Provide testing utilities and mock implementations

[coolguy1771]

Proposal

The library currently lacks built-in testing support, forcing users to implement their own mocks or manage complex infrastructure for integration tests. I propose adding a testutil package to provide a flexible MockClient and a TestServer helper.

Proposed Solution

The testutil package will focus on two areas:

1. Unit Testing: A function-based MockClient that allows users to inject custom behavior per test case without implementing the entire interface manually.
2. Integration Testing: A TestServer that wraps httptest.Server to simulate the scaleset API, allowing users to verify their full networking stack.

Mock Implementation

package testutil

import (
    "context"
    "github.com/actions/scaleset"
)

// MockClient provides a flexible mock for the listener.Client interface.
type MockClient struct {
    GetMessageFunc    func(ctx context.Context, lastID, cap int) (*scaleset.RunnerScaleSetMessage, error)
    DeleteMessageFunc func(ctx context.Context, messageID int) error
    SessionFunc       func() scaleset.RunnerScaleSetSession
    CloseFunc         func(ctx context.Context) error
}

func (m *MockClient) GetMessage(ctx context.Context, lastID, cap int) (*scaleset.RunnerScaleSetMessage, error) {
    if m.GetMessageFunc != nil {
        return m.GetMessageFunc(ctx, lastID, cap)
    }
    return nil, nil
}

// ... other interface methods implemented similarly ...

Integration Helper

// TestServer simulates the scaleset API for end-to-end testing.
type TestServer struct {
    URL    string
    Client *scaleset.Client
}

func NewTestServer(t *testing.T, options ...TestServerOption) *TestServer
func (s *TestServer) EnqueueMessage(msg *scaleset.RunnerScaleSetMessage)
func (s *TestServer) GetReceivedDeleteRequests() []int

Benefits

• Reduced Boilerplate: Users don't need to redefine interfaces or mock logic in every downstream project.
• Reliable Integration: The TestServer ensures that user-implemented scalers handle real HTTP interactions, including headers and status codes, correctly.
• Consistency: Provides a standard way to test scaleset logic across different implementations.