[arch-projects] [initscripts][netcfg] deprecating advanced network functionality from initscripts

[Dan McGee]

On Tue, Apr 26, 2011 at 1:37 PM, Tom Gundersen <teg at jklm.no> wrote:
> On Tue, Apr 26, 2011 at 7:55 PM, Dave Reisner <d at falconindy.com> wrote:
>>> I'd vote for #2, keeping old timers (and me) happy. It seems like the
>>> right line to draw so that someone using a "normal" desktop computer
>>> setup, a VM, a VPS, etc. over Ethernet doesn't need to suddenly dive
>>> into netcfg. This would eliminate a lot of code still: wi_up,
>>> bond_up/down, bridge_up/down.
>>
>> I'll vote for #2 as well.
>
> Sounds like #2 it is then :-)
>
>> Since this involves some amount of dramatic
>> change, can I be annoying and harp on my iproute2 patchwork as well?
>> Seems like it would be an appropriate time to merge, since we're
>> breaking other things as well.
>
> Yes, that would be great! May I suggest that we no longer allow "open
> ended" configuration variables, but do something like this:
>
> interface="eth0"
> ip="192.168.0.10"
> broadcast=""
> netmask""
> gateway=""
>
> (removing the ones that are uncommon to change, and adding anything I forgot)
>
> Only allow one interface, if ip is unset then use dhcp, and if the
> other variables are unset use sensible standards. Parse the variables
> for correctness to stop people from doing weird things.
>
> Does anyone object to this plan?

I do- making a system unbootable without a config change is never a
good idea. My hope with selection number 2 was that no one using a
"normal" config would have to change anything, e.g. those using
headless systems would not have to tweak any dials.

If I'm of the minority opinion here I will concede. I will also be
sure to assign the "my system isn't responding and I can't SSH in" bug
reports to you too though. :P

-Dan