On 12/06/15 22:15, Anto wrote:
>
> On 12/06/15 18:34, Marlon Nunes wrote:
>> On 2015-06-12 10:03, Steve Litt wrote:
>>> On Fri, 12 Jun 2015 07:50:25 -0300
>>> Marlon Nunes <nunes@???> wrote:
>>>
>>>> Hi, i've been testing connman for a while and found it to handle very
>>>> well my network connections.
>>>>
>>>> https://01.org/connman
>>>
>>> The following sentence from the preceding link made me sweat a little
>>> bit:
>>>
>>> =====================================================
>>> ConnMan is optimized through open source for embedded and client
>>> focused Intel® Quark technology, Intel® Atom™ processors and Intel®
>>> Core™ processors.
>>> =====================================================
>>>
>>> I'm an AMD guy.
>>
>> I found it ok just for the fact that its completely independent of
>> systemd.
>>
>>>> https://wiki.archlinux.org/index.php/Connman
>>>
>>> Those Arch guys are the biggest bunch of systemd jingoists on the
>>> planet but you've got to admit, they write far and away the best
>>> documentation on the planet.
>>
>> Their wiki help pages are almost complete.
>>
>>>>
>>>> In my view, we can forget about network-manager completely for
>>>> desktop usage.
>>>
>>> Whether we stay with Wicd, which Devuan Alpha 2 does such a great job
>>> with, or switch to ConnMan, either way, you're right: network-manager
>>> is too entangled in dbus and systemd to be useful on Devuan, and it
>>> also requires you be in X, and that's not always true.
>>>
>>> I think that whether Wicd or ConnMan is our default network "make it
>>> easier machine", it should be easy to switch between the two, and part
>>> of that ease could be good documentation.
>>>
>>> By the way, I was going to answer Bardot Jérôme's query about Devuan
>>> Network-Manager similarly: Better to be rid of Network-Manager than to
>>> wonder if it's going to drag in systemd on an update. Network-Manager's
>>> wonderful for the one use case Debian envisions, but turns into a
>>> stumbling block when you go offroad.
>>
>> That's why a wrote about it. =)
>>
>
>
> I doubt that connman is free of systemd. I was just in the middle of
> preparing to compile connman 1.29, and I saw 2 files which are
> definitely meant for systemd. And they are in http://git.kernel.org.
>
> http://git.kernel.org/cgit/network/connman/connman.git/tree/src/connman.service.in
>
>
> [Unit]
> Description=Connection service
> Requires=dbus.socket
> After=dbus.socket
> Before=remote-fs-pre.target
> Wants=remote-fs-pre.target
>
> [Service]
> Type=dbus
> BusName=net.connman
> Restart=on-failure
> ExecStart=@sbindir@/connmand -n
> StandardOutput=null
>
> [Install]
> WantedBy=multi-user.target
>
>
> http://git.kernel.org/cgit/network/connman/connman.git/plain/vpn/connman-vpn.service.in
>
>
> [Unit]
> Description=ConnMan VPN service
> Requires=dbus.socket
> After=dbus.socket
>
> [Service]
> Type=dbus
> BusName=net.connman.vpn
> ExecStart=@sbindir@/connman-vpnd -n
> StandardOutput=null
>
> [Install]
> WantedBy=multi-user.target
>
>
> I know that they are harmless. But that tells me the intention to
> allow it to be locked-in into systemd. So I always want to remove
> everything related to systemd including the unit, service and socket
> files. I got the impression that a lot of people find my attempt to do
> that ridiculous. But I really don't care :)
>

I just purged all connman files that I downloaded tonight. I think t is
not worth trying to compile and install it. The title of the commit
below clearly says that connman is definitely being locked-in to systemd.

machine: Integrate ConnMan with systemd-hostnamed

http://git.kernel.org/cgit/network/connman/connman.git/commit/?id=d5acb39e80b40d2b21eed37506523e73fcd8956f