Re: [gentoo-user] recent trouble with wicd and systemd (non-global ctrl_ifname)

[Canek Peláez Valdés]

On Fri, Dec 13, 2013 at 1:59 AM, Alan McKinnon <alan.mckinnon@×××××.com> wrote:
> On 13/12/2013 00:47, Canek Peláez Valdés wrote:
>> On Wed, Dec 11, 2013 at 4:52 PM,  <gottlieb@×××.edu> wrote:
>>> At home I use a wired connection so did notice the following problem
>>> until I traveled and tried to connect wirelessly.
>>> The problem must have started sometime within the past month.
>>>
>>> If I have wicd started by systemd, i.e.
>>> systemctl enable wicd
>>> The wired network is started fine but not the wireless.  Instead, I see
>>> in the systemd journal
>>>
>>> wicd[290]: Failed to connect to non-global ctrl_ifname: wired  error: No
>>> such file or directory
>>> wicd[290]: Failed to connect to non-global ctrl_ifname: wireless  error: No
>>> such file or directory
>>>
>>> If I instead systemctl disable wicd, reboot, and then manually type
>>> wpa_supplicant -i wireless -c /etc/wpa_supplicant/wpa_supplicant.conf -B
>>> it works.
>>>
>>> Indeed after I have booted I can start wicd and cannot get the error
>>> above, but the actual behavior is not consistent.
>>>
>>> My system is ~amd64, profile gnome/systemd
>>>
>>> My wireless driver is from the package broadcom-sta (wl)
>>
>> I have never used wicd, so I can't say exactly what it's the problem;
>> but I was under the impression that wicd is basically dead. Its last
>> release was more than a year and a half ago.
>>
>> Regards.
>>
>
> release more than a year and a half ago != dead

In this particular case I think it is.

> the code the user has still works whether the devs adds upstream commits
> or not.

Well, apparently not [1].

> It hasn't bit-rooted, is not incompatible with everything else and
> doesn't have outstanding security bugs with little chance of being fixed.

Checking [1] and [2], I would think that wicd satisfies (or *at least*
starts to satisfy) the very definition of bitrot.

> So what's the problem?

If the code worked perfectly, none. But apparently it doesn't; I don't
know, I don't use it myself. The usual signs of bitrot are there,
though.

> By that logic, zenity needs to have died 5 years ago but it's still around

That's a really bad example. Zenity didn't had a 3.10 release, but it
had a 3.8 [3] in march, so it's 9 months since the last release, not
18. Also, now zenity has a 3_10 tag in git [4]. And lastly, its lats
commit was 6 days ago, and it had several bugfixes committed not three
weeks ago [5]. On the other hand, wicd only has had translations
committed in the last 6 *months* [6], and the "development" branch for
2.0 hasn't been touched in *3 years* [7].

This is only after a quick search through wicd and zenity repositories
(and Gentoo bugzilla). Perhaps wicd has reached perfection and it
doesn't need an upstream since everything simply works and there is
nothing else to do with it. That would be a first in software history,
though.

I would simply not use it, and I will recommend any of its users to
change to either NetworkManager [8] or connman [9], like pronto.

[1] https://bugs.launchpad.net/wicd/+bugs
[2] https://bugs.gentoo.org/show_bug.cgi?id=486440
[3] ftp://ftp.gnome.org/pub/gnome/sources/zenity/3.8/
[4] https://git.gnome.org/browse/zenity/tag/?id=ZENITY_3_10_0
[5] https://git.gnome.org/browse/zenity/log/
[6] https://code.launchpad.net/~wicd-devel/wicd/experimental
[7] https://code.launchpad.net/~wicd-devel/wicd/aqua
[8] http://projects.gnome.org/NetworkManager/
[9] https://connman.net/

Regards.
-- 
Canek Peláez Valdés
Posgrado en Ciencia e Ingeniería de la Computación
Universidad Nacional Autónoma de México