[nightshift] test-gap: Test Coverage Analysis

[nightshift-micr]

Test Coverage Gap Analysis: Microck/tailstick

Executive Summary

The tailstick project is a Go application for managing Tailscale-based USB/network device leases. It has 10 source files and 5 test files. While the existing tests are high-quality and cover several critical paths, there are significant gaps in test coverage across multiple packages.

Source-to-Test Mapping

Source File	Test File	Status
internal/tailscale/client.go	internal/tailscale/client_test.go	✅ Partial
internal/app/workflow.go	internal/app/workflow_test.go	⚠️ Partial
internal/app/cli.go	(none)	❌ No tests
internal/app/gui.go	(none)	❌ No tests
internal/state/store.go	(none)	❌ No tests
internal/config/config.go	internal/config/config_test.go	⚠️ Minimal
internal/crypto/secret.go	internal/crypto/secret_test.go	⚠️ Minimal
internal/gui/server.go	internal/gui/server_test.go	⚠️ Partial
internal/logging/logger.go	(none)	❌ No tests
internal/platform/exec.go	(none)	❌ No tests
internal/platform/platform.go	(none)	❌ No tests

4 of 10 source packages have zero test files. ~30 exported functions have no tests.

P0 — Critical (No test coverage for core logic)

P0-1: internal/state/store.go — Zero tests for all exported functions

All 4 exported functions (Load, Save, UpsertRecord, AppendAudit) have zero test coverage. State persistence is the source of truth for all lease data. A bug here means lost leases, double-cleanup, or missed cleanup.

P0-2: internal/platform/exec.go — Zero tests for Runner

Runner.Run() and Runner.RunWithTimeout() — empty command guard, dry-run mode, timeout behavior — all untested. Runner is the primary interface used by the tailscale client.

P0-3: internal/platform/platform.go — Zero tests

Detect(), sanitizeHost(), IsElevated(), RequireSupportedLinux(), path resolution functions — all untested. sanitizeHost edge cases: empty string, special characters, Unicode, all-invalid chars.

P1 — High (Missing tests for important business logic)

P1-1: internal/app/cli.go — RunCLI() completely untested

Command dispatch (run, agent, cleanup, version, help), runEnroll(), runAgent(), runCleanup() — all untested.

P1-2: internal/app/workflow.go — Enroll() has zero test coverage

The most complex function in the codebase (~125 lines) has no tests. Missing: privilege check, config loading, preset lookup, password validation, exit node validation, tailscale up, secret encryption, audit logging, agent installation.

P1-3: internal/config/config.go — Minimal test coverage

Only 1 test exists (38 lines). Missing: Validate() error paths, FindPreset() resolution, ResolvePath(), ResolvePresetSecrets() env var handling.

P1-4: internal/crypto/secret.go — Missing error path tests

Only happy-path roundtrip tests exist. Missing: wrong password, corrupted input, empty key material, corrupted nonce/salt/cipher.

P1-5: internal/tailscale/client.go — Multiple untested methods

IsInstalled(), EnsureInstalled(), Down(), Logout(), Status(), Uninstall(), ParseDurationDays(), Future() — all untested.

P2 — Medium (Missing edge cases and error paths)

• internal/gui/server.go: Missing enroll success path, non-POST method, invalid JSON body, Run() port validation
• internal/logging/logger.go: Zero tests for New(), Close(), Info(), Error()
• internal/tailscale/client.go: deleteDevice edge cases — empty apiKey/deviceID silent nil return

P3 — Low (Nice-to-have)

• No CI integration for sandbox E2E tests
• Duplicated fake tailscale binary pattern across test files → extract to internal/testutil/
• Several functions would benefit from table-driven tests: sanitizeHost(), ParseDurationDays(), validateExitNode(), FindPreset()

Quantitative Summary

Category	Count
Packages with zero test files	4
Source files with zero test coverage	5
Exported functions with no tests	~30
Error handling paths untested	~25
Edge cases untested	~20

Recommended Priority Actions

1. [P0] Add internal/state/store_test.go — Test Load, Save, UpsertRecord, AppendAudit with real temp files
2. [P0] Add internal/platform/exec_test.go — Test Runner.Run() with dry-run, empty command, RunWithTimeout()
3. [P0] Add internal/platform/platform_test.go — Test sanitizeHost() with edge cases, path resolution
4. [P1] Expand config_test.go — Cover Validate, FindPreset, ResolvePath, ResolvePresetSecrets with table-driven tests
5. [P1] Add crypto error path tests — wrong password, corrupted input, empty key material
6. [P2] Add internal/logging/logger_test.go — Basic New/Close/Write test with temp file
7. [P3] Create internal/testutil/ package to consolidate shared test helpers

---

Generated by Nightshift v3 (GLM 5.1)

[Microck]

Partially addressed by #43 with added GUI and tailscale regression coverage. Closing this broad generated test-gap report in favor of opening focused test tasks when specific behavior needs coverage.