| rrq | generally speaking it's a farily costly way of access butu it does provide the end-point shading of tor access. | 00:00 |
|---|---|---|
| rrq | it's "expensive" due to being tcp-in-tcp tunneling that involves a number of otherwis unrelated intermediates, and it "helps" bogging down pkgmaster.devuan.org ... certainly not one of my favourite access paths | 00:03 |
| darwin | how do I stop /etc/resolv.conf getting overwritten? I thought his had been a thing on OS with systemd, but now seems to be happening on most/all? | 07:28 |
| gnarface | darwin: most likely network-manager doing it. you can flag it as immutable, or you could just fix your DHCP configuration so that it gives network-manager the right data. | 07:39 |
| gnarface | (or you could just uninstall network-manager and do the whole network configuration by hand) | 07:39 |
| gnarface | (same advice goes for any other graphical network administration utility) | 07:40 |
| darwin | i don't want to remove that right now, just configure it to stop doing that | 07:40 |
| darwin | but occasionally I edit it | 07:41 |
| gnarface | to be clear, the right place to fix this information is in your DHCP server, which is most likely also your router and cable modem if you are wondering | 07:41 |
| gnarface | usually but not always, your ISP won't force you into any particular settings for it | 07:41 |
| gnarface | you can just make it immutable with i think chattr (check the man page) but that won't work on every filesystem | 07:43 |
| gnarface | i think it should for ext4 though so if you're using defaults it should be fine | 07:43 |
| gnarface | but that's a quick and dirty fix, not really "the right way" | 07:44 |
| Xenguy | darwin, I made /etc/resolv.conf immutable years ago, and it still works fine for me | 07:49 |
| darwin | good | 07:49 |
| Xenguy | However I don't use NM | 07:50 |
| Xenguy | So YMMV | 07:50 |
| darwin | yes, I have DNS servers in my router, but I want extra. I put 20+ in resolv.conf only to read the standard is no more than two are used | 07:51 |
| Xenguy | 2 nameservers yes | 07:52 |
| darwin | then my ISP's two were discontinued so we were out of DNS for a while... and sometimes public ones were inaccessible. The standard is horrible | 07:52 |
| rwp | darwin, If you want something better then you have to run a local caching nameserver. | 07:53 |
| Xenguy | heh, just checked and I have 3 nameservers configured | 07:53 |
| darwin | what about recompiling? | 07:53 |
| rwp | What are you asking about recompiling? | 07:53 |
| * Xenguy uses unbound ... | 07:53 | |
| darwin | the definition that only two are used rather than all in resolv.conf or larger default | 07:53 |
| rwp | Only the first three are used. But there are long timeouts if the first one is down and it has to fall through to the 2nd and 3rd. So, not good. | 07:54 |
| Xenguy | rwp, So 3 nameservers is the limit, or 2? | 07:55 |
| rwp | Three. | 07:55 |
| Xenguy | Aha | 07:55 |
| rwp | The libc resolver is a relatively simple bit of code. The strategy is that if you want smart DNS then run a local nameserver. Then have ONE in resolv.conf pointing to 127.0.0.1 | 07:55 |
| Xenguy | So my current config is good | 07:55 |
| rwp | This works great everywhere but... mobile laptops connecting to captured portals. Captured portals force you to use their dns in order to either pay or accept their EULA. | 07:56 |
| Xenguy | That's the first line in my config, with the idea that the other 2 are fallback nameservers in case the first doesn't work for some reason | 07:56 |
| rwp | So in those cases you probably will have to flip back and forth between two configurations. | 07:56 |
| Xenguy | Hrm, good to be aware of that use case | 07:57 |
| rwp | Waiting 30 seconds or whatever it is (I forget) to fallback to the next nameserver is pretty painful though. | 07:57 |
| gnarface | darwin: i also recommend running a local DNS server instead if you're having problems like that. i use bind9 here, but dnsmasq and unbound are also popular. | 07:58 |
| rwp | That's why we all should run a local nameserver like unbound or bind. Unless talking to a captured portal. | 07:58 |
| Xenguy | Fortunately unbound seems to work quite nicely | 07:58 |
| gnarface | darwin: (of those, dnsmasq is probably the easiest to configure because it just reads your /etc/hosts file) | 07:59 |
| Xenguy | I also found when I switched to it that it seemed to speed up resolution of domain names | 07:59 |
| Xenguy | Less lag | 07:59 |
| Xenguy | yw | 08:14 |
| Alverstone | evdev: OpticalMouse: Unable to query fd: No such device | 11:22 |
| Alverstone | happens when I unplug and plug back | 11:22 |
| Alverstone | so only one user can use the mouse | 11:23 |
| Alverstone | workaround: never unplug | 11:23 |
| Alverstone | i'm angry | 11:23 |
| gnarface | Alverstone: i'm sure there's a better workaround | 11:29 |
| gnarface | you get angry too quick | 11:29 |
| Alverstone | https://bbs.archlinux.org/viewtopic.php?id=201708 | 11:29 |
| gnarface | no, running it as root isn't it | 11:30 |
| Alverstone | I have no idea why no such device | 11:30 |
| Alverstone | it's in /dev/input | 11:30 |
| gnarface | maybe it changed orders | 11:30 |
| Alverstone | (II) seatd_libseat opened /dev/input/event3 | 11:30 |
| gnarface | the best option is to just dial it into your xorg.conf i think | 11:31 |
| Alverstone | (**) evdev: OpticalMouse: Device: "/dev/input/event3" | 11:31 |
| Alverstone | evdev: OpticalMouse: Unable to query fd: No such device | 11:31 |
| Alverstone | (EE) PreInit returned 2 for "PixArt A4Tech OpticalMouse" | 11:31 |
| gnarface | also "unable to query fd" may be referring to /dev/fd/ instead of /dev/input/ | 11:31 |
| Alverstone | (II) UnloadModule: "evdev" | 11:31 |
| Alverstone | can i force x to try it again? | 11:32 |
| Alverstone | :) | 11:32 |
| Alverstone | i DO have xorg.conf entry for my mouse and keyboard | 11:35 |
| Alverstone | that sets evdev driver instead of libinput | 11:35 |
| Alverstone | it might be elogind issue, which drives me mad | 11:35 |
| Alverstone | how am i supposed to debug that moster | 11:35 |
| rrq | Xorg might rescan on VT switching (away and back) | 11:42 |
| Alverstone | doesn't work | 11:42 |
| rrq | otherwise it's the role of libinput I think | 11:42 |
| Alverstone | Could not revoke evdev on device fd: No such device | 12:03 |
| Alverstone | [seatd/seat.c:338] Could not revoke evdev on device fd: No such device | 12:03 |
| Alverstone | [ERROR] [seatd/seat.c:338] Could not revoke evdev on device fd: No such device | 12:03 |
| Alverstone | [ERROR] [seatd/seat.c:222] Could open device: client is not active | 12:03 |
| Alverstone | [ERROR] [seatd/client.c:238] Could not open device: Operation not permitted | 12:03 |
| Alverstone | [ERROR] [seatd/seat.c:222] Could open device: client is not active | 12:03 |
| Alverstone | [ERROR] [seatd/client.c:238] Could not open device: Operation not permitted | 12:03 |
| Alverstone | Okay so I'm going to remove elogind completely and try again | 12:04 |
| Alverstone | nope | 12:15 |
| Alverstone | same as always | 12:15 |
| Alverstone | it's libseatd issue then | 12:15 |
| Alverstone | whose else | 12:15 |
| rrq | I think i's the evdev module that doesn't discover the usb path change. libinput should do that | 12:20 |
| rrq | when the mouse is plgged in it gets a new eventN devnode | 12:20 |
| Alverstone | nope | 12:21 |
| Alverstone | libinput same behaviur | 12:21 |
| Alverstone | libinput same behavior | 12:21 |
| Alverstone | sometimes even worse, mouse disappears for all users | 12:21 |
| Alverstone | I posted log somewhere above | 12:22 |
| Alverstone | no? | 12:22 |
| Alverstone | sec | 12:22 |
| Alverstone | libinput: PixArt A4Tech OpticalMouse: Failed to create a device for /dev/input/mouse2 | 12:23 |
| * Alverstone shrugs | 12:24 | |
| rrq | do you have seatd o logind or user has devnode access? | 12:25 |
| rrq | o = or | 12:25 |
| Alverstone | tried everything | 12:26 |
| Alverstone | just elogind | 12:26 |
| Alverstone | just seatd | 12:26 |
| Alverstone | elogind along with seatd | 12:26 |
| Alverstone | always the same | 12:26 |
| rrq | hmmm and /dev/input/mouse2 is the right devnode ? | 12:27 |
| Alverstone | yep, it picks it up from udev | 12:28 |
| rrq | what's the seatd related error line just prior to thet libinput errorline? | 12:29 |
| Alverstone | posted above | 12:30 |
| Alverstone | it can't open device | 12:30 |
| Alverstone | for some reason | 12:30 |
| rrq | hmm but it can open the mouse devnode before unplug+plug ? | 12:31 |
| Alverstone | yes | 12:31 |
| Alverstone | work before unplugging | 12:31 |
| Alverstone | works* | 12:32 |
| Alverstone | replug and only active user picks it up | 12:32 |
| Alverstone | others loose | 12:32 |
| Alverstone | and can't reclaim | 12:32 |
| Alverstone | I'm gonna do something ostensibly ugly right now | 12:34 |
| rrq | ah the mouse is picked up by the active Xorg, but after VT switch the target xorg doesn't pick it up... as if that Xorg's libinput is not signalled by udev | 12:34 |
| Alverstone | I'll report back if it works | 12:34 |
| Alverstone | Yep | 12:37 |
| Alverstone | just as I thought | 12:37 |
| Alverstone | I installed xserver-xorg-core from trixie and everything just works | 12:37 |
| Alverstone | Devuan's libseat patch is broken | 12:37 |
| Alverstone | need to file bug report | 12:37 |
| Alverstone | I don't have time currently | 12:37 |
| Alverstone | But can help | 12:37 |
| Alverstone | Better if you discuss it on the forum thought | 12:38 |
| Alverstone | I can't hang out in IRC 24/7 | 12:38 |
| rrq | ok.. please describe the use case with VT switch etc in some detail | 12:38 |
| Alverstone | what do you mean | 12:38 |
| Alverstone | I use several users for different tasks | 12:39 |
| rrq | what did you mean saying "others loose" ? | 12:39 |
| Alverstone | no graphical login manager | 12:39 |
| Alverstone | OK later | 12:39 |
| Alverstone | is Discuss/Devuan an appropriate forum | 13:10 |
| Alverstone | or where do I report it at all | 13:10 |
| Alverstone | https://git.devuan.org/ ? | 13:11 |
| Alverstone | register and report there? | 13:11 |
| gnarface | forum link in channel topic | 13:11 |
| gnarface | incidentally, i'm curious if it works if you just disable seatd instead and run Xorg from startx directly | 13:12 |
| gnarface | no elogind or whatnot | 13:12 |
| gnarface | though i'm not completely sure i understand your use case | 13:12 |
| Alverstone | yes, forum | 13:13 |
| gnarface | and i'm still only assuming the issue is path changes on hotplug... | 13:13 |
| Alverstone | but where exactly? | 13:13 |
| gnarface | oh, uh... not sure really | 13:13 |
| Alverstone | gnarface, I'll describe it in the post and we'll figure out all questions there | 13:13 |
| gnarface | just pick what seems to make sense then link it here | 13:13 |
| Alverstone | gnarface, https://dev1galaxy.org/viewtopic.php?pid=52656#p52656 | 13:31 |
| Alverstone | How do you mount external media? | 19:15 |
| Alverstone | (block devices) | 19:15 |
| djph | (1) plug it in, (2) mkdir -p /mnt/whatever, (3) mount -t filesystem-type /dev/whatever /mnt/whatever | 19:17 |
| Alverstone | requires root | 19:18 |
| user71 | I use dmesg|tail to find out the device name afer plugging it in | 19:18 |
| user71 | using root is fine for this | 19:19 |
| Alverstone | nope, I don't allow regular users to execute anything as root | 19:19 |
| Alverstone | I disabled all ways to escalate privileges except some parts of the polkit | 19:19 |
| debdog | create an ebtry for that device in /etc/fstab to allow users to mount it | 19:22 |
| debdog | USB, I assume? | 19:22 |
| Alverstone | bad, because relies on UUID/LABEL/etc | 19:22 |
| Alverstone | it's not flexible at all | 19:22 |
| user71 | I have accidentally deleted / as root only once, mustn't be that common | 19:22 |
| debdog | where etc is what? | 19:24 |
| debdog | install some desktop environment which takes care of that | 19:25 |
| Alverstone | don't wanna udisks2 | 19:26 |
| debdog | still, USB? | 19:26 |
| debdog | is this for a certain computer or generally? | 19:26 |
| debdog | you can still use /dec/sdX (regarding etc) | 19:26 |
| debdog | where X is one letter higher than the installed drives | 19:28 |
| debdog | I ahve entries for that in fstab for up to three USB devices (which you still have not confiremd or not) with up to five partitions on them. works | 19:30 |
| debdog | or install Windows | 19:33 |
| debdog | SCNR | 19:33 |
| Alverstone | udisks is interesting because it also allows you to setup loopback interfaces | 19:38 |
| Alverstone | pmount and udevil -- not really | 19:38 |
| Alverstone | on the other hand, it depends on polkit and elogind | 19:39 |
| Alverstone | I struggle to understand why though | 19:40 |
| Alverstone | polkit doesn't really have use cases, it's policy system is stupid and unmanageable | 19:40 |
| Alverstone | its* | 19:40 |
| Alverstone | elgoind... is there really no way which user invoked the damn executable and whether his session is currently active? | 19:41 |
| * Alverstone rants | 19:41 | |
| Alverstone | no way to know* | 19:41 |
| fsmithred | Alverstone, use pmount for user mounting external devices. | 19:44 |
| fsmithred | you can also specify in the config file internal block devices that are allowed to be mounted. | 19:45 |
| fsmithred | *I use* | 19:46 |
| fsmithred | udevil is another one | 19:46 |
| Alverstone | fsmithred, looks pretty good to be honest, but doesn't cover the whole scope of possible use cases. If pmount could loopback mount files though... It isn't even very hard I think, you'd need to set nosuid and the like, same way FUSE does, to achieve security | 19:46 |
| Alverstone | as for the rest, it's just a losetup call | 19:47 |
| Alverstone | it's one of them DIY moments when you realize you can't do everything in this life, so you have to put up with bloated udisks | 19:47 |
| Alverstone | I have a bunch of ISOs that use that stupid UDF file system extension, they can't be mounted with fuseiso | 19:49 |
| Alverstone | udisks kinda does the job | 19:49 |
| fsmithred | ouch | 19:49 |
| fsmithred | what's the advantage of mounting as a loopback? | 19:50 |
| Alverstone | how else would I mount them? | 19:50 |
| Alverstone | the ISOs are stored as regular files | 19:50 |
| fsmithred | oh, I thought you were mounting external media | 19:50 |
| Alverstone | well, since I have udisks anyway, I use it for external media too | 19:51 |
| Alverstone | to be 100% honest with you, I got a bit carried away thinking how cool the world could be if seatd could mediate access to video and audio devices, so X11/pulse/pipewire would become free of systemd completely | 19:52 |
| Alverstone | dream world | 19:52 |
| Alverstone | :D | 19:52 |
| Alverstone | need to go cool down I guess | 19:53 |
| fsmithred | archivemount will mount isos | 19:55 |
| Alverstone | fsmithred, hm really/ | 20:06 |
| Alverstone | ? | 20:06 |
| Alverstone | it has O(4) reading algo though iirc | 20:07 |
| Alverstone | wait a sec | 20:07 |
| fsmithred | what's that? | 20:07 |
| Alverstone | O(2) actually | 20:07 |
| Alverstone | see here https://github.com/google/fuse-archive | 20:07 |
| Alverstone | performance comparison | 20:07 |
| fsmithred | I use it to mount isos and I've never even though about performance or speed. I don't think I'd be able to tell the difference with a stopwatch. | 20:10 |
| fsmithred | if you're truncating big archives, then maybe it's important. | 20:10 |
| Alverstone | you would with large files | 20:10 |
| fsmithred | between 1-2G usually | 20:10 |
| fsmithred | only time I use it is when making a live-usb and copying to the usb is infinitely longer than mounting it. | 20:12 |
| Alverstone | I try to use archivemount, it creates README.txt file owner by root and with 400 permissions | 20:13 |
| Alverstone | wtf | 20:13 |
| Alverstone | owned* | 20:13 |
| Alverstone | I can't read what's inside | 20:13 |
| Alverstone | and I don't have allow_others | 20:13 |
| fsmithred | where does it make that? | 20:14 |
| Alverstone | Inside mountpoint | 20:15 |
| Alverstone | the mountpoint | 20:15 |
| Alverstone | google's fuse-archive tell me the following | 20:15 |
| Alverstone | This disc contains a "UDF" file system and requires an operating system | 20:15 |
| Alverstone | that supports the ISO-13346 "UDF" file system specification. | 20:15 |
| Alverstone | since they both use libarchive | 20:15 |
| Alverstone | I assume I'm out of luck here | 20:15 |
| Alverstone | FUSE filesystems have a way of creating a README.txt file when they can't actually mount and putting the error msg there | 20:16 |
| Alverstone | this UDF is a monster | 20:17 |
| Alverstone | iirc mostly used by windows | 20:17 |
| Alverstone | and there are probably no reasons for it to exist | 20:17 |
| Alverstone | so nobody wrote a fuse implementation afaik | 20:17 |
| fsmithred | I see there is a package called udftools | 20:19 |
| Alverstone | yeah it's cli | 20:19 |
| Alverstone | not actually an fs | 20:19 |
| Alverstone | not bad | 20:20 |
| Alverstone | until your ISO actually contains software to be installed :) | 20:20 |
| Alverstone | unpacking 20+GB of data just to install it | 20:20 |
| Alverstone | I don't have the nerve | 20:20 |
| Alverstone | I google this periodically and always find nothing but udisks | 20:21 |
| Alverstone | heh | 20:21 |
| dvbst | hello, is getting my printer working on this realistic or just a dream? | 20:27 |
| Alverstone | dvbst, getting your printer to work is very easy. 1. Check if there is a manufacturer provided driver. 2. If there is, install and enjoy. 3. If there isn't, check if it works out of the box. 4. If it doesn't, you don't need a printer | 20:29 |
| dvbst | where do i check | 20:32 |
| Alverstone | https://github.com/sjpotter/udf-fs "quick and dirty" | 20:33 |
| Alverstone | dvbst, manufacturers website, actually | 20:33 |
| Alverstone | iirc CANON printers shipped some linux drivers | 20:33 |
| Alverstone | they even worker | 20:33 |
| Alverstone | they even worked | 20:33 |
| Alverstone | closed source ofc | 20:33 |
| Alverstone | kinda scary :) | 20:33 |
| debdog | sometimes just mentioning the printer's designation can have a huge impact on the replies | 20:35 |
| dvbst | it says hp laserjet 1018 on the printer but idk if thats really it | 20:36 |
| Alverstone | https://support.hp.com/us-en/drivers/hp-laserjet-1018-printer/1814092 | 20:36 |
| Alverstone | does it look like this | 20:37 |
| debdog | then you might want to have a look at hplip | 20:37 |
| Alverstone | https://support.hp.com/us-en/document/ish_7795955-7796023-16 | 20:38 |
| Alverstone | https://developers.hp.com/hp-linux-imaging-and-printing | 20:38 |
| Alverstone | https://developers.hp.com/hp-linux-imaging-and-printing/supported_devices/index | 20:38 |
| Alverstone | says your is supported | 20:39 |
| Alverstone | apt show hplip | 20:40 |
| dvbst | awesome | 20:40 |
| Alverstone | seems like it might solve your problem | 20:40 |
| dvbst | wow it works | 20:54 |
| dvbst | i love you | 20:54 |
| Alverstone | aw thanks | 21:08 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!