[Mageia-dev] Proposal: Deprecate draknetcenter+network init scripts after systemd becomes default.

[Colin Guthrie]

'Twas brillig, and Thierry Vignaud at 22/08/11 13:32 did gyre and gimble:
> On 22 August 2011 11:57, Colin Guthrie <mageia at colin.guthr.ie> wrote:
>> This is just a heads up to suggest that we officially adopt a policy of
>> moving to NetworkManager after systemd becomes our default (and likely
>> only) init system - i.e. likely for mga3.
>>
>> As you may know, GNOME is moving towards a systemd user session and
>> systemd itself is very much aligning itself to be the one true init
>> system on linux.
>>
>> I also have it on good authority that many of the features lacking in
>> NetworkManager (such as bridging configuration) will be available in the
>> not too distant future and many other more advanced networking features
>> such as fast-start DHCP, per-interface DNS, 4-8's DNS fallback and
>> several other nice features will ultimately be possible too.
>>
>> So with all this in mind, I think the writing is very much on the wall
>> with regards to the future integration of NetworkManager into the
>> desktop. I believe we will be able to concentrate our resources better
>> by going along with the rest of the world in this regard. I have
>> personally switched to it now (as I do not need advanced features of
>> draknetcenter et al) as a step towards this.
> 
> err... drakx-net's code is heavily used by other tools such as:
> - drakx-installer
> - mgaonline
> - ...
> 
> So I think we still want to keep drakx-net

Perhaps, but then a more interesting question is "what bits of drakx-net
are used in mgaonline and drax-installer?" Just stating that they are
used there as a reason to keep it isn't really a good argument IMO.

Asbestos is used in fire retardant wall insulation, but it's not a good
argument to keep it considering the world has moved on and we now know
the dangers of working with that material.


So what does mgaonline need to work with drakx-net? Does it just need to
know if you are online or not? If so why not use NM's dbus service to do
that? It could eventually know whether you are on a paid/metered
connection (this is not yet supported by NM but there has been some talk
about it) or a "free" connection - a fast vs. slow connection etc. It
could make intelligent, informed decisions, not just "is the user online
- no matter how crappily?"

Then of course the wider question... should mgaonline be deprecated in
favour of more integrated options suck as package-kit? Should we focus
resources on that? This is IMO a more open question, even if I do feel
the writing is ultimately on the wall here too.

I'd very much like to use much more standard things. I think some of the
mdv tools were differentiators in the past but those days are pretty
much slipping by. Some of the tools are still best of bread, but things
like network management and handling updates etc. is perhaps something
that has no progressed to the stage that us having independent update
mechanisms is a hindrance to users rather than a benefit.

It's all too easy to get stuck in a rut and get emotionally attached to
such things (I do it all the time in my day job) but sometimes it helps
to look at things from a slightly detached view point.

Cheers

Col


-- 

Colin Guthrie
mageia(at)colin.guthr.ie
http://colin.guthr.ie/

Day Job:
  Tribalogic Limited [http://www.tribalogic.net/]
Open Source:
  Mageia Contributor [http://www.mageia.org/]
  PulseAudio Hacker [http://www.pulseaudio.org/]
  Trac Hacker [http://trac.edgewall.org/]