Xenguy | sokan: My understanding is that the Debian project would continue: they have protocols for who assumes what in the event no one runs for DPL election | 00:03 |
---|---|---|
Xenguy | *assumes what duties | 00:03 |
sokan | is sddm that gray ugly mess by default or did I do something wrong? :\ | 00:22 |
dtigue | So I come from an Arch background, but I've been (forced) to use Windows for the past few years and am now trying to get back up to speed with Linux. Is there any documentation someone can point me to that has info about devuan apt sources so I can add more applications? | 04:08 |
dtigue | Specifically chromium | 04:08 |
Jjp137 | dtigue, here: https://devuan.org/os/etc/apt/sources.list | 04:27 |
dtigue | Thanks Jjp137 | 04:31 |
Jjp137 | np | 04:31 |
golinux | There's also https://pkginfo.devuan.org if you're looking for a package. | 04:31 |
dtigue | Cool, thanks golinux | 04:35 |
dtigue | I'll check them both out | 04:35 |
nailyk | hello ! Do you know if a raspberrypi version of devuan exist please ? | 11:29 |
KatolaZ | nailyk: sure they do | 11:32 |
sokan | How can I get me sddm-kcm installed? I don't find it via apt search sddm-kcm | 11:32 |
KatolaZ | https://files.devuan.org/devuan_ascii/embedded/ <- nailyk | 11:32 |
nailyk | thanks!!! | 11:33 |
KatolaZ | sokan: it's in kde-config-sddm | 11:33 |
KatolaZ | sokan: you should install apt-file | 11:33 |
KatolaZ | then `apt-file update` | 11:33 |
KatolaZ | and `apt-file search SPEC` will tell you the packages which install a file whose path contains SPEC | 11:34 |
KatolaZ | sokan: ^^^ | 11:34 |
sokan | SPEC is whicher package? | 11:35 |
sokan | whichever* | 11:35 |
sokan | got it working! | 11:36 |
sokan | Why on earth can't add a theme on sddm them? :S | 11:36 |
KatolaZ | sokan SPEC is whatever you are looking for | 11:42 |
KatolaZ | apt-file search sddm-kcm | 11:42 |
KatolaZ | kde-config-sddm: /usr/share/sddm-kcm/main.qml | 11:42 |
KatolaZ | sokan: ^^^ | 11:42 |
sokan | foud it | 11:43 |
sokan | found* | 11:43 |
sokan | https://store.kde.org/p/1214121/ | 11:44 |
sokan | based on this, I should be able to change theme on sddm, but I don't see any option at all. Is there another way to "impose" a theme? Do you have any idea KatolaZ ? | 11:44 |
KatolaZ | sokan: I am not using KDE, sorry | 11:45 |
KatolaZ | or sddm | 11:45 |
sokan | no worries | 11:45 |
KatolaZ | or any other DE | 11:45 |
KatolaZ | :) | 11:45 |
sokan | xinit for logging in then I presume? ;)\ | 11:45 |
KatolaZ | wdm | 11:46 |
KatolaZ | it's more than enough | 11:46 |
sokan | any panels/bars? | 11:46 |
KatolaZ | sokan: ? | 12:00 |
sokan | like xfce4 panel. To show battery, cpu usage etc | 12:00 |
KatolaZ | you mean if they are available in the repo? | 12:01 |
KatolaZ | sure they are | 12:01 |
sokan | no no if yuo use any | 12:01 |
KatolaZ | oh | 12:01 |
KatolaZ | :D | 12:01 |
KatolaZ | nope | 12:01 |
sokan | :O | 12:01 |
sokan | hardcore ricing | 12:01 |
KatolaZ | why? | 12:02 |
sokan | just curious is all | 12:02 |
KatolaZ | no, I mean, why you define that "hardcore ricing"? :D | 12:02 |
KatolaZ | I check battery every now and than with acpi | 12:03 |
sokan | just one wm no extra bar programs == hardcore in my eyes :P | 12:03 |
KatolaZ | used to have a small bar, but it is a waste of space on my screen | 12:03 |
KatolaZ | :) | 12:03 |
KatolaZ | and anyway, there is a script which puts the laptop on standby when the battery is low | 12:06 |
KatolaZ | so I don't need to check the battery level | 12:06 |
KatolaZ | :) | 12:06 |
epergny | hello community, very old time Linux user and very likely future Devuan convert here! | 12:14 |
epergny | s/old/long/ | 12:14 |
KatolaZ | welcom epergny | 12:14 |
sokan | joinb the club epergny | 12:14 |
epergny | my needs are simple: my "workstation" (just a fat PC with no sound: no speakers, no internal speaker, no sound configure) runs basically Emacs compiled from source, a few terms and browsers. I think Devuan would be totally fine here ^ ^ | 12:16 |
epergny | and I've got a few headless Raspberry Pis running simple things. sshd, no X, private bare git repositories. I may test the water first with a RPi. | 12:18 |
KatolaZ | epergny: you should be fine | 12:18 |
epergny | I also have a thing of beauty: one of the Debian server with the longest uptime in the world. Now I keep it on (in a datacenter) just to see how many years it reaches. | 12:19 |
epergny | I did patch everything I could but without rebooting, so it's not secure. It only has sshd open though. It's of course running pre-systemd Debian. | 12:19 |
epergny | An old Debian of yore, back when Debian didn't have that weird PID 1 rube-goldberg machinery. | 12:20 |
KatolaZ | how long has it been running? | 12:21 |
epergny | two sec, let me paste the uptime | 12:21 |
epergny | 11:09:24 up 2150 days, 21:04, 1 user, load average: 0.07, 0.04, 0.05 | 12:22 |
KatolaZ | :) | 12:22 |
KatolaZ | nice one | 12:22 |
epergny | Linux ksXXXXX.kimsufi.com 3.2.13-grsec-xxxx-grs-ipv6-32 #1 SMP Thu Mar 29 09:43:21 UTC 2012 i686 GNU/Linux | 12:22 |
KatolaZ | so it's a wheezy | 12:23 |
epergny | it's hosted at OVH, in France. I just replaced its "ks number" with XXXXX because I wouldn't want someone to DDOS it or something : ) | 12:23 |
epergny | 7.11 | 12:23 |
epergny | basically I forgot I had it and one day I noticed it was there still running (paying about 10 EUR / month for it) and had a big uptime, like 3 years or so. And I decided to keep it just for fun. | 12:24 |
epergny | does Devuan ASCII offer easy full LUKS disk encryption during install like Debian? | 12:26 |
KatolaZ | epergny: the installer is the same as devuan | 12:27 |
KatolaZ | unless you install from a live | 12:27 |
KatolaZ | sorry, I meant the same as debian | 12:27 |
epergny | : ) | 12:28 |
KatolaZ | arm-images are pre-built though | 12:28 |
epergny | I'm running a Core i7 / 4.9-something-amd64 kernel | 12:28 |
KatolaZ | epergny: https://files.devuan.org | 12:28 |
KatolaZ | look for devuan_ascii | 12:29 |
KatolaZ | under installer-iso you have the usual netinst, cdrom, dvd | 12:29 |
epergny | ascii is stretch right? | 12:29 |
KatolaZ | ascii is ascii :) | 12:29 |
epergny | ofc | 12:29 |
KatolaZ | https://devuan.org/ | 12:30 |
epergny | and the release follow the same cycle as Debian? For example when Buster shall be stable, Beowulf shall basically come out? Or is there a long delay? | 12:31 |
KatolaZ | there has been a longish delay in the past | 12:32 |
KatolaZ | but it looks like we might catch up with beowulf | 12:33 |
KatolaZ | maybe with a small delay | 12:33 |
epergny | I'm a Linux (Debian even) power user (since 20 years) but I'm no sysadmin and not a Linux guru. So I may ask silly questions. | 12:33 |
Bjornn | I can one up you on the silly | 12:34 |
epergny | heck, even more than 20 years. Started with Slackware 3.0, that was before 1999. | 12:34 |
epergny | Bjornn: ah ah :) | 12:34 |
gnarface | hey slackware! ^5 | 12:35 |
KatolaZ | epergny: you are in good company here I guess | 12:35 |
sokan | KatolaZ: do you know if I can install this https://packages.debian.org/sid/keepassxc in devuan? | 12:35 |
epergny | gnarface: ^ ^ back then we were still all on dialup and d/l ing Linux was too crazy. So I bought the Slackware book and got a Linux CD with it. | 12:35 |
KatolaZ | sokan: in general you should avoid mixing releases | 12:36 |
KatolaZ | sokan: especially if you don't know your way around De??uan particulary well | 12:36 |
KatolaZ | (no offence intended, but dealing with sid requires a bit of experience with debian) | 12:36 |
sokan | thing is I have my passwords in keepassxc and I'd like to stay there | 12:37 |
KatolaZ | sokan: but it's available in ascii as well | 12:37 |
sokan | keepassxC? | 12:37 |
KatolaZ | why do you need the unstable version? | 12:38 |
KatolaZ | o_O | 12:38 |
epergny | First silly question: I compile my Emacs "--without-dbus --without-gtk --without-sound". I don't use any DE: I just use a simple WM (Awesome WM). Do I even need dbus on my system? | 12:38 |
sokan | KatolaZ: oooh. It's in backports | 12:38 |
gnarface | epergny: probably not | 12:38 |
sokan | how do i add stuff from backports? :S | 12:38 |
gnarface | epergny: you don't use a graphical login either, right? you just startx? | 12:38 |
epergny | gnarface: yup exactly | 12:39 |
KatolaZ | epergny: it depends on whether any of the app you use needs dbus | 12:39 |
KatolaZ | or was compiled with dbus support | 12:39 |
epergny | old school text mode login (and I miss my runlevels) | 12:39 |
gnarface | epergny: some WMs might need dbus. Steam needs dbus. most of the other stuff that wants dbus can do without it | 12:39 |
KatolaZ | sokan: you should add ascii-backports repos in /etc/apt/sources.list | 12:39 |
epergny | emacs, xterm, browsers. That's basically it. | 12:39 |
epergny | gnarface: ah good to know. | 12:40 |
gnarface | epergny: (thunderbird and firefox like dbus but they don't do anything particularly useful with it) | 12:40 |
KatolaZ | epergny: some apps need libdbus but won't complain if you don\t have the dbus daemon running | 12:40 |
epergny | oh that is very interesting. | 12:40 |
epergny | so it wasn't that silly of a question after all ^^ | 12:40 |
gnarface | so if you're missing dbus, then thunderbird will make you pick a browser when you click on a link from an email, or something like that. you lose very little functionality | 12:40 |
gnarface | conversly firefox will make you pick an email client when you click a mailo: link | 12:41 |
gnarface | *mailto | 12:41 |
sokan | if I get it straight keepassxc (and other backpots packages) are tested but not quite unstable yet? | 12:41 |
epergny | another silly question: can I run Devuan under Stretch inside a Docker container? (just as a test) | 12:41 |
gnarface | barring Docker bugs... | 12:42 |
KatolaZ | sokan: backports are packages not present in the stable release which have been ported there | 12:42 |
KatolaZ | sokan: backports should work without problem | 12:42 |
epergny | gnarface: I should switch back to Emacs for email but somehow I fell on the darkside and I'm using a web-based mail atm. | 12:42 |
epergny | but Docker is a bit special in that it does something special with PID 1: and systemd too. I don't really understand how that plays out. | 12:43 |
gnarface | epergny: someone had done this recently with Docker and ran into trouble with a bug in Docker but i think they found a fix | 12:44 |
epergny | I mean: inside a Docker, PID 1 is not systemd (even if you run Debian stretch say) | 12:44 |
gnarface | the Docker bug might have even been unrelated | 12:44 |
epergny | ok ok | 12:44 |
gnarface | but i remember a discussion about it | 12:44 |
gnarface | i'd recommend qemu-kvm instead | 12:44 |
gnarface | of course | 12:44 |
KatolaZ | epergny: just test it | 12:44 |
epergny | well, at least trying with Docker is really easy. | 12:44 |
epergny | KatolaZ: yeah | 12:44 |
KatolaZ | docker is a moving target | 12:44 |
eyalroz | Would someone be willing to help me out with GLX woes? | 12:48 |
epergny | eyalroz: I would if I could but I simply use the mobo's integrated gpu and all I know about GLX is "glxgears" : ( | 13:08 |
eyalroz | epergny: Thanks for taking the trouble to apologize :-) | 13:09 |
eyalroz | Anyway, a description of my problem appears here: | 13:09 |
eyalroz | https://unix.stackexchange.com/q/254377/34868 | 13:09 |
epergny | and moreover glxgears in full screen on my system shows tearing, so even here it's either not v-synching or not using double-buffering correctly ^ ^ | 13:09 |
eyalroz | And I had it both with Debian Stretch and with Devuan ASCII. | 13:09 |
epergny | but then I've got a 3840x1600 screen so I guess that the mobo's integrated GPU may be in trouble ^ ^ | 13:10 |
eyalroz | @epergny: I don't even want to do anything demanding, I just want the HW acceleration of various run-of-the-mill desktop stuff to not fail. | 13:10 |
eyalroz | Like, Cinnamon crashes into fallback mode. | 13:10 |
gnarface | eyalroz: which nvidia gpu? | 13:16 |
gnarface | the nvidia-driver pack age is in non-free | 13:16 |
gnarface | *package | 13:16 |
eyalroz | gnarface: GeForce GTX 1050 Ti, but - does it matter? I want to have Xorg completely ignore it... | 13:17 |
gnarface | oh, you're trying to get glx NOT on that? | 13:17 |
gnarface | everything else uses mesa | 13:17 |
gnarface | which is in main | 13:17 |
eyalroz | gnarface: The situation is that I have a GPU, and I have on-board graphics. My monitor is fed from my board's output connector, since I only want to use the GPU for compute work | 13:18 |
eyalroz | but | 13:18 |
eyalroz | the nVIDIA drivers are installed (not the Debian package - I used nVIDIA's CUDA installer for drivers and the toolkit), | 13:19 |
eyalroz | and are loaded, since I do use them | 13:19 |
gnarface | well that shell script shouldn't be used | 13:19 |
eyalroz | and somehow this situation gets a bunch of software mixed up about itself | 13:19 |
gnarface | it runs roughshod over the package dependencies, it doesn't care about your distro | 13:20 |
gnarface | it breaks things | 13:20 |
gnarface | normally you should be using the packages from the repo | 13:20 |
gnarface | but i thought for compute tasks you want to just install CUDA alone, don't you? | 13:20 |
eyalroz | 1. I realize that might be the case; if I know what's possibly broken, I could try and fix it. | 13:20 |
eyalroz | 2. I don't let the setup script change my X configuration | 13:20 |
gnarface | i guess i don't know for sure, i've never been in that situation | 13:20 |
gnarface | but the onboard card, if it is not nvidia, it would be using mesa | 13:21 |
gnarface | you might have to blacklist the nvidia driver or at least make sure xorg isn't loading it | 13:21 |
gnarface | you probably will have to have some amount of manual xorg.conf | 13:21 |
gnarface | which i assume you do | 13:21 |
gnarface | normally it figures out what to do if you are not missing any packages | 13:21 |
eyalroz | gnarface: in my modprobe.d, I have: | 13:22 |
eyalroz | blacklist nouveau | 13:22 |
eyalroz | options nouveau modeset=0 | 13:22 |
eyalroz | blacklist lbm-nouveau | 13:22 |
eyalroz | alias nouveau off | 13:22 |
eyalroz | alias lbm-nouveau off | 13:22 |
gnarface | nouveau is the open source nvidia driver | 13:22 |
gnarface | it ALSO conflicts with the non-free nvidia driver | 13:22 |
gnarface | but in this case you may need to blacklist both of them | 13:22 |
eyalroz | So, it looks like my X server is trying to load the nvidia driver, and the nvidia module, | 13:23 |
eyalroz | and the nouveau module... | 13:23 |
gnarface | yea, so you could actually have more than one problem going on there | 13:23 |
eyalroz | and an "nv" module which it can't find | 13:23 |
gnarface | nv is deprecated, long long gone | 13:23 |
gnarface | just delete that | 13:23 |
eyalroz | I'm looking at the log output; not sure where to delete things. | 13:24 |
gnarface | but it would have conflicted with nouveau and nvidia's official drivers as well | 13:24 |
gnarface | no i mean just delete it from your xorg.conf | 13:24 |
eyalroz | I don't have an xorg.conf file, I think it's auto-generated. | 13:24 |
gnarface | so it stops trying to load it | 13:24 |
gnarface | oh hmmm | 13:24 |
eyalroz | I remember there's something I can do to generate it into a file, and for X to default to loading that | 13:24 |
eyalroz | but I forgot how that works | 13:24 |
gnarface | well it does try to auto-config when it can't find a xorg.conf, that is the default behavior now. but it's not very smart and it tends to fail with multi-head setups | 13:24 |
gnarface | my guess is the closed-source nvidia driver loads first and then tries to take over | 13:25 |
gnarface | but you're probably also missing some mesa glx stuff | 13:25 |
gnarface | what is the onboard card? | 13:25 |
eyalroz | just a sec. | 13:25 |
eyalroz | " Intel Corporation Device 5912 (rev 04)" | 13:26 |
gnarface | well it's intel | 13:26 |
eyalroz | it's on a Z170 Intel board | 13:26 |
gnarface | it can probably make at least limited use of mesa | 13:26 |
eyalroz | I think it uses the i915 kernel module | 13:26 |
gnarface | yea that should already be present, but mesa probably isn't | 13:27 |
gnarface | check for packages: mesa-utils, mesa-va-drivers, libgl1-mesa-lgx | 13:28 |
eyalroz | Ok, so if I do: | 13:28 |
eyalroz | Xorg :1 -configure | 13:28 |
eyalroz | it creates a new xorg.conf file for me. | 13:28 |
gnarface | i mean libgl1-mesa-gls | 13:28 |
gnarface | i mean libgl1-mesa-glx | 13:28 |
gnarface | sigh | 13:28 |
gnarface | sorry it's late | 13:28 |
gnarface | yea you'll probably have to edit that xorg.conf by hand a bit | 13:28 |
eyalroz | All three of these packages are installed | 13:29 |
gnarface | oh hmmm | 13:29 |
eyalroz | let's me reinstall them just to be on the safe side | 13:29 |
gnarface | oh i wonder if you're missing firmware-linux-nonfree | 13:29 |
eyalroz | is there a way to verify the installation of a package? i.e. check if all the files are in place? | 13:29 |
gnarface | only by hand, that i know of | 13:29 |
gnarface | but you can reinstall them | 13:29 |
eyalroz | So, installing firmware-linux-nonfree now and reinstalling the other 3. | 13:30 |
gnarface | have you seen this? https://wiki.debian.org/Mesa | 13:30 |
eyalroz | then I'll go read that xorg.conf, edited some stuff out, and then get disconnected when I bring X down then up again. | 13:30 |
gnarface | that wiki page for mesa from debian doesn't contain any systemd specific info | 13:31 |
gnarface | should be still relevant for you | 13:31 |
gnarface | oh and it can be a permissions error, depending on your setup | 13:31 |
gnarface | (xorg doesn't run setuid root anymore) | 13:31 |
gnarface | (not by default, anyway) | 13:31 |
eyalroz | gnarface: Wouldn't I see permissions errors in the Xorg log then? | 13:32 |
eyalroz | And - Xorg _does_ load, it's just other things that have weird failures. | 13:32 |
eyalroz | Like Cinnamon crashes to fallback mode, | 13:33 |
eyalroz | and the Libreoffice spalsh screen usually being black | 13:33 |
eyalroz | and so on. | 13:33 |
eyalroz | gnarface: The xorg.conf that was generated for me has _no_ lines regarding nvidia at all. | 13:33 |
eyalroz | But | 13:34 |
eyalroz | it does have two Device sections, one with "intel" for the driver and one with "nouveau". I'll try removing the second. | 13:34 |
eyalroz | Well, wish me luck... | 13:37 |
eyalroz | gnarface: So here's what happened. When I removed the nouveau driver, and an associated screen and monitor - X would not load. It said that no screens were found. | 13:44 |
eyalroz | So on one hand, it seems like it's using the nouveau driver. On the other handm this doesn't make sense, since the signal gets to the monitor from the motherboard. | 13:44 |
gnarface | too many possibilities still | 13:49 |
gnarface | eyalroz: the nouveau driver should be blacklisted though.... mabye comment out all the lines in your modprobe.d files referring to it other than the one that says "blacklist nouveau" and comment out the lines referring to it in xorg.conf | 13:51 |
gnarface | and make sure you're looking at the right Xorg.0.log | 13:51 |
gnarface | if you're running it as root or suid root it will be in the usual place in /var/log/ but otherwise it will be in ~/.local/share/xorg/ | 13:52 |
gnarface | then pastebin that xorg log for me, i'll look at it | 13:53 |
eyalroz | gnarface: So, comment out just that one xorg.conf line which says what the driver is? That doesn't sound right... | 14:00 |
eyalroz | And I think I'm running X as root, since the only Xorg log is at /var/log | 14:01 |
gnarface | eyalroz: no, not just that line, the whole stanza | 14:13 |
gnarface | the whole "Device" section associated with it | 14:14 |
gnarface | there should only be one driver referred to in each Device section | 14:14 |
gnarface | you will need one for each card | 14:15 |
gnarface | but be careful not to have multiple for any card | 14:15 |
gnarface | nouveau, nv, and the official nvidia drivers are all mutually exclusive | 14:15 |
eyalroz | gnarface: so, should I use the nvidia driver for the second device? | 14:21 |
eyalroz | oh, sorry, you said delete that whole section. | 14:22 |
eyalroz | ok, but it'll fail like before. | 14:22 |
eyalroz | doing this now. | 14:22 |
gnarface | eyalroz: well, for the second device, try it both ways | 14:25 |
gnarface | eyalroz: but it's probably easier to debug the onboard video card alone than to debug them both at once | 14:25 |
gnarface | so for the first try, just comment out the nvidia one | 14:26 |
gnarface | yes, when you get it working, you'd have two "Device" sections, one for the intel card and one for the nvidia card | 14:26 |
gnarface | make sure there is also a "Screen" section and a "Monitor" section | 14:26 |
poeinklum | gnarface: Initial X server log (no xorg.conf): http://paste.debian.net/1073346/ | 14:29 |
gnarface | you're also eyalroz? | 14:29 |
eyalroz | Sorry. | 14:29 |
eyalroz | gnarface: X server log with the edited xorg.conf: http://paste.debian.net/1073347/ | 14:29 |
eyalroz | The edited xorg.conf: http://paste.debian.net/1073348/ | 14:30 |
gnarface | the first one is just a basic failure of the intel device i've seen before - it tries to load a framebuffer driver and chokes | 14:30 |
eyalroz | gnarface: But X does load, somehow, with the first one. Not with the second one. | 14:31 |
gnarface | yea but that's why your compositor fails | 14:31 |
gnarface | because it's just in fallback mode | 14:31 |
gnarface | like with a generic driver | 14:31 |
gnarface | [347488.301] (EE) No devices detected. | 14:32 |
gnarface | did you add yourself to the "video" group? | 14:33 |
gnarface | there might be a missing package or two still... it's right on the tip of my brain... | 14:33 |
gnarface | make sure you're in the video group | 14:34 |
gnarface | for your second try, the driver clearly loads and can't find any devices | 14:36 |
gnarface | but i think that might just be a normal permissions thing as easily as a missing xorg package | 14:36 |
gnarface | i'm booting an intel laptop for reference, stand by | 14:36 |
eyalroz | checking the video group | 14:38 |
eyalroz | also, use my nickname so I know you've replied... | 14:38 |
gnarface | thanks, that's confusing when it starts changing around | 14:39 |
eyalroz | Yes, I'm in the video group | 14:39 |
gnarface | alright... can you pastebin me the output of: dpkg -l |grep xorg | 14:39 |
eyalroz | sure | 14:42 |
eyalroz | http://paste.debian.net/1073354/ | 14:43 |
gnarface | hmmm | 14:44 |
gnarface | xserver-xorg-video-intel is present | 14:44 |
gnarface | xserver-xorg-legacy does suggest you'd be running it suid root | 14:44 |
gnarface | damn | 14:44 |
gnarface | i guess there's no guarantee this card is supported | 14:45 |
gnarface | i don't think they all are, but it's been a really really long time since i ran into one that wasn't | 14:45 |
gnarface | what do you have for: dpkg -l |grep intel | 14:45 |
eyalroz | hmm | 14:46 |
eyalroz | 6 things actually | 14:46 |
gnarface | is libdrm-intel1 in there? | 14:47 |
eyalroz | http://paste.debian.net/1073355/ | 14:47 |
eyalroz | yes, it is in there | 14:47 |
gnarface | that looks normal | 14:48 |
gnarface | i must be missing something still | 14:48 |
gnarface | wait | 14:48 |
gnarface | is this from within Docker? | 14:48 |
eyalroz | I don't use docker, ugh. | 14:49 |
eyalroz | why do you ask? | 14:49 |
gnarface | sorry, you weren't the guy who was gonna try this in Docker first, that was the other guy i guess | 14:49 |
* eyalroz shrugs | 14:50 | |
* eyalroz doesn't know about other guys | 14:50 | |
gnarface | can you check ps to make absolutely sure that xorg is running as suid root? | 14:50 |
gnarface | on my intel laptop here i'm actually not using it that way | 14:50 |
gnarface | i got it running as the current user | 14:50 |
eyalroz | gnarface: It's running as root, yes. | 14:50 |
eyalroz | root 25556 3.5 1.5 718924 254700 tty7 Ssl+ 15:24 0:54 /usr/lib/xorg/Xorg :0 -seat seat0 -auth /var/run/lightdm/root/:0 -nolisten tcp vt7 -novtswitch | 14:51 |
eyalroz | I really don't remember choosing this to be the case though. | 14:51 |
gnarface | it's because you have that xserver-xorg-legacy package | 14:51 |
gnarface | but it should work either way | 14:52 |
gnarface | do you also have the firmware-linux-free package? | 14:53 |
gnarface | check for that | 14:53 |
eyalroz | Yes | 14:53 |
eyalroz | but | 14:54 |
eyalroz | it looks like that package is empty | 14:54 |
gnarface | hmmm | 14:54 |
eyalroz | it just has stuff in /usr/share/doc/firmware-linux-free | 14:54 |
eyalroz | a copyright and a changelog | 14:54 |
gnarface | oh that's because the actual firmware i think it's downloading from somewhere on the fly after install | 14:54 |
gnarface | or generating it somehow | 14:54 |
gnarface | dirty tricks, i know | 14:55 |
gnarface | hmmm | 14:56 |
gnarface | ok at this point i have two guesses | 14:56 |
gnarface | first i need a clarification: was the nvidia kernel module loaded this whole time while testing? | 14:56 |
eyalroz | Yes | 14:57 |
eyalroz | Umm | 14:58 |
eyalroz | I mean, I'm pretty sure it was. | 14:58 |
eyalroz | Can check for extra safety I guess | 14:58 |
gnarface | blacklist it AND the nouveau driver, comment out the nvidia-related sections in the xorg.conf, then reboot | 14:58 |
gnarface | if that doesn't work, i'm blaming it on the nvidia shell script installer | 14:58 |
eyalroz | gnarface: Ok, but I do need it running though. | 14:58 |
gnarface | understood, but one thing at a time here | 14:58 |
* eyalroz proceeds to edit scripts | 14:58 | |
gnarface | the nvidia shell script installer may very well have replaced key files belonging to some of these other packages, which might be choking it | 14:59 |
eyalroz | gnarface: should I just blacklist nvidia, or also other nvidia_XXX modules? | 14:59 |
gnarface | there are others????? | 14:59 |
eyalroz | nvidia_uvm, nvidia_drm, nvidia_modeset | 14:59 |
eyalroz | and a module which only nvidia depends on: ipmi_msghandler | 15:00 |
gnarface | oh, yea, those. uh, they shouldn't load if you've blacklisted the main one. don't bother unless they do | 15:00 |
eyalroz | Yes, you're right. Ok, here goes nothing. | 15:00 |
gnarface | good luck | 15:00 |
sokan | how does sysvinit work? | 15:01 |
sokan | I'm used to `rc-service sshd start` and so on | 15:02 |
eyalroz | So, my "blacklist nvidia" line was apparently ignored | 15:05 |
eyalroz | since when X came up, the module was loaded | 15:05 |
gnarface | yea, did you remove the other lines referring to that in there? | 15:07 |
gnarface | when you pasted it, i noticed that *after* the "blacklist nvidia" line you had options nvidia .... | 15:08 |
gnarface | i don't know if that'll just override the previous blacklist line | 15:08 |
gnarface | in all my files, it comes last in the file and alone | 15:08 |
gnarface | nothing else referring to nvidia | 15:08 |
gnarface | oh | 15:08 |
gnarface | hmmm | 15:08 |
gnarface | you didn't tell me the name of the file either... at some point they changed it so only files ending with ".conf" are acknowledged in /etc/modprobe.d | 15:09 |
gnarface | you're aware of that, right? | 15:09 |
eyalroz | I added it in a file which the nvidia installer created, | 15:09 |
eyalroz | which was originally intended to blacklist nouveau | 15:10 |
eyalroz | /etc/modprobe.d/nvidia-installer-disable-nouveau.conf | 15:10 |
gnarface | right | 15:10 |
eyalroz | and | 15:10 |
gnarface | comment out the lines in there that aren't the blacklist ones | 15:10 |
eyalroz | The X log says: | 15:10 |
eyalroz | ernel command line: BOOT_IMAGE=/vmlinuz-4.9.0-8-amd64 root=UUID=acad7fc8-7ba2-4c41-a0d9-ba73f5a4a87c ro quiet nomodeset nouveau.blacklist=1 nouveau.modeset=0 | 15:10 |
eyalroz | ... but somehow no nvidia.blacklist=1 | 15:11 |
gnarface | those are set elsewhere | 15:11 |
gnarface | look in /etc/defaults/grub | 15:11 |
eyalroz | Yeah. | 15:11 |
eyalroz | It's there. | 15:11 |
eyalroz | I could blacklist through there...? | 15:11 |
gnarface | yes, it should work as far as i know but i can't say from first-hand experience | 15:12 |
gnarface | also worth a try | 15:12 |
eyalroz | Ok, trying again. | 15:12 |
eyalroz | gnarface: X rudely ignored my kernel line... and loaded the nvidia module. | 15:18 |
gnarface | this is a fully up-to-date ascii install, right? | 15:18 |
eyalroz | I apt-get dist-upgrade'd about a week ago I think | 15:19 |
gnarface | well, this is an entirely separate problem we need to get to the bottom of before anything else is gonna work | 15:19 |
gnarface | but at this point i'm running out of ideas other than maybe try the live image, see if it behaves better. if it does, i'm pretty sure the nvidia shell script installer must have broke something here | 15:19 |
eyalroz | I don't want to take so much of your time if it's inconvenient. I mean, I know this isn't a very typical setup and everything. | 15:20 |
eyalroz | Where's the ASCII live image? | 15:20 |
gnarface | stand by i'll find you a link | 15:20 |
gnarface | it's more just that i've been up all night and my mental usefulness is waning | 15:20 |
gnarface | https://files.devuan.org/devuan_ascii/desktop-live/ | 15:20 |
gnarface | live images in here | 15:20 |
eyalroz | Do you think the module could be loaded through the initrd.img maybe? | 15:23 |
gnarface | hmmm. possibly | 15:23 |
gnarface | that's what the initrd.img is for | 15:23 |
eyalroz | gnarface: This says the same: https://unix.stackexchange.com/a/90692/34868 | 15:24 |
eyalroz | I'll try that before the live image. | 15:24 |
gnarface | oh, right | 15:24 |
gnarface | because you used the shell script installer, it doesn't trigger the update-initramfs automatically either | 15:24 |
gnarface | that makes sense | 15:24 |
gnarface | i also note that it seems support for the 5912 card first appears in kernel 4.5, which is not so much older than the stock ascii kernel | 15:25 |
gnarface | so as a last shot if the live image doesn't work, maybe try the kernel from backports | 15:25 |
gnarface | it could just be a bug in the driver | 15:25 |
gnarface | but that's pretty unlikely | 15:26 |
gnarface | that nvidia shell script has a history of mangling stuff | 15:26 |
gnarface | i would say it is the prime suspect at this point | 15:26 |
eyalroz | gnarface: No such luck. Downloading the image now. | 15:28 |
gnarface | damn man, alright | 15:28 |
gnarface | good luck | 15:28 |
gnarface | something is definitely off if it's ignoring all those changes.... | 15:29 |
gnarface | there's just too many things it can be for me to remember them all | 15:29 |
gnarface | i suppose it's probably worth checking the output of dmesg for any weird errors at boot or after xorg startup | 15:30 |
gnarface | but we got all the firmware you might need so i can't even think of what | 15:31 |
eyalroz | Not seeing any weird errors | 15:34 |
eyalroz | in dmesg I mean | 15:34 |
gnarface | ok | 15:34 |
eyalroz | Ah, oh | 15:35 |
eyalroz | there's: | 15:35 |
gnarface | once you get it obeying you, it might just come down to passing the right PCI BusID to each "Device" stanza in the xorg.conf | 15:35 |
eyalroz | nvidia: unknown parameter 'blacklist' ignored | 15:35 |
gnarface | heh | 15:35 |
eyalroz | I guess it doesn't work that way. | 15:35 |
gnarface | ok so i guess that means maybe the official binary driver doesn't pay attention to the kernel command-line like that | 15:35 |
eyalroz | and nouveau is willing to "blacklist itself" | 15:35 |
gnarface | well, they should both obey the /etc/modprobe.d files | 15:36 |
eyalroz | But who's the culprit? It's not the modules that each have to look themselves up in modprobe.d , right? It must be something else | 15:36 |
gnarface | the culprit still could be that you have nvidia driver options *after* the "blacklist nvidia" line in modprobe.d | 15:37 |
gnarface | and you may need to run that update-initramfs line again after changing those files too | 15:38 |
gnarface | i don't specifically remember having a problem with that but i've also not gone this direction with these changes before | 15:39 |
gnarface | usually i'm blacklisting nouveau, or *completely uninstalling the nvidia packages* (something the shell script can't cleanly do) | 15:39 |
gnarface | i've never actually tried to make an intel and nvidia card coexist in the same machine | 15:40 |
eyalroz | I did rerun update-initramfs | 15:40 |
eyalroz | Do you really think the order of blacklist directives could matter? That seems rather unlikely. | 15:40 |
gnarface | to be clear | 15:41 |
gnarface | no | 15:41 |
gnarface | but i think the order of the *options* directives in respect to them might | 15:42 |
gnarface | having options for the nvidia driver *after* the blacklist nvidia line might just actually nullify the blacklist line | 15:42 |
gnarface | i AM however POSITIVE that it's not worth it to specifically avoid testing that theory | 15:43 |
eyalroz | I commented-out the options lines in that file after the blacklist directives, so hopefully that shouldn't be an issue | 15:45 |
gnarface | if they're commented out with a # then it should be fine | 15:45 |
gnarface | and the entire nvidia Device stanza is commented out in the xorg.conf right? | 15:46 |
gnarface | config snippets in theory could be conflicting with the xorg.conf in /etc/X11/xorg.conf.d or /usr/share/X11/xorg.conf.d, but the packages in the repo would not have done that. i can't tell you for sure the shell script wouldn't either. | 15:47 |
eyalroz | Not right now it isn't. I'm dumb. | 15:47 |
eyalroz | Although - it's a nouveau stanza | 15:48 |
eyalroz | gnarface: Brought the system up via liveimage | 16:02 |
eyalroz | Will send a Xorg.0.log paste soon. | 16:02 |
gnarface | any different? | 16:02 |
eyalroz | Yes. | 16:03 |
eyalroz | https://paste.debian.net/1073371/ | 16:03 |
eyalroz | Well, sort of | 16:03 |
eyalroz | nouveau is loaded | 16:03 |
eyalroz | But not mentioned in Xorg.0.log | 16:03 |
eyalroz | ... and it can be unloaded, meaning it's not used. | 16:04 |
gnarface | hmmm | 16:06 |
eyalroz | If I now generate an xorg.conf, it looks just like before: An "Intel" device section and a "nouvea" device section. | 16:07 |
eyalroz | And: | 16:08 |
eyalroz | # glxinfo | grep "OpenGL render" | 16:08 |
eyalroz | OpenGL renderer string: Mesa DRI Intel(R) Kabylake GT2 | 16:08 |
gnarface | Note: Some (Debian & Ubuntu, Fedora, KDE) recommend not installing the xf86-video-intel driver, and instead falling back on the modesetting driver for fourth generation and newer GPUs. See [1], [2], Xorg#Installation, and modesetting(4). However, the modesetting driver can cause problems such as Chromium Issue 370022. Also, the modesetting driver will not be benefited by Intel GuC/HuC/DMC firmware. | 16:09 |
gnarface | (that's from the arch linux wiki but i think that's what is going on here) | 16:09 |
gnarface | it is falling back on kms after fbdev and vesa fail | 16:10 |
gnarface | no evidence it tried the intel driver | 16:11 |
eyalroz | gnarface: Is it actually failing to load those? | 16:11 |
eyalroz | It doesn't report failure | 16:11 |
gnarface | well it always tries them first. they just load then unload normally | 16:12 |
gnarface | so this looks in better shape than your install | 16:12 |
gnarface | but try it with the intel driver now | 16:12 |
eyalroz | so apt-get install xserver-xorg-video-intel ? | 16:13 |
gnarface | if it's not already there, yea | 16:13 |
eyalroz | already installed. | 16:13 |
eyalroz | How do I force that? | 16:13 |
gnarface | don't | 16:14 |
gnarface | just use the xorg conf | 16:14 |
eyalroz | Ah. ok. | 16:14 |
gnarface | i frankly don't know if it will work better for you than KMS is now | 16:16 |
gnarface | this will be a learning experience for both of us | 16:16 |
gnarface | don't forget that the graphical login manager will also have to be restarted now | 16:17 |
gnarface | or it will feed the old xorg config to the new xorg instance and ignore your new xorg.conf | 16:17 |
eyalroz | I remembered I can just keep the IRC session live on my laptop... | 16:19 |
gnarface | heh | 16:19 |
eyalroz | https://paste.debian.net/1073373 | 16:20 |
eyalroz | This is with the nouveau section commented-out | 16:20 |
gnarface | there we go | 16:21 |
gnarface | that looks right | 16:21 |
gnarface | note all these lines: (II) intel(0) | 16:21 |
gnarface | those are the driver initializing | 16:22 |
eyalroz | Yes, I noticed those | 16:22 |
gnarface | as far as i can tell it's working right now | 16:22 |
gnarface | how does it actually look? | 16:22 |
eyalroz | Well, it's running Xfce, so I can't check whether cinnamon crashes | 16:22 |
eyalroz | but I'll do that soon | 16:22 |
eyalroz | Other than that - it's not like I can tell. | 16:23 |
eyalroz | If glxinfo doesn't complain, I'm good I guess. | 16:23 |
gnarface | vblank_mode=0 glxgears | 16:23 |
gnarface | glxinfo |grep direct.rendering | 16:23 |
gnarface | glxinfo |grep renderer | 16:24 |
gnarface | that type of stuff | 16:24 |
eyalroz | Cinnamon works with HW rendering. | 16:25 |
gnarface | glxgears isn't for benchmarking, but you will still normally see a notable framerate difference between it in hardware rendering mode vs software rendering mode on modern hardware, as long as vsync is disabled | 16:25 |
gnarface | hooraaaaay! | 16:25 |
gnarface | ok | 16:25 |
eyalroz | I see the gears | 16:25 |
gnarface | cool | 16:25 |
eyalroz | about 9760 FPS | 16:26 |
gnarface | so, what did we learn here? | 16:26 |
gnarface | this is the quiz section of the class | 16:26 |
eyalroz | That the introduction of the nvidia driver - and possible other things I've done - have messed stuff up on my system, | 16:26 |
gnarface | yea, that's pretty much it | 16:26 |
eyalroz | so that the intel driver doesn't get loaded, and the nvidia driver does | 16:26 |
gnarface | more succinctly i would simply say that the nvidia shell script crippled your install | 16:27 |
gnarface | you might have thrown a couple other monkey wrenches in there with it | 16:27 |
gnarface | so the lesson to be learned is, don't use that shell script | 16:27 |
gnarface | however | 16:27 |
eyalroz | Ok, but: | 16:28 |
eyalroz | 1. Xorg still has the correct opinion about what hardware is available | 16:28 |
eyalroz | 2. Since I'm not running the shell repeatedly, it should be possible to force X to load what we want it to load | 16:28 |
gnarface | right | 16:28 |
gnarface | so as i was about to say | 16:28 |
gnarface | you might be able to make your existing install behave if you explicitly pass the correct BusID for each video card to the corresponding "Device" stanza in your xorg.conf | 16:28 |
gnarface | now, you're gonna have to find the correct PCI bus ids for both of those on your own, i'm out of steam here | 16:29 |
gnarface | i'll be back eventually though | 16:29 |
gnarface | but right now i've just gotta sleep | 16:29 |
eyalroz | But it's the _same_ device stanza both with the live image and in my original system | 16:30 |
gnarface | actually i think that last xorg log you just pasted from the live image even listed both the bus IDs | 16:30 |
eyalroz | and the same BusID value AFAICR | 16:30 |
gnarface | BusID | 16:30 |
gnarface | well this is my only theory as to a correctable issue | 16:30 |
gnarface | if that's not it, the nvidia script did something to your install that i can't undo | 16:30 |
gnarface | (or... unspecified other changes... ) | 16:31 |
eyalroz | So, do you suggest I go through the nvidia shell script and see what it does? | 16:32 |
gnarface | but the live image proves it should work | 16:32 |
gnarface | no, i don't. but if it sounds fun to you feel free | 16:33 |
gnarface | TBH what i did when i realized how badly behaved it was, was i reinstalled from scratch, using only packages from the repo | 16:34 |
gnarface | and then i made a mental note to NEVER use the shell script from nvidia.com again | 16:34 |
gnarface | but you gotta be true to yourself, ya know? | 16:34 |
eyalroz | but - the Devuan repos don't have the CUDA versions I need, | 16:36 |
eyalroz | and the newer versions can't work with older drivers. | 16:36 |
gnarface | not even in backports? | 16:36 |
eyalroz | Actually, just the second point is the problem. | 16:36 |
eyalroz | Didn't check bacports, TBH. | 16:36 |
specing | > reinstalled | 16:36 |
gnarface | check backports | 16:36 |
specing | no btrfs? | 16:36 |
gnarface | eyalroz: if the one you need is no in backports, consider doing the CUDA stuff from inside a chroot - you can debootstrap an entire install into the chroot, and then probably inside that the nvidia shell script can't break your base install (probably) | 16:37 |
gnarface | heh, and yea, i'd advise making a backup first | 16:37 |
eyalroz | gnarface: Not sure if this isn't more pain than what I'm experiencing right now. | 16:38 |
specing | you should always have backups | 16:38 |
gnarface | well, as a long-time customer of nvidia, i can say i'm pretty much sick of their shit and this is a perfect case study in such shit | 16:38 |
specing | but btrfs snapshots are not proper backups, they are snapshots | 16:38 |
specing | and they are highly useful in OS-is-broken situations | 16:39 |
specing | especially as you can reboot straight into them | 16:39 |
eyalroz | gnarface: I can't find any references to backports on pkginfo.devuan.org | 16:39 |
specing | gnarface: NVidia is fine, they make best graphics for Linux | 16:39 |
gnarface | eyalroz: no it works like debian backports | 16:39 |
gnarface | eyalroz: ascii-backports | 16:39 |
gnarface | (instead of stretch-backports) | 16:40 |
* eyalroz scratches head to jog /me's memory | 16:40 | |
eyalroz | I need an apt source line for it I guess | 16:40 |
gnarface | yea | 16:40 |
_abc_ | Hello. Is there an efficient way to locate packages and package by file from the devuan cli? I have dlocate installed, does not help with that. | 16:40 |
gnarface | _abc_: apt-file might still be broken | 16:41 |
gnarface | i haven't checked in a while | 16:41 |
eyalroz | _abc_: How about dpkg -S ? | 16:41 |
gnarface | eyalroz: (that only works for currently installed packages) | 16:41 |
fsmithred | apt-file works | 16:41 |
_abc_ | That searches installed packages, I need to search all packages, obviously not installed | 16:41 |
_abc_ | Another: sometimes my middle mouse click (button3) paste is out of whack on X11 xfce on devuan ascii. Where does one fix this? | 16:42 |
_abc_ | Clipboard and middle button are out of sync. | 16:42 |
fsmithred | eyalroz, just copy your ascii main line and change ascii to ascii-backports | 16:44 |
_abc_ | apt-file found what I need but package xinput is not in the list offered by aptitude. | 16:44 |
* gnarface afk | 16:44 | |
_abc_ | apt-file search xinput ... xinput: /usr/bin/xinput | 16:44 |
eyalroz | fsmithred: Ok. TBH though, I doubt this will let me recover from the "horrors" of the nVIDIA shell script. On a fresh install perhaps. | 16:44 |
_abc_ | conclusion: aptitude is out of whack | 16:45 |
man_in_shack | who does what now? | 16:45 |
eyalroz | fsmithred: Oh, it seems I already _have_ backports lines in my sources.list:-P | 16:45 |
fsmithred | eyalroz, are you switching from nvidia binaries to nvidia packages or the other direction? | 16:45 |
eyalroz | fsmithred: For now, I'm not switching anywhere. But my current state is nvidia binary installation over _no_ nvidia Devuan packages | 16:46 |
fsmithred | the shell script has an uninstall option | 16:46 |
eyalroz | fsmithred: Probably not very useful when you've run it for different CUDA versions; | 16:47 |
eyalroz | and - I will actually need CUDA 10.0, so I think I'll have to work out how to make the script not bust Devuan. | 16:47 |
_abc_ | Okay, thanks, problems solved so far. The out of sync middle mouse click paste? Any ideas what can bring that out of sync with the clipboard? | 16:48 |
fsmithred | I've never had problems installing nvidia with the shell script except when I was trying to switch from using the packages | 16:48 |
fsmithred | _abc_, no, but if you figure it out, please let us know | 16:48 |
eyalroz | Oh, it installs fine. The problem is that I want to _not_ use it in X | 16:48 |
_abc_ | fsmithred: :) :) | 16:48 |
_abc_ | What was that mechanism in X11 for multiple selection choosing again? | 16:49 |
_abc_ | My current behavior is, the middle click pastes the "previous" selection. | 16:49 |
eyalroz | fsmithred: And right now, I can't prevent the nvidia drivers from being loaded, and X is refusing to properly load the intel drivers and recognize my on-board graphics device; while with a live-image and no CUDA install, X behaves well. | 16:49 |
fsmithred | would setting the vid card in the bios work? | 16:50 |
fsmithred | _abc_, you mean selecting multiple files at once or you mean having multiple clipboards? | 16:52 |
_abc_ | fsmithred: I have one clipboard, one mouse. double clicking a word in a xterm/etc highlights it and the cliboard app shows it current and selected | 16:52 |
fsmithred | triple-click highlights a line | 16:53 |
_abc_ | fsmithred: middle clicking pastes the previously selected word in the clipboard. If I re-select (click on clipboard app on word) the word, and middle, click it will be pasted. If I highlight another with double click, the next middle click again pastes the previous selection, not what I lighted recently. | 16:53 |
eyalroz | fsmithred: Setting in what sense? It's not that X does not recognize the existence of the card - it generates the same xorg.conf for my system both from a live image and with CUDA installed; but somehow, it is not willing to use the intel device and avoid the nvidia one. | 16:53 |
eyalroz | gnarface: Drats! it seems the nvidia installer is actually a binary; and the shell script is just for decompressing the installer before really running it. | 16:54 |
_abc_ | I unplugged and replugged the mouse just to see, it does not help | 16:54 |
_abc_ | Try to use nouveau not nvidia eyalroz | 16:54 |
_abc_ | nvidia is a can of worms as you discovered. | 16:54 |
eyalroz | _abc_: I'm trying to _not_ use any nvidia-card-related display drivers. Unfortunately I have to, at the same time, use the nVIDIA drivers themselves, since I'm using the GPU for compute work. | 16:55 |
_abc_ | nouveau is the open source reverse engineered version | 16:56 |
fsmithred | triple-click then hold down shift and use down arrow highlights multiple lines | 16:56 |
_abc_ | You can always run in VESA mode up to 1024x768 | 16:56 |
fsmithred | he wants to use the on-board intel video for video, nvidia for crunching numbers | 16:57 |
_abc_ | fsmithred: that is not what I do. Simple double click or triple click highlights, appears in clipboard as expected, with arrow pointing at it, but middle paste pastes the /previously/ pasted item | 16:57 |
_abc_ | fsmithred: ah, iirc there is a PCI level switch for that somewhere. I do not remember the details. Maybe in the "run dual screen" settings? | 16:58 |
fsmithred | yeah, you have to right-click and paste | 16:58 |
eyalroz | _abc_: What fsmithred said. | 16:58 |
_abc_ | fsmithred: right click and paste works but middle click direct paste pastes the previous selection, not the current one. | 16:58 |
_abc_ | Wrong, not even right click and Paste pastes the previous selection | 16:58 |
fsmithred | no, that would paste the current selection | 16:59 |
_abc_ | pasteIt should but it does not | 16:59 |
_abc_ | That is the problem | 16:59 |
fsmithred | oh, you have to right-click copy first | 17:00 |
_abc_ | I've had this once before but I am stumped as to how it is possible. X11 has a multiple selection mechanism in the server. | 17:00 |
_abc_ | fsmithred: double click-1 is "copy", single click-3 (middle) is paste | 17:00 |
fsmithred | hm... it's working here. Double-click to highlight a word, go to another tab in text editor and "middle" click (left and right buttons) and it pastes | 17:04 |
_abc_ | I know. It used to work and now it's crazy | 17:05 |
fsmithred | used to work how long ago? | 17:06 |
fsmithred | with a two-button mouse it only works sometimes - have to click both buttons just right | 17:07 |
_abc_ | This morning. | 17:07 |
fsmithred | did you change anything today? | 17:08 |
_abc_ | fsmithred: there's a setting in the xserver, usually settable with a gui option too, to determine the allowed delay for mouse chording | 17:08 |
_abc_ | fsmithred: no, the laptop slept while I was away, and had the problem when I got back after a few hours. | 17:08 |
fsmithred | slept = suspend to ram? | 17:09 |
_abc_ | I'm going to put it to sleep and wake it again, to see. After all, it once ran windows. | 17:09 |
_abc_ | fsmithred: yes | 17:09 |
_abc_ | rebooting heals everything. In theory. | 17:09 |
_abc_ | brb | 17:09 |
fsmithred | ok | 17:09 |
fsmithred | maybe just restart desktop is enough | 17:09 |
_abc_ | I have too many open windows to restart the desktop | 17:10 |
_abc_ | And no it did not help. Drat. | 17:10 |
fsmithred | you in xfce? | 17:10 |
_abc_ | I'll see when I de-clutter the system a bit how I can suss out what causes this. | 17:11 |
_abc_ | Yes | 17:11 |
_abc_ | I think it's in the server not in the desktop | 17:11 |
_abc_ | https://www.jwz.org/doc/x-cut-and-paste.html maybe-relevant | 17:12 |
_abc_ | Ok, I am not going to do this right now, was hoping for a shortcut. | 17:13 |
fsmithred | oh, I need to read that page | 17:13 |
_abc_ | There are several on that theme, that one is decent. | 17:14 |
fsmithred | but first I need to go paint a door frame. bbl. | 17:14 |
_abc_ | https://en.wikipedia.org/wiki/X_Window_selection another view on the same issues. The ICCCM standard linked in the 1st link above is the definitive answer. | 17:17 |
_abc_ | <aside>One wonders whose "free" linux these nice people copied and are flogging for $35+$8P&P with no links to open source / references. https://www.xtra-pc.com/ | 17:27 |
_abc_ | I assume they copied the more popular ubuntu. not devuan related... | 17:27 |
_abc_ | I raised that link on ##electronics yesterday, there was a mini discussion. | 17:27 |
_abc_ | Good news: using double click selection and middle click paste ONCE in firefox FIXED it. | 17:29 |
_abc_ | I assume ff's internals are more carefully debugged than the innards of other tools I am using and it does the "Right Thing" wrt the multiple selection jungle in X11 explained in the links above. | 17:30 |
DonkeyHotei | https://static1.squarespace.com/static/56306979e4b052b6d2eea4c0/t/5639853ae4b0ad72a19f88b4/1446610234316/Source-Code-1.pdf | 17:30 |
_abc_ | to add to the haze fsmithred : https://unix.stackexchange.com/questions/11889/pasting-x-selection-not-clipboard-contents-with-keyboard | 17:32 |
_abc_ | DonkeyHotei: good one. I assume they print it on paper and send it by courier? :) | 17:32 |
* _abc_ will now take a tea and biscottae break | 17:33 | |
_abc_ | re | 17:42 |
_abc_ | I report that the middle click paste in terminals is still not okay, only in firefox | 17:58 |
_abc_ | Meaning, the bug, if any, is in xfcr4-terminal which is reasonable as an assumption. | 17:58 |
_abc_ | Please suggest a way to get the id of a X11 window without user interaction, for a xfce4-terminal, from the terminal running, itself. | 18:08 |
_abc_ | xprop and xwininfo do not seem to have an option to show the "current" window they are running in (as a cli in the terminal) | 18:09 |
KatolaZ | _abc_: lsw | 18:18 |
_abc_ | ? | 18:20 |
KatolaZ | a way to get the id of a X11 window | 18:22 |
KatolaZ | it's in suckless-tools | 18:26 |
_abc_ | xdotool getactivewindow does it | 18:30 |
_abc_ | Hi. Again. Suggest a way to switch (activate) a specific tab in xfce4-terminal preferrably using a cli command, not compiled code? | 18:34 |
_abc_ | just finished my single instance irssi under xfce4-terminal script, with window raise etc: http://paste.debian.net/1073389/ | 19:02 |
_abc_ | http://paste.debian.net/1073397/ fixed, v1, launches it's own xterm | 19:43 |
_abc_ | http://paste.debian.net/1073398/ fixed v1.1 small bash scripting fixes with help from #bash -- final, done | 20:05 |
_abc_ | moving on, bbl | 20:05 |
_abc_ | fsmithred: re: selection and such: xsel is a cli tool which permits access to the x server's selection as it is currently. | 20:48 |
xrogaan | KatolaZ: I just noticed dbus got an update, and that there is no changelog provided for the packages managed by devuan. Is it a limitation of the mirror system? | 20:50 |
xrogaan | I also noticed a difference in version between debian's dbus and devuan's dbus (1.10.26-0+deb9u1 vs 1.10.22-1+devuan2). Are we staying on .22 on purpose? I'm asking because I'm curious :) | 20:52 |
xrogaan | also, why is it so hard to get information through the dyne.org interface? | 20:58 |
fsmithred | _abc_, thanks. I was trying to remember that command. | 20:59 |
golinux | xrogaan: What info are you trying to get that you couldn't get here? | 21:16 |
_abc_ | golinux: likely a normal "list of projects" alphabetical or whatever. I only see a huge page of icons and vague textes describing the items. | 21:21 |
xrogaan | golinux: relative to? If it's about dyne.org, I find everything related to the lists quite confusing. | 21:23 |
golinux | xrogaan: Sorry I wandered off | 22:04 |
golinux | I agree about the Dyne site - not well organized, poor navigation etc. Previous iterations were more accessible. I don't even like to send people to the Dyne site anymore. | 22:06 |
_abc_ | My opinion: they at dyne went windows 10 look and feel. Barf. | 22:19 |
_abc_ | Ok time to go and zzz a bit soon | 22:19 |
golinux | <_abc_> My opinion: they at dyne went windows 10 look and feel. Barf. | 22:35 |
golinux | I can't imagine that anyone at dyne even knows what Win 10 looks like. It's it that stupid mobile design - one long page with lots of images and no info that looks dreadful on a real computer.. | 22:36 |
epergny | golinux: at one point my bank's website went "mobile first" but it was basically "mobile only" and suddenly I had a tiny rectangle for input/ouptut in the middle of a big blank webpage. Was completely silly. | 22:56 |
epergny | golinux: they fixed it since then | 22:56 |
Nematocyst | did an apt-upgrade today and X didn't start. after dbus-x11 and related files updated. startx from shell works, but slim.log gives: slim: open_session: Unable to open session: Failed to execute program org.freedesktop.ConsoleKit: Permission denied | 23:29 |
Nematocyst | ugh. checking for newer updates 'apt update' partially fails with: Error: GDBus.Error:org.freedesktop.DBus.Error.Spawn.ExecFailed: Failed to execute program org.freedesktop.PackageKit: Permission denied | 23:38 |
xrogaan | did you reboot? | 23:39 |
Nematocyst | yeah | 23:39 |
xrogaan | well, I don't know. I use lightdm | 23:39 |
xrogaan | with elogind | 23:40 |
golinux | Nematocyst: Sounds like a session management mismatch. Have you read the release note? | 23:50 |
sokan | Anyone has any idea about a good VPN provider? To cover IP etc | 23:50 |
golinux | Nematocyst: https://files.devuan.org/devuan_ascii/Release_notes.txt | 23:51 |
Nematocyst | golinux, well i see that XFCE uses slim + consolekit. i'm using slim + elogind. but i haven't changed those since installing devuan. just booted after doing updates earlier today. | 23:55 |
Nematocyst | been running ascii since it came out. though possibly i did an upgrade from whatever devuan 1 was called. don't remember | 23:57 |
golinux | Try it with console kit. | 23:58 |
golinux | After all these months I still am confused what session manager needs what backend | 23:59 |
epergny | golinux: that's why I use none ^ ^ | 23:59 |
Nematocyst | golinux, hm. maybe i don't know what i'm using. dmesg has entries that begin with elogind. but 'apt install consolekit' says it is at the latest version | 23:59 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!