Skip to content

Latest commit

Β 

History

History
638 lines (449 loc) Β· 16.9 KB

File metadata and controls

638 lines (449 loc) Β· 16.9 KB

submod

Crates.io Documentation Static Badge Rust codecov Crates.io Downloads (latest version)

Git submodules solve a real problem. Managing submodules is a pain. You use them infrequently enough that you always forget which command does what β€” and when something breaks, the recovery steps are a small nightmare. New contributors hit this especially hard: onboarding onto a project that uses submodules is its own obstacle course.

submod wraps the whole lifecycle in one consistent CLI. Sixteen commands, including nuke-it-from-orbit for when you're done being reasonable. Built on gitoxide and git2, with automatic fallback so operations don't fail silently.1 It's actively used across @knitli and @plainlicense, where submodules handle shared functionality between repos.

πŸš€ Features

  • TOML config β€” define submodules, sparse-checkout paths, and defaults in one file
  • Sparse checkout β€” clone only the parts of a submodule you actually need
  • Global defaults with per-submodule overrides β€” set it once, customize where it matters
  • Fallback chain β€” tries gitoxide first, falls back to git2, then CLI
  • Clear status and errors β€” you'll know what broke and why

πŸ“‹ Table of Contents

πŸ”§ Installation

Using Cargo

cargo install submod

Using Mise

Mise is a project management tool and package manager that can manage your development environment.

# Global installation
mise use -g cargo:submod@latest

# Project-specific installation
mise use cargo:submod@latest

From Source

git clone https://github.com/yourusername/submod.git
cd submod
cargo install --path .

πŸš€ Quick Start

  1. Initialize a config file in your git repository:

    # Create a basic submod.toml configuration
    cat > submod.toml << EOF
    [defaults]
    ignore = "dirty"
    
    [my-submodule]
    path = "vendor/my-lib"
    url = "https://github.com/example/my-lib.git"
    sparse_paths = ["src/", "include/", "*.md"]
    EOF
  2. Initialize your submodules:

    submod init
  3. Check status:

    submod check

βš™οΈ Configuration

Create a submod.toml file in your repository root:

# Global defaults applied to all submodules
[defaults]
ignore = "dirty"          # ignore dirty state in status
update = "checkout"       # update method
branch = "main"          # default branch to track

# Individual submodule configuration
[vendor-utils]
path = "vendor/utils"
url = "https://github.com/example/utils.git"
sparse_paths = ["src/", "include/", "*.md"]
ignore = "all"           # override default ignore setting
active = true            # whether submodule is active

[my-submodule]
path = "libs/my-submodule"
url = "https://github.com/example/my-submodule.git"
sparse_paths = ["src/core/", "docs/"]
branch = "develop"       # track specific branch

Configuration Options

Global Defaults

  • ignore: How to handle dirty submodules (all, dirty, untracked, none)
  • update: Update strategy (checkout, rebase, merge, none, !command)
  • branch: Default branch to track (. for current superproject branch)
  • fetchRecurse: Fetch recursion (always, on-demand, never)

Per-Submodule Settings

  • path: Local path where submodule should be placed
  • url: Git repository URL
  • sparse_paths: Array of paths to include in sparse checkout
  • active: Whether the submodule is active (default: true)
  • All global defaults can be overridden per submodule

πŸ“– Commands

submod add

Add a new submodule to your configuration and repository:

# Basic add
submod add https://github.com/example/my-lib.git --name my-lib --path libs/my-lib

# With sparse checkout paths and extra options
submod add https://github.com/example/my-lib.git \
  --name my-lib \
  --path libs/my-lib \
  --sparse-paths "src/,include/" \
  --branch main \
  --ignore all \
  --fetch on-demand

Options:

Flag Short Description
<URL> (required) URL or local path of the submodule repository
--name -n Nickname for the submodule used in your config and commands
--path -p Local directory path where the submodule should be placed
--branch -b Branch to track
--ignore -i Dirty-state ignore level (all, dirty, untracked, none)
--sparse-paths -x Comma-separated sparse checkout paths or globs
--fetch -f Recursive fetch behavior (always, on-demand, never)
--update -u Update strategy (checkout, rebase, merge, none)
--shallow -s Shallow clone (last commit only)
--no-init Add to config only; do not clone/initialize

submod check

Check the status of all configured submodules:

submod check

alias: submod c

submod init

Initialize all missing submodules:

submod init

alias: submid i

submod update

Update all submodules to their latest commits:

submod update

alias: submod u

submod reset

Hard reset submodules (stash changes, reset --hard, clean):

# Reset all submodules
submod reset --all

# Reset specific submodules (comma-separated)
submod reset my-lib,vendor-utils

alias: submod r

submod sync

Run a complete sync (check + init + update):

submod sync

alias: submod s

submod change

Change the configuration of an existing submodule:

submod change my-lib --branch main --sparse-paths "src/,include/" --fetch always

submod change-global

Change global defaults for all submodules:

submod change-global --ignore dirty --update checkout

aliases: submod cg, submod chgl, submod global

submod list

List all configured submodules:

submod list
submod list --recursive

aliases: submod ls, submod l

submod delete

Delete a submodule from configuration and filesystem:

submod delete my-lib

alias: submod del

submod disable

Disable a submodule without deleting files (sets active = false):

submod disable my-lib

alias: submod del

submod nuke-it-from-orbit

Delete all or specific submodules from config and filesystem, with optional reinit:

# Nuke all submodules (re-initializes by default)
submod nuke-it-from-orbit --all

# Nuke specific submodules permanently
submod nuke-it-from-orbit --kill my-lib,old-dep

aliases: submod nuke-em, submod nuke-it, submod nuke-them

Use nuke-it-from-orbit was created because sometimes submodule just... don't cooperate. You're done being nice and just want to get back to work. Nuke it.

submod generate-config

Generate a new configuration file:

# From current git submodule setup
submod generate-config --from-setup .

# As a template with defaults
submod generate-config --template --output my-config.toml

aliases: submod gc, submod genconf

submod completeme

Generate shell completion scripts:

bash

mkdir -p ~/.bash_completion.d
submod completeme bash > ~/.bash_completion.d/submod

aliases: submod comp, submod complete, submod comp-me, submod complete-me

Add completions to other shells

zsh

# zsh has an fpath array with possible function directories. You can
# put your competions in any of these; we use the first one here:
ZSH_DEFAULT="${XDG_DATA_HOME:-~/.local/share}/zsh/site-functions"
ZFUNCDIR=""${fpath[1]:-$ZSH_DEFAULT}"
mkdir -p "$ZFUNCDIR"
submod completeme zsh > "${ZFUNCDIR}/_submod"

fish

mkdir -p "${XDG_CONFIG_HOME:-~/.config}/fish/completions"
submod completeme fish > "{XDG_CONFIG_HOME:-~/.config}/fish/completions/submod.fish

powershell

mkdir -p $Home\Documents\PowerShell\completions
submod completeme powershell > $Home\Documents\PowerShell\completions\submod.completion.ps1

elvish

mkdir -p ~/.config/elvish/completions
submod completeme elvish > ~/.config/elvish/completions/submod.elv

nushell

submod completeme nu > "$NUSHELL_CONFIG_DIR/scripts/completions/submod.nu"
echo 'use completions/submod.nu' >> "$NU_CONFIG_PATH"

πŸ’» Usage Examples

Basic Workflow

# Start with checking current state
submod check

# Initialize any missing submodules
submod init

# Update everything to latest
submod update

# Or do it all at once
submod sync

Adding Submodules with Sparse Checkout

# Add a submodule that only checks out specific directories
submod add https://github.com/company/react-components.git \
  --name react-components \
  --path src/components \
  --sparse-paths "src/Button/,src/Input/,README.md"

Working with Different Configurations

# Use a custom config file
submod --config my-custom.toml check

# Check status with custom config
submod --config production.toml sync

Handling Problematic Submodules

# Reset a problematic submodule
submod reset my-problematic-submodule

# Check what's wrong
submod check

# Re-sync everything
submod sync

If that doesn't work, try nuke-it-from-orbit

πŸ› οΈ Development

Prerequisites

  • Rust 1.87 or later
  • Git
  • Mise (recommended) - for tool management and task running

Quick Setup with Mise (Recommended)

# Clone the repository
git clone https://github.com/bashandbone/submod.git
cd submod

# Install mise if you haven't already
curl https://mise.run | sh

# Install all development tools and dependencies
mise install

# Build the project
mise run build
# or: mise run b (alias)

# Run tests
mise run test

# Run the full CI suite (build + lint + test)
mise run ci

Available Mise Tasks

# Build the project
mise run build          # or: mise run b

# Run tests
mise run test

# Lint with clippy
mise run lint

# Run full CI pipeline
mise run ci

# Clean build artifacts
mise run clean

# Cut a new release (maintainers only)
mise run release

Git Hooks with hk

This project uses hk for automated git hooks that ensure code quality:

# Install git hooks (done automatically with mise install)
hk install

# Run pre-commit checks manually
hk run pre-commit

# Run all linters and checks
hk check

# Auto-fix issues where possible
hk fix

# Run CI checks locally
hk run ci

The pre-commit hooks automatically run:

  • cargo fmt - Code formatting
  • cargo clippy - Linting
  • cargo test - Test suite
  • typos - Spell checking
  • prettier - TOML/YAML formatting
  • cargo deny - Security and license auditing

Manual Setup (Alternative)

If you prefer not to use mise:

# Clone the repository
git clone https://github.com/bashandbone/submod.git
cd submod

# Install Rust if needed
curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh

# Build the project
cargo build

# Run tests
cargo test
# or hk run test

# Or use the comprehensive test runner
./scripts/run-tests.sh --verbose

Running Tests

# Using mise (recommended)
mise run test           # Run all tests
mise run ci             # Run full CI suite

# Using hk
hk run test                 # Run tests only
hk run ci                   # Run CI checks

# Using cargo directly
cargo test              # Run all tests
cargo test --test integration_tests  # Integration tests only

# Using the test script
./scripts/run-tests.sh --verbose     # Comprehensive reporting
./scripts/run-tests.sh --performance # Include performance tests
./scripts/run-tests.sh --filter sparse_checkout  # Filter tests

Project Structure

submod/
β”œβ”€β”€ src/
β”‚   β”œβ”€β”€ main.rs              # CLI entry point
β”‚   β”œβ”€β”€ commands.rs          # Command definitions (clap)
β”‚   β”œβ”€β”€ config.rs            # TOML configuration handling
β”‚   β”œβ”€β”€ git_manager.rs       # High-level submodule operations
β”‚   └── git_ops/             # Git backend abstraction
β”‚       β”œβ”€β”€ mod.rs           # GitOpsManager (gixβ†’git2β†’CLI fallback)
β”‚       β”œβ”€β”€ gix_ops.rs       # gitoxide backend
β”‚       └── git2_ops.rs      # libgit2 backend
β”œβ”€β”€ tests/                   # Integration tests
β”œβ”€β”€ sample_config/           # Example configurations
β”œβ”€β”€ scripts/                 # Development scripts
└── docs/                    # Documentation

🀝 Contributing

We welcome contributions! Please see our Contributing Guide for details.

Quick Contributing Steps

  1. Fork the repository
  2. Create a feature branch: git checkout -b feature/amazing-feature
  3. Set up development environment: mise install (installs all tools and git hooks)
  4. Make your changes and add tests if applicable
  5. Commit your changes: git commit -m 'Add amazing feature' (hooks run automatically)
  6. Push to your branch: git push origin feature/amazing-feature (they'll actually run again in check mode, so they need to pass)
  7. Open a Pull Request

Development Guidelines

  • Follow Rust best practices and idioms
  • Add tests for new functionality. I'm not big on unit tests, but integration tests are essential.
  • Update documentation for user-facing changes
  • Use conventional commit messages
  • Run mise run ci or hk run ci before submitting PR
  • Pre-commit hooks will automatically format code and run basic checks
  • All automated checks must pass before PR can be merged

πŸ” Troubleshooting

Common Issues

Submodule not initializing:

# Check if the URL is accessible
git ls-remote <submodule-url>

# Verify your configuration
submod check

Sparse checkout not working:

  • Ensure paths in sparse_paths are relative to the submodule root
  • Check that the submodule repository contains the specified paths
  • Verify sparse checkout is enabled: git config core.sparseCheckout in the submodule

Permission issues:

  • Ensure you have proper SSH keys set up for private repositories
  • Check if your Git credentials are configured correctly

πŸ“‹ Motivation

Managing git submodules, especially with sparse checkouts, can be complex and error-prone. Traditional git submodule commands require multiple steps and careful attention to configuration details.

This tool was created to:

  • Reduce barriers to contribution - Make it easier for new developers to work with projects using submodules
  • Simplify complex workflows - Handle initialization, updates, and sparse checkout configuration automatically
  • Provide better tooling - Clear status reporting and error messages
  • Leverage modern Git libraries - Use gitoxide for better performance and reliability

The tool is actively used in multiple projects at @knitli and @plainlicense, where submodules are essential for sharing core functionality across repositories.

πŸ“„ License

This project is licensed under the Plain MIT License.

πŸ™ Acknowledgments

  • gitoxide - Fast and safe pure Rust implementation of Git
  • git2-rs - Rust bindings to libgit2
  • clap - Command line argument parser

Homepage β€’ Documentation β€’ Crate

Made with ❀️ for the Rust and Git communities

Footnotes

  1. The fallback architecture is more a reflection of the status of gitoxide and git2 submodule support than a stability concern. Their features do not consistently provide the full lifecycle of submodule operations. Together they cover >90%, but sometimes immaturely. The fallbacks architecture handles that gracefully and lets native operations grow with those libraries. ↩