57[00:57:02] <randombit> has anyone any suggestion why might be
happening that every time i click on a menu or such (with system
sounds - so with click sound) the speakers make this loud
"""blob""" sound... i got this when
the system boots, yet it started to happening even during normal
usage... anyone?
67[01:02:27] *** Quits: freebench_ (~freebench@replaced-ip) (Remote host closed the connection)
68[01:03:25] <aaro> randombit: try disabling sound power save
69[01:03:46] <randombit> sound works, and even that
"event" sound works. only the speaker makes this
irritating ""blob"" sound-thing... it happens
during boot, and even with menu-clicking.
70[01:04:07] <randombit> sound power save? i dont know that one
71[01:05:19] <aaro> randombit: do you have this file
/sys/module/snd_hda_intel/parameters/power_save ?
105[01:46:47] <randombit> well im on pc... so it doesnt affect
me. i was really annoyed because it was not happening before. idk
when it changed, but while i was "ok" with the startup
""blobbing"" when the menu-clicking started to
do it... idk why i didnt ask earlier :D
106[01:48:29] <randombit> especially because i use subwoofer
too.. and with that it _is_ _really_ annoying.
107[01:48:48] *** Quits: kreyren (~kreyren@replaced-ip) (Remote host closed the connection)
307[06:07:42] <bolt> wikan: if you're trying to download a
movie from a streaming website, check the bug reports for your tool
of choice. That stuff is a whole cat-and-mouse game. The website
devs keep changing stuff around to confuse the downloaders, because
they don't actually want you to download the videos without
seeing all the advertising, and because it breaks their licensing
agreements.
321[06:20:20] <n4dir> youtube-dl can for for me or can not. But
if i start the video with vlc i got those 100 MB too, so besides not
commercials no advantage. And you probably thought of that already
322[06:20:33] <n4dir> can work for me. not can for me. Oh my god
449[08:43:44] <kats99> excuse me, I accidently overwrote a
usbdrive with zeros using dd. Now It's not recognised in
gparted and the space has shrunk from 4gb to 64mb. how do I recover
it?
466[08:56:45] <efloid> does anyone know if there is some issue
with Linux and certain M.2 drives? i have a drive that gives fatal
I/O errors when trying to mkfs.ext4 on a nvme partition. tried with
a 5.7 and a 5.8 kernel. windows is able to format the nvme partition
fine.
467[08:56:59] *** Quits: CyberManifest (~CyberMani@replaced-ip) (Remote host closed the connection)
468[08:57:51] *** Quits: jvwjgames (uid290762@replaced-ip) (Quit: Connection closed for inactivity)
572[10:06:25] <judd> No package named 'drbd' was found
in amd64.
573[10:06:43] <saveNexit> hello, I would like that the computer
hibernates when the battery is critically low, but it shuts down
instead, warning accordingly at the said battery level. UPower.conf
is set to "HybridSleep", and my device normally is capable
of hibernation (i.e. when I manually suspend it), so even if
"HybridSleep" is not available (which I dont know whether
is), it should be able to select hibernate as Upower.conf says, but
it doesnt.
608[10:13:11] <Haohmaru> then you power up the computer, it will
start booting (bios and whatnot) then when it goes to debian
it'll see that it's been hibernated, so it'll restore
the RAM contents from disk
609[10:13:41] <Haohmaru> roughly something like that
610[10:14:08] <azeem> right, you need to test that hibernation
works before you can bother about hybridsleep
611[10:14:09] <saveNexit> but after having a swap partition the
size of RAM, what else do I need to do? i.e. how can I have it
hibernate then?
614[10:14:29] <another> ratrace: actually, by default you need
about 2/5 of the RAM
615[10:14:30] <Haohmaru> in suspend, the RAM contents don't
get written to disk, because the RAM is kept powered, most other
things go in standby (low power or off)
616[10:15:02] <ratrace> another: how do yo figure that
617[10:15:05] <Haohmaru> saveNexit i'd first test if
hibernation works by manually doing a hibernation
620[10:15:33] <Haohmaru> i know suspend works on my computers
because i use it every day
621[10:15:36] <saveNexit> the question is, how do I do a
hibernation? where is that option on debian?
622[10:15:45] <ratrace> saveNexit: GNOME?
623[10:15:46] <Haohmaru> saveNexit from the.. start menu thing?
624[10:15:52] <saveNexit> ratrace, yes
625[10:15:59] <Haohmaru> aww, gnome
626[10:16:13] <Haohmaru> it SHOULD have a button in the menu
somewhere
627[10:16:15] <ratrace> ah, it's hidden, needs some
settings to show in the quick power options, lemmesee...
628[10:16:23] <Haohmaru> wut
629[10:16:24] <saveNexit> Haohmauru, on the start menu thing, I
dont even have "suspend", it just suspends when I press
power button shortly, or close the lid
630[10:16:32] <ratrace> right. hold down the ALT key and suspend
button will turn to hibernate before you click it
631[10:16:40] <Haohmaru> toss that gnome IMO ;P~
632[10:17:09] <jmcnaught> saveNexit: click on the system menu
(which has the power button) and hold Alt
633[10:17:45] <saveNexit> ah ok with alt held, power icon turns
to suspend
634[10:17:56] <ratrace> and hovering it says hibernate?
635[10:18:12] <saveNexit> it says nothing on hover
636[10:18:27] <ratrace> does the button icon look like download?
arrow downward to disk?
637[10:18:44] <Haohmaru> aww, who knows his icon theme
638[10:18:48] <saveNexit> ratrace, no it is the
"pause" symbol; like 'II'
639[10:18:55] <Haohmaru> maybe use commandline for this
640[10:18:57] <ratrace> not sure that's it then
641[10:19:07] <saveNexit> it is the default one, I have no theme
of my own
642[10:19:19] <Haohmaru> or install lxpanel and run it
647[10:20:55] <Haohmaru> weren't these commands to
systemctl btw?
648[10:20:59] <saveNexit> still dont understand, when I have its
matching swap space, will I magically have "somewhere" on
my desktop a hibernate option?
653[10:22:48] *** Quits: BaffinMan (~BaffinMan@replaced-ip) (Remote host closed the connection)
654[10:23:08] <kats99> That showed up while installing a package
and the same error is shown when I restart sshd.service..says Unit
-.mount is masked..Unmasking it doesn't work
655[10:23:13] *** Quits: Kuttenbrunzer (~Kuttenbru@replaced-ip) (Remote host closed the connection)
656[10:23:24] <ratrace> another: that doesn't make sense
and that probably means something else. where does all the RAM go
when you power off? compression? how do you guarantee compression
ratio? you can't...
708[10:32:53] <ratrace> saveNexit: when you hibernate, your RAM
goes to disk, to swap partition, plain and simple. so you need swap
size to be RAM size. ignore all other edge theoretical maybe somehow
whatever use cases. swap >= RAM . done.
709[10:33:21] <Haohmaru> at least
710[10:33:31] <saveNexit> back to above; where will I find
"hibernate"? will the "Alt+power" show
"hibernate" then?
711[10:33:34] *** Quits: CyberManifest (~CyberMani@replaced-ip) (Remote host closed the connection)
712[10:33:41] <saveNexit> or even if it doesnt show, will it
work like that?
713[10:33:50] <Haohmaru> maybe there should be some kind of a
dedicated "swap-like" partition specifically for
hibernation
722[10:35:09] <r0b0> hello, what is creating and directories in
/run/user/NNNN and mounting tmpfs there? and how can I have schroot
bind mount them to my schroot jail?
723[10:35:18] <saveNexit> all right, I will try these, thank you
all so very much!
724[10:37:38] <ratrace> r0b0: PAM and systemd integration create
those. what _exactly_ are you trying to achieve here?
726[10:38:47] <r0b0> ratrace, I run a centos in a schroot jail,
it works fine, just /run/user/1000 is drwx------ root root instead
of owned by my user id
727[10:38:57] <kats99> ah yes I ran raspi-config on debian host
accidently and get this error when I try to install any new packages
Error:GDBus.Error:org.freedesktop.systemd1.UnitMasked: Unit -.mount
is masked.
728[10:39:04] <another> Haohmaru: yop. but if you're using
both RAM and swap near 100%, you have other problems
729[10:39:17] <ratrace> r0b0: /run/user inside the container is
root?
730[10:39:31] <efloid> i'm wondering if one uses a nand
device as the primary system drive is it better to put logs on the
mechanical drive?
731[10:39:31] <another> realisticly nobody does that
732[10:40:24] <ratrace> another: there's no rule of thumb,
it can be any range of usage. thus swap_size >= RAM.
733[10:40:33] <r0b0> ratrace, /run/user is drwxr-xr-x root root
in both debian and in the container, that's OK, just
/run/user/1000 is wrong - I guess that's because it's
another tmpfs
734[10:40:37] <themill> r0b0: if you want this for just one
user, you can add "/run/user/1000 /run/user/1000 none rw,bind 0
0" to /etc/schroot/desktop/fstab
735[10:40:59] <ratrace> r0b0: that's what I meant yes,
/run/user/1000 ... so you're talking about that dir _inside_
the container or on the host?
736[10:41:14] <r0b0> ratrace, on the host it is correct
737[10:41:24] <r0b0> ratrace, in the container it is wrong
738[10:41:46] <r0b0> themill, ok, I guess that will do it for
one user - enough for my case
739[10:41:57] <Haohmaru> another what problem?
740[10:42:01] <themill> (assuming that your schroot is
type=desktop)
741[10:42:09] <r0b0> themill, it is
742[10:42:13] <Haohmaru> it happens to me often, ram goes full,
swap starts going up to the ceiling too
743[10:42:20] <ratrace> question is do you actually want to bind
host's with all the data and tokens in it, or you just want a
plain /run/user/1000 owned by non-root uid inside the container?
744[10:42:36] <efloid> here's a useful chart on recommended
swap sizes:
replaced-url
745[10:42:41] <r0b0> ratrace, i want to bind it - it's the
same user
746[10:42:48] <Haohmaru> there'd be a problem with
hibernate in this scenario, no matter how big the swap is
747[10:42:50] <ratrace> r0b0: okay
748[10:42:57] <r0b0> preferably bind all of the users if
possible
749[10:42:58] <another> ratrace: if you want to make it
foolproof, then yes. go with the simple answer. if you know about
your usage pattern, you can optimize
750[10:43:01] *** Quits: saveNexit (~saveNexit@replaced-ip) (Remote host closed the connection)
752[10:43:23] <ratrace> another: why wouldn't you want to
make it fulproof? wth? if you want to hibernate, swap_size >=
RAM. why are we even having this nonsense discussion?
753[10:43:29] <another> Haohmaru: performance? looming oom
killer?
754[10:43:37] <ratrace> who cares about edge cases and MAYBE
situations and blah blah blah. wanna hibernate? swap_size >= RAM.
done
755[10:43:42] <Haohmaru> another sure.. what can i do tho
761[10:44:36] <oiaohm> another: to be currect really large swap
can result in oom killer not working so instead you getting stuck in
icestation thrashing between swap and ram.
762[10:44:51] <Haohmaru> i got 8GB ram at home, and also 8GB
swap i think
774[10:47:34] <ratrace> I currently have 4GB of RAM available on
this machine, because 12GB is locked into hugepages for Windozer VM.
FF works just fine, there's actually 2.3 free and swap uses
maybe 100MB
775[10:47:42] <oiaohm> ratrace: chrome can still eat you out of
memory.
776[10:47:49] <another> ratrace: *I* care about edge cases, for
example.
777[10:47:51] <ratrace> oiaohm: you don't sa
778[10:47:53] <ratrace> y
779[10:48:02] <oiaohm> I use both chrome and firefox
780[10:48:17] <Haohmaru> ratrace i usually open a lot of .pdfs
with firefox
781[10:48:17] <oiaohm> And both have been causes of my 4G of ram
system having memory problems.
784[10:49:01] <Haohmaru> and there are some websites that make
firefox sh*t on its pants, you may close those pages afterwards -
but the memory won't be freed
785[10:49:01] *** debhelper sets mode: +l 1136
786[10:50:05] <oiaohm> Haohmaru: I have seen the same thing with
chrome as well. I have got to know how to restart both firefox and
chrome really well.
787[10:50:16] <Haohmaru> oiaohm basically, it doesn't
matter how much RAM you have.. go to imgur and have fun there for
enough time and your RAM will get filled
789[10:50:44] <Haohmaru> i got computers with RAM from 2GB up to
8GB
790[10:50:45] <ratrace> another: right, but those edge cases
aren't generalizable to everyone. you can't know how much
RAM someone needs and uses at any given point, so any answer other
than swap_size >= RAM is wrong and misleading
794[10:51:51] <Haohmaru> ratrace the answer (afaiu) should be
"to hibernate, you gotta have swap, and when you go to
hibernate, you better have enough free space in the swap to hold the
used RAM"
795[10:52:08] <makara1> i heard swap should be double your RAM
size
803[10:53:20] <ratrace> with RAM regularly being 2GB or more,
that rule of thumb really doesn't apply
804[10:53:28] <Haohmaru> ratrace yes, certainly the swap should
be at least as big as the RAM, but my answer sorta includes it
805[10:53:48] <ratrace> Haohmaru: it does but vaguely.
806[10:54:13] <Haohmaru> i'd explain it ;P~
807[10:54:35] <oiaohm> makara1: in a system like mine with 32G
of ram I have just over 32G of swap. Once I start using swap I
normally have performance trouble.
808[10:54:56] <Haohmaru> i really think there should be a
dedicated partition for hibernation, if anyone seriously wants to
rely on it
809[10:55:08] <Haohmaru> then things would be simpler
817[10:57:19] <ratrace> Haohmaru: note one more thing, as oyu
can't change swap size on the fly (it's locked upon
creating the swap partition or file), then RAM usage at any given
moment is irrelevant and useless information, because you can't
alter swap to adapt (unless it's a swapfile and you
swapoff&swapon and blah blah, but you get the gist of it)
830[11:01:14] <another> ratrace: you can change swap size on the
fly on lvm
831[11:01:17] <Haohmaru> so, do you agree with me that perhaps a
dedicated partition for hibernation could maybe solve this?
832[11:01:36] <ratrace> another: not without swapoff&swapon
833[11:01:51] <oiaohm> makara1: welcome to horrible over those
traceroute. traceroute 2.1.0-2 is from 2016-03-08
inetutils-traceroute version 2:1.9.4-11 is from 2015-06-10 Basically
neither one is highly maintained.
835[11:02:21] <oiaohm> makara1: you would expect the slightly
newer one to be better.
836[11:02:23] <Haohmaru> oiaohm or maybe they work flawlessly
already?
837[11:02:24] <ratrace> Haohmaru: frankly I don't see the
difference, as the same problem would exist: how much.
838[11:03:06] <Haohmaru> ratrace i mean, then you'd have a
dedicated partition for hibernation, you'd only have to make
that one big enough to fit both the RAM and swap
839[11:03:08] <another> sure. but you can add a second,
temporary swap partition
840[11:03:10] <ratrace> makara1: mtr-tiny is what I use for all
the tracerouting needs of mine
841[11:03:10] <Haohmaru> right?
842[11:03:30] <oiaohm> Haohmaru: I would say not thinking
traceroute 2.1.0-2 the -2 there are debian fix to faults that are
not fixed updstream so have not been properly peer reviewed.
843[11:03:33] <ratrace> Haohmaru: right but as another said,
that's just like having two swap partitions
844[11:03:35] <ElNomReal> Why not leave the swap in swap and
have a partition just for the ram for hibernate?
845[11:03:47] <ratrace> because it wouldn't matter
847[11:04:04] <Haohmaru> ratrace yes, except the 2nd one is
never gonna be used as swap, it'll be empty untill hibernation
is wanted
848[11:04:38] <ratrace> Say your current RAM usage is X for
active pages, Y for unused. your swap size has to be >=Y for
non-hibernate situation, and >= X+Y for hibernate situations. if
you split those into two, you still have X + Y total space on disk
needed
852[11:05:23] <ratrace> so even if you designate one for
"active pages from RAM when I hibernate" and the other for
"unused pages from RAM in normal operation" you end up
with exactly the same on-disk space demands
853[11:05:31] *** Quits: saveNexit (~saveNexit@replaced-ip) (Remote host closed the connection)
855[11:05:47] <oiaohm> Haohmaru: you can run into the current
nightmare where hybernate fails because you have used enough swap
that there is no enough space in swap to dump memory to swap.
856[11:05:50] <ElNomReal> It seems to be just a matter of, if
you want to use hibernate like this, make sure that the ram can
always fit into the free space of swap, or if not when you
hibernate, remove the oldest swap until it fits?
858[11:05:59] <Haohmaru> okay, my coffee levels are too low for
this discussion, i quit
859[11:06:26] <oiaohm> I would like to see some better swap
management.
860[11:06:29] <ratrace> Haohmaru: in such technical discussion,
"swap" as a single term disappears. you really must talk
about active pages and unused pages.
862[11:06:51] <Haohmaru> yeah, i'm unfamiliar with the
details, so i quit
863[11:07:07] <Haohmaru> i've never needed hibernate so
far, suspend is gud enuf
864[11:07:22] <ratrace> because hibernation is really just
swapping out 100% of active pages. so if you already have unused
pages ther due to overcommit, you run into trouble if your
overcommit is such that active+unused > RAM >= swap_size
865[11:07:25] <Haohmaru> i don't have/like laptops or
battery-powered crapputers
866[11:07:47] <oiaohm> There is a horrible one. No swap at all
with Linux kernel can cause increased memory fragmentation because
new memory defragmentation methods with the Linux kernel use push to
swap as part of the process.
870[11:09:20] <oiaohm> I do have systems with zero disc based
swap but with a small ram based swap instead so keeping kernel
defragmentation of memory happy.
871[11:09:21] <ratrace> Haohmaru: well the
tl;dr;wheres;my;coffee answer is: the linux kernel doesn't
differentiate "hibernation RAM". it's just a
"normal" swap-out operation where swappiness is turning
from heuristic into "PAGE THEM ALL OUT!!"
873[11:10:30] <oiaohm> Items you have in swap stuff you have
hibernated. stuff that has had to be swapped out. Stuff the Linux
kernel has sent to swap because it currently moving its location in
ram thinks nothing going to access it.
874[11:10:39] <ratrace> as for memory fragmentation, it's
bigger problem than one may think.
876[11:12:53] <oiaohm> ratrace: Yes memory fragmentation is a
really nasty problem. Thinking zero swap equals Linux kenrel using
less methods to deal with it the problem is now even worse with swap
disabled.
881[11:16:11] <oiaohm> I have like 32 and 128G of ram systems.
Stall hard is normally not my issue from no swap. Memory
fragmenetation resulting in massively degraded performance that
magically disappearing as soon as you issue 128meg zram swap as in
swap in ram.
882[11:16:47] <oiaohm> There are many ways no swap can get your
ass with a Linux kernel.
887[11:21:30] <ratrace> another: one problem I have with VMs and
frag is reserving hugepages of 1G. so I have to reserve them on
boot, because there's no guarantee I'll find enough
contiguous 1G blocks even with several GB apparently
"free"
888[11:21:55] <ratrace> (several GB _more_ than what I need for
hugepages)
889[11:22:03] <oiaohm> another: it can. Its one of the things
you need to check your vm workloads for.
890[11:22:28] <oiaohm> Yes a degraded VM hosted in the cloud can
be expensive for pocket.
891[11:22:34] <Haohmaru> luckily i don't use virtual sh*t
892[11:22:52] <Haohmaru> it sounds like a big hack ;P~
893[11:23:10] <H4ndy> Baremetal virtualization is great
894[11:23:19] <ratrace> Haohmaru: VMs are a great tool. solves
so many problems...
895[11:23:34] <hwm4rgs> VMs are OP. So are containers.
896[11:23:39] <Haohmaru> i've tried it once or twice.. to
run linux on crapdows, with 2GB RAM on the computer
897[11:24:29] <oiaohm> Haohmaru: yes the push to swap by the
Linux kernel when defraging memory is a hack to provide a safe
guard. When you are moving memory you are trying to make sure
nothing is accessing at the time. Push to swap if something has not
be isolated yet and it access it the swap system is able to bring
the memory back.
898[11:24:39] <ratrace> Haohmaru: I'm using VMs to run
crapdows on linux :) I wouldn't allow crapdows to touch my
hardware except pci passthrud' dedicated hardware, for no
amount of money in the world!
902[11:24:59] <Haohmaru> i have 3 computers at home, the oldest
one isn't used, the bestest one has debian10 only, the middle
one has crapdows and debian10
903[11:25:35] <Haohmaru> the goal is to get rid of the
crapdows.. my music-related stuff is holding me back
904[11:25:45] <Haohmaru> looking into resolving that
906[11:26:26] <ratrace> wish I could get rid of it, but alas I
can't. what seemed like linux gaming picking up on steam in the
past two years, melted away and the future of it looks very, very
bleak.
908[11:26:36] <oiaohm> With music stuff I am watching pipewire
and wanting it complete to finally fix the jackaudio/pulseaudio
issue and finally get down to 1 sound server.
928[11:30:51] <Haohmaru> i wanna make new plugins too, i
don't want to make them windows-only thus i've paused for
a long time untill i abandon the crapdows ship
929[11:31:04] *** Quits: BenNZ (~Ben__@replaced-ip) (Quit: Everytime I think IQ's must have dropped recently, I
remember that this is the internet)
943[11:36:52] <Haohmaru> i gotta look into this ardour thing to
see if it's good enough for my usage, or else i'm thinking
to buy Renoise which afaik comes for linux too
944[11:37:40] <Haohmaru> i know renoise is eggcelent, but needs
learning
946[11:39:30] <oiaohm> another: its not easy to metric.
Different partly filled structs that could be merge could be causing
memory fragmentation. Unless you know those structures should be
merged they can look no fragmented just using a lot of memory.
947[11:41:14] <oiaohm> another: there are a stack of different
memory defragmentors in the Linux kernel each designed to deal with
different structures the Linux kernel users. Basically there is no
simple metric.
951[11:43:05] <oiaohm> Normally I do a test comparing with and
without swap if I see major difference I know I need to dig though
all the different metrics for the difference between the two runs.
952[11:43:45] <oiaohm> and that will normally lead me to what
ever defragmentor was not running to note down for when it improved
in the Linux kernel source to no longer need swap.
985[11:46:11] <Haohmaru> coz when even the swap runs out, you
can get a situation when you can't even kill stuff, you
can't get a terminal or taskmanager
986[11:46:16] <another> oiaohm: do you have any more reference
material ?
999[11:47:27] <oiaohm> another: there was a video from Linux
conference Australia covering it and I just have a stack of personal
collected notes from being aware the beast is there.
1013[11:52:22] <oiaohm> So yes some of those odd behavours you
can get out of vm and baremetal at times when you have no swap can
lead to the fact you have disabled one or more of the memory
defragment systems.
1022[12:04:19] <solrize> hi, i want to run pdftotext and all docs
say it is in poppler-utils but i have that installed and pdftotext
isn't there. any idea where to look next? i'm on buster
1023[12:05:20] <solrize> hmm now it's there. weird
1024[12:05:28] <solrize> i did apt-get update etc
1025[12:06:28] <ElNomReal> Hmm, There's likely something
that had to be restarted.
1086[13:05:43] <TuxCrazy> on Virtual Machine Manager, I am not
able to access other partitions on my computer's HDD.
1087[13:05:51] <TuxCrazy> please help
1088[13:06:08] <ratrace> TuxCrazy: what do you mean other? give
an example please
1089[13:07:23] <TuxCrazy> I am able to create a virtual machine
with an iso that is copied on the Home partition only. But, if the
iso is on any other partition, I can't create that virtual
machine.
1090[13:08:54] *** Quits: greknod (~greknod@replaced-ip) (Remote host closed the connection)
1093[13:10:31] <TuxCrazy> It says that - the emulator may not
have search permissions for the path.
1094[13:11:09] <ratrace> TuxCrazy: well yes, if your ISO is in a
location not accessible to the user the VM is running as, it
can't be read
1095[13:11:10] <lovemytux> @TuxCrazy: check the permissions to
the relevant directory then
1096[13:11:32] <ratrace> not just location (parent directory),
but all parts of that path, including the ISO itself.
1097[13:11:33] <TuxCrazy> lovemytux, how
1098[13:11:36] <TuxCrazy> ?
1099[13:11:46] <ratrace> with ls
1100[13:12:19] <TuxCrazy> ratrace, could you specify the full
command?
1101[13:12:21] <ratrace> TuxCrazy: or better yet, start with
learning the unix permission model and how it relates to files and
directories. then with that knowledge, check if the path of the ISO
is readable to the VM user
1110[13:14:23] <ratrace> TuxCrazy: on linux "user" is
more than just a physical person running the machine
1111[13:14:35] <TuxCrazy> ratrace, ok
1112[13:14:56] <ratrace> again, please learn about the unix
permission model. good place to start, the holy grail of linux
wikis, archwiki:
replaced-url
1165[14:00:11] <oxek> I messed up. I need to purge firefox and
reinstall it. When I do 'apt purge firefox-esr', it not
only want to purge firefox, but it also wants to install
epiphany-browser and other packages.
1166[14:00:23] <oxek> how can I just purge firefox and not
install anything else?
1167[14:01:22] <bendem> does it list them under unused
dependencies? If so you can `apt install` them to mark them as
manually installed
1168[14:01:38] <nvz> oxek: you need to make a package that
provides
replaced-url
1169[14:01:51] <nvz> oxek: or uninstall any and everything which
depends on a browser
1204[14:07:24] <nvz> if it wants firefox-esr with equivs you can
give it firefox-esr or w/e it wants
1205[14:07:39] <nvz> you can very easily make an empty package
with whatever headers you like
1206[14:07:56] <nvz> install that package, the system is none the
wiser
1207[14:10:47] <nvz> however you should note that equivs was
designed to tell the package manager you've satisfied the
dependency in your own way.. if you /haven't/ some things may
not work properly
1208[14:11:03] <oxek> I'll see what I can do with equivs
1209[14:11:11] <nvz> the maintainers dont put depends lines in
there just do be aggrivating
1210[14:11:12] <oxek> and hopefully do stuff in a way that can be
reverted 30 seconds later
1211[14:11:26] <jelly> package A that only Recommends B
won't get removed if B is removed.
1212[14:11:46] <oxek> in this case, the maintainers have not
considered that I want to remove firefox and immedially install it
after
1215[14:12:49] <nvz> yeah.. if you want to resintall it from the
repos and not manually like compling it or something.. then just do
it all in one step
1216[14:13:06] <azy> hello! im trying to compile
'cherrytree', but i get the error: Requested
'glibmm-2.4 >= 2.64' but version of glibmm is 2.58.0. i
have a newer version it seems. what should i do? debian10
1217[14:14:03] <nvz> ,v glibmm-2.4
1218[14:14:04] <judd> No package named 'glibmm-2.4' was
found in amd64.
1226[14:14:47] <jelly> oxek, use a lower level tool to purge a
package if you need that might help things, and then let apt correct
the state afterwards
1228[14:14:51] <n4dir> there is an old thread at
forums.debian.net how to deal with the problem at hand as far
metapackages are concerned. If it is good advice or not i
wouldn't know
1231[14:15:37] <jelly> oxek, in your case, it would be a RARE
example where usage of --force- options for dpkg is warranted. dpkg
--purge --force-depends firefox-esr
1232[14:15:39] <ratrace> but the program needs 2.64 or newer,
while buster packages 2.58
1237[14:16:43] <jelly> azy, you would need to backport the newer
glibmm-2.4, and it's hard to say how that might break other
software that uses the same library
1247[14:19:20] <ratrace> no rabbit holes. just use a sid
container.
1248[14:19:32] <azy> should i use sid instead of buster in
general?
1249[14:19:35] <nvz> sorry, I was in the middle of checking the
other package when you made me laugh at the idea you think something
that requires a wrapper for a toolkit can magically work without it
1250[14:19:46] <nvz> azy: no, probably not
1251[14:19:55] <ratrace> containers were invented fora reason you
know :) it's bad when people overabuse them for things that
don't really require them, but if you do? there they are.
1252[14:20:19] <nvz> azy: worst case scenario you build it in
some kind of chroot or container
1253[14:20:35] <alex11> sid means continual updates and potential
breakage compared to stable
1261[14:21:46] <ratrace> nvz: because fighting with dependencies
is worse than debootstrapping a sid container and building/running
your app like that.
1262[14:22:29] <ratrace> and glib is very much used in many
places on a typical linux desktop, so yeah... or else you'd
"slot" it (build with different soname, in parallel with
existing glibmm)
1286[14:31:22] <oxek> I got the equivs thing to work, so that I
no longer get offered epiphany-browser when I purge firefox. Went
ahead with the purge, and apt happily did: update-alternatives:
using /usr/bin/chromium to provide /usr/bin/x-replaced-url
1287[14:31:30] <oxek> so why did apt want to install epiphany...
1288[14:31:39] <oxek> when I already have both lynx and chromium
1294[14:33:06] <ksk> oxek: is it installed right now? Then you
can ask aptitude why it was installed: "aptitude why pkg"
(Didnt we talk about that in here like three weeks ago??)
1295[14:33:24] * nvz said it less than an hour ago
1296[14:33:43] <n4dir> pretty much the only reason why i install
aptitude at all
1306[14:35:23] <oxek> ksk: I don't remember if I discussed
this 3 weeks ago.
1307[14:35:24] <nvz> ratrace: I only use FF to appify things, I
use a custom userChrome.css to make icons for things like netflix
and have them run in FF with absolutley no UI.. no scrollbars or
anything
1308[14:35:28] <n4dir> i'd have a look at the page at
forums.debian.net anyway. This won't be the last time the
problem occurs.
1327[14:38:08] <ratrace> I nver install package python libs,
unless they're automatic dependency for something else. all my
python apps are managed in virtualenvs, and installed with deps from
pypi
1328[14:38:31] <nvz> I've used pip but not virtualenv yet
1329[14:39:01] *** debhelper sets mode: +l 1153
1330[14:40:15] <jelly> ratrace, well that's just giving up!
1331[14:40:35] <jelly> might as well curl | sudo bash -
1332[14:41:05] * nvz curls jelly
1333[14:41:14] <ratrace> not at all. gives me the ability to fine
tune deps and upgrade them when and if I want.
1334[14:41:43] <nvz> ew, anyone have a wet wipe? :P
1335[14:41:45] <ratrace> waiting for debian to package something
up is not good enough for python in our case. we're a python
shop, our apps are in python and we need 100% control of the
ecosystem.
1434[15:03:58] <ksk> ratrace: "it does not work" if
your email localpart matches procmail rcs its "I forgot"
Variable - somehwere on bugs.d.o that is stated..
1435[15:04:16] <ksk> (and "mail" is in that variable on
a default debian config)
1436[15:04:19] <oxek> in fact, there are no dirs created by the
firefox-esr package that would just be called 'firefox',
it's always 'firefox-esr'
1437[15:04:22] <ratrace> ksk: oof!
1438[15:04:32] <ratrace> even more reason bts is stoopid.
1439[15:04:53] <ksk> some grey cat told me its stuped to have an
email with "mail@" ;)
1442[15:05:53] <nvz> they're a lil grumpy sometimes too
1443[15:06:01] <oxek> hmm, perhaps another bug - firefox-esr
package creates /etc/firefox-esr/firefox-esr.js in its postinst, but
it does not get loaded because firefox expect the dir
/etc/firefox/firefox-esr.js
1444[15:06:15] <oxek> and /etc/firefox-esr/firefox-esr.js
contains settings that debian should load by default
1445[15:06:18] <ratrace> ksk: may be, but bts should have no
business policing what's left of @ as long as it's RFC
compliant.
1451[15:07:38] <oxek> nvz: that would be great, but I think I
should first somehow verify if firefox 78esr (the version coming to
debian-stable this month) behaves differently
1452[15:07:57] <ratrace> 78 \o/
1453[15:08:06] <oxek> unforutnately, 78esr is not maintained in
testing
1454[15:08:13] <oxek> so I don't know how to get 78esr into
stable
1455[15:08:23] <oxek> upstream package of course uses different
dirs
1456[15:08:31] <nvz> oxek: I'd think if the maintainer made
this kinda mistake they are just totally overlooking this issue..
its an oversight
1457[15:08:41] <nvz> and they'll probably continue doing
things this way
1458[15:09:04] <Haohmaru> if no one reports anything..
1459[15:09:51] <nvz> I don't use FF and I dont know what
these things do.. but I could very simply explain that they're
creating a path that firefox doesnt actually use.. or forward a
pre-made report
1461[15:10:20] <Haohmaru> nor do i use policy stuff, i don't
even..
1462[15:10:29] <oxek> I think it's pointless to report this
for the current firefox-esr, since the patch would never land
because a newer esr is coming soon
1463[15:11:06] <nvz> yeah well I think that way sometimes too
with bugs in stable.. but if they don't know about it, it can
just keep happening
1464[15:11:08] <oxek> and since I can't see where the newer
esr is maintained, I think it's best if i just wait and then
test on newer version
1465[15:11:12] <oxek> I have my workaround in the meantime
1466[15:11:21] <nvz> and this def seems like one of those things
that is gonna keep happening
1468[15:12:23] <nvz> as for the comments on email and the bts.. I
/hate/ email but I love the simplicity of using it for the bts :P
1469[15:13:09] <oxek> it's probably nice & simple when
it works, but when it doesn't it's annoying. Plus it
definitely is a barrier for new devs & maintainers.
1472[15:13:56] <nvz> I've always just used gmail to shoot a
simple email with pseudoheaders to submit@bugs.debian.org and get a
nice confirmation email back.. real painless
1473[15:14:15] <oxek> I found out that there's launchpad
section for debian packages, and the maintainer for ubuntu &
debian is sometimes the same person, so I might use that one day.
1478[15:15:12] <oxek> nvz: ah, that helps me - so there's a
confirmation email sent back when a bug is posted? I never received
one when I managed to send it.
1497[15:19:46] <nvz> wth is that about? debian-edu-config?
1498[15:19:53] <dka> I am on debian, each time I want to use the
sound, I need to type: pulseaudio --start in the terminal, how can I
make it persistent accross restart ?
1499[15:20:06] <dka> I have tried sudo systemctl enable
pulseaudio, but it tells me pulseuadio.service does not exist
1500[15:20:31] <nvz> dka: are you using a DE or some kind of
Xsession ?
1506[15:21:44] <oxek> nvz: it might be a big problem, since
firefox might not be respecting various cert overrides and other
things, if they are to be in /etc/firefox instead of
/etc/firefox-esr
1519[15:23:58] <n4dir> dka: i am not much in systemd or such in
general, but i just ran in something where i had to do
"systemctl --user status <name-of-service" as user
1520[15:24:02] <ratrace> afaik PA is socket-activated, and
sockets are used by programs that want PA
1521[15:24:07] <n4dir> you might wanna investigate in that
direction
1522[15:24:14] <dka> and it stopped working will I upgraded my
kernel, I did a bad manip and removed the whole kernel, I had an old
one and I was able to reinstalled the removed dependencies not on
purpose
1523[15:24:17] <dka> but I lost this
1524[15:24:23] <n4dir> at least here doing that with pulseaudio
and status does give a result
1541[15:26:43] <nvz> dka: perhaps check running as your user
systemctl --user status pulseaudio
1542[15:26:47] <dka> sudo systemctl --user enable pulseaudio
return Failed to connect to bus: No such file or directory
1543[15:26:51] <jelly> huh.
1544[15:27:01] <dka> Failed to connect to bus: No such file or
directory
1545[15:27:03] <jelly> how are you starting X, dka
1546[15:27:04] <nvz> dka: you don't use root when doing
stuff for your user's services
1547[15:27:21] <ratrace> yeah don't use sudol
1548[15:27:23] <ratrace> *sudo
1549[15:27:23] * n4dir writes down dbus-x11 and dbuser-user-session to
check later if that stops qjackctl from starting jack
1550[15:27:43] <dka> systemctl --user status pulseaudio => ●
pulseaudio.service, Loaded: masked (Reason: Unit pulseaudio.service
is masked.), Active: inactive (dead)
1551[15:27:50] <dka> systemctl --user enable pulseaudio =>
Failed to enable unit: Unit file
/etc/systemd/user/pulseaudio.service is masked.
1552[15:28:03] <jelly> are you using a display manager of some
sort (the service providing a graphical login screen), or startx, or
what?
1553[15:28:08] <nvz> I've read about masked services before
but I dont recall :P
1554[15:28:15] <dka> I use the default cinnamon install
1555[15:28:22] <Haohmaru> nvz pandemic measures are getting out
of hand
1556[15:28:25] <ratrace> if it's masked, then I'm
guessing that didn't happen by itself
1557[15:28:29] <nvz> Haohmaru: hah
1558[15:28:32] <jelly> !which dm
1559[15:28:32] <dpkg> To figure out which Display Manager you are
using, open a terminal and type the following command: `systemctl
status display-manager` see also <which de> and <which
x>.
1560[15:28:59] <nvz> whoo.. someone uses my factoids :P
1562[15:29:12] <dka> As I said, I once removed the kernel and
some other packages not on purpose, I was able to start up using an
old kernel, I upgraded the kernel, and I reinstalled the package I
removed I could think off like cinnamon, but sound was not starting
up afterwards
1563[15:29:21] <dka> I need to manually run pulseaudio --start
1583[15:37:31] <ratrace> "Error opening PCM device hw:0,0:
Device or resource busy" -- you sure something else is not
running and (ab)using ALSA directly?
1588[15:39:55] <nvz> dka: the masked service is probably due to
there being a symlink in /usr/lib/systemd/pluseaudio.service to
/dev/null as to /why/ this happened idk
1589[15:40:07] *** Quits: nsegkos (~Thunderbi@replaced-ip) (Remote host closed the connection)
1590[15:40:14] <nvz> dka: simply removing that symlink if it
exists as such pointing to /dev/null would unmask it
1591[15:41:13] *** Quits: niko (~niko@replaced-ip) (Ping timeout: 622 seconds)
1592[15:41:17] <nvz> you'd need to then do systemctl
daemon-reload for the changes to take effect while its running
1593[15:41:36] <n4dir> jelly: did you mean that the problems
might occur if not running a display manager?
1594[15:42:07] <n4dir> cause as said i have similar problems, but
didn't investigate any further yet. Only a message bout
"dbus" got stuck
1595[15:42:36] <ratrace> I think daemon-reload is not required
for symlink modifications (for en/dis/abling or masking)? Only for
modifying unit files
1596[15:42:46] <dka> no sym link => -rw-r--r-- 1 root root 915
Aug 15 2019 /usr/lib/systemd/user/pulseaudio.service
1645[15:55:46] <nvz> how/why it got masked in the first place is
plenty curious, but also irrelevant to the issue
1646[15:56:11] *** Quits: xmattx__ (~matt@replaced-ip) (Remote host closed the connection)
1647[15:56:31] *** Quits: saveNexit (~saveNexit@replaced-ip) (Remote host closed the connection)
1648[15:56:36] <nvz> what seems painfully obvious to me is
somewhere in one of systemd's paths you got a file called
pulseaudio.service linked to /dev/null and your user doesnt have
permissions to it
1649[15:56:53] <nvz> problem is, systemd is all over the f'n
place :P
1682[16:05:37] <nvz> never did finish setting up all the
different DE in virt-manager so I don't have to go in and get
the other machine to fire up virtualbox
1687[16:09:39] <tw> I don't think masked is the default for
any service. afaik, debian doesn't drop any service units in
/etc/systemd; all the os provided stuff is in /usr.
1688[16:10:25] <tw> but that's neither here nor there, `sudo
systemctl unmask pulseaudio.service`
1689[16:10:58] <nvz> tw: there are links to /dev/null galore in
/lib/systemd/system/ for me
1758[16:39:37] <nvz> tux_de_botina: you'd only need to
reload the modules, but feel free to reboot.. perhaps after
installing a backports kernel, if you like :P
1762[16:40:05] <diogenes_> try reboot if won't work then see
what nvz suggests.
1763[16:40:07] *** Quits: n4dir (~n4dir@replaced-ip) (Remote host closed the connection)
1764[16:41:01] <nvz> I been compiling kernels for nearly two
decades.. and the first 5.x kernel I made, booted, things worked..
but then I realized I only had two CPUs showin up.. so..
1765[16:41:06] <nvz> yeah.. you can make mistakes
1766[16:41:24] <tux_de_botina> true
1767[16:41:31] *** Quits: saveNexit (~saveNexit@replaced-ip) (Remote host closed the connection)
1813[17:16:39] <petn-randall> dka: Have you tried 'sudo
chattr -i /etc/resolv.conf'?
1814[17:16:48] <nvz> dka: from the fine manual of chattr..
"The operator '+' causes the selected attributes to
be added to the existing attributes of the files; '-'
causes them to be removed
1815[17:17:07] <petn-randall> Yeah, that's also like the 3rd
sentence in the man page.
1816[17:17:08] <jelly> petn-randall, that sounds too obvious
1817[17:17:12] <nvz> heh
1818[17:17:27] <nvz> I was just gonna say it.. but I figured
I'd quote the manual
1819[17:17:44] <nvz> as it did seem painfully obvious
1887[18:15:57] <tga> I've been using Borg for backing up
individual directories, but I don't have a good strategy for
full running system backup and recovery
1889[18:17:18] <petn-randall> tga: Ah, I'm also using
borgbackup for that. I created a simple shell script that creates
LVM snapshots, mounts them, backups it to borg, unmounts and then
deletes the LVM snapshots.
1894[18:18:13] <petn-randall> tga: Nice thing about borg is the
deduplication feature, so you don't really need to delete
backups. Downside is that single bit flips might destroy your only
copy.
1895[18:18:45] <petn-randall> nkuttler: It does, if you backup to
the same repo. However, only a single backup can be done into a repo
at the same time. So if you have a large number of hosts, that
becomes the bottleneck.
1896[18:18:51] <tga> petn-randall: apparently many people put
together a few scripts but I was hoping for a canned solution, so I
don't have to reinvent the wheel
1897[18:19:03] <ElementalWindX> Does anyone know how to get
debian to boot with kde plasma starting immediately? I've
installed kde plasma, but when the vm boots, it goes to command
line, and I have to login and then type startx to get a GUI
1898[18:19:23] <petn-randall> ElementalWindX: You need to install
a login manager.
1899[18:20:04] <tga> afaik borg breaks everything in the repo up
into small pieces and saves them in hashed chunks
1900[18:20:08] <petn-randall> It should actually be automatically
installed if you installed the complete kde suite.
1901[18:20:29] <tga> so it will even deduplicate parts of
different files that are the same
1916[18:26:24] <ElementalWindX> @petn-randall how do I install a
login manager? The last 4~ vm's I've made all act this way
for some reason :/
1917[18:26:25] <petn-randall> tga: Do full disk backups then.
1918[18:26:33] <tga> can you maybe dump the lvm snapshot as a
single binary, and backup deltas?
1919[18:26:38] <petn-randall> ElementalWindX: How did you install
KDE?
1920[18:27:31] <tga> petn-randall: or what's a good way of
doing full disk backups of running systems?
1921[18:27:34] <petn-randall> tga: You'll have the same
problem since you're not backing up the metadata (LVM,
partition tables)
1922[18:28:02] <petn-randall> tga: That in itself is impossible.
Either do full disk backups, or online backups. Both is not
possible.
1923[18:28:53] <CyberManifest> tga: there's time shift and
deja-dup
1924[18:29:02] <ElementalWindX> @petn-randall I used the debian
installer wizard when initially installing debian
1925[18:29:04] <petn-randall> tga: You really got to ask yourself
how much work you're willing to do upfront to make your life in
case of a recovery easier. For me, just documenting the steps needed
is good enough for me.
1926[18:29:24] <tga> oh I don't mind the work, I'm
mostly worried I'll screw something up
1927[18:29:35] <tga> and backups have to be online, otherwise
they'll never get done
1928[18:29:53] <tga> can lvm do an online backup of a whole lv
and then dump it on any pv of at least the same size?
1929[18:30:04] <petn-randall> tga: Then restore your backup to a
different machine/VM, and write down the steps needed. IMHO if time
isn't of the essence, you can also just figure it out as you
go.
1970[18:36:42] <tga> damn I think I was referring to them
swapped. I want to backup the *l*v as an image
1971[18:36:48] <petn-randall> tga: You need to understand how LVM
snapshots work, and then you'll understand that backing up the
disk from beginning to end during operation will give you an
incosistent backup.
1972[18:37:12] <tga> how is that a snapshot then?
1973[18:37:19] <petn-randall> ElementalWindX: Are you using the
Debian stable installer? Or a different one?
1974[18:37:34] <tga> when you read it, isn't the snapshot
the same as a frozen copy of the lv?
1975[18:37:59] <petn-randall> tga: Yes, but you're then not
backing up any UUIDs, so they will be different.
1976[18:38:19] <petn-randall> tga: The snapshot doesn't have
the same UUID as the original, for sanity reasons.
1977[18:38:54] <petn-randall> tga: But unless I'm
understanding your idea wrong, your plan was to back up the PV?
1978[18:38:58] <tga> so then when you restore it e.g. fstab would
have to be changed?
1981[18:39:31] <ElementalWindX> yes the debian stable installer
1982[18:39:32] <tga> no, I said that wrong, I'm thinking
pv=disk lv=virtual machine disk -- so I want to make a full copy of
the lv, as you would copy paste a vm
1983[18:39:55] <tga> and I was thinking I could restore as easily
as copying back the lv on a new pv
1984[18:40:20] <petn-randall> tga: ... on a new LV, you mean?
1985[18:40:30] <tga> no, on a new physical volume
1986[18:40:47] <tga> new disk, create new lvm partition, new
physical volume, and then copy back the whole logical volume with
data
1992[18:42:31] <Bushmaster> Hello folks, i am having a problem, I
created a bootable USB dflash drive of LMDE (Linux Mint Debian
Edition) ISO. It was fruitless and now I cant get rid of any of the
files from that flash drives, neither I can format the drive in
gparted either here is the output of the flash drive
replaced-url
1993[18:42:41] <petn-randall> The contents of a LV are usually a
filesystem, the contents of a PV is usually metadata for LVM, and
also the actual data of whatever might be on there right now (might
be nothing, might be *some* of a LV, might be several LVs).
1994[18:43:01] <ElementalWindX> I guess when I go to do the
install I should uncheck the first checkbox for regular desktop, and
only select the KDE desktop checkbox
2001[18:44:28] <Bushmaster> petn-randall, yes, I used this site
step 4 to create ISO
replaced-url
2002[18:44:49] <petn-randall> tga: Have fun! I'd go with
backups from LVM snapshots, and then just document the 1-3 files you
need to fix on a restore. Perfection is the enemy of doable. ;)
2003[18:45:11] <petn-randall> Bushmaster: Ah, so you finally
installed Debian?
2004[18:45:13] <Bushmaster> petn-randall, it did not work, in
terms of booting LMDE from flash drive, so I need to delete all its
content but it wont allow me
2005[18:45:52] <Bushmaster> petn-randall, no I am using Debian
for many years, but alongside Debian, I have PCLinuxOS which I want
to get rid of and install LMDE
2006[18:46:06] <Bushmaster> petn-randall, i am using Debian
2007[18:47:04] <Bushmaster> petn-randall, all I need help is to
format that flash drive if anyone can help me with that, i would
appreciate, I tried with gparted and it wont do it
2010[18:48:21] <petn-randall> Bushmaster: You can't delete
the contents, because it's an ISO filesystem. You need to
create a new partition table with gparted.
2012[18:49:01] <Bushmaster> petn-randall, I tried, it wont do it,
let me try again now, flash drive is already inserted into USB port
hence mounted
2013[18:49:14] <ElementalWindX> "debian desktop
environment" is the first checkbox. I just wiped the whole vm
and starting over again without that checked....
2014[18:50:01] <petn-randall> Bushmaster: You might want to be
more specific about what exactly fails.
2015[18:50:59] <Bushmaster> petn-randall, let me send you screen
shot
2031[19:00:00] <dpkg> To figure out which Display Manager you are
using, open a terminal and type the following command: `systemctl
status display-manager` see also <which de> and <which
x>.
2036[19:01:52] <Bushmaster> petn-randall, this time it was
successful
2037[19:02:15] <petn-randall> yay \o/
2038[19:03:01] <Bushmaster> now I have downloaded the LMDE ISO
from mint website, what is the best way to make it bootable USB ISO,
I tried with Unetbootin and when I went to BIOS and changed to boot
from USB device, it says ISO.bin corrupted or something like that
2039[19:03:54] <sney> I have heard that mint uses hybrid images,
like debian. so you can probably write that iso directly to the usb
drive, without converting it
2040[19:04:08] <sney> however, mint isn't supported here, so
any follow-up questions will have to go to the mint people
2066[19:13:03] <sney> Bushmaster: I told you 'cp file.iso
/dev/sdX; sync' which is the same instructions for writing a
bootable iso as contained in the Debian install guide.
2067[19:13:12] <sney> Bushmaster: if that's not good enough,
I'm not sure what to tell you.
2069[19:13:53] <oxek> ElementalWindX: if all fails, checkout
ventoy
2070[19:14:05] <petn-randall> ElementalWindX: Maybe you want to
switch terminal with ctrl + alt + F1 to F6 to find it. I think
it's on F1.
2071[19:14:33] <Bushmaster> its okay sney thanks, I did not see
that, perhaps cos my nick was not attached to it, not sure about the
command you provided but i will get back some other time, i need to
go now, thanks for all the help petn-randall and sney too
2072[19:14:42] <oxek> Bushmaster: if all fails, checkout ventoy
2073[19:14:44] <sney> X isn't on TTY7 anymore?
2074[19:14:46] <oxek> sorry ElementalWindX didn't mean to
tag you
2114[19:36:35] <mutante> upgrade all packages except the kernel
package?
2115[19:36:52] <GNU\colossus> in that case (if you are using the
host kernel), your debian installation probably doesn't even
have a kernel-image package installed
2116[19:37:10] <beo> mutante: it was the other package that
required updated kernel
2163[19:53:20] <somiaj> Whyvn: since apt is just a piece of
software...and it depends on the version of debian you are using.
Debian stable, new kernels are only security fixes, and seem to come
1-2 per point release (multiple months). Backports get new version
about every month or so.
2194[20:26:07] <hussar> package management on linux upsets me so
much. This has to be the dumbest system ever created
2195[20:26:09] *** Quits: bocaneri (sauvin@replaced-ip) (Remote host closed the connection)
2196[20:27:40] <hussar> how the hell am I supposed to know which
packages I actually installed and what commands the gave me.
There's no record of what I installed with the order I
installed them that I can look at
2197[20:27:45] <mutante> seems to me like it's a pretty
complex problem to solve
2200[20:28:15] <dpkg> Files in package mutante-pr0n:
/srv/mutante-pr0n/mutante.jpg
2201[20:28:21] <mutante> lol
2202[20:28:29] <mutante> wtf is that :)
2203[20:28:53] <somiaj> hussar: you might want to clarify, but in
debian the package manager is very nice, there is full record
(including md5 sums) of ever file that you installed, its exact
location, and lists of all the packages currently installed.
2204[20:29:01] *** debhelper sets mode: +l 1173
2205[20:29:05] <hussar> they could just have apt write append a
line to a log file every time you install and build stats and
history info from it. That's not that complex
2206[20:29:06] <somiaj> mutante: dpkg gives snarky responses when
you start a line with dpkg.
2207[20:29:21] <hussar> doesn't md5sum have a lot of
problems now?
2211[20:29:43] <somiaj> hussar: /var/log/apt contains a full
history of every file installed by apt, there is also
/var/log/dpkg.log
2212[20:30:32] <somiaj> hussar: not really, it does a good job
for checking intergurity of individual files inside a package. This
isn't the hash being used to verify the idtergurity of the
package, that is different.
2213[20:30:38] <hussar> somiaj: great, now how do i get the list
of commands I created and order I installed them
2221[20:32:24] <somiaj> oh you wanted the binaries installed.
usually the binary name is similar to the package name, but as
mutante said, you can get a full list of every file installed by the
package with the command they shared.
2223[20:32:55] <hussar> somiaj: that doesn't answer my
question very well
2224[20:33:16] <somiaj> hussar: your question is vague. What
package did you install that you want to know the binaries of?
2225[20:34:04] <hussar> now i'm supposed to look the all the
directories in the path variable and guess which packages created
them with no context of when or why I installed the packages. What
if I accidentally delete the wrong package?
2226[20:34:21] <somiaj> If you want to know which package owns a
file, dpkg -S /path/to/file
2237[20:36:36] <hussar> Also, a file belonging to a package
doesn't mean it was a package I manually installed. It could
have been created by a dependency of a manually installed package.
Which is the problem I'm having now with qemu and it's 100
different commands
2240[20:37:03] <somiaj> hussar: What is your actual problem?
2241[20:37:35] <somiaj> hussar: yes debian is a shared libary
system, so lots of packages get installed, but this shouldnt'
be an issue, as apt will take care of the resolution for you.
2243[20:38:10] <hussar> viewing what packages i've
installed, their dependencies, and their related commands with the
context of when and how they were created is very complex and in
some cases is impossible
2244[20:38:13] <somiaj> hussar: are you here just to complain
about debian, or do you have an actual issue on debian stable you
would like help solving?
2248[20:38:55] <somiaj> yes, the package dependencies are
complex, apt/dpkg has lots of tools that you can access any of that
information you need, but you don't need to know all the
dependicies.
2249[20:39:01] *** debhelper sets mode: +l 1167
2250[20:39:06] <somiaj> hussar: what is your actual issue? What
isn't working?
2251[20:40:15] <hussar> i've installed some packages a
tutorial recommended for getting qemu running. Now I don't
remember what those packages were and how the relate to all the
commands I now see. I think some of the packages were unnecessary
and want to delete them to get the minimal install and not gui tools
that are unnecessary for what i'm doing
2259[20:42:44] <somiaj> hussar: most likely you installed
virt-manager manually, but 'aptitude why vrit-manager'
will let you know if/what package pulled it in
2267[20:45:12] <somiaj> logs are rotated, check the other ones.
Also rotation will often delete really old logs, so it is good to
get into habbit of backing up the logs
2270[20:45:54] <hussar> this is already unreasonably complicated.
this is something that should be built into apt or dpky
2271[20:45:57] <hussar> *dpkg
2272[20:46:02] <oxek> apt logs are rotated on a monthly basis,
and keep 12 months
2273[20:46:25] <oxek> hussar: and it is built in. Clearly
describe your issue and we might be able to help.
2274[20:46:40] <hussar> this is my first time using qemu or a vm
on this machine i have no reason to think that virt-manage will be
in another log file, but now I have to look through all of them
2275[20:47:18] <oxek> is your question "I did not tell apt
to install virt-manager. What package installed virt-manager?"
2276[20:47:22] <sponix2ipfw> hussar: cat the logs and grep them
2277[20:47:34] <hussar> oxek: yes, that's basically it
2282[20:48:27] <hussar> not just virt-manager but some other
commands too but let's start with that. At least I know which
packages i've installed now but looking through logs to check
is really a dumb system that should be changed
2283[20:49:07] <somiaj> hussar: The logs let you know when you
installed stuff, dpkg -l will list what is actually installed.
2284[20:49:08] <oxek> hussar: is your question "Which
packages did I explicitly tell apt to install on this system?"
2291[20:51:20] <hussar> i know there is a way to list manualy
installed packages but that isn't the same as listing packages
the user manualy installed
2292[20:51:41] <somiaj> hussar: except for the base system it is
the same thing
2293[20:52:10] <oxek> hussar: that command shows you which ones
are manually installed
2301[20:58:10] <hussar> i ran that command, I don't believe
it's accurate because it's thousands of lines long. It
also doesn't give the related commands, or the order and time
they were created. It's just a list of nonsense to me
2302[20:58:52] <oxek> 'apt-mark showmanual' is accurate
2304[20:59:24] <oxek> if you want a list of all binaries from all
manually installed packages, use something like: apt-mark showmanual
| apt-file list -f- | grep '/bin/'
2311[21:01:06] <hussar> that's better but still has a lot of
the same problems. I installed this os from debian minimal and there
are libs in that list I doubt were part of the base install
2318[21:02:51] *** Quits: j7k6 (~j7k6@replaced-ip) (Remote host closed the connection)
2319[21:03:12] <somiaj> you are getting all worked up instead of
learning the tools that are there. part of the reason there are so
many debian based distros is the quality of the package manager.
2323[21:05:03] <somiaj> have you ever built software, often times
building software causes one to manually install libaries
2324[21:05:26] <hussar> this isn't my first time having this
problem. I've tried understanding the tools and there
doesn't seem to be a way to solve it. No one in this chat has
given a realistic solution either
2325[21:05:44] <somiaj> you haven't given us an actual
problem, except the fact you don't like that there are just
lots of packages.
2331[21:08:39] <hussar> no, i've explained my problem a few
times. I want to see the packages I've manually installed and
the context they were installed so I have an idea of why I installed
them. This means I want to thee time time and order in a manner
that's readable. The best option I've gotten was reading
through the logs, but this is complex, has no filtering options, and
still doesn't give me their relevant
2340[21:11:28] <somiaj> keep backups of your logs, all of that is
logged (as told), and the manual install flag is mostly useful (just
the base system gets marked as manual installed by debootstrap)
2341[21:11:53] <oxek> I don't think there's a simple
tool to do exactly what hussar wants
2343[21:12:06] <oxek> but it is doable with some data
manipulation
2344[21:12:32] <oxek> even filtering out the base install
packages, by using e.g. debootstrap to get a list of base packages
and then excluding them from apt-mark showmanual
2345[21:12:43] <oxek> but that's not a simple tool,
you'd have to learn things
2346[21:13:12] <somiaj> taking good notes is useful too, and book
marking 'random guides' you follow
2348[21:15:04] <sponix2ipfw> oxek: probably not one simple tool,
but you both have presented many options to do what he is asking,
even if they aren't "fun"
2349[21:15:09] <hussar> oxek: that's what i think too.
bookmarking random guides and looking through logs creates stress
and time wasted. Reinstalling is a lot faster. It's suprising
to me that this isn't something that's been addressed yet.
You would think this would be a huge annoyance for debian devs
2350[21:16:18] <hussar> my first time running into this thing was
when i installed a bunch of packages for header files and
dependencies to build stuff from source. Then found myself totally
confused about what packages I had installed and no longer needed
2351[21:16:53] <somiaj> with things like aptitude why
packagename, apt show, looking at dependencies, and manual install
packages, most people can figure out if something was manually
installed or a dependency
2352[21:17:46] <somiaj> but this is why everything is logged, and
if you have a good backup system of your logs, you can track down
this info if you need it.
2353[21:17:46] <hussar> somiaj: finding out if something was
manually installed is not the problem. I want to find ou what the
'thing' is in the first place
2354[21:18:06] <oxek> hussar: 'apt show PACKAGENAME'
2355[21:18:53] <oxek> hussar: the issue is that you're
presenting a scenario that is not a problem for anyone but you - I
am not exaggerating here, I've never seen anyone have the issue
you describe.
2356[21:18:56] <karlpinc> hussar: Often, a good strategy is to
build in a debootstrap-ped chroot. Then you delete the whole thing
when you're done and don't worry about build dependences
cluttering things up. YMMV.
2357[21:19:01] <oxek> hence a tool does not exist yet
2359[21:19:46] <hussar> oxek: that's not what i mean. I mean
I want to find out what PACKAGENAME is based on the commands I have
available and time I made the changes
2360[21:20:06] <oxek> example scenario - I find I have this
command 'equivs-control' and have no idea what it is or
where it came from
2361[21:20:07] <karlpinc> hussar: There might, possibly, be a way
to leverage something like ansible to keep the history you're
looking for.
2362[21:20:14] <oxek> so I do 'command -v
equivs-control'
2363[21:20:27] <oxek> it tells me /usr/bin/equivs-control
2364[21:21:00] <oxek> I then do 'dpkg -S
/usr/bin/equivs-control'
2365[21:21:05] <oxek> it tells me: equivs
2366[21:21:13] <oxek> I then do 'apt show equivs'
2367[21:21:18] <hussar> It's not a problem people often
complain about because most user just install a bunch of stuff and
never worry about it or care that gigabytes of space on their drive
is being wasted by packages they used only once
2368[21:21:19] <oxek> that tells me what that package is
2369[21:21:20] <sponix2ipfw> karlpinc: umm, I do similar with my
own applist.txt notes file. Has what I install, why, and so on - and
also links to the guides I follow when doing it
2370[21:21:32] <oxek> if I want to know what else the package
installed, I do 'dpkg -L equivs'
2372[21:22:01] <oxek> if I want to find out when I installed it,
I do grep on logfiles for the package name, and it gives me a date
2373[21:22:11] <oxek> I look around the date what else I might
have installed around that time
2374[21:22:15] <oxek> and that jogs my memory
2375[21:22:20] <karlpinc> sponix2ipfw: Yeah. A system logbook
that is maintained by a human.
2376[21:22:42] <oxek> if I want to find out why some package is
installed, I do 'aptitude why PACKAGENAME'
2377[21:22:44] <oxek> etc.
2378[21:22:50] <somiaj> hussar: I regurally go though what I have
installed and clean up stuff. But I do it based off the package
descriptiosn and what I use vs when it was installed. deborphan is
nice to clean up libs
2379[21:22:52] <oxek> there are lots of tools, each doing a
specific jobs
2380[21:22:59] <sponix2ipfw> karlpinc: I got use to that routine
working in the Army server room... Had to list your apps, updates,
patches and so on
2381[21:23:09] <karlpinc> oxek: Maybe I'm missing something.
"command"? What package is that in?
2390[21:25:02] <oxek> 'which firefox-esr' will give you
/usr/bin/firefox-esr
2391[21:25:12] <karlpinc> oxek: Interesting. I've always
used "which".
2392[21:25:14] <hussar> somiaj: That's hard for me to
believe. The command to show manually installed packages that was
suggested was thousands of lines. You regularly go through thousands
of lines of packages carefuly to see what you actually use?
2393[21:25:27] <oxek> karlpinc: 'command -v' is
preferable to 'which'
2394[21:25:42] <karlpinc> oxek: "command" must be
posix, because it's in dash.
2395[21:25:47] <oxek> hussar: I regularly go through billions of
lines as part of my day-job
2396[21:26:03] <oxek> karlpinc: it is posix
2397[21:26:15] <oxek> karlpinc: which is why it is preferable to
'which'
2398[21:26:28] <karlpinc> oxek: I was just going to ask that. :)
2399[21:26:34] <hussar> oxek: but there's no way you scan
through those line manualy
2400[21:26:49] <oxek> hussar: that's why there are tools
like grep, cut, awk, perl
2401[21:26:50] <somiaj> hussar: I use dpkg -l | less and go
through all packages every so often, partly because it gets me use
to what packages debian provides and I use.
2402[21:27:02] <sponix2ipfw> oxek: Yep... "grep" is
golden
2403[21:27:28] <jelly> hussar, there are workaround tools like
deborphan, and there's mk-build-deps command to be used to
avoid apt-get build-dep leftovers
2404[21:27:47] <karlpinc> sponix2ipfw: (But perl is just
"shiny"? ;-)
2413[21:32:23] <jelly> ElDiabolo, I want to say xxd but unsure
about the options used
2414[21:33:02] <jelly> echo foo | xxd -i ?
2415[21:33:09] <karlpinc> I graviated to aptitude, once upon a
time, instead of apt-get because aptitude removed dependent packages
when you remove the depending package. Which has never helped with
"suggests", but.... Since then I've not kept track of
which apt tool does what in this regard.
2429[21:38:33] <ElDiabolo> jelly, No, output should be a string
literal, something like "\tfoo\n"
2430[21:39:03] <jelly> and what's the difference between a
char[] in one format or the other
2431[21:40:42] <ElDiabolo> jelly, Readability.
2432[21:40:46] <jelly> (except, perhaps, length and some amount
of legibility if your input is ASCII-alike)
2433[21:40:55] <jelly> right
2434[21:42:34] <ElDiabolo> jelly, I need to output a given
makefile, so tabs are important and it should still be readable.
This just looks like a "somebody already did it"
situation.
2435[21:42:57] <ElDiabolo> I just can't find it :-/
2436[21:43:05] *** Quits: thiras (~thiras@replaced-ip) (Remote host closed the connection)
2474[22:17:42] <c-c> Any ideas how to fix horizontal tearing in
browser video? I attempted to reconf X11 by setting section
"OutputClass" with option "TearFree"
"true":
replaced-url
2482[22:27:49] *** Quits: ElNomReal (~E@replaced-ip) (Remote host closed the connection)
2483[22:28:00] <c-c> Perhaps of interest here is also to note
that I do run steam games, and they do not seem to have horizontal
tearing, even some 2016 games that are pretty open gl 3d titles
2484[22:28:05] *** Quits: Jerrynicki (~niklas@replaced-ip) (Remote host closed the connection)
2485[22:28:11] <sney> ok, I was just about to ask something like
that/
2495[22:38:45] <c-c> Strangely, firefox-esr seems to have
horizontal tearing at random position on the video, chromium looks
to have h-tearing around the same vertical level on the top 10% or
so
2496[22:39:17] <c-c> - hard to find a clip like the one I
originally noted it in (live stream from YLE areena)
2508[22:43:49] <c-c> Chromium seems to have tearing near the top
only at higher bitrates >1.5 mbps. Firefox seems to have tearing
that jumps up and down, and even at low bitrate 480p has occasional
tearing every 1-2 Hz.
2531[22:52:23] <c-c> I was trying to figure out why, but there
was nothing special in Xorg.0.log. Monitor found,
"glamorgl" loads, the intel GPU detected, etc. I wonder if
there is a package "SNA" or "TearFree" needs
installed? Or maybe some typo here:
replaced-url
2535[22:55:12] <sney> let us know if it works, you shouldn't
need to add any xorg config stuff for something this basic, and it
may indicate a bug/regression somewhere
2537[22:55:44] <c-c> heh thats what I get for copypasting
directly from internet forums
2538[22:55:51] <sney> also, what intel cpu is it specifically? if
it's particularly new the firmware files in buster might not be
sufficient. there's a newer update in buster-backports.
2540[22:56:02] <somiaj> c-c: some intel cards run better with
firmware-misc-nonfree (skylake and various generations do have
non-free firmware)
2541[22:56:40] <somiaj> c-c: also use the 'modesetting'
driver not the intel driver, the intel driver isn't as good in
lots of situations (not having an xorg.conf would probably help out
too, it is often not needed)
2554[22:58:53] <somiaj> oxek: the older kernel might not know of
the newer firmware files (they are often foo-24, foo-25, foo-26,
older kernels don't know to look for newer firmware
2555[22:59:12] <somiaj> you may not need the firmware, but many
cards will preform better using it
2556[22:59:28] <oxek> oh, so newer firmware packages always
contain all the older versions of firmware too?
2564[23:00:52] <sney> firmware packages are built from the
contents of
replaced-url
2565[23:00:55] <oxek> I was thinking of the case where AMD cpus
required a fixed firmware to boot, and then if you updated the
firmware again without updating the kernel, the kernel would
suddenly not find any firmware files (because it's not aware of
the newer ones) and the machine won't boot (because the cpu
microcode is not updated)
2566[23:01:31] <sney> cpu microcode is a different animal. this
is closed source driver components, *usually* not something required
to boot the system.
2568[23:02:12] <oxek> don't all the firmware-sth-nonfree
packages pull in cpu microcode?
2569[23:02:16] *** Quits: szorfein (~daggoth@replaced-ip) (Remote host closed the connection)
2570[23:02:27] <somiaj> no cpu micrcode is in a different package
2571[23:02:30] <somiaj> !microcode
2572[23:02:30] <dpkg> Microcode are instructions/structures for
implementing high-level machine code within processors. The Linux
kernel can load updated microcode on most x86 processors. Microcode
patches for Intel and AMD64 CPUs are packaged for Debian as
intel-microcode and amd64-microcode respectively; installing the
relevant package is recommended to ensure system stability. Ask me
about <non-free sources>.
replaced-url
2573[23:02:42] <sney> intel-microcode and amd64-microcode are
packaged separately, I don't remember the exact dependency
relationships
2627[23:22:26] <alex11> the security repo in bullseye is a
different url but i think it probably doesn't even exist yet so
it's not relevant until it becomes stable
2628[23:22:36] <johnjay> ok
2629[23:22:47] <sney> alex11: yes, testing security is currently
empty and will be until after the freeze
2636[23:31:09] <sney> afaict it isn't as dire as some recent
blog coverage made it sound, but all debian developers and
maintainers do their work on a volunteer basis
2637[23:31:17] <sney> so there's always room for more
2638[23:31:56] <sney> !wnpp
2639[23:31:56] <dpkg> For information on packages which are not
in Debian but you think should be, check the Work-Needing and
Prospective Packages list at
replaced-url
2640[23:33:53] <somiaj> c-c: in what sense? In general the
modesetting driver is the prefered and one that often works better.
There are some generations of intel cards that do work better with
the old intel driver (though this is less and less since 2016), as
for the particular issues in that thread, it depends a lot on the
exact intel gpu (such as if you get a performance increase for the
non-free firmware depends on exactly
2641[23:33:58] <somiaj> which generation of intel gpu being used)
2644[23:34:31] <somiaj> c-c: I personally suggest people switch
to the modesetting driver, and at least try it out. You can always
switch back if you have a card that doens't work as well with
it.
2645[23:34:55] <somiaj> c-c: users making good bug reports, and
providing patches is also very helpful beyond matainers and devs.
2647[23:35:50] <somiaj> though those users should run testing (At
least on a development box) to help improve the quality of packages
by reporting (and fixing if they can) bugs
2656[23:40:52] *** Quits: platvoeten (~platvoete@replaced-ip) (Remote host closed the connection)
2657[23:41:06] <jhutchins> oxek: One issue is that they are not
all employees of a single, centralized company. They don't have
a corporate imposed common platform.
2661[23:44:39] <c-c> somiaj: I see. There's quite confusing
and conflicting info on this topic. Seems one doesn't need to
fiddle with drivers if modern intel integrated GPU. And that implies
modesetting with buster.
2662[23:44:53] <sney> debian *does* use salsa, which is a gitlab
instance, for many many packages' version/source control. issue
tracking already has a good home at the debian BTS.
2663[23:45:29] <sney> making some big sweeping infrastructure
change to "get more devs/maintainers" would, in the real
world, more likely start a flame war bigger than the systemd one.
2664[23:47:31] <jhutchins> A sudden broad increase in developers
would likely cause a lot of chaos and a lot of submissions that
didn't meet the standards and conventios that have been
established.
2667[23:48:17] <sney> and the help debian needs is on the lower
levels. solving small-scale problems, maintaining packages with
userbases in the 1000s, little un-glamorous stuff. not coming in as
a manager and shaking things up or whatever
2669[23:48:44] <jhutchins> Again, since there's no central
controlling corporation, it's a lot harder to impose those
standards and conventions.
2670[23:49:30] <jhutchins> This is probably off topic for this
channel.
2671[23:49:32] <sney> some day I'll finish learning C and
make myself truly useful
2672[23:49:37] <sney> it's borderline, I agre
2673[23:50:10] <jhutchins> sney: There are projects in things
like ruby and perl and python that need work.
2674[23:50:26] <c-c> C is easy to learn, but its hard to learn
the solutions using it to use it right
2675[23:51:07] <c-c> jhutchins: can you name a project in ruby
that needs work?
2676[23:51:13] <sney> I'm ok at python but I don't like
it very much, so I wouldn't want to be debugging in it. I like
the lower level technical stuff. just need to get over the hump.
2677[23:51:32] <c-c> - I could have a look if there's a
corner I could help with
2678[23:51:46] <dvs> sney, I don't like Python either.
2679[23:51:51] <sney> I believe debian has a "ruby
team" that is a general umbrella of all of the ruby packages
2680[23:51:55] <jhutchins> I learned to code in six languages,
but never found any useful work to do with them.
2691[23:54:42] <somiaj> c-c: the basic idea is due to kms (kernel
mode setting) drivers are moving away from the old format of driver
+ xorg driver, to a kms standard, and the modesetting driver is the
driver that uses kms and just speaks directly to the driver to
implment the features.
2692[23:55:07] <somiaj> c-c: so just as with lots of things open
source, you see a lot of the development cycle, but the modesetting
is getting fairly well worked out and standard these days
2693[23:55:32] *** Quits: yuta (~pi@replaced-ip) (Quit: WeeChat 2.9)
2694[23:58:11] <somiaj> c-c: this also means the same driver
being used at boot (in the tty consoles) is the same being used in
xorg, so switching thigns are smoother and no more need for
framebuffers
2695[23:58:50] <somiaj> c-c: from y limited view point, getting
all the requirements to build with rust into debian has been a bit
of a pain.