libera/#devuan-dev/ Friday, 2018-10-26

LeePenHi. I am considering adopting Debian systemd-shim. Not because I am interested in it, but because the lack of version 10.4 in testing seems to be a block on maintaining alternative inits in Debian.19:11
LeePenSee #903295.19:12
LeePenI see there is a sysvinit upload to Debian experimental today. Does somebody else have an idea to take on systemd-shim too? I don't want to duplicate.19:17
kilobyteLeePen: -shim doesn't need to be adopted, it needs to be shot, burned, buried, receive a tombstone written in Comic Sans, with an organic grave marker you can produce with your lower back20:14
kilobyteLeePen: we plan to replace it with elogind20:14
kilobyteshim never worked (at least for me) in the first place -- and its purpose is to run systemd when it's not pid 120:15
LeePenThanks. I get that from Devuan's POV, but there is no elogind in Debian (yet?)20:43
LeePenWould elogind in Debian enable a fix for #903295 et al by replacing -shim?20:46
Centurion_DanLeePen: you don't want systemd, then you should run devuan20:48
kilobyteCenturion_Dan: bad thinking -- you guys don't have anything close to the manpower to fix regressions20:49
LeePenI am and enjoying it.20:49
Centurion_Dankilobyte: whatever...20:49
LeePenMy thinking is that if we keep alternative inits viable in Debian, Devuan will have less work to do with limited person-power.20:50
kilobyteit's Debian who by default put an alternative, extremely bogus, poor excuse for an init.  Sysvinit is the standard one; other alternatives being openrc and so on.20:53
kilobyteyou have a point that getting GUIs installable again is pretty urgent, though20:53
LeePenYes, I am not defending Debian's decision which is why I have switched to Devuan. I am just trying to be pragmatic in terms of reducing Debian's excuse for further reduction in init diversity which will cause Devuan more problems.20:56
kilobyteright20:57
kilobyteas opposed to sysvinit where somewhere around 10-20 people(!) started contributing after the call on debian-devel@lists.d.o, one person sent me some packages, and that's all20:58
kilobyteI planned to depart for the family home for the weekend, but I can do that tomorrow morning; looks like elogind won't do itself...21:00
kilobyteunless you have a _trivial_ patch for the meantime21:01
Centurion_Dankilobyte: for elogind just fork the source from git.devuan.org...21:02
Centurion_Danit should build unchanged in debian21:02
kilobyteCenturion_Dan: that's what I'm using on this machine -- but that's only good for a person like us21:02
kilobytethe vast majority of work is making it upgradeable from systemd21:03
LeePenkilobyte: Nothing trivial, I have just spent the afternoon trying to get my head around the various threads over the last 10 days and see where it is worth spending time and effort.21:03
kilobyteyeah :(21:04
kilobytelooks like the packages need to be at least briefly coinstallable with systemd21:05
kilobyteunless someone beats into systemd maintainers that systemd should be removable while it's running21:05
Centurion_Danyou may find people have to switchout systemd first and then install elogind...21:09
kilobytecan't do that without removing most of the packages (on GUI systems)21:10
LeePenI notice the is a RFP for elogind: #90538821:10
Centurion_Dankilobyte: We can migrate debian -> devuan with one reboot.21:10
Centurion_DanI really don't know if it's possible to do that in debian..21:12
* kilobyte looks at Lorenzo Puliti's packaging. It looks... weird.21:20
LeePenI have just been looking at elogind. I agree replacing systemd with logind is difficult.22:11
LeePenHowever, #903295 is a box without systemd. So if elogind and libpam-elogind (provides libpam-systemd) were available in Debian, I think that would be a fix.22:13

Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!