31[00:45:48] <GigaFlow> need to set 2 ip addresses on teh same
interface. Using enp0:0 doesn't work for me, networking service
can't run (while 'ip a' shows both addresses
assigned).
213[05:10:11] <dannylee> ok in Debian repository there a editor
emulator call yudit...most of the time i use emacs...but yudit will
let you edit any thing,,,its a unix editor,,,
235[05:43:43] <Renari> I'm trying to install virt-manager
and it seems one of its dependencies isn't available?
236[05:43:44] <Renari> virt-manager : Depends: virtinst (>=
1:2.0.0-3) but it is not going to be installed
237[05:43:49] <Renari> Does anyone know what's up with
this?
238[05:45:15] <Renari> Actually drilling down and trying to
manually install the error packages it ends with this:Package
libpython3.5 is not available, but is referred to by another
package.
247[05:48:10] <tomreyn> you probably have mixed up apt sources
then
248[05:48:26] <themill> Renari: sounds like you might have a
mixture of two releases there. Can you pastebin "apt-cache
policy virtinst libpython3.*; apt-cache policy"
249[05:49:32] <themill> (nothing in buster has a dependency on
libpython3.5)
322[06:42:01] <Renari> if you have none of them you can probably
create mpv.conf to override the volume
323[06:42:07] <awal1> problem solved Renari, themill
324[06:42:09] <awal1> thanks
325[06:43:24] <awal1> I created mpv.conf in home .config.mpv and
added inserted this line "volume=50"on it and problem
solved . now it starts with volume at 50%
331[06:46:24] <awal1> i mean : i added "volume=50" in
mpv.conf file in .config/mpv dir in user home and the issue is
solved now. volume when mpv starts is at 50 %. thanks Renari,
themill for your attention :)
332[06:52:17] *** Quits: dez (uid92154@replaced-ip) (Quit: Connection closed for inactivity)
400[09:11:19] <hiptobecubic> I have 3 ssds. I am currently
booting off of sda1, but have made an lvm raid1 (/dev/vg/lv) out of
sdb1 and sdc1. I am trying to figure out how to install debian to
the raid1 but I have kind of made a mess.
401[09:11:58] <hiptobecubic> specifically, i can boot to grub,
but can't get to the root fs and it drops me in initramfs
402[09:12:32] <hiptobecubic> I think this is because it's
not activating the vg's and lv, but i'm not sure why
403[09:13:39] <hiptobecubic> i installed by using debootstrap
and then grub-install --root-directory /mnt /dev/sdb, where
there's a 1MB partition /dev/sdb2 of type BIOS_boot
404[09:13:43] <hiptobecubic> it said it installed without errors
406[09:14:06] <hiptobecubic> it seems more like it didn't
correctly detect what would be needed in the initramfs and then
couldn't do anything when it got there
407[09:14:15] <hiptobecubic> but i'm not sure how to fix
that
414[09:23:48] <ratrace> hiptobecubic: and also the kernel, grub,
create your unpriv user and/or set root password (and locale and
timezone but that's not gonna block boot if you didn't)
416[09:24:23] <ratrace> debootstrap does a very, very
rudimentary installation, suitable for containers, thus without
grub, kernel, specific filesystem utils, etc...
417[09:26:59] <hiptobecubic> ratrace, good question.
418[09:27:17] <hiptobecubic> rebooting now, if it fails,
i'll boot back to first os and then chroot and see
419[09:27:40] <hiptobecubic> now it's just printing
"mdadm: No arrays found in config file or automatically"
repeatedly
420[09:27:50] <hiptobecubic> welp... reboot again
421[09:28:01] *** debhelper sets mode: +l 975
422[09:28:02] <ratrace> hiptobecubic: are you using mdadm? you
should've installed that too. I thought oyu said RAID was LVM
based
423[09:29:03] <hiptobecubic> i installed mdadm after flailing
around on the wiki and various blogs. I don't know why i would
need it. Yes my raid is lvm based
424[09:29:25] <ratrace> ok. but most importantly, initramfs and
grub (in that order) shoudl've been set up from within the
chroot so devices and paths are autodetected.
428[09:31:56] <ratrace> hiptobecubic: debootstrap is actually
awesome once you get the hang of it. I only use debootstrap to
install debians and derivatives, but primarily due to rootfs on zfs
or btrfs on luks :)
429[09:32:32] <ratrace> hiptobecubic: it could be possible to
use the installer and do the manual partitioning step where you
prepare the LVM arrays yourelf and then just select LV devices for
rootfs and other mounts
431[09:32:36] <hiptobecubic> it would be nice if it was more
complete then, like there are at least 5-6 steps you have to
remember afterwards or you're dead in the water
434[09:33:23] <hiptobecubic> I have a luks encrypted drive that
I think got corrupted. I spent 180 gpu hours trying every variation
of what i remember my password to be but no luck :(
435[09:33:27] <hiptobecubic> or should i say, no luk
436[09:33:37] <ratrace> hiptobecubic: it is something like that.
deboostrap; prepare chroot; chroot; install kernel, grub, filesystem
utils, lvm|mdadm if needed, set up root pass, networking if
that's a remote server, locale, timezone.
437[09:34:30] <ratrace> fstab if you got mounts like /boot or
anything else; not strictly needed for rootfs, but for some
filesystems it may be required to flip root from RO to RW
438[09:34:32] <hiptobecubic> Ok here we go, rebooting again!
447[09:35:22] <hiptobecubic> with some firmware errors
apparently
448[09:35:33] <hiptobecubic> but still, a login screen,
that's a big step up
449[09:35:37] <ratrace> yay!
450[09:35:44] <hiptobecubic> The problem was apparently not
having had lvm2 installed in the chroot
451[09:35:49] <ratrace> $1M question: did you remember to set up
root pass? :)
452[09:36:04] <hiptobecubic> Not at first no. Already did the
reboot and fixed that 2 hours ago
453[09:36:09] <ratrace> ah, k :)
454[09:36:17] <hiptobecubic> but that's a good example of
something debootstrap should really do
455[09:36:45] <ratrace> assuming it's needed
456[09:36:55] <hiptobecubic> maybe also mention that you are
trying to mount lvm volumes but aren't installing lvm? that
would be cool too
457[09:37:03] <hiptobecubic> ratrace, when does root not need a
password?
458[09:37:05] <ratrace> so maybe an option you need to remember
to switch on... so it doesn't really solve much
459[09:37:13] <ratrace> hiptobecubic: containers
460[09:37:31] <hiptobecubic> I suppose, although it's then
harmless isn't it?
461[09:38:12] <ratrace> what, setting root pass when it's
not needed? sure
462[09:38:29] <hiptobecubic> or... prompt the user at the start
of install?
463[09:38:50] <hiptobecubic> "Do you want to be stranded?
[y/n]"
464[09:38:58] <ratrace> heh
465[09:39:55] <ratrace> but then why stop there. it could also
autodetect filesystem, lvm, mdadm and install those, ask about
networking, locale, timezone. maybe even a ncruses menu..... and
you've just invented the installer :)
467[09:40:32] <hiptobecubic> i think there's a market for
three levels
468[09:40:42] <ratrace> so I'd rather debootstrap was as
raw and functional as it is now (adding all that stuff also adds
bugs), and instead beef up the installer for things like LVM, mdadm
installations, better LUKS support, maybe even zfs/btrfs
469[09:41:07] <hiptobecubic> "i don't care if anything
works because this is for a container" "i want just enough
that i'll be able to boot and take it from there" and
"just install debian please. i hate this"
477[09:45:13] <hiptobecubic> I'm thinking about just the
minimum required to boot the system you're setting up. Like if
you tell it you have a particular fs and it doesn't have the
module needed to mount that fs, what's the usecase really?
480[09:48:14] <ratrace> debootstrap ist just a shell script that
bootstraps a base debian ostree and leaves the configuration to the
user. if it added higher level options and functions, it's
really just reinventing the installer.
496[10:11:31] <hiptobecubic> For some reason google's gpg
key for their chrome apt repo doesn't work for me
497[10:12:25] <hiptobecubic> Not really sure what to make of it.
Says it finds two keys and they import fine, but then gpg -k to list
them says there are no keys and apt says the repos are unsigned.
539[11:04:30] <jmd> How do I install a package from
buster-backports?
540[11:05:54] *** Quits: Nokaji (~Nokaji@replaced-ip) (Quit: "... when the freedom they wished for most was
freedom from responsibility then Athens ceased to be free and was
never free again.” ~ Edward Gibbon (1737-1794) - Decline and
Fall of the Roman Empire, 1909)
556[11:20:57] <ratrace> jmd: apt install -t buster-backports
<package name> ; having previously added the backports repo to
sources.list and having run apt update
569[11:25:41] <ratrace> jmd: just remember not to use that -t
for apt upgrade. the packages from backports will upgrade normally
if there's a newer version, with regular apt upgrade.
570[11:25:57] <ratrace> adding -t to upgrade will force all
packages to look up upgrades in backports.
571[11:26:11] <jmd> okay. I will bear that in mind.
572[11:26:25] *** Quits: szorfein (~daggoth@replaced-ip) (Remote host closed the connection)
600[11:50:40] <redcheckers> if the debian project has dropped
support for an architecture, like say, 32 bit powerpc / ppc , is it
still possible to install it from older supported versions, or are
all the dist files probably gone from the mirrors too?
605[11:53:54] <dpkg> An archive is a collection of files.
'tar', 'ar', 'cpio' are all archiving
tools. This is *not* the same as compression, which is a separate
operation. Debian Archives is the repository for old Debian
releases, see
replaced-url
606[11:54:22] <redcheckers> nkuttler: ah thank you
613[11:59:58] <MWM> I need a hand with proxmox (pretty much
custom debian). I was wondering if anyone here might lend a hand?
614[12:01:01] *** Quits: earthundead (~earthunde@replaced-ip) (Remote host closed the connection)
615[12:01:04] <MWM> proxmox cant find the network adapter, si I
tried manually loading the drivers (rtl8125) during installation,
and I get a complaint that make isnt there
620[12:01:29] <MWM> Is there a way to get make onto a machine
without network?
621[12:01:32] <ratrace> !based on debian
622[12:01:32] <dpkg> Your distribution may be based on and have
software in common with Debian, but it is not Debian. We don't
and cannot know what changes were made by your distribution (compare
replaced-url
623[12:02:17] <MWM> ill try proxmox then. ty
624[12:03:00] <MWM> honestly had the same problem with vaniall
debian though
635[12:11:27] <MWM> however I cannot get that package because my
network will not work.
636[12:11:46] <sfinks> Hi, is it possible to configure an IP6
address alongside IP4 in the initrd environment for dropbear? It
seems like the part in the linux kernel that picks up the ip=
setting only supports IP4
638[12:12:10] <MWM> SO I downloaded the tarball from realtek but
got a make error that lead me to beleive that make was not present
and would need to be installed
639[12:12:42] <ratrace> don't install tarballs. if you can
download, then find your distro's .deb package with the
firmware
640[12:13:04] <ratrace> sneakernet it to your machine and
install with apt install ./the-filename-of-the.deb
641[12:13:32] <MWM> I am not aware of how to accomplish that
from another machine? I can just grab the .deb from somewhere and
put on usb, or is there another way?
643[12:14:34] <MWM> let me explain: fresh install on new
hardware cannot find the network, but I do have another working
machine elsewhere in the house to download stuff with
829[15:33:30] <dob1> I have this rules in OUTPUT chain of
iptables
replaced-url
830[15:34:12] <yahal> I am under Debian Buster system and I try
to enter a systemd-nspawn container (also debian) with
"machinectl login <machinename>" which fails with
"Failed to get login PTY: Protocol error" if I stop the
container and login with systemd-nspawn it works fine
853[15:41:39] <ratrace> can you reverse the logic? allow only
root and specific services on outbound, drop all else?
854[15:41:46] <rudi_s> dob1: You need to match the uid of each
user
855[15:41:47] <ratrace> (careful you don't lock yourself
out of ssh)
856[15:41:57] <dob1> rudi_s, there is the gid too
857[15:42:05] <dob1> ratrace, I don't know if I am able
858[15:42:18] <rudi_s> Or just use something like -m owner
--uid-owner x -j user_chain; --uid-owern y -j user_chain; ...
859[15:42:30] <rudi_s> dob1: What rat<tab> said, it's
most likely just the primary group.
860[15:42:48] <dob1> it's useless in this way
861[15:43:14] <rudi_s> dob1: To you, yes.
862[15:43:17] <rudi_s> So don't do it that way.
863[15:43:41] <ratrace> dob1: the iptables-extensions(8) manpage
doesn't specify, but various posts online suggest that due to
nftables backend, it's limited to primary group
864[15:43:47] <dob1> I have to do any specific rule for every
user so n users -> n rules (n * 3 in my case but the logic is
this)
865[15:43:59] <dob1> I was thinking 3 rules for a gruop, add
users to gruop
866[15:44:04] <ratrace> dob1: or reverse the logic to whitelist
instead of blacklist
867[15:44:32] <rudi_s> dob1: You can use a separate chain, that
reduces it to n rules. And n rules will - most likely - not be an
issue.
868[15:44:48] <ratrace> dob1: also what users are those?
services or humans?
869[15:44:58] <dob1> humas
870[15:45:00] <dob1> humans
871[15:45:04] <rudi_s> Or you could use nftables which permits
using a single rule with a table lookup.
872[15:45:21] <ratrace> dob1: can you reverse the logic to
whitelist instead of blacklist?
873[15:45:35] <dob1> rudi_s, I have to think about this, not
practice about chains, you then create a rule for every user in
output chain that jump to the chain and the chain has 3 rules that i
want ?
884[15:49:52] <rudi_s> As I wrote above ("-m owner
--uid-owner x -j user_chain; --uid-owern y -j user_chain;
..."), user_chain is the separate chain with your rules
886[15:50:16] <dob1> I would like to try the chain idea, seems
more simple to manage
887[15:51:34] <rudi_s> Yeah. It's generally a good idea if
you want to group stuff.
888[15:52:29] <rudi_s> Btw. if you setup completely new firewall
you might want to look at nftables. Personally I'm still mostly
using iptables but nftables has some nice features and it's
where firewalls on linux is migrating to (in the future at least).
895[16:01:09] *** Quits: zmcf (~user@replaced-ip) (Remote host closed the connection)
896[16:01:18] <ratrace> I'm not bothering learning whole
new nftables becuse: iptables interface JustWorks(tm) and
they'll replace nft with ebpf anyhoo, so what's the point
897[16:01:48] <ratrace> I prefer the If it ain't broken,
don't fix it paradigm. So far, the iptables interface
ain't broken.
899[16:03:37] <MegaCarp> trying to set up an ssh connection to
my home pc. can connect to locahost but remote connection
doesn't work - "connection timed out". network is on
router, i've forwarded the ports, different ports in the course
of my trying to make it work. i've a firewall installed on the
pc, first i've disabled it (since i have a router, duh), now
i've turned it on to, again, add rules for ssh ports. i've
edited /etc/ssh/sshd_config - i've
900[16:03:37] <MegaCarp> wanted to use different ports, then
i've allowed the default ports for the sake of the experiment
and to just get it working once. since it keeps working with
"ssh localhost -p [my custom port]" i've since
returned to my custom ports. i see in the journal of gufw that i
have zerotier-one vpn installed for w/e reason - could it be
interfering in any way?.. i also have hamachi running
905[16:09:35] <rudi_s> ratrace: I mostly concur (and that's
the reason why all my systems currently still use iptables).
However, the ebpftables project seems to be dead and nftables has
some nice properties (e.g. faster lookups, verdict maps, no
unnecessary hooks) which are useful for high-performance firewalls.
906[16:10:14] <rudi_s> *and some nice filtering properties
(tables, maps, etc.)
928[16:24:37] <rudi_s> ratrace: Well, it can use nftables as
backend. - Still, some features of the interface are not possible
with nftables and can be quite useful (and better performing).
929[16:25:06] <rudi_s> But sure, it your decision. Nothing is
forcing you to change (especially with the compatibility mode
nftables provides).
962[16:47:24] <ratrace> jack2019: are you talking about sector
size of your disk?
963[16:47:54] <jack2019> ratrace, I have a ssd which support
4096
964[16:48:05] <ratrace> are you talking about sector size?
965[16:48:30] <jack2019> ratrace, Yes
966[16:48:38] *** Quits: Night-Shade (~TimF@replaced-ip) (Quit: My MacBook has gone to sleep. ZZZzzz…)
967[16:49:01] *** Joins: jerry (~jerry@replaced-ip)
968[16:49:15] <ratrace> so what exactly you want to change? you
can't change your disk sector size, that's baked into it.
you could change your filesystem block size or other layers',
but......
969[16:49:19] <ratrace> !xy
970[16:49:19] <dpkg> Slow down for a bit! Are you sure that you
need to jump through that particular hoop to achieve your goal? We
suspect you don't, so why don't you back up a bit and tell
us about the overall objective... We know that people often falsely
diagnose problems because they are too close to them -- it's
easy to miss that there is a better way to proceed. See
replaced-url
971[16:49:46] <ratrace> actually.... I think some drives DO
allow sector size change..... via hdparm if I'm not mistaken
972[16:49:52] *** Quits: MegaCarp (~carp____@replaced-ip) (Remote host closed the connection)
978[16:51:45] <ratrace> I don't know, never tried it. I
doubt it, though. if you change sector size, all the mappings will
be off. so.. probably better do it assuming no data on it
979[16:53:02] <jack2019> ratrace, thank you, I will read the
manual (hdparm)
980[16:53:18] <ratrace> jack2019: btw, is that sata or nvme ssd?
986[16:57:39] <pankaj> I have never experienced this before on
my laptop. After installing debian I have noticed that the system
shuts down intermittently.
994[17:07:41] <fpombal> hi, I have 2 questions about packaging
backports. This is in the context of uploading to Ubuntu PPAs, but
so far I have not obtained a response from #ubuntu, so I'm
reposting here in hopes someone else can help out. Thanks
1004[17:16:42] <de-facto> How can i use
replaced-url
1005[17:17:29] <de-facto> cant find it in
replaced-url
1006[17:17:52] <fpombal> ratrace: well, at least the part of
getting sources should be the same right? I just want to know how to
use git properly instead of dget/apt source to get the sources +
dsc/tarballs + apply debian packages automatically - this is not
Ubuntu specific, right?
1009[17:18:51] <ratrace> de-facto: it says so right there on the
link you posted, under "Source code and support"
1010[17:19:11] <ratrace> fpombal: that's the thing,
it's not. launchpad has LP specific tools for that.
1011[17:19:44] <ratrace> fpombal: and also, are you building
packages for ubuntu, or trying to use PPAs for debian....?
1012[17:20:11] <de-facto> it seems its contained in
replaced-url
1013[17:20:27] <fpombal> ratrace: I'm building for Ubuntu
1014[17:20:49] <ratrace> fpombal: so you should really ask in an
ubuntu dev channel, how to use LP tools to build software for
ubuntu.
1015[17:20:59] <fpombal> ratrace: so does `apt source` only
output this line on Ubuntu? "NOTICE: <package> packaging
is maintained in the 'Git' version control system at:
(...) Please use: `git clone
replaced-url
1016[17:21:07] <ratrace> that's not really even tangentially
related to debian. it's very much LP specific.
1017[17:24:36] <fpombal> I already know how to build and upload
to Ubuntu PPAs, I'm just asking if I can use git to get the
sources in the first place. apt is telling me it is possible, and
apt is Debian tool, right? So at least that part is still a valid
question for this channel
1018[17:24:48] <fpombal> s/if/how
1019[17:25:47] <oerheks> fpombal, hence the different build
procedure, you should ask in #ubuntu-packaging or #ubuntu main
channel.
1020[17:26:23] <ratrace> fpombal: actually, that's not true.
apt in random derivative distro is not on topic for this channel.
that said. iirc you could use git to fetch the sources from debian,
but the packaging and build procedure is LP specific.
1021[17:26:47] <ratrace> it's very much better to get the
sources using those same LP tools. I forgot the name of the suite,
it's been a few years since I built ubuntu pakcages.
1023[17:27:29] <ratrace> the reason for that is that using those
same tools, you use the LP building machinery for each of the
flavors you want to or need to support
1024[17:27:49] <ratrace> and each flavor might require separate
patches, and all that is managed by LP tools.
1025[17:27:55] *** Quits: b_reino (~b_reino@replaced-ip) (Remote host closed the connection)
1028[17:29:43] *** senux_druid is now known as senux
1029[17:29:49] <fpombal> It is my understanding that at least
until the step of actually uploading the package to the PPA, only
standard Debian tools are used, and the process is identical to
preparing source or binary packages in the context of using Debian.
That's my reading of the basic documentation at least:
replaced-url
1030[17:30:14] <isapgswell> xscreen do not deal with the
diference between same monitor and dual gpu
1033[17:30:41] <fpombal> ratrace: for instance, the main tools
are dch and debuild. The manpage for debuild says "debuild -
build a _Debian_ package" (emphasis mine)
1037[17:32:40] <fpombal> (these tools are all part of the
devscripts package:
replaced-url
1038[17:33:31] <ratrace> our questions from your original post,
lines 15 and 16 have nothing to do with debian unless you're
fetching debian packages proper, and not .debs from ubuntu sources
(with ubuntu patches)
1040[17:34:25] *** black_ant is now known as Guest29614
1041[17:35:16] <ratrace> in that you can also clone a git repo,
but the building process is not specific to having sources from git.
also patches are applied during building
1044[17:37:04] *** Quits: de-facto (~de-facto@replaced-ip) (Quit: See you around. 💤)
1045[17:38:39] <fpombal> ratrace: alright, so let's say I
want to build the source package. Forget about the PPA. 1. I clone
the debian sources: `git clone
replaced-url
1046[17:40:01] <ratrace> you don't run git apply on
debian/patches. they're applied during normal package build
process
1050[17:41:32] <ratrace> fpombal: that git repo is basically teh
same thing you obtain via apt source <package>, except
it's the whole git repo, and thus main contain changes,
branches, that aren't released
1056[17:46:56] <fpombal> ratrace: right, so if I just want the
source package, is it enough to run `debuild -S` after cloning and
it will build with all packages from debian/patches applied?
1057[17:48:41] <ratrace> fpombal: probably not, you'd want
specific release tag most likely
1058[17:49:08] <ratrace> unless you want to build from the git
master for that package
1061[17:55:23] <fpombal> ratrace: ok, please assume I checkout my
preferred tag after cloning the repo, such as debian/1.0.0-1. Then
is it sufficient to run `debuild -S` for it to build with all
packages from debian/patches applied?
1067[18:04:35] <fpombal> ratrace: alright, I just tried it out
(cloning, checking out the tag and running `debuild -S
--no-check-builddeps`), and this is the error I get:
1071[18:05:52] <fpombal> This package has a Debian revision
number but there does not seem to be an appropriate original tar
file or .orig directory in the parent directory; (expected one of
foo_1.0.1.orig.tar.gz, foo_1.0.1.orig.tar.bz2,
foo_1.0.1.orig.tar.lzma, foo_1.0.1.orig.tar.xz or foo.orig) continue
anyway? (y/n)
1072[18:06:52] *** Quits: senux (~senux@replaced-ip) (Remote host closed the connection)
1073[18:07:03] <fpombal> obviously it fails after continuing. So
I need to get the original tarball from somewhere. Is this done with
some debuild flag or some other deb* tool? (of course I can always
do it manually, but I would expect there would be some automatic way
of getting this tarball)
1077[18:08:26] <fpombal> ratrace: I want to do it with git,
because even apt source is advising me to do so
1078[18:08:49] *** Quits: null1337 (~WhoAmI@replaced-ip) (Quit: Give a man a fish and he will eat for a day. Teach him how
to fish, and he will sit in a boat and drink beer all day)
1079[18:08:54] <ratrace> I suppose with git cloning you
don't have the same directory structure, with source tarball,
debianized tarball and the dsc.... so I could've been wrong
about git workflow not being specific
1080[18:09:37] <ratrace> fpombal: the git repo is really for
maintainers. the code should be the same, in the apt source
installed pacakge, and corresponding git repo tag
1081[18:09:53] <ratrace> thus... the git repo might have
additional branches, and modifications pending for next tagging
1082[18:11:06] <ratrace> and so we go back to the original
premise that building for ubuntu might be different than building
for debian. on debian, unless, again, you're a maintainer and
want to push your changes to salsa... the recommended way is apt
source
1083[18:12:15] <fpombal> ratrace: but apt source is recommending
git; and my interpretation of the notice (the way it's worded)
does not make it seem like it is only for maintainers
1084[18:12:34] <ratrace> also, dunno for ubuntu's apt, but
on debian, it's not telling you to use git. it's stating
to use git [if you want] to retreive latest (possibly unreleased)
updates to the package
1088[18:15:06] <ratrace> fpombal: looks like there's git
specific tooling, on debian:
replaced-url
1089[18:16:16] <fpombal> I mean if you want _any_ changes more
recent than the one available in the distro's default repos,
unreleased or not, you either have to use git and checkout the tag
you want or call dget pointing to a URL with the appropriate dsc
file.
1091[18:16:57] <fpombal> apt source just gets the source package
of the package in the distro. I want a more recent source from the
repos of a more recent version of the distro
1093[18:17:25] <ratrace> you can use sources URL for the distro
you want, like testing or sid.
1094[18:17:49] <fpombal> hence why I think " [if you want]
to retreive latest (possibly unreleased) updates to the package
" is not as restrictive as I think you're implying
1098[18:18:54] <fpombal> alright I'm checking out
replaced-url
1099[18:18:58] <ratrace> depends on what you really want. again,
ubuntu process is different. on debian, you typically use the src
repo for the release you're on, eg. stable, but that's not
mandatory. you can use testing or sid's repo if you want to
backport
1100[18:19:19] <ratrace> on ubuntu, PPAs usualy (must) have all
the flavors covered and the process is very much different
1102[18:20:46] <ratrace> at any rate, apt is not _recommending_
to use git over itself. for good number of use cases you want to
base on a released version. for some use cases you might want the
tip of debianized package, which is git master, although personally
I'd still use released versions, from sid.
1103[18:21:06] <ratrace> git master could easily have breaking
changes, which unless you're maintainer of that package, are
not obvious how to fix.
1105[18:24:20] <fpombal> ratrace: ok, the assume I really want
git master. after cloning, how do I get past the error I posted
above? I still have to get the dsc and such from somewhere
1108[18:25:39] <ratrace> fpombal: well did you check that wiki
page? it shows you the procedure step by step, first of which is
importing orig tarballs
1111[18:33:23] <fpombal> ratrace: seems like running gbp clone
<repo> && cd <repo> && gbp buildpackage
nearly works. that last commands automagically generates the orig
tarball. The problem now is that it's actually trying to build
the binary package, which I don't want
1112[18:35:48] <fpombal> I suppose gbp buildpackage takes some
kind of flag to instruct it to do a source build, I'll try to
figure it out
1113[18:37:45] <ratrace> there's extensive documentation
linked from that wiki page
1115[18:40:48] <fpombal> ratrace: looks like the manpage has what
I'm looking for: (...) passing along all arguments given to gbp
buildpackage on the command line that don't start with --git-.
1116[18:41:30] *** Quits: Vizva (~Vizva@replaced-ip) (Remote host closed the connection)
1117[18:42:13] <fpombal> so gbp buildpackage -S
--no-check-builddep does what I want. Now I just need to figure out
the rest of the workflow (handling changelog, patches, etc), and I
should be able to use git/gbp for everything instead of
dget/dch/debuild etc
1118[18:42:26] <filePeter> My new desktop computer has an SSD
drive. As I recall the fstab option „discard“ was used
for my crypttab and ext4 filesystem. Is this option still relevant
these days? (Can I erase my 10 year old memories about having to use
»discard«?)
1119[18:43:13] <fpombal> (well I guess the last step for
uploading to the ppa will still be the classic dput command, but
that's not Debian-related)
1120[18:43:24] <fpombal> ratrace: thanks for bearing with me
1126[18:46:31] <jmcnaught> filePeter: I think instead of discard
in fstab it is recommended to use the fstrim.service and
fstrim.timer which does discard once a week.
1226[20:18:04] <SponiX> Growly: you are on Debian 10 Buster ?
1227[20:18:10] <Growly> bullseye
1228[20:18:28] <petn-randall> !debian-next
1229[20:18:28] <dpkg> #debian-next is the channel for
testing/unstable support on the OFTC network (irc.oftc.net), *not*
on freenode. If you get "Cannot join #debian-next (Channel is
invite only)." it means you did not read it's on
irc.oftc.net. See also
replaced-url
1230[20:18:40] <Growly> ok. thanks.
1231[20:18:57] <Growly> in any case, it turns out i had to
install graphviz, thanks to some comment on this bug:
replaced-url
1232[20:20:58] *** Quits: rwp (~invalid@replaced-ip) (Remote host closed the connection)
1277[21:05:12] <hiya> I want latest drivers, firmware, gnome etc
1278[21:05:31] <hiya> Is Debian testing stable vs let us say
Arch?
1279[21:05:40] *** Quits: Vizva (~Vizva@replaced-ip) (Remote host closed the connection)
1280[21:05:50] *** Quits: Grldfrdom (uid391113@replaced-ip) (Quit: Connection closed for inactivity)
1281[21:06:00] *** Quits: earthundead (~earthunde@replaced-ip) (Remote host closed the connection)
1282[21:06:14] <ratrace> !sns
1283[21:06:14] <dpkg> Shiny New Shit Syndrome is a serious
disorder, which usually breaks out into an epidemic every time
something new is released. If you have SNS, ask me about
<backports> and <ssb>; these are better options than
upgrading to <testing> because it is a <moving target>.
1284[21:06:48] <ratrace> if you're looking for "latest
<xyzzy>" then maybe debian is not for you. if you want to
stay in .deb land, consider ubuntu.
1285[21:06:52] <jhutchins> hiya: No, testing is not stable. It is
meant to be broken so things can be fixed.
1286[21:07:10] <sney> ^ that said, bullseye is frozen, so
it's likely more stable than some chaotic rolling distros
1288[21:08:12] <jhutchins> sney: Yeah, but freeze is when they
sometimes tackle the big problems that keep important things from
running, and they take the time to really fix them instead of just
putting on bans-aids.
1289[21:08:37] <jhutchins> F'r instance, the Intel graphics
driver, which took more than a month to fix.
1292[21:11:01] <hiya> ratrace: but I have a new laptop
1293[21:11:13] <hiya> Or else I am using Debian stable since last
three years
1294[21:11:23] <riff-IRC> I'm running testing on my main
computer and I've never seen anything majorly busticated.
1295[21:11:23] <sney> the only comparable things I'm aware
of this time around are the dkms stuff that was handled during 5.9,
and the gcc10 upgrade mess.
1297[21:11:40] <sney> bullseye has been (relatively) smooth
compared to some testing cycles in the past
1298[21:11:52] <hiya> bullseye is 11 correct?
1299[21:12:01] <hiya> Debian is releasing it now?
1300[21:12:01] <sney> it will be, yes
1301[21:12:10] <hiya> wow
1302[21:12:13] <sney> release will probably be 6-8 months from
now
1303[21:12:16] <riff-IRC> I know that Nautilus was passing
invalid parameters to ffmpegthumbnailer for a bit, which killed
thumbnailing
1304[21:12:18] <hiya> :(
1305[21:12:28] <hiya> But I can use bullseye?
1306[21:12:34] <sney> dpkg: buster->bullseye
1307[21:12:34] <dpkg> In /etc/apt/sources.list, change
"buster" to "bullseye", change
"bullseye/updates" to "bullseye-security" on the
security line; remove lines like buster-backports, debian-multimedia
<dmo>, and other 3rd party repos as they are known to cause
issues then do: apt update && apt upgrade && apt
full-upgrade. Note that testing is a <moving target> and may
be buggy, and read the sid FAQ:
replaced-url
1308[21:12:58] <riff-IRC> spent about 8 hours tracing the root
cause of that one, then filed a bug report
1309[21:13:20] <riff-IRC> and boom, issue fixed with the next
round of updates
1310[21:13:31] <hiya> I think Fedora is the next best thing then.
Most free OS if you want a working Wifi. I <3 Debian so much
1311[21:13:46] <rudds> what dictates whether a given package gets
updated via backports or not? i'm a bit new to Debian land, and
i'm curious why lxc (for example) doesn't have newer
versions packaged in buster-backports
1312[21:13:59] <sney> rudds: demand and difficulty, mostly
1313[21:14:35] <sney> sometimes a package that could be
backported isn't just because nobody requested it. other
packages are too complex with too many moving parts to make it
worthwhile.
1315[21:15:02] <rudds> interesting. the version of lxc that
appears to have been frozen for Bullseye stable has some major
issues with systemd/v2 cgroups
1316[21:15:25] <rudds> so i wonder what will happen down the line
with Debian packaging if that gets fixed on the lxc side
1317[21:15:39] <sney> check the changelog, package versions in
debian don't always correspond identically to the upstream
version. patches are backported all the time.
1319[21:16:35] <sney> that's kind of the main thing about
stable, the program version/abi is frozen but security and bugfix
patches are applied as long as the stable release is supported
1320[21:16:50] <rudds> ah, that makes sense. where's the
best place to look at the notes for Debian-specific patches?
1321[21:17:08] <sney> 'apt changelog packagename' on
your system, or tracker.debian.org has all of that info and then
some
1322[21:17:13] <rudds> i'm mostly coming from FreeBSD where
the ports tree is full of custom patches
1336[21:29:00] <rudds> hmm, the current problem is that core lxc
functionality fails on distros with pure unified cgroups (e.g.
Bullseye), unless you do some manual systemd-run workarounds
1338[21:29:15] <jhutchins> rudds: Backports have fature upgrades,
while stable only has fixes. A backport happens when someone who is
capable of building a binary .deb is interested enough in doing it.
They should be motivated to maintain the backport, not just do a
one-off.
1341[21:30:17] <rudds> that makes sense, i'll keep an eye on
things as Bullseye is getting closer
1342[21:31:04] <rudds> i'm harping on lxc so much because i
was planning to move my home server/NAS from BSD to Bullseye once it
goes final, but i rely on containers extensively so if something is
broken there it'll be harder to do what i want
1377[22:30:59] <dpkg> #debian-next is the channel for
testing/unstable support on the OFTC network (irc.oftc.net), *not*
on freenode. If you get "Cannot join #debian-next (Channel is
invite only)." it means you did not read it's on
irc.oftc.net. See also
replaced-url