freenode/#devuan/ Saturday, 2018-06-09

barnyardwhy is sdl2 gone from the repositiories00:30
barnyardthere seems to be a lot of packages missing00:31
barnyardfrom the normal repos00:31
KatolaZBesnik_b: ?00:31
KatolaZbarnyard: what are you talking about?00:31
KatolaZdefine "normal"00:31
barnyardfor example00:31
barnyardwhen I try to install irssi00:31
barnyardI get  Depends: libval14  but it is not installable00:31
KatolaZbarnyard: which repos are you using?00:32
KatolaZare you mixing repos?00:32
muep_I guess you did run apt update (or equivalent) already?00:32
KatolaZbarnyard: irssi installs just fine on ascii00:32
KatolaZand on jessie00:32
barnyardyeah I have jessie00:33
KatolaZare you sure you don't have non-devuan repos?00:33
barnyardI'm only using the stardard repos00:33
KatolaZapt-cache policy irssi00:33
KatolaZapt-cache policy libval1400:34
barnyardwhat do those do00:34
KatolaZblows your computer off...00:34
KatolaZgotta go sorry00:35
barnyardand the latest kenerl I can get is 3.5?00:36
barnyardthere has to be a mistake00:36
barnyardare there more packages on ascii?00:37
barnyardbecause if so how can I update from jessie to ascii?00:37
golinuxHold on a day or so an there will be plenty of documentation to guide you.00:49
barnyardso ascii hasn't been released yet?00:49
barnyardalso I have no make00:50
barnyardbuild-essentials doesn't exist00:50
golinuxYour sources must be mixed up00:55
golinuxThis is what you should have for ascii:00:57
golinuxdeb ascii          main00:57
golinuxdeb ascii-updates  main00:57
golinuxdeb ascii-security main00:57
barnyardseems like it00:57
barnyardI have jesse00:57
barnyardthe ones here00:57
barnyardare good, right?00:58
barnyard  deb jessie main deb jessie-updates main deb jessie-security main deb jessie-backports main00:58
barnyardyeah, seems to be working great now00:59 is a round robin of maybe a dozen mirrors.00:59
barnyardshould I switch ascii anytime soon?01:00
golinuxThat's up to you.01:00
golinuxand what your needs are.01:00
xrogaanso, logged in today and mate didn't had a window manager (didn't know what was up).02:57
xrogaanI started mate tweak and it informed me so.02:57
xrogaanit is weird02:58
barnyard_does devuan have amdgpu drivers04:58
barnyard_or does that come with linux standard04:58
barnyard_have some bad perfomance with minetest, runs usually good when I ran it on manjaro04:59
golinuxMaybe it's in non-free?05:01
Jjp137barnyard_, the package you're looking for is named firmware-amd-graphics, which is indeed in non-free05:16
Jjp137further more, if you're still on jessie, it can only be found in jessie-backports05:16
barnyard_is there a free alternative?05:16
Jjp137oh uh there should be but let me see05:17
Jjp137although I'm not really familiar with amd/ati as much (I get to bang my head against nvidia's nonsense instead) but here's a page on Debian's wiki:
Jjp137what's your GPU?05:19
barnyard_RX 48005:19
barnyard_also, is the transtion to ascii relatively painless? I'm thinking of upgrading if its stable enough05:21
Jjp137barnyard_, the xserver-xorg-video-amdgpu package is in jessie-backports too, but you might also need a newer kernel from backports as well; unless you really want to stay on jessie, you might as well upgrade to ascii06:00
Jjp137the transition should be mostly painless from what I've heard, and I have a feeling it's going to be released pretty soon anyway :)06:00
xrogaanthere is no more free/proprietary package for recent amd GPU as the two projects have merged.06:13
xrogaanIt's based from AMD's will to unify the experience, they actually contribute to the opensource thing instead of releasing their own proprietary installer (fglrx).06:13
xrogaanit's called something "AMDGPU-UNIFIED-LINUX"06:14
xrogaanSo, it's in non-free because the opensource driver uses AMD proprietary blobs. Which means the code isn't in the kernel. Overall, it's way easier this way than to install fglrx.06:34
KatolaZDocScrutinizer05: /topic07:26
ipetime to install07:35
HestbenCongratulation to the stable ascii release07:44
HestbenAlso, that announcement on devuan mailing list, that support for devuan jessie will end June 2018, was that correct, not just a mistake writing 2019? Then there seems to be some hurry to upgrade...07:46
KatolaZHestben: which announcement?07:46
golinuxI think Jessie goes for 2 more years LTS07:48
HestbenKatolaZ: from leftoft to on May 21 2018.08:29
HestbenThere is no year there, so maybe they mean 2021? Or have I misinterpreted it?08:34
Jjp137Hestben, that is pretty much based on:
Jjp137there will still be LTS support for Devuan Jessie until at least some time in 2020 (source:
Jjp137to be more exact, Debian's LTS support for Jessie ends on June 30, 2020, so at least that long08:42
HestbenJjp137: Thanks for calming me down :)08:47
Jjp137you're welcome :)08:49
AlexLikeRockhappy  holly  days09:11
AlexLikeRocki go to sleep09:11
AlexLikeRockwe celebrate  SPANISH comunity  too09:11
batmoreCongratulations all, and thanks much for the stable ascii release!09:13
alexandruhi everyone, I understand that you can replace OpenRC with runit in Devuan. How can I do that?09:35
alexandruactually, nvm, found this
xrogaanwait, is it better to use elogind?10:36
xrogaanoh, not for mate, alright.10:36
aitoralexandru: today i want to share an image with runit10:47
aitori'm fixing an issue with the gpg key while building debian-installer10:48
aitordebmirror is taking to long with, and i don't know why (but it works):10:50
aitorUpdating remote trace files (using rsync) ...10:50
aitorsaturday's coffee, bb in a few minuts :)10:52
xrogaanW: The repository ' ascii-proposed Release' does not have a Release file.10:55
xrogaanoh, it's /devuan and not /merged10:56
DocScrutinizer05re /topic, I guess you want that improved, ASCII replacing Jessie?10:57
DocScrutinizer05and really10:59
infoboti guess isodate is USE *ONLY*, HELL AND BRIMSTONE, or "In representations for interchange, dates and times are arranged so the largest temporal term (the year) is placed to the left and each successively smaller term is placed to the right of the previous term.", or
alexandruaitor: interesting, to say the least :D11:10
aitorhi again11:15
DocScrutinizer05the "imperial" MMDDYY is so damn puzzling, it's close to trumpism to use it in international context11:47
DocScrutinizer05KatolaZ: shall ASCII replace the part about Jessie in /topic ?11:48
DocScrutinizer05KatolaZ: I gather both are "stable", though iirc Jessie is LTS, so... no clue here11:49
eyalrozIs someone around to answer a newbie question?13:13
gnarfaceyou'll never know unless you ask the question13:20
eyalrozOk. So - I've used Debian, Lubuntu and Linux Mint over the past 2 decades, currently mostly the latter two. I won't go into details about my experiences which have had ups and downs, but right now I'm mostly ok with them as a lay user, up to issues with CUDA and an nvidia GPU not used for display.13:44
eyalrozAnyway, I was quite annoyed with the choice by major Linux distributions, Debian in particular, w.r.t. systemd, and I want to put my PC where my mouth is - by switching to Devuan. However, I'm worried about user experience, and developer experience, deterioration. What can you tell me, or what can I read, which would assuage my reservations enough to brave a switch to Devuan when I next upgrade my desktop's distro in a few months (from Mint 18.3)?13:44
KatolaZeyalroz: define user experience13:45
KatolaZand, possibly, define "deterioration" :)13:45
eyalroz@KatolaZ: A bit difficult to define. Apps and/or UI elements not behaving as you would expect / not working, the need for lots of configuration file tweaking just to get daily lay user activity to work properly, issues with the support for common hardware, stuff like that.13:47
aitorarrrrrrrrr !!13:49
eyalroz@aitor: ?13:49
aitorsources.list.udeb.local -> deb [trusted=yes] jessie main/debian-installer13:49
aitori am an idiot13:49
aitoradd me to the list of idiots :)13:49
aitorneed to reboot, see you13:51
gnarfaceeyalroz: gnome is busted, most of the rest is still fine13:58
gnarfacewas busted?  did anyone ever make a patch for that?13:59
eyalroz@gnarface: Oh, I forgot to mention I'm an anti-gnomer. I use Cinnamon (if it doesn't crash on me) or XFCE13:59
gnarfacei think people are using mate instead13:59
gnarfacexfce is the default13:59
gnarfacesome little stuff is broken with temperature sensors and hibernate functionality14:00
gnarfacein places where support was removed14:00
gnarfacehere and there14:00
gnarfacemost stuff will still work with lm-sensors14:01
eyalroz@gnarface: But generally, I can expect it to be as "smooth-feeling" as Debian (Which is perhaps less than Mint, but still pretty good)?14:06
gnarfaceyou won't notice any performance difference if that's what you mea14:06
gnarfaceit's the same kernel, the same drivers, the same packages (mostly)14:07
gnarfacewhere systemd grants blanket permissions to the logged-in user, devuan is still using the same permissions setup as debian was in wheezy and before, for the most part with only minor updates here and there14:08
gnarfaceit should look very familiar to you14:08
gnarfaceeyalroz: i guess the short version is that it's not deteriorated very much yet.  it's still very much salvageable.  you're getting in early enough.14:37
eyalroz@gnarface: Wait, so, you're saying there will gradually be more, rather than less, deterioration / breakage? I would have assumed you took a hit initially turning off systemd, and gradually things would get better15:34
gnarfaceeyalroz: you failed to account for asshattery.  that's what you should hope for, but what you should plan for is a steadily increasing amount of vandalism until they eventually force a full fork, which can be staved off longer the more developers join the project to help15:42
gnarfacenonetheless, with your help i'm confident we can still save it15:42
gnarfacethe debian that was15:42
gnarfaceso far it's working well15:43
eyalroz@gnarface: I'm actually not much of an OS distro guy. I've written couple of Thunderbird extensions,15:43
eyalrozyes, another distressted project...15:43
eyalrozbut next time I'm employed I'll definitely make a donation15:43
gnarfaceit's really the only serious attempt to preserve the general distro design/behavior of debian wheezy and earlier15:44
gnarfacefor software going forward15:44
gnarfaceas that, it's doing really well, but obviously more help would mean less delay15:45
grillonhi there15:45
grillonI need arm-none-eabi-gcc6 how can I install it on devuan ascii? I guess it's in testing?15:46
gnarfaceeyalroz: but what i meant about you getting in early enough is primarily that as of ascii, they really haven't broken much15:46
grillonHi MinceR15:46
gnarfaceeyalroz: and for the most part it's all been fixed for the ascii release15:46
eyalrozDo we know that there's planned additional breakage?15:46
eyalroze.g. by systemd taking over more of the system,15:47
eyalrozor stronger dependencies on its facilities?15:47
gnarfacethere's been a worrisome pattern of them removing stuff that "nothing was using anymore" like /etc/resolv.conf ...15:47
gnarfacebut for the most part those types of things get corrected in devuan15:47
MinceRhay grillon15:48
gnarfacegrillon: it might be named slightly differently15:48
grillonit's been a long time MinceR, nice to see you15:48
grillonhmmm, so how could I find it?15:49
gnarfaceapt-cache show gcc-arm-none-eabi15:50
gnarfacewhat does that show you?15:50
grillononly the version 5.4.115:50
gnarfacei see 6.3.1 in ceres15:51
gnarfacecheck ascii-backports, otherwise you'll have to backport it yourself or install it in a ceres chroot or something like that...15:51
grillonnot in ascii-backports15:52
grillonhmm let's try chroot, thank you for your help :)15:54
gnarfaceno problem15:54
barnyardascii is looking good16:26
muep_which common GNU/Linux distribution does not use resolv.conf anymore? afaik at least all of debian, fedora, ubuntu still use it just fine16:44
gnarfacemaybe it was rc.local?16:55
muep_I think that one still works as well on all of those16:56
muep_it's a bit less of a free-for-all than it used to, though, because there is an expectation there that the processes started from rc.local also terminate in some reasonable time17:00
DocScrutinizer05where to report a ssh bug? (upstream, also after checking source)19:57
DocScrutinizer05>>client: no timeout in negotiation after server stops responding during negotiation<<19:58
yann-kaeligelo. How long should I wait for a new stable release of a package like opennebula and last stable release 5.4.13
DocScrutinizer05any devel around who could point me to ssh client upstream?20:49
DocScrutinizer05sorry I suck at this20:49
KatolaZDocScrutinizer05: ?20:55
DocScrutinizer05KatolaZ: I think I found a bug in ssh client, want to check upstream source20:55
DocScrutinizer05and bugtracker20:56
KatolaZapt-cache source PACKAGE20:56
KatolaZwe haven't forked ssh20:56
KatolaZthe package comes form Debian20:56
DocScrutinizer05yeah, I'm more interested about where to find upstream20:56
DocScrutinizer05the genuine bugtracker and upstream repo20:57
KatolaZit should be in the corresponding orig.tar.gz20:57
KatolaZor in the upstream branch20:57
KatolaZdepending on how it's maintained20:57
DocScrutinizer05:-) I'll try to find my way along that20:57
DocScrutinizer05though *maybe* aunt google can help me out20:57
DocScrutinizer05or - DUH - a man ssh20:58
DocScrutinizer05hmm no, manpages are not helpful20:59
Jjp137DocScrutinizer05, ssh is OpenSSH I believe, so here:
DocScrutinizer05oh great! thanks a lot21:01
Jjp137(yea it is OpenSSH:, see the right sidebar)21:03
DocScrutinizer05I found the bug on a pretty old version of ssh (2015) on a Suse distro, but I know of this bug since 10+ years and I want to check if it's fixed meanwhile. Since it's hard to reproduce I can't check the easy way, I was lucky to encounter it while running ssh -vvv today, first time ever21:08
bill-auger2015 is pretty old huh?21:09
bill-augerlemme think now .... when did jessie come out?21:09
bill-augers'a tough one eh doc :)21:09
DocScrutinizer05I *guess* the devels simply didn't think of a reply never arriving so are sitting idly waiting happily for reply until hell freezes over21:13
DocScrutinizer05alas with packet loss it's not exactly unlikely that a reply during negotiations gets lost21:13
yann-kaeligvery good choice openrc as init, now you can move on S6 ^^21:18
Wonkabut... shouldn't tcp resend stuff?21:35
nacelletcp should retry and time out21:47
nacelleopenssh is well thought out, beat up and tested pretty hard at this point21:48
aitorhi all21:49
nacelleDocScrutinizer05: do you have the ClientAliveInterval and ClientAliveCountMax settings set?  (do those even come into play if the connection isnt established?)21:54
nacelleor ServerAliveInterval 1521:54
nacelleServerAliveCountMax 321:54
DocScrutinizer05no, aiui they don't21:55
nacellei got 'em backwards21:55
nacelleclient should have the ServerAlive...21:55
nacellenot sure if that applies when connection isnt up21:55
DocScrutinizer05funny enough in latest sources I can find a "SSH2_MSG_NEWKEYS sent"  but no "SSH2_MSG_SERVICE_REQUEST sent"21:55
aitorDocScrutinizer05 has been in Manhattan21:57
DocScrutinizer05Wonka: I guess the scenario is: TCP packet client->server succeeds: reply server->client gets eaten completely by packet loss and server gives up on resending21:57
DocScrutinizer05nacelle: ^^^21:58
nacellethis seems dubious21:59
nacelletcp retransmits21:59
nacelleyou wouldnt have it do that21:59
DocScrutinizer05so what? when connection has 100% packet loss21:59
nacellelike you're able to get a few packets out and then your internet connection fails?21:59
DocScrutinizer05no, in this case the server internet connection failed with 98% packet loss22:00
DocScrutinizer05so fact is my client got stuck forever on
nacellei know i've hit it before22:01
DocScrutinizer05obviously no timeout on application protocol layer22:01
nacellethere might be one you can set and a reason why its not by default22:01
nacellebut there could be a bug22:01
nacelle(i'd be cautious about calling it a bug outright, but you know you have questionable behavior for now)22:02
DocScrutinizer05I guess it's not showing up too frequently, since connection has to be up and delivering to send the request, then goes down completely before response arrives at client22:03
nacelleit smells bug-esque22:03
nacellelike a gateway transition could cause it if they're not using state so well22:03
nacellefirewall failover, etc.22:03
DocScrutinizer05or random packet loss like in my case22:04
nacellemight happen if you're trying to do state over a common port they give up on tracking state over, like 80 or something22:04
nacelleeh i doubt it22:04
nacelletcp retransmits22:04
DocScrutinizer05dosn't help22:04
nacelleit doesnt just send one thing out and stop22:04
nacelleit does!22:04
DocScrutinizer05random does whatever you want, you just have to wait for it22:05
nacelleright, this gets around it22:05
nacellesome things shoudl randomly get through22:05
nacelleso tcp retrying packets it sent off will randomly get them through, in theory22:05
nacelleotherwise you end up with a socket timeout22:05
nacelleand that will be seen22:05
DocScrutinizer05fact is the connection went down after first ping when I ran a mtr against server, so what are we discussing here?22:06
DocScrutinizer05the socket timeout is on server side22:06
nacellesocket is formed in the client, yo22:06
DocScrutinizer05client side doesn't see any connectivity ussues22:06
nacellethat particular socket will timeout22:06
DocScrutinizer05no obviously not22:07
nacelletcp will not see responses and will stop22:07
DocScrutinizer05why would it?22:07
* nacelle sighs22:07
nacellego read this:
DocScrutinizer05we're talking application layer protocol here, not a SYN/ACK whatnot TCP layer issue22:08
DocScrutinizer05TCP has no keep-alive pings built in22:08
DocScrutinizer05that's why a ssh connection idles for hours even when server rebooted22:09
nacelleuh, no22:09
DocScrutinizer05unless you have keep-alive22:09
* nacelle wanders off22:09
DocScrutinizer05maybe a established TCP connection tomes out, after hours or days22:11
DocScrutinizer05times out, even22:11
DocScrutinizer05or maybe there's no TCP connection established in this case, and my ssh client waits for a new inbound connection from server? No, can't be since this wouldn't fly for a number of reasons, NAT being only one of them22:12
DocScrutinizer05so ssj has --keep-alive for a reason: to avoid NAT closing the session on *idle* TCP connections, when neither client nor server send data. Server can get nuled off the globe and client wouldn't even notice until next time tryong to send a packet22:14
aitori updated the repo of gnuinos, doing d-i compatible with linux-4.922:14
aitorneed to reboot again: live-sdk :)22:17
DocScrutinizer05I can ssh from PC into my phone via WLAN, then take the phone out for a walk for half an hour with no WLAN connection for sure, come back home and the ssh session still is up and running as if server (phone) never vanished22:17
DocScrutinizer05no timeout on any sockets22:18
DocScrutinizer05since no packet got sent either direction22:18
nacellessh reestablishes if you have keys22:20
nacelleits some pretty magic stuff22:20
nacelleyou'll still tcp timeout (your client doesnt see an ack in time and cools off)22:20
DocScrutinizer05ACK for what?22:21
nacellewhatever it sent22:21
DocScrutinizer05when there's no data to send, there's no ACK to reply to such data sent22:21
nacelleit sent a request?22:22
DocScrutinizer05so? thst arrived22:22
DocScrutinizer05and got its ACK22:22
nacellecan you take this to #openssh?22:22
nacellei'll join you there22:22
aitorkmod replaces to module-init-tools?22:39
aitori had the sources.list set to ascii23:09
aitori'm building gnuinos jessie with the live-sdk with the new repo; it'll take some  time23:11
tom13note to community:  if running devuan ceres, gimp upgrade from 2.8 to 2.10 requires you to add an additional package gimp-python, to maintain scripting functionality23:26
DocScrutinizer05ssh topic "solved":  HMMM there's something in there evein in 6.2p2, but I fail to understand where timeout var gets set to a non-null value, and for sure I don't see the logit("Connection to %.200s timed out while " "waiting to read", get_remote_ipaddr());23:28

Generated by 2.17.0 by Marius Gedminas - find it at!