libera/#devuan/ Friday, 2019-08-23

klaushi all01:11
scutihi, can i ask a quick question? i have to leave in about 20 minutes03:08
scutii cant get wicd-gtk to display wireless networks; i can get wicd-curses to display wireless networks, but none appear on wicd-gtk03:09
scutiabout the last months i've used this devuan install, i had to use iw to get wireless on my laptop03:10
poontangmessiahscuti:03:10
poontangmessiahsudo apt purge wicd03:10
poontangmessiahsudo apt install network-manager03:10
scutiso, just dont use wicd?03:11
poontangmessiahthis is the only advice i can give because wicd is truly trash on every distro i used03:11
scutii suppose since i'm having to much issues with it currently03:11
scutionly used it since it came with devuan's dvd install i recall03:12
scutiok thanks this is less hassle than wicd03:20
gnarfacepoontangmessiah: at least make sure they're updated first.03:42
gnarfacepoontangmessiah: known bugs have been fixed, and it's not like network-manager doesn't also suck.03:43
golinuxnetwork-manager sucks more03:44
poontangmessiahgnarface golinux explain why03:45
poontangmessiahi've never had netman screw up on me, not even once03:45
poontangmessiahexcept maybe once but that was my router's fault anyway03:45
gnarfacei used to be a redhat user from back when they first introduced it.  i think you just don't have enough experience to fill the chart with enough data points to draw accurate lines.03:51
gnarfacemost the issues users have had with wicd have been self-inflicted03:52
gnarfacethat said, i'd advise users to learn to manipulate the /etc/network/interfaces file directly instead of either03:52
gnarfacei mean, if you're going to take a stance based on reliability, take the right stance at least03:53
onefangThe only issue I have with wicd is that it'll reorder lines in it's own config file each time I reboot, which wouldn't be a problem except I also use etckeeper.03:54
gnarfacebut it's really important to make sure people have at least got the latest updates and try to diagnose the problem if it's unknown03:54
gnarfacebecause now we don't actually know if there's a new issue with wicd or not03:54
gnarfaceand it's something that enough people have used successfully that it was very problably either self-inflicted or a really weird corner-case03:55
onefangThough scuti wanted a quick answer, coz they had to leave in 20 minutes.  Scuti got a quick answer that satisfied them.  shrugs04:24
gnarfacewell, i was thinking maybe just user error, like not turning the wifi on first, or maybe there was actually a bug but one that had been fixed in the repos, and he just hadn't updated yet04:31
gnarfaceit rang a bell, like something that other users had gotten confused about04:31
klauswhat would be the equivalent of /etc/systemd/logind.conf on a systemd free system ?06:20
klausi used to use that file to modifiy behavior of my power key: HandlePowerKey=ignore06:20
klausin case of me pressing it by mistake so the system wouldn't go to immediate shutdown without warning..06:21
gnarfaceklaus: should be handled by acpi06:22
gnarfaceit feeds off bios defaults though, so you might be getting this behavior even if it's not installed06:23
gnarfaceand it's typical that this is a feature you can change in the bios too06:24
gnarface(some really old machines use apm for this)06:24
klausi don't have bios access on this machine, it's a macbook06:24
gnarfaceoh, well that makes it trickier06:25
klausyes, mac are a pain about that06:25
klauson a systemd system i can use that systemd file logind.conf and define: HandlePowerKey=ignore06:26
klausi'm searching equivalent .. but nothing pops up as evident substitue06:26
klaus*substitute06:26
gnarfacedid you look in /etc/acpi or /usr/share/doc/acpi ?06:30
klausnot yet, i do now06:30
klaussomeone forgot to delete a line in /etc/acpi/handler.sh , last line i see is:06:32
klaus# vim:set ts=4 sw=4 ft=sh et:06:32
klausthis /usr/share/doc/acpi path doesn't exists, the closer i have is /usr/share/doc/acpid/  and there's nothing in the beside changelog, readme, todo06:35
klaus /etc/acpi might have something interesting though06:35
klausnot sure yet i'm reading06:36
klaushum, no, it's just log input lines06:37
klausoh hey :)06:38
klaus--> /etc/elogind/logind.conf06:38
klaus:)06:38
gnarfaceoh, the gui login manager could take over that maybe06:39
gnarfacei had forgotten06:39
klausi don't use one06:39
klausi removed the display manager06:39
gnarfacethen that might not help you06:39
klausi log to tty06:39
klaushum06:39
gnarfacebut the readme in /usr/share/doc/acpid might actually be useful06:39
gnarfacehere, i have a /etc/acpi/powerbtn-acpi-support.sh though, not sure why you don't.06:40
gnarfacethis could be new or old cruft06:40
klausturns out trying to restart elogind was a bad idea06:44
klausbut otherwise it worked06:45
klauspressing my power key doesn't immediately shudown the system without warning anymore06:45
gnarfacewell that's good06:47
gnarfaceif you don't have a graphical login, you don't need elogind06:47
gnarfaceyou could just uninstall it06:47
gnarfacethough then you might get new and exciting behavior06:47
klausbut that's the one thing that change the behavior of my powerkey, if i remove it i'll get back to that behavior of immediate shutdown if i press it by mistake ..06:49
gnarfacewell, i'm not convinced that this couldn't be done with acpi, too06:50
klausi'll try out things a bit, if i can remove it and find a way to prevent that behavior then i'll remove it, but meanwhile i find an alternate way i'll just keep it06:50
gnarfaceyou might need some other package too like acpi-support-base or something06:50
klausi'll try it out, thanks for the help !06:51
gnarfaceno problem06:51
gnarfacewhat's important is that it works how you want it to06:52
klausdevuan is a blessing, i should have gone that way sooner06:54
klausi like artix too, it's arch without systemd06:55
FlibberTGibbetWindows Subsystem for Devuan still behaving very well on my work Windows laptop :D12:35
poontangmessiah_gnarface: it's been some time since i've used devuan, does startup ifup still hang when the ethernet cable is unplugged?15:30
Walexpoontangmessiah: there are various tools rthat detect cable presence and run 'ifup' or 'ifdown' depending.17:10
poontangmessiahWalex: that doesn't answer the question17:10
Bureknew devuan when?17:57
jonadabBurek: As soon as Perl6 reaches the point where we can drop Perl5 support.18:02
Burekkk, thanks18:03
jonadabHTH.HAND.18:03
golinuxjonadab:  Where did you get that?  If we had the (wo)man power to do the remaining work it could be out next week.  But alas . . .18:24
golinuxBut we want to do an ascii point release before beowulf.18:24
r3bootjonadab: :D :D :D18:25
golinuxMaybe your response was sarcasm.  Dunno . .18:25
r3bootgolinux: it's sarcasm; Perl6, eventhough ppl have been writing for it for eons, is still not released18:26
r3booterr, *has not replaced perl5 yet18:27
Burekwow i got bamboozled18:30
djphBurek: eh?18:44
* golinux knows very little about perl except for Netanel Rubin�s Perl Jam 2 talk at CCC a few years ago.19:31
golinuxhttps://media.ccc.de/v/32c3-7130-the_perl_jam_219:35
koollmanI intend to try installing devuan on a rather modern laptop (t495). what's the current 'best' installer, refracta from a liveusb ?19:35
golinuxIt was hysterical!19:35
golinuxDepends what you want to have installed19:36
koollmanhm. what are my options ? :)19:36
koollmanI suppose I want devuan ascii, with a few things from backport. and I want to have at least Xorg running19:37
fsmithredlive installer gives you what is in the iso - full xfce desktop19:38
fsmithredand takes 10 minutes19:38
fsmithredinstaller isos give you more software choices and take a lot longer to install (usually 1-2 hours for me)19:38
koollmanI'm fairly sure installer iso won't work on this laptop. Although I could try and open bugs :)19:39
fsmithredshould say full desktop for the desktop-live19:39
fsmithredif you think the hardware is too new, live iso probably won't help.19:39
fsmithredwhy do you think it won't install?19:39
koollmandetection process. as I remember it, devuan installer is rather basic. I remember it getting stuck in various situations when there are no cdrom and it cannot find the usb key19:40
koollmanI mean, I can try it right now, should be fast :)19:41
fsmithredoh, ryzen. You might need a newer kernel19:41
fsmithredI have something for that.19:41
fsmithredkinda old, but should still work.19:41
fsmithredThis is ascii with backports kernel and a minimal desktop: https://get.refracta.org/files/experimental/refracta-test-oblx-bpo-20180913_1952.iso19:42
koollmanand ... yeah. doing basic text install, get stuck on 'detecting cdrom'. and doing default graphical install, no touchpad, but also get stuck on detecting cdrom.  :)19:42
koollman(from netinst iso)19:42
fsmithrediso is on a usb?19:42
koollmanyes19:42
fsmithredyou can play tricks with finding the cdrom19:43
koollmanshould I try mounting it before the detect step ?19:43
fsmithrednot sure if I can remember exactly, but I'll try19:43
fsmithredI think you have to let it fail to find the cdrom, then switch to another vt BEFORE you dismiss the error screen19:43
fsmithredthen make a symlink19:43
koollmandoesn't get to an error. it is just 'stuck'. no logs on alt-f4 terminal19:44
fsmithredthen go back to vt1 and tell it to try again19:44
fsmithredoh19:44
fsmithredother option would be to boot the refracta iso I just posted above and do a debootstrap install of devuan19:45
fsmithredthen you really get to choose what you want19:45
koollmanwell, that would be equivalent to doing a debootstrap install from any other env, right ? (assuming recent enough kernel)19:45
fsmithredyeah, with devuan's debootstrap. Not sure if that makes a big difference.19:46
fsmithredyeah, ascii has a devuanized debootstrap. The debian version doesn't include us in any of the scripts.19:47
fsmithredThere might still be a working beowulf mini.iso.19:48
fsmithrednot sure if the kernel has updated since the last build.19:48
koollmanI can probably manage to do that. I remember doing install from debootstrap a while ago. I'm still going to try the live iso first19:48
fsmithredok, here's the last mini.iso that was made, in case the live doesn't work: http://pkgmaster.devuan.org/devuan/dists/unstable/main/installer-amd64/current/images/netboot/19:50
koollmanlive is promising ... reached xfce desktop :)19:52
koollman(still probably a few bugs given the delays in boot/udevd and errors mentionning iommu : https://dpaste.de/Kbn6#L228 )19:55
fsmithredif it booted, it should install no problem20:00
fsmithredbrb20:01
koollmanI think there are some problems with udev,still. I can see 'udevadm settle' (launched by gparted on validation) that was stuck for maybe 4/5 minutes. but gparted seems to continue now. (udevadm was polling /run/udev/queue I think)20:06
koollmanoh ... it's going to do it for each partition. that may take a while :)20:09
koollman(guess I will pre-partition next time :) )20:09
fsmithredgpt or msdos partition table?20:10
koollmangpt20:11
koollmanand, it's not the disk that is slow. nvme :)20:11
fsmithredmake sure to create the bios_grub partition20:11
fsmithredat least 1mb with no filesystem on it, with bios_grub flag20:11
fsmithredoh20:11
fsmithredunless you're using efi20:11
fsmithredthen you just need efi partition20:12
fsmithreddo you know if you booted uefi or bios?20:12
koollmanuefi. and I'm reusing existing efi partition20:13
fsmithredok20:13
fsmithredI've noticed that gparted is very slow on a gpt disk with many partitions20:13
fsmithredslow to come up.20:14
fsmithredonce it's up, it's fine20:14
koollmandifferent problem there, I think20:14
koollmancame up fast, but writing partitions is very slow. And if I look at what gparted is doing, it's waiting for 'udevadm settle', which takes a few minutes, then do one operation, then call it again20:15
fsmithredgdisk might work better20:16
koollmanmaybe I will kill it and start over doing partitions with another tool first20:16
fsmithredif you pre-format, make sure to select "Do not format" in the options menu.20:16
fsmithredand I'm talking about making the filesystems, not just the partitions.20:17
koollmanok. took 30s in cgdisk :)20:18
koollman(and that was mostly human time)20:18
fsmithredthat's sounds better20:18
koollmanstill indicates a bug in udevadm or gparted in the live image20:19
fsmithredI think gparted20:19
koollmanwell, it was really a long-lived udevadm process that was slowing it down.20:20
fsmithredI've had some trouble with partitioning on uefi with it. Can't remember specifics, but I tend to use gdisk now.20:20
koollmanalthough it is clearly a bug to call that from the main thread in a graphical application, too ...20:20
fsmithredkoollman, when you get to the window for grub install, you might want to go to a root terminal, chroot /target and then add backports to install the bpo kernel20:32
fsmithredthen go back and let it install the bootloader20:33
fsmithreddon't choose the chroot button20:33
koollmannah, I was impatient. now it's booting on the newly installed devuan. But, yeah, I will get new kernel first thing after boot is finished. seems to get some slowdowns like the liveusb (which I expected)20:34
koollmanand ... login prompt \o/20:34
fsmithredconsole login?20:34
koollmanyeah. but that's ok. I can work from there :)20:35
fsmithredgood :)20:35
koollmanhm. the generated fstab is wrong, though20:36
koollmanI suppose since I preformatted everything, I was not asked which fs to pick ... so the fs column is missing20:36
fsmithredI will look at that now20:37
fsmithredpretty sure it pulls the fs type from blkid20:38
koollmanand ... rmmod pcspkr. I forgot how loud it was in this laptop :)20:40
fsmithredit should have printed the fs type in fstab20:42
koollmanmaybe I did mess up something. but it's fine, easy to correct on my side20:43
fsmithredlook over the log. Should be in your home dir.20:43
fsmithredif [[ $no_format = "yes" ]]; then20:44
fsmithredfs_type_os=$(blkid -s TYPE "$install_dev" | awk -F"\"" '{ print $2 }')20:44
koollmancurrently booting with backports kernel. still seem to get a delay at boot. I will investigate that later. once it is booted I'll try blkid command20:45
fsmithredfind that place in the log and make sure that it saw the type20:46
koollmanbut, hey ... now I get to a graphical login manager :)20:47
koollmanand xfce is working.20:47
fsmithredcool20:47
koollmanfs_type_os=20:48
koollmanso ... yeah. not detected when it ran20:48
fsmithredweird20:48
koollmanI will probably try some more installs. I suspect it was my fault. And if I run it now it does gives back 'ext4'20:49
koollmanmaybe at least add a check for empty string20:49
fsmithredand have it do what?20:50
fsmithredopen fstab in a text editor?20:50
koollman:)20:50
koollmanthat's an option. or error with an option to re-run detection maybe20:50
koollmannow I need to find out why I lose 2 minutes with no logs in the middle of my boot. although I do strongly suspect unsupported hardware and udevd :)20:51
fsmithredif you're going to do more installs, you might consider installing beowulf20:52
koollmanmaybe, yes20:52
fsmithredupgrading to it with full xfce desktop right now would be tricky, because some of the task-* packages aren't ready20:53
fsmithreddo standard system install then add desktop parts individually (xfce4 metapackage should be ok)20:53
fsmithredbut don't select it from tasksel20:53
koollmanif I do standard install I generally don't even bother with X during install. it comes later. and if ion3 is working I'm all set :)20:55
fsmithredgreat20:55
koollmanhttps://dpaste.de/zyDH around lines 1007, 1008, 1009 ... looks like two pauses, of about 1 minute each. do you have any idea what would fail with a timeout of 1min ? (or maybe two minutes ... but with a udevd message in the middle)20:58
koollman(and of course before that, I have at least one problem with a null dereference in kernel. which is ugly but maybe need a much newer kernel for this cpu/chipset :) )20:59
fsmithreddo you know what device is cited there?21:02
fsmithred /devices/pci0000:00/0000:00:01.7/0000:05:00.021:02
koollmanwhat pci id would that be. 05:00.0 ?21:03
koollman10ec:522a, realtek card reader RTS522A21:04
fsmithreddevices/pci0000:00/0000:00:01.7/0000:05:00.0/rtsx_pci_sdmmc21:04
fsmithredsdmmc... is that your nvme drive?21:04
fsmithredyes, look for that number in lspci output21:05
fsmithredoh21:05
koollmanit was the line just after my question ;)21:05
fsmithredyeah, finally saw that21:05
koollman(first bit is pci id if you want exact model)21:06
fsmithredmaybe you could turn it off in bios21:06
fsmithredand wait for a few kernel updates21:06
koollmanwell, I can try a more recent kernel, too21:06
koollmandoesn't seem very new, though21:08
fsmithredthe card reader or the kernel?21:08
koollmancard reader21:09
koollman(by the way thank you very much for all the help :) )21:11
koollmanand I even got my wireless card working. great :)21:12
koollman(although having an ethernet port is pretty great when trying this kind of things. was hard to find a decent model that still had one)21:13
gnarfacepoontangmessiah: the answer is it never did, if you set it up right.21:13
fsmithredkoollman, lsmod |grep rtsx21:14
koollmanrtsx_pci is loaded21:15
fsmithredI'm out of ideas21:15
koollmanI'll probably try a few kernels to see if something bleeding-edge behaves the same way21:16
fsmithredyou could probably run the 5.2 from ceres on beowulf21:17
koollman(and blobwars run. my work laptop is finally ready ;) )21:43
jonadabgolinux: Made it up out of whole cloth.22:23
jonadabAs a rule of thumb, any answer followed by "HTH.HAND." is probably not actually the real answer.  This has been true since usenet was invented, I think?  Or for most of that time, anyway.22:24
golinuxHave never run across that (that I can remember) on any Devuan channels.22:37
drawkulaI see that for the 1st time now22:39
golinuxAnd you've been around a veeeery long time.22:40
drawkulamaybe jonadab and me read different groups...22:43
drawkulaglobalising own experiences is often ... hmmmm ... misleading22:43
drawkulaor our memory is cheating... lossy data compression in our grey matter...22:44
jonadabgolinux: Eh, I didn't get on the internet until 1993.22:47
jonadabThere are folks who have been around way longer.22:47
jonadabBut yeah, I thought "HTH.HAND" would be understood in technical circles.  I guess I should switch to saying "kappa" like the gamer kids.22:48
golinuxDon't know that one either.  I play solitaire.22:49
jonadabOr just stop making jokes.22:49
Akulii play solitaire too <322:49
jonadabI think "kappa" originated on Twitch?  Which is a change of pace, because most of the really terrible memes come from 4chan.22:49
Akulii wrote a curses-based solitaire that i can play with terminal22:49
Akuliit draws the cards with unicode line-drawing characters, and it even has colorz22:50
golinuxProbably not a good idea to joke about official Devuan news22:50
jonadabI have seriously thought about writing a terminal-based turn-based shmup.22:50
golinuxCan we please take this over to #debian fork?22:50
golinuxno space22:51
golinux#debianfork22:51
golinuxThat's the chat channel22:51
r3bootghe, hth.hand, thats old :D23:05
drawkulaits k3wl 2 b cryptic23:23
sixwheeledbeastcr3wl may b23:45

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