Enable resolving languages without CodeQL CLI

.github/workflows/validate-resolved-languages.yml

@@ -0,0 +1,80 @@
+name: Validate Resolved Languages File
+env:
+  GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+  GO111MODULE: auto
+on:
+  push:
+    branches:
+      - main
+      - releases/v*
+  pull_request:
+    types:
+      - opened
+      - synchronize
+      - reopened
+      - ready_for_review
+  schedule:
+    - cron: '0 5 * * *'
+  workflow_dispatch:
+    inputs: {}
+  workflow_call:
+    inputs: {}
+defaults:
+  run:
+    shell: bash
+jobs:
+  validate-resolved-languages:
+    name: Validate Resolved Languages File
+    permissions:
+      contents: read
+      security-events: read
+    timeout-minutes: 10
+    runs-on: ubuntu-latest
+    steps:
+      - name: Check out repository
+        uses: actions/checkout@v5
+      - name: Install Node.js
+        uses: actions/setup-node@v4
+        with:
+          node-version: 20.x
+          cache: npm
+      - name: Install dependencies
+        run: npm ci
+      - name: Prepare test
+        id: prepare-test
+        uses: ./.github/actions/prepare-test
+        with:
+          version: linked
+          use-all-platform-bundle: 'false'
+          setup-kotlin: 'false'
+      - name: Initialize CodeQL
+        id: init
+        uses: ./../action/init
+        with:
+          tools: ${{ steps.prepare-test.outputs.tools-url }}
+      - name: Verify that the resolved languages file is up to date
+        env:
+          CODEQL: ${{ steps.init.outputs.codeql-path }}
+          RESOLVED_LANGUAGES_FILE: ../action/src/resolved-languages.json
+        run: |
+          if [ ! -f "${RESOLVED_LANGUAGES_FILE}" ]; then
+            echo "::error file=src/resolved-languages.json::The resolved-languages.json file is missing."
+            exit 1
+          fi
+          #
+          # Please update getStoredSupportedLanguageMap() when you update the
+          # following call to use --filter-to-languages-with-queries.
+          #
+          "${CODEQL}" resolve languages --format=betterjson --extractor-include-aliases > resolved-languages.json
+          if ! diff -u resolved-languages.json "${RESOLVED_LANGUAGES_FILE}"; then
+            echo "::error file=src/resolved-languages.json::The saved resolved-languages.json file is out of date."
+            exit 1
+          fi
+      - name: Upload resolved languages file for debugging
+        if: ${{ failure() }}
+        uses: actions/upload-artifact@v4
+        with:
+          name: resolved-languages
+          path: resolved-languages.json
+    env:
+      CODEQL_ACTION_TEST_MODE: true

src/codeql.ts

@@ -237,12 +237,10 @@ export interface BetterResolveLanguagesOutput {
     [alias: string]: string;
   };
   extractors: {
-    [language: string]: [
-      {
-        extractor_root: string;
-        extractor_options?: any;
-      },
-    ];
+    [language: string]: Array<{
+      extractor_root: string;
+      extractor_options?: any;
+    }>;
   };
 }
 

src/config-utils.test.ts

@@ -1163,14 +1163,18 @@ const mockRepositoryNwo = parseRepositoryNwo("owner/repo");
           }),
       });
 
+      const languageMap = await configUtils.getSupportedLanguageMap(
+        codeQL,
+        features,
+        mockLogger,
+      );
       if (args.expectedLanguages) {
         // happy path
         const actualLanguages = await configUtils.getLanguages(
-          codeQL,
+          languageMap,
           args.languagesInput,
           mockRepositoryNwo,
           ".",
-          features,
           mockLogger,
         );
 
@@ -1180,11 +1184,10 @@ const mockRepositoryNwo = parseRepositoryNwo("owner/repo");
         await t.throwsAsync(
           async () =>
             await configUtils.getLanguages(
-              codeQL,
+              languageMap,
               args.languagesInput,
               mockRepositoryNwo,
               ".",
-              features,
               mockLogger,
             ),
           { message: args.expectedError },

src/config-utils.ts

@@ -16,7 +16,7 @@ import {
 } from "./analyses";
 import * as api from "./api-client";
 import { CachingKind, getCachingKind } from "./caching-utils";
-import { type CodeQL } from "./codeql";
+import { type BetterResolveLanguagesOutput, type CodeQL } from "./codeql";
 import { shouldPerformDiffInformedAnalysis } from "./diff-informed-analysis-utils";
 import { Feature, FeatureEnablement } from "./feature-flags";
 import { getGitRoot, isAnalyzingDefaultBranch } from "./git-utils";
@@ -27,6 +27,7 @@ import {
   OverlayDatabaseMode,
 } from "./overlay-database-utils";
 import { RepositoryNwo } from "./repository";
+import * as resolvedLanguages from "./resolved-languages.json";
 import { downloadTrapCaches } from "./trap-caching";
 import {
   GitHubVersion,
@@ -331,6 +332,30 @@ export async function getSupportedLanguageMap(
       `The CodeQL CLI supports the following languages: ${Object.keys(resolveResult.extractors).join(", ")}`,
     );
   }
+  return buildSupportedLanguageMap(
+    resolveResult,
+    resolveSupportedLanguagesUsingCli,
+  );
+}
+
+// This function serves the same purpose as getSupportedLanguageMap(), but it
+// uses the stored resolved-languages.json file instead of calling out to the
+// CodeQL CLI.
+export function getStoredSupportedLanguageMap(): Record<string, string> {
+  return buildSupportedLanguageMap(
+    resolvedLanguages,
+    // resolveSupportedLanguagesUsingCli is false because we currently generate
+    // resolved-languages.json without the --filter-to-languages-with-queries
+    // flag (see .github/workflows/validate-resolved-languages.yml). Once that
+    // changes, we should set this to true.
+    false,
+  );
+}
+
+function buildSupportedLanguageMap(
+  resolveResult: BetterResolveLanguagesOutput,
+  resolveSupportedLanguagesUsingCli: boolean,
+): Record<string, string> {
   const supportedLanguages: Record<string, string> = {};
   // Populate canonical language names
   for (const extractor of Object.keys(resolveResult.extractors)) {
@@ -418,11 +443,10 @@ export async function getRawLanguagesInRepo(
  * then throw an error.
  */
 export async function getLanguages(
-  codeql: CodeQL,
+  languageMap: Record<string, string>,
   languagesInput: string | undefined,
   repository: RepositoryNwo,
   sourceRoot: string,
-  features: FeatureEnablement,
   logger: Logger,
 ): Promise<Language[]> {
   // Obtain languages without filtering them.
@@ -432,8 +456,6 @@ export async function getLanguages(
     sourceRoot,
     logger,
   );
-
-  const languageMap = await getSupportedLanguageMap(codeql, features, logger);
   const languagesSet = new Set<Language>();
   const unknownLanguages: string[] = [];
 
@@ -575,12 +597,12 @@ export async function initActionState(
     analysisKinds.push(AnalysisKind.CodeQuality);
   }
 
+  const languageMap = await getSupportedLanguageMap(codeql, features, logger);
   const languages = await getLanguages(
-    codeql,
+    languageMap,
     languagesInput,
     repository,
     sourceRoot,
-    features,
     logger,
   );