libera/#devuan/ Friday, 2020-05-01

Garb0Anyone on beowulf having issues with pulseaudio?03:25
Garb0audio for tor browser or firefox in particular03:26
fsmithredGarb0, yes. I don't use pulseaudio. Firefox-esr does not need it, but tor-browser does.03:35
fsmithredI can use apulse with tor-browser03:36
Garb0There's the thing, pulseaudio is well set up, waterfox works fine with it but not tor browser.03:36
Garb0apulse not working with TB either.03:37
fsmithredI added apulse to the command in the tor-browser launcher03:37
Garb0Beowulf really did break a buncha stuff didnt it04:01
jungleprimesGarb0 - What are you observing?04:04
Garb0You got a laptop?04:05
Garb0run redshift, wait for a few hours, switch your TTY then go back to Xorg04:05
Garb0it will freeze04:06
jungleprimesI currently use Devuan as a server, not as a workstation/GUI build. Although I have a GUI VM of devuan beowulf and a GUI VM of devuan jessie for testing purposes only.04:09
jungleprimesNot on laptop04:09
Garb0I had to upgrade to beowulf because some of my software only runs on >2.28glib04:10
Garb0otherwise ASCII was one solid son of a gun, i run everything in it, literally everything, Xorg, browsing, sshd and webserver, steam and wine, quite literally ran everything like a champ.04:11
Garb0all in the same setup04:11
jungleprimesnice04:11
Garb0Anyway, i will try reporting bugs when i can confidently reproduce them, for now i guess i'll suffer04:14
jungleprimesGarb0 - Have you considered keeping ASCII as primary and just booting into Beowulf for this specific application you are talking about? Or run Beowulf inside of ASCII as a vm?04:28
Garb0it's waterfox04:28
Garb0i'm not VMing just for a browser04:29
meep_____» [18:25:55] <Garb0> Anyone on beowulf having issues with pulseaudio?04:30
meep_____Always having issues with pulseaudio04:30
furrywolfI stopped having issues with pulseaudio when I started apt-get purging it by default.04:30
djph^04:31
ulletanybody have a life-migration list of tips05:15
ulletlive05:15
ulletfrom debian to devuan without re-imaging05:15
golinuxhttps://beta.devuan.org/os/documentation/dev1fanboy/en/buster-to-beowulf05:19
meep_____Mentos to cheetos06:14
meep_____Basketballs to broke06:14
Garb0should i just say fuck it and upgrade to ceres07:04
ulletit depends on which package mix will cause you least work and problems07:04
ulletand that mix is different for each administrator07:04
Garb0maybe it will unscrew the paudio issue with tor07:05
Garb0guess i got nothing to lose07:05
ulletif you do a lot of work with building recent projects from git, i think ceres makes sense07:05
ulletwhy do you use pa Garb0 ?07:05
Garb0tor and ff usually doesn't work without it, mumble too i believe07:06
ulletff works fine with alsa here07:06
Garb0quantum?07:06
Garb0you use apulse?07:07
ulletno07:07
Garb0esr?07:07
ulletyes07:07
ullet68.4.1esr-1~deb10u107:07
ulletis that a bad one?07:07
ulletfirefox dropped alsa for a while but whatever debuan gives me works fine with alsa07:08
Garb0not me though, hm.07:09
Garb0ESR does work with alsa i think07:09
Garb0but the tor bundle and quantum don't, regardless of distro07:09
ulletah07:10
ulletwell i don't view media in browser anyway07:10
meep_____When will ascii become EOL?07:24
blebif you down an interface07:44
bleblike ip link set wlan0 down07:44
blebdoes wicd put it back up?07:44
ullettry it?07:45
blebwell07:48
blebsomething puts it back up07:48
ulletidk07:49
jungleprimesyou could use a syscall tracer tool, so you can identify which pid made the request07:49
jungleprimesor try using a different network connection, shutting down wicd, and see if it still happens07:50
jungleprimesI'm not very familiar with wicd tho07:51
ulletanybody install qtcreator lately without dependency problems ?08:33
ulletbeowulf here08:34
Garb0_alright, terminus font is broken in ceres08:52
ulletuse unscii font!09:00
ullethttp://pelulamu.net/unscii/09:00
nemoso... something keeps deleting /tmp in my devuan ascii at work.  (with no reboot) causing my tmux socket and other things to be lost16:02
nemoI was wondering how devuan configured /tmp and if anyone had any ideas for what might be at fault16:03
masonnemo: You could use inotify to watch what's going on in there.16:03
gnarfacenemo: should have been chosen by you at install time.  what does the output of "mount" say?16:03
masongnarface: He's not talking about clearing at boot/reboot, though.16:04
nemoO_o /tmp is not a mount16:04
nemoI thought for sure I'd see tmpfs16:05
nemowhaaaaaat16:05
nemoI never even bothered to review that when installing16:05
gnarfacethat's what i was looking for, tmpfs.  but this is also an important revelation.   it means your /tmp directory is in your / partition16:05
nemoyep16:05
nemoshocker to me too16:05
gnarfacestill don't know why it's being purged, but it might be important to note the amount of free space on / when that happens16:06
nemognarface: seems to happen on all the devuans regardless of available space16:06
nemobut the one that is annoying me is definitely low on space16:06
gnarfacehmmm. so it's gotta be something you installed16:06
gnarfaceyea maybe inotify is the key16:06
nemognarface: well. could perhaps be something the network admins are running. they insist on an admin account with ssh access16:07
nemohmmm maybe it purges /tmp if it is short on space16:07
gnarfacepossible16:07
nemoI should see if I can move my tmux to $HOME16:07
gnarfacewell16:07
gnarfaceone thing is that the policy for /tmp is supposed to be such that random purges are safe16:07
gnarfacenot always true, but that's *supposed* to be the case16:08
nemognarface: yeah, but a lot of stuff relies on it for long-running sockets16:08
nemotmux for example16:08
nemoprobably based on assumption that purges are reboot linked16:08
gnarfacebut, there is /var/tmp, which has a different policy about persistence16:08
nemokeychain is another that bugs me16:08
gnarfaceso, in theory you are supposed to use /var/tmp instead when someone administratively above you is doing this to you16:08
nemognarface: well then the debian tmux package needs a different config. lemme see if I can move the file16:08
nemook16:08
gnarfacenow, if this starts happening to you in /var/tmp/ too, you either have a rogue process or clueless admins16:09
nemohttps://unix.stackexchange.com/questions/496454/how-to-attach-to-a-tmux-session-after-deleting-its-session-file  hm. welp. here's the tmux fix at least16:10
nemowill remember that one.16:10
gnarfaceeven the change to purging /tmp on reboot is relatively recent in the scope of my experience.  the policy about it has never changed, but for a long time it was also not enforced16:11
gnarfacesomeone must have noticed that openbsd was doing it16:12
buZzif only we had a tmpfs for that16:12
buZz:D16:12
buZzoh we do, since 200116:13
nemobuZz: heh. I used the devuan install defaults.  I *assumed* those would include a tmpfs for /tmp by default16:18
nemoI mean, I can set one up now, but...16:18
nemo(also doesn't fix the base problem there, gotta see how to move keychain and tmux to /var/tmp without having to build my own package)16:18
nemohmmmm16:19
nemoI could hardlink them in /var/tmp perhaps. amusingly because devuan does not have tmpfs mounts on /tmp and /var/tmp that would actually work 😃16:19
nemothen at least I could recover them easily when they are destroyed16:19
buZz:)16:20
gnarfacesomething unnerves me about that approach, but i can't think of any technical reason why it wouldn't work16:20
buZzactually, my devuan desktop does use tmpfs for /tmp16:20
buZzquite sure i didnt set that up myself, but could be wrong16:21
gnarfacenemo: if you let it do automatic partitioning, the logic may do different things based on the size of the disk16:21
gnarfacenemo: i've literally never been satisfied with automatic partitioning.  i haven't done a non-expert install since the very first one.16:21
nemognarface: I don't think I chose automatic partitioning, but hard to remember.16:35
nemoit might have been16:35
nemognarface: I'm a little confused as to why partitioning would have anything to do with the mount points16:35
nemobuZz: my devuan workstation here does not seem to have tmpfs on /tmp either16:35
nemognarface: ok. yes, I can see a relation for the base mount for each partition, but not so much for this tmpfs thing ☺16:36
nemohuh. interesting... I just checked an oooold gentoo system and an oooold ubuntu and neither had /tmp tmpfs either16:40
nemoso maybe this is standard behaviour for a long time16:41
gnarfacenemo: sorry, did not mean to to confuse you; yes, the manual deleting of the contents of /tmp as a policy in the distro has been a feature of the debian distro for a few releases before devuan forked, that's is a separate thing from the tmpfs issue16:42
gnarfaceeither one could have been a potential culprit16:42
gnarfacesometimes when i'm helping people debug stuff, i have them check tangentially relevant possibilities in case they forgot something (changing /tmp to tmpfs is feasibly something someone might do for performance and then forget 3 years later)16:43
gnarfaceand no, it still shouldn't have been an issue unless you were facing some sort of additional hardware or software failure, but when someone is having a problem nobody else has ever had, these are the types of possibilities we need to have on the checklist to be thorough16:44
nemognarface: yeah. uptime for this devuan ascii has been 50 days now. but I seem to lose /tmp almost every day16:44
gnarfaceheh, yea maybe check the system for cron jobs that run around that time every day...16:45
gnarfacehopefully you're not facing some ext4 filesystem corruption, but that's also a theoretical possibility here16:46
gnarfacei think it may have been in the jessie->ascii transition where if you use the older e2fsprogs on the newer ext4 version for fsck or resizing at any point, it causes a silent corruption that slowly unravels the whole filesystem16:47
nemohmmm16:56
nemothat would be bad16:56
nemohave not done any resizing16:56
nemooh. also this was a clean ascii install16:56
gnarfaceyea, probably not the issue then16:57
masonnemo: Had meetings, but... Why not just run inotify and find out unambiguously what's clearing your /tmp ?17:00
ham5urgI just installed unattended-upgrades and took a look at vim /etc/apt/apt.conf.d/50unattended-upgrades and found "origin=Debian,codename=${distro_codename},label=Debian";17:16
ham5urgIs this correct?17:16
ham5urgI found this: https://bugs.devuan.org/db/78/78.html17:18
ham5urglooks like it ahs to be edited.17:18
nemomason: heh. meetings here too.17:30
nemomason: sounds like a good idea actually. I've just never done inotify from commandline, so have to look up how the utility works17:31
nemo*DDGs for a guide*17:31
nemolooks like inotify-utils inotifywatch/wait17:33
nemohm17:34
masonHrm, it's been a while. Maybe inotify isn't idea. Not seeing how to wring a process ID out of it at a quick glance.17:34
nemoyeah17:34
masonSo, perhaps auditd then.17:35
nemomason: also was rather surprised.  did inotifywaych -r /tmp/test  and then rm /tmp/test/foo and rmdir /tmp/test in another window17:35
masonThat'll show processes. It's heavier-weight to configure and the logs are verbose.17:35
nemoand nothing showed up in the watch17:35
nemowhich is kinda wtf17:35
gnarfacei suddenly thought of the possibility this could be caused by a poor edit to the php garbage collection (debian uses a cron script for it, for security purposes, in defiance of upstream)17:35
nemoah. only reports on abort17:35
nemomaybe works on a SIG too17:35
masoninotifywait -m /tmp is probably what I'd say17:36
nemognarface: oooooooh17:36
masonor inotifywait -rm /tmp17:36
nemognarface: these problems are recent and on machines where php is installed17:36
nemoer. wait. you're talking about my problem?17:36
gnarfacenemo: you got php on there?  yes. i'm talking about your problem.  it just occurred to me that php session storage is in /tmp17:36
nemognarface: where's this garbage collection?17:37
gnarfacelike i said, a cron script17:37
nemognarface: yeah, it's my dev server so I have a ton of stuff17:37
nemognarface: tomcat, weblogic, apache with mod_php, X...17:37
nemognarface: ok. fair 'nuff. was hoping you could help narrow it down a bit17:38
gnarfaceif you used the the php packages in the repo, it should have put a cron script somewhere that parses the php.ini manually for the session path.  it would be easy to make a mistake and have it deleting all of /tmp17:38
ham5urgI changed /etc/apt/apt.conf.d/50unattended-upgrades to17:38
nemoI will now search all the locations where crons are ☺17:38
ham5urg        "origin=Devuan,codename=${distro_codename},label=Devuan";17:38
ham5urg        "origin=Devuan,codename=${distro_codename},label=Devuan-Security";17:38
nemognarface: was absolutely the repo php yes17:38
ham5urga dry run looks good but at this moment the system is up to date. So no real test.17:38
masonnemo: here: https://bpaste.net/OA2Q17:45
nemothanks17:47
nemolooks like something that could get big fast17:47
nemomason: those rules seem to be global?17:47
nemodon't see any mention of /tmp17:47
masonnemo: You can add -F dir=/tmp17:48
masonhalf a sec17:48
nemomason: ok. thanks. actually to cut down on noise I'll probably just do the tmux socket17:48
nemo'cause I'm sure to forget this was setup17:48
masonnemo: https://bpaste.net/XJXQ17:48
masonnemo: you'll remember when you run out of disk space. =cough=17:49
nemolol17:49
nemomason: that won't be long since this vm is at 99%17:49
nemostill haven't figured out what I can delete to free it up ☹17:49
nemoooooh17:49
nemoapt cache!17:49
nemo2.7 gigs17:50
nemoa start17:50
nemo93% yay17:50
masonargh, service auditd restart17:52
masonsystemctl warps my memory17:52
masonbut otherwise the paste ought to work17:52
nemoheh17:53
nemomason: don't worry, I'm not in the habit of copying and pasting things into shell anyway17:53
nemothey were all done piecemeal17:53
nemobut yes, I noticed that ☺17:53
nemothanks though17:53
nemonow to wait for that rogue process to step into the trap. bwahahaha17:53
masonheh17:53
masonauditd is useful17:54
openbsdtai123hi, devuan ascii amd64 x64 stable cannot format mkfs.extXX anything. it just hangs17:54
openbsdtai123I have tried with slackware stable, current 4.4.12 it just works fine (it is a x86_64). Same for openbsd, it can format fine with live openbsd usb stick (mkfs.ext2). So it comes from devuan.17:55
buZzsounds like a hardware issue18:01
buZzbut not sure anyone should even be still on ascii :P18:02
buZzdid you try mkfs.extX a loopback filesystem yet?18:02
masonopenbsdtai123: I don't have an ASCII machine to test, but it seems to work fine on Beowulf. I'd be surprised if it broke on ASCII, though. Please try this:18:02
buZzdd if=/dev/zero of=./this.fs bs=1M count=1024 ; mkfs.ext2 ./this.fs18:02
buZzdoes that hang too?18:03
masonOh, heh.18:03
buZz:)18:03
masonbuZz: https://bpaste.net/REXQ18:03
masonwas just getting that ready18:03
masonyours was faster18:03
buZzheh, yours is more verbose :)18:03
buZzif loopback fs' work but real disks dont, checkout dmesg and smartlogs18:04
masonopenbsdtai123: try this: https://bpaste.net/REXQ18:04
masonopenbsdtai123: or buZz's version, either way18:04
openbsdtai123(I put first slackware in sda1, with kde, and after sda2 with ascii x64, thank you !!)18:05
buZzthats not meaningful or helpful :P18:06
buZzif you were able to install devuan at all, that means it did run mkfs at some point :)18:06
openbsdtai123I used CP/M, Unix at bit, started Linux with slackware 2. maybe I should know what I say...18:08
masonI often hope for that, but whether I end up spewing confusion or not is too random for my comfort.18:10
buZzso i guess that mkfs issue just magically went away? :D19:02
masonbuZz: You have to understand what's going on under the hood. Then it works.19:20
buZz:)19:29
buZzi'm excited about popcorn linux coming to mainline kernel19:30
ulletpopcorn linux?19:39
ulletapulse i don't find in repository20:15
ulletdid it get killed?20:15
ulletoh there it is. i was in wrong window20:15
ulletdebian removed it though hah ah hah20:15
radHello everyone. Do the mouse speed and acceleration settings work for you? For me both sliders do nothing. I tried in a fresh live-cd installation with XFCE and then on a netinst installation with MATE.21:52
radI noticed that XFCE allowed to configure the sensitivity on both my mouse and my touchpad (I'm on a Dell laptop) while MATE doesn't show which device it is configuring.21:53
gnarfacerad: maybe try a different mouse driver21:55
radPossibly related: https://www.reddit.com/r/debian/comments/6mdtdo/mouse_preferences_dont_work_on_debian_9_mate/ I'll try installing that xserver-xorg-input-evdev package.21:55
radIs xserver-xorg-input-evdev a mouse driver or are you talking about something different?21:56
ulletiduno21:56
radI'm going for a quick restart, brb. :-)21:56
ulletxfce has it's own channel and problems21:56
gnarfacewell, i meant xserver-xorg-input-synaptics but xserver-xorg-input-evdev is worth a try too21:57
gnarfacealso make sure you have xserver-xorg-input-libinput21:57
gnarfaceyou should have all 3 of those already actually21:57
gnarfacebut you might have to put a custom xorg.conf snippet in place to select one by default21:57
furrywolfI had to create an xorg.conf to make mouse settings work at all.21:59
radI only had libinput. I just installed the other two, I remember having to install synaptics on Linux Mint in the past otherwise the touchpad worked but was extremely jittery and was jumping around.21:59
radBe right back.21:59
furrywolfhttps://paste.ubuntu.com/p/f9K8Gky9NF/   (do we have our own pastebin somewhere?)22:01
gnarfacei've been using paste.debian.net22:01
furrywolfI need to get ready for work...  if they come back and end up needing an xorg.conf, they can steal mine.  bbl.22:04
gnarfacerad: furrywolf left this example snippet for you while you were gone: https://paste.ubuntu.com/p/f9K8Gky9NF/22:09
radAh thank you!22:09
radIs it normal that this directory doesn't even exist for me /etc/X11/xorg.conf.d/ ?22:11
ulletit was introduced quite a few years ago22:12
gnarfacedoes seem weird, but i wouldn't worry much22:12
gnarfacecreate it as root with default permissions22:12
gnarfacemaybe it is missing because you did a livecd install, i'm not sure22:13
rad_No this is a netinst installation with MATE. The live-cd was the previous installation and it was with XFCE and had the same issue.22:15
gnarfaceit shows up in the x11-common package here, but this is a upgraded system22:16
gnarface(upgraded from debian to ceres)22:17
rad_https://paste.ubuntu.com/p/ck2KWwfKCn/ I think those config files for X11 exist in /usr/share, not in /etc22:19
radSolved! Here's what I did. Went in /usr/share/X11/xorg.conf.d as su and then I did mv 10-evdev.conf 75-evdev.conf22:32
radThese files are parsed alphabetically and the last driver to claim some device will be the one handling it. So moving it to 75 will cause evdev to handle my mouse etc. Now the acceleration for my mouse works but the sensitivity slider doesn't do anything. Still, with acceleration at minimum it's exactly how I want it.22:33
radThanks for the help. :-)22:34
gnarfacerad: to be clear, Xorg will check both locations, the /etc/ one and the /usr/share one.  you should put your edits in the /etc one because the /usr/share one is supposed to be overwritten by the package manager on updates22:34
gnarface(whether it will or not after you've made edits i'm not sure, but either situation could be a disaster)22:35
radoh I see. ok I will copy the file to etc.22:38
user282069have there been many problems with X going from ASCII  to Beowulf ? seems like the change from root owning it is big22:40
gnarfaceuser282069: depends on your hardware.  not all drivers are affected equally.  the more common tripping point is the permissions backends22:42
gnarfaceuser282069: (check the https://files.devuan.org/devuan_beowulf/Release_notes.txt for details)22:43
user282069thank you. ati here. I'll read on22:44
user282069slim seems to always come through22:45
radHmmmm no 32bit wine or wine32 on devuan repos because devuan is 64 bit only?23:06
debdograd: you prolly need to add multiarch i38623:16
openbsdtai123the wine 32 works well on a kernel x86 32bits. there is a kernel and modules and methods: http://openbsdtai123.shell.ircnow.org/devuan.html    I use 32bits usually with wine devuan for oldies, old apps. On amd64 x64 kernel, wine32 does not work.23:16
debdoghttps://wiki.debian.org/Multiarch/HOWTO#Usage23:16
radHow was I able to run 32bit windows apps on 64bit Linux Mint? I remember addint the i386 arch like debdog says, and then installing something like wine:i386...23:18
openbsdtai123sounds good... nice. I used 4 partitions, wiht 32bits and 64bits for wine.23:19
gnarfacerad: ignore openbsdtai123 he's wrong23:19
gnarfacerad: you just need to enable multiarch and remember to re-run "apt-get update" first23:19
raddpkg --print-foreign-architectures gives me i38623:20
radso I think I added it just fine.23:20
gnarfaceshould be working then.  the 32-bit packages have the suffix :i38623:20
gnarface(you still won't get any of them by default unless you install something that requires them like steam or wine)23:20
radI'm sure I did update before, from within synaptic, now I did it from the console and it worked!23:21
gnarfacei don't know shit about synaptic other than i could never make it work either23:21
gnarface(i didn't try very hard though; i prefer doing this from a terminal)23:21
gnarfacethere might be some synapic config panel you have to explicitly tell it to not filter certain packages out in23:27
gnarfacei seem to remember it hiding a lot of dependencies and meta-packages by default23:28
gnarfaceso maybe other arches too23:28
meep_____hello?23:58
gnarfacewe see you meep_____23:58
meep_____good23:58
meep_____testing out mcabber23:58

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