fsmithred | linux-image-5.8.0-1-amd64 | 00:02 |
---|---|---|
brocashelm | ok, it was kinda confusing me as to which one it actually was | 00:03 |
fsmithred | apt-cache search linux-image-5 | 00:07 |
gnarface | brocashelm: most likely you can exclude the *-dbg ones and the *-unsigned ones | 00:14 |
gnarface | brocashelm: and the *-pae ones are only for 32-bit boards that can handle more than 4GB of ram, which is far more rare than the incidents of people installing this kernel | 00:18 |
xrogaan | It's not so much being replaced, projects have a natural turnover. | 00:19 |
xrogaan | The systemd thing accelerated people leaving IIUC | 00:19 |
gnarface | brocashelm: (though if you really want to boot 32-bit linux on a amd64 board, the *-pae* ones may actually have a good use case. | 00:19 |
gnarface | ) | 00:19 |
Soltis | Okay so ... that doesn't solve the immediate problem: vim without full throttle stupidity. | 00:30 |
gnarface | Soltis: can the package be rebuilt without canberra? or maybe there's another package that doesn't have the graphical dependencies? for example there is emacs-nox | 00:32 |
fsmithred | vim-tiny needs less | 00:33 |
Soltis | vim-nox also depends on it | 00:33 |
Soltis | ... holy god, have these guys been drinking mercury since birth? | 00:33 |
Soltis | I saw someone offer the tissue paper thin argument that it was fine because alsa isn't x. | 00:33 |
fsmithred | uh, vim-tiny does not need libcanberra is what I meant | 00:33 |
fsmithred | lol | 00:34 |
gnarface | Soltis: probably something worse and far more sinister; they're probably completely competent and consciously committing evil. i just noticed who the libcanberra maintainer is. | 00:34 |
danuan | why would a commented out line after localhost in /etc/hostname actually help in setting the hostname from dhcp repply and not set it if there is no commented out line | 00:35 |
gnarface | fsmithred: you're right, though the fact that vim-nox does is pretty counter intuitive | 00:35 |
fsmithred | you're being polite | 00:35 |
gnarface | and i just noticed emacs-nox depends on libasound2 directly :-/ | 00:35 |
Soltis | gnarface: Well, yes, just between us chickens the odds this is accidental stupidity is effectively zero. | 00:36 |
gnarface | at least in ceres, anyway | 00:36 |
fsmithred | so your editor can make sounds. Oh goody. | 00:36 |
brocashelm | lol | 00:36 |
Soltis | So I do have vim-tiny installed, but it doesn't create a 'vim' command, just 'vi'. Let's see if it'll accept (and correctly handle) my .vimrc .... | 00:38 |
gnarface | Soltis: there might still be some difference between vim and vi, but there might also be something about the binary just behaving differently depending on what it is named | 00:40 |
Soltis | gnarface: Yeah, I'm used to that, but in this case it doesn't look like it works correctly. | 00:49 |
djph | Soltis: so, install vim? | 00:49 |
Soltis | djph: I'm not putting alsa on a damned server. That's slovenly administration. | 00:50 |
djph | Soltis: O-o ALSA for _vim_ ? | 00:50 |
Soltis | djph: Yes! Even vim-nox has it as a dep now | 00:51 |
djph | o_O | 00:51 |
djph | jeebus | 00:51 |
djph | well, that's annoying and stupid | 00:52 |
Soltis | djph: Welcome to the modern world of Linux. If you don't have a whiskey bottle at your desk, you're either an idiot or a better man than I. | 00:52 |
djph | Soltis: I mean, I know at least enough to not run systemd ... | 00:52 |
Soltis | djph: Hahaha | 00:52 |
Soltis | You know, I used to have _fun_ administering Linux systems. | 00:53 |
rrq | speech interface is quite useful for anyone with reduced eyesight | 01:05 |
Soltis | rrq: That's true, but it's also irrelevant. While non-zero, the number of blind server administrators is quite low - much too low to justify contaminating a dep tree like that. | 01:17 |
Soltis | rrq: Especially since the correct place to handle that would be in the desktop environment they're using | 01:18 |
fsmithred | or make it optional | 01:18 |
mason | Hrm, I'm not seeing vim-nox depending on ALSA. What's the chain? | 01:24 |
mason | And which ALSA package? | 01:25 |
fsmithred | libcanberra - libasound - alsa-utils | 01:26 |
mason | But, that's not a dependency. | 01:26 |
fsmithred | in ceres | 01:27 |
mason | Ah, I don't run that anywhere. | 01:27 |
mason | I guess I should spin up a test box. | 01:29 |
mason | (No pun intended.) | 01:29 |
rwp | OMG! Am I reading this right? vim depends upon ALSA? WAT?! | 01:32 |
rwp | I think I never noticed this before because I am not a vim user. I install nvi everywhere. [ That's right, and keep way off my lawn! :-) ] | 01:37 |
rwp | Also I am typing this in Emacs. For which I am complaining about the hard dbus dependency even in emacs-nox. | 01:38 |
gnarface | rwp: though most stuff that depends on dbus will work fine without it if you uninstall dbus later | 01:41 |
rwp | Emacs then emits a warning message. It works. But the warning message is annoying. | 01:41 |
rwp | A friend of mine found a way to disable that warning message. But I have didn't fight long enough and have dbus installed now. | 01:42 |
mason | upstream and stow can fix that for you :) | 01:45 |
rwp | mason, Yes. I can do the "Linux From Scratch" thing. Again. But I have been enjoying timely security upgrades appearing in the queue without my panic to get them created. | 01:51 |
mason | That's fair. Welcome to dbus. :P | 01:51 |
rwp | It's just so nice to be able to apt-get upgrade and avoid most security vulnerabilities. | 01:51 |
rwp | By the time I read about some vulnerability most of the time when I look I find that I already have the mitigation patch installed. That's just so nice. | 01:52 |
mason | sure | 01:52 |
mason | agreed | 01:52 |
rwp | When I tried to figure out why Emacs wanted dbus about the only thing I found was that someone could change a font globally and it would ripple into Emacs too. | 01:53 |
rwp | That is not compelling for me. Running emacs-nox inside an XTerm. I mean... How does that work there? | 01:53 |
sunshavi | emacs-nox does need dbus | 02:06 |
rwp | sunshavi, What does it need it for? Or are you simply confirming the Depends upon it? | 02:51 |
rennj | it needs that sweet sweet machine-id | 03:07 |
rennj | you know /var/lib/dbus/machine-id /etc/machine-id chrome,firefox,openoffice all check | 03:07 |
sunshavi | rwp: emacs-nox does NOT need dbus | 03:22 |
sunshavi | rwp: Perhaps a question on the mailing list (emacs-devel). Could solve the question ' what it is required dbus for emacs-nox | 03:25 |
sunshavi | rwp: on the tty or Xterm. It could be used for notifications with the var DBUS_SESSION_BUS_ADDRESS | 03:28 |
sunshavi | rwp: also on emacs-nox you could configure emacs with '--without-dbus' | 03:30 |
kreyren_ | How do i get all the dependencies required to build emacs on devuan? | 03:35 |
* kreyren_ remembers there being a cool command to install all required packages~ | 03:35 | |
fsmithred | apt-get build-dep emacs | 03:36 |
kreyren_ | fsmithred, thanku! | 03:36 |
rwp | rennj, Right. That machine-id that I just learned existed a few days ago. | 03:59 |
rwp | sunshavi, I don't think it should hard Depend upon dbus either. But emacs-nox Depends: libdbus-1-3 and so it is now. | 04:00 |
rwp | I could rebuild the package for myself. But dbus just sitting there not being used is less of a problem than other things that must be more actively avoided. | 04:01 |
sunshavi | rwp: '... actively avoided'. Like which ones? | 04:15 |
linux_n | My devuan time was correct but when dual booted to windows and back to devuan now my devuan time is incorrect. Anyway to fix devuan time? | 05:24 |
yanmaani | linux_n: Oh yeah, it's the hardware clock setting wrongly | 05:27 |
yanmaani | Linux and Windows use different systems | 05:27 |
yanmaani | https://mashtips.com/fix-linux-windows-dual-boot-clock-different-time/ | 05:28 |
linux_n | I get this error: timedatectl: command not found | 05:40 |
linux_n | Maybe those instructions are for os that uses systemd like debian but devuan uses a different one. | 05:44 |
yanmaani | yeah, it seems like it's harder to do in systemd. You can change it in Windows too. Or you can just use ntp to set the time on boot | 05:46 |
linux_n | I read this: timedatectl is part of systemd package. | 05:49 |
linux_n | so wont work on devuan only debian i think | 05:50 |
yanmaani | oh yeah, I mean outside of systemd | 05:50 |
yanmaani | maybe you can just run ntp on boot | 05:51 |
linux_n | devuan uses sysvinit instead of systemd. | 05:52 |
linux_n | ok thaks | 05:52 |
linux_n | thanks | 05:52 |
systemdlete2 | xfce weather widget was working until... about a week ago maybe? At least, that's when I started noticing. There was an API update for met.no -- did xfce devs update the weather applet? | 06:37 |
systemdlete2 | (I'm getting so tired of these API changes... why don't they just make everything upgradeable, but backwards compatible?) | 06:38 |
systemdlete2 | Oh, yeah. This is xfce on ascii, not beowulf. | 06:38 |
systemdlete2 | also, on another ascii, the isc-dhcp-client does not send the host-name parameter it seems. | 06:39 |
systemdlete2 | Originally, the line in the dhclient.conf file read: "send host-name = gethostbyname()" | 06:40 |
systemdlete2 | The docs for the version (4.3.5) indicate that the syntax does not indicate an equals sign. I removed it, just to see what would happen. No change. The dhcp server does not record the host-name for the client, but does for other clients. | 06:41 |
gnarface | systemdlete2: try it like this: send host-name "andare.example.com"; | 06:51 |
gnarface | systemdlete2: no = | 06:51 |
gnarface | systemdlete2: and with a hardcoded string first (test the gethostbyname call separately) | 06:51 |
systemdlete2 | As I said, above, I tried that and no luck. I also tried it with a static string like you suggest, and that did not work either. | 06:51 |
gnarface | hmmm | 06:51 |
systemdlete2 | Yeah. hmmmm. | 06:52 |
gnarface | sure it's the right file? | 06:52 |
systemdlete2 | That's the only conf file there, and it is the one containing the send command. | 06:52 |
gnarface | i vaguely recall this coming up a long time ago and there needing to be some command-line flag too | 06:52 |
systemdlete2 | oh great. | 06:52 |
systemdlete2 | I thought computers were supposed to make things simpler? | 06:53 |
systemdlete2 | OK, I'll take a look at that. thanks. | 06:53 |
gnarface | so some of this information may be inaccurate because the memory is dusty but it had to do with some protocol, standard, RFC, or something like that being disabled by default in isc-dhcp-client, and that being the part that included clients specifying their own hostname to the dhcp server (some microsoft concoction) | 06:54 |
Jjp137 | systemdlete2, I use the xfce weather widget on beowulf and it's also broken there; here's an issue about it: https://gitlab.xfce.org/panel-plugins/xfce4-weather-plugin/-/issues/27 | 06:55 |
Jjp137 | nothing we can do I guess | 06:55 |
systemdlete2 | Jjp137: This happens EVERY time that met.no chaanges the API. | 06:56 |
gnarface | systemdlete2: (maybe one of those RFCs mentioned at the bottom of the dhclient.conf manpage) | 06:56 |
systemdlete2 | gnarface: RFC2132, RFC2131 | 06:56 |
gnarface | systemdlete2: yes, one of those perhaps | 06:57 |
systemdlete2 | I don't understand why met.no can't just leave the standing API in place, and just add new functions as needed. That way they'd be compatible forwards and back. | 06:57 |
systemdlete2 | https://bugzilla.xfce.org/show_bug.cgi?id=14972 | 07:00 |
rwp | sunshavi, Why systemd of course! | 07:08 |
systemdlete2 | gnarface: On beowulf, which uses 4.4.1, the same exact send command works. | 07:12 |
systemdlete2 | send host-name = gethostbyname() | 07:13 |
systemdlete2 | ; | 07:13 |
systemdlete2 | btw, now the weather applet is working again. | 07:13 |
systemdlete2 | but the UI is not working. There is missing info in the forecast data, I can't switch between tabs, and I can't close it either. | 07:14 |
systemdlete2 | Had to brute-force kill the weather app. | 07:16 |
systemdlete2 | Restarted; waiting for data update. | 07:16 |
systemdlete2 | Actually, more vexing than either of these problems, making a system unusable: On my laptop (Cherry Trail) when the screen is rotated to landscape, the screen is leaving artifacts which do not clean up even if I move some windows around, trying to sort of "erase" the artifacts. They appear on the portion of the screen that would not be there in portrait mode, that is, the right hand section (about 1/3) of the screen. | 07:19 |
systemdlete2 | That's with beowulf. Ascii did not have this issue. | 07:20 |
systemdlete2 | I should also mention that xfce seems to work ok on beowulf (no artifacts), but mate has problem. | 07:23 |
systemdlete2 | And I still can't get multi-tap functions to work, which kind of puts the kabosh on using the machine as a tablet, practically speaking. | 07:24 |
systemdlete2 | nvm. On xfce, the double-tap does appear as an option I can change after all. | 07:27 |
systemdlete2 | I guess I'll just use xfce instead. | 07:27 |
systemdlete2 | It's OK, all I do with the tablet is read, and as long as I can do that much, I don't care too much about the other functions and issues. | 07:28 |
gnarface | systemdlete2: hmmm, i wonder if it's possible you're just missing the package for gethostbyname()? | 08:04 |
gnarface | systemdlete2: that might be implemented by a separate package | 08:05 |
systemdlete2 | wouldn't dhclient catch an error since it is launched with -v? | 08:05 |
systemdlete2 | but I'll look, thanks | 08:05 |
gnarface | i'm not sure, i just thought i recall that being in a separate package | 08:06 |
gnarface | systemdlete2: you wouldn't be missing the "hostname" package on that system, would you? | 08:10 |
danuan | systemdlete2: if we are talking about hostname not setting from dhcp , on my netboot systems for some reason /etc/hostname wants a #oldhostname as second line after localhost in first | 08:27 |
danuan | i have no idea why though | 08:28 |
ottavio | Is this article still valid https://devuan.org/os/documentation/dev1fanboy/en/stretch-to-beowulf.html | 13:15 |
syco- | hello | 13:47 |
Joril | hello! | 13:50 |
sammi` | bon jo | 13:52 |
user1041 | привед | 13:53 |
SuaveDandy | Hello. | 14:48 |
SuaveDandy | Is there such branch as contrib in Devuan? | 14:49 |
SuaveDandy | Examples only show that there are only main and non-free. | 14:49 |
systemdlete2 | gnarface, danuan: Just FYI. After some further investigation, I found out what was going wrong. Somehow, the checkbox setting hostname in wicd was getting unset, maybe after I switched from fixed to dhcp addresses. | 16:12 |
systemdlete2 | It turns out that wicd does not actually use the /etc/dhcp/dhclient.conf directly, it has its own versions in /etc/wicd. This must be why we all love networkmanager and wicd tools so much. | 16:13 |
systemdlete2 | It's actually self-documenting! The generated file ends up in /var/lib somewhere, and in it, it says that it will replace the next (last?) line with an appropriate send command... except that explanation is not clear until you run wicd with the checkbox checked. | 16:14 |
systemdlete2 | I couldn't tell if it wanted me to just plop my hostname in the next line, or if it was a name=val pair, or the actual invocation. Turns out the latter, but... well, anyway. It's fixed. | 16:15 |
ottavio | Is it possible to choose openrc as init in the installer? | 16:52 |
specing | yes | 16:53 |
ottavio | Has anybody converted LMDE3 (based on Stretch) to Beowulf? | 20:45 |
gnarface | thanks for the update systemdlete2 | 22:51 |
systemdlete2 | gnarface: Yvw, actually. You are a huge help around here, as are many of the others. | 23:46 |
yanmaani | Any way to automatically start zram on boot without systemd? | 23:48 |
buZz | sure, the normal way, with a init script | 23:51 |
buZz | lacking the knowledge to do so, you could just plonk the commands to start it in rc.local | 23:51 |
rennj | the goodstuff 4 lines for me | 23:52 |
rennj | # 400MB swap on zram | 23:52 |
rennj | modprobe zram | 23:52 |
rennj | echo 419430400 > /sys/block/zram0/disksize | 23:52 |
rennj | mkswap /dev/zram0 > /dev/null 2>&1 | 23:52 |
rennj | swapon /dev/zram0 -p 10 > /dev/null 2>&1 | 23:52 |
rennj | been using in vm for couple of months now seems to work | 23:53 |
buZz | zram frw :) | 23:53 |
rennj | embedded fu | 23:53 |
rennj | this vmware vm run in ram, so never really hits disk unless i fdisk/mkfs vmware.vmdk | 23:54 |
gnarface | nice, rennj | 23:54 |
rennj | 5 year old ssd 17TB of writes to 500MB sdd | 23:55 |
rennj | my win10 gaming vm is the pig | 23:56 |
buZz | another of my fav blockdev modules in kernel is 'bcache' | 23:56 |
buZz | just -amazing- what performance you get caching a LVM pool on spinning rust with a NVMe :D | 23:56 |
rennj | yeah looks nice, like zil/arc of zfs | 23:59 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!