xrogaan | so, err, my /run/user is empty | 13:58 |
---|---|---|
xrogaan | so is /var/run/user | 13:58 |
xrogaan | I don't think it's normal, is it? | 13:58 |
xrogaan | what would it take to get a XDG_RUNTIME_DIR? | 14:01 |
xrogaan | `When looking at how to implement XDG_RUNTIME_DIR for non-systemd inits, I couldn't find any actual specification for how to do this. That's because there isn't one, just some loosely-worded descriptions; it only exists in the systemd implementation. And the semantics of it are very poor indeed; it hasn't been developed with safety, security or flexibility in mind.` | 14:02 |
xrogaan | Nevermind then. | 14:02 |
xrogaan | So instead, where should I put my runtime things? | 14:03 |
xrogaan | using /tmp isn't really a great idea. | 14:03 |
amesser | xrogaan: sysvinit, no system and I have: | 14:04 |
amesser | set | grep XDG | 14:04 |
amesser | XDG_RUNTIME_DIR=/run/user/1100 | 14:04 |
amesser | so it works without systemd | 14:04 |
xrogaan | I don't though. | 14:04 |
xrogaan | I used to have something un that folder | 14:05 |
amesser | /lib/x86_64-linux-gnu/security$ grep -R XDG_RUNTIME_DIR . | 14:08 |
amesser | Übereinstimmungen in Binärdatei ./pam_elogind.so | 14:08 |
amesser | Übereinstimmungen in Binärdatei ./pam_gnome_keyring.so | 14:08 |
xrogaan | §yeah? | 14:08 |
amesser | so chance is high it is set by elogind session management daemon | 14:09 |
amesser | if you dont want to use elogind, maybe you can workaround this somehow usiing pam_env | 14:10 |
amesser | however, have to leave now, sorry. | 14:11 |
DocScrutinizer05 | purpy theme ;-D https://youtu.be/69FUuB7rh74?t=340 | 14:13 |
xrogaan | I use elogind | 14:14 |
xrogaan | huh, might be bugged | 14:16 |
xrogaan | loginctl doesn't show my session | 14:18 |
xrogaan | okay, so, if I start my session through lightdm there is no /run/user/1000 | 14:35 |
xrogaan | if I start a session though tty, I get one. | 14:35 |
xrogaan | doesn't lightdm make use of /etc/pam.d/common-session ? | 14:44 |
xrogaan | this fucking shit: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=717297 | 14:53 |
xrogaan | I don't know what I did, but it seems fixed. | 15:12 |
xrogaan | restarted elogind and lightdm | 15:12 |
xrogaan | I removed systemd from the greeter pam.d thing | 15:13 |
xrogaan | and now, I don't have transparency in my virtual consoles anymore | 15:19 |
xrogaan | I get these weird issues, I have no clue why they happens. | 15:32 |
xrogaan | Sometimes, dnsmasq/resolvconf don't get configured properly at network start and I have to reload /etc/networking | 15:33 |
xrogaan | err | 15:33 |
xrogaan | I meant /etc/init.d/networking* | 15:33 |
xrogaan | is it sysvinit that manage all that? | 15:34 |
ExtraCrispy | Hello | 16:26 |
ExtraCrispy | So Devuan on QubesOS is not a go to | 16:27 |
ExtraCrispy | I think sysvinit is the caus | 16:27 |
KatolaZ | ExtraCrispy: what is QubesOS? | 16:27 |
ExtraCrispy | casue | 16:27 |
ExtraCrispy | KatolaZ: A distro.Google it. | 16:27 |
KatolaZ | it's based on Fedora, apparently... | 16:28 |
ExtraCrispy | Supposedly is the most secure distro out there. | 16:28 |
KatolaZ | which might indeed be the cause, right? | 16:28 |
KatolaZ | which one? Fedora? | 16:28 |
ExtraCrispy | KatolaZ: Yeah by default it is. | 16:28 |
KatolaZ | :D | 16:28 |
ExtraCrispy | What do you mean? | 16:29 |
KatolaZ | how do you plan to migrate a Fedora do something that is based on Debian? | 16:29 |
KatolaZ | s/ do / to / | 16:30 |
ExtraCrispy | Oh no you misunderstood. | 16:31 |
KatolaZ | sorry then | 16:31 |
ExtraCrispy | I am migrating a Debian 9 VM to Devuan 2 | 16:31 |
nemo | ExtraCrispy: huh. why does qubesos care about systemd? | 16:31 |
KatolaZ | but you mentioned QubesOS | 16:31 |
ExtraCrispy | nemo: No idea. | 16:32 |
nemo | I mean for the guests that is | 16:32 |
nemo | ExtraCrispy: maybe #qubes or #qubes_os would know | 16:33 |
ExtraCrispy | nemo: Yeah i just asked in #qubes | 16:34 |
nemo | ExtraCrispy: are you setting up as Standalone VM? | 16:34 |
nemo | ExtraCrispy: 'cause seems there's no limits on that... BSD... who cares | 16:34 |
nemo | or are you actually trying to hack the distro to use devuan everywhere? | 16:34 |
nemo | in which case not so suprising it fails | 16:34 |
ExtraCrispy | No im just migrating from Debian | 16:35 |
nemo | rypervenche: slt | 16:35 |
rypervenche | nemo: :P | 16:37 |
fsmithred | ExtraCrispy, can you describe how it fails? | 16:38 |
ExtraCrispy | fsmithred: Just while installing | 16:40 |
ExtraCrispy | The VM just shuts down | 16:40 |
fsmithred | any error logs catch that? | 16:41 |
fsmithred | and at what point does it shut down? What exactly did you do? | 16:41 |
ExtraCrispy | JUst sudo apt-get install sysvinit-core | 16:42 |
ExtraCrispy | As for the ogs please wait cause i deleted the VM | 16:42 |
fsmithred | ok. I can't imagine that qubes polices what packages you install in a VM. | 16:44 |
ExtraCrispy | fsmithred: So... | 18:02 |
ExtraCrispy | It seems,not 110% sure though,that the problem exists in 3.2 verwion of QubesOS. | 18:03 |
ExtraCrispy | The problem is that sysvinit-core removes some needed packages. | 18:05 |
ExtraCrispy | And that makes the VM unbootable. | 18:05 |
ExtraCrispy | And i dont know how to keep those packages with apt | 18:06 |
KatolaZ | ExtraCrispy: you are not the first one to migrate a stretch to ascii, so this sounds strange | 18:07 |
fsmithred | what order did you do things? | 18:08 |
ExtraCrispy | KatolaZ: Well it is also a QubesOS related problem. | 18:08 |
fsmithred | the way I've done it was 1. install sysvinit in debian, 2. changes sources to devuan, update, dist-upgrade 3. reboot into devuan and remove sytemd | 18:08 |
KatolaZ | ExtraCrispy: I don't get why QubesOS should be concerned of what you run inside a VM... | 18:09 |
fsmithred | 1 also reboot | 18:09 |
ExtraCrispy | fsmithred: Since there is no GUI i just went straight to installing sysvinit-core? | 18:09 |
fsmithred | gui is irrelevant | 18:09 |
fsmithred | install sysvinit-core, reboot, then change sources | 18:09 |
ExtraCrispy | Then installing xfce4 is also irrelevant | 18:09 |
fsmithred | it shouldn't remove stuff | 18:09 |
fsmithred | I don't think (don't really remember) | 18:10 |
fsmithred | your debian is cli only? | 18:10 |
ExtraCrispy | The problem is sysvinit-core removes some important qubess specific packages | 18:10 |
ExtraCrispy | fsmithred: Ye | 18:10 |
fsmithred | there are cubes packages in debian? | 18:10 |
ExtraCrispy | Oh i just got an idea.Why not apt download sysvinit-core dpkg it and do the rtest? | 18:11 |
ExtraCrispy | fsmithred: Yes and all templates | 18:11 |
fsmithred | can you tell me a package name? | 18:12 |
fsmithred | I'm looking in debian repo | 18:12 |
fsmithred | or are they third-party packages? | 18:12 |
ExtraCrispy | Third party | 18:12 |
buZz | ExtraCrispy: then they arent in debian ;) | 18:13 |
ExtraCrispy | They are in QubesOS repos | 18:13 |
ExtraCrispy | buZz: Yeah? | 18:13 |
fsmithred | so you either need to edit the templates or maybe get devuan to identify as debian | 18:13 |
fsmithred | depends on what the templates test for | 18:13 |
fsmithred | and maybe play some trickery with package dependencies | 18:13 |
fsmithred | figure out why those packages get removed | 18:14 |
ExtraCrispy | Yeah maybe they hard-depend for some reason | 18:14 |
fsmithred | what are you using for virtualization? | 18:14 |
fsmithred | some qubes built-in stuff? | 18:15 |
ExtraCrispy | Xen hypervisor | 18:15 |
fsmithred | I guess qubes has some special requirements for VMs, and I have no idea what that might be. | 18:18 |
ExtraCrispy | The only issue i see is the removal of some specific packages | 18:19 |
ExtraCrispy | If i could bypass that shit then ok | 18:19 |
ExtraCrispy | But idk if tyhere is ant apt command to do that | 18:20 |
fsmithred | well, paste the error messages you get when you try to install sysv | 18:20 |
fsmithred | track down the deps | 18:21 |
ExtraCrispy | fsmithred: https://paste.debian.net/plain/1050126 | 18:22 |
ExtraCrispy | This shows up when i do install systemd-shim as well | 18:23 |
ExtraCrispy | Which brings some different packages,sysvinit-core being one or 'em. | 18:23 |
KatolaZ | ExtraCrispy: just get rid of qubes-core-agent and qubes-gui-agent | 18:26 |
KatolaZ | they might have hard deps on systemd | 18:26 |
fsmithred | and get rid of systemd-shim | 18:26 |
ExtraCrispy | fsmithred: LOL?W: Download is performed unsandboxed as root as file '/home/user/sysvinit-core_2.88dsf-59.9_amd64.deb' couldn't be accessed by user '_apt'. - pkgAcquire::Run (13: Permission denied) | 18:26 |
fsmithred | actually, I don't know - I've never used it | 18:26 |
ExtraCrispy | And the weird thing is...It is downloaded | 18:26 |
fsmithred | yup. download as unpriv user | 18:27 |
fsmithred | ??? | 18:27 |
ExtraCrispy | FFS Debian.. | 18:27 |
ExtraCrispy | KatolaZ: Withoiut those VM gets unbootable | 18:27 |
fsmithred | why consolekit? | 18:28 |
fsmithred | you got a desktop? | 18:28 |
ExtraCrispy | No | 18:28 |
ExtraCrispy | IDK why it exists | 18:28 |
ExtraCrispy | Maybe for the app shortcuts? | 18:29 |
KatolaZ | ExtraCrispy: uh> | 18:30 |
KatolaZ | ? | 18:30 |
ExtraCrispy | fsmithred: https://paste.debian.net/plain/1050139 | 18:30 |
KatolaZ | ExtraCrispy: are you using other repos on top of devuan's? | 18:30 |
fsmithred | KatolaZ, he hasn't gotten to devuan yet | 18:31 |
ExtraCrispy | ExtraCrispy: The repos should be added after installing sysvinit no? | 18:31 |
KatolaZ | ExtraCrispy: are you using other repos on top of debian's ones? | 18:32 |
KatolaZ | (i.e., where did you get the qubes-agent packet?) | 18:32 |
ExtraCrispy | Yes QubesOS has their own repos | 18:33 |
ExtraCrispy | Which are essential | 18:33 |
KatolaZ | oh then good luck | 18:34 |
KatolaZ | it looks like those packages have hard deps on systemd | 18:34 |
ExtraCrispy | Yeah | 18:34 |
ExtraCrispy | KatolaZ: Thanks | 18:42 |
KatolaZ | sorry ExtraCrispy | 18:43 |
KatolaZ | but why don't you risk an actual install on real hardware? (or on a qemu VM?) | 18:43 |
ExtraCrispy | I am not changing QubesOS. | 18:44 |
ExtraCrispy | Maybe i could make it work. | 18:44 |
fsmithred | qemu or virtualbox or usb (live or installed) | 18:45 |
fsmithred | or repackage the qubes stuff and hack the deps | 18:45 |
fsmithred | but the fact that you might be able to install it like that, it still might not work | 18:45 |
ExtraCrispy | Qemu and VBox is not an option. | 18:46 |
ExtraCrispy | Maybe on my laptop. | 18:46 |
ExtraCrispy | fsmithred: The problem is within those packages | 18:47 |
ExtraCrispy | BTW is wicd migration mandatory? | 18:47 |
fsmithred | wicd is default if you install task-whatever-desktop for some of the desktops | 18:55 |
fsmithred | but if you build stuff up manually, you can have pretty much whatever you want | 18:56 |
ExtraCrispy | I am building | 19:01 |
ExtraCrispy | Damn sysvinit | 19:01 |
ExtraCrispy | KatolaZ: fsmithred I got another idea. | 19:14 |
ExtraCrispy | I will pin them with hold as manually installed so they wont get removed | 19:15 |
KatolaZ | ExtraCrispy: if they need systemd, this is probably not gonna work | 19:24 |
ExtraCrispy | KatolaZ: Yah | 19:28 |
ExtraCrispy | KatolaZ: Well idk how to bypass that ard-deps thing | 19:31 |
ExtraCrispy | hard-deps | 19:31 |
* helios21 waves "Just like to say hi. I had contact one mailinglist with some of you. Just see my whois in case you like to know who is speaking here :)" | 21:20 | |
KatolaZ | hi helios21 | 21:23 |
KatolaZ | welcome | 21:23 |
helios21 | Hi KatolaZ. Thank you a lot for your work on sysvinit. | 21:23 |
golinux | helios21: Nice to see you here. Yeah, KatolaZ is a wizard . . . | 21:34 |
helios21 | golinux: :) | 21:37 |
eyalroz | Is there an apt repository for newer versions of GCC/g++ for installation on ASCII? Or should I somehow install from experimental? | 22:24 |
eyalroz | Something like Ubuntu's "ppa:ubuntu-toolchain-r" ? | 22:26 |
golinux | Try backports | 22:27 |
golinux | You can search here: pkginfo.devuan.org | 22:27 |
golinux | (I didn't) | 22:28 |
eyalroz | golinux: gcc-7 is available in beowulf and ceres, but I don't see any "backports" | 22:47 |
eyalroz | golinux: Oh, and my sources.list already has ascii-backports | 22:48 |
KatolaZ | eyalroz: do you need gcc-7? | 22:58 |
gnarface | it's not in backports and i'm skeptical that it can be successfully backported in the sense that it would generate binaries that don't crash randomly | 23:01 |
gnarface | but it will work fine in a ceres chroot | 23:01 |
gnarface | so you could just deboostrap ceres into a chroot | 23:02 |
gnarface | then build in that | 23:02 |
eyalroz | gnarface: I don't want it to replace gcc-6 as the system compiler, I just want it installed alongside it | 23:02 |
gnarface | eyalroz: that's not the part that's holding it back | 23:03 |
eyalroz | to be honest, I feel it's easier to just build it from sources into /opt/gcc-x.y.z than to debootstrap things | 23:03 |
eyalroz | (which I've not done before) | 23:03 |
gnarface | i've attempted both approaches and i've directed you to the one that was easier | 23:04 |
eyalroz | KatolaZ: I need something which supports most of c++17 | 23:04 |
eyalroz | gnarface: I didn't mean to sound ungrateful, I'm sorry. | 23:04 |
gnarface | eyalroz: it's ok, i know it may not be clear just how much more effort backporting will be, because it seems like you'd download a lot less. but you can't just get away with building one package for something like this. you'll have to recursively backport all it's dependencies too, by hand. *that's* where the real drama comes in - many of these will directly conflict with ascii in various ways | 23:05 |
KatolaZ | eyalroz: you could use clang also | 23:06 |
gnarface | i've got gcc 5 6 and 7 all installed concurrently on ceres, in multiple architectures, but i'm certain there's a good reason gcc7 isn't already in backports | 23:06 |
KatolaZ | 4.0 is in ascii | 23:07 |
KatolaZ | and supports the whole C++17 | 23:07 |
gnarface | i actually have several 4.x versions too, i just noticed | 23:07 |
KatolaZ | you also have clang 5 and clang 6 in backports | 23:07 |
KatolaZ | ascii-backports | 23:07 |
eyalroz | KatolaZ: Hmm, maybe | 23:08 |
eyalroz | gnarface: I know what libs I need, it's ok. | 23:08 |
gnarface | so, i've been having a problem with ascii on one machine for a while. it started a month or two ago after an update i think. i thought i had it locked down but it's still happening intermittently. the ethernet device basically just self disables if i don't attach the ssh session quick enough after boot (during boot?) | 23:50 |
gnarface | anyone seen this behavior lately? | 23:50 |
gnarface | the machine isn't going to sleep, and disabling the powermanagement flag on the ethernet module doesn't seem to do anything | 23:51 |
gnarface | if i log in to the machine locally the ethernet device looks up, has a valid route, just can't actually send traffic | 23:52 |
gnarface | but if i down it then up it again, it starts working | 23:52 |
gnarface | the ip address is statically configured | 23:52 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!