Created Homebrew formula & release hash workflow

[LaurentFough]

Hey;

Very interesting project. Been using/testing it for a a couple of days; and started picking off items on the roadmap.

I've created a working Homebrew formula (+makefile), plus a Github Actions release file, to generate shasums, for releases.

• for the Homebrew file, you should institute some form of versioning.

You can test it first, then I can create a proper PR.

• Repo here: https://github.com/LaurentFough/lazyshell

• ** There are some other small code tweaks that I can push; as well.

[not-poma]

Thanks for the contribution, I never worked with homebrew before

1. Shouldn't the homebrew file reference the dependencies (curl, jq)?
2. I feel like it should try to add the lines to .zshrc automatically (and remove them upon uninstall), and only list caveat about OpenAI key. Some users might even have the API key in their environment as the number of different AI tools grows rapidly. Although in this case we probably should suppress the API key warning on startup.
3. What this regex does?
4. Even the smallest improvements are welcome!

[LaurentFough]

Thanks for the contribution, I never worked with homebrew before

1. Shouldn't the homebrew file reference the dependencies (curl, jq)?
2. I feel like it should try to add the lines to .zshrc automatically (and remove them upon uninstall), and only list caveat about OpenAI key. Some users might even have the API key in their environment as the number of different AI tools grows rapidly. Although in this case we probably should suppress the API key warning on startup.
3. What this regex does?
4. Even the smallest improvements are welcome!

@not-poma Great feedback, thanks for taking a look.

• 1. Actually yes; I should add some depends_on for curl & jq.
• 2. I can't recall the current Homebrew stance on modification of files, etc — but I do know that most of the tools that I install ask/require the user to make the mods themselves.
• • we definitely don't want to be in a situation where we make changes to a users shell init, those changes get edited and during uninstall we can't uninstall correctly as we don't know what to look for.
• • we should leave the users system as clean as possible, when we interact with it.
• 3. That regex is there to remove build-status references from the README.md, if you go that route.
• 4. I have some other tweaks in my fork
• • creation of a $XDG_CONFIG_HOME/lazyshell/config file; and reading in the API_KEY, and keybindings from it, etc.

[LaurentFough]

I also have a clean method of using wget for the API call; then failing over to curl if needed.

[xPMo]

• we definitely don't want to be in a situation where we make changes to a users shell init, those changes get edited and during uninstall we can't uninstall correctly as we don't know what to look for.

imho: Asking, then adding as [[ -r $pathto/lazyshell.zsh ]] && ... is reasonable.

[not-poma]

I also have a clean method of using wget for the API call; then failing over to curl if needed.

Forgot to mention that I've implemented this fallback after you mentioned it.

imho: Asking, then adding as [[ -r $pathto/lazyshell.zsh ]] && ... is reasonable.

brew installs are usually non-interactive, not sure where would be a good place to ask this.