| Guest6809 | i'm getting this error when doing sudo apt upgrade: | 00:35 |
|---|---|---|
| Guest6809 | Err:1 http://deb.devuan.org/merged daedalus-security/main amd64 fonts-opensymbol all 4:102.12+LibO7.4.7-1+deb12u5 | 00:35 |
| Guest6809 | 404 Not Found [IP: 106.178.112.231 80] | 00:35 |
| gnarface | Guest6809: how many times did you try? deb.devuan.org is a dns round-robin, so sometimes if one mirror fails you just have to retry and it'll work | 00:37 |
| Guest6809 | ok, i will try again | 00:39 |
| Guest6809 | i tried it just now three times | 00:39 |
| Guest6809 | same error | 00:39 |
| gnarface | hmm, odd... | 00:41 |
| gnarface | did you remember to run "apt update" once first? | 00:42 |
| gnarface | the download link here worked for me https://pkginfo.devuan.org/cgi-bin/package-query.html?c=package&q=fonts-opensymbol=4:102.12+LibO7.4.7-1+deb12u5 | 00:42 |
| gnarface | try "apt update" if you haven't yet, then try it again | 00:42 |
| gnarface | maybe your package cache is just stale | 00:43 |
| Guest6809 | yes i run apt update first | 00:43 |
| Guest6809 | ok will try again | 00:43 |
| gnarface | does this link work for you? http://deb.devuan.org/merged/pool/DEBIAN-SECURITY/updates/main/libr/libreoffice/fonts-opensymbol_102.12+LibO7.4.7-1+deb12u5_all.deb | 00:43 |
| gnarface | that one worked for me | 00:43 |
| gnarface | and it should be the same... | 00:44 |
| Guest6809 | yep still the same | 00:44 |
| Guest6809 | same error | 00:44 |
| gnarface | you still get an error on that too? | 00:44 |
| gnarface | huh | 00:44 |
| Guest6809 | exactly | 00:44 |
| Guest6809 | ok will try that link now | 00:44 |
| gnarface | when you retried with apt, did the error come up from the same ip address (106.178.112.231) every time, or was it a different one on each try? | 00:47 |
| Guest6809 | 404 Not Found [IP: 106.178.112.231 80] | 00:50 |
| gnarface | so, same IP every time | 00:50 |
| Guest6809 | yes | 00:51 |
| Guest6809 | honestly i don't know what file to put the link you gave me | 00:51 |
| Guest6809 | i tried to put it in sources.list | 00:51 |
| gnarface | this link? i meant to just try to download that with your web browser http://deb.devuan.org/merged/pool/DEBIAN-SECURITY/updates/main/libr/libreoffice/fonts-opensymbol_102.12+LibO7.4.7-1+deb12u5_all.deb | 00:51 |
| Guest6809 | and then the apt update didn't work | 00:52 |
| Guest6809 | oh | 00:52 |
| Guest6809 | let me try | 00:52 |
| Guest6809 | @gnarface thanks man the apt upgrade after i downloaded that file | 00:57 |
| Guest6809 | it worked | 00:57 |
| gnarface | alright, well i'm glad it worked anyway | 00:58 |
| gnarface | that doesn't really tell us what was going wrong, but sometimes there are temporary failures during mirror replication | 00:58 |
| Guest6809 | alright. thanks. bye for now. | 00:59 |
| systemdlete | gnarface, just got my c920s logitech webcam. what did you say I need to load for that? | 02:27 |
| systemdlete | I've plugged it in, and made sure uvcvideo was loaded, but so far, nothing | 02:28 |
| gnarface | systemdlete: install uvcdynctl and issue uh... "uvcdynctl -c" i think? make sure it's /dev/video0 | 02:29 |
| gnarface | maybe just to disambiguate, and make sure your kernel is seeing it, issue a "v4l2-ctl --list-devices" first | 02:30 |
| gnarface | oh, and sorry it's uvcdynctrl that you want to nistall | 02:30 |
| gnarface | *install | 02:30 |
| gnarface | it should also be able to list cameras with "uvcdynctrl -l" but afaik maybe only the uvcvideo ones, which aiui is a subset of the v4l2 ones | 02:31 |
| systemdlete | oh, I saw a mint article which said uvcvideo... not your fault I got confused | 02:32 |
| gnarface | oh, uvcvideo is the kernel module, you'll need that too but i don't think you'll have to do anything special for it to autoload | 02:32 |
| systemdlete | ok... | 02:32 |
| systemdlete | so I am running cheese, but see no image. However, the blue leds on the sides are on | 02:32 |
| gnarface | what did "uvcdynctrl -l" say? maybe you've just got something else showing up as /dev/video0, which would be the typical default | 02:33 |
| gnarface | cheese may be trying to read from your tv capture card by default or something like that | 02:33 |
| systemdlete | oh | 02:33 |
| systemdlete | wait I did not install uvcdynctl yet... | 02:34 |
| systemdlete | (doing now) | 02:34 |
| gnarface | uvcdynctrl | 02:34 |
| systemdlete | ^ what you said | 02:34 |
| systemdlete | I see it also wants to install libwebcam0 | 02:34 |
| systemdlete | I can see where that might be useful... | 02:34 |
| systemdlete | :D | 02:34 |
| gnarface | hmm, yea did that on mine too, i think | 02:35 |
| gnarface | mostly i just use it with ffmpeg, but basically anything should work as long as you have the right path and permissions | 02:35 |
| gnarface | uvcdynctrl just adds access to LED and autofocus controls | 02:36 |
| gnarface | sorry, not even auto-focus, just auto exposure | 02:36 |
| gnarface | but that can be helpful | 02:36 |
| Guest574 | i'm trying to do automated repository configuration as per the posgresql website: sudo /usr/share/postgresql-common/pgdg/apt.postgresql.org.sh | 02:40 |
| gnarface | systemdlete: here, the setup script i'm using for timelapse... not sure if this helps or not: https://paste.debian.net/1329666/ | 02:40 |
| Guest574 | Here's the error message i got: https://paste.debian.net/1329667/ | 02:42 |
| systemdlete | gnarface, is there a simple video and audio program I can use for testing? Including audio input? (I'm assuming this device has input?) | 02:42 |
| gnarface | systemdlete: (probably could use adjustments, but my lighting situation isn't ideal and i don't really know what i'm doing so...) | 02:42 |
| systemdlete | (understood) | 02:42 |
| systemdlete | (still trying to make this baby burp a bit) | 02:42 |
| systemdlete | e.g., does ffmpeg have a front end UI? | 02:43 |
| gnarface | systemdlete: yea, it's got a mic, you should be able to read from it just fine with arecord. any v4l2 supporting program should be able to get images out of it... ffmpeg does not but vlc is probably a good simple place to start | 02:43 |
| systemdlete | (all gui hate aside) | 02:43 |
| systemdlete | does vlc have audio and video input? I never noticed that | 02:43 |
| gnarface | i think it should have them both, yea, i just find arecord easier to use | 02:43 |
| systemdlete | I only use vlc for video out | 02:43 |
| gnarface | it has tv card support, and the v4l2 interface should make webcams and tv cards basically the same thing | 02:44 |
| gnarface | VLC -> Media -> Open Capture Device | 02:44 |
| systemdlete | ok, switching back over to the PC with the software... | 02:44 |
| systemdlete | (and thanks) | 02:45 |
| gnarface | no problem | 02:45 |
| Guest574 | hey gnarface | 02:45 |
| Guest574 | i'm trying to do automated repository configuration as per the posgresql website: sudo /usr/share/postgresql-common/pgdg/apt.postgresql.org.sh | 02:45 |
| Guest574 | Here's the error message i got: https://paste.debian.net/1329667/ | 02:45 |
| gnarface | Guest574: i'm not familiar with that process, but from this error this seems to be an issue with their repository not recognizing Devuan as a valid distro. you could try the Debian equivalent to daedalus, which would be bookworm, but do so at your own risk; this could hose your install | 02:46 |
| Guest574 | yeah i'm not doing it. thanks anyway. | 02:47 |
| gnarface | Guest574: it might be a simple thing for them to fix, you could ask them about it... | 02:47 |
| Guest574 | i see. thanks. | 02:47 |
| gnarface | Guest574: (it might not be, and either way they might cop an attitude about it too, so also do that at your own risk...) | 02:48 |
| Guest574 | yes, thanks for the heads up | 02:49 |
| gnarface | no problem | 02:49 |
| Guest574 | bye for now. | 02:49 |
| systemdlete | gnarface, when I open the dialog, I get a dropdown for the device type, which is pre-selected as video camera, but there are no devices listed for the sound and video . Also, any thoughts on choice of video standard for my testing? | 02:59 |
| systemdlete | (I've only recorded video once before, about 20 years ago, with an AIPTEC camera (worked, but only B&W) | 03:00 |
| systemdlete | and I should mention, I'm working on chimaera, not daedalus. | 03:02 |
| gnarface | systemdlete: the system i'm using mine on is a raspberry pi running ascii. what did you get for the output of "uvcdynctrl -l" ? | 03:05 |
| gnarface | maybe you need to load the uvcvideo kernel module manually after all? maybe you forgot to add yourself to the video group? ... maybe you just forgot to hit the arrow to open the pulldown? (it should be a freeform text field you can just type /dev/video0 into manually too, but there's clearly something simple missing here) | 03:07 |
| gnarface | damnit | 03:08 |
| rrq | systemdlete: you have v4l-utils installed? | 03:13 |
| gnarface | he already disconnected | 03:14 |
| gnarface | ping timeout 248 seconds at 18:08:05 | 03:14 |
| gnarface | hopefully the camera didn't explode... | 03:16 |
| rrq | right... I also have guvcview installed which is a bit helpful | 03:16 |
| gnarface | well, it shouldn't be difficult, and i still think it's probably a pretty simple permissions or path issue, but i haven't actually tested this camera since ascii, and it's busy right now doing timelapses so i can't yet | 03:18 |
| gnarface | well, if he comes back, someone get him to verify that it's creating a /dev/video# device when it is plugged in, first, then we can go from there | 03:28 |
| systemdlete | It's starting to work, a bit, here and there. | 03:50 |
| systemdlete | I'm getting a very sketchy video, if it works. The sound is awful, feedback keeps getting louder and louder. | 03:51 |
| systemdlete | I'm guessing this unit has a powerful input. | 03:51 |
| systemdlete | (is that "high gain?" idk) | 03:51 |
| systemdlete | That's in a VM. But on hardware/host, it does not play video. Tailing kern.log, I see that there are zillionsof radeon error messages. | 03:52 |
| systemdlete | *zillions of | 03:52 |
| gnarface | systemdlete: you disconnected before i could answer your last questions, but i guess you must have found the right /dev/video* device? | 03:58 |
| systemdlete | well, nvm. It works perfectly for zoom, which is sufficient for now. | 03:58 |
| systemdlete | yes, I did, and thanks for your patience. | 03:58 |
| gnarface | it should have its own volume control in alsamixer | 03:58 |
| gnarface | it might help if you turn it down a bit | 03:58 |
| systemdlete | Not sure why, but Comcast does weird stuff to my connection intermittently. | 03:59 |
| systemdlete | oh, sure, for vlc. | 03:59 |
| systemdlete | My immediate need is for several zoom classes I am taking this semester as part of an adult CE program | 03:59 |
| gnarface | well, the alsamixer controls should work on anything, but zoom might have it s own conflicting volume controls, i'm not familiar with it | 03:59 |
| systemdlete | But I will definitely fool around with the controls a bit and see if I can get vlc working full-fledged. | 04:00 |
| gnarface | improving the video quality is mostly an issue in my experience of just dropping the resolution until you can attain a suitable framerate, then changing the auto-exposure setting to something appropriate for your lighting environment | 04:00 |
| systemdlete | gnarface, oddly, zoom seems to have the magical ability to get everything working automatically. | 04:01 |
| gnarface | however the camera can take both mjpeg and yuyv (raw) recordings, so if you're having trouble with performance in zoom make sure you're set to mjpeg (i only know how to do that with v4l2-ctl) | 04:01 |
| systemdlete | But I think you are right; zoom probably does things differently. | 04:01 |
| gnarface | oh, well zoom might be programmed by someone competent with the v4l2 interface so it might just have sane defaults | 04:02 |
| gnarface | i use that setup script i linked just because it was easier than figuring out how to make ffmpeg configure the camera right | 04:02 |
| systemdlete | competent? When so many millions depend on zoom every minute of the day? | 04:02 |
| systemdlete | why, that would be downright... appropriate. | 04:02 |
| gnarface | afaik vlc doesn't really have any defaults, so you'd have to dial them in in the "advanced" pulldown, in some cases manually typing them in as a text string in some vlc specific format | 04:03 |
| systemdlete | I chose mp3/mp4 | 04:04 |
| systemdlete | I found that not choosing something caused vlc to hiccup | 04:04 |
| systemdlete | vlc is very buggy, I find | 04:04 |
| gnarface | yea, vlc can be picky, and it doesn't really error check settings, so you have to know which codecs/containers work on your hardware and are compatible with each other too, because vlc will happily let you set nonsensical combinations that just crash | 04:05 |
| systemdlete | If I add videos to the ones I've already chosen, it doesn't always work. It is ignorant of webp files (but not webm or others) | 04:05 |
| gnarface | or that its requisite libraries are missing transcoding support for, or whatever | 04:05 |
| systemdlete | very linux/open source | 04:06 |
| systemdlete | heheheh | 04:06 |
| systemdlete | a dig on linux, not dig on devuan | 04:06 |
| gnarface | while you were gone, rrq mentioned that guvcview might be useful | 04:06 |
| systemdlete | ooh, thanks rrq! | 04:06 |
| systemdlete | the file chooser dialog in vlc is a horror. It doesn't handle embedded spaces in file and directory names correctly. | 04:08 |
| systemdlete | as well as some other characters | 04:08 |
| systemdlete | installing guvcview now | 04:08 |
| gnarface | the important tip i have is just to work on a simple setup script that will work with anything's generic defaults, partially because not everything has the ability to control all the camera's features like uvcdynctrl does, partially just so you can remember what settings you like and are best for the camera, and partially because it's gonna lose all those settings as soon as you unplug it... | 04:09 |
| gnarface | if zoom is smart enough to keep its own settings though, then good for it | 04:10 |
| systemdlete | great idea, thanks for that tip | 04:11 |
| systemdlete | guvcview is great, but I can't find a way to listen or record. | 04:11 |
| gnarface | it might only do picture | 04:11 |
| systemdlete | no, it actually does have audio controls | 04:11 |
| gnarface | oh, huh | 04:12 |
| systemdlete | there is a filter called "echo" but I think that might be for the input side only | 04:12 |
| systemdlete | special effect, or the like | 04:12 |
| systemdlete | not sure | 04:12 |
| systemdlete | first time working with it | 04:12 |
| gnarface | "arecord -D hw:1,0 test_output.wav" should work | 04:12 |
| gnarface | if you just want to get a wav file from the mic | 04:13 |
| gnarface | or hw:#,0 whichever is the right one for the camera | 04:13 |
| gnarface | "arecord -l" should list them all | 04:13 |
| systemdlete | arecord is only for audio though, right? | 04:13 |
| gnarface | yea | 04:13 |
| gnarface | but the mic in there just shows up as a generic alsa mic, nothing special about it being in a camera | 04:14 |
| gnarface | any alsa recording tools should work | 04:14 |
| gnarface | again, like the video device though you gotta make sure you actually pick the right one (because it almost certainly won't be the default) and you need to make sure you have permissions too | 04:15 |
| gnarface | in my experience it does default to max volume, unmuted though, so i don't think you should have to open alsamixer to set a volume first just to get sound | 04:15 |
| gnarface | you might need to turn it down to get rid of feedback if you're not using headphones though | 04:15 |
| systemdlete | not quite to max, about 1/2 way in the white zone | 04:15 |
| gnarface | hmm, that might be a change zoom made | 04:16 |
| systemdlete | getting the headphones working has been a chore | 04:16 |
| systemdlete | they are USB also | 04:16 |
| gnarface | (it will hold whatever volume settings you set until you unplug it) | 04:16 |
| gnarface | ah, i see, so your output device is also not default | 04:16 |
| systemdlete | yes, looked at alsamixer after I ran zoom on same system | 04:16 |
| gnarface | yea, i can see how that would be annoying | 04:16 |
| systemdlete | right. So there is this magical config file that I once had working for the headphones, but I lost it in numerous updates | 04:17 |
| systemdlete | for alsa I mean | 04:17 |
| gnarface | so, "aplay -l" should also list your USB headphones as their own device, and since they're a hotplug device they might be in a different order | 04:17 |
| systemdlete | I don't quite get the whole syntax of the files, and the ones in /etc are not the same as the ones in $HOME | 04:18 |
| systemdlete | I can see why a lot of people prefer to just use pulse and be done with it. But I try to avoid such fluff if I can. | 04:18 |
| systemdlete | yeah, good idea. aplay -l | 04:19 |
| systemdlete | let me go try that | 04:19 |
| gnarface | i've been working on learning the syntax of the alsa configs a bit, i've gotten a lot of the important basics down, but i still have to often swap ~/.asoundrc files for different purposes | 04:19 |
| systemdlete | I think I ran into that also. | 04:20 |
| systemdlete | but it has been a few years since I had it working | 04:20 |
| systemdlete | once again, and as always, thanks for the assistance. | 04:21 |
| systemdlete | You've given me more than a few clues here | 04:21 |
| gnarface | no problem | 04:23 |
| onefang | Just a quick visit between todays various journeys. I just feed ALSA to JACK if I need to swap things around for different reasons. Makes it easy to switch inputs and outputs of everything JACK known about. JACK is more a pro tool though. | 06:16 |
| onefang | And naturally the very first thing that happened after I said "no breaky breaky mirror mirror" is someone managed to break one somehow. | 06:17 |
| rustyaxe | i been quite happy with pipewire/qpwgraph but i've got weird things going on (i've got multiple sound devices - speakers for notifications/whatever watching in browser, etc and a headset for talking to people. qpwgraph makes that much more pleasant to switch around on the go | 06:57 |
| freem | Hi. Regarding coredumps (I mentioned receiving help here a while ago, and did an update on this 2 days ago), apparently oom-killer will *not* generate coredumps by default on debian (probably the kernel's silly default config, really... let's allow programs to allocate more memory than there is, kill them randomly when the memory allocated is used, and *not* generate a coredumps... linux is indeed great :'(), even if the program does not have signal | 12:55 |
| freem | handlers (I made it sure to get rid of that). A quick research led me to this link https://serverfault.com/questions/662201/get-a-core-dump-of-debugging-a-process-killed-by-oom-killer, which itself hints at /proc/sys/vm/oom_dump_tasks, "Enables a system-wide task dump (excluding kernel threads) to be produced when the kernel performs an OOM-killing" | 12:55 |
| freem | I'm not fond of having a "system wide dump" really... it's going to freeze system and fill storage, it's a terrible behavior! I'll try to find an alternative, and use this if no other choice. I wonder how many changes I'll have to do to the system's configuration to behave sanely... | 12:56 |
| freem | sometimes I wonder how many programmers receive more economics lessons than electronics ones | 12:57 |
| buZz | well, why would you need a coredump of a program thats shut down? | 12:57 |
| buZz | its not crashing, its killed | 12:57 |
| freem | a program leaking memory enough to fill the system's ram is affected by a bug, obviously, and a coredump would help fixing the bug | 12:58 |
| buZz | freem: but oom_killer isnt about 'program using too much memory' | 12:59 |
| buZz | its about 'computer has too little free memory' | 12:59 |
| buZz | imho debuggin memoryleaks is not oom_killer's job, but valgrind or similar | 13:00 |
| freem | considering the fact I had many, many system freezes which led to hours of work lost in the past because of overcommit and other shits like that, I do disagree | 13:00 |
| buZz | why not just save regularly? :P | 13:00 |
| freem | you can't run a program under valgrind 100% time, it's way too slow, and to cover stricly all paths? won't happen any day soon. | 13:01 |
| buZz | obviously, you dont debug memoryleaks in programs 100% of the time too | 13:01 |
| buZz | typically you just run debian's programs on devuan , and they're nearly all well tested to not have memoryleaks | 13:01 |
| freem | but you do want informations about the process when it happens | 13:01 |
| buZz | oom_killer does log what process it killed and how much memory it was using, no? | 13:02 |
| freem | in a perfect world where debian have packaged all the software and where all the softwares are bug free, sure | 13:02 |
| buZz | been a short decade since i ran into it :) | 13:02 |
| buZz | freem: ah ok, what programs are you running that leak often? | 13:02 |
| freem | but here I'm debugging a game dedicated engine which not packaged in debian :) | 13:02 |
| buZz | they're from outside repo? | 13:02 |
| buZz | ah, which engine? | 13:03 |
| freem | daemon-engine, it's a quake3 successor kind of like ioquake3, with the main difference you don't write gamelogic in quakeC, but in any lang supported by LLVM, up to C++14 because gNaCl | 13:03 |
| buZz | cute :) | 13:03 |
| buZz | you'd think a q3 descendant didnt introduce many memoryleaks :) | 13:04 |
| freem | gNaCl being also the reason why debian does not packages it (I'm fine with that, really, the dep is weird, old, unmaintained, etc) | 13:04 |
| freem | yeah, except... q3 uses their own VM system, which loads their own, historically non-standard and non-free C variant | 13:05 |
| buZz | right | 13:05 |
| freem | the leak happens in a setup which is normally used for debugging purposes, too, not in the normal runs | 13:05 |
| buZz | ah hehe | 13:05 |
| freem | but I need that debugging (and improved perfs compared to the in-VM runs) ability since I am modding myself | 13:05 |
| buZz | i've had such issues on arduinos (notoriously low memory mcu setup) | 13:05 |
| buZz | where my debugging introduced so much errors that i just switched to 'blink a led' as debugging ;) | 13:06 |
| freem | https://forums.unvanquished.net/viewtopic.php?t=2471 get a video about some of the upcoming changes in that engine, featuring the only game using it | 13:06 |
| freem | if you are interested :) | 13:07 |
| buZz | freem: gigantic offtopic, but have you seen those AIs that can write 'full games' nowadays? like on https://websim.ai/ | 13:07 |
| buZz | heh i already had the unvanquished website open ;) | 13:07 |
| freem | well... no, but then I also know how AIs are trained | 13:07 |
| buZz | its kinda amazing stuff | 13:07 |
| freem | take everything humans produced, good or bad (more bad than good, sadly), mix it, take the average | 13:08 |
| buZz | sure :) | 13:08 |
| freem | so AI will only do sub-par quality imo :) | 13:08 |
| buZz | try it sometime ;) | 13:08 |
| freem | and this for a very huge resource waste | 13:08 |
| freem | there are also quite the legal grey around this stuff | 13:08 |
| buZz | its not perfect, its not ideal, its just great fun | 13:08 |
| freem | I agree it can be fun to toy with such tools though | 13:09 |
| buZz | i'm not suggesting you sell whatever it produces ;) | 13:09 |
| freem | right | 13:09 |
| buZz | most often, i use AIs outputs as inspiration to a topic | 13:09 |
| buZz | here's some effect i had websim.ai write for me ; https://space.nurdspace.nl/~buzz/trippy.html | 13:10 |
| freem | well, in short, unvanquished is a fork of tremulous, which was a quake3 mod. It's graphics are more modern, at least | 13:10 |
| buZz | freem: whats that XreaL ? | 13:10 |
| freem | I think an AI based tool was actually used by a contributor to refresh some old foss textures | 13:10 |
| buZz | ah https://www.phoronix.com/review/xreal_engine | 13:11 |
| freem | I don't know the whole ancestry, but IIRC tremulous run on ioquake3, Xreal was an attempt at improving the rendering. Daemon-engine is apparently a continuation of that effort. | 13:11 |
| buZz | bit hard to find :) | 13:11 |
| freem | old stuff :) | 13:12 |
| freem | https://wiki.unvanquished.net/images/0/0d/Unvanquished_and_Daemon_history.png | 13:12 |
| buZz | quake3 isnt the newest either ;) | 13:12 |
| freem | will be more exact than my memories | 13:12 |
| freem | tbh unvanquished was almost dead for a long while, but got new activity since a few years (3 or 4?) including non negligible bot improvements, and currently rendering improvements which are for next release, theoretically soon, but really when it's ready | 13:14 |
| freem | the biggest problem for distro maintainers is really that google-NaCl is a quite old and complicated piece of software, and it makes boostrapping very complicated. There are plans (but plans are just that, plans) to go wasm instead, though | 13:16 |
| buZz | webonly? | 13:16 |
| freem | no | 13:16 |
| freem | nothing web related | 13:16 |
| buZz | oh, i thought wasm was typically web oriented | 13:17 |
| freem | the point is to have a way to distribute the games accross network, while keeping that foreign binary in a sandbox | 13:17 |
| freem | it is, but it does not means it's the only way to use it | 13:17 |
| buZz | right, ok | 13:17 |
| buZz | fancy stuff :) | 13:17 |
| freem | yeah, it's a bit too complicated to me :) | 13:17 |
| freem | my own plans are more about getting bots smarter and funnier | 13:18 |
| buZz | i loved that story about the quake3 demo having bad bots when released | 13:19 |
| freem | but it's not really a trivial stuff neither, so being able to debug is important | 13:19 |
| buZz | and someone in the community made 'really good bots' and then ID software hired them to port the bots to the final version | 13:19 |
| freem | there was a thesis on q3's bots | 13:19 |
| buZz | cute :) | 13:19 |
| freem | pretty interesting read, if you ask me | 13:20 |
| buZz | i did my oral english exam on quake1 :P | 13:20 |
| freem | https://www.researchgate.net/publication/240430519_The_Quake_III_Arena_Bot | 13:20 |
| freem | neat | 13:20 |
| buZz | we had to 'pick a article and do a oral report on it' , i picked some 10-15 page article on quake1 , it was just released | 13:20 |
| freem | nice | 13:21 |
| freem | well, there were interesting improvements on game bots area since those times, notably this: https://web.archive.org/web/20230912145018/https://alumni.media.mit.edu/~jorkin/goap.html | 13:22 |
| buZz | cute | 13:23 |
| freem | the big idea of that stuf is: you implement a bunch of actions, give them pre-requisites and expected effects (both are "atoms"), describe a goal (with atoms), and give the bot a model of the world (again based on those atoms). The stuff builds a graph, and runs A* on it. This gives the plan the bot will try to follow. | 13:24 |
| freem | the nice thing is that it means you do not have to care about other actions implemented when you want to add one, it will find it's place automatically. | 13:24 |
| freem | I've spent enough hours (more like months...) on the behavior tree system implemented in unvanquished to understand the value of that idea :D | 13:25 |
| buZz | :) | 13:25 |
| freem | so I've spent some time working at reimplementing things to use this, but motivation and times varies | 13:26 |
| freem | and I started that right before those were a bit hard to find | 13:26 |
| DFP | I respect the approach, but it irks me that there isn't a way for the bots to anticipate the opponent's actions. I was thinking of a system where opponents get a turn too. | 13:28 |
| freem | I did other researches on the topic, ofc. There are many ways to implements AIs, but this one seems the most recent, and compared to others, the less painful to implement, when you want really complex behaviors. And a team-based game which allows building your base, and other stuff, is clearly complicated | 13:28 |
| DFP | So you plan your first move, then consider the moves that the opponents might take, pick best, and plan your responses... So like a minmax. | 13:28 |
| DFP | Still thinking about it though, no concrete results. | 13:29 |
| freem | it is hard to anticipate a human's behavior, but I would say that bots usually do, on very short term (1 or 2 seconds) | 13:29 |
| DFP | I was thinking to use the bots' own algorithm to predict humans. | 13:29 |
| DFP | "If I was this character, what would I do?" | 13:29 |
| freem | ah | 13:30 |
| freem | yeah, that seems very hard... | 13:30 |
| DFP | If you restrict the problem to turn-based domain, seems a bit easier. | 13:30 |
| freem | but I have watched some F.E.A.R. videos after finding goap stuff, and it *looks* as if bots are trying to do nasty stuff | 13:30 |
| buZz | i'd love to see GANs or LLMs applied to 'be game AI' | 13:31 |
| freem | such as trying to catch their enemy from 2 sides | 13:31 |
| freem | I'm certain for on-line games, this will come | 13:31 |
| buZz | i saw unreal et al are deploying 'npcs can talk through a LLM' | 13:31 |
| freem | for off-lines games though? Dunno. LLMs are quite resource intensive, right? | 13:32 |
| buZz | the future seems fascinating :) | 13:32 |
| buZz | freem: well, some are even cpuonly and not superslow | 13:32 |
| buZz | like https://github.com/ggerganov/llama.cpp | 13:32 |
| DFP | My latest thought was giving each 'actor' in the plan their own timeline, and have each action give a time estimate, so actions fill that timeline. So the planning essentially becomes almost turn-based, just with potentially multiple actions per turn depending on 'time points' available. | 13:32 |
| buZz | and 'webgpu' is a thing too | 13:32 |
| freem | I recently learnt (yesterday?) that modern CPUs have LLM-dedicated chips, too, they call that NPUs apparently? | 13:33 |
| buZz | yeah basically the modern version of hw opencv accelerators ;) | 13:33 |
| buZz | (like that myriad asic that intel bought) | 13:33 |
| freem | I don't think I know any game which is *not* turn based, technically | 13:33 |
| freem | the only question is about how long the turn is. In unvanquished's case, the default is 0.1s | 13:34 |
| freem | "real time" in IT is not actually real time, but about running frames in a specific, well defined duraction (not too fast, not too long). Which is kind of close to turn based, except for performance constraints. | 13:35 |
| freem | well, IT have many variants of "real time" anyway | 13:36 |
| DFP | Yeah, but in this case it's more of a 'frame of mind' thing. You could look at something in different ways to find a perspective that makes it easier to solve a problem. | 13:37 |
| freem | indeed | 13:37 |
| freem | guess I'll write a "daemon" which watches my process and "kill -9" it when it goes above a threshold of memory... but that won't give me the info about which call generates the problem :'( | 13:39 |
| freem | oh... silly me! | 13:41 |
| freem | I apparently forgot to disable overcommit up there | 13:42 |
| freem | or did I enabled it for some reason now... | 13:42 |
| buZz | freem: cant you make that NaCL stuff analyze the stuff it runs? | 13:42 |
| freem | if I get clean crash, then I might be able to get more data | 13:42 |
| freem | apparently, it's possible, but I know nobody hosting a mod who managed | 13:43 |
| freem | it's easy to say everything's fine, when you only host the stable release :) | 13:43 |
| freem | apparently you can't even gdb in that stuff | 13:44 |
| freem | which is why mod hosters run in one of the 2 native modes, one being "executable" and the other "dll", the dll one is the only one which allows to gdb in there and thus get intel about the problems, easily | 13:44 |
| freem | why is it impossible to ask the kernel to send a -9 seriously | 13:45 |
| freem | and is it really impossible? | 13:45 |
| freem | oh, perhaps I could use earlyoom | 13:51 |
| freem | no, it's silly | 13:51 |
| freem | I just need to disable overcommit and I'll see what happens. It's a toy server anyway, ain't a problem if something bad happens | 13:52 |
| freem | yay! it worked! | 14:04 |
| buZz | yay! | 14:04 |
| freem | so the solution to debug a damn memory leak without keeping gdb-server 24/7 is: disabled overcommit_memory (set to 2), and make sure the buggy process does not have crash handlers | 14:05 |
| freem | disabling overcommit_memory means: asking the system to *not* lie to processes | 14:05 |
| freem | about remaining resources... | 14:05 |
| buZz | hehe | 14:05 |
| buZz | seems important yeah ;) | 14:05 |
| freem | which I was persuaded to have done on my vps, but turns out I didn't... despite I do this on all my systems | 14:06 |
| freem | so that's the story... overcommit_memory not only makes your system thrashing to death if you dared to have swap (I disable it as well), but also prevents you to debug buggy memory code | 14:07 |
| freem | I wonder if freebsd have similar feature... if not, then I regret Debian/kFreeBsd was dropped (or better, a kOpenBSD or kNetBSD?) | 14:08 |
| Guest3374 | Warning: Failed to open the file /home/devuan/.local/bin/yt-dlp: No such file | 14:28 |
| Guest3374 | Warning: or directory | 14:28 |
| Guest3374 | does Devuan doesn't have .local directory? | 14:28 |
| freem | devuan is likely your username | 14:31 |
| freem | you can create your .local directory if missing. System packages will not touch the home directories of users | 14:32 |
| freem | I more suspect the missing directory is /home/devuan/.local/bin though | 14:32 |
| Guest3374 | that's correct i guess | 14:33 |
| freem | and this, because many, many programs (all those following freedesktop's only sane recommendation) will create it | 14:33 |
| Guest3374 | i'm trying to install yt-dlp | 14:33 |
| freem | just "apt-get install yt-dlp" should do it? | 14:34 |
| Guest3374 | and the command is this: curl -L https://github.com/yt-dlp/yt-dlp/releases/latest/download/yt-dlp -o ~/.local/bin/yt-dlp | 14:34 |
| freem | why do you not use the one provided by devuan or debian? | 14:34 |
| Guest3374 | i though the one from github is the official | 14:34 |
| Guest3374 | thought* | 14:35 |
| freem | this command was probably given to you by someone who made their custom setup, adding ~/.local/bin to their $PATH. I personnally use ~/.bin for this | 14:35 |
| Guest3374 | let me try | 14:35 |
| freem | oh, it possibly is, but it might have requirements which are not yet in devuan/debian | 14:35 |
| freem | I would advise you to use devuan/debian packages as much as possible, you will have less troubles. Using upstream binaries can be useful, though, when debian/devuan ones are lagging too far behind | 14:36 |
| Guest3374 | same error | 14:36 |
| freem | what is the command you run, exactly? | 14:36 |
| Guest3374 | ok i'll try the apt install yt-dlp | 14:37 |
| freem | % yt-dlp 2>&1 | upl | 14:37 |
| freem | https://p.mort.coffee/dKs | 14:37 |
| freem | this is what I get when I try to run yt-dlp without URI, from debian's repo | 14:38 |
| Guest3374 | yes, done installing using apt install | 14:39 |
| Guest3374 | Oh no. It looks like Youtube is disallowing downloading now | 14:41 |
| Guest3374 | it's being THROTTLED | 14:41 |
| freem | ah, that is very possible. You could try to go through some invidiuos instance, though | 14:41 |
| freem | also, I believe I only have yt-dlp installed because it's an mpv recommended dep, so perhaps mpv will know how to magically work around this kind of stuff | 14:42 |
| freem | I really don't know, though, I'm not exactly a youtube user | 14:42 |
| Guest3374 | mpv | 14:44 |
| Guest3374 | is it a player? | 14:44 |
| sixwheeledbeast | yt-dlp is often out of date from the repos | 14:46 |
| freem | yes, mpv is a player, that is launched from command-line | 14:46 |
| freem | akin to mplayer2 | 14:47 |
| Guest3374 | i just want to download Youtube vids | 14:48 |
| Guest3374 | let me try to update yt-dlp | 14:49 |
| freem | well, sorry, I do not know enough to help more. Maybe someone knows how to install the upstream binaries | 14:49 |
| freem | oh! | 14:49 |
| freem | you could also try with the backports | 14:49 |
| freem | --\ Versions de yt-dlp (3) | 14:49 |
| freem | p A 2023.03.04-1~bpo11+1 | 14:49 |
| freem | i A 2023.03.04-1 | 14:49 |
| freem | p A 2024.08.06-1~bpo12+1 | 14:49 |
| freem | this ^ is what I have in aptitude | 14:49 |
| freem | the backported version seems very fresh, here is a link about debian's backports (which works for devuan as well, pretty sure): https://backports.debian.org | 14:50 |
| freem | https://backports.debian.org/Instructions/ | 14:50 |
| Guest3374 | thanks. let me check on this | 14:51 |
| Guest3374 | Here's the output: https://paste.debian.net/1329707/ | 14:54 |
| freem | those to really thank are the maintainers who handle that burden :) | 14:54 |
| freem | hm... can you upload the result of `cat /etc/apt/sources.list /etc/apt/sources.list.d/*`? | 14:55 |
| freem | so that we can check you have the backports enabled | 14:55 |
| Guest3374 | sure | 14:55 |
| freem | and after adding them, did you ran `apt-get update`? | 14:56 |
| Guest3374 | a moment please | 14:56 |
| freem | (if you added it) | 14:56 |
| freem | no hurry, I'm trying to debug something on the side anyway :D | 14:56 |
| Guest3374 | Here's the output of cat: https://paste.debian.net/1329708/ | 14:58 |
| sixwheeledbeast | While I wouldn't generally recommend it, you would have been fine following the instructions from the git wiki. | 14:59 |
| sixwheeledbeast | the current version is 2024.08.06 | 14:59 |
| Guest3374 | yes but it throws an error regarding a directory missing | 15:00 |
| Guest3374 | Warning: Failed to open the file /home/devuan/.local/bin/yt-dlp: No such file or directory | 15:01 |
| Guest3374 | sixwheeledbeast | 15:01 |
| sixwheeledbeast | only unstable has that version. | 15:02 |
| sixwheeledbeast | in the repos | 15:02 |
| sixwheeledbeast | sid/ceres | 15:02 |
| freem | no, it's also in backports | 15:02 |
| freem | https://p.mort.coffee/bAU this is my own sources.list, and I do have the 2024.08.06 version available | 15:03 |
| Guest3374 | check this @sixwheeledbeast https://paste.debian.net/1329709/ | 15:03 |
| freem | Guest3374: you have not added the backport repos | 15:03 |
| sixwheeledbeast | I mean if the repo has the latest versions great you would be best to use that but i always find it was out of date and throws an error when I needed it. | 15:03 |
| freem | you need a line looking like: "deb http://deb.devuan.org/merged/ daedalus-backports main" I think | 15:04 |
| Guest3374 | ok let try to add it on my sources.list | 15:04 |
| freem | not using devuan, I can't tell the exact line, though | 15:04 |
| freem | (the main difference between my system and a devuan is likely the name of the release though, I don't have systemd stuff installed neither :D) | 15:05 |
| freem | (so I'm in the same boat) | 15:05 |
| Guest3374 | yep same error even after adding it in the sources.list | 15:07 |
| sixwheeledbeast | don't try and solve it two ways, get it from backports if it's available. | 15:08 |
| Guest3374 | yes i already tried to add the backports in the sources.list | 15:08 |
| Guest3374 | and same error | 15:08 |
| freem | how does one install a backport thing from CLI already? | 15:08 |
| freem | I always install stuff from aptitude's ncurses interface, it makes picking a specific version and seeing problems trivial... but it makes it harder for me to give commands to people :/ | 15:09 |
| freem | since basically, to install something from backports, you need to 1) enable backports 2) update the repos and 3) install the backported version, considering apt will default on the stable one | 15:10 |
| sixwheeledbeast | not if the version number is higher | 15:10 |
| rrq | Guest3374: it's in daedalus-backports/main ... maybe you forgot "apt-get update"? | 15:11 |
| freem | I'm pretty certain backports won't be installed even if version is higher | 15:11 |
| freem | except if you worked your way with apt-pinning, but I would not recommend that to a beginner | 15:12 |
| sixwheeledbeast | apt policy yt-dlp | 15:13 |
| Guest3374 | does ~/.local/bin/ directory exist in devuan? | 15:15 |
| freem | only if user creates it | 15:16 |
| Guest3374 | i see | 15:16 |
| sixwheeledbeast | No, I wouldn't expect it to be in PATH either | 15:16 |
| freem | debian or devuan will *never* touch anything in your home directory | 15:16 |
| freem | programs will, ofc, but not debian nor devuan themselves | 15:16 |
| Guest3374 | then why is the installation instruction in git points to that directory? | 15:17 |
| freem | the installation instructions probably are "it works for me, on my system" stuff | 15:17 |
| sixwheeledbeast | I did spot that, you can put it anywhere but would need to be PATH | 15:17 |
| freem | a classic. | 15:17 |
| Guest3374 | what does PATH mean | 15:17 |
| sixwheeledbeast | /usr/local/bin is usual | 15:17 |
| freem | and this classic is the reason why distros are so useful, they avoid trying to guess what the dev's system looks like | 15:18 |
| Guest3374 | so should i just put the yt-dlp in /usr/local/bin then? | 15:18 |
| sixwheeledbeast | PATH is where the system looks for binaries when you call for them | 15:18 |
| Guest3374 | i see | 15:18 |
| sixwheeledbeast | if you echo $PATH you will see | 15:19 |
| freem | you can add folders to the $PATH variable, as well | 15:19 |
| sixwheeledbeast | I add personally add $HOME/.bin to PATH | 15:20 |
| freem | /home/USERLOGIN/.bin:/usr/local/bin:/usr/bin:/bin:/usr/local/games:/usr/games:/home/USERLOGIN/.bin:/home/USERLOGIN/.bin | 15:20 |
| Guest3374 | here's the output for the echo of PATH | 15:20 |
| freem | this is the PATH I use, with USERLOGIN replaced by my real login, obviously | 15:20 |
| sixwheeledbeast | so put it in /home/USERLOGIN/.bin | 15:20 |
| freem | sixwheeledbeast: you misread :) | 15:20 |
| sixwheeledbeast | oh | 15:20 |
| sixwheeledbeast | you do the same | 15:21 |
| freem | I was giving an example of what it looks on my system, I'm not the one who's asking | 15:21 |
| sixwheeledbeast | :nod: | 15:21 |
| freem | yeah, it's handy to put some scripts in there, and I do have a few. I also have a workaround for the fact some clang versions can't be debugged with ASAN/gdb because of binary format changed :/ | 15:21 |
| freem | (the last one only, that is) | 15:22 |
| Guest3374 | i got it | 15:28 |
| Guest3374 | i just removed the ~ after the -o | 15:28 |
| freem | nice | 15:31 |
| Guest3374 | thanks guys | 15:32 |
| Guest3374 | i learn some things tonight | 15:32 |
| freem | thank us by helping someone when you have the occasion :) | 15:33 |
| fsmithred | FYI: Here a line from my ~/.profile that I did not add: | 15:37 |
| fsmithred | PATH="$HOME/.local/bin:$PATH" | 15:37 |
| fsmithred | it also adds $HOME/bin | 15:37 |
| Guest3374 | thanks fsmithred. good to know. | 15:38 |
| fsmithred | That's in daedalus. I don't think it was like that in chimaera or earlier. | 15:38 |
| Guest3374 | @freem i will if can | 15:38 |
| freem | :) | 15:39 |
| Guest3374 | if i can* :)) | 15:39 |
| freem | it's how things can be built | 15:39 |
| * freem also can remember learning stuff *while* trying to help | 15:39 | |
| Guest3374 | yes, i agree that's the best way to learn | 15:42 |
| marcfp | hi | 16:31 |
| marcfp | i'm MarcFP, finally with my "new computer" installed with Devuan | 16:31 |
| marcfp | thank's to all people, whose help me, and whose not helped too :) you are always welcome :) | 16:32 |
| Xenguy | MarcFP, Enjoy : -) | 17:14 |
| MarcFP | thank's :) | 17:16 |
| freem | is there is a way to use apt-file to search for a *source code* file, when the deb-src repo is enabled? | 18:13 |
| fsmithred | freem, what file are you looking for? | 18:40 |
| freem | I was searching for ./nptl/futex-internal.c | 18:41 |
| freem | which is the name given by gdb. Basically I was trying to find out why the memleak kills the program I'm debugging when either __futex_abstimed_wait_common64, __libc_accept, __GI___clock_nanosleep or __recvmsg_syscall were called | 18:42 |
| freem | I don't see why the 3 last syscalls would allocate, so was searching for the 1st ones code | 18:43 |
| fsmithred | I have src enabled for daedalus, and that file isn't showing up with apt-file | 18:43 |
| freem | yeah, me neither... a feature request perhaps :D | 18:43 |
| fsmithred | download the source and use find | 18:44 |
| freem | well, I moved to another way to find my problem, since anyway this would not help | 18:44 |
| freem | the app crashed when there was 1gig free according to free -h, so I smelled something like a fork() call | 18:44 |
| fsmithred | fwiw, you're talking way over my head | 18:45 |
| freem | since I'm curious and this surprised me, I wanted to peek | 18:45 |
| freem | add some procrastination since I'm a bit tired of the hunt... | 18:46 |
| fsmithred | source code would be at salsa.debian.net | 18:46 |
| freem | ah, thx | 18:46 |
| fsmithred | sorry, salsa.debian.org | 18:46 |
| freem | or could simply apt-get source glibc, it seems likely related | 18:46 |
| freem | glibc, libphread and kernel | 18:47 |
| fsmithred | searching just futex gives me texlive_publishers as a package | 18:47 |
| freem | still weird for a "wait" instruction to allocate though | 18:47 |
| freem | anyway, I *need* some beer now | 18:47 |
| freem | so I'll afk :) | 18:47 |
| freem | brb in ~30min I guess | 18:48 |
| paculino | I think I accidentally made a fork bomb (by a mismatch in brackets around the conditional which would cause a fork). As a result I cannot use my shell... I'll have to reboot, won't I? | 20:07 |
| paculino | I didn't have a system monitor open and only had one terminal open, so cannot end the process. | 20:08 |
| paculino | Taskbar just vanished, so I'm going with yes | 20:11 |
| freem | can you use the TTYs? | 20:15 |
| freem | ctrl+alt+f[1-6] on traditional systems | 20:15 |
| freem | I don't remember if I had to work around stuff to keep that or not | 20:16 |
| paculino | The actual answer is rebooting is not possible, but a crash will be imminent anyway, so it doesn't matter much | 20:24 |
| fsmithred | alt-sysrq keys | 20:25 |
| fsmithred | alt-sysrq with R S U B to reboot | 20:26 |
| fsmithred | or alt-sys-K to kill X session | 20:26 |
| paculino | I remapped alt to altgr since I lack altgr | 20:26 |
| paculino | would mod4 work? | 20:26 |
| fsmithred | I don't know what that is | 20:26 |
| paculino | hyper I think? | 20:27 |
| fsmithred | or, can you ssh in from another machine? | 20:27 |
| paculino | I crashed, started again, crashed again, then realized that I had to use awesome-wm so that the program would not start within a minute of booting thanks to me putting it in the startup programs list that mate has | 20:28 |
| fsmithred | ouch | 20:29 |
| paculino | I am now recompiling what I did | 20:29 |
| paculino | I thought that the fork would only regenerate the precalculated points if the old list was too short, so removed the line that ended the program (returning a segfault) after regenerating it (so it would start the loop over) | 20:30 |
| paculino | I am glad that I sed an alias to change the screen brightness; otherwise I would be blinded by the light right now | 20:31 |
| paculino | I am, however, extremely impressed by the ram useage now. This is less than half my prior record (651 MiB) | 20:33 |
| onefang | Well done. | 20:33 |
| paculino | I forget; what is the method to stop the starting of X when booting? I have the delay so that I can just barely login before it starts if I try five times. | 20:34 |
| freem | one can't ssh from another system when under fork bomb, since ssh would need to fork | 20:36 |
| freem | (to fsmithred) | 20:36 |
| freem | this -> https://p.mort.coffee/2Cb <- is the "alt gr" key | 20:38 |
| freem | (according to xev) | 20:38 |
| freem | ffs.. ok, I need a dev tool which requires kde | 20:40 |
| paculino | Would a chroot work in a system with a fork bomb program as autostart (fortunately I only had it start from mate and be delayed) | 20:40 |
| freem | which means systemd | 20:40 |
| freem | which means time to say fuck you debian | 20:40 |
| freem | can someone point me at a sources.list to move to devuan? | 20:41 |
| freem | I happen to *need* kcachegrind, and this stuff is tied to KDE, which some smartass decided must require systemd bullshit | 20:41 |
| paculino | kde requires systemd now? | 20:41 |
| freem | on debian, yes | 20:41 |
| paculino | Do you just need the default debian sources.list? | 20:42 |
| freem | ain't systemd a pretty lovely worm? | 20:42 |
| freem | no, the devuan ones. I'll migrate today (again, technically, since I was an early adopter :p) and be happier | 20:42 |
| freem | I don't feel in mood to fight the maintainer's stupid ideas today | 20:43 |
| paculino | So you temporarily migrated to debian? | 20:43 |
| freem | plus, they never even answer when you ask a question on #debian, so... I guess time to embrace | 20:43 |
| freem | a long while ago, I moved to devuan out of curiosity. Aptitude was broken, though, so I moved back after a while | 20:44 |
| paculino | Here's my sources.list, freem: | 20:44 |
| freem | thx in advance :) | 20:45 |
| paculino | Wait, I now remember that the pastebin is preferred | 20:45 |
| freem | :) | 20:45 |
| freem | % upl .bin/upl -k | 20:45 |
| freem | https://p.mort.coffee/Ihf | 20:45 |
| freem | is what I use ;) | 20:45 |
| freem | I do need to improve it to put the URI in both stdout and xclip though | 20:46 |
| freem | just | tee | xclip or something I suppose | 20:46 |
| freem | or maybe socat, since tee only prints to standard output and *files* | 20:47 |
| freem | oh, silly me | 20:47 |
| freem | % echo foo | tee /dev/stdout | 20:47 |
| freem | foo | 20:47 |
| freem | foo | 20:47 |
| freem | solution's so obvious | 20:47 |
| freem | (the point to print in both is to allow both selection for some reason *and* to avoid grabbing mouse to copy it) | 20:48 |
| paculino | https://www.devuan.org/os/documentation/install-guides/daedalus/bookworm-to-daedalus is better than just showign my sources.ist | 20:50 |
| paculino | sources.list * | 20:50 |
| paculino | Well, I'll log out so I can return to mate; hopefully it works this time | 20:51 |
| freem | thx | 20:52 |
| paculino | Oh, it appears I never set up a log out widget when trying the bare minimum setup in case I needed this as backup | 20:54 |
| freem | what is the devuan equivalent of bullseye? | 21:22 |
| freem | I still have programs from then, because they'd segfault if I take the stable ones | 21:22 |
| freem | despite the code source of the author was *not* changed, the version nubmer being exactly the same (so it's a debian patch which fucks everything... what changed, I wonder... hm... might start with a s, and end with a d?) | 21:23 |
| freem | wxHexEditor being the name, and I do like this tool, a graphical hexeditor comes in handy at some times | 21:23 |
| freem | I have bullseye, bookworm, and buster sources | 21:24 |
| yeti | why graphical? | 21:24 |
| freem | because I'm not masochist | 21:24 |
| yeti | some editors do it well | 21:24 |
| freem | I do use a lot of CLI/TUI tools, but it is always when no graphical alternative can be used by a keyboard lover | 21:25 |
| freem | when a GUI tool *can* though, it outclasses TUI, which itself always outclasses CLI when it comes to interactively manipulate large amounts of data | 21:25 |
| yeti | BS | 21:26 |
| freem | if you say so | 21:26 |
| yeti | _o/" bye. | 21:26 |
| freem | but I will have you know that the reason I stay with debian/devuan is because aptitude is far better than any CLI package tool, for example | 21:26 |
| freem | otherwise I'd probably be a void-linux user | 21:27 |
| freem | a *resilient* rolling release distro, unlike arch | 21:27 |
| yeti | aptitude is good for some things, apt-get is much faster for others | 21:27 |
| freem | but I felt like I had hardly any ability to gain control on my system | 21:27 |
| freem | yes, ofc | 21:27 |
| freem | for scripting, I'd pick apt-get anyday | 21:27 |
| freem | (well, technically false, I would then pick dpkg...) | 21:28 |
| freem | (did that, already, in the past, for a job) | 21:28 |
| freem | in any case, I'm perfectly happy mixing 3 or more debians, it's much easier than what I used to do | 21:31 |
| freem | which is apt-pinning. I even remember moving from stable to testing to unstable+experimental, then back to stable. Without reinstall. | 21:31 |
| freem | this was only possible because of aptitude's TUI, the same way I only really master my system and learnt that much because of that same TUI, despite it's defaults (if only I could get rid of the solver!) | 21:32 |
| EarlGrey | Are the Devuan repositories down?? ive been trying to update my system but its not working. stuck on dev.devuan.org. | 21:33 |
| freem | now that I think about it, I could just do a frankenvbuan | 21:36 |
| freem | (no typo there) | 21:36 |
| freem | just need to replace one of the sources this way | 21:36 |
| freem | https://www.devuan.org/os/documentation/install-guides/daedalus/bookworm-to-daedalus this page is missing an important element for me though | 21:37 |
| freem | where does devuan mirrors the dbgsyms? | 21:37 |
| freem | also daedalus is a great name, it kinds of evoques to me the troubles one have to suffer to *not* be annoyed by systemd | 21:38 |
| fsmithred | EarlGrey, first, make sure you're using deb.devuan.org and not dev.devuan.org | 21:39 |
| freem | which makes me think, are there devuan people trying to take over the effort for a kFreeBSD? | 21:39 |
| fsmithred | second, sometimes you catch it when a mirror is updating and it'll work again a few minutes later. | 21:39 |
| freem | no, apparently, it's the DNS resolution which is failing, fsmithred, EarlGrey | 21:43 |
| freem | it says to be a temporary, so yeah, let's retry in a few moment... even if DNS errors are usually not clear | 21:44 |
| fsmithred | I just did an upgrade in the last five minutes | 21:45 |
| freem | % LANG=C ping deb.devuan.org | 21:46 |
| freem | ping: deb.devuan.org: Temporary failure in name resolution | 21:46 |
| fsmithred | ok, but why would that affect EarlGrey? | 21:47 |
| freem | I'm not a network expert though. I know there are ways to get more data but... | 21:47 |
| freem | I used *ping* not apt-get | 21:48 |
| freem | so we probably have common DNS somewhere in the pyramid | 21:48 |
| fsmithred | maybe. ping works here. | 21:49 |
| fsmithred | US East coast | 21:49 |
| freem | western europe here | 21:49 |
| fsmithred | I hit a .de mirror | 21:49 |
| fsmithred | 64 bytes from mirror.checkdomain.de (46.4.50.2): icmp_seq=1 ttl=40 time=106 ms | 21:50 |
| freem | I know there are some stuff about dnsmask and the likes to help on this, but I ain't network admin | 21:50 |
| fsmithred | me neither | 21:50 |
| freem | I have some basics of system admins, and relatively decent C++ coding skills, but is all | 21:50 |
| freem | I'm much more skilled at drinking beers, but that's likely offtopic :) | 21:51 |
| freem | you can't train in everything after all | 21:51 |
| EarlGrey | im trying to do a netinstall | 21:51 |
| EarlGrey | of peppermint os devuan, which is based on devuan's main repos | 21:52 |
| freem | what's that, out of curiosity? | 21:53 |
| fsmithred | it's a debian-based distro. I'm familiar with it. | 21:53 |
| fsmithred | EarlGrey, maybe try selecting a specific mirror from the list. | 21:53 |
| EarlGrey | its like linux mint, but with debian and devuan | 21:53 |
| EarlGrey | fsmithred i did | 21:54 |
| freem | linux mint? | 21:54 |
| freem | I know mint, not peppermint | 21:54 |
| freem | IIRC they have 2 variants, one on ubuntu and other on debian | 21:54 |
| freem | but this is about mint... | 21:55 |
| freem | oh, thx, EarlGrey | 21:55 |
| EarlGrey | this isnt mint, its peppermint | 21:55 |
| * freem brain lags a bit after all that leak hunt | 21:55 | |
| EarlGrey | it has 2 variants, debian and devuan. | 21:55 |
| freem | interesting | 21:55 |
| freem | do you have a link? I'm curious | 21:55 |
| freem | nvm, I'll get one myself | 21:55 |
| fsmithred | EarlGrey, which mirror are you using? | 21:56 |
| fsmithred | EarlGrey, which mirror are you using? | 21:57 |
| EarlGrey | i tried deb.devuan.org and us.deb.devuan.org | 21:58 |
| fsmithred | apt-panopticon shows deb.devuan.org had errors | 21:59 |
| fsmithred | on http | 21:59 |
| fsmithred | That is a round-robin. If you look at the list I linked, there are all the specific mirrors listed. Pick one close to you and try that. | 22:00 |
| EarlGrey | did somebody trip on the server plug? -_- | 22:00 |
| fsmithred | No, I think what happened is our mirror herd said he would be offline for a bit and there better not be any problems. | 22:00 |
| EarlGrey | and yet there are problems | 22:01 |
| fsmithred | Murphy strikes again. | 22:01 |
| EarlGrey | im guessing this isnt the first time something like this happened? | 22:02 |
| fsmithred | No, mirrors go down and up all the time. But as a said before, deb.devuan.org worked for me just a few minutes after you first posted. Is it still failing for you? | 22:03 |
| EarlGrey | i cant tell right now. im tryna reinstall it on my laptop and i failed like 5 times | 22:04 |
| EarlGrey | ill keep u guys updated i guess | 22:04 |
| fsmithred | wait | 22:04 |
| EarlGrey | hm? | 22:04 |
| fsmithred | if you can't easily select another mirror in the installer, choose Expert install | 22:04 |
| fsmithred | and you will get more choices. But I thought mirror selection was available in regular install. | 22:05 |
| fsmithred | ok, that's all | 22:05 |
| EarlGrey | i can do that if i install pure devuan, but i was trying to install peppermint os | 22:05 |
| fsmithred | oh | 22:06 |
| EarlGrey | it doesnt have expert install sadly | 22:06 |
| fsmithred | I'm not sure what installer they use or how you would hack their sources.list in the installer. | 22:06 |
| yeti | it'd add --download-only if it were a manual update via CLI, but apps sure do rock... :-Þ | 22:08 |
| yeti | repeating with that until everything was dl'ed | 22:08 |
| golinux | EarlGrey: Peppermint uses the Calamares installer. | 22:11 |
| EarlGrey | they added a mini installer based on the net installer recently. gives you more choices | 22:11 |
| golinux | And here is the announcement for it. https://dev1galaxy.org/viewtopic.php?pid=44698#p44698 | 22:14 |
| * golinux wipes egg from face for the second time today . . . | 22:15 | |
| EarlGrey | i would use mx linux but it broke on me like 3 times | 22:19 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!
