libera/#devuan/ Friday, 2021-07-23

SoltisOkay so first weird problem - closing the lid and reopening causes the keyboard to stop responding and mouse gets confined to a small area in the center of the screen.00:43
gnarfacesounds like a problem with the video driver failing to resume from suspend properly00:52
gnarfaceor something like that00:52
gnarfaceis it a nvidia card?00:53
gnarfaceyou can typically get better behavior out of nvidia cards by using the backports kernel and driver versions00:53
gnarfacediminishing returns on that for other hardware though...00:54
gnarfacebut i wonder if there might be a way to kick it back into shape just by restarting xorg00:54
EHeMNow to see how this is setup versus previous home...00:57
EHeMWhile the source package for the Linux kernel got its security update, the kernel builds didn't get their updates (linux-source was updated, linux-image-* weren't updated).00:58
gnarfaceit's normal for the devuan build servers to be a couple hours or so behind the debian ones01:02
gnarfacealthough it's also normal for it to take some time for the binary packages to sync up too, even if they're not the ones being rebuild01:03
gnarfacerebuilt*01:03
EHeM>24 hours at this point, I believe the update was Tuesday.01:03
gnarfaceuh oh01:03
gnarfacehmmm01:03
EHeMHmm, might merely be linux-image-amd64.01:03
gnarfaceyou're using deb.devuan.org in the sources.list?01:06
EHeMAlso have pkgmaster.devuan.org.01:07
fsmithredI got this one today: https://dev1galaxy.org/viewtopic.php?id=233001:09
fsmithrednooooo01:09
fsmithredshit01:09
fsmithred4.19.0-17-amd64 #1 SMP Debian 4.19.194-3 (2021-07-18) x86_64 GNU/Linux01:09
EHeMHmm, looks like "pkgmaster.devuan.org" is missing linux-image-4.19.0-17-amd64 and the updated linux-image-amd64.01:28
EHeMThere is also an issue with `lynx https://pkginfo.devuan.org/`.01:28
XenguyEHeM, What's the issue?01:43
rrq.. seems to be a spurious "200 OK" header01:43
EHeMXenguy: Try loading (and using) that page with `lynx` instead of your favored browser.01:43
XenguyInteresting, I see source code01:44
fsmithredsame here, but it works ok in w3m01:46
XenguyWorks in 'links'01:47
XenguyEHeM, Thanks for the report BTW01:53
rrq.. no, lynx has a problem also without that spurious header... wonder why01:54
XenguyI've noticed that lynx will also complain about 'https' problems, when other text browsers are fine...01:55
XenguySo also not sure01:55
rrqit renders the page from a local file.. is there a way to see its headers in lynx?01:57
Xenguy\   ?01:58
Xenguyor01:58
Xenguy= ?01:58
fsmithred \02:00
fsmithredexcept it doesn't02:01
rrqthere's a -head command lne argument.. but it sees the headers fine.02:03
rrqa lynx -dump adds an extra, final newline .. probably just its dump function02:05
rrqhmm there's a "\342" character ...02:07
XenguyWonder if there's an existing bug report?02:08
XenguyNot really sure how to search that out in this case02:09
rrqactually utf-8 code ’02:10
rrqnope, not that02:13
Soltisgnarface: It's Intel graphics, but it may be an issue with the kernel version being too old, apparently it requires a pretty recent kernel.02:26
rrqEHeM: perhaps you could lodge a bug report to debian about lynx; it works fine on a dump of that page but not upon direct connection.. maybe it's unhappy that css is gotten from a different server(?)02:40
fluffywolf...  lynx knows css exists?02:45
rrqright; probably not the problem... it should know to ignore css02:52
rrq.. but I expected it to show an html page the same way as it shows a file that is its own dunp of that page02:56
fluffywolfjust glancing at it, the bug looks to be the script on the devuan server is not returning a valid content-type header.03:05
fluffywolfI'm not sure who's responsible for maintaining that.03:06
gnarfaceSoltis: it's worth trying the backports kernel for newer models of any laptop for sure03:08
fluffywolfI think the developer thought that a meta http-equiv would be sufficient, but it is not.03:08
fluffywolfwho maintains the pkginfo search script?03:08
* gnarface thought it was Xenguy 03:08
gnarfaceto be fair, there's a lot of misinformation about the validity of using meta tags to override http headers floating around03:09
gnarfaceit's one of the things i fought hardest against before i burned out completely03:10
fluffywolfin this case, the server isn't sending back a content-type header at all, and lynx is correctly assuming the result is plain text.03:10
fluffywolfwhile other browsers assume it must be html.03:10
gnarfaceyea, the other browsers actually behave out-of-spec because of bullying by Microsoft03:11
rrqfluffywolf: great.. thanks.03:12
fluffywolfinterestingly, if you hit = twice, it renders it as html.  that might be a bug.  lol03:13
fluffywolfwhoops, \ twice03:13
rrqgot changed now03:14
fluffywolf  Content-Type: text/html03:15
fluffywolflol03:15
fluffywolfor, it got fixed between the two times I hit \.  :)03:15
EHeMNext issue is it relies on JavaScript and thus Lynx can't do the query.03:17
fluffywolfthat's probably harder to fix.03:18
* gnarface would have strongly advised against that too, but was not consulted 03:19
fluffywolfdoes it really use javascript?  *checks*03:21
fluffywolfit doesn't use javascript, but it lacks a submit button.03:23
fluffywolfthat's easy to fix.03:23
gnarfaceoh, lynx can't submit a form with the enter key?03:23
gnarfaceanother deviation from standard that other browsers have already routed around, i would assume03:23
gnarfacei wonder if a hidden field named "submit" would be enough to make the enter key submit the form without a button03:24
fluffywolfor just have a submit button labeled "submit query" or such, like most forms.03:25
gnarfacewell, that would work obviously but i assumed the lack of a visible button was an artistic choice03:26
Soltisgnarface: I'm already running Chimaera; pretty sure that doesn't have a backports repo yet03:26
gnarfaceit should have a submit button, every form is supposed to but it is common for them to be omitted03:26
fluffywolffunctionality trumps artistic choices.  :)03:26
gnarfaceSoltis: yea, that's true.  what kernel is it?03:27
Soltisgnarface: 5.1003:27
Soltisgnarface: Need at least 5.13 I think.03:27
gnarfacedamn, i'm not sure that's in the repos yet03:27
gnarfacei think backports is also 5.10 or maybe 5.1103:28
SoltisYeah, I don't think it is either.03:28
Soltis... how much of a pain is building kernels these days? I haven't done it since circa 2.6 on Slackware.03:28
gnarfaceSoltis: probably easier than it was then03:29
* fluffywolf pokes whichever person here is maintaining that script03:29
gnarfaceSoltis: you can probably copy the running config from 5.10 then build it with kernel-package fairly easy03:29
fsmithredcan't submit a search pattern in w3m, either03:29
fluffywolfrather than "push enter", add an <input type="submit">, for better browser support.03:29
gnarfaceyou can even <input type="submit" name="submit" value="push enter" />03:30
Soltisgnarface: ... pretend my process back then was hand configuring the entire kernel via `make menuconfig` in a tarball. I've never used a script to do it.03:30
fluffywolfa button saying "push enter" is just silly.  :P03:30
gnarfaceSoltis: make menuconfig should have an option to load the old config.  feed it the one in /boot03:31
SoltisOkay.03:31
SoltisAh yes, I do see the config there.03:32
fsmithredSoltis, are you making your own new kernel for the xps?03:33
Soltisfsmithred: I'm thinking it may be necessary, yes.03:33
fsmithredhow new do you need? 5.10 in chimaera and ceres03:34
Soltisfsmithred: Haven't yet, trying to remember all the steps to do it + not sure if those steps have changed in the past 15 years03:34
Soltisfsmithred: 5.13 I believe.03:34
fsmithredwow. I think liquorix has 5.1203:34
SoltisThe Intel docs have you cloning from a github repo to enable the drivers.03:35
SoltisSo that is probably the direction I'm going to be going03:35
* fluffywolf notes the script still says "push enter", scowls at whomever should be fixing it03:41
Soltis... corporate Linux distros. I feel so dirty inside.03:41
Soltiss/distros/repos/03:41
rrqfluffywolf: yeah I'll look at it eventually.. pity though lynx limits its 'return action' to forms of single input elements03:44
fluffywolfit's a simple fix...  just remove the push enter div and replace it with an <input type="submit">.03:45
fluffywolfshould work in all browsers.03:46
fluffywolfprobably more compatible with phone browsers and such too.03:46
rrqyes, pretty much... comes with the "alternative cost" of redirecting me :)03:47
fluffywolfalso, a design note, I really hate the styling of the dropdown box.03:47
rrqfluffywolf: thanks... I think I'll need redesign that form .. I'm sure I can make something that everyone hates ;)04:02
Xenguyrrq, That could be the ultimate goal of course04:03
XenguyThat's a hard problem, haha04:04
XenguyBut I digress04:04
SoltisI don't recall a kernel taking this long to build even on an Athlon 100004:04
fsmithredare you using more than one core?04:07
Soltisfsmithred: make -j $(nproc --all) targz-pkg LOCALVERSION="-xe-max"04:07
* fsmithred hasn't compiled a kernel since Athlon 200004:09
Xenguymenuconfig is a long ways away in this memory bank ; enjoyed it at the time ...04:11
fluffywolfrrq:  heh.  I like my form elements to look like form elements, not plain text.04:12
Xenguyfluffywolf, You should try some pull requests, seems like you have that knowledge04:13
fluffywolfspecifically, I had absolutely no idea that was a form element, until I loaded the page with lynx.  :P04:13
fluffywolfI always used make menuconfig too...   but it's been a long time.  stock kernels have had the features I've needed lately, although it seems I essentially always have to use a backports kernel to get them.04:18
XenguyFor a long time I've been using stock kernels too04:18
XenguyIt's fun to explore, but it only scaled so far for me04:19
golinuxI dug out the css for the iso docs04:20
golinuxPM?04:20
XenguySure04:20
SoltisApparently I was smoking crack. There IS a module loaded for this device - I just misread. So .... mouse confined to a rectangular area 1/4 the size of the screen and keyboard unresponsive after closing and reopening laptop lid. Any ideas?05:16
SoltisOh, one other note: this did NOT happen the one time some kind of error prevented the screen lock program from running.05:16
gnarfacecan xrandr change the resolution back?05:17
Soltisgnarface: All input methods are unresponsive so I don't know. Even shortcut for killing X fails. Note that the resolution itself is not changed; the screen displays normally at its full resolution, but the mouse can't be moved outside the smaller region in the center of the screen.05:20
SoltisI can MOVE the mouse, but I can't click, and keyboard inputs are ignored entirely.05:21
gnarfacei can't be sure there's a fix for this05:22
gnarfaceyou can't ctrl+alt+f2 out to a virtual terminal?05:22
gnarfacethe ctrl+alt+backspace thing is off by default these days, you can re-enable it05:23
SoltisOh, I didn't think I could, but I can - I forgot I had to use Fn to actually do F205:23
SoltisNot used to laptops.05:23
Soltis... I thought I had enabled it, perhaps not.05:23
gnarfacei would still try setting the resolution again05:23
SoltisYeah, let's see.05:23
gnarfaceDISPLAY=":0.0" xrandr...05:24
gnarfaceif you're using a graphical login it could complicate this05:24
SoltisI'm not.05:25
SoltisI'm logged in to the shell05:25
Soltis... just gotta remember how to do this shit. I stopped using desktop Linux in 2007 or so; this is my first foray back into it.05:25
Xenguywb, should be easier now05:26
gnarfacewell you can probably just kill Xorg if you have to to get it back but i was hoping we might be able to find a way to reset it without literally resetting it05:26
SoltisYeah, lemme try xrandr05:26
gnarfacethere is some chance the issue could be related to the compositor but usually it's a bug in power management, sometimes bios related05:27
fluffywolfI've had no end of problems with power management, but I've yet to see that specific one.05:28
gnarfaceyea, usually only windows ever gets those fixes05:28
fluffywolfrefusing to suspend, immediately resuming, or crashing on suspend or resume seem to be the most common ones...05:29
SoltisJust doing `DISPLAY=":0.0" xrandr` throws "can't open display :0.0" - and doing `xrandr --listmonitors` just says "can't open display"05:30
Soltis... ugh.05:30
Soltisgnarface: Okay so DISPLAY=":0.0" XAUTHORITY=/home/user/.Xauthority xrandr --output eDP-1 --mode 1920x1200 didn't throw an error, but it also didn't change anything.05:50
gnarfacedrat05:54
gnarfacewhat if you change it down a size then back up a size?05:54
gnarfacedoes the display update visually or does it seem to ignore the change?05:54
SoltisIf I try any other mode (so far I've tried 5 or so) it says "configure crtc 0 f ailed"05:55
gnarfacehmm, interesting, it might just be actually froze05:56
SoltisYeah, the only thing that responds is the mouse cursor05:56
SoltisIt recovers fine if I kill xorg, so at least it's not unrecoverable.06:18
eyalrozDoes the debian installer save a log somewhere?10:47
gnarfacei think in /root/10:48
gnarfaceeither there or /var/log/ i thought10:49
rrqeyalroz: the installer keeps a /var/log/syslog while installing, but that's in the installer (RAM) file system... you'll need to copy it to /target it before the first reboot11:10
pandom79Can i download a devuan iso runit edition? or How can i switch to runit from a default bootstrap?11:16
gnarfacei think you just install runit11:22
fsmithredthat works.11:22
gnarfacei couldn't speak for how well all the services are supported though11:23
fsmithredif you just want to take it for a test drive, I made a live-iso with runit: https://get.refracta.org/files/experimental/refracta_runit_test_amd64-20210614_1136.iso11:23
fsmithredthere are sample runscripts in the user's home and some have been enabled.11:23
eyalrozrrq: Would it not make sense to do so automatically? Or at least do so for non-final-release installers?11:37
eyalrozIt would have definitely helped me jog my memory regarding what I did exactly during installation.11:37
rrqgood point.. probably comin in the next build if I can get my development team interested11:38
eyalrozrrq: Should I file that idea as a bug against something?11:41
eyalrozAnyway, I'm trying to use reportbug for the issues I'm noticing where I believe I can decide which package is to blame.11:41
rrqthanks. it's fine without, but maybe useful to possibly garner some othet thoughts on it11:42
rrqdevuan-installer11:43
pandom79fsmithred: i'll try it. thx ;)11:44
fsmithredrrq, eyalroz see /var/log/installer/syslog and other files in the installed system11:50
rrqah, great!11:51
eyalrozfsmithred: ... but will this tell me what I chose during intallation?11:51
fsmithrednot sure11:52
fsmithredI'm booting a VM now to look11:52
eyalrozWhat happened to the v4l2ucp package in Chimaera?11:53
eyalrozhttps://pkginfo.devuan.org/cgi-bin/policy-query.html?c=package&q=v4l2ucp11:53
eyalrozit's just gone11:53
Jorileyalroz: https://packages.debian.org/search?keywords=v4l2ucp11:53
JorilIt looks like it's gone from Bullseye11:54
eyalrozJoril: Does anybody know why?11:56
eyalroz... maybe I'll just ask them on their channel.11:56
aplainzetakindI can't autostart pulseaudio. Neither with a script fragment in /etc/X11/Xsession.d/ nor from .xinitrc.11:56
Jorileyalroz: maybe https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=875221 ?11:57
fsmithredaplainzetakind, is it beowulf or chimaera?11:57
aplainzetakindbeowulf11:57
fsmithreddid you look in /etc/pulse/ ?11:58
fsmithredneed to edit the autostart file in there.11:58
fsmithredcomments tell you what to do. (comment or uncomment a line.)11:58
fsmithredautospawn not autostart11:59
aplainzetakindI see.11:59
aplainzetakindLet's try11:59
eyalrozJoril: Thanks. Unfortunately, the upstream github repo has not been updated in 7 years :-(12:00
fsmithredI think there might be a package that does it12:00
fsmithreddebian-pulseaudio-config-override12:00
aplainzetakindIt worked. Thanks.12:01
fsmithredyw12:01
aplainzetakindDo you have any idea why I ended up having to launch xmonad with dbus-launch in my .xinitrc in order to enable xmobar to find sockets to work with?12:02
aplainzetakindI hadn't had to do that in my previous set-up attempt a couple of days ago (which I broke messing with nvidia drivers).12:03
fsmithredno, sorry. I don't know those words.12:04
aplainzetakindXmonad is a wm and xmobar is a status bar.12:04
aplainzetakindShouldn't be particular to them. I somehow didn't have dbus sockets launching my wm.12:05
aplainzetakindUsing only startx/.xinitrc12:05
fsmithredeyalroz, installer/syslog has a list of packages to be installed. You can search through that for package names12:06
fsmithredaplainzetakind, does startx work without dbus-launch?12:06
aplainzetakindYes.12:06
aplainzetakinddbus-launch only prevents the error from the status bar.12:07
aplainzetakindWhich is launched by the wm, after xinitrc is done.12:07
aplainzetakindI had to change my last line from `exec xmonad` to `exec dbus-aunch xmonad` basically.12:08
fsmithredis elogind installed?12:08
aplainzetakindYes.12:08
fsmithredall that stuff makes my head hurt12:13
aplainzetakindOh well. I'll just hope it's not a syptom of a significant misconfiguration.12:17
aplainzetakindWhat do I need for usb drives to automount with sane permissions?12:18
fsmithrednot sure about automount, but pmount or udevil let users mount external drives12:19
aplainzetakindLet me check udevil.12:20
aplainzetakindudevil just works apparently.12:29
brocashelmfsmithred: someone recently did these instructional videos on refracta (but they are in portuguese): https://www.youtube.com/watch?v=zOFUsqhaxjo https://www.youtube.com/watch?v=u6q3w9IGnyA16:54
brocashelmgoing through them, looking pretty cool16:55
brocashelmcool, i see another new video on devuan premiering in an hour: https://www.youtube.com/watch?v=9nOfYKC_0YI (not familiar with this channel, but i'm interested in checking this out)17:03
brocashelmwell, that video (that "premiered") was pointless18:33
brocashelmdidn't really show off anything; waste of time18:33
fluffywolfreset: pkginfo from yesterday, just confirmed it now works fine in lynx.19:18
fluffywolfgrr, stupid auto completion.  that was re:19:18
man_of_the_oceanI tried to install Devuan Beowulf. I am using lvm + luks encryption. When I try to boot I get dropped into a grub shell. When I `ls` I cannot see the lvm partitions even after `insmod lvm`.22:13
gnarfacemissing step i think22:30
gnarfacesomething about cryptsetup22:30
gnarfaceyou have to add a flag to the grub commandline for it maybe?22:31
gnarfaceor some config file, i'm always forgetting22:31
fsmithredif /boot is part of the encrypted volume, you need to enable grub_cryptodisk or something (I'd have to look at it.)22:32
sadsnorkman_of_the_ocean, is this what you are getting?  https://dev1galaxy.org/viewtopic.php?pid=2969422:34
sadsnorkThe answer being to remove grub-efi-amd64-signed?22:34

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