fix(cli): prevent OOM crash by limiting file search traversal and adding timeout (#16696)

packages/core/src/utils/filesearch/crawler.test.ts

@@ -503,14 +503,14 @@ describe('crawler', () => {
       });
     });
 
-    const getCrawlResults = (maxDepth?: number) => {
+    const getCrawlResults = async (maxDepth?: number) => {
       const ignore = loadIgnoreRules({
         projectRoot: tmpDir,
         useGitignore: false,
         useGeminiignore: false,
         ignoreDirs: [],
       });
-      return crawl({
+      const paths = await crawl({
         crawlDirectory: tmpDir,
         cwd: tmpDir,
         ignore,
@@ -518,6 +518,7 @@ describe('crawler', () => {
         cacheTtl: 0,
         maxDepth,
       });
+      return paths;
     };
 
     it('should only crawl top-level files when maxDepth is 0', async () => {
@@ -571,4 +572,34 @@ describe('crawler', () => {
       );
     });
   });
+
+  it('should detect truncation when maxFiles is hit', async () => {
+    tmpDir = await createTmpDir({
+      'file1.js': '',
+      'file2.js': '',
+      'file3.js': '',
+    });
+
+    const ignore = loadIgnoreRules({
+      projectRoot: tmpDir,
+      useGitignore: false,
+      useGeminiignore: false,
+      ignoreDirs: [],
+    });
+
+    const paths = await crawl({
+      crawlDirectory: tmpDir,
+      cwd: tmpDir,
+      ignore,
+      cache: false,
+      cacheTtl: 0,
+      maxFiles: 2,
+    });
+
+    // fdir returns files and directories.
+    // In our filter, we only increment fileCount for files.
+    // So we should have 2 files + some directories.
+    const files = paths.filter((p) => p !== '.' && !p.endsWith('/'));
+    expect(files.length).toBe(2);
+  });
 });

packages/core/src/utils/filesearch/crawler.ts

@@ -16,6 +16,8 @@ export interface CrawlOptions {
   cwd: string;
   // The fdir maxDepth option.
   maxDepth?: number;
+  // Maximum number of files to return.
+  maxFiles?: number;
   // A pre-configured Ignore instance.
   ignore: Ignore;
   // Caching options.
@@ -43,6 +45,9 @@ export async function crawl(options: CrawlOptions): Promise<string[]> {
 
   const posixCwd = toPosixPath(options.cwd);
   const posixCrawlDirectory = toPosixPath(options.crawlDirectory);
+  const maxFiles = options.maxFiles ?? Infinity;
+  let fileCount = 0;
+  let truncated = false;
 
   let results: string[];
   try {
@@ -51,7 +56,21 @@ export async function crawl(options: CrawlOptions): Promise<string[]> {
       .withRelativePaths()
       .withDirs()
       .withPathSeparator('/') // Always use unix style paths
+      .filter((path, isDirectory) => {
+        if (!isDirectory) {
+          fileCount++;
+          if (fileCount > maxFiles) {
+            truncated = true;
+            return false;
+          }
+        }
+        return true;
+      })
       .exclude((_, dirPath) => {
+        if (fileCount > maxFiles) {
+          truncated = true;
+          return true;
+        }
         const relativePath = path.posix.relative(posixCrawlDirectory, dirPath);
         return dirFilter(`${relativePath}/`);
       });
@@ -72,7 +91,7 @@ export async function crawl(options: CrawlOptions): Promise<string[]> {
     path.posix.join(relativeToCrawlDir, p),
   );
 
-  if (options.cache) {
+  if (options.cache && !truncated) {
     const cacheKey = cache.getCacheKey(
       options.crawlDirectory,
       options.ignore.getFingerprint(),

packages/core/src/utils/filesearch/fileSearch.test.ts

@@ -7,6 +7,7 @@
 import { describe, it, expect, afterEach, vi } from 'vitest';
 import { FileSearchFactory, AbortError, filter } from './fileSearch.js';
 import { createTmpDir, cleanupTmpDir } from '@google/gemini-cli-test-utils';
+import * as crawler from './crawler.js';
 
 describe('FileSearch', () => {
   let tmpDir: string;
@@ -481,6 +482,33 @@ describe('FileSearch', () => {
     expect(results).toEqual(['src/', 'src/main.js']);
   });
 
+  it('should respect default maxFiles budget of 20000 in RecursiveFileSearch', async () => {
+    const crawlSpy = vi.spyOn(crawler, 'crawl');
+
+    tmpDir = await createTmpDir({
+      'file1.js': '',
+    });
+
+    const fileSearch = FileSearchFactory.create({
+      projectRoot: tmpDir,
+      useGitignore: false,
+      useGeminiignore: false,
+      ignoreDirs: [],
+      cache: false,
+      cacheTtl: 0,
+      enableRecursiveFileSearch: true,
+      disableFuzzySearch: false,
+    });
+
+    await fileSearch.initialize();
+
+    expect(crawlSpy).toHaveBeenCalledWith(
+      expect.objectContaining({
+        maxFiles: 20000,
+      }),
+    );
+  });
+
   it('should be cancellable via AbortSignal', async () => {
     const largeDir: Record<string, string> = {};
     for (let i = 0; i < 100; i++) {

packages/core/src/utils/filesearch/fileSearch.ts

@@ -24,6 +24,7 @@ export interface FileSearchOptions {
   enableRecursiveFileSearch: boolean;
   disableFuzzySearch: boolean;
   maxDepth?: number;
+  maxFiles?: number;
 }
 
 export class AbortError extends Error {
@@ -109,7 +110,9 @@ class RecursiveFileSearch implements FileSearch {
       cache: this.options.cache,
       cacheTtl: this.options.cacheTtl,
       maxDepth: this.options.maxDepth,
+      maxFiles: this.options.maxFiles ?? 20000,
     });
+
     this.buildResultCache();
   }