New broken link report (github#16412)

* add linkinator npm package

* add new script that uses Linkinator

* reorg the excluded links file and update comments

* replace blc artifacts with linkinator artifacts in .gitignore

* update the scheduled workflow to use the new script

* dismantle BLC scripts

* add workflow_dispatch event so we can test this manually

* npm uninstall broken-link-checker

* use different exit codes depending on whether broken links are found

.github/workflows/check-all-english-links.yml

@@ -1,6 +1,7 @@
 name: Check all English links
 
 on:
+  workflow_dispatch:
   schedule:
     - cron:  "40 19 * * *" # once a day at 19:40 UTC / 11:40 PST
 
@@ -10,21 +11,16 @@ jobs:
     if: github.repository == 'github/docs-internal'
     runs-on: ubuntu-latest
     steps:
-    - uses: actions/checkout@5a4ac9002d0be2fb38bd78e4b4dbde5606d7042f
-    - name: npm ci
-      run: npm ci
-    - name: npm run build
-      run: npm run build
     - name: Run script
-      run: script/check-external-links en > broken_links.md
+      run: script/check-english-links > broken_links.md
     - name: Check if any broken links
       id: check
       run: |
-        if [ "$(grep 'All links are good' broken_links.md)" ]; then
+        if [ "$(grep '0 broken links found' broken_links.md)" ]; then
           echo ::set-output name=continue::no
         else
           echo "::set-output name=continue::yes"
-          echo "::set-output name=title::$(grep 'found on help.github.com' broken_links.md)"
+          echo "::set-output name=title::$(head -1 broken_links.md)"
         fi
     - if: ${{ steps.check.outputs.continue == 'yes' }}
       name: Create issue from file

.gitignore

@@ -4,8 +4,6 @@
 node_modules
 npm-debug.log
 coverage
-
-# blc: broken link checker
-blc_output.log
-blc_output_internal.log
-dist
+.linkinator
+broken_links.md
+dist

lib/excluded-links.js

@@ -1,28 +1,20 @@
+// Linkinator treats the following as regex.
 module.exports = [
-  //  GitHub search links fail with "429 Too Many Requests"
-  'https://github.com/search?*',
+  // Skip GitHub search links.
+  'https://github.com/search?.*',
   'https://github.com/github/gitignore/search?',
 
-  // LinkedIn links fail due to bug: https://github.com/stevenvachon/broken-link-checker/issues/91
-  'https://www.linkedin.com/*',
-
-  // blc returns "BLC_UNKNOWN"  on this link, even though cURL returns "302 Found"
-  'https://www.ilo.org/dyn/normlex/en/f?p=NORMLEXPUB:12100:0::NO::P12100_ILO_CODE:P029',
-
-  // the codercat link works but blc reports a false 404
-  'https://github.com/Codertocat/hello-world-npm/packages/10696?version=1.0.1',
-
-  // this URL started returning 403 to blc and cURL even though it works in a browser; see docs-internal #10124
-  'https://haveibeenpwned.com/',
-  'https://haveibeenpwned.com/*',
-
-  // this is a private repo customers are given access to when they purchase Insights; see docs-internal #12037
+  // These links require auth.
+  'https://github.com/settings/profile',
+  'https://github.com/github/docs/edit',
   'https://github.com/github/insights-releases/releases/latest',
 
-  // developer content uses these for examples; they should not be checked
-  'http://localhost:1234/*',
+  // Developer content uses these for examples; they should not be checked.
+  'http://localhost:1234',
   'localhost:3000',
 
-  // this URL works but blc reports a false 404
+  // Oneoff links that link checkers think are broken but are not.
+  'https://haveibeenpwned.com/',
+  'https://www.ilo.org/dyn/normlex/en/f?p=NORMLEXPUB:12100:0::NO::P12100_ILO_CODE:P029',
   'http://www.w3.org/wiki/LinkHeader/'
 ]