Re: [PR PATCH] [Merged]: [RFC] use alternatives to manage initramfs regeneration

[ahesford <ahesford@users.noreply.github.com>]

[-- Attachment #1: Type: text/plain, Size: 1314 bytes --]

There's a merged pull request on the void-packages repository

[RFC] use alternatives to manage initramfs regeneration
https://github.com/void-linux/void-packages/pull/39284

Description:
Yes, I know that XBPS alternatives have some problems, but this seems to be the most sensible approach on balance. Rather than dealing with a bunch of potentially competing initramfs hooks, we can let all packages install custom hooks in `/usr/libexec` and use the alternatives system to symlink the preferred hooks in `/etc/kernel.d`. At that point, we can use the generic alternative script in the XBPS regenerator trigger and avoid the need for a conf file, although it falls back to legacy behavior if the alternative isn't set.

Another option I considered, but decided against, was creating a centralized `initramfs-kernel-hooks` package that installs a common hook and reads the `/etc/default/initramfs-regenerate` file used in the existing XBPS trigger to determine what to execute. All initramfs generators (currently only dracut and mkinitcpio) would have to depend on the common package. This has the drawback that all of the generation knowledge must be centralized instead of located in the package that adds the functionality.

This is not really tested at this point, but I want to get some early comments.