al1r4d | https://mastodon.social/@nixCraft/111658754855580408 xD | 02:36 |
---|---|---|
brocashelm | how to fix the broken desktop icon fonts in xfdesktop4 (beowulf)? this happened after installing beowulf-backports | 04:32 |
gnarface | did you enable backports and do a full upgrade from it or something? that's usually not advised | 04:36 |
gnarface | the first thing to try would be to put all the packages back except just the one you actually needed from backports | 04:37 |
gnarface | one or ones... there is sometimes groups of multiple packages in there that need to be all upgraded together, but in general the whole raft isn't tested against itself like the regular stuff | 04:38 |
brocashelm | i reckon it might be one of the mesa packages | 04:40 |
gnarface | like, for example, if you install nvidia drivers from backports you usually also want the kernel and any relevant non-free firmware versions from backports as well so they all match, and stuff often breaks if you don't... but on the other hand if you just install everything from backports you're almost certain to get a hodgepodge of stuff that both is and isn't meant to work together | 04:40 |
gnarface | if you got a mesa package from backports you'd definitely want to get all the mesa packages from backports, but i'm not sure that's a likely candidate for breaking fonts | 04:40 |
HimeHaieto | if this isn't already what you did, I might recommend learning some basic apt pinning, and pinning the backports repo below the main beowulf ones | 04:40 |
HimeHaieto | you should have a history in /var/log of exactly what packages got upgraded | 04:41 |
gnarface | it's supposed to be at a lower priority by default but i've seen it go sideways before both from user error and from apparent package maintainer error | 04:42 |
gnarface | yea, checking the apt log is a good idea | 04:43 |
gnarface | i think all the backports packages will have "bpo" in the version string as well | 04:45 |
gnarface | there was a font rendering change a while back, maybe around then too though... so it's possible what you're seeing is actually corrected behavior, in which case the fix you really want is to just run "dpkg-reconfigure fontconfig-config" and adjust the settings to your preference | 04:48 |
Nietz | gnarface: do you use apt-listchanges for paging through logs? | 04:49 |
HimeHaieto | could even get the full list from the log, and copy/paste it into something akin to `echo <packages> | awk '{gsub(/\ |$/, "/beowulf"); print}' | xargs apt-get install` (do *not* actually run this as is though without carefully checking everything yourself) | 04:50 |
gnarface | Nietz: nah i just usually use less or emacs | 04:51 |
HimeHaieto | something like that could revert the changes though | 04:51 |
gnarface | Nietz: that doesn't mean apt-listchanges is bad though | 04:51 |
HimeHaieto | oops, already saw at least one obvious error, "/beowulf" should have been "/beowulf " | 04:52 |
brocashelm | thanks, i'll see what the culprit was | 04:58 |
brocashelm | it's not a big deal; just a cosmetic bug showing two instances of text (e.g. "Home", "File System", "Trash" in white and gray texts) | 04:59 |
brocashelm | i created a test account on the same system, and the icons and fonts were displaying correctly | 05:06 |
brocashelm | which leads me to think this must be some css issue that newer versions from backports aren't handling right | 05:06 |
gnarface | brocashelm: my best guess is it's font the hinting settings (controlled by fontconfig) | 06:32 |
gnarface | s/font the/the font/ | 06:32 |
gnarface | but i'd have to actually see a before/after screenshot to be sure | 06:33 |
al1r4d | hello | 09:02 |
al1r4d | can someone help me please? | 09:03 |
al1r4d | gnarface, are you there? | 09:07 |
rrq | wassup? | 09:08 |
al1r4d | hello | 09:08 |
al1r4d | i need help, rrq | 09:08 |
al1r4d | would you help me? | 09:09 |
rrq | sure | 09:09 |
rrq | if I can | 09:09 |
al1r4d | i want to migrating from debian trixie to devuan excalibur | 09:09 |
al1r4d | https://paste.debian.net/1302497/ | 09:09 |
al1r4d | logs here | 09:09 |
al1r4d | how to fix? | 09:09 |
al1r4d | i'm on this phase: | 09:09 |
al1r4d | Once this is done eudev needs to be installed. Note that if Gnome is installed it will be removed by this command, but can be installed again after the migration. | 09:09 |
al1r4d | root@debian:~# apt-get install eudev sysvinit-core | 09:09 |
al1r4d | https://www.devuan.org/os/documentation/install-guides/daedalus/bookworm-to-daedalus | 09:10 |
rrq | first try would be to add "udev-" to the install line ... ending in - to indicate removal | 09:11 |
al1r4d | sorry, i dont understand, rrq | 09:12 |
rrq | apt-get install eudev udev- | 09:12 |
al1r4d | "dpkg: error processing archive /var/cache/apt/archives/eudev_3.2.14-1_amd64.deb (--unpack): | 09:12 |
al1r4d | trying to overwrite '/usr/share/pkgconfig/udev.pc', which is also in package systemd-dev 255.2-2" | 09:12 |
rrq | I saw the paste; did you get the same with udev- added? | 09:13 |
al1r4d | i did based you said -> https://paste.debian.net/1302498/ | 09:14 |
rrq | mmm maybe systemd-dev- | 09:14 |
al1r4d | sorry bad english | 09:14 |
al1r4d | okay, rrq, let me check | 09:14 |
al1r4d | rrq, https://paste.debian.net/1302499/ (systemd-dev-) | 09:15 |
al1r4d | looks like devuan repository doesnt have systemd-dev | 09:15 |
DelTomix | al1r4d: the next instruction says "The last command may cause package breaks but they will be resolved as part of the migration process." and so next step shows to use force option "apt-get -f install" | 09:15 |
al1r4d | oh, it has.. sorry | 09:16 |
al1r4d | DelTomix, i did it | 09:16 |
al1r4d | DelTomix, https://paste.debian.net/1302500/ | 09:16 |
DelTomix | the -f will force it to overwrite the file, (make sure you also installed sysvinit-core sucessfully ) then after reboot the instructions show to use the dist upgrade and then the purge steps | 09:17 |
DelTomix | hmm sorry I am confusing with fix | 09:18 |
al1r4d | should i do sudo apt-get remove systemd-dev? | 09:18 |
DelTomix | I don't think it will let you until you have rebooted | 09:18 |
al1r4d | let me check | 09:19 |
DelTomix | don't just yet - and you don't want to reboot unless sysvinit-core and eudev are installed. | 09:19 |
al1r4d | yes | 09:20 |
al1r4d | then, what should i do? | 09:20 |
DelTomix | I would use 'apt -f --force-yes' you are want it to boot with those files replaced, then when you get to the step to remove systemd you might have a similar issue | 09:21 |
DelTomix | but we can fix that I think | 09:21 |
al1r4d | sorry, i confused | 09:22 |
al1r4d | apt -f --force-yes install? | 09:22 |
DelTomix | wait | 09:22 |
DelTomix | sorry my habit is apt - it should be apt-get | 09:23 |
DelTomix | so do "apt-get -f --force-yes" | 09:23 |
rrq | and maybe: -o Dpkg::Options::="--force-overwrite" | 09:23 |
al1r4d | `sudo dpkg -i --force-overwrite /var/cache/apt/archives/eudev_3.2.14-1_amd64.deb` | 09:23 |
al1r4d | let me try | 09:23 |
DelTomix | ^ or that | 09:23 |
al1r4d | thank you ::)) | 09:24 |
al1r4d | now working | 09:24 |
DelTomix | make SURE sysvinit-core was installed properly though | 09:24 |
gnarface | it was just a dev package, couldn't it have been simply removed ahead of time? | 09:25 |
DelTomix | 'dpkg -l sysvinit-core' should show 'ii' at the start of the line before the name | 09:25 |
al1r4d | oh no | 09:25 |
al1r4d | https://paste.debian.net/1302501/ | 09:26 |
al1r4d | what should i do? | 09:26 |
DelTomix | oh! gnarface is right - I missed the -dev | 09:26 |
gnarface | hmm, well now it seems to have a different issue though | 09:27 |
DelTomix | I think the apt-get -f part covers that and then the after the reboot with sysvinit-core it allows the systemd to be purged (after the dist upgrade step) | 09:28 |
al1r4d | https://paste.debian.net/1302502/ | 09:28 |
al1r4d | DelTomix | 09:28 |
gnarface | al1r4d: what happens if you run: apt-get install sysv-rc sysvinit sysvinit-core sysvinit-utils | 09:28 |
al1r4d | gnarface, https://paste.debian.net/1302503/ | 09:29 |
gnarface | al1r4d: what happens if you run: apt-get install elogind | 09:30 |
al1r4d | gnarface, https://paste.debian.net/1302504/ | 09:31 |
DelTomix | what is the output of "ls -al /sbin/init" < And I don't know what the path is for usrmerge | 09:32 |
gnarface | al1r4d: what's in your sources.list right now? | 09:32 |
al1r4d | alif@debian:~$ ls -al /sbin/init | 09:33 |
al1r4d | -rwxr-xr-x 1 root root 52400 Dec 7 01:03 /sbin/init | 09:33 |
al1r4d | gnarface, deb http://deb.devuan.org/merged excalibur main | 09:33 |
gnarface | hmmm | 09:33 |
gnarface | libpolkit-agent-1-0 (= 123-3devuan1) but 123-3 is to be installed | 09:33 |
gnarface | this seems to be the problem | 09:33 |
al1r4d | yeah | 09:34 |
al1r4d | what should i do? | 09:34 |
gnarface | not sure, but i'd be tempted to just remove all the polkit stuff and reinstall it correctly after the upgrade | 09:34 |
gnarface | what might work though is explicitly specifying every package in the dependency loop including version | 09:35 |
rrq | "apt-get install elogind libelogind0" would sort that bit I think | 09:35 |
rrq | polkitd remains an issue | 09:35 |
gnarface | the problem is you're doing this against excalibur and i don't know for sure it's not excalibur's fault (it's testing after all) | 09:35 |
al1r4d | but "alif@debian:~$ sudo apt-get install eudev sysvinit-core | 09:36 |
al1r4d | Reading package lists... Done | 09:36 |
al1r4d | Building dependency tree... Done | 09:36 |
al1r4d | Reading state information... Done | 09:36 |
al1r4d | eudev is already the newest version (3.2.14-1). | 09:36 |
al1r4d | sysvinit-core is already the newest version (3.08-5devuan1)." | 09:36 |
gnarface | another potential option is to just use aptitude and let it suggest a solution | 09:36 |
al1r4d | should i reboot? | 09:36 |
gnarface | not sure it's safe yet, but if you have a rescue disk or live image it should still be recoverable either way... | 09:36 |
gnarface | brb | 09:37 |
DelTomix | ok I checked - the byte size of his sbin/init file matches current excalibur sysvinit-core sbin/init | 09:40 |
DelTomix | (and date) | 09:45 |
rrq | ( md5sum = 8c7c8ae287379d4639dd1375e8f7b3af ) | 09:48 |
DelTomix | al1r4d: output of 'md5sum /sbin/init' ? | 09:49 |
DelTomix | hmm mine does not match rrq's md5sum = 8e0c01073150541f21d3a9a880ddaa5c for sysvinit-core_3.08-5devuan1_amd64 | 09:51 |
rrq | I get that latter in ceres | 09:53 |
rrq | maybe my excalilbur is old... hmmm | 09:53 |
DelTomix | hmm ok pkginfo.devuan.org shows same version for ceres and excalibur | 09:54 |
rrq | yes it has sysvinit-core=3.08-3devuan1 | 09:54 |
rrq | old | 09:54 |
rrq | my excalibur I mean | 09:54 |
rrq | I've scheduled an upgrade into some future | 09:55 |
DelTomix | understandable | 09:55 |
DelTomix | he must have rebooted | 09:55 |
DelTomix | although still in irc | 09:56 |
gnarface | al1r4d: trying aptitude? | 09:57 |
al1r4d | :/ | 09:58 |
al1r4d | just revert to debian bookworm | 09:58 |
al1r4d | and i reboot | 09:59 |
al1r4d | :) | 09:59 |
gnarface | and it booted? | 09:59 |
al1r4d | Looks like working | 09:59 |
gnarface | so now you're upgrading from bookworm to daedalus? | 09:59 |
al1r4d | Yup | 09:59 |
gnarface | that might be easier | 09:59 |
al1r4d | Based on fuise | 09:59 |
al1r4d | *guide | 09:59 |
al1r4d | Yep! Working | 09:59 |
gnarface | the problem with testing and unstable is the packages themselves are often broken in non-obvious ways | 10:00 |
al1r4d | service: command not found :/ | 10:00 |
gnarface | it's in a package | 10:00 |
al1r4d | Okay | 10:00 |
al1r4d | :) | 10:00 |
gnarface | i forget which one | 10:00 |
gnarface | you don't actually need it thoguh | 10:00 |
gnarface | you can just run the scripts in /etc/init.d/ directly | 10:00 |
gnarface | al1r4d: ah, yes, it's in the "init-system-helpers" package, but note that it's also in /usr/sbin/ which isn't in root's default path anymore | 10:02 |
gnarface | you can add it back to the ENV_SUPATH line in /etc/login.defs if you want | 10:02 |
* cousin_luigi could not rename network interfaces by adding a rules file inside /etc/udev/rules.d: is it necessary to rebuild initrd also? | 12:28 | |
fsmithred | cousin_luigi, are you struggling with old interface names (eth0) vs new names (enp-whatever) or are you trying to make custom names? | 12:45 |
cousin_luigi | fsmithred: Well, right now I wish to create matching rules and then I will decide. Plenty of vlans and bridges make clutter. | 12:47 |
fsmithred | ok, if you just wanted to use the new "predictable" names with eudev, you can boot with net.ifnames=1. But I guess you have a lot of other names to deal with. | 12:48 |
fsmithred | I can't help there. | 12:49 |
cousin_luigi | fsmithred: I successfully created a few usb ones, but I struggle with pci devices. | 12:51 |
cousin_luigi | Will try recreating mkinitrd to see if that's the problem. | 12:51 |
al1r4d | i'm happy with devuan :) | 12:53 |
fsmithred | al1r4d, did you solve your migration problem? | 12:54 |
al1r4d | yes | 12:54 |
fsmithred | cool. What did you do? | 12:55 |
al1r4d | i have 2 devuan installed now: laptop and computer | 12:55 |
al1r4d | fsmithred, revert to bookworm and migrated | 12:55 |
fsmithred | ouch. | 12:55 |
fsmithred | In case you try it again, there's an old post on forum where someone had similar trouble migrating to chimaera (I think) and he was able to install sysvinit-core from debian repo and then upgrade in devuan repo. | 12:58 |
al1r4d | :) thank you | 12:59 |
n4dir | searching the dpkg #debian boot how to replace systemd with sysvinit, i get the info that sysvinit wasn't able to boot the linux kernel since 2003. And i never realized! | 13:01 |
n4dir | i can't find the info though, it once was there, the debian wiki probably has it still | 13:01 |
n4dir | if you want to switch to devuan after that, it probably isn't too exciting. | 13:01 |
cousin_luigi | So, if anyone uses custom udev rules for integrated devices, please shout back | 13:02 |
* fsmithred hates udev rules | 13:03 | |
gnarface | cousin_luigi: what do you mean by "integrated devices" exactly? like the onboard ethernet card? i've literally never had to mess with the /etc/udev/rules.d/70-persistent-net.rules file that is auto-generated to store them, except once to change the NAME field on two of them to switch their order. its never given me problems aside from that, and that was expected behavior. is it possible you're adding your own file in | 13:16 |
gnarface | there and it's fighting with the auto-generated one? | 13:16 |
gnarface | cousin_luigi: i know it keys off MAC address, so if you're using MAC address randomization that will hose it good... | 13:17 |
al1r4d | btw | 13:17 |
al1r4d | how to enable dnssec and dnsovertls on devuan? | 13:18 |
al1r4d | hmm -> https://dev1galaxy.org/viewtopic.php?pid=19950 | 13:19 |
gnarface | al1r4d: dunno about general solutions, but last i checked dnssec was on by default in bind9 and dnsovertls was on by default in firefox (firefox calls it "dns over https" though) | 13:22 |
gnarface | i doubt you have to do anything special in current stable but it should be easy to check | 13:23 |
al1r4d | let me check later, thank you, gnarface | 13:27 |
cousin_luigi | gnarface: I was being too selective | 13:41 |
cousin_luigi | The MAC address is more than enough to zero on a device. | 13:42 |
peterrooney | well this is weird. aqemu (0.9.2-3, daedalus) provides deprecated "-soundhw" options to qemu-system-x86_64 (7.2.7) which uses the -audiodev semantics. hence no way to use aqemu without disabling sound. is the problem in aqemu, or devuan packaging? | 20:31 |
fsmithred | peterrooney, we don't package qemu or aqemu. Those come from debian unchanged. | 21:29 |
n4dir | peterrooney: i asked something like that, sound in aqemu, a few weeks ago. I for one couldn't get it solved, but i was not too engaged. | 21:30 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!