| Alverstone | Nope, using evdev hasn't solved anything | 09:44 |
|---|---|---|
| Alverstone | something lagged and now I can only use keyboard in one session | 09:45 |
| Alverstone | switch vts and it stops working | 09:45 |
| Alverstone | or rather | 09:45 |
| Alverstone | works on tty1 and tty>2 | 09:45 |
| Alverstone | and not on tty2 | 09:45 |
| Alverstone | unplug plug back on tty2 | 09:45 |
| Alverstone | will stop working on tty1 | 09:46 |
| Alverstone | this is so bizzare | 09:46 |
| Alverstone | bizarre* | 09:46 |
| Alverstone | restarted elogind to "magically" fix | 09:47 |
| Alverstone | rootless xorg has always been full of stupid bugs | 09:47 |
| Alverstone | in my debian I would have my X session frozen with no way to cope but to kill | 09:48 |
| Alverstone | at least it doesn't do it here, *seemingly* | 09:48 |
| DingoSaar | This might be a newbie question, but I haven't found a definite answer for it. I want to read (and possibly write) from a NetBSD FFS partition. However, I haven't found yet an utility package. Can anybody point me in the right direction? ddg.gg hasn't yielded any results I can work with. | 10:49 |
| rrq | maybe this is a help? https://wiki.netbsd.org/tutorials/how_to_mount_ffs_partition_under_linux | 10:53 |
| rrq | afaict there are some caveats for writing | 10:55 |
| DingoSaar | Thanks, but that only meant that ufs isn | 10:56 |
| DingoSaar | t installed on my newly installed devuan system. And I haven | 10:56 |
| DingoSaar | t yet found a way to enable it. | 10:56 |
| DingoSaar | On this VM, I cannot compile my own kernel, alas | 10:56 |
| rrq | right; and apparently CONFIG_UFS_FS_WRITE is not set for debian kernel 6.1.0-25-amd64 so there's no direct writing... you might need to set up a freebsd chroot | 10:58 |
| rrq | or netbsd chroot | 10:59 |
| rrq | though that might not be sufficient if it actaully requires kernel support | 11:00 |
| rrq | there seems to be some "fuse ffs" option but I don't know anything about it ... https://github.com/avinashshenoy97/FFS | 11:06 |
| rrq | (may even be something else) | 11:07 |
| rrq | btw there is an "ufs" module, so "modpobe ufs" would load that, but apparently it doesn't include write support | 11:10 |
| drizzt | Alverstone: "restarted elogind to "magically" fix" <-- there's nothing magical in this : elogind may have taken control of the keyboard and mouse on the tty, which leaves nothing for any other programm | 11:13 |
| DingoSaar | Thanks, so it seems it is there by default | 11:20 |
| DingoSaar | Thanks, from all options I looked into, "it'a a module in the standard kernel" didn't come to mind. It seems this solves my problem. Thanks again! | 11:23 |
| Alverstone | drizzt, I don't have the barest idea. The only helpful log I found is that i failed to close an input device, presumably the keyboard. Happens with the mouse too. Probably it's because I switch too quickly between ttys. Need to hold a pause of 4-5 seconds. It's empirical data though | 11:31 |
| freem | rootless xorg would be a good idea, *if* another process didn't had to be root instead... | 17:48 |
| drizzt | the real question is : why does the system need a root process to run an X server ? | 17:49 |
| * freem still uses xorg as root, because still have not found any good reason to switch to a way which just requires another daemon to do the job, and unlike elogind & consort, Xorg received decades of bugfixes | 17:49 | |
| drizzt | freem: +1, same here | 17:49 |
| freem | from what I remember (not much), it's because xorg embeds some drivers, so have to work along with the kernel on some stuff | 17:49 |
| drizzt | freem: any user can work with the kernel | 17:50 |
| freem | also I believe openBSD have rootless Xorg without requiring a 3rd party daemon, but I'm not sure at all | 17:50 |
| drizzt | from the kernel point of view, root IS a user | 17:50 |
| freem | but if this is true, then the problem comes down to a bad software architecture on linux's xorg server | 17:51 |
| drizzt | (and there's no other point of view than the kernel point of view on a Linux kernel based system | 17:51 |
| freem | problem is that historically, you had to be UID0 to access some files, since posix does not have anything akin to capabilities or jails AFAIK | 17:52 |
| drizzt | root is a user which has ID 0, and as such can do anything a user can do without having to become another particular user | 17:53 |
| drizzt | but the linux system also provides groups, and users can be asigned different groups, and files can get specific rights for a given group | 17:54 |
| drizzt | this solves almost all situations. | 17:54 |
| freem | indeed | 17:54 |
| drizzt | and when this is not enough, then there's ACL to the rescue :) | 17:54 |
| freem | are ACLs portable, though? | 17:55 |
| drizzt | (but I do not know if this would be useable on /dev content) | 17:55 |
| freem | xorg is not just a linux software, it's used on other OSes | 17:55 |
| drizzt | freem: define "portable" | 17:55 |
| drizzt | accross architectures, yes | 17:55 |
| freem | was more thinking about accross OSes | 17:56 |
| drizzt | accross OSes ... I don't know | 17:56 |
| drizzt | They are called "POSIX ACL" | 17:56 |
| * freem does not knows anything about ACLs | 17:56 | |
| freem | oh | 17:56 |
| drizzt | Should be rinning on any posix system | 17:56 |
| freem | well then I guess that should be reasonably portable yes | 17:57 |
| drizzt | running | 17:57 |
| drizzt | mind, we may not be able to use them anymore in a near future, as "posix" seems to be something which devs and distribution maintainers seem to forget, or do not care about .... | 17:58 |
| freem | dunno. I personnally care about using stuff from standards as much as possible | 17:59 |
| drizzt | freem: that's a good point for you :) | 18:00 |
| freem | but I guess most linux distro maintainers indeed don't care that much | 18:00 |
| drizzt | systemd devs do not seem to care as much as we do | 18:01 |
| freem | otherwise systemd would not have taken that much | 18:01 |
| freem | systemd people never gave a damn about portability | 18:01 |
| freem | they do document the fact they *require* glibc6, because they use some of it's extensions, and rejected patches to get muslc portability | 18:02 |
| Alverstone | is it possible to put eudev under runit supervision? does it make sense? | 19:42 |
| fsmithred | checking... | 19:45 |
| Alverstone | i ask because i often embark on some autistic journeys of 'fixing' something that isn't supposed to be fixed | 19:46 |
| Alverstone | i waste a lot of time | 19:46 |
| Alverstone | use my crippled solution for a year or so | 19:47 |
| Alverstone | roll back | 19:47 |
| Alverstone | and then try to forget it as a nightmare | 19:47 |
| Alverstone | and then roll back* | 19:47 |
| fsmithred | I don't have one and I don't see one in the runit-service package | 19:48 |
| fsmithred | I also would assume that it's possible. | 19:49 |
| Alverstone | it's because eudev is started very early | 19:49 |
| Alverstone | it's S02 in rcS.d | 19:49 |
| fsmithred | and if you're interested, I'm pretty sure gnuinos has a build with runit and vdev instead of udev/eudev | 19:49 |
| Alverstone | pros of vdev? | 19:49 |
| fsmithred | I don't know, but in the last build, he got it to boot fast. It used to be very slow. I think it's faster than eudev now. | 19:50 |
| fsmithred | disadvantage is that you might have hardware that vdev does not know what to do with. | 19:51 |
| Alverstone | https://github.com/jcnelson/vdev | 19:51 |
| Alverstone | says to get my hands dirty D: | 19:51 |
| fsmithred | I managed to write a few files for some hardware when vdev was new. It was not too difficult. I'm sure I could not write a udev rules file. | 19:51 |
| fsmithred | it's come a long way since jude messed with it. | 19:52 |
| fsmithred | aitor has been working on it. (gnuinos dev) | 19:52 |
| fsmithred | gnuinos.org/mirror/daedalus/runit | 19:56 |
| Alverstone | looks quite involved | 19:58 |
| Alverstone | these barebones things require skills and knowledge i don't have, and i tend to avoid software i don't understand and i am not sure is entirely safe/reliable | 19:58 |
| Alverstone | runit is simple to understand, so i switched to it because i don't like what systemd/gnomeOS is doing to linux | 19:59 |
| Alverstone | but udev... | 20:00 |
| Alverstone | i don't understand any of what it does or how | 20:00 |
| fsmithred | ok, don't go there. | 20:01 |
| Alverstone | :) | 20:02 |
| fsmithred | I checked the forum. Back in 2021 Lorenzo said he wasn't sure he could do udevd. | 20:07 |
| Alverstone | fair enough | 20:08 |
| Alverstone | void folks start udev in stage 1 same way devuan does and then just restart it upon entering stage 2 | 20:10 |
| Alverstone | idk how good that is | 20:10 |
| Alverstone | or how bad | 20:10 |
| Alverstone | exec 2>&1 | 20:11 |
| Alverstone | udevadm --exit | 20:11 |
| Alverstone | udevadm control --exit * | 20:11 |
| fsmithred | why? | 20:11 |
| Alverstone | exec udev | 20:11 |
| Alverstone | three lines | 20:11 |
| Alverstone | that's void's run file for udev | 20:11 |
| fsmithred | oh, so runit takes over | 20:11 |
| Alverstone | the result is that udev is controlled by runit | 20:11 |
| Alverstone | but whether it makes sense to do so, I don't know | 20:12 |
| fsmithred | I don't either | 20:12 |
| Alverstone | can udev fail? if it does, is it good to restart it? or maybe there's a good reason why it exited? | 20:12 |
| Alverstone | too many questions | 20:12 |
| fsmithred | no, you're not really supposed to restart udev, but I've done it plenty of times | 20:13 |
| fsmithred | sorry I don't remember why | 20:13 |
| fsmithred | not for normal computer usage. I must have been messing with stuff. | 20:13 |
| freem | <Alverstone> is it possible to put eudev under runit supervision? does it make sense? | 20:16 |
| freem | Yes, it is. | 20:16 |
| golinux | Alverstone: You seem to have a lot of "fringe" problems . . . | 20:17 |
| freem | Alverstone: https://p.mort.coffee/61A and https://p.mort.coffee/yVK are the files I use | 20:17 |
| freem | and yes, it makes sense. Udev can crash, for example | 20:20 |
| Alverstone | freem, but what about stage 1? | 20:22 |
| freem | no need | 20:22 |
| Alverstone | how so | 20:22 |
| freem | why would you need it at stage 1? | 20:23 |
| freem | udev is, IIRC, embedded in initrd, and does some work there already | 20:23 |
| Alverstone | say, S06cryptdisks-early | 20:23 |
| Alverstone | S07cryptdisks | 20:23 |
| Alverstone | don't they need /dev | 20:24 |
| freem | well, stage 1 will run cryptdisks except if you named a folder cryptdisk in your /etc/sv folder | 20:24 |
| freem | err... /etc/service, sorry | 20:25 |
| Alverstone | i didn't | 20:25 |
| Alverstone | i don't see a point of overloading runit with one-off scripts just to sv down them forever | 20:25 |
| freem | same | 20:25 |
| freem | but udev is a daemon sitting around waiting for some stuff to do, it's not a one-script thing | 20:26 |
| freem | one-off* | 20:26 |
| freem | besides, if your disks are encrypted, you need the kernel modules load by initrd, when you land in your normal system, it means there's already enough things loaded to read /etc and /usr at minimum | 20:27 |
| freem | especially with usrmerge around | 20:27 |
| freem | I don't remember when I have found that run file for udev though. When I tried myself, udev would not work because this crappy daemon requires commands to be sent it after it have done it's init in order to start actually working | 20:29 |
| freem | most of that script is not from me, and it is clearly perfectible, as if for some stupid reason someone decided udev binary is a library (hello systemd crowd...) and you naively just tried to summon it, the script would spam processes endlessly | 20:31 |
| freem | instead of noticing something's wrong | 20:31 |
| freem | I'll take a look at this vdev thing though. udev is a component I want to replace since *long* | 20:31 |
| Alverstone | i don't kernel modules to be loaded by initrd actually because it's just encrypted swap as of now | 20:33 |
| Alverstone | i don't need8 | 20:33 |
| Alverstone | i don't need* | 20:33 |
| Alverstone | at least that's what i think... | 20:34 |
| Alverstone | i don't even really know who does the loading, scripts in /etc/rcS.d are not very explicit, S09kmod doesn't too much seemingly | 20:35 |
| Alverstone | if all most modules are loaded in initramfs, then you're correct | 20:36 |
| Alverstone | all or most | 20:36 |
| freem | I'm no expert on that matter, mind you | 20:41 |
| freem | the /etc/rcS.d/ files made me run away screaming when I tried to read them, long ago | 20:42 |
| freem | really, it's a good thing this stuff will be less and less used: to understand _one_ of those you need to find sourced scripts from a multitude of random places | 20:43 |
| freem | as if shell was not already a nightmare by itself | 20:43 |
| freem | what I do know though, is that I use that udev run script on all my computers, including that VPS with an encrypted root partition :) | 20:44 |
| Alverstone | how to update alternatives? | 20:53 |
| Alverstone | i want mousepad to provide editor | 20:53 |
| Alverstone | not vim | 20:53 |
| Alverstone | ig i want too much sorry | 20:54 |
| n4dir | usually with "update-alternatives --config editor" ; but using mousepad doesn't seem possible, probably because it is a gui editor | 20:55 |
| debdog | update-alternatives --config item | 20:55 |
| debdog | double tab gives you a list of all available items | 20:57 |
| debdog | or look at /etc/alternatives/ to see the items and the targets they're linked to right now | 20:58 |
| debdog | https://wiki.debian.org/DebianAlternatives | 21:00 |
| debdog | oops, sorry, totally missed n4dir's comment | 21:00 |
| n4dir | no need to be sorry, there is quite some extra info you gave | 21:02 |
| debdog | https://wiki.debian.org/DebianAlternatives#Adding_a_new_alternative_to_an_existing_group mayhap following this mouspad can be added? | 21:03 |
| Alverstone | debdog, n4dir has a point about GUI | 21:07 |
| Alverstone | i won't die typing 'mousepad', which i almost never do anyway, and some things may rely on /bin/editor and i don't want it to explode in a text session | 21:07 |
| Alverstone | didn't think about it | 21:08 |
| n4dir | all i can say is mouspad which is installed isn't listed here. More i don't know | 21:09 |
| debdog | Alverstone: there is x-terminal-emulator for that | 21:11 |
| rwp | Really everyone should be setting EDITOR in their ~/.bashrc, ~/.xsessionrc, or equiv file anyway. Pick one: EDITOR=emacs EDITOR=vim EDITOR=nano | 21:11 |
| fsmithred | you can set the default graphical editor inside your desktop | 21:11 |
| fsmithred | I don't see mine in the alternatives | 21:12 |
| Alverstone | i don't use desktop :D | 21:12 |
| fsmithred | window manager? | 21:13 |
| Alverstone | i3wm with i3bar | 21:13 |
| Alverstone | well i use xfce4-appfinder actually | 21:13 |
| Alverstone | but my file manager is worker | 21:13 |
| fsmithred | oh, I don't know how to make a menu item in that. Does it use .desktop files? | 21:13 |
| Alverstone | in worker, you configure everything yourself | 21:13 |
| Alverstone | including how certain file types are handled | 21:13 |
| Alverstone | so I just made editor.sh to exec mousepad and that's it | 21:14 |
| Alverstone | if I want to change it, I just edit editor.sh | 21:14 |
| fsmithred | yeah, that's probably the easiest | 21:14 |
| Alverstone | fsmithred, appfinder uses .desktop files. very handy | 21:14 |
| Alverstone | haven't found anything more straightforward to use | 21:14 |
| fsmithred | key combination? | 21:15 |
| fsmithred | can't you set those in i3? | 21:15 |
| Alverstone | don't wanna clutter i3 config with those | 21:15 |
| Alverstone | i use a few for file manager and the like | 21:15 |
| Alverstone | but that's all | 21:16 |
| rossignol | I just installed the raspberry pi flavor of devuan. It's got a cute little login that shows basic info about the host, kind of list fastfetch or neofetch. But it's does not appear to be any of those packages. | 23:11 |
| rossignol | Anyone here know what's giving the output on login? I can't find anything in .bashrc or .profile. | 23:11 |
| user71 | I'm confused, you want either /etc/issue or /etc/motd depending if the message is displayed pre or post login | 23:14 |
| golinux | You might also want to ask in #devuan-arm | 23:25 |
| rossignol | @golinux, thanks - I'll check in there | 23:26 |
| drizzt | debdog: double tab gives a list of items only if you have bash-completion installed and loaded (or equivalent for other shells) | 23:46 |
| rwp | rossignol, I think it is pam_motd.so a PAM module in /etc/pam.d/login and /etc/pam.d/sshd and the message can be silenced by touching a file ~/.hushlogin | 23:51 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!