Prevent Link Checker from erring on long file lists (github#24269)

* Add '--list' parameter to rendered-content-link-checker script

* Utilize trilom/file-changes-action output files and pass to rendered-content-link-checker script

* Rename job to unique name

* Correct module path for import

* Move import into local imports section

* Dump the contents of the file listing

* Fix typo

Author: JamesMGreene

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

@@ -22,7 +22,7 @@ concurrency:
   cancel-in-progress: true
 
 jobs:
-  build:
+  check-links:
     runs-on: ${{ fromJSON('["ubuntu-latest", "self-hosted"]')[github.repository == 'github/docs-internal'] }}
     steps:
       - name: Checkout
@@ -37,26 +37,15 @@ jobs:
       - name: Install
         run: npm ci
 
+      # Creates file "${{ env.HOME }}/files.json", among others
       - name: Gather files changed
         uses: trilom/file-changes-action@a6ca26c14274c33b15e6499323aac178af06ad4b
-        id: get_diff_files
-
-      # Necessary because trilom/file-changes-action can't escape each file
-      # name for using in bash. So, we do it ourselves.
-      # trilom/file-changes-action will, by default produce outputs
-      # in JSON format. We consume that and set a new output where each
-      # filename is wrapped in quotation marks.
-      # Perhaps some day we can rely on this directly based on;
-      # https://github.com/trilom/file-changes-action/issues/130
-      - name: Escape each diff file name
-        id: get_diff_files_escaped
-        uses: actions/github-script@2b34a689ec86a68d8ab9478298f91d5401337b7d
         with:
-          github-token: ${{ secrets.DOCUBOT_READORG_REPO_WORKFLOW_SCOPES }}
-          script: |
-            const input = JSON.parse('${{ steps.get_diff_files.outputs.files }}')
-            const files = input.map(filename => `"${filename}"`)
-            core.setOutput('files', files.join(' '))
+          fileOutput: 'json'
+
+      # For verification
+      - name: Show files changed
+        run: cat $HOME/files.json
 
       - name: Link check (warnings, changed files)
         run: |
@@ -66,7 +55,7 @@ jobs:
             --check-anchors \
             --check-images \
             --verbose \
-            ${{ steps.get_diff_files_escaped.outputs.files }}
+            --list $HOME/files.json
 
       - name: Link check (critical, all files)
         run: |

script/rendered-content-link-checker.mjs

@@ -19,6 +19,7 @@ import { languageKeys } from '../lib/languages.js'
 import warmServer from '../lib/warm-server.js'
 import renderContent from '../lib/render-content/index.js'
 import { deprecated } from '../lib/enterprise-server-releases.js'
+import readFileAsync from '../lib/readfile-async.js'
 
 const STATIC_PREFIXES = {
   assets: path.resolve('assets'),
@@ -65,13 +66,33 @@ program
     }
     return parsed
   })
+  .option(
+    '--list <file>.json',
+    'JSON file containing an array of specific files to check (default: none)',
+    (filePath) => {
+      const resolvedPath = path.resolve(filePath)
+
+      let stats
+      try {
+        stats = fs.statSync(resolvedPath)
+      } catch (error) {
+        // Ignore
+      }
+
+      if (!stats || !stats.isFile()) {
+        throw new InvalidArgumentError('Not an existing file.')
+      }
+
+      return resolvedPath
+    }
+  )
   .arguments('[files...]', 'Specific files to check')
   .parse(process.argv)
 
 main(program.opts(), program.args)
 
 async function main(opts, files) {
-  const { random, language, filter, exit, debug, max, verbose } = opts
+  const { random, language, filter, exit, debug, max, verbose, list } = opts
 
   // Note! The reason we're using `warmServer()` in this script,
   // even though there's no server involved, is because
@@ -89,6 +110,19 @@ async function main(opts, files) {
   const filters = filter || []
   console.assert(Array.isArray(filters), `${filters} is not an array`)
 
+  if (list && Array.isArray(files) && files.length > 0) {
+    throw new InvalidArgumentError('Cannot specify both --list and a file list.')
+  }
+
+  if (list) {
+    const fileList = JSON.parse(await readFileAsync(list))
+    if (Array.isArray(fileList) && fileList.length > 0) {
+      files = fileList
+    } else {
+      throw new InvalidArgumentError('No files found in --list.')
+    }
+  }
+
   if (random) {
     shuffle(pageList)
   }