feat(core): share file list patterns between glob and grep tools (#6359)

Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
Co-authored-by: Arya Gummadi <aryagummadi@google.com>

packages/core/src/tools/glob.test.ts

@@ -29,6 +29,9 @@ describe('GlobTool', () => {
     getFileFilteringRespectGitIgnore: () => true,
     getTargetDir: () => tempRootDir,
     getWorkspaceContext: () => createMockWorkspaceContext(tempRootDir),
+    getFileExclusions: () => ({
+      getGlobExcludes: () => [],
+    }),
   } as unknown as Config;
 
   beforeEach(async () => {

packages/core/src/tools/glob.ts

@@ -155,7 +155,7 @@ class GlobToolInvocation extends BaseToolInvocation<
           stat: true,
           nocase: !this.params.case_sensitive,
           dot: true,
-          ignore: ['**/node_modules/**', '**/.git/**'],
+          ignore: this.config.getFileExclusions().getGlobExcludes(),
           follow: false,
           signal,
         })) as GlobPath[];

packages/core/src/tools/grep.test.ts

@@ -39,6 +39,9 @@ describe('GrepTool', () => {
   const mockConfig = {
     getTargetDir: () => tempRootDir,
     getWorkspaceContext: () => createMockWorkspaceContext(tempRootDir),
+    getFileExclusions: () => ({
+      getGlobExcludes: () => [],
+    }),
   } as unknown as Config;
 
   beforeEach(async () => {
@@ -258,6 +261,9 @@ describe('GrepTool', () => {
         getTargetDir: () => tempRootDir,
         getWorkspaceContext: () =>
           createMockWorkspaceContext(tempRootDir, [secondDir]),
+        getFileExclusions: () => ({
+          getGlobExcludes: () => [],
+        }),
       } as unknown as Config;
 
       const multiDirGrepTool = new GrepTool(multiDirConfig);
@@ -308,6 +314,9 @@ describe('GrepTool', () => {
         getTargetDir: () => tempRootDir,
         getWorkspaceContext: () =>
           createMockWorkspaceContext(tempRootDir, [secondDir]),
+        getFileExclusions: () => ({
+          getGlobExcludes: () => [],
+        }),
       } as unknown as Config;
 
       const multiDirGrepTool = new GrepTool(multiDirConfig);
@@ -367,6 +376,9 @@ describe('GrepTool', () => {
         getTargetDir: () => tempRootDir,
         getWorkspaceContext: () =>
           createMockWorkspaceContext(tempRootDir, ['/another/dir']),
+        getFileExclusions: () => ({
+          getGlobExcludes: () => [],
+        }),
       } as unknown as Config;
 
       const multiDirGrepTool = new GrepTool(multiDirConfig);

packages/core/src/tools/grep.ts

@@ -21,6 +21,7 @@ import { makeRelative, shortenPath } from '../utils/paths.js';
 import { getErrorMessage, isNodeError } from '../utils/errors.js';
 import { isGitRepository } from '../utils/gitUtils.js';
 import { Config } from '../config/config.js';
+import { FileExclusions } from '../utils/ignorePatterns.js';
 import { ToolErrorType } from './tool-error.js';
 
 // --- Interfaces ---
@@ -58,11 +59,14 @@ class GrepToolInvocation extends BaseToolInvocation<
   GrepToolParams,
   ToolResult
 > {
+  private readonly fileExclusions: FileExclusions;
+
   constructor(
     private readonly config: Config,
     params: GrepToolParams,
   ) {
     super(params);
+    this.fileExclusions = config.getFileExclusions();
   }
 
   /**
@@ -281,7 +285,6 @@ class GrepToolInvocation extends BaseToolInvocation<
 
   /**
    * Gets a description of the grep operation
-   * @param params Parameters for the grep operation
    * @returns A string describing the grep
    */
   getDescription(): string {
@@ -391,7 +394,27 @@ class GrepToolInvocation extends BaseToolInvocation<
       if (grepAvailable) {
         strategyUsed = 'system grep';
         const grepArgs = ['-r', '-n', '-H', '-E'];
-        const commonExcludes = ['.git', 'node_modules', 'bower_components'];
+        // Extract directory names from exclusion patterns for grep --exclude-dir
+        const globExcludes = this.fileExclusions.getGlobExcludes();
+        const commonExcludes = globExcludes
+          .map((pattern) => {
+            let dir = pattern;
+            if (dir.startsWith('**/')) {
+              dir = dir.substring(3);
+            }
+            if (dir.endsWith('/**')) {
+              dir = dir.slice(0, -3);
+            } else if (dir.endsWith('/')) {
+              dir = dir.slice(0, -1);
+            }
+
+            // Only consider patterns that are likely directories. This filters out file patterns.
+            if (dir && !dir.includes('/') && !dir.includes('*')) {
+              return dir;
+            }
+            return null;
+          })
+          .filter((dir): dir is string => !!dir);
         commonExcludes.forEach((dir) => grepArgs.push(`--exclude-dir=${dir}`));
         if (include) {
           grepArgs.push(`--include=${include}`);
@@ -474,13 +497,7 @@ class GrepToolInvocation extends BaseToolInvocation<
       );
       strategyUsed = 'javascript fallback';
       const globPattern = include ? include : '**/*';
-      const ignorePatterns = [
-        '.git/**',
-        'node_modules/**',
-        'bower_components/**',
-        '.svn/**',
-        '.hg/**',
-      ]; // Use glob patterns for ignores here
+      const ignorePatterns = this.fileExclusions.getGlobExcludes();
 
       const filesStream = globStream(globPattern, {
         cwd: absolutePath,

packages/core/src/tools/read-many-files.test.ts

@@ -16,6 +16,10 @@ import { Config } from '../config/config.js';
 import { WorkspaceContext } from '../utils/workspaceContext.js';
 import { StandardFileSystemService } from '../services/fileSystemService.js';
 import { ToolErrorType } from './tool-error.js';
+import {
+  COMMON_IGNORE_PATTERNS,
+  DEFAULT_FILE_EXCLUDES,
+} from '../utils/ignorePatterns.js';
 import * as glob from 'glob';
 
 vi.mock('glob', { spy: true });
@@ -77,6 +81,13 @@ describe('ReadManyFilesTool', () => {
       getTargetDir: () => tempRootDir,
       getWorkspaceDirs: () => [tempRootDir],
       getWorkspaceContext: () => new WorkspaceContext(tempRootDir),
+      getFileExclusions: () => ({
+        getCoreIgnorePatterns: () => COMMON_IGNORE_PATTERNS,
+        getDefaultExcludePatterns: () => DEFAULT_FILE_EXCLUDES,
+        getGlobExcludes: () => COMMON_IGNORE_PATTERNS,
+        buildExcludePatterns: () => DEFAULT_FILE_EXCLUDES,
+        getReadManyFilesExcludes: () => DEFAULT_FILE_EXCLUDES,
+      }),
     } as Partial<Config> as Config;
     tool = new ReadManyFilesTool(mockConfig);
 
@@ -484,6 +495,13 @@ describe('ReadManyFilesTool', () => {
         }),
         getWorkspaceContext: () => new WorkspaceContext(tempDir1, [tempDir2]),
         getTargetDir: () => tempDir1,
+        getFileExclusions: () => ({
+          getCoreIgnorePatterns: () => COMMON_IGNORE_PATTERNS,
+          getDefaultExcludePatterns: () => [],
+          getGlobExcludes: () => COMMON_IGNORE_PATTERNS,
+          buildExcludePatterns: () => [],
+          getReadManyFilesExcludes: () => [],
+        }),
       } as Partial<Config> as Config;
       tool = new ReadManyFilesTool(mockConfig);
 

packages/core/src/tools/read-many-files.ts

@@ -15,7 +15,6 @@ import { getErrorMessage } from '../utils/errors.js';
 import * as fs from 'fs';
 import * as path from 'path';
 import { glob, escape } from 'glob';
-import { getCurrentGeminiMdFilename } from './memoryTool.js';
 import {
   detectFileType,
   processSingleFileContent,
@@ -98,49 +97,13 @@ type FileProcessingResult =
     };
 
 /**
- * Default exclusion patterns for commonly ignored directories and binary file types.
- * These are compatible with glob ignore patterns.
+ * Creates the default exclusion patterns including dynamic patterns.
+ * This combines the shared patterns with dynamic patterns like GEMINI.md.
  * TODO(adh): Consider making this configurable or extendable through a command line argument.
- * TODO(adh): Look into sharing this list with the glob tool.
  */
-const DEFAULT_EXCLUDES: string[] = [
-  '**/node_modules/**',
-  '**/.git/**',
-  '**/.vscode/**',
-  '**/.idea/**',
-  '**/dist/**',
-  '**/build/**',
-  '**/coverage/**',
-  '**/__pycache__/**',
-  '**/*.pyc',
-  '**/*.pyo',
-  '**/*.bin',
-  '**/*.exe',
-  '**/*.dll',
-  '**/*.so',
-  '**/*.dylib',
-  '**/*.class',
-  '**/*.jar',
-  '**/*.war',
-  '**/*.zip',
-  '**/*.tar',
-  '**/*.gz',
-  '**/*.bz2',
-  '**/*.rar',
-  '**/*.7z',
-  '**/*.doc',
-  '**/*.docx',
-  '**/*.xls',
-  '**/*.xlsx',
-  '**/*.ppt',
-  '**/*.pptx',
-  '**/*.odt',
-  '**/*.ods',
-  '**/*.odp',
-  '**/*.DS_Store',
-  '**/.env',
-  `**/${getCurrentGeminiMdFilename()}`,
-];
+function getDefaultExcludes(config?: Config): string[] {
+  return config?.getFileExclusions().getReadManyFilesExcludes() ?? [];
+}
 
 const DEFAULT_OUTPUT_SEPARATOR_FORMAT = '--- {filePath} ---';
 const DEFAULT_OUTPUT_TERMINATOR = '\n--- End of content ---';
@@ -172,7 +135,11 @@ ${this.config.getTargetDir()}
       .getGeminiIgnorePatterns();
     const finalExclusionPatternsForDescription: string[] =
       paramUseDefaultExcludes
-        ? [...DEFAULT_EXCLUDES, ...paramExcludes, ...geminiIgnorePatterns]
+        ? [
+            ...getDefaultExcludes(this.config),
+            ...paramExcludes,
+            ...geminiIgnorePatterns,
+          ]
         : [...paramExcludes, ...geminiIgnorePatterns];
 
     let excludeDesc = `Excluding: ${
@@ -230,7 +197,7 @@ ${finalExclusionPatternsForDescription
     const contentParts: PartListUnion = [];
 
     const effectiveExcludes = useDefaultExcludes
-      ? [...DEFAULT_EXCLUDES, ...exclude]
+      ? [...getDefaultExcludes(this.config), ...exclude]
       : [...exclude];
 
     const searchPatterns = [...inputPatterns, ...include];