xrogaan | Is there a netinstall ISO with proprietary network drivers loaded? | 07:23 |
---|---|---|
gnarface | there should be i think | 07:32 |
gnarface | did you try the devuan_chimaera_4.0.0_amd64_netinstall.iso? | 07:32 |
rwp | Also IIRC the refracta installer has all of the drivers too. | 08:13 |
xrogaan | gnarface: I believe I got that one yeah, the boot menu is a bit confusing with all the entries | 09:35 |
gnarface | find the non-graphical expert mode | 09:37 |
gnarface | if they're not loaded automatically they'll be in the "additional installer components" menu | 09:38 |
chomwitt | how can i now if a package is in stable backports ? | 11:20 |
fsmithred_ | chomwitt, search the package name at pkginfo.devuan.org | 12:15 |
fsmithred_ | or... | 12:15 |
fsmithred_ | if you have a line for chimaera-backports in sources.list, do 'apt policy <package>' | 12:16 |
chomwitt | fsmithred_, thanks | 16:49 |
chomwitt | i found an example of a backported package and use it as guide : https://pkginfo.devuan.org/cgi-bin/policy-query.html?c=package&q=0ad&x=submit | 16:50 |
chomwitt | unfortunately there is no backport for qutebrowser | 16:51 |
chomwitt | anyway i dont complain. 0ad is an excellent game . if i have time i would like to package oolite game and backport qutebrowser. | 16:58 |
chomwitt | is there a entry guide to start experimenting? | 17:03 |
chomwitt | xmm.. looking at devuan packapots Package.gz file i discover duf !! indeed a better df. cool | 17:19 |
chomwitt | s/packpots/backports | 17:19 |
chomwitt | wow even a nes emulator is backported! some guys in deb land love retrogaming!!! | 17:25 |
chomwitt | sorry i should go to offtopic | 17:25 |
fsmithred_ | chomwitt, debian wiki has backporting instructions | 19:24 |
chomwitt | ok | 19:30 |
gnarface | i just debootstrapped daedalus and something might be wrong with the locales | 20:09 |
gnarface | i mean something that might not be me forgetting a step | 20:10 |
gnarface | can anyone else confirm or did i actually do something wrong? | 20:11 |
gnarface | when i run "locale charmap" i get errors like locale: Cannot set LC_CTYPE to default locale: No such file or directory | 20:12 |
gnarface | but when i run "dpkg-reconfigure locales" there's only utf8 locales available in the list | 20:12 |
fsmithred_ | gnarface, make sure locales-all is installed | 20:30 |
fsmithred_ | hm... maybe not. I don't have it. I do have util-linux-locales | 20:30 |
gnarface | hmm, nope, added it but utf8 is still the only available locale, and i'm still getting the errors at nearly every package install as well as after running "dpkg-reconfigure locales" itself | 20:34 |
gnarface | i also note that after running "dpkg-reconfigure locales" the /etc/default/locale file gets updated to en_US.UTF-8 but the output of the "locale" command itself still says plain old en_US | 20:37 |
gnarface | though i guess maybe part of the issue could be because i'm chrooted | 20:37 |
gnarface | should i boot it to see if it behaves differently? | 20:37 |
gnarface | https://paste.debian.net/1259992/ | 20:38 |
gnarface | fsmithred_: what about libencode-locale-perl? | 20:45 |
gnarface | fsmithred_: it wouldn't be krb5-locales would it? | 20:57 |
gnarface | it doesn't seem to be libencode-locale-perl | 20:57 |
eyalroz | Hello devuaners, | 21:02 |
eyalroz | Recently, there was a large update of daedalus (including libc). After the update, my self-built celluloid media player has started acting up: If I run /opt/celluloid/bin/celluloid /path/to/file.mp4 , it works fine (more or less), but trying to start it by double-clicking a file in the nemo file manager - fails. | 21:02 |
eyalroz | I was wondering whether you might have any leads I could follow to try and figure out why I can't launch the player with my file. (I've already tried rebuilding of course.) | 21:03 |
rwp | gnarface, What environment variables are you setting? Try "env | grep -e LANG -e LC_" and see what is set. Seems to me those are overriding. | 21:07 |
rwp | eyalroz, Are there any errors in the ~/.xsession-errors file for this action? That file is where X logs all of the output from graphical applications. | 21:08 |
rwp | eyalroz, For a deep dive debugging one could run nemo under strace logging all system calls, trigger the failing action, and look at the file trace. | 21:09 |
rwp | Something like "strace -o nemo.strace.out -v nemo &" or similar. | 21:09 |
rwp | That will produce a lot of data to grep through. Can reduce that somewhat by restricting to just files "strace -e trace=file -o nemo.strace.out -v nemo &" | 21:10 |
gnarface | rwp: i'm not setting any | 21:11 |
gnarface | oh, LANG=en_US on the host though, hmm | 21:13 |
gnarface | so that is probably what's making that happen, but i think something else is still broken or i'd see something other than just the utf8 options in dpkg-reconfigure | 21:13 |
eyalroz | rwp: I'll check the .xsession-errors file | 21:18 |
eyalroz | ... nope, nothing there, it seems | 21:27 |
rwp | gnarface, A normal thing is to run "locale" and see what it says. There is a subtle way to tell which variables are set in the output. | 21:29 |
rwp | Here is an example. https://paste.debian.net/plain/1259999 I set three variables. In the locale output those are NOT quoted with double quotes. | 21:30 |
rwp | gnarface, When I dpkg-reconfigure locales on my machine and page down to the en_* listing I see this: https://www.proulx.com/tmp/dpkg-reconfigure_locales.png | 21:34 |
rwp | eyalroz, That path for /opt/celluloid/bin/celluloid is in a custom location. How is nemo being configured for it? I might suspect a problem there. Because if it calls "celluloid" it likely is not finding it on PATH. | 21:36 |
rwp | Sorry that I do not use nemo myself so I can only guess at how it is configured. | 21:36 |
rwp | gnarface, Another thing... The bash shell loads ~/.bashrc and that can hide problems because it will load variables when the terminal and shell are started. | 21:37 |
rwp | But if one launches an X application for example that never has bash in the process series and will not load any of those variables. | 21:37 |
rwp | That has snagged people before. The ~/.bashrc equiv for X is ~/.xsessionrc but careful that .xsessionrc is /bin/sh syntax not /bin/bash syntax. | 21:38 |
eyalroz | rwp: I right-click the file, choose "Open with", | 21:39 |
rwp | Changes there need one to log out and then log back in again so that /etc/X11/Xsession can source the ~/.xsessionrc file. | 21:39 |
eyalroz | and using the dialog, define a new application I guess | 21:39 |
rwp | That last was meant for gnarface of course. Parallel interleaved conversations are always fun! :-) | 21:39 |
eyalroz | Yeah, np. | 21:40 |
rwp | eyalroz, I would try starting from a terminal with "strace -e trace=file -o nemo.strace.out -v nemo &" and then trigger the problem then "grep file.mp4 nemo.strace.out" hoping for some interesting clue there. | 21:41 |
gnarface | rwp: do you think this will stop it from booting? the thing is, i considered overriding LC_ALL and such manually but i saw specific advice not to do that | 21:44 |
gnarface | i'm wondering if maybe i can fix it if i boot it on bare metal and the problem is just that i was chrooted from a less up-to-date copy of testing | 21:45 |
rwp | gnarface, The only thing interacting with booting is that init scripts load /etc/default/locale to set the "system locale" that you set with dpkg-reconfigure. Personally I recommend always setting dpkg-reconfigure of the system local to None so that system logged messages are in the C/POSIX locale by default. | 21:46 |
rwp | However if that is set to a non-existent locale then one thing I notice is that perl complains loudly at every perl invocation making things noisy. | 21:47 |
rwp | I have not seen it otherwise affect booting. But I think at most to have a local take effect for your environment logging out and logging back in should be sufficient. | 21:48 |
rwp | I have been known to be wrong though. :-) | 21:48 |
rwp | If you chroot'd in then your environment would have copied in. Exit the chroot. Enter the chroot again using this "chroot /path/to/chroot su -" and it will run su with the - to have it spawn a login shell. Then you will get root's login environment in the chroot. Or use "su - gnarface" to get the gnarface login environment in the chroot. | 21:50 |
rwp | And the equiv with sudo would be chroot /path/to/chroot sudo -i -u gnarface" (or sudo -i for plain root login shell). | 21:51 |
gnarface | ah, i should have done that | 21:52 |
gnarface | i used chroot /mnt /bin/bash | 21:52 |
gnarface | didn't think of su | 21:52 |
gnarface | that might have done the trick | 21:52 |
rwp | You can see what is passed through with "chroot /mnt /usr/bin/env" and have it dump the environment. | 21:53 |
gnarface | the one thing that seems weird to me is that the chroot only had utf8 but the host is set to 8859-1 | 21:53 |
rwp | And you can handle the environment explicitly with "chroot /mnt env -i /bin/bash -l" to initialize a clean environment and then bash will act as a login shell. | 21:54 |
gnarface | i would have expected the chroot to be merely limited to the host setting in this case | 21:54 |
rwp | I would expect the chroot to be completely independent, other than the environment variables passed through. | 21:54 |
rwp | Since the chroot has a complete copy of libc and all of the libc locale files. | 21:54 |
gnarface | well this is all based on the hypothesis that the host's environment variables were causing some unknown conflict | 21:55 |
gnarface | but it didn't quite behave as i'd expect even in that case | 21:55 |
rwp | True. And who knows what evil lurks in the heart of the host system. Only the shadow knows... | 21:55 |
gnarface | anyway i've already disconnected it from the other machine, so i guess we'll see what happens after it's booted on bare metal | 21:55 |
onefang | /etc/shadow? | 21:55 |
rwp | The Shadow is an old time radio show and every episode started with the above jingle. | 21:56 |
onefang | I know. | 21:56 |
rwp | Okay then! It needed a smiley so we would know who was in on the joke. Sorry. | 21:56 |
onefang | lol | 21:56 |
gnarface | rwp: do you think i am gonna have to reinstall all the packages that depended on this, or should they inherit the change automatically? | 21:58 |
rwp | No re-install is needed. Everything happens dynamically when programs are launched. | 21:58 |
rwp | Programs that *were* already launched and running will not change though. Most programs look at these at program start time. | 21:58 |
rwp | Aside: I am going to note that unexpected environment inheriting is one of the reasons for using "service" to interact with the /etc/init.d/* scripts. Since "service" cleans the environment before invoking the init scripts. Avoids environment inheriting issues. | 22:02 |
eyalroz | rwp: So, this is interesting... when I start nemo like you suggested, from a shell session - double-clicking _does_ work, and opens the file fine. | 22:05 |
eyalroz | Maybe it's because of some environment variables? Like GSETTINGS_SCHEMA_DIR or XDG_DATA_DIRS ? | 22:06 |
norayr | hey devuan people | 22:31 |
norayr | i was trying to turn my debian to devuan | 22:35 |
norayr | eventually some apt process hunged and i lost the connection to the board, and i had no keyboard. | 22:35 |
norayr | now i have keyboard, booted from live usb, mounted the debian system in chroot. | 22:36 |
norayr | it had no kernel, i am installing the kernel now. | 22:36 |
norayr | from debian repos. | 22:36 |
norayr | back then i was not able to install sysvinit-core from devuan repos, so i tried to install it from debian repos. | 22:36 |
norayr | now i have it installed, and i am installing the kernel. | 22:36 |
norayr | i guess next step is to switch to devuan repos again. | 22:36 |
norayr | and install something. | 22:36 |
norayr | gnarface: it's me! | 22:38 |
rwp | eyalroz, Or perhaps PATH is being inherited from the command shell. | 22:38 |
rwp | norayr, These instructions work well for me. https://www.devuan.org/os/documentation/install-guides/chimaera/bullseye-to-chimaera | 22:39 |
gnarface | norayr: you probably should have been using the devaun repos already for any missing packages in the chroot | 22:41 |
gnarface | norayr: for the kernel package itself though it may not matter on x86 and x86_64 | 22:41 |
gnarface | (since that's not a rebuilt package) | 22:41 |
norayr | i installed linux-image-5.10.0-19-amd64 | 22:42 |
norayr | because apt-cache search brought it in the results and it felt last. | 22:42 |
norayr | is it last? now i switched to devuan repos, maybe i can install kernel again? | 22:42 |
rwp | 5.10.0-19-amd64 is the latest kernel for Stable Chimaera. | 22:43 |
norayr | yay | 22:44 |
norayr | i'll install mate and reboot then. | 22:44 |
norayr | hmm i have this error, i guess because of efi partition on fat: ln: failed to create hard link '/boot/initrd.img-5.10.0-19-amd64.dpkg-bak' => '/boot/initrd.img-5.10.0-19-amd64': Operation not permitted | 22:45 |
gnarface | oh, i forget how that's supposed to be fixed | 22:46 |
norayr | task-mate something brings too many packages | 22:51 |
norayr | is there a package just for pure mate, without let's say gimp? | 22:51 |
norayr | trying mate-desktop-environment-core | 22:51 |
norayr | so cool this project exists! | 22:52 |
norayr | now, what will i have for wifi, instead of network manager? | 22:52 |
norayr | mate network applet will talk to other service? | 22:52 |
norayr | is it wicd?? | 22:52 |
gnarface | i think network manager has been forked, you should be able to use it | 22:52 |
gnarface | wicd might be there too | 22:52 |
gnarface | in theory you could even just use wpa_supplicant bare | 22:53 |
norayr | that i can do. | 22:53 |
norayr | just this device will be accessed via mouse and display only. | 22:53 |
onefang | wicd for chimaera had issues due to PYthon 2 -> 3. | 22:53 |
norayr | should i install network-manager? i see there is no network-manager after installing mate-desktop-environment-core | 22:54 |
onefang | I switched from wicd to bare wpa_supplicant. Also coz I got tired of wicd always rewriting it's own config file and upsetting etckeeper. lol | 22:54 |
norayr | there are only traces of old network-manager, with rc... | 22:54 |
gnarface | norayr: this is with devuan sources? | 22:55 |
norayr | yes | 22:55 |
gnarface | norayr: yes, you should try to install network-manager again | 22:55 |
norayr | already | 22:55 |
norayr | and also, i see network-manager-gnome, but i am searching for something like network-manager-mate | 22:55 |
gnarface | just try to reinstall everything it had removed on you before you got disconnected, i should all fit it just needed to be reinstalled | 22:56 |
rwp | The heir apparent for wicd is connman. So either that or network-manager. (I hate network-manager though. But connman is odd. No great favorite.) | 22:56 |
norayr | yes, i don't remember what was that. | 22:56 |
norayr | rebooting | 22:57 |
norayr | lets see | 22:57 |
norayr | hmmm, i have no bootloader. i have some uefi menu here, and it is empty. | 23:00 |
norayr | will try to boot to livecd and generate grub.cfg, but shouldn't it be generated automatically when i installed kernel again? | 23:01 |
gnarface | doh, forgot grub when you installed the kernel | 23:01 |
norayr | i checked, the grub package was there | 23:01 |
norayr | and some other grub packages. | 23:01 |
gnarface | hmm, did you run grub-install from within the chroot? | 23:01 |
norayr | no, that i didn't do | 23:01 |
gnarface | you might have to | 23:02 |
norayr | maybe i need grub-mkconfig? | 23:02 |
gnarface | i don't think so | 23:02 |
gnarface | however some efi thing could be related | 23:02 |
gnarface | try to run grub-install on to the first drive (not a partition of the drive) from within the chroot then reboot | 23:03 |
gnarface | i guess that's two reboots since you'll have to get back into the live image first | 23:03 |
eyalroz | rwp: Well, the path shouldn't really matter, since I specified the executable with the full path; but maybe I should try it nonetheless. Do you know where I can "inject" path elements for applications started via the desktop environment launcher? In my case, cinnamon? | 23:03 |
gnarface | norayr: make sure you bind mount /proc /sys /dev /dev/pts in the chroot (last one may not be needed for this) | 23:04 |
norayr | thank you, yes yes i did that last time as well | 23:09 |
norayr | hmmm, so apparently last time | 23:09 |
norayr | i mounted first partition to /mnt/boot | 23:09 |
norayr | but it contains some EFI directory | 23:09 |
norayr | and symlinks didn't get generated because of that | 23:09 |
norayr | but the installed image on the second drive | 23:09 |
norayr | had /boot in / | 23:09 |
norayr | and it contained grub directory, etc | 23:10 |
norayr | so apparently that was a mistake to install kernel on the first partition | 23:10 |
norayr | btw this first partition has also two directories in EFI, debian and ubuntu | 23:10 |
norayr | this device came with ubuntu preinstalled, so that ubuntu directory is because of that. and i can see it from bios. | 23:10 |
gnarface | yea something went wrong in there with regards to distro names, i'm sorry i don't ever remember the fix... fsmithred_? | 23:10 |
norayr | i wonder how to generate new, devuan directory. or i hope EFI/debian will boot devuan somehow. | 23:10 |
gnarface | someone around here knows | 23:11 |
gnarface | i don't, i don't trust efi and i've been avoiding it | 23:11 |
gnarface | maybe you just need to copy or symlink EFI/debian to EFI/devuan? | 23:11 |
norayr | yeah, i would like to avoid it, but i have no idea how, this bios doesn't seem to allow that. | 23:11 |
norayr | now doing apt-get --reinstall install linux-image-5.10.0-19-amd64 | 23:12 |
gnarface | if i recall, the thing that went wrong is about just the literal string debian/devuan | 23:12 |
norayr | without that partition mounted | 23:12 |
norayr | hope the symlinks will be ok this time. | 23:12 |
gnarface | there might be a optional grub package you need | 23:12 |
gnarface | grub-efi or something like that | 23:12 |
onefang | Some BIOSes are fussy about what directories you use in the EFI partition. | 23:13 |
norayr | so debian directory contains: search.fs_uuid 42978f48-b7df-4ead-8b5e-c8bf1d4c75df root | 23:13 |
norayr | set prefix=($root)'/boot/grub' | 23:13 |
norayr | configfile $prefix/grub.cfg | 23:13 |
norayr | actually that was a grub.cfg config in debian directory. three lines only. | 23:13 |
rwp | eyalroz, The ~/.bashrc equiv for X applications is ~/.xsessionrc but careful that .xsessionrc is /bin/sh syntax not /bin/bash syntax. | 23:14 |
rwp | eyalroz, Changes there need one to log out and then log back in again so that /etc/X11/Xsession can source the ~/.xsessionrc file. | 23:14 |
norayr | i have grub-efi* packages installed | 23:14 |
eyalroz | rwp: Yes, I realize I'll need to restart the X session. I wonder if root can inject changes into a running process' environment... | 23:15 |
gnarface | norayr: i think it might work then | 23:21 |
norayr | i guess i indeed need to do grub-install. | 23:24 |
norayr | but doing it to first partition i guess is wrong. | 23:24 |
norayr | i don't know how to do it with efi. debian installer did the setup instead of me. | 23:24 |
gnarface | you don't just install it to the disk? | 23:25 |
norayr | i was getting empty grub> prompt. but then i entered bios, chosen debian, and ... i was getting devuan prompt!!! | 23:25 |
gnarface | like /dev/sda instead of /dev/sda1 doesn't work? | 23:25 |
rwp | eyalroz, Nope. (Though I should never say never. Linux acquires new behaviors unknown to me all of the time.) | 23:26 |
norayr | my guess was to install it to /dev/mmcblkp2, where is the / | 23:26 |
norayr | and /boot | 23:26 |
norayr | so... but it booted and i had no password, booting to livecd again, to set pw, i don't remember it. i was connecting via network with key. | 23:26 |
norayr | and lightdm had autologin configured. | 23:26 |
norayr | sorry for being so verbose. | 23:26 |
norayr | ok it works! | 23:33 |
norayr | it booted! | 23:33 |
norayr | oh even xscreensaver started working!!! | 23:33 |
norayr | i just don't understand why the directory 'debian' works. i did cp -a debian devuan there, but bios doesn't see it. | 23:34 |
gnarface | hmm, i don't know either but someone around here surely does | 23:34 |
gnarface | glad it's working anyway | 23:34 |
norayr | whatever. it boots and it works. i got only the message in mate that some multipanel applet didn't start, and suggestion to remove it. probably mate-something-core doesn't bring it. | 23:34 |
norayr | and i have no idea what kind of applet is it. probably i don't need it. | 23:35 |
norayr | so glad it works now! | 23:35 |
norayr | and thank you for suggesting to not reinstall, but repair. | 23:35 |
norayr | (also ventoy boot manager on my usb didn't boot devuan installer, says it's not efi compliant or something) | 23:35 |
gnarface | no problem... this is probably mentioned in the forums somewhere | 23:37 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!