init observations

[Avi Kivity]

Kenneth Porter wrote:

> --On Tuesday, November 15, 2005 3:44 PM -0500 Luke Macken 
> <lmacken at redhat.com> wrote:
>
>>  o Still supports the /etc/rc.d/rc$i.d framework (which is horrible by
>>    design)
>
>
> What do packagers of services need to do to natively support the 
> alternatives? Presumably they need to supply the native equivalent to 
> the init.d script.
>
It would be excellent if the glue code could be done in python:
- fast compared to shell scripts since there is no need to continually 
fork/exec utilities
- non-intrusive: less patches to packages
- easy to write and maintain

I imagine an initng plugin should be feasible.

[there might be opposition to linking init with /usr/, but I consider it 
outdated. we have initrd to mount our filesystems today]

-- 
Do not meddle in the internals of kernels, for they are subtle and quick to panic.