OneManHateGroup | It gets all the way up to starting X and then goes blank | 00:05 |
---|---|---|
OneManHateGroup | The livesub starts and works with "nomodeset" boot option | 00:06 |
ServiceRobot | ello ello | 09:24 |
MinceR | i | 09:29 |
ServiceRobot | i? | 09:29 |
MinceR | "hi" without the 'h' | 09:32 |
ServiceRobot | ah, I see. I must be getting old, lingo is getting shorter and shorter these days | 09:32 |
MinceR | :> | 09:32 |
MinceR | you started it :> | 09:32 |
ServiceRobot | yeah, well I'm about to start something else... | 09:32 |
ServiceRobot | a conversation | 09:33 |
ServiceRobot | I can't find any updates on the current status of devuan. I've been following the mailing list but it's a bit of a mess to figure out where this distro is going | 09:34 |
ServiceRobot | I've been using it for quite a while now, and I'm holding out hope more init systems will get better support, even if putting that support would screw up "muh backwards compatibility" | 09:35 |
MinceR | ¯\_(ツ)_/¯ | 09:37 |
ServiceRobot | if you want me to be honest, if you really want this to set itself as different to debian's way of doing things, you got to get a bit dirty. and by that I mean make init systems interchangable | 09:37 |
MinceR | that's probably already a given | 09:38 |
ServiceRobot | but from what I can gather, it's debian but with sysvinit installed by default with certain blacklisted systemd packages | 09:38 |
MinceR | some of them are even packaged, afaik | 09:38 |
MinceR | they may not have scripts/whatever for many packaged services, though | 09:38 |
Evilham | You can also pick OpenRC at install time | 09:38 |
ServiceRobot | that's a biggy I want to help improve | 09:38 |
Evilham | But you are rught that more work has to be done there | 09:39 |
ServiceRobot | in any case, onto my question. has beowulf improved AT ALL this year? | 09:39 |
ServiceRobot | ascii is too old for me | 09:39 |
Evilham | I use beowulf on anything new and have migrated a couple things to it as well | 09:40 |
Evilham | So, yeah, it's quite alright | 09:41 |
ServiceRobot | well, with beowulf comes the introduction to things like openrc-init and runit-init. I want to replace sysvinit completely | 09:41 |
ServiceRobot | and I mean tear sysvinit out | 09:41 |
ServiceRobot | not leave crap packages behind that are needed as "studs" or whatever you call it | 09:41 |
Evilham | :-D it's one of those things, it requires work, doesn't mean it's not desired just hasn't been done to that extent | 09:42 |
ServiceRobot | but has it been in the works, put on the backburner, or at the very least been talked about at all? | 09:43 |
Evilham | Yup | 09:43 |
Evilham | There were also 2 talks about s6 on dev1-conf | 09:44 |
ServiceRobot | well, what was discussed, agreed upon, etc? or are we stuck in limbo? ah okay | 09:44 |
* redrick has nothing against studs, but suspect you mean stubs. | 09:44 | |
ServiceRobot | yeah, stubs | 09:44 |
ServiceRobot | my bad | 09:44 |
Evilham | s6 looks promising :-) | 09:44 |
Evilham | But that requires a lot of work | 09:44 |
redrick | No, don't apologise. It added a deliciously surreal tinge to my evening. | 09:44 |
ServiceRobot | I tried s6 once. didn't like it. I rather use something a bit simpler like runit | 09:45 |
ServiceRobot | but runit support in debian is.... well runit-init support at the least, is awful | 09:45 |
ServiceRobot | also in devuan as well since it uses the same setup | 09:45 |
Evilham | So far, I do too, but then I use sysvijit for init and runit for services | 09:45 |
ServiceRobot | I don't like halfing the work like that. runit or nothing for me | 09:46 |
Evilham | Ysah, tbf I don't think it's possible to make runit a very awesome *init* | 09:46 |
ServiceRobot | I did get it to work without sysvinit scripts though | 09:46 |
Evilham | It is a very awesome service manager and | 09:46 |
ServiceRobot | but the dependencies prevent me from removing it | 09:46 |
Evilham | Not saying it's not doable | 09:46 |
ServiceRobot | which is what needs to be worked on the most I think. making it so I can remove core init packages like sysv-rc, sysvinit-utils, etc | 09:47 |
ServiceRobot | and just use runit instead | 09:47 |
ServiceRobot | and I'm saying this after I've tinked with devuan for a couple months | 09:47 |
Evilham | :-D sure, but someone has to do al that work so you still have a bootable and usable system | 09:48 |
Evilham | And so far there have been more urgent matters | 09:48 |
ServiceRobot | sure sure. getting it bootable wasn't the easiest task, since testing it is such a pain | 09:48 |
ServiceRobot | what are the current urgent matters? | 09:48 |
Evilham | Then you'd understand why there is some reticence to make it an option: people will expectbit to work awesomely out of the box | 09:49 |
ServiceRobot | well I'm not asking all the work done for me. I just think modularity is very important if we're going for init freedom | 09:50 |
Evilham | So: you are right, the state of that is not what it "should". But also: it is something that is desired to accomplish | 09:50 |
Evilham | It just hasn't gotten there yet | 09:50 |
ServiceRobot | but that brings up the question: what are the current roadblocks before init is worked on? | 09:50 |
ServiceRobot | where are we currently, what needs to be done before we can get there, etc? | 09:51 |
Evilham | There was a thread a couple months ago about runit specifically on the ML, look for the archive | 09:51 |
ServiceRobot | I think I started that thread actually | 09:52 |
ServiceRobot | and read the responses | 09:52 |
Evilham | And before that other things have to be done/figured out | 09:52 |
ServiceRobot | it didn't go further after that | 09:52 |
ServiceRobot | I'm curious to know exactly what these "things" are. a roadmap of sorts | 09:53 |
Evilham | :-D why don't you come back to this in about a week? | 09:54 |
Evilham | Please do, really :-) | 09:54 |
ServiceRobot | come back to what? | 09:55 |
Evilham | To this exact conversation | 09:55 |
ServiceRobot | why in a week? | 09:55 |
Evilham | Because it's not too much time that it can't wait, yet enough for other matters to be figured out and this to make more sense | 09:59 |
ServiceRobot | huh, this must be a busy week then | 09:59 |
ServiceRobot | so I guess next tuesday then? | 10:00 |
Evilham | Yup | 10:01 |
jiefk | Hum... I updated my laptop with ASCII and Beowulf in sources.list. Now I can't get sddm, not lightdm to start (I installed task-kde-desktop metapackage). | 19:29 |
gnarface | why would you do that? | 19:30 |
jiefk | For minimalism, I installed fluxbox also and tried startx startfluxbox, and got some strange error : | 19:30 |
gnarface | why would you use both repos? | 19:30 |
gnarface | they're not meant to be mixed and matched | 19:31 |
gnarface | it's supposed to be a one or the other type of thing | 19:31 |
jiefk | Well I don't know actually. I was under the impression that one could complete the other (especially when Beowulf was very young) | 19:32 |
jiefk | I'll tri to comment ascii then, thanks :) | 19:32 |
jiefk | try* | 19:32 |
gnarface | there might have been a situation earlier where some needed patches had only been pressed into ascii but that should be considered an anomalous situation, not expected behavior | 19:33 |
jiefk | I just tried to apt update/upgrade/dist-upgrade, no packages were added nor removed. | 19:34 |
gnarface | in general you should use ascii, then only mix in some newer packages from ascii-backports as necessary. beowulf is only advised if you're interested in contributing to testing | 19:35 |
gnarface | that said, the issue you're having in beowulf might be fixable by manually adding or removing the right packages | 19:37 |
gnarface | i couldn't tell you off the top of my head which ones, but make sure that you're not holding both those permissions backend packages (consolekit/elogind?) that are mentioned in the ascii release notes as mutually exclusive | 19:38 |
jiefk | Well I'll check that, thanks | 19:38 |
jiefk | Can someone help me please ? I have no more screen on the laptop. My laptop (on Beowulf) starts fine until -supposely- X starts | 21:49 |
jiefk | From then the screen turn off (not even backlight is on) | 21:49 |
jiefk | I still can login via SSH | 21:49 |
gnarface | pastebin xorg log? | 21:50 |
gnarface | or, just try it without compositing first | 21:50 |
jiefk | Even on Recovery mode, i have no access to TTY | 21:50 |
jiefk | I'll try to get a Xorg.log | 21:50 |
gnarface | there could be a lot of things causing it, but usually it is some sort of driver issue | 21:53 |
jiefk | https://privatebin.net/?c7fb0128ffd6434b#OeOM2866+iZPDD8zz3bWlJzIh12odpPAhllwVVPFNDw= | 21:53 |
jiefk | ^ Xorg.0.log | 21:54 |
gnarface | uh, could you use paste.debian.net please? i hate having to risk attaching to a new domain every time i want to help someone new | 21:54 |
jiefk | ok, I'll upload it there :) | 21:54 |
gnarface | there's no ads on paste.debian.net | 21:54 |
debdog | do you by chance have a monitor at hand? jiefk. if so, do you see something if you plug it in? | 21:55 |
jiefk | an no sorry debdog I have no monitor at hand. :/ | 21:56 |
jiefk | http://paste.debian.net/1080935/ | 21:57 |
jiefk | ^ gnarface | 21:57 |
gnarface | jiefk: yea, it's using the wrong driver | 21:58 |
jiefk | debdog: if I find a VGA cable, I might be able to connect to another screen by tomorrow :) | 21:58 |
gnarface | jiefk: you might have other problems too, but you're not gonna get to the bottom of them while you're using that driver | 21:58 |
jiefk | gnarface: which line ? for the wrong driver ? | 21:59 |
gnarface | jiefk: pretty much every line matching the substring "modeset(0)" | 22:00 |
gnarface | jiefk: that's a generic driver, a sorta new one. | 22:00 |
gnarface | this line here might be the smoking gun: [ 878.661] (EE) modeset(0): failed to set mode: Invalid argument | 22:01 |
gnarface | it's your only error, basically telling you it couldn't find a valid resolution | 22:01 |
gnarface | everything else seemed to work right | 22:01 |
gnarface | up to that point | 22:01 |
jiefk | Before it happens, I remember having installed installed elogind, this action removed libsystemd0 I think ... Is there some king of consequence ? | 22:02 |
gnarface | i just happen to know that this driver also doesn't belong to any specific hardware. you most likely want to force it to try the intel or nouveau driver instead, whichever is appropriate for your hardware | 22:02 |
gnarface | load the right driver first, then we'll see about possible permissions problems. elogind shouldn't be relevant to running xorg with startx though, i thought that was just for graphical logins | 22:04 |
gnarface | i've seen it cause a red herring that looks something like this, but that also amounts to it basically loading the wrong driver | 22:05 |
gnarface | this was beowulf, right jiefk? | 22:05 |
jiefk | gnarface yes beowulf | 22:06 |
gnarface | yea, force it to use the right driver first. there's too many unanswered questions without that. | 22:06 |
jiefk | gnarface: here is the output of `X -configure` | 22:06 |
jiefk | http://paste.debian.net/1080938/ | 22:06 |
gnarface | it thinks you should use the intel driver. i trust it. | 22:07 |
* debdog is astonished that actually created a .conf file | 22:07 | |
jiefk | yes, IIRC, it's a i915 chipset | 22:07 |
gnarface | you'll probably also want to make sure you've got mesa installed | 22:08 |
gnarface | and maybe some libva packages too, for video acceleration | 22:08 |
gnarface | however this should work even without those things as long as your WM isn't using compositing | 22:08 |
jiefk | oh crap :p, I already issued init 6 :p | 22:09 |
jiefk | But should the system start normally to TTY in recovery mode ? | 22:09 |
gnarface | i'm not sure. maybe | 22:09 |
gnarface | even if this new config doesn't work, you should at least get a more relevant error | 22:10 |
jiefk | the screen still turns off | 22:10 |
gnarface | pastebin the new xorg log | 22:11 |
jiefk | http://paste.debian.net/1080940/ | 22:12 |
gnarface | are you copying this config while xorg is running? | 22:14 |
gnarface | or did you exit it first? | 22:14 |
gnarface | this one says it worked | 22:14 |
jiefk | Heh can't tell, I SSH'd from my other compy :p | 22:14 |
gnarface | this one says it worked and exited successfully | 22:15 |
jiefk | I'll htop to see what's going on | 22:15 |
gnarface | see if you still have a login manager running | 22:15 |
gnarface | like slim or lightdm or kdm or whatever | 22:15 |
gnarface | it does seem possible that maybe it's sending the output to the wrong display | 22:18 |
gnarface | you said this was a dual-gpu laptop though, right? if so, it's possible it's sending it to the wrong card | 22:18 |
gnarface | though usually in those cases, the intel card is the default | 22:18 |
jiefk | mesa-utils ans mesa-utils-extras were not installed | 22:19 |
jiefk | I installed those packages | 22:20 |
jiefk | also, I `update-rc.d lightdm remove` | 22:20 |
gnarface | oh, it was running? | 22:20 |
jiefk | no, from ssh still | 22:20 |
gnarface | you could check that from ssh | 22:20 |
gnarface | ps aux --forest | 22:21 |
jiefk | --forest ? this is quite an argument :) | 22:21 |
jiefk | I'll try that after next reboot | 22:21 |
jiefk | OK I still have no screen :/ | 22:27 |
jiefk | http://paste.debian.net/1080942/ | 22:28 |
gnarface | hmmm. well i don't recognize everything you're running there, but nothing sticks out at first glance | 22:29 |
gnarface | any change to the xorg log? | 22:29 |
gnarface | hmmm. i wonder if gpm could be causing an issue | 22:30 |
gnarface | i doubt it though | 22:30 |
jiefk | gnarface : http://paste.debian.net/1080945/ | 22:35 |
jiefk | I got this file from `startx` as root from ssh | 22:36 |
gnarface | jiefk: the last line still says it's terminating successfully: [ 841.893] (II) Server terminated successfully (0). Closing log file. | 22:36 |
jiefk | I CTRL-C it after about 30 secs | 22:37 |
gnarface | oh | 22:37 |
gnarface | well it thinks it was working | 22:37 |
jiefk | but the screen is still off on the laptop | 22:37 |
jiefk | (ie no backlight) | 22:38 |
gnarface | outputting to "eDP1" at 1366x768 | 22:38 |
gnarface | something is still missing but it's not clear to me what | 22:38 |
jiefk | the screen works OK as I view GRUB and the starting sequence | 22:38 |
gnarface | this is kdm? | 22:38 |
gnarface | er, kde? | 22:38 |
gnarface | kde plasma? | 22:39 |
jiefk | Yes KDE, but I removed every *dm packages I found | 22:39 |
jiefk | to get to TTY directly | 22:39 |
gnarface | it could be failing differently now even though it looks the same. it could be now giving you a black screen due to compositing. try disabling compositing in kde (sorry don't know how specifically) or try a different WM, one that doesn't composite by default | 22:40 |
jiefk | fluxbox ? | 22:40 |
jiefk | but how do I change it ? | 22:40 |
gnarface | i'm not sure about fluxbox | 22:41 |
gnarface | but startx will obey ~/.xinitrc | 22:41 |
gnarface | to whatever WM you set there | 22:41 |
gnarface | i know blackbox won't composite by default, and e17 will ask you at first startup | 22:41 |
gnarface | kde WILL composite by default, but there's some way to turn it off in a config, i just don't know exactly what | 22:42 |
gnarface | i think if you search for disabling plasma effects or something like that, it should tell you | 22:43 |
gnarface | "plasma desktop" maybe? | 22:43 |
gnarface | brb | 22:46 |
jiefk | gnarface: yes, it's sending screen to the HDMI output. | 23:11 |
jiefk | I just plugged it to the TV screen | 23:11 |
jiefk | now I need to know how to get back everything to the main laptop screen (including TTYs) | 23:12 |
jiefk | but it's for tomorrow, now is time to walk with $DOG and sleep :) | 23:13 |
jiefk | Thanks again gnarface and debdog, see ya ! | 23:14 |
fsmithred | jiefk, maybe F8 or some other key? | 23:36 |
gnarface | jiefk: no problem... tomorrow try using xrandr on it while xorg is still running. that's supposed to work | 23:52 |
gnarface | jiefk: you should be able to use xrandr to see the list of current display ports and tell it to switch to a different one | 23:53 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!