project

Overview ¶

Package project provides the core domain logic for managing multi-repository workspaces.

It is designed to cleanly separate the command-line interface (CLI) concerns from the underlying business logic of fetching, pulling, syncing, and setting up projects. By abstracting git operations behind the RepoClient interface, this package enables fast, deterministic, and reliable operations across multiple repositories concurrently.

Key concepts and features:

• **Options Structs:** Each primary operation (Setup, Sync, Pull, Fetch, List) is configured via an `Options` struct. These structs contain pure data, cleanly decoupling the package from global configuration parsers like Viper.

• **RepoClient Abstraction:** All external side effects (e.g., executing git clone, git fetch, checking for dirty working trees) are abstracted behind the `RepoClient` interface. If no client is explicitly provided in the options, a default OS-based client is used. This architecture enables high test coverage without hitting the network or creating real repositories on disk.

• **Concurrency & Resilience:** Operations such as Sync, Pull, and Fetch are designed to run in parallel using an `errgroup`, capped at a safe concurrency limit. Interactive prompts and network timeouts are strictly disabled to prevent the CLI from hanging indefinitely on network or authentication issues.

• **Safety First:** The default workflow respects the developer's working tree. Pull and Sync operations skip repositories that are "dirty" (contain uncommitted changes) to prevent merge conflicts or interrupted rebases mid-sync. Furthermore, operations pull the *current* branch rather than forcing checkouts to the default branch.

Example Usage:

opts := project.SyncOptions{
	DevPath:  "/path/to/dev",
	Projects: myConfiguredProjects,
	DryRun:   false,
}

// Executes a concurrent sync, pulling the latest code for all projects
// without uncommitted changes.
project.Sync(context.Background(), opts)