libera/#devuan/ Friday, 2021-02-19

foresterHow today I decided add chimera in repositary and suddenly have understand there is http: . I added 's' https://deb.devuan.org/merged/ and it is imposible to download. QEUSTION: Why there is no https varaint?  Because there were many times when I was downloading in synaptic some packages, and something another with them was downloaded in parrallel. Synaptic did show rate only 20-40 Mb, but System Monitor did show full rate about 200 Mb (09:38
forestermobile usb modem). And package about 20 Mb finished downloading process, where as System Monitor did show that there was downloaded about 100-200 Mb. WHAT COULD YOU SAY?09:38
forester* not How, but Hullo09:39
onefangdeb.devuaan.org is a round robin DNS, pointing to a bunch of package mirrors.  So we can't use HTTPS for that.09:40
foresterThen as I could suppose there is allowed to use this situation for attack (do not remember how it names, something 'between' it is possible to replace packages, or add something malicious). In any case something was downloaded in parallel many times (4-5). I have Sublime Text in repository, and it has 'https'09:44
onefangLook at https://pkgmaster.devuan.org/mirror_list.txt and choose a mirror that HTTPS if you want to use HTTPS.09:45
ltsThey usually need to hear the answer "packages are signed"10:07
xinomilosigned & reproducible (most of them), but https is a good thing too.10:24
ltsSure, signing provides integrity and https some confidentiality14:57
furrymcgeePlease note that Debian can neither confirm nor deny whether the certificate authorities whose certificates are included in this package have in any way been audited for trustworthiness or RFC 3647 compliance.15:09
lts"mozilla/Hongkong_Post_Root_CA_3.crt" is there indeed, for example15:13
unixbsddo you know maybe an easy alternative to LyX for easy use, aka. for "TeXlamers / beginners?17:24
MinceRvim :>17:25
unixbsdI mean with a frontend17:26
unixbsdvim is evilwm for begin17:26
unixbsdMinceR: ed? ted?17:27
MinceRmeh17:27
masongnarface: Oddly, I went from a pixel-based xscreensaver to an OpenGL one and it hasn't crashed - the opposite of what one might expect.18:32
gnarfacemason: you might have gotten lucky... or they might have actually fixed it after all these years (not holding my breath though - i'd give it another 48 hours before i'd even be comfortable wagering on how much longer it would last)18:41
gnarfacemason: and also i'd make sure vsync was working and check my gpu temps18:41
masonHrm, makes me wonder if I can query nVidia temps via the command line.18:43
masonnvidia-settings shows 32°C at present18:43
gnarfacei gave up on them a good decade ago, but once about 4 or so years ago i was just testing on ceres idly and managed to lock it up within moments using the one that has whales swimming around (that was using a either a nvidia 560 or 650 on whatever nvidia drivers were in the repo on ceres at the time)18:43
masonnouveau or proprietary?18:44
gnarfaceproprietary for sure18:44
gnarfacethere is a way to get temps by command-line18:44
gnarfacenvidia-msi or something like that18:44
masonI'd go for AMD but you need to chase the cutting edge kernel.18:44
gnarfaceyea i switched to AMD and it has been better18:44
masonhm, right, remembering something like that now, but I don't seem to have it installed. Will dig later.18:45
gnarfacei don't know for sure that the nvidia-settings binary can't dump values to command-line too though, i haven't looked carefully but there are command-line switches for a number of the other settings18:45
gnarfacei'm having trouble finding the package for the nvidia msi stuff18:47
gnarfaceit used to be there for sure but it wasn't ever a required package18:48
gnarfacefrankly i'm glad to be free of it18:48
gnarfaceoh shit18:50
gnarfacethat's why18:50
gnarfacebecause it's nvidia-smi not nvidia-msi18:50
gnarfacesorry18:50
gnarfacesmi for "system management interface"18:50
gnarfacea command-line tool they made primarily for cluster use, i think18:51
gnarfacehttps://pkginfo.devuan.org/cgi-bin/policy-query.html?c=package&q=.*nvidia-smi.*18:52
gnarfacemason: ^18:52
masonAh, didn't have that install, but that rings the bell authoritatively.18:53
gnarfacemsi is the thing you enable on intel onboard audio and ethernet kernel modules18:53
masons/install/&ed/18:53
nemohm19:49
nemohttps://metadata.ftp-master.debian.org/changelogs//main/t/tomcat9/tomcat9_9.0.31-1~deb10u3_changelog19:49
nemo* Removed the SysV init script19:49
nemowhy do they do crap like that? ☹19:50
nemoisn't that violation of the last Debian meeting where they said they wouldn't go out of their way to desupport SysV?19:50
nemodoes that mean Devuan has to patch a package like this, or make a fixup subpackage, or force the installer to maintain the init?19:50
nemoand. is this sort of thing a general issue across Debian that Devuan now has to worry about?19:51
nemoif I searched changelogs across the board, would I see more "Removed the SysV init script" ?19:51
nemois this a thing worth complaining to Debian about?19:51
gnarfacenemo: yes, yes, yes, no19:51
gnarfacenemo: and the devuan package full of their abandoned sysvinit scripts already exists19:52
nemognarface: oh? good!19:52
nemognarface: what is it?19:52
nemois it automatically pulled in?19:52
gnarfacei forget the name19:52
gnarfaceno, but fsmithred knows where it is19:52
nemoI feel like I need to make sure it is on the work servers19:52
nemojust-in-case19:52
gnarfacenot a standard package yet, i think it might be in proposed-updates or something19:52
nemoalthough so far haven't run into this19:52
fsmithredwhat?19:52
gnarfacefsmithred: they pulled another sysvinit script apparently19:53
fsmithredwhich one?19:53
gnarfaceapache tomcat i think?19:53
nemosadness19:53
fsmithredwe forked it19:53
fsmithredpretty sure19:53
nemoso... they deliberately ignored their own vote. it was just a show.19:53
nemo"ok, you guys voted for support, we'll just passively aggressively defer this to package managers with no hand slapping or guidelines"19:54
fsmithredCandidate: 9.0.31-1~deb10u2+devuan119:54
fsmithredtomcat9 ^^^19:54
nemofsmithred: that could get problematic no? if every time a tiny text file is pulled you guys now need yet another massive binary ☹19:54
fsmithredin beowulf19:54
nemooh. or do you do a dependency .deb ?19:55
fsmithredI forget the name, but there's a package for orphaned init scripts in debian19:55
nemoI guess you still need to fork then to add the dependency so that doesn't help19:55
fsmithredI don't think we will fork all the packages that get their init script dropped19:57
nemofsmithred: so you'll make some meta package for devuan and then some generic script that copies in inits on install of new packages?20:00
nemoinitmanager ?20:00
fsmithredmaybe. I'm not sure what the debian package will do20:00
gnarfaceluckily the total amount of storage required by all of the sysvinit scripts in the entirety of time and space is trivial20:01
fsmithredIt will only put files into /etc/init.d/ when the relevant package is20:02
fsmithred installed on your system.20:02
gnarfaceall we have to do is put them all in one place and fix them to prove their efforts to be a worthless lie20:02
fsmithredthat's in the Description20:02
fsmithredit's in ceres and chimaera20:02
nemognarface: oh. so they are pulling them because "too much effort to maintain" ?20:03
fsmithredthey may not even need fixing in many cases20:03
nemois that the reason?20:03
fsmithredtoo much fucking effort to hear someone else say "I tested it and it works, please keep it in."20:03
nemofsmithred: yeah tomcat is incredibly self-contained. I can't imagine either its init or the script for systemd are much more than "launch tomcat"20:03
gnarfacenemo: yes.  that's the whole reason systemd was created - the garbage attitude of package maintainers towards maintaining their own packages20:03
nemomildly related... on both Devuan and Redhat (well Amazon Linux 2) I've had a bug with Tomcat where random library errors appear after distro package update20:04
nemoThis is resolved on restarting tomcat20:04
nemowhich makes me think neither Redhat nor Debian packages restart tomcat service on update20:04
fsmithreddoesnt debian policy dictate how that should be done?20:05
gnarfacedo we think the people left over there still read their own policies?20:05
fsmithredI see services restarting when they get upgraded20:05
fsmithredoh, maybe they don't read20:06
nemoI haven't actually read the .deb or .rpm to see what they do20:06
nemoit's not impossible it's some other issue w/ tomcat, although I'm not clear what20:07
nemomaybe wrong order of operation..20:07
nemoor restart vs reload20:07
gnarfacelast i checked they used to restart, but that was before systemd.   after systemd i had a disturbing conversation with one of the maintainers about how they no longer saw fit to properly unmount disks "because that's what journaling is for"20:07
fsmithred!!!!!!!20:07
infoboti heard !!!!! is \"And all those exclamation marks, you notice? Five? A sure sign of someone who wears his underpants on his head.\" -- Terry Pratchett, Maskerade20:07
fsmithredwell, it was in response to something that makes as much sense as wearing underpants on your head20:08
masongnarface: Weird. There are explicit services to gate on clean block availability.20:08
gnarfacei should have saved the log but this person also told me in no uncertain terms that the only meaningful use of linux was in a VM on a azure cluster anyway20:09
fsmithredlol20:09
masongnarface: I'd be curious to know what camp they were in. That's an unusual perspective.20:09
gnarfacekernel development, iirc20:09
gnarfacecould have just been a troll masquerading as a maintainer to trigger me, which worked well, but none of the ops chimed in to say anything20:10
gnarfaceand they were definitely watching20:10
gnarfacethis was back when we were also having that failure to unmount disks before shutdown problem here20:12
gnarfacethe same problem trickled down to us but we got it fixed at some point afterwards20:12
gnarfacei'm sure you guys will recall that incident20:12
gnarfacei don't reboot my machines often enough to have a clear track on how long after we inherited that bug from them that we got it fixed here, but i have no particular reason to believe they ever bothered fixing it for debian.  i assume there's still a race condition where spinning-platter harddrives and slow usb keys still don't unmount fast enough20:14
gnarfaceand if you bring that up in #debian they will just haze you for being poor20:15
nemognarface: heh. VM in an azure cluster with your own personal ubuntu salesman 😉20:27
nemohttps://news.ycombinator.com/item?id=26114194  context for folks who missed that20:27
dagalhey, i'm experimenting with a zfs-on-root installation of devuan23:17
dagali've just snatched the dkms and relevant libs from buster-backports, and it looks like it's working for now23:17
dagalwill update when i'm sure23:18
numzobdagal: mixing debian and devuan packages?  that's a thing?  does it work?23:39
kiwi9Yes. You can even take Ubuntu ppas & use the debs they contain as long as you have the dependencies.   I installed Audio-recorder that way this week on Devuan Ceres.   https://launchpad.net/~audio-recorder/+archive/ubuntu/ppa/+packages.23:43
gnarfacedangerous and risky but possible23:49

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