[SMO-522] Fix heap overflow in large monorepo scans

[mikolalysenko]

Summary

• Add streaming-based filtering to globWithGitIgnore to prevent heap overflow when scanning large monorepos with 100k+ files
• Instead of accumulating all file paths and filtering afterwards, files are now filtered during streaming which dramatically reduces memory usage

Changes

• Add filter option to globWithGitIgnore for early filtering during streaming
• Add createSupportedFilesFilter helper to create filter from supported files
• Update getPackageFilesForScan to use streaming filter
• Add comprehensive tests for the new filter functionality

Testing

• Added unit tests for the new filter functionality
• All glob-related tests pass (21 tests)

Related Issue

Fixes SMO-522

🤖 Generated with Claude Code

---

Note

Massive project overhaul to a TypeScript + Rollup-based distribution with new tooling, packaging, and CI.

• Introduces rollup.base/dist/sea configs, Babel setup, and SEA bootstrap; adds transform/modify plugins and env inlining
• Adds bin entrypoints bin/cli.js, bin/{npm,npx,pnpm,yarn}-cli.js and dist/external bundling pipeline; vendors and patches deps under external/ and patches/
• Replaces old cli.js and lib/** with new src-driven build and outputs under dist/
• Revamps package.json (new name socket, engines, scripts, exports, overrides, lint-staged) and adds extensive build/test scripts
• Adds GitHub workflows for lint, test, e2e, provenance publish, auto PRs, and Claude automation
• Adds linting/formatting/type configs (eslint.config.js, biome.json, oxlint), EditorConfig, pnpm config, Husky hooks, env files, and ignore files
• Updates docs and metadata (README.md, CHANGELOG.md, CLAUDE.md, LICENSE)

^{Written by Cursor Bugbot for commit fe7d65d. Configure here.}

[cursor]

Comment @cursor review or bugbot run to trigger another review on this PR

[jdalton]

Ah ha! Good find in the negated glob!

[graphite-app]

The test expectation contradicts the negated gitignore pattern. The .gitignore contains ignored/ !ignored/package.json which means the ignored/ directory is ignored BUT ignored/package.json should be un-ignored and included in results. However, the test expects NO files from ignored/ to be present (expect(relativePaths.some(p => p.startsWith('ignored/'))).toBe(false)). This test will fail because ignored/package.json should be included due to the negated pattern.

Fix: Change the expectation to verify that ignored/package.json IS included:

expect(relativePaths).toContain('ignored/package.json')

Suggested change

	it('should handle negated gitignore patterns', async () => {
	const result = await globWithGitIgnore(['**/*'], {
	cwd: testDir,
	})
	const relativePaths = result.map(p => path.relative(testDir, p))
	// The ignored directory should be excluded.
	expect(relativePaths.some(p => p.startsWith('ignored/'))).toBe(false)
	})
	it('should handle negated gitignore patterns', async () => {
	const result = await globWithGitIgnore(['**/*'], {
	cwd: testDir,
	})
	const relativePaths = result.map(p => path.relative(testDir, p))
	// The ignored directory should be excluded except for the negated file
	expect(relativePaths).toContain('ignored/package.json')
	})

Spotted by Graphite Agent



Is this helpful? React 👍 or 👎 to let us know.