| rwp | amarsh04, My Unstable system now has /run/systemd/stuff https://paste.debian.net/plain/1310879 and this is confusing things like /etc/cron.daily/spamassassin which uses that as a test for systemd running. Have you seen this on your Unstable system too? | 04:31 |
|---|---|---|
| rwp | I am not sure what is creating those directories. It's new behavior. | 04:32 |
| rrq | rwp: I think those come from elogind | 04:57 |
| rrq | at least after purging elogind (bringing in consolekit and stuff) and rebooting my /run got cleansed of systemd subdir | 05:11 |
| amarsh04 | rwp, my /run/systemd has files/directories dates starting at 14 March 2024 at 19:41 - will check aptitude log for what was running then | 08:19 |
| rwp | rrq, (amarsh04) I think you might be right about elogind creating those as etckeeper's git log shows it was recently installed in a combination of things on the 13th. | 18:20 |
| rwp | That's new behavior for elogind I am pretty sure because I have had it installed other places and did not have the problem of "if [ -d /run/systemd ]; then systemctl ..." stuff before. | 18:21 |
| rwp | That seems like an insufficient test though so I might file a bug report asking for it to be changed to "if [ -d /run/systemd/system ]; then" which is what I have been using elsewhere. | 18:22 |
| rwp | For spamassassin I am reading the script and I see they provide an avoidance-condition of this "if [ -d /run/systemd ] && [ ! -e /etc/spamassassin/skip-timer-conversion ]; then" so if one creates /etc/spamassassin/skip-timer-conversion then it avoids this error I am seeing. | 18:25 |
| CueXXIII | rwp: how about checking for existence of the command systemctl? there is a standalone "systemctl" package, too, but by description it should run with any initsystem (except systemd itself) | 19:18 |
| rwp | CueXXIII, That would be less good than what they are attempting in that daily script. One might have installed the docker's package systemctl but still not want to use systemd timers. | 19:23 |
| rwp | I think they just missed the test by a little bit when they used -d /run/systemd instead of -d /run/systemd/system and that is the root of the problem. I am going to file a bug report and suggest doing that instead. | 19:24 |
| CueXXIII | ah, true, for timers that wouldn't help | 19:25 |
| rwp | Here is the entirety of the /etc/cron.daily/spamassassin script. https://paste.debian.net/1310943/ | 19:26 |
| rwp | Here is the current error output of the script: https://paste.debian.net/1310944/ | 19:27 |
| rwp | On line 24 they have a correct test for for systemd and I am simply going to propose they change the second test on 34 to be exactly the same. | 19:28 |
| CueXXIII | ah, with all these hacks a global config which init system is used would be useful… | 19:29 |
| CueXXIII | but try to get all vendors to agree on a global config…^^ | 19:30 |
| rwp | It is strange but https://salsa.debian.org/debian/spamassassin/-/commit/211e60e8bfec56f31e4fe54100438a6a899953fa added that to line 24 but not to line 34. | 19:31 |
| rwp | Line 34 was last changed 3 years ago. Which is strange I have only been seeing these errors the last two days on this Unstable system. (It's not a problem in the current Stable.) | 19:31 |
| rwp | Do I have skip-timer-conversion in my Stable system? I do. That's how I wasn't seeing this there. | 19:32 |
| CueXXIII | yeah, clearly an oversight | 19:34 |
| rwp | Speaking of docker's systemctl package... I have not looked at it in detail. I need to. But the description made me think it was rather docker container specific. Which scared me off from thinking it was a general purpose replacement. Yet they took over the namespace anyway. Which felt annoying. | 19:37 |
| rwp | The problem I always dread about submitting these types of bug reports is that there are several systemd folks who monitor every submitted bug report and simply because I am doing to need to mention systemd somewhere in the report this will cause them to jump into the discussion to argue against making things work without systemd and they will tell me I am a bad person for not complying with their wishes. | 19:51 |
| rwp | I spend a lot of time thinking about how to word things to be as neutral as possible. I know ahead of time that it will not be enough. | 19:51 |
| Xenguy | .oO( 'Dread' must be the word of the week ) | 20:34 |
| Xenguy | .oO( Better 'dread' than 'dead' I suppose ) | 20:34 |
| Xenguy | .oO( 'Don't spread the dread' ) | 20:36 |
| Xenguy | I'm out of ideas | 20:36 |
| CueXXIII | .oO(what the dread are you thinking?) | 20:36 |
| Xenguy | .oO( Not bad at all ) | 20:37 |
| Xenguy | But these are thought crimes now | 20:37 |
| rwp | After attempting reportbug I found I had a DNS misconfiguration to fix first. Then sent in the bug report. | 20:51 |
| rwp | spamassassin: Insufficient test condition in /etc/cron.daily/spamassassin https://bugs.debian.org/1067007 | 20:51 |
| rwp | I was almost able to file the entire report without mentioning systemd anywhere but of course the actual test condition itself mentions it so in the end I had to as well. | 20:52 |
| rwp | I tried my best to word things in such a way as to not trigger the systemd folks to jump in and argue against fixing this. We will see. Noah is a good maintainer and I have high hopes that he will simply see the problem and fix it kindly. | 20:53 |
| CueXXIII | rwp: i'd just tell them you are suggesting a bugfix and they should keep their religios opinion to themselves | 20:54 |
| rwp | It's just exhausting to always be a second class person. | 20:56 |
| Xenguy | rwp, I always felt that way as a Linux user in the first place | 23:15 |
| Xenguy | But worth the work I think; at least I was willing to do it | 23:16 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!