libera/#devuan/ Wednesday, 2023-12-27

joergmy wine has its own app menu, I think that came OOTB00:15
joergwell, or it came with KDE  https://i.imgur.com/d7wNaIv.jpg00:17
fluffywolfI've had really bad luck with wine.  It usually doesn't work, and when it does work, I don't know why and it's not repeatable.00:33
rwpI am not a fan of wine because other than for specific games there is nothing I need from the windows side of things.  Everything else I need is available natively.00:36
onefangIt's holiday season, or Silly Season as we call it here.  Wine is only for drinking.  B-)00:39
fluffywolfI don't know anything about games...  I usually trying to run specialized software that only exists for windows, like automotive or CAD/CAM stuff.00:41
fluffywolfonefang:  I don't drink, and haven't seen wine (or other alcohol) being anything other than harmful to society.00:41
fluffywolfI wasted much of yesterday trying to get rhino3d to install.  did not succeed.00:43
systemdleteso I am still trying to figure out how to do this... it turns out that adding a script to /etc/boot.d will be too late; I need to run my script before rsyslogd starts due to a file dependency.00:48
systemdleteIn 2023, what is the politically correct way to add a script to /etc/init.d?   update-rc.d complains because my script is missing LSB.  But LSB is... uh... old?  And I thought not actually used anymore.  But maybe I misunderstand the ways of sysvinit in 2023.00:49
joergwell, my menu has: Agilent, Fujifilm, Keysight, TP-link. All 4 are for apps that dealwith external devices like cameras or digital multimeters, scopes etc. All 4 are windows-only apps00:50
systemdleteI mean, I know HOW to do it.00:50
systemdleteln -s /etc/init.d/myscript /etc/rcS.d  or the like.00:50
systemdletebut that seems just a little hackerish00:50
systemdleteAccording to a wiki article on the subject, LSB is no longer supported anymore00:51
fluffywolfjoerg:  yep...  specialized software that only exists for windows.00:52
systemdletewhat would be perfect is if one of the scripts in /etc/rcS.d were linked to a script that called run-parts on some directory for customizing bootup.01:03
systemdletebut I haven't noticed any01:03
joergsystemdlete: I *think* `ln -s /etc/init.d/myscript /etc/rcS.d ` is only half of the pudding. It's prolly more like `ln -s /etc/init.d/myscript /etc/rcS.d/S05myscript` (change rcS to whatever meets yyour needs best,  and the S05 part ot S01 or S66 depending on S_tart prio)01:04
systemdletejoerg, yes I know.  I was just saying I know how to create a symlink for this purpose!  The problem is that I am trying to avoid hackish solutions.01:04
systemdleteI mean, it probably doesn't matter all that much, but I still try to do things the best practices way01:05
systemdlete(what I meant by "politically correct"01:05
systemdlete)01:05
joergthe ppint is that prio and thus the sequence in which stuff runs is determined by the name, particularly the lexical sorting of that name, so you need a (S00 ?) prefix to make your start script run early01:07
joergthat's not any hack01:07
systemdleteright, I know that01:07
systemdleteupdate-rc.d is usually the p.c. way of doing things01:07
systemdletebut that doesn't work without a LSB it seems.01:07
systemdleteBut LSB is no longer a thing these days it seems.01:08
joergright, since that's per definition what it does01:08
systemdleteso what part of what I say do you not understand?01:08
systemdleteI'm not sure I can explain it better.01:08
onefangLast I checked you can still install LSB.  It's still a thing if something uses it.01:09
systemdleteonefang, is there a package for LSB?  I thought it was just a standard for init scripts01:10
onefangThe package is lsb-release.01:11
systemdleteI think it is installed already anyway01:11
joergthe "standard" I know of used a comment header in script files that told what the script depends on and which runlevels it is meant for (iirc). And whatever job is installing the script looks for those comments, "solves" the dependencies and **renames the symlink to the script accordingly**. Then the init process just runs the scripts in the alphabetical sort order of their (symlink) names01:13
systemdletejoerg:  Yes, that's how I understand it too01:14
systemdletebut LSB is no longer really supported, so I wonder why new scripts should have to comply with a not so existent thing anymore.01:15
onefangLSB Linux Standard Base is for more than init scripts.  Though I think that's the only part still being used.  Ironically I read up on LSB long ago so I could write my own LSB compliant sysv-init init system.01:15
onefangDon't think I have seen any init scripts WITHOUT the LSB header.  I see things that tell me this LSB stuff is stil being used.  Specifically apt complains when those init script LSB headers have something not matching the spec precisely.01:17
systemdletejust as ironically, I think, is that this LSB (iirc) was an extensive documentation effort, one of few I can think of in linuxland.  And it seemed to be completed about the same time as our arch nemesis was insidiously creeping into linux.01:17
systemdleteso, to sum it up:  We still have to have a LSB in init scripts in order for the existing toolset to continue working and maintain some kind of discipline.01:19
onefangWhich might explain why LSB went for RPM for package management instead of APT.01:19
systemdleteI've been reading dozens of pages today on various sites re init scripts and LSB and the like, and I recall (vaguely now) that one of those pages was discussing this very issue.01:20
onefangIt's the "this init script depends on these other ini scripts" part of the LSB headers that is used to figure out what order to make the init script links, and which run level to put them in.01:21
systemdleteof course it is.01:21
systemdleteno disagreement with anything said here.01:22
* systemdlete wonders how much worse this whole init system business can get...01:25
onefangBTW my sysv-init could run init "scripts" written in any language.  So if you have a string in your init "script" written in C that looks like the LSB headers, it'll find that and do the correct LSB thing.01:25
systemdleteonefang, go find your code, package it, and push it out on the mirrors!  :D01:26
onefangI have considered that.  lol01:26
systemdletemaybe it could become the standard for some future devuan release, where all traces of yecccch have been completely and utterly removed01:27
onefanghttps://sledjhamr.org/cgit/urunlevel/  "Clean room implementation of runlevel/init.d/SYS V init applets for busybox, aiming for LSB compliance."  Busybox never did accept it.  I should rewrite it for ToyBox.01:28
systemdletenice job, onefang.  I salute you.01:29
systemdletedid busybox devs ever say why they rejected it?01:29
systemdlete(like, "it works TOO well" or something?)01:30
onefangThat was looong ago, I'd have to search my email.  Though I do recall the maintainer at the time would often complain that he can't read code, and would tend to reject anything that wasn't fed to him in bite sized chunks.01:31
systemdleteSo he could handle bite-sized, but not byte-sized, software?01:32
onefangI know he rejected my text editor for that reason.01:32
onefanglol01:32
onefangAnyway, we are now off topic here.01:33
systemdletetrue01:33
systemdleteok, well, I might just have to force a symlink to avoid all of this...01:33
systemdleteI already know exactly where and how and what to do.01:34
rwpsystemdlete, I only skimmed the scrollback but you really should add those LSB headers to your initscript.  It isn't hard.  And then insserv will work.  If you don't then insserv will be complaining about it upon every run.  Just because systemd says LSB is dead doesn't mean it isn't a good idea for helping insserv do the right thing.02:12
onefangI agree.02:13
* onefang wanders off shopping, now that the shops are open again.02:24
systemdleterwp:  I tend to agree.  This will allow me to do things the "standard" way, which I much prefer.  I've gone the hack route in the past and usually ended up getting bitten some way or another.02:52
rrqsystemdlete: did you check "man insserv" ?05:35
rrqand while I'm at it: "man update-rc.d" ?05:40
rrq(given that you prefer doing it "right")05:40
brocashelmdoes the amd gpu firmware loading issue primarily affect excalibur/ceres? i use daedalus-backports and haven't had any issues, although i noticed i don't have /usr/lib/firmware (even just rebooted an install to check; all good)08:09
brocashelmdoes the kernel version also matter? i'm using 5.19 instead of 6.1-6.508:09
rrqthe kernel looks in /lib/firmware .. firmware packages for excalibur and ceres install their goodies into /usr/lib/firmware ... you can check the file lists in pkginfo.d.o08:12
brocashelmk08:17
brocashelminteresting. i installed the 20230515-3 build of firmware-amd-graphics (i think right around the time bookworm was released), so it's only available locally on my system08:19
rwpWhere did it install the firmware?  /lib/firmware or /usr/lib/firmware?  If it was /usr/lib/firmware and you have not converted to UsrMerge then the kernel will not find it as the kernel only looks in /lib/firmware and not /usr/lib/firmware.08:44
rwp"dpkg -L firmware-amd-graphics | less" will display where it installs things.08:44
brocashelmi never converted to usrmerge08:51
brocashelmusrmerge and usr-is-merged is blacklisted on all my apt systems08:51
brocashelmalthough i guess that means i haven't been affected08:52
rwpYou have not been affected as long as the firmware-amd-graphics you installed installs files into /lib/firmware and not the newer version of the package which installs into /usr/lib/firmware.08:53
brocashelmcorrect, i can confirm they are in /lib/firmware instead of /usr/lib/firmware08:53
brocashelmalthough the version i installed is from ceres just after daedalus was released08:53
brocashelmso it wasn't the case then08:54
rwpExcellent!  Then you should be good to go.08:54
brocashelm:D08:55
|cos|Darn! Now my ThinkPad get stuck on that "Waiting for /dev to be fully populated." persistenty. It first failed to wake up from suspend and now it fails to boot, given a handful of attempts.09:20
|cos|Rescue mode boots thankfully, but I'm not sure what to do next. No log entries in /var/log/* since pre-suspend yesterday.09:22
|cos|...except for boot.log, which isn't saying much useful.09:23
|cos|Given that smartctl expectedly reports Power_On_Hours for the only disk to be 178, I highly suspect this is a software issue rather than hardware.09:31
|cos|Rebooting after booting into rescue mode made the next normal boot, oddly enough.09:33
rwpThat's a strange error.  AFAIK it is waiting for hardware devices to settle at that point.  I would suspect that a full power off power cycle would reset things.09:37
rwpNormally with suspend it does not matter but is it possible there was a kernel upgraded in between?  I have seen strange things happen in that case.09:39
gnarfacecold boot might help, unless the issue is the power supply is failing09:41
rwpAlong with a failing battery.09:41
rwpI have two Thinkpad batteries that immediately drop the power if they are inserted into the laptop even with the laptop plugged into AC power.09:42
|cos|any ideas on how to increase logging verbosity? i did add ' udev_log="info"' to my kernel arguments already.10:11
|cos|my batteries might be flaky, at least the internal one. but those don't add anything to /dev do they? i though all acpi stuff was accessed through /sys?10:14
XenguyIs there a way to config/tune my fan to be less active?10:21
XenguyLike with hdparm, or some similar idea?10:21
XenguyHrm, $ apt-cache search thinkfan10:25
Xenguythinkfan - simple and lightweight fan control program10:25
brocashelmfancontrol, maybe?10:27
brocashelmcorectrl could also work (i use it more for my gpu)10:27
XenguyAppreciate the pointers, I noticed that I could do most things (except for say twitch/streaming) without problems, even when the fan was inactive10:28
Xenguyi.e. things did not actually overheat, except for more extreme streaming/video activities10:30
data41201Mashing Enter to bypass full disk encryption with TPM, Clevis, dracut and systemd - https://pulsesecurity.co.nz/advisories/tpm-luks-bypass12:19
onefangAll I have to do is mash the Enter key to bypass having systemd on my system?  Wish it was that easy.  lol12:24
gnarface|cos|: if the device isn't getting enough power but only barely, it could cause intermittent boot delays initializing various unrelated devices, it doesn't have to be the battery device itself causing the delay... it could be some usb device, or your soundcard, or... anything really. i think you're right that battery devices themselves typically only show up in /sys though12:26
gnarfaceXenguy: from what i understand, with modern hardware (fans with 3 or 4 pins) you have whatever fan speed controls are provided by the BIOS/EFI only. if you have older style fans where the speed is controlled directly by current, you can use the fancontrol package on some of those successfully, but it's not advised to do so and even the fancontrol docs themselves i believe basically say "don't use me, don't do this,12:29
gnarfaceyou're insane"12:29
gnarfaceXenguy: to be clear, if you have these style of fans they'll only have 2 wires leading to them12:29
[-_-]Hi guys14:02
[-_-]you don't use cdemu?14:02
[-_-]what else do you use?14:03
rustyaxeseems kinda pointless since you can just mount an iso somewhere14:07
[-_-]yeah, I did that, yet, it would be nice if it was managed14:20
[-_-]writing a mounting script is not very hard, I think the problem will be the unmounting script14:20
[-_-]I used fuseiso14:21
fluffywolfI've figured out why X stops responding after lid close...  xscreensaver is running.  but not displaying anything.  xscreensaver, xscreensaver-gfx, xscreensaver-systemd (!), and a random screensaver (cwaves in the test I just did), are all running, X is using cpu, etc...  but absolutely nothing is showing.15:04
fluffywolfnever seen this before.15:04
fluffywolfkilling xscreensaver from a vt restores the x session to normal functionality.15:04
onefangHas xscreensaver turned off your screen?  It does two things, blanks (runs the pretty graphics) the screen, and turns off monitors.  You can set it to turn off monitors after it has shown the screen "saver" for some time.15:07
onefangNo idea if it turns off the screen on lappies, I don't have one.15:08
onefangBut, it should turn on the monitor / screen again after input.15:08
fluffywolfit's very hard to know on a laptop...  the screen could be turning off entirely in hardware.15:17
fluffywolfwhen I open the lid, screen comes on, X looks normal.  mouse pointer moves, things being displayed keep moving, dialog boxes pop up, etc, but I can't interact with anything via either mouse or keyboard.  I can switch to a different VT normally.15:18
fluffywolftop shows xscreensaver is running, and everything is using ~50% cpu, suggesting it really thinks it's displaying pretty colors.15:19
onefangAh xscreensaver grabbing the mouse and keyboard.15:20
fluffywolffor now, I have fixed the problem by unchecking screensaver in xfce's startup settings thing.  heh.15:24
fluffywolfthis is a very stock install.  hitting a whole lot of problems for an install where I've pretty much not installed anything and configured nothing...15:26
fluffywolfI mean, I didn't even create a user account with my user name on this box, just a generic one with the name of the laptop, because I don't plan on doing anything on it.  lol15:27
fluffywolfI forget how much I hate desktop environments until I use one and promptly remember.15:28
onefangYou wont be needing the screen, keyboard, or mouse then.  B-)15:28
onefangI'm not even running a DE, just a window manager.15:30
fluffywolfsame here, on the machine I actually use.15:31
fluffywolficewm, to be specific.15:31
fluffywolfthat laptop is utterly terrible, and I just wanted a quick install for doing quick things on, mostly with a browser...  but I keep having to fix and fuck with shit.15:32
onefangAwesome for me.15:32
onefangRecently switched to it coz it's configured and scripted with Lua, and I love Lua.15:33
onefangPlus I needed hybrid tiled / not tiled.  One monitor is tiled, the other isn't.15:33
fluffywolfI also need to figure out how to permanently disable kernel lockdown.  also, how to beat the person who decided it should break everything and be on by default.15:35
onefangHow?  With a big stick, nine inch nails in the stick are optional.15:36
fluffywolflol15:36
fluffywolfwhat is the official way to modify the kernel command line parameters these days, now that there's huge blobs in /etc/grub.d?15:44
onefangNo idea, I don't use grub.15:44
fluffywolfah, looks like /etc/default/grub15:46
onefangSo I didn't know that "these days, now that there's huge blobs in /etc/grub.d".15:46
fluffywolflockdown breaks too much stuff, such as hibernation, to be default, imho...  it's a level of paranoia that people should need to actively want.  heh.  being unable to hibernate your laptop without non-obvious kernel command line options, for a theoretical security risk, is silly.15:51
fluffywolfthe argument is that if you allow a system to hibernate, someone with physical access to the system can then boot it off removable media and inspect the swap partition to extract secrets.15:52
onefangHibernate doesn't encrypt?15:53
fluffywolf...  people who have people after them who will be covertly gaining access to machines to hibernate them and extract secrets from the swap partition will know they have such people after them, and can fucking turn it on themselves.  lol15:53
onefangNot really, "Oh, Bob left his laptop in the lounge today.  He'll be at work for the next few hours.  Let's see what we can find."15:55
onefangOpportunistic scum exists.15:56
fluffywolfextracting secrets from the swap partition is non-trivial enough that the average company joker won't be doing it for a laugh.15:56
fluffywolffar more likely they'll just steal the laptop.  lol15:56
onefangJust saying, shit happens.15:56
onefangOr that.15:56
fluffywolfmeanwhile, hibernation is a feature that everyone has expected to work for like 25 years now, that suddenly doesn't, for very non-obvious reasons with a non-obvious diagnostic path and an even less obvious fix.15:57
onefangOne of my old clients had his laptop stolen.  Then wondered how some scam phone caller knew his wife's name.15:57
onefangCan we blame systemd?15:57
fluffywolfno15:57
onefangDOH!15:58
fluffywolfit's a compile time kernel option, that you have to disable by adding an additional kernel command line option.15:58
fluffywolfyou can find things like wife's names trivially googling people's names.  heh.15:58
onefangWell he did ask me how they might know, and this was not so long after it was stolen, so I mentioned it.15:59
fluffywolfthat's my first approach when I get a package with a bad address to deliver...  google the person's name and the town.16:00
onefangTrue, which is why my "secret security question" answers, which are often things like "What is your mothers middle name?", are randomly generated by my password manager.  Try looking THAT up on google.  Pffft16:01
fluffywolfheh, I got a great one last week...  it was addressed to 1929 on a street that I knew didn't have a 1929.  I googled his name, he lived at 1924.  went to 1924, asked if he was the person the package was addressed to, said yes, I mentioned it was addressed to the wrong address.16:01
fluffywolfhe walked down his driveway, looked at his 1924 address sign for a few seconds,...  "OHH!  29 is my PHONE NUMBER!"16:02
fluffywolfsigh...16:02
onefanglol16:02
fluffywolfI really, really hate those security questions...  I usually just enter random nonsense and assume I'll never be asked them.16:03
onefangMy new doctor had my street address as number 56, not 58.  There is no 56.  Not sure if it was my typo or theirs.16:04
fluffywolfmuch of the time these days they make you pick from a list of things none of which I have a correct answer to.16:04
onefangLike the local government site, which has mandatory questions from their fixed list, and they do ask each time you log on.  I have to use it at least once a fortnight.16:05
fluffywolfthey don't ask for mother's maiden name and such anymore, as they're too easy to look up.  now they're all "what color was the jewlery of the second person you asked to your highschool prom?" and other innanity.16:05
fluffywolfI had to sign up for a local government site last week...  I hope it never asks me, because I sure don't remember if I entered glashgsdhgsdf or lkjshfjkafhg....16:06
onefang"Favourite subject at school?"  Likely the one you did best at, there are records.16:06
fluffywolf(county is moving building permits and such to an online portal next month, had to sign up for it)16:07
onefangThat's why I used my password manager.   It's job is to generate long strings of random shit and remember them.16:07
fluffywolfmy password manager only does the first half.16:07
onefangAfter all, they are basically passwords.16:08
fluffywolfit generates long strings of random shit just fine, but I sure don't remember them.  :)16:08
onefangI'd suggest there's an app for that, but I doubt you can side load it into the "device" you run your password manager in.  B-)16:09
fluffywolfwell, there's always the hans niemann approach...16:10
onefangHave we stepped over #devuan-offtopic yet?  No idea who that is.16:10
fluffywolfchess master who was suspected of cheating by receiving instructions from a wireless vibrating buttplug.16:11
onefangAh I saw that headline, but have not read the article.16:12
fluffywolfI should probably figure out password management one of these days.16:12
onefangThe keepass famiily are great.  I use KeePaasXC.16:14
fluffywolfI use unique passwords everywhere, but low-importance ones are a stock password with site-specific letters in front of it...  prevents automated attacks nicely, but a determined human attacker specifically targetting me could figure it out.16:14
onefangAnd KeePassDroid on my smart phone.16:15
onefangIf I ever bothered to keep passwords on my Windows development VM, there's KeePass for that as well.16:16
onefangIt's 1 AM, my meds are kicking in, and I should probably try to sleep.  G'night.16:17
fluffywolfcyas16:18
Nietzhello @movedmactodev1 welcome!16:41
NietzIf you are having trouble, I suggest running the browser appimage16:41
Nietzlocated here https://codeberg.org/librewolf or as a flatpak16:42
NietzI like it as the latter for keeping the browser isolated from the system.16:43
fluffywolfor, you know, not?16:48
fluffywolfcontainers should fucking die.16:48
Nietzok.  just a moment please @movedmactodev1 .  i think i have more insight on the original issue16:51
golinuxWhat the heck was that about?17:01
Nietzno idea17:02
golinuxWrong forum?17:03
NietzI thought this was a posi place . . . back to @movedmactodev1 's question... I see that you say you manually updated the distro detection as a workaround17:03
Nietzif you are not using the proper signing key for that repo as well, it will fail to verify the signature17:03
golinuxHe is not here to see your response . . .17:04
Nietzbut it will be in the logs, right?  I check them all the time17:04
NietzI always look back your time @golinux has been so valuable.17:05
golinuxI see your point.17:06
golinuxBut according to the logs, he has never been on this channel http://reisenweber.net/irclogs/libera/_devuan/search?q=movedmactodev117:13
golinuxAnd this is probably OT at this point . . .17:13
Nietzok, thank you for keeping the chat focused.  Wishful thinking on my part.17:14
Nietzquestion on something I think is happening with users that I would like to help with, do you have a moment for a DM? golinux17:15
joerggolinux: http://reisenweber.net/irclogs/libera/_devuan/search?q=movedmactodev1%28%3Fi%2917:43
joerghttp://reisenweber.net/irclogs/libera/_devuan/_devuan.2023-12-21.log.html#t2023-12-21T09:20:4017:43
sixwheeledbeastThat script is terrible to be fair. it will be signing issue but that user is long gone from what it looks like.18:02
Nietzoh well.  I have seen a few users doing this in the forums.  Maybe it will help the future generations.18:36
joerg:-)18:41
joergNietz: much appreciated18:50
golinuxjoerg . . .I stand corrected. Don't know the other instances didn't pop up on my search . . .18:50
golinuxBut them my mind was focused on putting b'fast together . . .18:51
joergyes, you did a case sensitive search which alas is the default for the more capable regex search, I don't know how to make ignore-case the default while keeping all the versatility18:52
golinuxOy . . . bad typing day . . .18:52
rustyaxeDevuan has snuck into my router :o18:54
joergI found the "(?i)" can also be the first thing in search term, so maybe eventually I'll look into the code to have the search textbox pre-populated with a "(?i)" you could delete when you don't want it18:54
golinuxjoerg thanks for the explanation19:00
joergyw :-)19:00
golinuxI've never had a problem with it before . . . that I noticed . . .19:01
joergI changed that a few years back, previously it did a simple (case-insensitive) plain string-compare19:02
joergcode originated back to Marius Gedminas (mgedmin)19:03
joerghttps ://mg.pov.lt/irclog2html/19:05
XenguyThanks gnarface , I decided to uninstall fancontrol19:27
CERAPIOhello hello23:39
plasma41hello23:40
CERAPIOplasma41 They bother me because of my Windows7u nickname, ha, you have to upgrade to 523:55
CERAPIOI am Argentine, sorry if my joke has bothered you23:57
CERAPIOI'll be back, restart23:57

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