'
+        run: npm run link-check

script/check-internal-links.js

@@ -2,14 +2,16 @@
 
 const linkinator = require('linkinator')
 const checker = new linkinator.LinkChecker()
-const { deprecated } = require('../lib/enterprise-server-releases')
+const { deprecated, latest } = require('../lib/enterprise-server-releases')
 const englishRoot = 'http://localhost:4002/en'
+const allowedVersions = ['dotcom', 'enterprise-server', 'github-ae']
 
 // [start-readme]
 //
-// This script runs in CI via GitHub Action to check all *internal* links in English content,
-// not including deprecated Enterprise Server content. This is different from script/check-english-links.js,
-// which checks *all* links in the site, both internal and external, and is much slower.
+// This script runs in CI via GitHub Action to check all *internal* links in English content for a
+// given version, not including deprecated Enterprise Server content. This is different from
+// script/check-english-links.js, which checks *all* links in the site, both internal and external,
+// and is much slower.
 //
 // If you want to run it locally, you must have a local server running. You can use `npm run link-check`.
 //
@@ -31,14 +33,34 @@ const config = {
     // Skip dist files
     '/dist/index.*',
     // Skip deprecated Enterprise content
-    `enterprise(-server@|/)(${deprecated.join('|')})/?`
+    `enterprise(-server@|/)(${deprecated.join('|')})(/|$)`
   ]
 }
 
+// Customize config for specific versions
+if (process.env.DOCS_VERSION === 'dotcom') {
+  // If Dotcom, skip Enterprise Server and GitHub AE links
+  config.linksToSkip.push('^.*/enterprise-server@.*$', '^.*/enterprise/.*$', '^.*/github-ae@latest.*$')
+} else if (process.env.DOCS_VERSION === 'enterprise-server') {
+  // If Enterprise Server, skip links that are not Enterprise Server links
+  config.path = `${englishRoot}/enterprise-server@${latest}`
+  config.linksToSkip.push('^((?!enterprise-server@).)*$')
+} else if (process.env.DOCS_VERSION === 'github-ae') {
+  // If GitHub AE, skip links that are not GitHub AE links
+  config.path = `${englishRoot}/github-ae@latest`
+  config.linksToSkip.push('^((?!github-ae@latest).)*$')
+}
+
 main()
 
 async function main () {
+  process.env.DOCS_VERSION && allowedVersions.includes(process.env.DOCS_VERSION)
+    ? console.log(`Checking internal links for version ${process.env.DOCS_VERSION}!\n`)
+    : console.log('Checking internal links for all versions!\n')
+
+  console.time('check')
   const result = (await checker.check(config)).links
+  console.timeEnd('check')
 
   const brokenLinks = result
     .filter(link => link.state === 'BROKEN')