Merge pull request #3083 from github/henrymercer/resolve-languages-default-queries

Resolve supported languages using CodeQL CLI

src/codeql.ts

@@ -127,7 +127,9 @@ export interface CodeQL {
   /**
    * Run 'codeql resolve languages' with '--format=betterjson'.
    */
-  betterResolveLanguages(): Promise<BetterResolveLanguagesOutput>;
+  betterResolveLanguages(options?: {
+    filterToLanguagesWithQueries: boolean;
+  }): Promise<BetterResolveLanguagesOutput>;
   /**
    * Run 'codeql resolve build-environment'
    */
@@ -736,13 +738,22 @@ export async function getCodeQLForCmd(
         );
       }
     },
-    async betterResolveLanguages() {
+    async betterResolveLanguages(
+      {
+        filterToLanguagesWithQueries,
+      }: {
+        filterToLanguagesWithQueries: boolean;
+      } = { filterToLanguagesWithQueries: false },
+    ) {
       const codeqlArgs = [
         "resolve",
         "languages",
         "--format=betterjson",
         "--extractor-options-verbosity=4",
         "--extractor-include-aliases",
+        ...(filterToLanguagesWithQueries
+          ? ["--filter-to-languages-with-queries"]
+          : []),
         ...getExtraOptionsFromEnv(["resolve", "languages"]),
       ];
       const output = await runCli(cmd, codeqlArgs);

src/config-utils.test.ts

@@ -1092,6 +1092,13 @@ const mockRepositoryNwo = parseRepositoryNwo("owner/repo");
     expectedLanguages: ["javascript", "csharp", "cpp"],
     expectedApiCall: true,
   },
+  {
+    name: "unsupported languages from github api",
+    languagesInput: "",
+    languagesInRepository: ["html"],
+    expectedApiCall: true,
+    expectedError: configUtils.getNoLanguagesError(),
+  },
   {
     name: "no languages",
     languagesInput: "",
@@ -1121,57 +1128,71 @@ const mockRepositoryNwo = parseRepositoryNwo("owner/repo");
     expectedLanguages: ["javascript"],
   },
 ].forEach((args) => {
-  test(`getLanguages: ${args.name}`, async (t) => {
-    const mockRequest = mockLanguagesInRepo(args.languagesInRepository);
-    const stubExtractorEntry = {
-      extractor_root: "",
-    };
-    const codeQL = createStubCodeQL({
-      betterResolveLanguages: () =>
-        Promise.resolve({
-          aliases: {
-            "c#": KnownLanguage.csharp,
-            c: KnownLanguage.cpp,
-            kotlin: KnownLanguage.java,
-            typescript: KnownLanguage.javascript,
-          },
-          extractors: {
-            cpp: [stubExtractorEntry],
-            csharp: [stubExtractorEntry],
-            java: [stubExtractorEntry],
-            javascript: [stubExtractorEntry],
-            python: [stubExtractorEntry],
-          },
-        }),
+  for (const resolveSupportedLanguagesUsingCli of [true, false]) {
+    test(`getLanguages${resolveSupportedLanguagesUsingCli ? " (supported languages via CLI)" : ""}: ${args.name}`, async (t) => {
+      const features = createFeatures(
+        resolveSupportedLanguagesUsingCli
+          ? [Feature.ResolveSupportedLanguagesUsingCli]
+          : [],
+      );
+      const mockRequest = mockLanguagesInRepo(args.languagesInRepository);
+      const stubExtractorEntry = {
+        extractor_root: "",
+      };
+      const codeQL = createStubCodeQL({
+        betterResolveLanguages: (options) =>
+          Promise.resolve({
+            aliases: {
+              "c#": KnownLanguage.csharp,
+              c: KnownLanguage.cpp,
+              kotlin: KnownLanguage.java,
+              typescript: KnownLanguage.javascript,
+            },
+            extractors: {
+              cpp: [stubExtractorEntry],
+              csharp: [stubExtractorEntry],
+              java: [stubExtractorEntry],
+              javascript: [stubExtractorEntry],
+              python: [stubExtractorEntry],
+              ...(options?.filterToLanguagesWithQueries
+                ? {}
+                : {
+                    html: [stubExtractorEntry],
+                  }),
+            },
+          }),
+      });
+
+      if (args.expectedLanguages) {
+        // happy path
+        const actualLanguages = await configUtils.getLanguages(
+          codeQL,
+          args.languagesInput,
+          mockRepositoryNwo,
+          ".",
+          features,
+          mockLogger,
+        );
+
+        t.deepEqual(actualLanguages.sort(), args.expectedLanguages.sort());
+      } else {
+        // there is an error
+        await t.throwsAsync(
+          async () =>
+            await configUtils.getLanguages(
+              codeQL,
+              args.languagesInput,
+              mockRepositoryNwo,
+              ".",
+              features,
+              mockLogger,
+            ),
+          { message: args.expectedError },
+        );
+      }
+      t.deepEqual(mockRequest.called, args.expectedApiCall);
     });
-
-    if (args.expectedLanguages) {
-      // happy path
-      const actualLanguages = await configUtils.getLanguages(
-        codeQL,
-        args.languagesInput,
-        mockRepositoryNwo,
-        ".",
-        mockLogger,
-      );
-
-      t.deepEqual(actualLanguages.sort(), args.expectedLanguages.sort());
-    } else {
-      // there is an error
-      await t.throwsAsync(
-        async () =>
-          await configUtils.getLanguages(
-            codeQL,
-            args.languagesInput,
-            mockRepositoryNwo,
-            ".",
-            mockLogger,
-          ),
-        { message: args.expectedError },
-      );
-    }
-    t.deepEqual(mockRequest.called, args.expectedApiCall);
-  });
+  }
 });
 
 for (const { displayName, language, feature } of [

src/config-utils.ts

@@ -316,16 +316,31 @@ export function getUnknownLanguagesError(languages: string[]): string {
 
 export async function getSupportedLanguageMap(
   codeql: CodeQL,
+  features: FeatureEnablement,
+  logger: Logger,
 ): Promise<Record<string, string>> {
-  const resolveResult = await codeql.betterResolveLanguages();
+  const resolveSupportedLanguagesUsingCli = await features.getValue(
+    Feature.ResolveSupportedLanguagesUsingCli,
+    codeql,
+  );
+  const resolveResult = await codeql.betterResolveLanguages({
+    filterToLanguagesWithQueries: resolveSupportedLanguagesUsingCli,
+  });
+  if (resolveSupportedLanguagesUsingCli) {
+    logger.debug(
+      `The CodeQL CLI supports the following languages: ${Object.keys(resolveResult.extractors).join(", ")}`,
+    );
+  }
   const supportedLanguages: Record<string, string> = {};
   // Populate canonical language names
   for (const extractor of Object.keys(resolveResult.extractors)) {
-    // Require the language to be a known language.
-    // This is a temporary workaround since we have extractors that are not
-    // supported languages, such as `csv`, `html`, `properties`, `xml`, and
-    // `yaml`. We should replace this with a more robust solution in the future.
-    if (KnownLanguage[extractor] !== undefined) {
+    // If the CLI supports resolving languages with default queries, use these
+    // as the set of supported languages. Otherwise, require the language to be
+    // a known language.
+    if (
+      resolveSupportedLanguagesUsingCli ||
+      KnownLanguage[extractor] !== undefined
+    ) {
       supportedLanguages[extractor] = extractor;
     }
   }
@@ -407,6 +422,7 @@ export async function getLanguages(
   languagesInput: string | undefined,
   repository: RepositoryNwo,
   sourceRoot: string,
+  features: FeatureEnablement,
   logger: Logger,
 ): Promise<Language[]> {
   // Obtain languages without filtering them.
@@ -417,7 +433,7 @@ export async function getLanguages(
     logger,
   );
 
-  const languageMap = await getSupportedLanguageMap(codeql);
+  const languageMap = await getSupportedLanguageMap(codeql, features, logger);
   const languagesSet = new Set<Language>();
   const unknownLanguages: string[] = [];
 
@@ -564,6 +580,7 @@ export async function initActionState(
     languagesInput,
     repository,
     sourceRoot,
+    features,
     logger,
   );
 

src/feature-flags.ts

@@ -73,6 +73,7 @@ export enum Feature {
   OverlayAnalysisSwift = "overlay_analysis_swift",
   PythonDefaultIsToNotExtractStdlib = "python_default_is_to_not_extract_stdlib",
   QaTelemetryEnabled = "qa_telemetry_enabled",
+  ResolveSupportedLanguagesUsingCli = "resolve_supported_languages_using_cli",
 }
 
 export const featureConfig: Record<
@@ -145,6 +146,12 @@ export const featureConfig: Record<
     legacyApi: true,
     minimumVersion: undefined,
   },
+  [Feature.ResolveSupportedLanguagesUsingCli]: {
+    defaultValue: false,
+    envVar: "CODEQL_ACTION_RESOLVE_SUPPORTED_LANGUAGES_USING_CLI",
+    minimumVersion: undefined,
+    toolsFeature: ToolsFeature.BuiltinExtractorsSpecifyDefaultQueries,
+  },
   [Feature.OverlayAnalysis]: {
     defaultValue: false,
     envVar: "CODEQL_ACTION_OVERLAY_ANALYSIS",

src/tools-features.ts

@@ -4,11 +4,12 @@ import type { VersionInfo } from "./codeql";
 
 export enum ToolsFeature {
   AnalysisSummaryV2IsDefault = "analysisSummaryV2Default",
+  BuiltinExtractorsSpecifyDefaultQueries = "builtinExtractorsSpecifyDefaultQueries",
   DatabaseInterpretResultsSupportsSarifRunProperty = "databaseInterpretResultsSupportsSarifRunProperty",
-  IndirectTracingSupportsStaticBinaries = "indirectTracingSupportsStaticBinaries",
-  SarifMergeRunsFromEqualCategory = "sarifMergeRunsFromEqualCategory",
   ForceOverwrite = "forceOverwrite",
+  IndirectTracingSupportsStaticBinaries = "indirectTracingSupportsStaticBinaries",
   PythonDefaultIsToNotExtractStdlib = "pythonDefaultIsToNotExtractStdlib",
+  SarifMergeRunsFromEqualCategory = "sarifMergeRunsFromEqualCategory",
 }
 
 /**