[Dev] [RFC] [nonsystemd] repository, packages built without systemd support

[Bruno Cichon]

I like to have 1 base and ask for init system.

bill-auger <bill-auger at peers.community> writes:

> i dont know how difficult this idea may be, but it seems like the
> logical user-friendly thing to do would be to have a separate package
> for each init system, including a new one with only systemd - each of
> which as a provider of 'your-init-freedom'; and base would depend on any
> one of those providers
>
> 'init-systemd'  provides('your-init-freedom')
>                 depends('systemd-sysvcompat')
>                 conflicts('init-openrc' 'init-runit' 'init-shepherd')
> 'init-openrc'   provides('your-init-freedom')
>                 depends(whatever 'base-openrc' depends now)
>                 conflicts('init-systemd' 'init-runit' 'init-shepherd')
> 'init-runit'    provides('your-init-freedom')
>                 depends('runit')
>                 conflicts('init-systemd' 'init-openrc' 'init-shepherd')
> 'init-shepherd' provides('your-init-freedom')
>                 depends('shepherd')
>                 conflicts('init-systemd' 'init-openrc' 'init-runit')
>
> as it is now, 'base' depends on 'systemd' and any other init package
> would need to remove 'systemd' so instead there is base-openrc that
> removes 'base' - but that seems very tacky to be removing 'base' or
> anything from it for any reason - it is most sensible to have only one
> 'base' package that allows for optional components
>
> # pacstrap /mnt 'base'
> There are 4 providers of 'your-init-freedom'. Please select one:
> 1) 'init-systemd'
> 2) 'init-openrc'
> 3) 'init-runit
> 4) 'init-shepherd')
>
> i dont know if a new repo is really need for this but i dont know why
> there is a separate repo for kernels either - maybe this is suggesting a
> similar schema
>
> _______________________________________________
> Dev mailing list
> Dev at lists.parabola.nu
> https://lists.parabola.nu/mailman/listinfo/dev