libera/#devuan/ Thursday, 2021-02-18

danyspin97I have done some packages00:00
danyspin97on the old gitlab00:00
danyspin97so I don't know if they are still available there00:00
danyspin9766 needed like 6-7 packages that didn't exists in debian00:00
danyspin97also, the scripting language written by skarnet, which is called execline00:01
kiwi9how is tt going ... you started in D but I saw C++ on the github.   Rust has a lot of momentum.00:01
fsmithredexecline is in repo00:01
danyspin97kiwi9: I ended up writing C++ like it was rust, just uglier and with a lot of bloat00:02
danyspin97so I am seriously rewriting it in Rust, but I am really busy with uni these days so I haven't got time00:02
danyspin97tt had a supervision working and i could start and stop services :D00:03
kiwi9:-D ... you should focus on the Uni ... tt can be taken up as time and enthusiasm is available.00:04
danyspin97you're right there :D00:04
danyspin97gn guys :)00:05
fsmithred'night00:05
kiwi9night00:06
XenguyDoes anyone recall offhand what package needs to be reconfigured to display larger text on the console?01:22
plubno01:22
plubi'd like multi-console though01:23
plubsomeone should write that for me01:23
plub2 or 3 or 4 virtual framebuffers01:23
XenguyI remember in the old days the default font was a nice large size, whereas nowadays some genius has decided to make the default console font a perfect size for 'young eyes'01:23
Xenguyplub: Don't hijack the question01:23
plubsorry i'll try to find an answer01:23
XenguyI don't need to hear if you don't know the answer01:23
fsmithredXenguy, dpkg-reconfigure console-setup01:24
XenguyThank you01:24
plub/etc/default/console-setup01:24
systemdleteyesterday's upgrade resulted in the creation of 2 new unix groups, render and kvm.  afaik, I don't have virtualization installed other than vbox client software.   Vbox has been working here for ages, so this is a bit odd.01:26
systemdleterender, I understand, is related to graphics, so it seems some features supporting X or other graphics tools now require this "render" ownership.01:27
systemdleteIt is the "kvm" group that has me a bit worried.   Why would that group suddenly be required when (a) I don't use kvm for virtualization (except for vbox which has been here for ages), and (b)01:28
systemdleteI did not install any new software that I can think of.01:29
fsmithredsystemdlete, eudev created it. Did you get an error message, or you just saw that it got created?01:30
systemdletethe apt history.log file only shows one entry yesterday (the previous update was on 2/6/2021) so the addition of kvm group must have been created by one of the package updates yesterday.01:30
fsmithredyeah, eudev01:31
systemdletefsmithred:  my nightly logwatch report.01:31
fsmithredcame from beowulf-proposed-updates along with a few other things01:31
systemdleteAny idea why?01:31
fsmithredsee bug#548 for more info01:31
systemdleteI find it strange.01:31
systemdleteok01:31
systemdletedevuan bug 548?01:31
fsmithredwhich has not yet been answered by the dev01:31
fsmithredyeah01:31
systemdletehmmm.01:31
systemdleteI take it this would originate "upstream"?01:32
systemdlete(debian)01:32
fsmithredno01:32
fsmithredwell, I don't know01:32
systemdleteok... thanks  for this info.01:32
* systemdlete is grateful for logwatch...01:32
fsmithredit's not related to any changes in eudev that I'm aware of, and I was involved in reporting the problem and testing the fix.01:33
fsmithredbut there were probably some other changes, too01:33
fsmithred"the problem" is completely different from what we're talking about now01:33
numzobodd, in another distro i only eeded to add a kvm group when installing qemu01:34
systemdleteof course it is   ;)01:34
numzobneeded*01:34
systemdleteyes, exactly.01:34
systemdletevbox USES kvm for guests that are configured to do so, but it is not needed by the guest itself.01:34
systemdletevbox offers kvm and other virt methods01:35
systemdleteif it were needed by the guest, it would have been configured for kvm long ago01:35
systemdletefsmithred:  I have not attempted to reboot since the update yesterday... from the sounds of the report, I wonder if my system is still bootable?01:38
systemdleteI'm guessing it is, since my update did not fail.01:38
fsmithredwhat???01:38
fsmithredwhy would it not be bootable?01:38
systemdleteI guess I'd only need to worry if the update had failed as per the report (548)01:39
systemdleteand even then, that might not be an issue, idk.01:39
systemdleteMessing around with / modifying such basic and essential parts of the system is generally a bad idea.  GNU/Linux has been around for over 25 years now, so my feeling is that this portion of the system should be very stable by now.  Updates/changes/fixes rare.01:40
systemdlete(Oh, yes.  I know why it isn't... )01:41
systemdleteI guess some people think that the software base of GNU/Linux is some sort of self-amusement park, and don't realize that there might be actual serious applications in use.01:42
systemdletebut alas...01:42
systemdletehttps://linux.slashdot.org/story/21/02/13/0115253/did-linux-kill-commercial-unix01:45
systemdletejust by coincidence, this is a slashdot story that broke recently.  "If something's genuinely mission-critical and it's working, you leave it working."  I agree.01:46
systemdleteIf you want to play, there's a sandbox for that somewhere...01:46
ltsJust noticed some 404s in apt upgrade. How common is it that several mirrors fail sync? The one I use is on the list https://pkgmaster.devuan.org/mirror_status.html08:11
djphlts: if your timing is good, it's fairly common11:03
djph("good" as in "you hit them mid-update")11:03
ltsRight, thanks for confirming11:28
onefangMy network has been down most of the day, I wonder what you are talking about?  Sounds a bit package mirror related to me.11:29
ltsI was just wondering how normal this is https://pkgmaster.devuan.org/mirror_status.html11:32
onefangdjph , lts ^^^11:32
onefanghttps://sledjhamr.org/apt-panopticon/results/Report-web.html shows more info.  And some of the errors in the Updated column are coz I haven't updated how often some of the mirrors update.11:34
onefangI need to crack the whip and get some mirrors to fix a few things as well.  lol11:35
onefangSoooo, for the Updated column, a warning means "the upstream master pkgmaster has been updated, this mirror hasn't caught up yet, but is still within its update time".  A mirror not being fully up to date 10 minutes after pkgmaster updates, but it still has a 20 minute window to get updated.  Error means it's now outside of it's update window and still not updated.11:37
ltsThanks. Looks quite minor. The mirror I use still shows there as FAILED but now has started to work (in the last 30 minutes I think)11:39
ltsTo work = now has the newest updates11:39
* enyc meows11:39
l41f3nHello, I can see embedded section for ascii devuan but not for beowulf. Is it normal ?13:33
l41f3nI would like install beowulf on my rapi3 :)13:33
fsmithredl41f3n, there are community-produced images for beowulf here - https://arm-files.devuan.org/14:14
fsmithredand a forum section for arm discussion: https://dev1galaxy.org/viewforum.php?id=2414:15
l41f3nthank you fsmithred :)14:18
fsmithredalso #devuan-arm irc14:18
rwpfsmithred, TIL: Devuan BTS does not support severity important! "severity 548 important; Severity level `important' is not known. Recognized are: critical, grave, normal, minor, wishlist."18:06
rwpSeems odd they did not like that level and did not implement it in the BTS.18:07
fsmithredI'll mention it at our meeting today18:07
fsmithredthanks. I never noticed that because I never bother with severity. I think mine are all normal.18:08
rwpI also don't see "serious18:08
rwpI also don't see "serious" in the list either.  So that one was dropped too.18:08
fsmithreddebian has all those?18:08
rwpSo now I must decide, is it grave?18:09
rwpYes.  See this list: https://www.debian.org/Bugs/Developer#severities18:09
fsmithreddoesn't really matter. The maintainer is aware of it and it will be taken care of.18:09
rwpYes.  But for us out in the field using the system we must maintain a list in our head of how everything is slightly different here versus there.18:10
fsmithredthe dev who got the bts up and running was the one who gave the talk on minimalism at the conference18:10
rwpHahaha!18:10
rwpI have often used "important" when I thought something was significant to raise up out of the see of sometimes thousands of "normal" bugs.18:12
rwpOh well.  Anyway...  Just thought I would pass that tidbit along.18:13
fsmithredthanks18:14
luser977be sure to ask for severity levels "systemd_tainted" and "ridiculous". if the 1st is not accepted then the 2nd could be used instead.18:23
luser977<jk>18:23
fsmithredsystemd_tainted might actually be a useful tag18:25
masonfsmithred: Less typing: unclean18:25
fsmithredas they fuck up more packages causing bugs18:25
fsmithredok, that works18:26
* tuxd3v is observing gladly the systemd_tainted tag :)18:30
user__Updates on bluez-tools and bluetooth @beowulf: the bluez-tools are a huge huge set of C sources which mostly manipulate strings related to dbus. Iow, a Perl, bash (!), tcl, etc approach might have done the job as needed (in the right programmer's hand), using nothing (much) more than dbus-monitor and dbus-send, 2 cli tools supplied with dbus.19:05
user__Just as a simple example, bt-agent source in bluez-tools has a ~80loc C function to read and hash (!) the known devices pin file which may have 10 entries or so on a busy machine, and that hash table is used in a 300 loc C function to authenticate (bluetooth pairing) a new device. Would I rewrite this? No.19:06
user__Is anyone aware of a project accessing dbus from scripts only?19:07
fsmithredI used to have a script for shutdown/reboot that used dbus commands19:11
Kittyhi, who would I need to talk to about a package being added to backports?19:20
fsmithredwhat do you want to backport?19:21
Kittydropbear19:21
gnarfaceit is present in ceres19:22
KittyI'm currently using beowulf-backports19:22
fsmithredall of it?19:23
Kittywell I stuck it in apt config, and did a dist-upgrade19:23
Kittywhere doeas ceres fit ?19:23
fsmithredwow19:23
fsmithredceres=sid19:23
fsmithredchimaera=bullseye19:23
Kittysid == testing?19:24
fsmithredsid=unstable19:24
fsmithredchimaera is testing19:24
KittyI can never track the names, so always just think of stable, unstable, testing19:24
fsmithredsame version of dropbear in both19:24
Kittyand backports... which IIRC is stable+some stuff from testing19:24
fsmithredyeah19:24
gnarfacehmm19:24
gnarfaceit is also in beowulf already... why do you need the backports one now?19:24
fsmithred...that has not *all* been tested together19:25
Kittybecause version 2018.76 doesn't support ED25519 keys,19:25
Kittybut version 2020.79 or later does19:25
fsmithredhave you backported it yourself yet?19:26
Kittyhow would I do that?19:26
gnarfaceKitty: a lot of times debian backports patches for important stuff, but does not backport the version number.  has it been verified that the build in beowulf doesn't support ED25519 keys whatever they are, or is this assumption being made based on the version number alone, compared to that version number in other unrelated distros?19:27
gnarfaceKitty: i'm just mentioning it because it is common for people to go chasing their tail for no reason over such issues19:28
KittyI installed the dropbear that is in beowulf, and it doesn't support ed25519 keys.19:28
gnarfaceok19:28
gnarfacethe backport process might be simple19:28
plasma41Just got Internet connectivity back. What software are we talking about?19:28
Kittythe upstream project only added it as a feature in version 2020.79 in jun 202019:28
Kittyhttps://matt.ucc.asn.au/dropbear/CHANGES19:28
gnarfaceKitty: 1) you just check out the build deps for the beowulf version with "apt-get build-dep dropbear" then you get the source package from ceres with "apt-get -t ceres source dropbear" and then you build it with ... i think something like: cd ./dropbear[tab] && dpkg-buildpkg -us -uc19:29
gnarfaceKitty: the process is well documented at debian but they bury it in chapter 6 of the new maintainer's guide.  someone has a condensed summary of the process around here somewhere but i don't have the link19:30
Kittyeek, that looks complicated19:30
gnarfaceit is the best option all things considered19:31
Kittywhat is the name for the equivilant of testing?19:32
gnarfacenothing is really stopping you from just upgrading to ceres instead, other than the significantly higher cost in luck19:32
gnarfacechimera is testing19:32
plasma41Kit19:32
plasma41Kitty: testing is chimaera19:32
fsmithredoldie but goodie: http://forums.debian.net/viewtopic.php?f=16&t=3897619:33
gnarfacein theory the backport process from ceres->beowulf should work just as well for chimera->beowulf19:33
fsmithredceres and chimaera are mostly the same now, since sid is in early freeze19:33
KittyE: The repository 'http://deb.devuan.org/merged chimera Release' does not have a Release file.19:34
fsmithredtry again in a little while19:34
Kittylooks like it's not a simple case of swapping beowulf for chimera in the sources file19:34
fsmithredfew minutes maybe19:34
fsmithredalmost simple19:35
plasma41Kitty: Looks like the testing version of dropbear requires the testing release of libtommath-dev to compile. All other build dependencies are available in stable or stable-backports.19:35
fsmithredyou don't need -updates or -security for chimaera or ceres19:35
gnarfaceKitty: nothing will stop you from just installing the ceres or chimera version directly into beowulf either, but that's not advised because the library version mismatches might cause instability19:35
plasma41afk19:35
KittyI tried that, but it went into dependency hell19:35
gnarfaceah.  well, backporting is the way you get around that19:36
gnarfacein theory, anyway19:36
gnarfacesometimes you have to backport some dependencies too19:36
gnarfaceand it is possible to run into a situation where you end up recursively backporting basically the entire next release19:37
gnarface(in which case you might as well have just gone ahead and upgraded to it)19:37
Kittywhat am I doing wrong with my sources.list file... deb http://deb.devuan.org/merged chimera main19:38
gnarfacethat's the binary line19:38
gnarfaceduplicate it with "deb-src" at the beginning instead of just "deb" for source packages19:38
gnarfacesource packages and binary packages are on different source lines19:39
Kittyexcept when I do apt-get update19:39
KittyI get E: The repository 'http://deb.devuan.org/merged chimera Release' does not have a Release file.19:39
gnarfacelike fsmithred just said.  wait a few minutes, it is updating19:39
Kittyoh right, I hadn't understood that19:39
gnarfacedebian has over 150 mirrors... we have like 12 or 1319:39
Kittyhttp://deb.devuan.org/merged/dists/chimaera/Release - the file exists...19:41
gnarfacedeb.devuan.org is a dns round-robin19:42
gnarfaceyou don't necessarily see the same server every time19:42
Kittyoh19:42
gnarface"nslookup deb.devuan.org" will show you the whole list19:43
Kittybash: nslookup: command not found19:43
Kittyor maybe not19:43
gnarfaceheh, if furnished19:43
Kitty131.188.12.211 - it's doing this site19:43
Kittyah well, patience19:44
fsmithredit's up again19:44
Kittystill complaining :(19:45
fsmithredoh, sorry19:45
Kitty 404  Not Found [IP: 195.85.215.180 80]19:46
KittyE: The repository 'http://deb.devuan.org/merged chimera Release' does not have a Release file.19:46
gnarfaceseems to be working here19:46
KittyPEBKAC19:47
Kittychimera != chimaera19:47
gnarfaceah19:47
plasma41Kitty: I was just about to point that out, but you realized it before I could type it out. Feel free to personally blame me for choosing a weirdly spelled codename.19:50
KittyI have dyslexia...19:50
fsmithredwow, I just tried installing the 2020 version in beowulf live iso, and it would be a mess. (remove a bunch and upgrade libc6 among others)19:50
fsmithredbackport or upgrade would be much better19:51
masonOr a chroot, if there end up being extensive build deps that are problematic.19:53
plasma41Kitty: I also proposed the codename for the release after chimaera: daedalus. Apologies for all the weird vowels.19:55
gnarfacegreat, google accepts both spellings19:56
gnarfaceno way to tell which is more proper due to sponsored links as responses19:56
* gnarface facepalms19:56
plasma41fsmithred: Now you've got me curious. I'm going to try doing a local backport to see how hard it is.19:57
plasma41Well so far I can say that the build dependencies themselves don't spring any recursive dependency traps.20:09
Kittythank you all for your help20:12
KittyI'm logging off for the night. Thanks! am looking forward to using devuan some more20:13
plasma41Kitty: Thanks for stopping by!20:13
user__I think this debian / devuan code name game should be taken to the next level, like those iq logic tests, where the green house is left of the one that has no dogs and not the last one in the row? One could win prizes for guessing the right code names in a short amount of time.20:30
fsmithredours are going in alphabetical order20:31
user__It is almost there, just needs some inpartial arbiter in the form of an irc trivia bot.20:31
user__Also, iirc, on devuan.org, there used to be a table with code names in devuan, an debian, testing whatever, and numbers. Which are yet another facet. Is it still there?20:33
user__I don't see it20:33
user__https://www.devuan.org/os/releases AHHH20:34
user__Kitty should see this. Left?20:34
user__Kitty: ^ tab completion let me down, sorry20:34
user__So, alphabetic: J A B C C ... well Debian also does it. J S B B S20:36
user__(this is so terrible)20:36
fsmithredjessie=jessie20:36
gnarfaceit makes more sense if you know more about how debian actually does it20:36
user__openwrt got a clue for the release names from somewhere and double the letter. Attack Attitude, Barrier Breaker etc20:37
fsmithredceres=sid and never changes, is not part of the alphabetical order20:37
fsmithredthey can wrap around to Attack Barrier?20:38
user__hm?20:38
user__AA BB CC ...20:38
fsmithredthat's good for 26 versions in english20:38
fsmithredafter that...20:38
fsmithredAB AC AD20:38
user__Yes. By version 13 it will become sentient and rename itself I guess.20:38
fsmithredshit, I forgot about skynet20:39
fsmithredand we are way OT now20:39
user__Also version HH is close now. I wonder if they'll skip it.20:39
user__yes, let's stop this20:39
plasma41Kitty: AFAICT, if you download the dropbear source from testing as well as the libtommath binary packages from testing and the dephelper binary package from stable-backports, you should be able to compile dropbear on stable and install the resulting binary without triggering any recursive dependency traps.20:39
user__One question: why does Kitty need dropbear instead of sshd? Small device with no room for full sshd?20:40
gnarfacei remember when they came up with the naming convention for debian, the idea of running out of toy story characters before the distro collapsed in upon itself due to human arrogance was laughable.  but here we are...20:40
Kittyuser__: dropbear can be installed in the initramfs,20:40
Kittywhich means that you can use it on boot to unlock a LUKS petition20:40
fsmithredOH!20:41
gnarfacegod i hope you've tied the decryption keys to a server you physically own20:41
Kittyhttps://www.cyberciti.biz/security/how-to-unlock-luks-using-dropbear-ssh-keys-remotely-in-linux/20:41
gnarfaceif that's some aws instance you're just asking for trouble20:41
user__Kitty: nice idea20:41
KittyI don't do cloud20:41
gnarfacewell, i wouldn't do this at all, and i would advise anyone not to, but it's a clever idea if you care more about the ability to lock people out on purpose than whether you lock them out on accident20:42
user__Well, technically, initramfs is a small device. Sort of.20:42
Kittygnarface: ?20:43
user__Kitty: have you considered importing the static dropbear usually compiled for use with busybox?20:43
gnarfaceKitty: lemme put it this way.  you'd better have unencrypted backups somewhere20:43
Kittybackups are encrypted as well. but we have multiple users with keys to decrypt20:43
user__I see packages for dropbear in beowulf including dropbear-initramfs20:44
Kittyyes, but they are not recent enough to support ED25519 keys20:44
user__I am not familiar with those20:44
Kittyecdsa and ed25519 keys are eliptic curve encryption keys20:45
user__ah ok20:45
KittyI wish it also supported ed25519-sk keys20:45
user__Ok, djb is behind this, then it has to be good :)20:46
plasma41Kitty: If you need any help with the above provided recipe, just let me know. I'd be happy to help.20:46
user__Kitty: you can always compile dropbear from scratch. Someone already have you the "debian way".20:47
user__Above.20:47
Kittyyeah, I was trying to avoid having to go the from scratch route...20:47
user__Well, debian and bleeding edge packages -- does that even fit in one line of text?20:48
user__2018.76 to 2020.81 is "just" 5 versions behind.20:48
Kittyed25519 was added in June last year20:49
user__Interesting he had .deb packages up till 2004. On the dropbear site.20:49
plasma41Kitty: It only took me about fifteen minutes to do. Granted, I frequently do quick 'n dirty backports for my own personally use.20:50
user__Sorry, till 200920:50
user__So .80 would work?20:51
KittyI solved it by dist-upgrading to chimaera20:51
user__Is there an automated way to debianize an arch package or such? Almost sure they'd be on the latest.20:51
user__Oh ok.20:51
plasma41Kitty: That's one way to do it20:51
user__What happens if one simply takes the .deb from the version one needs and dpkg installs it? Assuming no big changes?20:52
* user__ did this in the past quite a few times.20:52
Kittyright, at this point, I'm going to stop playing with my laptop, and go watch NASA lower a rocket pupping semi autonomous nuclear powered robot, onto mars from a rocket powered sky crane20:53
Kittys/pupping/pooping/20:53
user__:)20:53
plasma41user__: The close thing I can think of would be Bedrock Linux <https://bedrocklinux.org/>, but that's a whole other can of worms.20:53
lkjasdfyou switch to Klingon20:53
user__plasma41: nice link, thanks.20:54
Guest0does devuan provide screensavers or do they come with xfce?21:25
Guest0i remember using debian stable and not having any screensavers with xfce21:25
plasma41Guest0: Try the xscreensavers packages21:29
gnarfaceGuest0: devuan and debian aren't different in this regard.  the screensavers are in the "xscreensaver*" packages21:29
Guest0ah ok so i guess my install must have come with the xscreensaver package21:29
ltsxfce4-goodies maybe too?21:29
gnarfacethere are some non-stock optional ones that usually aren't included because of video driver stability issues;  "apt-cache search ^xscreensaver"21:30
Guest0alr21:30
gnarfacexlockmore still exists somewhere i think, but it was pulled from debian several releases ago, and it has all the same screensavers as xscreensaver anyway21:31
ltsNope, debian/devuan doesn't seem to have xfce4-screensaver in xfce4-goodies. Arch does, I remember using that. Sorry for misinformation21:31
gnarfacea couple window managers have static lockscreens of their own built-in, superseding the xscreensaver functionality as far as their developers are concerned21:31
gnarface(personally i disagree but you're talking about people who get free monitors from their jobs and have never had to make one last more than 2 years)21:32
lkjasdfhttps://www.jwz.org/blog/2021/01/i-told-you-so-2021-edition/21:33
gnarfaceif you're using nvidia drivers i strongly recommend to avoid the xscreensaver-gl and xscreensaver-gl-extra packages21:33
gnarface(hard i/o locks, going back decades.... )21:34
Guest0hahaha yeah21:34
lkjasdf"If you are not running XScreenSaver on Linux, then it is safe to assume that your screen does not lock." --- the Xscreensaver author, highly biased, but with legitimate reasons and evidence21:35
* rrq xautolock -time 5 -detectsleep -locker /usr/bin/xtrlock -b -corners 00-+ -cornerdelay 10 -notify 10 -notifier /usr/bin/play -v 0.1 /home/ralph/Sounds/860_1245242569.mp321:36
lts"physlock -l && <any fancy screensaver you want> && physlock -L" is nice. Failure of screensaver will still lock you to a dedicated empty TTY.21:36
gnarfaceoh yea, to be clear, xscreensaver is not actual security software.  everyone should know that already.21:36
gnarfacei merely advocate it for display wear-leveling21:37
gnarfacethe problem with any screensaver software is if you kill it somehow, you've unlocked the screen then21:38
user__First time I read about physlock21:44
user__lts: does physlock package install an /etc/init.d/ script?21:45
user__no21:46
lts..and I remembered wrong. That command string was for removing ability to change tty during screensaver. What I meant was "<your fancy screensaver command> || physlock"21:46
ltsYou don't run it from a service though21:46
user__what's the command again for cli auto timeout / logout ?21:46
user__lts: it has a -d option21:46
ltsHmm, never used that21:47
masongnarface: xscreensaver breaks on my workstation lately21:47
gnarfacemason: i assume it's just certain screensavers21:47
gnarfacemason: try stuff like space and rain21:48
gnarfacethe old pixel-based ones are the ones that suffer from the least video driver compatibility issues21:49
masongnarface: boxfit in this case - might be some interaction with nVidia, since there isn't a similar flakiness on my laptop, which has Intel graphics.21:49
masongnarface: I'll try something different. I've not been able to isolate a trigger, but the process just goes away.21:49
user__<ot> is that Mars landing on some live yt channel?22:03
user__yes, found22:03
masonuser__: NASA TV anyway, possibly YouTube as well - https://www.nasa.gov/multimedia/nasatv/#public22:04
masonuser__: Note that #devuan-offtopic exists.22:04

Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!