E2E Tests

End-to-end tests for gh-stack that compile and run the actual binary against isolated git repositories.

Running

# Run all E2E tests
make e2e

# Run specific test
go test -v ./e2e/ -run TestCascade

# Run with race detection
go test -race -v ./e2e/...

# Run unit tests only (faster)
make test-unit

# Run everything
make test

Architecture

Files

• main_test.go - TestMain compiles binary once before all tests
• helpers_test.go - TestEnv struct, Result type, core helpers
• scenario_helpers_test.go - Conflict and remote simulation helpers
• smoke_test.go - Basic smoke test
• init_create_test.go - Init and create command tests
• cascade_test.go - Restack tests including conflict recovery
• adopt_orphan_test.go - Adopt and orphan command tests
• push_test.go - Push command tests
• error_cases_test.go - Error handling tests
• chaos_manual_git_test.go - Manual git operation resilience
• chaos_remote_test.go - Remote sync scenario tests

How It Works

1. TestMain compiles gh-stack binary once to a temp directory
2. Each test creates an isolated git repo via NewTestEnv(t) using t.TempDir()
3. Tests execute the binary via subprocess and verify behavior
4. Temp directories are automatically cleaned up by Go's testing framework

TestEnv

TestEnv provides:

• Run(args...) - Execute gh-stack, returns Result (doesn't fail on error)
• MustRun(args...) - Execute gh-stack, fails test on non-zero exit
• Git(args...) - Execute git commands directly
• WriteFile(name, content) - Create files in the test repo
• CreateCommit(msg) - Create a commit with a unique file
• Assert*() - Various assertions (branch, ancestry, rebase state)

Test Categories

Category	Description
Happy path	Normal workflows (init, create, restack, push)
Error recovery	Conflicts, abort/continue, dirty working tree
Chaos tests	Manual git operations, remote sync edge cases

Writing Tests

func TestMyScenario(t *testing.T) {
    // Create isolated git repo
    env := NewTestEnv(t)  // or NewTestEnvWithRemote(t) for push tests

    // Initialize stack
    env.MustRun("init")

    // Use env.Run() for commands that might fail
    result := env.Run("some-command")
    if result.Failed() {
        // handle or check error
    }

    // Use env.MustRun() for commands that must succeed
    env.MustRun("create", "feature-1")

    // Use env.Git() for direct git operations
    env.Git("checkout", "main")

    // Use assertions
    env.AssertBranch("feature-1")
    env.AssertStackParent("feature-1", "main")
    env.AssertAncestor("main", "feature-1")
}

CI

E2E tests run as part of the standard test suite in GitHub Actions. The CI workflow configures git identity for the tests:

- name: Configure git for E2E tests
  run: |
    git config --global user.email "ci@github.com"
    git config --global user.name "CI"
    git config --global init.defaultBranch main