Skip to content

PUBLISH.md

Latest commit

 

History

History
175 lines (131 loc) · 3.75 KB

PUBLISH.md

File metadata and controls

175 lines (131 loc) · 3.75 KB

Publishing to crates.io

This document explains how to publish rust_tui_coder to crates.io.

Pre-Publication Checklist

All tests pass

cargo test
# Result: 94 tests passed across 9 test suites

No clippy warnings

cargo clippy --all-targets --all-features
# Result: Clean build with no warnings

Release build succeeds

cargo build --release
# Result: Successful compilation

Package metadata is complete

  • Name: rust_tui_coder
  • Version: 1.0.0
  • Description: Comprehensive and descriptive
  • License: MIT OR Apache-2.0
  • Keywords: Relevant and within limit (5)
  • Categories: Appropriate
  • README: Present
  • Authors: Specified
  • Rust version: 1.70 minimum

Binary configuration

[[bin]]
name = "rust_tui_coder"
path = "src/main.rs"

Installation After Publishing

Users will be able to install the binary using:

cargo install rust_tui_coder

This will install the rct binary to ~/.cargo/bin/ (which should be in PATH).

Usage After Installation

After installation, users can run:

rct

The application will look for a config.toml file in the current directory.

Publishing Steps

  1. Login to crates.io (one time only):

    cargo login <your-api-token>

    Get your API token from: https://crates.io/settings/tokens

  2. Verify package contents:

    cargo package --list

    This shows what files will be included.

  3. Dry run the publish:

    cargo publish --dry-run

    This verifies everything without actually publishing.

  4. Publish to crates.io:

    cargo publish

Post-Publication

After publishing, the package will be available at:

  • Main page: https://crates.io/crates/rust_tui_coder
  • Documentation: https://docs.rs/rust_tui_coder
  • Repository: Update the repository URL in Cargo.toml with actual GitHub URL

Version Updates

For future releases:

  1. Update version in Cargo.toml
  2. Run tests: cargo test
  3. Update CHANGELOG (if you create one)
  4. Commit changes
  5. Create git tag: git tag v1.0.1
  6. Publish: cargo publish
  7. Push tag: git push origin v1.0.1

Excluded Files

The following are excluded from the published package (see Cargo.toml):

  • config.toml (user-specific configuration)
  • plan.md (runtime generated file)
  • .git and .gitignore (version control)
  • tmp_rovodev_* (test artifacts)
  • *.png (images, already in README)

Important Notes

  1. Configuration Required: Users need to create a config.toml file with their LLM API credentials
  2. First Run: Include instructions in README for first-time setup
  3. Binary Name: The binary is named rct and can be renamed after installation if desired
  4. Dependencies: All dependencies are properly specified in Cargo.toml
  5. Platform Support: Works on Linux, macOS, and Windows (tested on Linux)

Testing the Installation

After publishing, test the installation in a clean environment:

# In a new directory
cargo install rust_tui_coder

# Create config file
cat > config.toml << EOF
[llm]
api_key = "your-api-key"
api_base_url = "https://api.openai.com/v1"
model_name = "gpt-4"
EOF

# Run the application
rct

Support and Issues

Direct users to:

  • GitHub repository for issues
  • README.md for documentation
  • config_example.toml for configuration template

Yanking a Version (if needed)

If you need to yank a published version due to critical bugs:

cargo yank --vers 1.0.0

To un-yank:

cargo yank --vers 1.0.0 --undo

Current Status

Ready for Publication

All prerequisites have been met:

  • Comprehensive test suite (94 tests)
  • No clippy warnings
  • Clean build
  • Proper package metadata
  • Documentation included
  • Binary configuration correct