Maintaining Arch Linux packages can be a time-consuming task, especially when you have to keep track of upstream updates across various platforms.

The Old Way: GitHub Watch

Previously, my main method for staying informed about new releases was using the Watch (watch -> custom -> releases) feature on GitHub. While this worked reasonably well for projects hosted on GitHub, it had significant limitations:

• It doesn’t scale well when you maintain many packages.
• It’s difficult to track projects hosted on other forges like GitLab, Codeberg, or self-hosted instances.
• It requires manual intervention for every single update.
• It does not work with packages repository like pypi of rubygems.

The New Workflow: nvchecker

My current workflow is much more automated and forge-agnostic. It relies on nvchecker and a set of custom scripts from my dotfiles.

1. Cloning and Updating Package Repositories

I use two main scripts to manage my local package repositories:

• clone_aur_package_by_maintainer: To bulk clone all packages I maintain.
• update_git_dir: To keep all my local clones up to date.

2. Version Tracking with .nvchecker.toml

Each package directory contains a .nvchecker.toml file that describes how to check for the latest upstream version. I generate this file once using gen_dot_nvchecker_toml.

3. Daily Update Checks

Every day, I run a routine to check for updates:

1. gen_nvchecker_config: This script scans my package directories and generates a global nvchecker configuration.
2. nvchecker: I run the tool to compare local versions with upstream.

4. Automation with nvupdate

When an update is detected, I use nvupdate. This script automates several steps:

• Updating the PKGBUILD with the new version and checksums.
• Building the package to ensure it still compiles.
• Running basic tests.

Conclusion

While the heavy lifting of detecting and preparing updates is now automated, the final validation and pushing to the AUR remain manual steps. This ensures that I still have a “human in the loop” to catch any potential issues before they reach other users.