libera/#devuan/ Sunday, 2018-08-26

g40systemdlete: badblocks is a good suggestion, will try it. thanks.00:14
pav5088Question : Does ifdown and ifup work on Devuan?  It seems on Debian Stretch it does not...  or at least the behaviour is undefined.  I reported this on the Debian mailing list, and it seems it isn't even an architecture-specific thing.  I'll say this again - even something as basic as ifup/ifdown doesn't work predictably on Debian.  I even got some (un)helpful systemd error messages (not saying the fault is necessarily systemd, but I'm suspicious).03:50
gnarfaceyea, sounds familiar03:55
gnarfacei don't think you're the first person to complain of this03:55
gnarfacei also think it's typical of systemd to break stuff they didn't know about03:56
pav5088gnarface, whaat?  That's just crazy.  Something so basic surely can't be so borked...03:56
pav5088...especially after so many releases03:57
pav5088(isn't Stretch onto the 5th release or something?)03:57
gnarfacecertainly you must be putting me on03:58
gnarfacethis whole distro only exists because a critical mass of people didn't think systemd exhibited basic competence03:58
gnarfacethey probably have some other way to do that which solves the imaginary problem of no tool existing to do that04:00
pav5088gnarface, well, I use Devuan on my personal PC...  and I haven't even had to report any bugs.  (Actually, I have had lockups when playing certain games though...   perhaps I should look into that).  The basics though are just traditional-Debian-solid....04:05
gnarfacedoesn't that make you wonder why they wanted to change it to a project that clearly wasn't being maintained competently?04:06
pav5088I just assumed Debian Stretch would be similar, except with annoying pointless hairballs.  I didn't expect to get snagged in my play environment almost instantly with failures of basic functionality.04:06
gnarfacelol i know right.  it's like they're not even using it themselves...04:07
gnarfacethis is more a rant for #debianfork though04:07
pav5088Didn't even know there was such a channel.04:08
golinuxThey are probably developing Linux on Macs.  Not joking04:10
gnarfacei believe they're just testing in virtual machines04:19
gnarfacemostly automated unit tests04:19
gnarfacethey might be using macs04:20
golinuxI have seen Linux presenters with Macs on the podium04:32
DocScrutinizer05systemdlete: badblocks doesn't work on MMC14:52
DocScrutinizer05oh holla beat me on that14:53
DocScrutinizer05I vab confirm that akk the defect uSD I hd either didn't show meaningful results on badblocks, or my system crashed (FU)BAR before badblocks even finished14:56
DocScrutinizer05"bad blocks" of MMC ar listed during boot (mount?) by kernel afaik.15:06
DocScrutinizer05damn, did "recent" ssh client updates (+ related libs) cghange the environment and/or behavior of *server* where I log in to with this ssh client? E.G. I can't start mc anymore, I get:15:36
DocScrutinizer05IroN900:~# mc15:36
DocScrutinizer05Error opening terminal: xterm-256color.15:36
g4570nCenturion_Dan: ping17:16
dethaDocScrutinizer05: could well be. I rarely look at what $TERM is set to, but it is now xterm-256color17:26
dethathat might explain why vi on servers suddenly started doing silly color-highlighting17:27
systemdleteDocScrutinizer:  So badblocks does not work on uSD cards?  Not surprising if so.18:03
systemdleteIt was my only idea, that's all.  Sorry for the bad input.18:03
DocScrutinizer05systemdlete: np. Good we cought it18:04
systemdleteare you booting off the card?  You said kernel reported bad blocks on boot18:04
DocScrutinizer05my experience: to test a uSD card. you need to write one large (or multiple smaller) files to it to fill it completely to its capacity18:04
DocScrutinizer05then read+compare for max certainty18:05
systemdletebadblocks does essentially the same thing, except it writes at the block device level, not via the fs18:05
systemdleteyour approach MIGHT not thoroughly test all the "overhead" blocks that are not part of the filesystem18:06
systemdletestill, yours is a pretty good way of testing the card, and I'll keep that in mind next time I use a mSD card18:07
DocScrutinizer05I often do dd if=whatever of=/DUT/testfile bs=1m count=1 && dd if=/DUT/testfile of=/DUT/testfile bs=1m count=<size_of_uSD_in_MiB_minus_a_few>18:07
systemdleteah, see, that is writing directly to the block device.18:08
systemdleteso badblocks ought to work also.18:08
DocScrutinizer05systemdlete: badblocks tests on a per-block basis which doesn't warant to test any particular amount of real physical blocks on uSD since those get moved around by the flash controller18:09
systemdleteI see.  so the dd trick is probably about the best bet -- you catch all blocks that way, regardless of the REAL ordering (per controller)18:12
DocScrutinizer05systemdlete: actually there are fake uSD / USB drives that have only a small fraction of the advertised (and reported in device info data structure) real physical memory and they simply mirror that fraction to cover the complete advertised address space. A badblocks would not even detect those devices18:20
systemdleteMirror?  how can you mirror blocks that don't exist?18:21
systemdleteor is the assumption that most people wont actually use all the space anyway, so they wont notice?18:21
DocScrutinizer05dd trick would, with readback+compare (unless you managed to use an exact base 2 fraction of claimed size for the initial bs=1m count=1)18:22
* systemdlete needs to make coffee but keep chatting if you want18:22
DocScrutinizer05yes, the latter18:22
systemdletepure sleaze!18:22
systemdletenext time I buy a USB, I will run a test before using it...  lol18:23
DocScrutinizer05fun to unnoticed have your 1st photo getting overwritten by #4096 photo18:23
systemdleteif actual < advertised, then start class action18:23
DocScrutinizer05btw kudos to your nick! nice :-)18:24
systemdleteI think it sums up my feelings pretty concisely18:25
DocScrutinizer05detha: how would we fix that $TERM etc shit?18:25
DocScrutinizer05and whom to shoot with owl feces for introducing that?18:26
DocScrutinizer05Leander: holla been correct, uSD *do* have exactly that type of wear leveling18:28
Leanderwell, I didn't know about that18:30
DocScrutinizer05NAND OTOH doesn't18:30
dethaDocScrutinizer05: good question. term comes somewhere out of the login stuffs18:31
DocScrutinizer05I think I noticed that shit with all my remote boxes (de??an, maemo, Suse) since upgrading my local PC ssh client (and complete system)18:33
DocScrutinizer05which now runs Leap1518:33
systemdletedid you remember to set $TERM-SYSTEMD ?18:34
dethahmm. first confirm that after 'export TERM=xterm' things work again18:34
* systemdlete ducks18:34
dethaif so, gotta trace back what tacks that stupid -256color onto it18:35
detha(or just bluntly put "if $TERM eq xterm* export TERM=xterm" in /etc/bash.bashrc)18:36
DocScrutinizer05systemdlete: not _that_ bizarre a comment. Quite possible something similar is needed18:38
unmyDocScrutinizer05, which xterminal you are using on your local PC?18:39
systemdleteweird is the new normal in Linux18:39
RyccardoDocScrutinizer05: check out ¨F3¨ (fight flash fraud), it´s a read-write test designed especially for capacity-hacked drives18:39
DocScrutinizer05detha: so you say it's all a question of local $ENV that get exported to host by ssh client?18:39
DocScrutinizer05Ryccardo: nice18:40
systemdleteDoes F3 automatically contact the FTC Fraud department when it finds these drives?18:40
RyccardoI think that´s bloat :P18:40
dethayup. ssh sends through whatever your local $TERM is to the other side18:41
systemdleteThe only thing that is bloated is their false claims!18:41
systemdletefor a couple days now, logwatch is reporting a LOT of logged packets on my Ascii VM18:42
DocScrutinizer05detha: does that also apply to keymaps? I used to invoke history completion oin bash via <substring+> PgUp. Now gives me a ~ only18:43
unmyDocScrutinizer05, which xterminal you are using on your local PC?18:44
fsmithredDocScrutinizer05, you talking about scrolling up your history with pgup?18:44
DocScrutinizer05Konversation, since ages18:44
unmyand what echo $TERM spit out?18:44
DocScrutinizer05fsmithred: completing a started line from histroy, by PgUp, yes18:45
fsmithredoh, I don't know that trick18:45
unmyremote hosts not always have whole database of terminfo18:45
DocScrutinizer05jr@saturn:~/Dokumente/Telegram> echo $TERM18:45
unmyif you wanna be safe just use something like: TERM=xterm-256color ssh user@host18:45
DocScrutinizer05of course :-S18:45
DocScrutinizer05ugh prolly not18:46
unmyDocScrutinizer05, did you export before or done that command right after login?18:46
DocScrutinizer05I guess you meant  TERM=xterm ssh user@host18:46
unmybasicly every distribution support xterm-256color right of the box18:47
DocScrutinizer05unmy: I never did anything like that before18:47
dethaActually don't know about keymaps. I have noticed that between devuan1 and devuan2 shift-arrow-up to scroll through xfce-terminal stopped working18:48
unmyblame your konversation which done something bad or you mess up somewhere else :\18:48
DocScrutinizer05and I'm sure I didn't have TERM=xterm-256color or otherwise I'd wonder why it worked on that maemo system that now spits "Error opening terminal: xterm-256color"18:49
dethahmm. /etc/ssh/ssh_config says SendEnv LANG LC_*, and doesn't say 'send TERM'18:51
DocScrutinizer05unmy: ys, I already insinuated that my local system broken it on upgrade. Either ssh client, or terminal, or bash startup scripts or whatever. I hoped some of you folks would have recognized that error/defect and know "ah this is the changed /etc/bash/foobarRC which been introduced with bash5.89"18:52
gnarfaceDocScrutinizer05: just set it to xterm18:52
DocScrutinizer05or sth like pretty much what detha just posted :-)))  I have massive issues with all LC_* struff18:53
gnarfacenobody universally agrees on what to set it to or how to set it.  "xterm" is ironically the most universally compatible setting right now18:53
gnarface(though it's supposed to be vt100 or something like that, xterm has become the defacto standard due to Linux saturation)18:54
DocScrutinizer05...and I guess I had "xterm" so far. Now it's TERM=xterm-256color18:54
gnarfacerxvt-unicode-256color is doing the right thing but nothing has been updated to actually pay attention18:54
DocScrutinizer05I asked in #suse, but... that channel is not even a 100th as helpful as e.g. this one here18:55
unmygnarface, ye but many remote systems don't have rxvt-unicode-256color in terminfo because many distribution need full rxvt package to install.....18:56
DocScrutinizer05it's weird how they seem to have fsckd up some pan-distro compatibility thing18:56
dethawait until someones brings adm3a back...... terminfo stuff has always been at some level of broken18:57
DocScrutinizer05or MAYBE I've changed that form $whatever666colors to xterm a decade_and_some ago, and it stuck until recent distro upgrade?18:57
DocScrutinizer05let me try the suggested fixes18:58
unmydetha, well you can always install own terminfo database as user on remote host and will work just fine18:59
DocScrutinizer05YEAH, mc is back with "export TERM=xterm"19:00
dethaunmy: yeah. but that is a royal pain. And it assumes I have a tic for that system19:00
unmydetha, yep :/19:00
DocScrutinizer05so "who" is setting those envs usually, where?19:00
DocScrutinizer05I *bet* the same culprit messed up my keymaps19:01
dethaused to come out of getty, or your X login stuff19:01
gnarfacei had assumed the remote system shell sets it to whatever the local term reports19:01
DocScrutinizer05obviously not19:01
dethait does.19:01
dethabut the remote understands xterm, but not xterm-666color19:02
DocScrutinizer05oh wait, I didn't finish reading the above sentence19:02
gnarfaceDocScrutinizer05: just add "-tn xterm" to your urxvt command-line19:02
DocScrutinizer05never heard of urxvt19:02
gnarfaceit's not worth going down the rabbit hole on this one just to come up with the conclusion that xterm is still the only reasonable value19:02
gnarfaceurxvt is rxvt-unicode-256color19:03
dethaIt would actually be nice to know where that is set19:03
gnarface(it's just a symlink)19:03
DocScrutinizer05gnarface: I only want to revert my local system back to what it been before the upgrade19:03
gnarfaceoh actually on ceres it's a binary now19:03
gnarfaceyikes that runs gid utmp19:04
systemdleteah, I had bumped up ufw logging to high. Set it back to low should quiet those AUDIT log messages -- they are coming from yahoo, google and other public sites19:04
DocScrutinizer05and my local system is suse leap1519:04
unmyhe using Konversation x terminal so first I would check other more sane terminals how working and after just blame Konversation :D19:04
gnarfaceDocScrutinizer05: on debian, i've never not had to override it19:04
gnarfacefor rxvt-unicode*19:04
DocScrutinizer05on suse, I also never had to, iirc19:05
DocScrutinizer05"just worked" until last system upgrade19:05
unmygnarface, depends from use, try login to openwrt router with rxvt* and run any ncurse based application :P19:06
* DocScrutinizer05 will have a deep look into Konsole settings (sorry, NOT Konversation, I'm not _that_ weird to run my shells from my IRC client)19:06
unmyin Konsole there are probably some setting about TERM to use and fallback or were there years ago...19:07
unmyor create hacky alias 'TERM=xterm ssh' :D19:09
gnarfacei just put the extra command-line options in my enlightenment toolbar icons19:10
DocScrutinizer05zilch like "ENV: ..." in Konsole settings. must be 100% heritage from shell startup / other (global) env inheritance19:13
DocScrutinizer05mhm!  [2018-08-26 19:22:56] <tacit> In particular, handling of the environment and terminal handling changed to improve security.19:25
DocScrutinizer05I'm pretty sure nuking a working system config doesn't do much good for security19:27
HazukiHello. I'm having trouble with Runit; installing it seems not to produce any service files (especially, nothing for agetty), which means it can't start a terminal. I've been googling for an hour or so on this and reading dev1galaxy but no luck20:37
HazukiIs this a case of "write your own, n00blet" or...?20:38
gnarfacenot sure but i'm guessing you're missing another package20:39
gnarfacesomething sysvinit also uses20:39
gnarfacefile-rc or sysv-rc or something like that maybe?  i don't know the name but i think this has come up recently20:39
gnarfacei could be wrong.  maybe you have to write your own stuff, but usually in that case they at least put decent examples in /usr/share/doc/[package]20:40 file-rc package and i've got sysv-rc already20:40
Hazukibeen combing through the docs a bit...i notice devuan's package has less stuff in it than debian's does20:42
unmyHazuki, dont you have anything in "/etc/sv/" ?20:44
Hazukinope, total blank. i used Void for a while and that's the first place i looked20:45
Hazukioh wow, there's a getty-run package and when installed it dumps stuff in /etc/sv20:47
Hazukiyou'd think this would be a dependency or at least a recommend...20:48
unmyHazuki, ye it definitely should be but isn't :/20:51
unmyHazuki, in Void could just do "xlocate /etc/sv" and in debian maybe that "apt-file search /etc/sv"20:55
Hazukino apt-file command on here...weird20:56
unmynah, have to install and after apt-file update20:57
Hazukiahh, okay. can you tell i'm not a debian girl, LOL? i started on Gentoo in 200420:58
Hazukiback when a Pentium M could complete a reasonably small emerge -uDavN world in a day20:58
unmyI remember gentoo and Pentium III 800mhz overclocked to 1ghz :P20:59
Hazukigoooood times21:00
gnarfacedo you have the "initscripts" package, Hazuki?21:01
Hazukias far as I know there's no package like "runit-world" on Artix that populates /etc/sv with scripts21:07
gnarfaceis there an /etc/ruinit?21:08
gnarfaceor /etc/runit i mean?21:09
gnarfacei found this on the debian wiki:
gnarfaceit says to copy the contents of /usr/share/doc/runit/debian/ to /etc/runit21:09
Hazukiyes, that was my first hit on google. there is no /debian/ folder though21:09
gnarfacehow about examples?21:09
gnarfaceor something like that?21:09
Hazukialso, the html doc for using runit as PID one references Debian Woody... >>;21:10
Hazukithat's so long ago i don't even bloody remember how long, pardon my Klatchian21:10
fsmithredI might still have some woody disks I could send you21:11
Hazuki appears to be a collection of generic run scripts...maybe the package ought to mention this on install21:11
gnarfacei have trouble believing those files aren't still in the repo somewhere21:12
gnarfacebut maybe you do have to get them from an older release...21:12
gnarfaceif so though, this is another case of inconspicuous vandalism21:12
fsmithredmight the missing files be called and getty.finish.x?21:13
Hazukihow so? o_O someone deliberately wiped out whatever runit-initscripts package there was, you think?21:13
fsmithredin the runit source, there's debian/runscripts/ containing those two files21:14
gnarfaceHazuki: we've been seeing some incidents of "nobody's using this, right?  woops, already deleted it... too late now" coming from debian21:14
gnarfaceit's not happening a lot yet but this wouldn't be the first incident21:15
Hazukioh, sod ><21:15
Hazukisounds like they've really drunk the systemd koolaid, eh?21:15
Hazukii can't believe this is happening....systemd is a transparent attempt to MS-ize Linux by RedHat21:16
Hazukii've been on Void and Artix for ages, was on Funtoo/OpenRC for years before that, as soon as systemd hit21:17
gnarfacemoney talks, they say.  but it doesn't just talk.  it sings, dances, cheats, steals, and it lies.21:17
Hazukiit's always a delicate money and no dev work happens. too much, and the devs end up doing dishonorable things21:18
Centurion_DanHazuki: there has to be both integrity and money... from a dev with no integrity you can expect them to sell out.. with no money devs will have be doing that work afterhours in competition with other things like family and social/leisure activities..22:26
ibanjaMy wife is moving to Linux and I was trying to find a way to put an icon on the desktop that ejects/unmounts usb sticks and such like on Windows, so she doesn't have to type unmount.22:36
ibanjaexcuse me, not on the desktop, but in the system tray22:37
xcmmaybe KDE is better suited for someone moving from windows?22:38
xcmotherwise depends on how you mount / what your desktop is22:38
gnarfacei don't think it would be so difficult with any window manager22:40
gnarfacedon't most of them support this natively if you have udisks2 installed?22:40
ibanjaI want to keep her with LXDE because I am familiar with it.22:41
gnarfaceworst case scenario you make a simple bash script and tie that to a toolbar icon instead22:41
ibanjagnarface: I was thinking that... just hoping someone knew of an existing solution22:41
gnarfacei assumed LXDE had something built in but i haven't really looked myself.  i still do it the old fashioned way22:42
ibanjahow does it support this with udisks2?22:42
gnarfaceis it installed?22:42
ibanjai do it the old fashioned way too...22:42
ibanjaudisks2 is installed22:42
gnarfacewell like in e17 and xfce for example, if udisks2 is installed, anything you plug in shows up as a desktop icon you can right-click on sorta apple style22:43
gnarfaceor it should anyway, i thought22:43
gnarfacei mean22:43
gnarfacei uninstalled udisks2 to make that stop happening22:43
gnarfaceit not working out of the box isn't a situation i was familiar with22:43
ibanjaI just tried putting a usbstick in and no icon22:44
gnarfacecould it be missing some other lxde packages?22:44
ibanjaGot it!22:45
gnarfacewhat was it?22:45
gnarfacelxpanel or something?22:46
ibanjaI just noticed that when PCManFM fires up after putting in the USB, there is an icon to umount... never noticed because I always navigate with the command line22:46
gnarfacethere might be some seek delay too, depending on power management states and device speeds22:46
ibanjaspeaking of PCManFM... my ~/.config/pcmanfm/LXDE/pcmanfm.conf file won't keep the "view_mode=list" setting. It keeps changing back when I start up PCmanFM... any ideas what's causing this?22:46
gnarfacei got nothing for that, sorry22:47
gnarfacenever used it22:47
gnarfacehang out though.  someone must have22:47
ibanjaright... thanks for the USB help22:47
gnarfaceno problem22:47
gnarfacesubdirectories of /media/ are the designated default mounting points for devices22:51
ibanjathat's where my USBs mount to22:53
ibanjaunless there's an fstab entry22:53
Centurion_Danibanja: by default xfce puts a disk/cd/dvd icon on the desktop when one is mounted and if you right click on it, it has an eject option - also thunar file manager shows and eject option for all removable medial.23:03

Generated by 2.17.0 by Marius Gedminas - find it at!