[jez]

Has Neovim ever considered to a rolling release model?

Where I work, it’s considerably less of a headache to install Homebrew bottles instead of source/HEAD packages. There was a bug in Neovim 0.4 that I ran into every day for years—it was fixed in master but unreleased for a long time. I just had the same thing happen in 0.5.0: there’s a bug, it’s already fixed in master, but I can’t upgrade to HEAD.

The 0.5.0 release was obviously a historic release (tree sitter integration, LSP built-in, lua config engine) but the release was delayed for a long time. I’d love to know the team’s thoughts for or against having some sort of rolling release schedule (or other strategy) to release earlier and/or more frequently.

[tjdevries]

Rolling release is kind of hard. We want to make sure that we keep our promises that we've set forth in `:help api-contract` and other expectations about semver. We think this is very beneficial for the plugin writing community. It's also hard because sometimes we merge something into master, before a release, and then recognize that a change to an API, an extension, etc. may be necessary for a feature to truly be great. Rolling release can make some of these things very hard.

However, we're hoping to release more often now to prevent this kind of situation.

I'm not 100% the person to talk to about releases, I primarily just try and write code to make Lua more fun to use :)

Side Note: You can install the HEAD of neovim easily with brew. `brew install neovim --HEAD` I think will do that for you. I'm not sure if that's what you mean by source/HEAD packages or not (I have no idea anything about Mac).