--- Log opened Fri Jul 06 00:00:27 2018
00:04 < jcarpenter2> linux namespaces are confusing
00:04 < searedvandal> oh?
00:05 < aBound> It's confusing I tell you. :P
00:10 < geckos> I was messing up with vm.dirty_ratio. What would be a good way to measure its effects?
00:12 < geckos> I know that vm.dirty_ratio controls I/O blocks on kernel, is there any counter of this I/O blocking? So that I can measure its frequence?
00:27 < jcarpenter2> say i want to have a JSON RPC service running on my machine, but I don't want to give access to it to random users that log in.  (In fact, I only want one user to be able to call the RPC server).  Can this sort of sandboxing be achieved using namespaces?
00:30 < jcarpenter2> My understanding was that anybody in a parent namespace can enter a child namespace needing no additional privileges.  That was from the quote "With PID namespace isolation, processes in the child namespace have no way of knowing of the parent process’s existence. However, processes in the parent namespace have a complete view of processes in the child namespace, as if they were any other process in the parent namespace" from htt
00:30 < jcarpenter2> ps://www.toptal.com/linux/separation-anxiety-isolating-your-system-with-linux-namespaces assuming that network namespaces are similar to PID namespaces in that regard.
00:31 < jcarpenter2> however, it looks like only root can enter a network namespace at least using the "ip netns exec" command
00:31 < jcarpenter2> ...
00:35 < jcarpenter2> man, it'll be nice when i get to close all these tabs
00:36 < lnnb> can you use unix sockets or this has to be on the ip network ?
00:37 < jcarpenter2> it has to be on IP
00:40 < jcarpenter2> "ip netns add [namespace]" and "ip netns exec [namespace] [command]" both fail with "Operation not permitted" when I run as non-root.  I'm not sure whether this is the `ip` command's doing an I could still connect by writing a C program and making a `setns(2)` call or something
00:42 < jcarpenter2> Also, I'm not sure this means that I couldn't call into an HTTP service running in the child namespace - especially after I connect the namespace to the internet as per https://blogs.igalia.com/dpino/2016/04/10/network-namespaces/
00:42 < lnnb> strace it
00:43 < lnnb> oh you didn't try the setns syscall directly yet?
00:43 < lnnb> it probably won't work
00:43 < lnnb> unless you use that user namespaces thing
00:45 < lnnb> you could use a launcher (or init) to spawn the users shell in the namespace
00:46 < caene> I'm having issues getting a NTFS USB drive to work with samba. I have a internal ext4 drive that is working flawlessly with samba, that I can access through windows. But I can't get access to the NTFS drive through windows. It mounts just fine, and the permissions on the mount are sorted (same permissions as the EXT4 drive). Anyone had this issue?
00:47 < jcarpenter2> just checkin' the strace output
00:49 < jcarpenter2> apparently "ip netns add [namespace]" uses unshare(2) and "ip netns exec [namespace] [command]" uses both that and setns(2), but that still doesn't entirely convince me that a program couldn't be written making those calls that successfully connects
00:51 < lnnb> jcarpenter2: the kernel is going to give you EPERM if you dont' have CAP_SYS_ADMIN
00:51 < jcarpenter2> awesome - now that you mention it, i see that in the setns(2) man page
00:53 < jcarpenter2> that means unprivileged users can't enter the network namespace, but i'm not 100% convinced they can't send IP packets to it
00:54 < lnnb> you would have to set up an interface, by default everything is disabled even the loopback
00:55 < LambdaComplex> what should i read to learn about DPDK?
00:56 < ayecee> what is a dpdk
00:57 < lnnb> democratic peoples dominion of korea, duh
00:57 < xamithan> I think they mean data plane development kit
00:57 < LambdaComplex> 10 points to xamithan
00:57 < LambdaComplex> good job :)
00:58 < ayecee> okay, what is a data plane development kit
00:58 < mawk> PCM is big endian ?
00:58 < ayecee> not by nature
00:58 < ayecee> the pcm encoding you're using may be
00:58 < LambdaComplex> ayecee: if i had a solid answer to that question then maybe i wouldn't be asking about it
00:58 < LambdaComplex> :D
00:59 < mawk> the standard LPCM used in the WAV format
00:59 < mawk> for instance
00:59 < ayecee> LambdaComplex: looks like you're in good company
00:59 < lnnb> there might even be a bit in the header that says what  it's encoded in
00:59 < ayecee> with other people who don't know what DPDK is
01:03 < allotrope94> Any Linux 2-in-1 tablets you all would recommend to run Lubuntu?
01:04 < allotrope94> I am considering the Asus T100T/T100TA
01:04 < allotrope94> There are a couple used units for $50 on eBay
01:23 < jcarpenter2> If I want to write a bash script that sets up my network namespace (running it via a systemd unit in /etc/systemd/system) each time i boot up the computer, where should I put my script according to the FHS?  I'm thinking /usr/local/bin, it's just that it's a script, not a binary
01:25 < bls> scripts can go in bin dirs
01:26 < jcarpenter2> ok-
01:26 < jim> jcarpenter2, is /usr mounted when the script is run?
01:26 < bls> the whole libexec or /usr/.*/share thing has fallen to the wayside
01:26 < jcarpenter2> yeah it is
01:26 < bls> or even /usr/lib
01:26 < jcarpenter2> if it weren't i guess it would go in /bin
01:27 < jim> howbout /usr/local/sbin
01:27 < bls> wouldn't want to put anything in /bin unless you've packaged it up
01:27 < jcarpenter2> sbin would make even more sense
01:28 < jcarpenter2> now i'm looking for how to give a user CAP_SYS_ADMIN on a network namespace i've created
01:28 < jim> I dunno about other dists. but debian's packages are not permitted to touch /usr/local (and a couple other places)
01:29 < jcarpenter2> i could of course use `sudoers(5)` for that but it would be cool if i could use the tools that setns(2) offers
01:29 < phogg> accidentally hit CTRL+L with xchat in focus... and it clears the scrollback. Ugh.
01:30 < jcarpenter2> so that i can have this one privileged user run commands in the namespace
01:30 < bls> that's a pretty unintuitive keybinding
01:31 < jim> jcarpenter2, well a script like that needs to run as root, and the caller (systemd?) runs as root
01:32 < jcarpenter2> yeah, i'm not talking about the setup script, but about the server and scripts that are to run in the namespace once it is created
01:32 < jcarpenter2> some terminals use Ctrl+L to clear the screen so it's got some precedent
01:33 < jim> jcarpenter2, oh, I guess you can do that if you wanted to
01:33 < lnnb> in gdb CTRL+L refreshes the screen,  useful when it starts glitching out
01:33 < jcarpenter2> i'd like the server and scripts to be run by a non-root user, and again (i think) i could use sudoers(5) to declare that this user can run `ip netns exec *` as root (not sure about the wildcard) but `setns` will run if you have CAP_SYS_ADMIN in the namespace, so i'd like to be able to grant that
01:34 < jcarpenter2> then i could just run `ip netns exec *` as the user themselves, not in sudo-form
01:34 < jim> I'm not completely understanding why the setup script needs to be run as a sudoing user, I guess for logs?
01:34 < lnnb> why does the user have to enter the NS?
01:35 < mawk> sudo looks like a bad idea to do that jcarpenter2
01:35 < mawk> I'd do a setuid binary that will enter the namespace
01:36 < mawk> it's very simple
01:36 < jcarpenter2> well, i should look more into that actually...  just running "ip netns add [namespace]" as an unprivileged user fails with "operation not permitted"
01:36 < mawk> indeed
01:37 < mawk> so you create the namespace as root, and make a setuid binary that will enter the namespace and execute a command in it
01:37 < mawk> it's 5 lines of C
01:37 < jcarpenter2> i know, and it's even already written
01:37 < lnnb> to check that it's entering the right namespace is way more than 5 lines
01:38 < mawk> what do you mean by entering the right namespace lnnb ?
01:38 < mawk> the binary would hardcode the path to the iproute2 namespace
01:38 < mawk> something like /var/lib/iproute2/mywonderfulnamespace
01:38 < lnnb> ok take shortcuts
01:38 < lnnb> see if i care
01:40 < jcarpenter2> `ip netns exec [namespace] [command]` could be added to sudoers(5) so that my user can run it through sudo
01:41 < jcarpenter2> afaik setuid on its own has some vulnerabilities, like PATH injection and IFS injection - it's generally recommended to use a command like sudo, or super
01:41 < mawk> IFS injection ?
01:41 < jcarpenter2> setting the IFS variable and then running a setuid binary
01:42 < mawk> IFS is a bash thing
01:42 < jcarpenter2> IFS only affects scripts that are run under setuid
01:43 < mawk> IFS is a bash variable that tells how to split input strings
01:43 < jcarpenter2> well, scripts in general, not excluding ones that are run with setuid
01:43 < mawk> IFS = input field separator
01:43 < mawk> yeah
01:43 < mawk> you can't have a setuid script
01:43 < mawk> it's a setuid binary
01:43 < mawk> and second, the binary would be simple: setns followed by setresuid
01:44 < mawk> but you can try with sudoers too if you can guarantee that there will be no injection too
01:45 < mawk> you can't allow arbitrary network namespaces, and you can't allow arbitrary commands
01:45 < jcarpenter2> about IFS i was just saying what i had read on this page http://manpages.ubuntu.com/manpages/artful/man1/super.1.html
01:45 < mawk> you need to downgrade to the regular user
01:45 < mawk> it looks pretty complicated to do that with sudoers
01:45 < jcarpenter2> i'm not sure what the attacks are
01:46 < bls> weird docs: "It is intended to be a secure alternative to making scripts setuid"
01:47 < mawk> yeah
01:47 < lnnb> lol were setuid scripts allowed at some point?
01:47 < jcarpenter2> a binary could run a script and IFS would carry over wouldn't it?
01:47 < bls> noticed that looking for the history section (there wasn't one)
01:48 < mawk> yes jcarpenter2
01:48 < jcarpenter2> then that's why
01:48 < lnnb> i don't have that manual :(
01:48 < mawk> well actually with my current bash version no
01:48 < mawk> IFS can't come from an environment variable it seems
01:51 < jcarpenter2> anyway, i could add `ip netns exec [namespace] [command]` lines to sudoers so that my unprivileged user can run commands in the namespace, but i would rather either A) figure out how to create and configure the namespace entirely as a non-root user, or B) give the non-root user the CAP_SYS_ADMIN capability inside the namespace so i don't have to add that stuff to sudoers
01:52 < mawk> ...
01:52 < mawk> just go with a setuid binary
01:53 < mawk> the sudoers thing can't be done safely at all, A) doesn't exit, B) is akin to giving plain root rights
01:53 < mawk> exist*
01:53 < mawk> I don't see what bothers you with the setuid binaryt
01:53 < jcarpenter2> again i could write a setuid binary but the binary code i would write is already written as `ip netns exec` so i would just be re-inventing the wheel
01:53 < lnnb> it's stigmatized by internet propaganda
01:54 < jcarpenter2> poorly and insecurely, i might add
01:54 < mawk> you don't consume any user input at all while being root: you just do a single setns(), then you transition as non-root, and only then you can start processing user input
01:54 < jcarpenter2> not sure what you mean about doesn't exit
01:54 < mawk> I meant doesn't exist
01:54 < jcarpenter2> oh
01:54 < mawk> you can't transition to a namespace if you don't have the CAP_SYS_ADMIN capability
01:54 < jcarpenter2> B is not akin to giving plain root rights, it's only CAP_SYS_ADMIN inside the namespace
01:54 < jcarpenter2> you have to have that inside the namespace
01:54 < mawk> it's not reinventing the wheel, ip netns exec does a single thing that takes a single line of C
01:54 < lnnb> though stack clash was a very real problem :(
01:55 < mawk> CAP_SYS_ADMIN is enough to do a lot of things to the system
01:55 < mawk> just look at the list
01:56 < mawk> it's akin to giving root rights
01:56 < jcarpenter2> "A process reassociating itself with a user namespace must have the CAP_SYS_ADMIN capability in the target user namespace." http://man7.org/linux/man-pages/man2/setns.2.html
01:56 < mawk> yes
01:57 < mawk> i.e. be some flavor of root on the main system
01:57 < mawk> because we're not talking about a network namespace coupled with a user namespace here
01:57 < mawk> or I missed something
01:57 < jcarpenter2> no, we're not talking about a user namespace + network namespace
01:57 < mawk> righty
01:59 < mawk> you created your network namespace with ip netns I assume jcarpenter2
01:59 < mawk> with a fixed name
01:59 < jcarpenter2> yes
01:59 < mawk> so the namespace lives in /var/run/netns/$NAME
01:59 < jcarpenter2> correct
02:01 < mawk> then there's no wheel to reinvent, you do int fd = open("/var/run/netns/mynamespace", O_RDONLY); setns(fd, CLONE_NEWNET);
02:01 < mawk> and that's all
02:02 < mawk> you get back to the regular user, and you can execvp(argv[1], argv+1);
02:02 < lnnb> if you're opening files in setuid program you really ought to check permissions
02:03 < mawk> yeah
02:03 < mawk> and add a little error checking
02:04 < lnnb> it's easier if you set effective uid, but it gets complicated
02:05 < lnnb> or drop {DAC,MAC}_OVERRIDE
02:05 < lnnb> s/or/and/
02:05 < jcarpenter2> so i guess i need to either couple the network namespace with a user namespace or add some stuff to sudoers
02:05 < mawk> ...
02:06 < mawk> you just keep listing incredibly complicated and/or insecure solution
02:06 < mawk> while disdaining real solutions
02:06 < mawk> I don't get it
02:06 < mawk> just try to do the sudoers thing for instance
02:06 < jcarpenter2> it's pretty easy actually
02:07 < mawk> yeah, it's easy to allow to run the `ip netns exec' command as root
02:07 < mawk> then the user can enter any namespace it wants, also he can be root now
02:08 < jcarpenter2> even list out specific arguments that they can run it with
02:08 < mawk> yeah, totally safe
02:09 < jcarpenter2> you can allow `ip netns exec` and you can also allow `ip netns exec [namespace] [command]` for particular namespace and command
02:09 < mawk> so the user will type in everytime ip netns exec mynamespace sudo -u myuser $command
02:09 < mawk> with sudo in front of that
02:09 < jcarpenter2> right
02:10 < jcarpenter2> give or take second sudo
02:11 < mawk> if you don't put the second sudo the user will be root
02:11 < mawk> that's not what you want
02:11 < jcarpenter2> oh right
02:11 < mawk> also you want sudo -u myuser --
02:11 < mawk> to prevent injection to sudo commandline
02:12 < jcarpenter2> i see
02:12 < lnnb> could always run a daemon in the namespace over a unix socket that checks creds to authenticate accept(), and acts on the users behalf
02:18 < my_mind> i know my cpu core 2 quad Q6600 is supported by vmware because it works in linux mint
02:20 < lnnb> my_mind: lscpu will also tell you that you have VT-x support
02:23 < my_mind> lnnb: yes, lscpu says I have VT-x
02:27 < my_mind> can i install vmware player on a live kubuntu usb to see if it will work when I click ignore?
02:27 < my_mind> I wanna keep my main machine as clean as possible
02:28 < my_mind> i'll boot into live usb, and then reach into the download dir of my main drive then run the vmware package from there
02:28 < bls> should be able to, but might have to set up persistence if you don't have enough RAM
02:29 < my_mind> 6gb
02:29 < my_mind> got it
02:31 < oerheks> all guides starts with "sudo apt install build-essential gcc" .. that would be fun on  al ive iso
02:32 < my_mind> what's gcc?
02:32 < bls> a C source code compiler
02:32 < my_mind> ok
02:32 < bls> likely needs it to build the vmware extensions
02:34 < stevendale> Hey all
02:35 < stevendale> Does anybody remember FreeCraft? It was a WarCraft clone by the Stratagus team that got shutdown personally by Blizzard, before Wargus became a thing
02:35 < my_mind> sounds complicated. i'll just use my laptop for vms
02:35 < my_mind> i don't care what happens to my laptops os
02:36 < sh1ro> vms shouldn't be able to fuck up your os
02:36 < sh1ro> that's like the whole point
02:36 < my_mind> no i mean installing vmware
02:37 < my_mind> if the install doesn't go right, i'm not good at linux yet to troubleshoot it
02:37 < sh1ro> uhh i've never had installing virualization software break anything on any os
02:38 < my_mind> if I install another kubuntu os on my second hard drive, will it mess with bootloader?
02:39 < my_mind> as in if i install a second os, ansd then delete it when i'm done testing, will i have bootloader issues?
02:39 < bls> there's not really any reason to install multiple
02:39 < sh1ro> yeah you're just overcomplicating things for yourself
02:40 < drakonan> So I'm trying to think of some ways to passively determine from a pi that i have internet or have very recently...
02:40 < drakonan> without actually actively pinging the internet somewhere
02:40 < drakonan> or otherwise wasting bandwidth have 5 minutes forever
02:40 < drakonan> every*
02:40 < bls> DNS lookup?
02:41 < drakonan> thats what i was thinking... but it would need to not actively query like i wonder if i can ask if there is a cache for it
02:41 < my_mind> yeah i over analyze every decision in my life. i wish i didnt care so much
02:41 < drakonan> for something like netflix is prob one that would never not be cached normally in my home
02:41 < drakonan> unless the net is down
02:41 < bls> depends on your distro, but most have something like dnsmask or nscd to cache queries
02:42 < bls> but you can use a tool like dig to bypass it
02:42 < drakonan> idk i guess it could be an active query too because the only time it would be would be when the net is down which wouldnt matter or recently back up or no one is home using bandwidth so no point in not wasting i suppose
02:42 < sh1ro> ping 1.1.1.1?
02:42 < drakonan> i just hate the idea of a ping every 5 minutes all the time
02:42 < bls> yeah, better off just configuring things with the knowledge it can go down
02:43 < drakonan> id like to set up a watchdog script to reboot my modem and check some basic stuff like my vpn is correctly forwarded (so i always have access)
02:43 < notmike> Do it!
02:43 < drakonan> but dont want to waste bandwidth pinging stuff all the time
02:44 < lnnb> 56 bytes is a serious concern?
02:44 < bls> if your router has a web interface, you could possibly query it to see if it's synced
02:45 < notmike> drakonan: solarwinds toolkit
02:45 < bls> because the chances of your access being down and the modem being in sync are pretty low
02:45 < drakonan> well one thing that is interesting is the upnp info... how does upnp know or the modem guesstimate its bandwith?
02:45 < stevendale> Texting here on IRC is using more bandwidth than pinging every 10 minutes would use in a whole day
02:46 < drakonan> it's not super acurate but close... wonder how often it reassesses... ie if my crap dsl is messing up wonder how quickly upnp would figure out its not right
02:46 < fattredd> Hey. I'm looking for a way to list all the disk drive attached to my system, except sda. I don't want to list the partitions. Right not all I can think of is:
02:46 < fattredd> $ ls /dev/sd* | grep -v a | grep -v 1 | grep -v 2
02:46 < fattredd> Which is hardly pretty
02:47 < bls> don't parse ls output, use something disk specific like lsblk
02:49 < bls> it has filters
02:50 < sveinse> Does anyone here now precisely why java requires the insane amount of PIDs that it does? My server runs a handful of services, and then comes a java app at staggering 888 PIDs. No, I'm not joking.
02:50 < kristina> do many things support exotic libcs like bionic?
02:50 < bls> getting a PID per thread due to its non-native threading implementation?
02:51 < sveinse> bls: yeah, but still, 888 threads?
02:52 < fattredd> Okay so maybe $ lsblk -o Name -p -P -S
02:52 < fattredd> But that still outputs NAME and /dev/sda
02:52 < kristina> linux doesn't have processes or threads, it only has tasks, TID is internally, the task id. pid is a leader of a task group for tasks that have usermode contexts.
02:53 < wr> https://en.wikipedia.org/wiki/List_of_Unix_commands any place can find bigger list than this of unix/linux commands?
02:54 < mawk> we're not in the kernel kristina
02:54 < mawk> don't be silly
02:54 < lnnb> mawk: be careful you're about to get supertrolled
02:54 < fattredd> ls /*bin
02:54 < mawk> by who
02:55 < kristina> well i'm not wrong and i try to get people to see that linux is not unix. and over time it'll be less and less unixey.
02:56 < mawk> yeah
02:56 < sveinse> well, regardless of being in userspace or in kernel scope, I really don't understand the threading model of java. I mean with so many threads/tasks in a small application, it means that you literally start a new thread/task to call printf
02:57 < kristina> the whole task unification happened with new style threads (LWP) being introduced, at that point process became a synonym for "leader of a task group". and tid became a synonym for "task".
02:57 < meyou> i know most of those words
02:58 < jim> kristina, some of it has its roots in how C handles stream I/O
02:59 < kristina> you mean the buffered io like fprintf etc.
02:59 < kristina> right?
03:00 < hatp> Anything quick I can try if my dns suddenly failed? I can ping addresses but not domains
03:00 < meyou> ping your nameservers
03:00 < jim> well partly, but mostly how |, < and > work to join certain file descriptors together
03:00 < meyou> do a manual lookup with dig or something, see what the error is
03:01 < kristina> file descriptors are a misnomer, they stopped being that a very long time ago.
03:02 < nekoseam> test1
03:02 < jim> kristina, internally, they still exist and are buried in the data structure referred to by the output of fopen()
03:02 < kristina> just call it a lx_port_t or lx_handle_t already.
03:02 < kristina> it's an abstract handle.
03:03 < kristina> fopen is not a syscall.
03:03 < kristina> and handles are obtained in so many ways in modern linux.
03:03 < kristina> memfd for example, or singalfd, or timerfd.
03:04 < jim> I didn't say it was, mostly referring to the data structure pointer it returns
03:05 < kristina> it's libc stdio, it's very similiar on anything that has unixey roots.
03:06 < jim> I'm not trying to fight with you, just discuss... I notice your temp rising a bit, and I invite you to relax and breathe instead
03:07 < drakonan> So when an employers asks "do you know linux" what is that supposed to mean
03:08 < lnnb> means they don't know what they're trying to ask you
03:08 < kristina> stdio works on "handles" to files but a handle (aka fd) is just an opaque integer, you can ask the kernel what it is, i think fprintf checks that the handle is in fact a handle that repsents an open file.
03:09 < drakonan> feel like its akin to do you know computers :)
03:09 < jim> kristina, exactly my point too ("it's libc/stdio, ... similar on that which has unixy roots")
03:09 < Loshki> drakonan: it means they have a checklist of items, and something called "linux" is one of them.
03:10 < drakonan> yeah probably gnu/linux resumes would get discarded because there's no space between the / :)
03:10 < jim> mac is unix (more lately), it has |, win isn't linux or unix and it has | too... so they share some of the mechanisms
03:10 < kristina> right and stdio is stdio. and i don't mind what they call fds (a typedef somewhere that distinguishes an fd from a handle just for static type checking would be nice though).
03:10 < Loshki> Your larger, crappier companies, pay 3rd parties to scan in resumes, run OCR and then pattern-match the keywords.
03:10 < xamithan> But do you know linux?
03:10 < meyou> everybody thought piping things was cool
03:11 < meyou> so they all stole it
03:11 < drakonan> i'm tryin to do 1337 preb for a coming interview hopefully in the next few days/weeks in infosec
03:11 < xamithan> Thats like asking someone if they know computers
03:11 < Psi-Jack> Heh.
03:12 < Psi-Jack> Fact is, if you have to prep, you don't know enough.
03:12 < jim> meyou, mostly it was C and libc and the output of fopen(), and how | connects fd 0 to fd 1
03:12 < xamithan> I prep those stupid behavorial questions
03:12 < kristina> jim: it's not unix, it's sort of bsd and and mach, now mach provides incredibly powerful ipc based stuff miles ahead of what unix has (merge bus1 already ffs).
03:12 < Loshki> Also, if you practice by doing enough interviews, you get better at them. Who knew?
03:13 < jim> kristina, what do you consider unix to be? bsd is -legally- unix (as if that means anything)...
03:13 < xamithan> I consider unix to be dead
03:13 < kristina> linux won't stop being unix-like, but it's not unix.
03:14 < lnnb> multics > unix
03:14 < jim> clearly
03:14 < jcarpenter2> hurd > multics
03:15 < kristina> ie. bus1 needs to be merged, it needs a syscall, we need an alternatie way of accessing process info like maps in a sane way even if it's over netlink.
03:17 < jim> I'm pretty much out of my depth when it comes to merging busses and process info
03:17 < kristina> bus1 is just port based ipc.
03:19 < jim> same comment applies :) I switched away from C, probably because it all got too detailed
03:20 < notmike> C is the master language
03:20 < jim> I mostly do python now (and occasionally tcl with naviserv/aolserver)
03:21 < jim> notmike, pretty much... and, I'm more comfortable not having to handle "object" memory allocation myself
03:21 < Loshki> I think we ran tops-20 on a DEC-10
03:23 < kristina> there will never be a year of the linux desktop until linux has these problems fixed and torvalds pulls his head out of his ass and rethink the stable versioned kernel driver apis/abi proposal. and we need any something like bus1 and something that provides an alternative to proc and sys (fun fact, bus1 being based on L4 IPC can serve as a less annoying netlink replacement)
03:23 < notmike> jim: you sell yourself short
03:25 < lnnb> i'm quite fond of being able to cat/echo files in proc
03:25 < lnnb> and sys
03:26 < lnnb> less shit devs can hide from us
03:26 < kristina> and let them stay there for humans to look at. programs should NOT have to decode ascii (?) strings.
03:27 < kristina> it's almost as insane as making syscalls by passing buffers with XML encoded parameters.
03:27 < lnnb> what is insane about it?
03:28 < kristina> syscall_xml(const char* buf, size_t sz);
03:29 < lnnb> is being too user friendly a bug?
03:29 < kristina> if there's no machine friendly alternative ABSOLUTELY 100%.
03:30 < lnnb> structs waste memory in alignment, thats not very machine friendly
03:31 < kristina> think about what you just said long and hard.
03:31 < lnnb> its convenient for programmers
03:31 < lnnb> i'll hand you that point
03:32 < WhiteDevil> python or C what do i study ?
03:32 < kristina> kernel isn't for the end user, it's for developers who build stuff that etc etc. and then you have the end user. kernel should not provide that data in ascii without a binary alternative.
03:32 < kristina> without a vfs-free alterntive.
03:33 < WhiteDevil> kristina is a fiesty one aint she
03:33 < WhiteDevil> so C vs python
03:35 < kristina> linux is not an OS unless you don't boot into usermode at all.
03:35 < kristina> it's a kernel.
03:36 < kristina> it's not meant to be super friendly to end users.
03:42 < kristina> arguably, with a support library and some fixes for c++ header issues (can make a quick thing with libtooling to rewrite the headers to be c++ friendly) it's a fairly nice environment to work in.
03:42 < kristina> as long as you're running in a preemptive thread.
03:42 < kristina> or task rather.
03:44 < kristina> i still remember this old abomination, behold, Darwin on Linux: https://github.com/christinaa/kernel_diff/blob/master/kernel.diff
03:46 < kristina> with lines like:
03:46 < kristina> -static int copy_io(unsigned long clone_flags, struct task_struct *tsk)
03:46 < kristina> +static int copy_io_ex(unsigned long clone_flags, struct task_struct *tsk, struct task_struct *from)
03:50 < kristina> also my naming conventions were pretty inconsistent heh: http://crna.cc/weird_naming.c
03:52 < nekoseam> erc is a pretty nice client
03:54 < kristina> i had to annotate a lot of functions to remember if they consumed or obtained references.
03:54 < kristina> thread_port_t* Native_get_thread(struct task_struct* task); /* [RetainPort] */
03:55 < kristina> ie. task_port_t was a handle to an address space and any thread assosciated with it, thread_port_t actually represented task_struct.
03:57 < WhiteDevil> whats a keystroke to get virt-manager or my vm out of full screen mode ?
03:58 < lnnb> "my vm" ?
03:58 < WhiteDevil> virtual machine in virt-manager
03:58 < SpeakerToMeat_> Hello good people
03:58 < SpeakerToMeat_> question, can you people recommend a very simple and SLIM OSS sms gateway?
03:58 < WhiteDevil> am i speaking gibberish ?
03:59 < lnnb> no i just don't know these terms
03:59 < SpeakerToMeat_> WhiteDevil: That'll depend on the client being used... virt-manager.....
04:00 < lnnb> in qemu its ctrl-alt-f
04:00 < SpeakerToMeat_> WhiteDevil: It seems, first ctrl-alt should bring your mouse pointed and in the top middle there should be one button for "exit fullscreen"
04:00 < michaelrose> so stupid question, I decided that /var ought to be on its own dataset, root is zfs, so I synced the file to a different dataset that I would really like to mount at var
04:01 < michaelrose> I get operation not permitted when trying to remove the current contents of var on various dirs in /var/lib/nfs
04:01 < michaelrose> however I've stopped nfs
04:02 < SpeakerToMeat_> The thing is, I have openwrt on my router with very little space left, and plugged in a 3g usb modem.
04:02 < SpeakerToMeat_> I want something I can use remotely (on lan) via api to send and receive sms.
04:02 < michaelrose> and I can't find these files being actually in use
04:02 < SpeakerToMeat_> preferably rest api, so that I can also do the same on the router (for warnings) via simple curl calls
04:03 < phogg> michaelrose: how are you checking for in use?
04:03 < michaelrose> lsof and fuser
04:04 < michaelrose> example /var/lib/nfs/rpc_pipefs/gssd/clntXX/gssd
04:07 < michaelrose> nevermind I could rename var it seems then mount the new dataset with all the data from var
04:08 < michaelrose> presumably after reboot I'll be able to remove the old junk
04:08 < phogg> michaelrose: swapping out /var on a running system is tricky
04:08 < phogg> you will run in to many weird reactions. I recommend doing this only from runlevel 1
04:08 < kristina> systemd builds with musl now days doens't it?
04:09 < notmike> F
04:10 < phogg> kristina: in the first place, who cares? In the second place, you can easily ask Google about that.
04:12 < kristina> i care and internet is vague on it, it used to not build and there were some patches that may have or may not have been applied or w/e, there's no definitive answer from a quick search, the best i could come up with is "probably".
04:13 < phogg> I believe In heard it had been made to work, but I can't say whether it was patched into mainline. I don't track systemd development. Maybe the best way to know is to try such a build.
04:13 < phogg> s/In heard/I heard/
04:16 < kristina> like annoyingly CrOS libc and bionic won't ever support systemd or half of the software linux uses even, since they're BSDlike.
04:18 < kristina> that said i think that's one of the things that makes them great, they're not abstract or cross platform, they're tied to the platform very closely.
04:19 < phogg> at least part of the patch series from 2014 was merged
04:23 < kristina> i want something like libsystem.so instead of all that libatomic/libgcc_s/(shim)pthreads/ld.so/libc/dl/rt mess they made with glibc.
04:24 < kristina> the core platform library is a weird thing. there's no need to split it to make things more annoying.
04:26 < kristina> -pthreads is a dumb idea like what's the point, the heavy lifting is not done bu libpthread, but by glibc. same for libdl. or librt.
04:27 < phogg> kristina: on the other hand there's little reason not to split it up.
04:27 < kristina> phogg: they're shims that expose glibc functionality.
04:27 < revel> I think that's a "POSIX dictates you have to have -lm and -ldl and -pthread and..."
04:27 < revel> thing.
04:27 < phogg> the price we pay for standards!
04:27 < phogg> and not the worst price, either
04:28 < revel> Going off of https://www.musl-libc.org/faq.html
04:28 < phogg> kristina: still, if you want to make a distribution that ships them all as a single .so that would be interesting. I would subscribe to your newsletter.
04:29 < kristina> why not just expose it anyway and if you don't use it it won't be imported. and if you load ld.so you load glibc since ld.so/glibc are basically inseparable because part of libc is in the linker.
04:29 < kristina> part of linker is in libc.
04:29 < my_mind> i'm trying to install vmware 14 on a q6600 machine, but it won't let me. I found out vmware supports CPUs 2011 and newer
04:30 < my_mind> is there a workaroud?
04:30 < bls> "it won't let me" isn't much for us to go on
04:30 < phogg> kristina: btw https://github.com/matijaskala/musl-patches looks like patches to musl to add the stuff to the libc that systemd refused to patch in on their side.
04:30 < bls> did you say please?
04:30 < my_mind> https://hastebin.com/raw/bidekekigu
04:30 < phogg> ugh, why is hastebin so popular? Is it the catchy name?
04:30 < my_mind> it's lightweight
04:31 < my_mind> to the point
04:31 < phogg> my_mind: don't quit your day job, you'll never make it as a comedian
04:31 < kristina> iirc it starts with a minimal libc that's statically linked and then loads the real libc, calls some weird apis in it where they exchange vtables and patch each other, if pthreads wasn't specified ld.so will do one thing if not it will do another.
04:32 < bls> kind of fits with shared attitudes of Drepper and Peottering that there's only one way to do things and no, we won't support or allow to support things we don't like
04:32 < triceratux> yep its not letting you alright
04:32 < my_mind> how do i make my computer run vmware player 14?
04:33 < my_mind> if I really can't do it, then what's stopping me from installing an older version? what's the drawback of older vmware player 12 or 13 ?
04:33 < phogg> bls: drepper isn't in charge any more. Time to change things?
04:33 < lnnb> my_mind: workaround, use qemu instead!
04:33 < phogg> my_mind: doesn't have the latest whiz-bang features
04:33 < kristina> phogg: yeah i'm considering fusing compiler-rt, heavily patched musl, libc++ libcxxabi and the dylinker into one blob, and loading libsystemd (i have a port of it called nanosd that's more sane) from a section because it's lgpl ugh.
04:34 < my_mind> is qemu type 1 or 2?
04:34 < kristina> either.
04:34 < phogg> kristina: it all sounds good until you get to libsystemd. What do you need out of it?
04:34 < lnnb> you have to say --enable-kvm  for whichever is teh fake VM type
04:36 < kristina> phogg: sd-bus and sd-daemon seem to be the absolute minimum for a desktop, sd-event is optional but it's a lightweight thing that's nice to have in the platform lib.
04:36 < my_mind> i want to a couple of windows VMs to test out samba and linux based domains
04:36 < kristina> for a server you only need sd-daemon.
04:36 < phogg> kristina: meh, I don't know that you need those for a nice GUI experience. But I don't care much for "desktops" anyway. I suppose sd-event has its uses, though.
04:39 < kristina> phogg: for example, nanosd.bin has sd-daemon and sd-event and some useful shared apis exposed. it's a very stripped down libsystemd designed for getting around lgpl and well, doing some stuff, i mean you may also want journal stuff since you need syslog facilities in the platform lib.
04:40 < kristina> the point is, like userspace is going to have it loaded regardless, it's in one object it can benefit from IPO and all sorts of crap.
04:40 < autopsy> Crap really?
04:41 < phogg> kristina: of all of lennart's stuff I dislike journald the most. I think a well configured rsyslogd gives you everything you could want.
04:41 < kristina> i mean its like a 300kb lib.
04:42 < my_mind> i can't use qemu yet, i'm not a linux expert at troubleshooting, i prefer something like vmware, i just need to know what version would work with Q6600 cpu, and will a older verion of vmware hinder performance and security,,,
04:43 < autopsy> my_mind, no run an older VMware.
04:43 < kristina> this assumes that you're running a systemd-based system at least the pid1 part. rest doesn't matter. musl needs proper bcopy/memmove/strcpy/strlen.
04:43 < phogg> my_mind: with virt-manager you can use qemu and still set things up by pointing and clicking. Isn't that what you really want out of vmware?
04:44 < autopsy> my_mind, try VirtualBox
04:44 < my_mind> phogg: don't get me wrong, i love working with linux cli
04:45 < my_mind> i work with virtualbox everyday, it's slow
04:45 < kristina> but ideally libSystem.C.so is some shared object you would have that would basically provide a platform library, and everything would link against it implicitly or explicitly.
04:46 < kristina> and likely the main part of the linker.
04:47 < autopsy> my_mind, you're on an old CPU what do you expect?
04:47 < my_mind> :(
04:48 < my_mind> i know its old
04:48 < my_mind> I'll get a poweredge T30 soon
04:48 < autopsy> my_mind, when?
04:48 < kristina> unless you're doing a static build disabling pthreads is idiotic because well, they're still in the shared object, whether you use them or not. don't forget that all these shims are ELF SOs so they cannot have IPO run over the whole thing because of this brain damaged design.
04:48 < my_mind> autopsy: within the next 3 months
04:49 < autopsy> my_mind, wait to run the VMware the 3 months.
04:50 < autopsy> my_mind, why don't you build a PC using Ebay I know you can get a 4 processor ASUS K8N for like $50 on Ebay.
04:50 < kristina> now recently, LLVM/Clang/LLD have been getting beyond amazing at IPO.
04:51 < my_mind> K8N supports 4 cpus?
04:51 < my_mind> i only see 2 sockets
04:52 < my_mind> oops, sorry, i'm getting out of topic
04:53 < kristina> also dl_load_object or something along the lines needs to be a documented and exposed API.
04:53 < autopsy> my_mind, it supports dual core CPU dies.
04:54 < kristina> instead of linking yourself and then patching glibc's structs in memory.
04:54 < autopsy> my_mind, 2 sockets 4 cores.
04:54 < my_mind> i'd rather go with a T30, i've had my eye on it for months
04:54 < kristina> and you have to find that __get_tls shit.
04:55 < autopsy> my_mind, what is the T30? What is it all about?
04:55 < my_mind> Xeon E3-1225 v5 3.3GHz
04:55 < kristina> so you end up hooking glibc in a million places so you can steal things like addresses of those "hidden" symbols.
04:56 < kristina> well not hooking but interposing.
04:56 < kristina> which is idiotic.
04:56 < my_mind> kristina: i have to ask, who are you talking to? i've been wondering this since i logged in
04:56 < kristina> no one.
04:57 < my_mind> ok cool good to know.
04:57 < my_mind> now that's out of my system. i feel at ease
04:58 < stevendale> Looks like my Intel GMA 900 handles glxgears at 59 fps
04:58 < autopsy> I'm reading what he's talking about but I like systemd so meh.
04:59 < autopsy> stevendale, thats still good.
04:59 < kristina> so musl: horribly hides linux specific functionality, has an awful builtin memory allocator, no attempt at integrating high performance arch specific copies of string.h routines, etc.
04:59 < lnnb> stevendale: turn off vsync
04:59 < lnnb> i can run glxgears in software at > 500FPS
04:59 < stevendale> How would I turn off vsync
05:00 < stevendale> o/
05:01 < autopsy> my_mind, Xeon is pretty good. How much is it?
05:01 < lnnb> i lied 471 FPS
05:02 < my_mind> autopsy: $418
05:02 < kristina> which is idiotic. considering google, apple, and some other vendors provide extremely well written routines for something like memmove, which are under free licenses.
05:02 < my_mind> https://www.amazon.com/Newest-Dell-PowerEdge-Server-E3-1225/dp/B0722J7NNX
05:02 < autopsy> My GPU does 60 FPS too.
05:02 < lnnb> o wait it jumped to 800 the second run
05:02 < kristina> or glibc or linaro.
05:03 < lnnb> stevendale: i guess it would be a setting in your GL driver, i really don't know  tbh
05:03 < Sapphirus> glxgears isn't accurately representative of your GPU's performance.
05:03 < lnnb> graphics are a cluster*
05:03 < lnnb> the hw accelerated kind at least
05:04 < autopsy> my_mind, $419 dollars not bad.
05:04 < autopsy> my_mind, run an older VMware for 3 months.
05:05 < kristina> is this for a home server or soemthing?
05:05 < autopsy> It's for experimenting with Samba and Windows domains.
05:06 < triceratux> kristina: what would you consider a sane distro these days for working productively with linux ? is ubuntu too good to be true & should be avoided at all costs ?
05:06 < my_mind> home testing server, which will be a small office server
05:06 < twainwek> if you want to get shit done, i've found opensuse to be the go to distro
05:06 < autopsy> my_mind, the DVDRW is a good choice to have in it.
05:07 < autopsy> triceratux, Fedora is mostly up to date.
05:07 < my_mind> autopsy: a joke? lol
05:07 < autopsy> my_mind, nah you always gotta have a RO medium to boot Live into.
05:08 < twainwek> i gave up on fedora since between their releases stupid thing would brake, like click and drag would 50% of the time not work
05:08 < my_mind> yeah i know, not all computers support booting from usb or network boot
05:08 < twainwek> break*
05:08 < autopsy> my_mind, The DVDRW is probably bootable.
05:08 < my_mind> most definetly
05:09 < autopsy> my_mind, why do you nlot get it now? Instead of in 3 months?
05:09 < my_mind> because o
05:09 < kristina> i use headless ubuntu on a build server just because they have more recent packages for some stuff, on workstation i use windows or test shit on VMs or on actual hardware w/ serial etc.
05:09 < my_mind> because i'm a little... cheap
05:09 < autopsy> triceratux, don't use Ubuntu.
05:10 < my_mind> when I get that T30, i'll install debian
05:10 < autopsy> my_mind, $419 ain't exactly cheap man.
05:10 < autopsy> That is 1/4 of what the PC costs in 1995 $2000
05:10 < my_mind> for a server yes it is
05:11 < my_mind> i was surprized when i found out Dell offered T30 for 400 bucks new with warranty
05:11 < my_mind> i was like whoaaaa
05:11 < autopsy> Well server smerver I can get 4 cores in an ASUS K8N to work at 3.4GHz a piece. Same thing as the server T30
05:11 < lnnb> autopsy: adjusted for inflation, that should be like $3500
05:11 < kristina> ubuntu's fine, imo. you bootstrap the toolchain (build clang then build everything again now without deps on libstdc++ etc), 2 stage bootstrap from a sane compiler. now no one gives a rat's ass what you run.
05:12 < autopsy> lnnb, yeah thank you.
05:12 < autopsy> Rats ass yay!
05:12 < triceratux> twainwek, autopsy, kristina: preciate it & youre right. i run all that stuff
05:12 < autopsy> I love rats ass.
05:12 < kristina> triceratux: what stuff?
05:13 < revel> autopsy: Should I tell Theo?
05:13 < autopsy> All that stuff man don't you dig?
05:13 < autopsy> revel, no please.
05:13 < revel> But how can love possibly blossom between you two?
05:14 < autopsy> revel, I am not sure anymore.
05:14 < triceratux> kristina: fedora, tumbleweed, & the most stripped down version of ubuntu i can tolerate. its just a good source of the software despite their spyware shenanegans
05:14 < kristina> llvm doesn't really care whether it's ubuntu or debian. does it run dropbear, does it have sane network configuration shit, does it have logind, does it have a toolchain that's not ancient.
05:14 < kristina> i don't use it as a desktop.
05:14 < my_mind> autopsy: good talk man, see ya
05:14 < autopsy> Yeah don't.
05:15 < autopsy> Yeah don't.see ya.
05:15 < my_mind> lol k
05:15 < kristina> well i'm not enough of a mascochist to use linux on a workstation.
05:15 < triceratux> kristina: i ran voidlinux for a long time & was really rooting for it. but it was clear there was something wrong
05:16 < autopsy> triceratux, Fedora has 6 month release cycles.
05:17 < triceratux> autopsy: yep. i keep up with them & have f28 xfce installed. theres no more korora either which simplifies things
05:17 < autopsy> triceratux, voidlinux korora those I've never heard of.
05:18 < kristina> windows or osx or cros for workstation, pick your poison, at least you'll have working drivers, and something to run chromium, a terminal and a text editor/ide, and a video player.
05:19 < lnnb> why would i run proprietary OS when all of that works fine on GNU/Linux
05:19 < autopsy> kristina, he could use Emacs as his OS.
05:19 < twainwek> windows ... terminal
05:19 < twainwek> what
05:19 < kristina> lxss.
05:19 < twainwek> only way to make windows semi usable is to effectively install another OS in it
05:19 < kristina> yes that's literally what lxss is, remember that NT is a microkernel and MS added another subsystem to it.
05:20 < triceratux> autopsy: void was a minimalist musl liveiso, but its maintainer disappeared. korora was a media-enabled fedora respin that used to come out just hours after each release. they just got bored & gave up
05:20 < twainwek> i'd rather have a full working OS than try to trick my OS into doing what it is that i really want it to do
05:21  * triceratux multiboots rather than distrohopps
05:21 < kristina> it's not tricking, lxss is a subsystem, win32k is a subsystem.
05:21 < autopsy> triceratux, oh ok. Sounds like Slackware.
05:21 < twainwek> so i go by the "anything but windows" philosophy
05:26 < kristina> it's not actually that bad: http://crna.cc/i_cant_believe_its_not_lunix.PNG
05:26 < lnnb> kristina: have you been contracted to come in here and spout a bunch of anti Linux nonsense?
05:27 < kristina> that's running a linux userland.
05:28 < kristina> in parallel to win32k userland, same scheduler, etc.
05:29 < Sapphirus> WSL is terrible imho. All the caveats of Windows running alongside a gimped userland.
05:29 < Sapphirus> I'd rather just run Windows in a virtual machine atop of Linux.
05:31 < kristina> oh wow i'm using mosh-client launched from zsh5 to log into my vps and attach to my session running irssi. and not really. windows redstone 5 is okay.
05:31 < Sapphirus> I mean, this is ##linux. I'd wager that Windows 10 has gotten progressively worse.
05:31 < twainwek> took them 30 years to upgrade their crappy cmd
05:32 < Sapphirus> Everything from UI/UX blunders, forced updated, to god awful telemetry.
05:32 < kristina> lxss does what it's meant to do quite well. and again, lxss is linux.
05:32 < lnnb> not really, but you are entitled to your own beliefs
05:32 < autopsy> Windows is off-topic here.
05:33 < autopsy> Slackware is on topic.
05:33 < Sapphirus> lnnb, sounds like something a Windows shill would say, to be frank.
05:33 < Sapphirus> Let's talk about some more Linux and less Windows.
05:33 < eletious> wait - are we calling is WSL or lxss?  Is there a difference?
05:34 < lnnb> Sapphirus: i mean they keep coming in here with the same old tired script, if they are contracted i think someone deserves their money back
05:34 < norkle> yeee!
05:34 < Sapphirus> lnnb, Where do you want to start then? I could backup those points very easily.
05:34 < lnnb> or at least find some new minor edge case flaws to bitch about
05:35 < twainwek> eletious: they'll probably change it to WindowsLinuxInTheCloudAzure soon
05:35 < lnnb> start what?
05:35 < Sapphirus> But, this is ##linux and not ##windows.
05:35 < lnnb> what are we starting?
05:35 < twainwek> i left out a 360 in there
05:35 < lnnb> you guys want to start a war?
05:36 < autopsy> Lets go to war.
05:36 < autopsy> I love war.
05:37 < eletious> ^^
05:37 < kristina> i mean you could argue that it has to run genuine (tm) linux kernel to be linux but http://crna.cc/this_is_fine.png
05:38 < kristina> CrOS/Android are disqualified too i guess since they run "Linux" on a BSD userland with custom facilities and custom IPC like mojo or binder.
05:38 < Sapphirus> https://img.plus/d9a6a
05:39 < Sapphirus> Great example of one of the many Win10 UI blunders.
05:39 < twainwek> i am not clicking these shady links
05:39 < eletious> I feel like CrOS counts as a linux distro though.
05:39 < eletious> a crippled one, but still...
05:39 < lnnb> seriously this is entry level trolling+phising going on here
05:39 < kristina> the links are fine.
05:39 < lnnb> yall need to work on your game
05:39 < Sapphirus> Not even MacOS dares to shove a touch-oriented UI at desktop users.
05:39 < kristina> you can click them they're screenshots.
05:39 < Sapphirus> They aren't that dumb.
05:40 < Sapphirus> Apple at least has some competent people left.
05:40 < autopsy> Apple is off topic too.
05:40 < twainwek> Windows and UI? i believe it was up until msvc 2017 that they finally figured out that people are not using 480x320 resolution anymore
05:40 < strive> lol
05:41 < twainwek> or studio or whatever the heck they call that crap
05:41 < autopsy> Studio? I'm lost.
05:41 < twainwek> msvisualstudio
05:42 < twainwek> their msvc compiler is garbage too by the way... first time i've experienced a compiler crash compiling a simple code
05:42 < kristina> it's not touch oriented, i have a terminal open, running mosh-client which is like ssh over udp a ton of chromium windows, and a few other terminals. blurry line.
05:42 < kristina> gcc has bugs too, i even filed a few.
05:43 < Sapphirus> kristina, look at how tall and large the title bar is.
05:43 < twainwek> bugs is fine. but when a compiler you commercially sell consistently crashes compiling trivial code, its not fine
05:43 < kristina> oh and wow yesterday i had a clang bug: http://crna.cc/lol_yes_generics.cpp
05:44 < Sapphirus> Not to mention, how there's literally zero differentiation between different elements of the UI.
05:44 < Sapphirus> Microsoft can't even manage to do a flat ui right
05:44 < kristina> i have all the ui pretty stuff off and a single color background.
05:44 < Sapphirus> Apple, Google, and even third-party GTK/QT themers on deviantart, can do a better job than whoever Microsoft is paying.
05:45 < stevendale> Ubuntu is a lot more user friendly than Arch or Debian
05:45 < twainwek> i feel sick talking about windows. im out
05:46 < kristina> we're talking about NT/LXSS technically.
05:46 < twainwek> ill just mention that i'd rather work on a toaster running templeos than ever touch windows again
05:46 < kristina> you're the ones shouting windows.
05:47 < autopsy> kristina, just quit.
05:47 < lnnb> because it is windows how can you have such in depth knowledge of linux edge case issues and not realize that?
05:47 < Sapphirus> twainwek, I second that sentiment.
05:49 < kristina> lnnb: because i've done a ton of compiler work on llvm and clang, i developed linux drivers, had odd hobby projects like porting mach over to linux before darbat, i brought up darwin (ios kernel) on several devices, i did that oss firmware for rpi thing, i'm not an idiot.
05:49 < kristina> so i would guess experience?
05:50 < Sapphirus> Yet you tolerate Windows 10 and boast about WSL in a ##linux channel.
05:50 < lnnb> that doesn't explain how you refuse to acknowledge that nt linux-like API is ****NOT***** Linux
05:50 < lnnb> you're just getting sloppy now, before you had good arguments
05:50 < revel> Isn't Mach a microkernel? How would porting that to Linux work?
05:50 < lnnb> now you seem too desperate, and for that reason, i'm out
05:51 < kristina> i said it runs linux userland, NT+LXSS is not Linux.
05:51 < Sapphirus> Then it's off-topic.
05:51 < kristina> Linux userland is offtopic?
05:51 < autopsy> kristina, it is off topic already.
05:52 < Sapphirus> kristina, when you bring Windows into the mix, it sur eis.
05:52 < kristina> I'm not the one who keeps bringing it up, I spoke about Linux exclusively, just casually mentioning LXSS on a workstation got everyone so upset over a personal choice?
05:53 < Sapphirus> Some people here, including myself, have legitimate grievances with Windows and it's a touchy subject.
05:53 < revel> kristina: Isn't Mach a microkernel? How would porting that to Linux work?
05:53 < Sapphirus> Easily can derail into off-topic discussion with ease.
05:54 < kristina> revel: Mach IPC/Tasks/Thread/Port rights/Message passing/Kqueue.
05:54 < stevendale> Can we just stop the Windows discussion now? If you want to continue you can get out and go to ##windows
05:54 < kristina> VM Ports (aka Task ports) and sending VM regions over Mach IPC.
05:55 < kristina> https://github.com/christinaa/kernel_diff/blob/master/kernel.diff
05:55 < kristina> ancient project of mine.
05:57 < kristina> it only targeted armv7 at the time,
05:57 < kristina> hence the weird names like Return_to_user_from_swi();
05:59 < kristina> just an old project for fun, it was based on linux 2.6,something.
05:59 < kristina> the kernel was, the userland was entirely Darwin.
06:00 < kristina> and the libc i put together with it, a lot of own code, a lot from bsd.
06:00 < kristina> https://github.com/christinaa/libSystem_and_linker/blob/master/core/lnk_macho.cpp
06:01 < kristina> I didn't ever finish this: https://github.com/christinaa/libSystem_and_linker/blob/master/core/lnk_elf.cpp
06:05 < kristina> i had a fairly interesting idea, back then regarding a unified linker/platform lib, you can read on from down there, some comments talk about it: https://github.com/christinaa/libSystem_and_linker/blob/master/core/lnk.cpp#L513
06:08 < kristina> oh yeah OSInitializeMainRuntime would kick off the mach/mig stuff from userland's perspective as well as set up the magazine allocator for zone support.
06:11 < revel> What an interesting bunch of stuff I don't understand :D
06:11 < autopsy> Me too.
06:11 < autopsy> Using fancy words like unified libsystem.so
06:12 < konimex> what's libsystem.so anyway
06:13 < bls> https://github.com/christinaa/libSystem_and_linker
06:13 < eletious> oooh, magenta work
06:14 < kristina> but anyway, regardless, it's very childish to cherry pick shit becuase a word you don't got mentioned esp because i do use linux on embedded and on servers, just not on workstations. even if i spent hours doing shit like fixing shitty HDMI drivers for TI OMAP5 or working on LLVM improvements for ELF targets or any of that stuff. what matters is what i use on the client machine i used to interact with
06:14 < kristina> other shit.
06:14 < ayecee> mind the language please
06:15 < autopsy> Yeah shit is a bad word. Please.
06:15 < [777]> Let's not repeat history but improve it.
06:16 < autopsy> [777] I don't believe history can be improved can it?
06:16 < ayecee> not with that attitude
06:17 < autopsy> You snap your fingers you snap your neck!
06:18 < kristina> and everyone in open source crowd wanted a oss rpi firmware, i mean i got arm to boot, which was like the only non-blob firmware to do that. but no one obviously wants to contribute or help, they just wait until it's all shiny for them. (and yeah for reference):
06:18 < kristina> https://github.com/christinaa/rpi-open-firmware/blob/master/drivers/BCM2708ArmControl.cc
06:18 < kristina> hardware bugs on undocumented hardware suck.
06:21 < kristina> well a lot of work went into writing a semi coherent CPR driver for power management and limited work on clock management but i gave up because i've had 2 people show interest in helping me and contributing out of 10s of thousands who cried over blobs on rpi.
06:21 < kristina> oss community in a nutshell.
06:23 < systemhalted> One thing about OSS. Most of the OSS, is developed by mostly one or at most three or four developer unless it is backed by a company. So, whatever you are doing is great.
06:24 < jak2000> how to fix: https://imgur.com/a/8498oN5   error installation?
06:25 < lordvadr> woot
06:25 < autopsy> jak2000, you have to select and install software from the menu.
06:28 < jak2000> selected ssh server and utilities
06:28 < jak2000> now cant select anything
06:28 < jak2000> how to start again?
06:29 < kristina> well my $dayjob is at a large software (incl OS) vendor that does a lot of proprietary stuff so yeah i mean it's a very different culture and i didn't do any OSS work at all until around 2011, only proprietary. as far as OSS goes, if i'm intersted in something i'll do it for fun like the challenge or REing SDRAM controller init/calibration on RPi but once i reach the goal unless someone else is working
06:29 < kristina> with me as far as i'm concerned it's done.
06:32 < kristina> i mean personally i see it as OSS has to be collaborative, or you have one person doing what they want and then it turning into abandonware, just another research thing.
06:33 < revel> kristina: So, wait, is it Kristina or Eric?
06:34 < revel> Oh, nvm
06:34 < autopsy> revel who is Eric?
06:34 < revel> Misunderstood some commits.
06:34 < kristina> Anholt did 3D usermode stuff, it's the GPU, I worked on the VPU which was a separate thing that was the main part of the SoC.
06:50 < Nautilus> small question, is there a way to tell "ls" to list the total file count in the folder?
06:51 < [R]> ls;ls|wc -l
06:52 < kurahaupo> find -printf x | wc -c
06:52 < iflema> [R]: goodone
06:53 < kurahaupo> That assumes you don't have filenames with newlines in them
06:53  * iflema starsr
06:53 < ayecee> doesn't seem like that would matter for that command
06:53 < ayecee> you're not even using the filename
06:54 < ayecee> oh, [R]'s command
06:54 < jim> Nautilus, count? you mean the number of files?
06:54 < ayecee> what kind of jerk puts newlines in filenames anyways
06:54 < Nautilus> yea, trying some of the suggestions now
06:55 < [R]> the same kind of jerk that puts a \1 in one?
06:55 < ayecee> it's you, isn't it
06:55 < [R]> haha
06:55 < Nautilus> I think my answer is "ls|wc -l", checking
06:55 < Nautilus> whats wc?
06:55 < [R]> man wc
06:55 < ayecee> water closet
06:55 < Nautilus> ac
06:55 < jim> is that a variation on "yer a \1"?
06:55 < ayecee> also word count
06:55 < kurahaupo> Worst case
06:55 < Nautilus> ahhh ty
06:56 < kerframil> Nautilus: are you only intending to count regular files - as opposed to directories, symlinks and the like?
06:56 < Nautilus> so no native "file count" for ls?
06:56 < [R]> Nautilus: man ls
06:56 < ayecee> maybe there is, but who has time to read manpages when you can oneline it.
06:56 < Nautilus> kerframil: not being that particular
06:57 < Nautilus> [R] I did
06:57 < Nautilus> not that I'm good at it
06:57 < [R]> Nautilus: then i guess you have your answer
06:57 < Nautilus> no, that's why I asked
06:57 < [R]> you think they're hiding it from you?
06:57 < Nautilus> i think theres a lot to read and missed it?
06:58 < jim> this is becoming an argument... please drop that aspect
06:58 < [R]> so read it again
06:58 < autopsy> Nautilus, you can search the manpager with /word count
06:58 < jim> either help him or don't comment.
06:58 < revel> `ls -A1q | wc -l` could work, maybe. Though this is probably GNU coreutils specific, so, might as well use something with `find`
06:59 < Nautilus> the suggestion of "ls|wc -l" is doing what I wanted, just trying to understand more
06:59 < revel> `wc -l` just counts lines of output.
06:59 < jim> but DO NOT make this a "shame him into reading the man page"
06:59 < Nautilus> I mean, I just think that a command like ls, a thing that lists files, could itself give me a total of what it listed.
06:59 < jim> there would be a year-long ban waiting
06:59 < autopsy> Nautilus, ls -A1q | wc -l
07:00 < ayecee> jim: can't be. he already did read it.
07:00 < jim> but that's what it was, even after that fact
07:00 < Nautilus> autopsy: trying it
07:00 < ayecee> mmkay
07:01 < autopsy> He would like to read the manual page on what is -A -1 and -q
07:01 < kerframil> Nautilus: following on from kurahaupo's answer, this is a safe equivalent to ls | wc -l: find -mindepth 1 -maxdepth 1 -printf x | wc -c
07:02 < kerframil> Nautilus: counting from ls may not give the correct answer (depending on the filenames printed and the quoting style that ls is configured to use)
07:02 < jim> I'm trying to encourage people into the man pages... please don't let that cause you to force the issue
07:02 < revel> ^ is probably better.
07:02 < revel> kerframil: -q drops formatting funkiness.
07:02 < jim> you might be able to get an accurate count with find and exec
07:02 < ayecee> encouragement with a year-long ban stick
07:03 < Nautilus> holy smokes, way more than I meant to get the channel talking about
07:03 < Nautilus> IOW, ls can't tell me how many files it just listed
07:03 < pingfloyd> Don't parse ls
07:03 < revel> This has just made me realize that I have no idea what ls's "total" at the top is supposed to be...
07:04 < Nautilus> revel: I was just wondering about that too
07:04 < bls> should it count only files? or files and directories? together or separate? what about symlinks? or when you recurse?
07:04 < autopsy> Total comes with ls -l only and doesn't list hidden dot files.
07:04 < Nautilus> bls: it seems to NONE of that
07:04 < bls> oh yeah, dotfiles too.
07:04 < kerframil> revel: yes, it should be ok with -q. I'm just generally reluctant to promote ls parsing.
07:05 < revel> Yeah, I agree with you there.
07:05 < Nautilus> autopsy: I see no total with ls -l
07:05 < bls> exactly, because you're expected to make that decision so they don't have to write all that extra code that most people aren't going to bother to use
07:05 < jim> I'm trying to make it work with find, not having luch
07:06 < autopsy> Nautilus, ls -l | more
07:06 < bls> use a tool to generate a list of the files you and combine it with a tool to count things
07:06 < kerframil> jim: what's the problem?
07:06 < jim> I tried this:
07:06 < Nautilus> autopsy: ths list of files fits in the window
07:06 < autopsy> Nautilus, at the top it says Total:
07:07 < jim> find . -exec x=$(expr $x + 1) \;
07:07 < Nautilus> autopsy: yes, it says 2632 when I have manually counted 50 or less
07:07 < bls> jim: because the $() is getting evaluated once, not repeatedly
07:08 < jim> before I did that, I ran: mkdir test ; cd test ; touch a b c d e ; x=0
07:08 < autopsy> Nautilus, yeah ok. I got you.
07:08 < kerframil> jim: also, to use shell code with -exec, you need to actually execute a shell. it's only evaluated once there, before the shell launches find.
07:08 < jim> so when I ran the find, I got: find: ‘x=1’: No such file or directory
07:08 < jim> five times
07:09 < kerframil> yeah, because you don't have a command named x=1
07:09 < kerframil> to execute
07:09 < Nautilus> is there something I should be using besides ls or find? List files, tell me how many that was
07:10 < jim> so counting files accurately is -hard-!!
07:10 < kerframil> jim: this can't actually be done with -exec in that fashion. I mean, you could do this: find ... -exec sh -c 'for f; do n=$((n+1)); done; echo $n' sh {} +
07:10 < kerframil> jim: but that would still be horribly wrong if find traverses more files than ARG_MAX
07:11 < Nautilus> in windows if I type "dir" I get:
07:11 < kerframil> jim: easiest thing to do is to use -printf x and wc -c as mentioned earlier. only problem is, it's not portable.
07:11 < Nautilus> 28 File(s)     19,621,975 bytes         10 Dir(s)  682,890,436,608 bytes free
07:11 < jim> yeah, I just didn't get it exactly right, and there's a few sets of lots of moving parts going here
07:11 < kerframil> jim: it could be posix-ified with: find ... -exec sh -c 'for f; do printf .; done' sh {} + | wc -c
07:12 < kerframil> jim: not pretty, but it works
07:12 < jim> probably a lot better in scsh
07:13 < bls> I've just got a dumb little 160 line C program I use to count files
07:13 < Nautilus> because ls can't count :(
07:13 < jim> that's clearly the case
07:14 < jim> but we should be able to get find to execute one line per item
07:14 < bls> ls is short for list, not list (and count and delete and sent email etc etc)
07:15 < kerframil> Nautilus: just put one of the proposed solutions into a shell function and call it a day. as you say, ls cannot generally count.
07:15 < jim> so we found a hole :P
07:15 < Nautilus> yep, looks like "ls|wc -l" is what I'm gonna use.
07:16 < Nautilus> fits for this use case anyways
07:16 < Nautilus> I thank all for the thoughts
07:17 < kerframil> Nautilus: that's wrong, though. if you insist on using (GNU) ls, at least use -q, as revel mentioned earlier.
07:18 < jim> Nautilus, just know that that won't work in all cases (filenames can have chars they probly shouldn'[t be allowed to have, like \n... space is iffy but permissible
07:18 < Nautilus> now I need to figure out how to grab all waybackmachine archive dates/time stamps for archives since 0161119014201
07:18 < autopsy> Lots of filenames have spaces.
07:18 < Nautilus> nods, just not these
07:19 < autopsy> Nautilus, ls -A1q is what you want seriously. That to wc -l
07:20 < jim> once you do ls | anything, you should realize that after the |, the system can't know they're files, and if there's a space, it gets counted as 2 when it's really not
07:21 < jim> if your use case guarantees there are no spaces or newlines in the filenames, probably you have something that will work as long as that guarantee holds up
07:21 < Nautilus> -A1q is just a bare list
07:21 < autopsy> Nautilus, yes | wc -l
07:21 < Nautilus> oh then pipe to wc?
07:21 < autopsy> Yes.
07:21 < Nautilus> yea
07:21 < kerframil> jim: that's not quite right. the default format is one entry per line. POSIX stipulates that the format is implementation-defined where stdout is attached to a tty. compare `ls` to `ls | less` for example.
07:21 < revel> jim: Wait, why would it not work with files with spaces? They're still on the same line.
07:22 < kerframil> jim: that said, vertical whitespace is still a problem and parsing ls is a fool's errand in general
07:22 < jim> kerframil, yes, and those are details that cause loopholes
07:22 < autopsy> He is not parsing ls output for anything except a listing count.
07:22 < bls> ls | wc -w
07:22 < kerframil> jim: yup
07:23 < Nautilus> autopsy: ok, going that way, thanks
07:23 < revel> wc -w is where you'd probably get problems with filenames with spaces.
07:23 < jim> now...
07:23 < kerframil> Nautilus: are you doing anything with the list of entries that you are counting, beyond actually counting them?
07:23 < autopsy> Nautilus, the real dot files are included in that -A1q for ls.
07:23 < jim> that we're here...
07:23 < revel> (-A also excludes . and .., which -a includes)
07:24 < [R]> autopsy: what about the fake dot files?
07:24 < autopsy> [R] yes fake dot files too.
07:24 < Nautilus> kerframil: trying to download files from archive.org, different archive dates, and just want a file count
07:24 < autopsy> Thank you.
07:24 < jim> this probably took longer and everything, but there wasn't a conduit for shame. that's how it should be
07:26 < ayecee> shame is a mechanism to correct behavior.
07:26 < jim> and it may have been how you came up
07:27 < jim> and how I came up
07:27 < jim> but I'm specifically not permitting it
07:27 < ayecee> also threats with a stick amirite
07:27 < notmike> jim: how does one center the mind on c after having been led astray by other languages like Python and Java
07:27 < Nautilus> i tend to think when a computer program can list files it could count them as well, shame is on ls (my opinion anyways)
07:28 < notmike> Nautilus: uh, do one thing well
07:28 < Nautilus> or not
07:28 < jim> notmike, C is just more detailed, and there are a lot of things you have to take care of yourself
07:29 < kerframil> Nautilus: while that's a fair criticism, even if it did, ls is not intended for machine parsing. there are other solutions that can be employed in the shell.
07:29 < notmike> jim: c++ seems to muddy the waters
07:29 < bls> for every feature other than listing files you think it absolutely should have, there are plenty more that people feel are pointless bloat
07:30 < jim> they're two totally differnt ways of thinking
07:30 < notmike> jim: why
07:30 < jim> well that's a long story
07:31 < pingfloyd> That total shown at the top of ls -l output represents total blocks allocated to files of the directory you're listing.
07:31 < jim> for one thing, c++ allows things that C does not
07:32 < pingfloyd> C++ is an abomination
07:32 < Nautilus> pingfloyd: ah, that makes a lot of sense.
07:32 < revel> Makes sense that a dir with just a socket file was marked as 0 then.
07:32 < jim> so if nation has abom, that's c++?
07:32 < pingfloyd> Sadly the man page doesn't tell you this,  but the info page for ls does.
07:33 < revel> GNU love those.
07:33 < sauvin> C supports anything, it's just that you probably don't want to do the work to make it do it.
07:33 < Nautilus> whats an info page? www docs?
07:33 < kerframil> this is to be expected for gnu software, unfortunately
07:33 < revel> No, `info ls`
07:33 < kerframil> they "generally abhor manpages"
07:33 < jim> cute_korean_girl, can we assist you debugging your connection?
07:33 < revel> They're like manpages, but worse.
07:34 < pingfloyd> Nautilus: they're like a book in electronic form
07:34 < pingfloyd> Like pdf before it existed
07:35 < pingfloyd> And not so crappy and proprietary
07:35 < autopsy> Nautilus, try info ls
07:35 < Nautilus> i did, looks a lot like man
07:35 < jim> revel, whether the format for info pages is better or worse doen't come into this part: I think the writing the gnu folks did in info pages, is easier to read
07:35 < bls> or info coreutils ls
07:35 < pingfloyd> Info coreutils ls
07:36 < autopsy> Nautilus, no they have different content than the man pages.
07:36 < Nautilus> understood
07:36 < pingfloyd> Info pages are more like a conventional book
07:37 < pingfloyd> They're designed to be printed too
07:38 < pingfloyd> To find anything in them,  use their index like a real book.
07:38 < kerframil> in general, one cannot master a given GNU utility by reading man pages alone. the bash man page is reasonably thorough, though.
07:38 < pingfloyd> Man pages were more meant as a quick reference
07:39 < jim> Nautilus, having done ALL this... I'd like to suggest you pick a different man page to read each day, because taken together they represent the "entire manual for YOUR installation of linux"
07:39 < pingfloyd> info pages are written in very old jargon
07:39 < Nautilus> jim: lolz
07:39 < notmike> sauvin: that's how I feel. C is super powerful. C++ repackages that into a high level language
07:39 < pingfloyd> They often refer to things in a strange manner,  because some of their concepts predate any jargon.
07:40 < notmike> Like std::string::toupper
07:40 < jim> Nautilus, reason for saying this, is when you do this, you'll get a feel for where things are in each man page, and also -how- the man page says things
07:40 < pingfloyd> C is the ultimate power for a high level language
07:41 < notmike> C is like unto Manna From Heaven.
07:41 < pingfloyd> Lower level than c and you're using assembler
07:41 < Nautilus> jim: ah, I get ya there, just that I'm nly single tasking, trying to grab a bunch of files from WaybackMachine. I dont do nix daily
07:41 < pingfloyd> No matter who you are,  you can benefit by knowing a little c.
07:41 < jim> pingfloyd, also, C has all the power of asm, with the expressiveness of asm... and the flexibility of... asm...
07:42  * Nautilus spent more than a decade with C. Like when you had to write your own SVGA library
07:42 < pingfloyd> asm is assembler
07:42 < notmike> C stood for Christopher, then Cambridge, then Combined
07:42 < pingfloyd> asm is a step up from machine code (in binary)
07:42 < jim> pingfloyd, to a degree, so is C :)
07:42 < Nautilus> 6502 asm, did plenty of that
07:43 < Nautilus> C on PIC devices too
07:43 < jim> I did forth on a 6502 machine
07:43 < sauvin> C is, in fact, assembler with no assembly required.
07:44 < sauvin> jim, which one? I think the one I had was from hesmon.
07:44 < Nautilus> C can inline ASM
07:45 < lnnb> if you're comparing it to ASM, all that added function call overhead is bloat city
07:45 < Nautilus> I still need to figure out all the waybackmachine archive timedate stamps so I can put them in my wget script and then count the results
07:45 < pingfloyd> lnnb: i know
07:45 < jim> sauvin, capn software figForth, plus we typed in an asm listing for 6502 forth into a atari 800 and I got it working and customized it
07:45 < notmike> Glad we all at least agree that slackware is master distro
07:46 < Nautilus> jim: neat
07:46 < jim> I wrote the compile words for an exception handling feature
07:46 < sauvin> One of the basic master baiters, maybe.
07:47 < jim> notmike, debian took some of its packaging ideas from slackware
07:49 < jim> back in a bit...
07:50 < notmike> Debian release cycle so long tho
07:51 < pingfloyd> Slackware is so old it lost relevance
07:52 < searedvandal> debian is just as old
07:52 < sauvin> Almost.
07:52 < searedvandal> a few months younger
07:52 < sudomoose_> Hello
07:52 < sauvin> And craploads better supported.
07:53 < Nautilus> dang, there's 50-some dates in 2017 WaybackMachine I want to try to grab from, but am not sure how to get the url without manual clicking
07:53 < sudomoose_> What are doing Nautilus if you don't mind me asking
07:53 < Nautilus> url with timedate stamp
07:54 < sudomoose_> Ah
07:55 < Nautilus> sudomoose_: trying to recreate this site: https://web.archive.org/web/20161122230201/http://iss.astroviewer.net:80/index.php
07:55 < thebigj> I am generating "dd" raw disk image of 16 GB SD card. Problem is the image is also 16 GB large.
07:55 < pingfloyd> But debian is maintained
07:55 < pingfloyd> Important things like security updates
07:55 < thebigj> What can be other options to generate "iso" is if it takes less space but expand while writing to SD card.
07:56 < thebigj> ?
07:56 < notmike> Either you can support slackware or you're using Windows.
07:56 < sudomoose_> Nautilus: ah, very cool
07:57 < RayTracer> thebigj: write the filesystem's remaining space with 0, then you can either create a sparse file with dd, or use compression on top which makes the 0-chunk small
07:57 < Nautilus> sudomoose_: That's why I'm trying :)   Problem is that WB only capstures images in the window at the time of crawl, and the iss's path over earth varies over time.
07:58 < Nautilus> sudomoose_: so I'm trying to crawl all of WB's crawls
07:58 < thebigj> RayTracer: Not getting the part of writing filesystem's remaining space with 0
07:58 < revel> I think Raspbian live-expands the ext4 root filesystem after you first boot it.
08:00 < RayTracer> thebigj: if there are large chunks of 0, it will compress better. Sparse files can also be used https://en.wikipedia.org/wiki/Sparse_file but they tend to be ephemeral, as you need to take special care to preserve sparseness on copying
08:01 < Triffid_Hunter> thebigj: you can use resize2fs or similar to shrink the filesystem to the minimum size that can hold all the data, then image the resulting shrunken image. after cloning, you can reinflate it to take remaining space
08:01 < RayTracer> thebigj: if you can do otherwise, like raspbian does it, you gain flexibility for smaller or larger cards
08:01 < ktokko> looking to buy a refurbished 14-15" screen size or less 64bit-usb 3.0+hdmi laptop that can handle most popular distro's with dual boot windows 7 pro, under $300 usd - recommendations to check? (storage capacity + battery life not as important)
08:01 < Triffid_Hunter> thebigj: probably vastly simpler to just compress though, run it through xzcat or gzip bzip2 or whatever
08:02 < ktokko> ram's also not too important ^^
08:03 < ktokko> and really more about finding a model-brand that has a few of the aforementioned specs to begin with
08:05 < RayTracer> ktokko: ##hardware maybe
08:07 < ktokko> thanks RayTracer
08:15 < snaphelp> SystemD is really interesting
08:16 < snaphelp> Especially things like systemd-nspawn
08:16 < nafg> Hi, I'm having trouble with a DNS cache issue. Is this the right place to discuss?
08:17 < autopsy> nafg, what is the trouble?
08:17 < ][_R_][> Try ##Networking (might be a single #)
08:19 < nafg> autopsy: I switched my webapp's nameservers back to google domains from cloudflare, but my computer is still resolving it to cloudflare. For a while it was working though.  Ubuntu 18.04 / systemd, on a mobile hotspot
08:19 < autopsy> nafg you need to set it in /etc/resolv.conf
08:19 < nafg> systemd-resolve --status shows the google DNS 8.8.8.8 and 4.4.4.4 as I configured in the mobile hotspot router, but funnily it's showing a 3rd entry, in IPv6 format
08:19 < nafg> and indeed it's resolving the site to an IPv6 address
08:20 < nafg> autopsy: "# This file is managed by man:systemd-resolved(8). Do not edit."
08:20 < Triffid_Hunter> nafg: may have to wait for intermediate caches to time out, the time is set by the fields in your DNS records' SOA
08:20 < nafg> It contains nameserver 127.0.0.1:53
08:20 < nafg> Triffid_Hunter: it's been a long time already
08:21 < nafg> Anyway the most of the internet sees the updated DNS AFAICT
08:21 < autopsy> nafg yeah but you can edit the file directly to set a nameserver to what you want using nameserver 8.8.8.8 if you will.
08:21 < Triffid_Hunter> nafg: longer than your server specified dns record timeout?
08:21 < nafg> e.g. stackdriver uptime check is telling me it's getting a 503
08:22 < nafg> Triffid_Hunter: most records I put in TTL of 2m ;) but really it's not a record problem, it's a nameserver problem
08:22 < paulcarroty> 8.8.8.8 sucks, 1.1.1.1 is too much faster
08:22 < nafg> that isn't the issue
08:22 < nafg> Partially I'm curious what the 3rd DNS entry is in systemd-resolve --status
08:22 < nafg> fe80::f8ec:d2ff:fedb:abc
08:22 < nafg> how would I find where it points to?
08:23 < nafg> ipv6 is chinese to me ;)
08:23 < Triffid_Hunter> nafg: fe80 is link local afaik
08:23 < paulcarroty> fe80 is a local address
08:23 < Triffid_Hunter> nafg: so probably your router
08:23 < nafg> aha
08:24 < paulcarroty> it can be any local device in your local subnet
08:24 < nafg> Strangely, when I first updated it to manual DNS it resolved the site correctly -- but then it stopped resolving itself (http://myhotspot/)
08:24 < Triffid_Hunter> nafg: or maybe even your local machine's link-local on the outbound interface, see ip addr list
08:24 < nafg> (I had to go to 192.168.128.1)
08:24 < nafg> later the dns name worked and the site started resolving to cloudflare again
08:25 < nafg> Triffid_Hunter: ip addr list doesn't have it (grepped)
08:25 < nafg> I'm guessing the router DNS is caching it
08:25 < nafg> any likelihood of there being a way to clear its DNS other than resetting it?
08:25 < Triffid_Hunter> hmm, do ipv6 addresses show up in arp -n? I'm not on a network with ipv6 here
08:26 < Triffid_Hunter> nafg: not really, cots routers make dns, dhcp pretty opaque
08:26 < nafg> no idea what that command does but the HWaddress column has ipv6
08:26 < nafg> ok I guess I'll try resetting it or something
08:26 < nafg> i should google first I guess
08:29 < nafg> hmm actually if I reset prob. have to reactivate :(
08:31 < pingfloyd> nafg: usually you want to turn down TTL as low as possible before you move
08:32 < pingfloyd> Like do it the day before.  Then when you move your domain,  the RRs will propagate fairly fast.
08:32 < nafg> pingfloyd: yeah I do. Like I said, it's the wrong nameserver, so record TTL irrelevant
08:32 < nafg> anyway I just changed the DNS in the control panel (duh, why did I forget about that)
08:32 < nafg> now ping gives the right ip
08:33 < pingfloyd> Should work for all nameservers
08:33 < nafg> the nameserver doesn't have a TTL
08:33 < nafg> NOW to figure out why haproxy is giving a 503!
08:36 < pingfloyd> Any working nameserver should resolve within that time
08:36 < nafg> that isn't the issue
08:39 < Pusteblume> Peace.
08:44 < quint> If I append extra data onto the end of an iso image (outside of the filesystem) will growisofs still burn the data found in the image after the filesystem?
08:50 < taaperotassu> Is it a bad practice to install programs from source from github by self made bash scripts? Using ubuntu server but not liking to install from ppa to get newer versions.
08:51 < sauvin> taaperotassu, depends on where and how you install the stuff you build.
09:03 < jcarpenter2> is there some rule that enables users to take ownership of root-owned files?
09:03 < michaelrose> taaperotassu, its poor practice and it doesn't scale
09:03 < jcarpenter2> because I can create a file "~/blah" as some user in vim, run "chown root:root blah" as root, and then type ":w!" in vim as my normal user and ownership goes right back to them
09:04 < michaelrose> for example recently gentoo saw their github compromised
09:04 < mcdnl> jcarpenter2: thats not possible
09:04 < taaperotassu> True
09:04 < jcarpenter2> mcdnl: but it's true
09:04 < taaperotassu> yup it was on news
09:05 < mcdnl> you're probably doing something wrong
09:05 < jcarpenter2> highly unlikely
09:05 < mcdnl> btw ~/blah as root it's not the same as ~/blah as any other user
09:05 < jcarpenter2> ik
09:05 < jcarpenter2> try it on your system
09:06 < lnnb> holy crap he's right
09:07 < mcdnl> that's because it's in a folder that you own
09:08 < mcdnl> still, i think its odd
09:08 < mcdnl> also, its not about vim
09:08 < lnnb> yeah and its mv command
09:08 < mcdnl> try # touch file $ rm file
09:08 < jcarpenter2> indeed, vim is just the test case
09:09 < lnnb> that's what vim does when it saves
09:09 < jcarpenter2> well that makes sense, because the swap file is owned by user:user
09:09 < jcarpenter2> makes sense sort of
09:10 < michaelrose> jcarpenter2, is this because you have the right permissions in ~ to overwrite the file with your own version?
09:10 < jcarpenter2> it does appear to be made possible because the directory is writable
09:10 < jcarpenter2> yeah, it's because the containing directory is writable by the user
09:10 < mcdnl> and its owned by you
09:10 < jcarpenter2> ty everyone
09:17 < lnnb> now try  (as root)  mkdir -p /home/user/root-test/hehe  then try to remove the dir as a reg user
09:18 < lnnb> you can't even mv it out of your home dir
09:18 < jcarpenter2> nope
09:19 < jcarpenter2> but if root-test is empty then you can remove it as a regular user
09:20 < Hasimir> interesting, but makes sense given it is your home dir
09:20 < Hasimir> re: the ~/blah example
09:21 < Hasimir> what I found slightly more interesting is that vim permits this, but emacs didn't
09:24 < jcarpenter2> can a user alter a file inside of a non-writable directory inside a writable directory?
09:24 < jcarpenter2> they can alter a non-writable file inside of a writable directory as well as an empty non-writable directory inside of a writable directory, but i'm not sure about a non-empty non-writable directory
09:26 < jcarpenter2> e.g. "<lnnb> now try  (as root)  mkdir -p /home/user/root-test/hehe  then try to remove the dir as a reg user" cannot be deleted but "mkdir -p /home/user/root-test" can be - is it true the first theoretically cannot be deleted or is "rm -rf /home/user/root-test" just not calling rm with enough flags?
09:28 < lnnb> i believe it is because the inner dir is in a directory owned by root, so when you try to unlink it recursively the permission check fails
09:38 < jcarpenter2> this kind of nonsense makes it very difficult to harden a server security-wise
09:39 < jnor> Is there a service/widget/something that can monitor for memoryleakage and alert before crash happen?
09:41 < jcarpenter2> jnor: i have the same question; free(1) can report on the amount of memory in use
09:44 < jnor> jcarpenter2: yeah so it should be easy to make a script or something, but there must already be things for this
09:44 < jcarpenter2> "must"
10:34 < Triffid_Hunter> jnor: how do you differentiate between memory leaks and normal memory usage patterns? you can use ulimit to prevent things using excessive memory
10:36 < MrElendig> you could use cgroups to limit processes
10:36 < MrElendig> before they crash the entire system
10:37 < jnor> I love you guys
10:37 < jnor> good ideas :)
10:37 < jnor> ty
10:39 < MrElendig> if it is one-off processes that runs out of ram, look at systemd-run
11:08 < epicmetal> Anyone running a Xfce or MATE on a ThinkPad X220?
11:08 < MrElendig> no, not a single person on the planet does
11:08 < epicmetal> I've noticed this infurating bug where the mouse focus certain panel applets (e.g. MATE menu, Whisker menu) flickers rapidly
11:08 < epicmetal> I think it's GTK3
11:08 < MrElendig> xfce is mostly gtk2
11:09 < MrElendig> anyway, check your logs
11:09 < epicmetal> MrElendig: for what
11:09 < epicmetal> MrElendig: nothing is output during the problem
11:10 < epicmetal> And yeah, Xfce panel seems to be GTK2
11:10 < epicmetal> So I guess it's not GTK
11:10 < MrElendig> intel igp?
11:10 < epicmetal> But if I use the normal Xfce menu applet it doesn't happen
11:10 < epicmetal> MrElendig: yeah
11:10 < epicmetal> It's the focus that's flickering... mouse focus
11:10 < MrElendig> try modesetting instead of xf86-video-intel, or the other way around
11:10 < MrElendig> and do check dmesg and the xorg logs
11:10 < epicmetal> Not sure that it's a graphics problem
11:11 < epicmetal> Seems to be more input device related
11:11 < epicmetal> So sometimes you click and the click doesn't register on the applet
11:11 < epicmetal> And you can see the highlight graphic turning off and on repeatedly
11:11 < epicmetal> There's no display corruption however
11:12 < MrElendig> so by flicker you mean "constantly opens and closes" instead of actual fliker
11:12 < MrElendig> flicker*
11:12 < epicmetal> MrElendig: the mouse focus flickers, yes
11:12 < epicmetal> not a visual artefact per se
11:12 < MrElendig> focus follow mouse?
11:12 < epicmetal> no just normal
11:12 < epicmetal> it's cross-distro, cross-desktop
11:13 < MrElendig> test with a usb mouse
11:13 < epicmetal> good idea, oen sec
11:13 < MrElendig> (disable the pad/nipple)
11:15 < epicmetal> MrElendig: trackpoint and trackpad disabled in firmware, but the problem still happens with USB mouse
11:15 < MrElendig> check xev / xinput
11:17 < epicmetal> MrElendig: well, if I move the panel out the way and put xev in the bottom left corner, the pointer co-ordiantate delta seems smooth as I edge the screen with the pointer
11:17 < epicmetal> is that what you mean?
11:17 < epicmetal> not sure what xinput is
11:18 < epicmetal> xinput lists Logitech Optical USB Mouse and Virtal core XTEST pointer
11:20 < epicmetal> On the plus side, the problem isn't related to Arch :)
11:20 < epicmetal> co-ordinate*
11:20 < jim> we're trying to discourage the use of acronyms like ik, ikr, which also looks like sms speak... so, please spell ik as I know, it will help new english speakers to understand better (or at least have a chance to)
11:21 < epicmetal> jim: was that to me?
11:21 < jim> well I didn't see any instances of that from you... I was scrolled back pretty far when I wrote that
11:21 < epicmetal> oh ok
11:22 < jim> having said that, it applies to everyone
11:22 < udkyo> orl korrect
11:22 < epicmetal> If only my desktop was as error free as my grammar
11:23 < epicmetal> Why does Linux torment me
11:23 < jim> it's as helpless as a kitten up a tree
11:23 < epicmetal> I even made a sacrifice to the systemd gods
11:25 < jim> with systemd, it tends to look like it's taking over more and more
11:26 < epicmetal> I guess I'll try the Intel/modesetting driver switch
11:26 < epicmetal> It's probably libinput though
11:26 < epicmetal> It's always pwning me
11:26 < rasputozen> anyone here use fifos a lot?
11:27 < epicmetal> jim: it really is... non-systemd choices are hard work
11:27 < jim> you mean named pipes? I use unnamed pipes :)
11:27 < rasputozen> yes named pipes
11:27 < autopsy> How do you name a pipe?
11:27 < rasputozen> autopsy: man fifo
11:28 < autopsy> rasputozen, GoodHashToker lolz.
11:28 < rasputozen> gottem
11:28 < jim> autopsy, you can make a file which is actually a pipe, and connect its ends
11:29 < autopsy> jim reading man fifo now I see that isn't that how systemd works of of FIFOs?
11:29 < jim> a named pipe, is actually two endpoints of a pair of sockets that go in each direction
11:30 < autopsy> No commas sorry.
11:30 < autopsy> No periods either.
11:32 < epicmetal> So the problem isn't intel/modesetting driver, nor Arch/Xubuntu, nor Xfce/MATE, nor touchpad/trackpoint/usbmouse, nor GTK2/GTK3.
11:32 < epicmetal> Kill meeeeeeee
11:33 < jim> there;s a way of making named pipes where you make the pipe then you fork the process, then each (child, parent process) has a copy of the pipe, and it becomes the two endpoints, not sure how that works or how you really do it
11:33 < MrElendig> might want to consider sockets instead
11:34 < epicmetal> MrElendig: ^ any other ideas...?
11:34 < jim> isn't a pipe really just two sockets?
11:35 < MrElendig> epicmetal: test with a new user
11:35 < epicmetal> It's a fresh OS install but OK
11:35 < rasputozen> MrElendig: were you referring to me with the socket comment?
11:36 < autopsy> No to jim I think.
11:36 < epicmetal> MrElendig: problem still happens for freshly created user
11:36 < MrElendig> rasputozen: no
11:37 < jim> how do you do that... make a socket in a process, then fork that process?
11:38 < Triffid_Hunter> Jim: socket(..); { maybe some dup2 in here }; int pid = fork(); ..
11:39 < Triffid_Hunter> then pid is the child pid in the parent process, and 0 in the child process
11:39 < doublehp> http://paste.debian.net/1032455/ what's wrong in my substitutions ?
11:39 < jim> ok, then how is that socket connected?
11:40 < epicmetal> It's got to be a problem with how MATE/Xfce are coding things if some applets have the problem and some don't
11:40 < jim> right, I umderstand that much about pid\
11:40 < epicmetal> Why can't I have nice things
11:44 < MrElendig> https://forum.xfce.org/viewtopic.php?id=12190
11:49 < epicmetal> MrElendig: nice find! Looks like there's no resolution, though. At least I'm not suffering alone.
11:49 < MrElendig> literally the second hit on google
11:49 < epicmetal> MrElendig: it was #5 for me now that I actually searched
11:51 < epicmetal> I guess I'll try Yet Another DE
11:52 < peetaur2> epicmetal: did you try plasma5? what do you like best so far?
11:53 < epicmetal> peetaur2: crossed off the list: GNOME, Plasma, MATE, Xfce, Budgie, Cinnamon, Deepin. I guess I can re-visit LXDE/LXQt...
11:53 < peetaur2> and did you reject each due to bugs or other reasons?
11:53 < epicmetal> peetaur2: yes
11:54 < peetaur2> yes bugs or which reason?
11:54 < epicmetal> peetaur2: performance and bugs
11:55 < epicmetal> I mean I know there are bugs in LXDE/LXQt that annoy me too
11:56 < rasputozen> can anyone reproduce jarbled output on subsequent 'echo some-random-text >myfifo' commands after running 'mkfifo myfifo; while read line; do echo $line; done <myfifo &; echo hello >myfifo; echo goodbye >myfifo; echo lol >myfifo'?
11:56 < rud0lf> i'm not sure where to ask, i guess you guys will know.. having set ftps server, can i force client to go by secure transport?
11:56 < rasputozen> im trying to figure out if its just my computer or some error in the logic or some other variable
11:56 < rud0lf> i mean, no unencrypted transfers
11:57 < peetaur2> epicmetal: and which distro and release is that? I find it hard to believe you can try so many and be annoyed by bugs in all unless it was something awful like fedora where bugs are features or something... unlike archlinux where bugs are frequent but fixed so it's just luck, not doom
11:57 < epicmetal> peetaur2: mostly Arch and Fedora
11:58 < epicmetal> peetaur2: list a distro or DE and I can tell you why I hate it ;)
11:59 < epicmetal> Plasma on Arch was probably the closest to good
11:59 < sh1ro> and then it exploded because arch + kde = kaboom
11:59 < epicmetal> But even that had bugs
11:59 < peetaur2> I've used kde forever basically, and any time I end up testing another, I find it unusable... (xfce I found not broken nor terribly designed, but not featureful enough)
12:00 < peetaur2> kde was only kaboom in the early days of every major release .... early kde4 was bad for a long time... early plasma5 was worse but for a very short time I think
12:00 < sh1ro> xfce is very nice, looking forward to one day having wayland that was thought through
12:00 < searedvandal> epicmetal, Trisquel, why do you hate that distro?
12:00 < epicmetal> searedvandal: lack of kernel firmware by default
12:01 < sh1ro> sles
12:01 < epicmetal> sh1ro: yast, snapper, and I'm cheap
12:01 < epicmetal> sh1ro: and the installer partitioner... eew
12:01 < sh1ro> you've used sles?
12:01 < epicmetal> yeah
12:02 < epicmetal> years ago
12:02 < epicmetal> sh1ro: used to have issues with yast automanaging config files
12:02 < sh1ro> yeah i havn't seen it in the wild since like 2007
12:02 < epicmetal> granted, sles isn't exactly a viable distro for personal use
12:02 < epicmetal> so it's not fair for me to slam it ;)
12:03 < sh1ro> yet somehow suse just sold for 2.5b usd
12:03 < epicmetal> sh1ro: used to work for an ISV that would sell software that they'd deploy to client sites that ran on Sles
12:03 < searedvandal> feren os
12:04 < epicmetal> searedvandal: ok i've not heard of that one
12:04 < day> damn. is suse even a thing these days?
12:04 < searedvandal> epicmetal, my goal is achieved then :)
12:04 < epicmetal> searedvandal: haha
12:04 < sh1ro> a 2.5 billion dollar thing day
12:04 < epicmetal> searedvandal: boutique obscure distros don't really count though... i could just slam ubuntu since that's what it's based on
12:04 < day> i iremember using it 20years ago, when it was rather big here but since the rise of ubuntu i havent heard much about it
12:05 < epicmetal> but i probably shouldn't be too negative
12:05 < searedvandal> epicmetal, fair enough :)
12:05 < sh1ro> i tried it like two years ago and it was really pretty average
12:05 < day> sh1ro: for some random investment group. that doesnt represent reality :^)
12:06 < sh1ro> btrfs was unstable but that's btrfs. package manager was better than it used to be but still very meh
12:06 < day> im not even sure what defines suse or any other linux os
12:07 < sh1ro> repos, package manager and dir structure for some stuff
12:07 < epicmetal> day: people
12:07 < peetaur2> sh1ro: you mean zypper? I find zypper is nearly the best... (hard to decide between featureful and functional zypper vs ultra fast simple pacman)
12:07 < peetaur2> but of course I left SuSE when they made the system unusable via systemd
12:07 < sh1ro> pacman is dope af
12:07 < epicmetal> peetaur2: ooh what are you running now
12:08 < peetaur2> epicmetal: if I wasn't putting it off, all my systems would be artixlinux now, but right now many are still on a manjaro-openrc even though the openrc part was discontinued and will inevitably blow up some day
12:09 < sh1ro> really they're all nice at this point. apt is very good until something breaks. emerge.. could use some more automation
12:09 < peetaur2> for my own servers, I use devuan mostly...and for work we use Ubuntu LTS and some CentOS (gah systemd)
12:09 < evanesoteric> I can't be using Linux as long as I have and no really know how to compile from source. A few questions if I may... If I compile say, Pidgin, from source, and I want a newer version. Should I be remove the previous version, then compiling the new version? Should I be compiling from source in /usr/bin/local?
12:10 < peetaur2> evanesoteric: you should always use packages and the package manager..so you simply install the new package, and the package manager knows how to remove the old stuff
12:10 < sh1ro> peetaur2: there is no escaping the systemd tentacle monster, it will assemilate all
12:11 < sh1ro> resistance is futile
12:11 < day> evanesoteric: if you just want to try something you can simply compile it somewhere. and run it locally
12:11 < peetaur2> so for example if you used arch based and for some reason you didn't like the binaries supplied by the distro, you can grab the PKGBUILD from https://www.archlinux.org/packages/extra/x86_64/pidgin/ and build that
12:11 < Triffid_Hunter> Jim: dup2 copies FDs from one number to another, you can use that to replace stdin/stdout with the pipe before exec'ing another process
12:11 < epicmetal> peetaur2: hmmmmm I looked at Artix briefly but can't remember now why I skipped over it. You'd go runit over OpenRC, right?
12:12 < peetaur2> epicmetal: the default in artix is openrc. It's by the same guys that made manjaro-openrc. But now they base it on arch instead of manjaro.
12:12 < peetaur2> and I don't use runit personally, or care about it or know anything about it
12:13 < epicmetal> peetaur2: why Devuan instead of Debian and switching the init? I found out recently that libsystemd0 is still in Devuan although at this point it time it apparently doesn't matter because that lib is useless if systemd isn't running
12:14 < epicmetal> I guess Devuan's overlaid packages would potentially mean less sysadmin work...
12:14 < peetaur2> devuan supports their init... like making init scripts ...why should I customize a system that was designed to force me into another choice and then fight me?
12:14 < epicmetal> I've been considering runnig Devuan Ceres
12:14 < epicmetal> Yeah, pretty much...
12:15 < epicmetal> I just wanted your view
12:15 < peetaur2> for me it's not a religious battle...I don't care if some systemd file is there if it's never actually used so it doesn't endlessly waste my time
12:15 < epicmetal> Ok
12:15 < epicmetal> The other one I like is Void... it's hard work though
12:15 < epicmetal> Although I have a strategy to get it working
12:16 < peetaur2> things like this make me never want to touch systemd....how can they break a thing like the emergency mode? http://brockmann-consult.de/peter2/systemd/2017-08-14%20ubuntu%2016.04%20emergency%20mode%20fail.png
12:16 < epicmetal> peetaur2: preaching to the choir ;)
12:17 < epicmetal> I recognise that it is featureful but it's also just nasty
12:17 < peetaur2> I finally solved that btw...turns out what it was doing is systemd would allocate 100% of all RAM on boot up, so it was a race condition... 8/10 times it would do that before the fs mounts or something like that, and kill it, and sometimes the important stuff would happen first
12:17 < epicmetal> Yikes
12:17 < epicmetal> Race conditions, what's the rush? :D
12:17 < peetaur2> so on machines like some of our processing machines where we tuned some sysctls for memory overprovisioning (where allocating 100% regardless of how much you have is a severely idiotic overprovision), it breaks
12:18 < epicmetal> It doesn't like corner cases much, does it
12:18 < peetaur2> so add that to the list of obviously insane things you should reject a project and all its deveopers for, along with umask 000 (equivalent to chmod 777) which I find one of the worst
12:19 < peetaur2> it throws away decades of best practices...it doesn't like even the general cases
12:19 < epicmetal> Yeah I just miss The Old Days
12:19 < epicmetal> The core of the OS used to be so rock solid
12:20 < epicmetal> The GUI has always been trash but at least you could depend on the basics
12:20 < epicmetal> Now it's like...
12:20 < peetaur2> yeah it was mostly wifi and xorg stuff breaking in the past... why'd that have to change
12:20 < epicmetal> It's funny because all my "safety blankets" are infected
12:20 < epicmetal> I keep getting the "oh, I'll just go back to Debian" instinct but then remember reality...
12:21 < epicmetal> It used to be that Fedora would burn me and I'd run back to Debian or something
12:25 < rasputozen> can anyone explain this behavior: https://ptpb.pw/Dwz4
12:26 < sh1ro> fedora has always been really stable for me things only go wrong during upgrades
12:27 < epicmetal> Fedora is relatively amazing
12:28 < epicmetal> I just feel like information about it is tucked away in corporate silos
12:29 < paulcarroty> Debian is good only for masturbation on old crap
12:31 < Misha_B> the best distro is the one you don't have to think about
12:31 < sh1ro> yes the arch wiki has become the be all end all of linux docs it's hard for fedora to compare
12:31 < epicmetal> Misha_B: so there is no best distro? :)
12:31 < epicmetal> paulcarroty: yeah it's a bit crusty for workstations
12:32 < sh1ro> just openbsd
12:32 < MrElendig> with the commercial support fedora has it would be easy for them to have a better wiki than the arch one
12:32 < MrElendig> same with ubuntu
12:32 < Misha_B> I use ubuntu since I haven't had to think about it since I installed it
12:32 < epicmetal> sh1ro: trueos wouldn't boot properly on this laptop
12:32 < sh1ro> which i can safely setup and leave for 5 years without once thinking about it
12:32 < epicmetal> sh1ro: haven't gone as hardcore as openbsd this decade
12:32 < peetaur2> with ubuntu, an infinite number of bloggers on an infinite number of blogs will outblog the archwiki ....if you can be careful enough to select the right ones
12:33 < epicmetal> MrElendig: i don't think it's a matter of money, but of incentive
12:33 < MrElendig> epicmetal: it is a matter of money
12:33 < epicmetal> peetaur2: signal to noise, though
12:33 < epicmetal> MrElendig: then why isn't it so
12:33 < MrElendig> you need good technical writers and enough of them
12:33 < epicmetal> i'm pretty sure archwiki is all volunteers
12:33 < Misha_B> tbh linux is just a platform to run emacs, so most distros are the same
12:34 < MrElendig> epicmetal: yes, and it sucks
12:34 < epicmetal> paid devs won't do it unless they're asked
12:34 < MrElendig> epicmetal: a lot of it is outdated/wrong/of bad quality
12:34 < epicmetal> MrElendig: did you just blaspheme against archwiki? ;)
12:34 < MrElendig> epicmetal: it is just that almost all of the others are even worse
12:34 < peetaur2> epicmetal: yes but it's so low cost that it's probably why all their funding didn't produce something better than archwiki
12:34 < epicmetal> Misha_B: i envy your simple life
12:35 < sh1ro> a life with no need of zfs
12:35 < epicmetal> heh
12:36 < epicmetal> BSD doesn't do LUKS though
12:36 < epicmetal> I know it has a native one but yeah
12:36 < azarus> dragonflybsd does luks I believe?
12:36 < epicmetal> Wow, right you are
12:37 < azarus> BSD is a bad generalization
12:37 < epicmetal> I wonder if it does LUKS2
12:37 < sh1ro> dragonfly is still a thing?
12:37 < azarus> yup
12:38 < sh1ro> i thought it just slowly died after the whole HAMMERII letdown thing
12:38 < epicmetal> azarus: I'm not sure whether to be happy or sad that you've added another OS to my shortlist :)
12:38 < azarus> hammerII letdown? what?
12:38 < azarus> epicmetal: i have many more
12:38 < epicmetal> Noooooo
12:38 < sh1ro> filesystem to replace zfs that never really panned out
12:39 < azarus> it's still active and the default filesystem for dragonfly
12:39 < azarus> if it matures OpenBSD might adopt it
12:39 < sh1ro> oh good, they might make better use of it
12:40 < sh1ro> i tried dragonfly the one time had it tell me no symbols allowed in password then laughed and shutdown the machine
12:40 < epicmetal> Do I run Void Linux or Devuan Ceres on my ThinkPad X220, and why?
12:40 < azarus> epicmetal: both work, i'm sure
12:41 < azarus> all a matter of preference
12:41 < epicmetal> azarus: which would you choose though
12:41 < azarus> neither
12:41 < sh1ro> install gentoo
12:42 < epicmetal> sh1ro: too much compile time
12:42 < sh1ro> it's an x220, it's only right
12:42 < epicmetal> sh1ro: X220i actually
12:42 < epicmetal> i3
12:42 < epicmetal> azarus: why not? And what would you run?
12:42 < sh1ro> yeah would be nice if the p50 were corebootable
12:42 < sh1ro> could have a thinkpad with a xeon
12:42 < azarus> epicmetal: i use alpine linux, mostly because of ease of development
12:43 < epicmetal> azarus: it wouldn't boot last time
12:43 < azarus> it's radically different from other distros tough
12:43 < epicmetal> as in the installer ISO hung on boot
12:43 < azarus> it takes a short while to boot
12:43 < azarus> and it is quiet during that time
12:43 < epicmetal> hmm
12:43 < user03> is linus online, does someone have his username? :)
12:43 < sh1ro> muscl is beautiful
12:43 < azarus> musl*
12:44 < epicmetal> i wouldn't run musl personally
12:44 < searedvandal> user03, what?
12:44 < sh1ro> which linus?
12:44 < user03> when?
12:44 < epicmetal> i have enough problems
12:44 < MrElendig> user03: windowslover69
12:44 < user03> :)
12:44 < sh1ro> torvalds you have to email
12:44 < user03> tux4711
12:44 < azarus> epicmetal: well i'd choose devuan then, if you want a glibc system
12:44 < searedvandal> user03, https://www.kernel.org/category/contact-us.html
12:44 < epicmetal> azarus: why that over void?
12:45 < azarus> epicmetal: dunno really, have experienced that more than void
12:45 < user03> ok thnnaks
12:45 < user03> thx
12:45  * sh1ro isn't saying it's highly compatible just nice code
12:45 < azarus> but then again, both package management systems (xbps and apt) are so overcomplicated for me
12:45 < azarus> i wouldn't run either personally
12:46 < luke-jr> [10:44:11] <epicmetal> i have enough problems <-- this is why I hacked my GCC/LLVM to use signed char :P
12:46 < epicmetal> luke-jr: I know what all of those things are individually, but I have no idea what you mean :)
12:47 < user03> hello from my investor and mother ;)
12:47 < epicmetal> one nice thing i will say about xubuntu, its fonts are gorgeous
12:47 < luke-jr> epicmetal: The C standard does not define whether the "char" type defaults to signed or unsigned. Some incompetent coders assume it is signed because that's what GCC does by default on ARM and x86.
12:47 < sh1ro> xubuntu is my favorite for non linux users
12:47 < luke-jr> Hence using signed char minimises problems
12:48 < epicmetal> sh1ro: my aforementioned panel blinking bug is making it unusable though
12:48 < sh1ro> windows and mac users alike can figure it out pretty fast and i can install an LTS version if it's appropriate
12:48 < epicmetal> luke-jr: so why force it to be signed? oh, for compatibility...?
12:49 < sh1ro> epicmetal: is it to do with the intelligent hide feature?
12:49 < luke-jr> epicmetal: right
12:49 < epicmetal> sh1ro: no https://forum.xfce.org/viewtopic.php?id=12190
12:49 < luke-jr> epicmetal: on my platform (POWER), GCC likes to default char to be unsigned
12:50 < epicmetal> luke-jr: you have a talos?
12:50 < sh1ro> huh that's the only xfce issue i've had in the past what 4 years
12:50  * sh1ro is now mad jelly of luke-jr
12:50 < epicmetal> sh1ro: you get this bug, too?
12:50 < BluesKaj> Hi folks
12:51 < epicmetal> It's pretty infuriating for the start button not to open when you click it
12:51 < sh1ro> no i had an odd bug with intelligent hiding on qubes
12:51 < searedvandal> blinking panel in xfce? never experienced
12:51 < epicmetal> searedvandal: MATE and Xfce suffer from it for some users
12:52 < luke-jr> epicmetal: yes
12:52 < searedvandal> I see. never been an issue on my systems, not even the nvidia one. but I can understand that it would make the system quite unusable
12:52 < luke-jr> sh1ro: there's a few "defective" Talos Lites available at $1600 FWIW
12:52 < luke-jr> (fully assembled machine)
12:53 < sh1ro> define defective
12:53 < luke-jr> the virtualisation stuff is broken
12:53 < epicmetal> luke-jr: what OS? any problems with it?
12:53 < luke-jr> epicmetal: Gentoo; only problem is Linux kernel issues with AMD GPUs
12:53 < sh1ro> that kinda kills the platform form me
12:54 < epicmetal> luke-jr: what issues with the gpu?
12:54 < luke-jr> epicmetal: for now, I am using fbdev and software 3D
12:54  * sh1ro will probably end up on an epyc in a couple years
12:54 < epicmetal> ah
12:54 < epicmetal> :\
12:54 < luke-jr> epicmetal: AMD cards only support 40-bit DMA, whereas Linux wants to use 32-bit or 64-bit DMA
12:54 < epicmetal> good thing i'm too cheap to buy one
12:54 < luke-jr> so it uses 32-bit DMA, which isn't enough address space for all my windows
12:54 < sh1ro> need something x86_64 so i can reasonable game in a VM
12:54 < luke-jr> sh1ro: hopefully POWER stuff gets cheaper over the next few years
12:55 < sh1ro> that would be nice i'd toatally throw the parts into my old power mac
12:55  * luke-jr just recompiles the games for POWER :P
12:55 < sh1ro> yeah but then i'd have to hack activision and ubisoft and that gets me on all kinds of shitlists in china
12:56 < sh1ro> my sesame score would be under -9000 x'D
12:56 < luke-jr> or play free games ;)
12:56 < oiaohm> luke-jr: https://www.systutorials.com/linux-kernels/47535/alsa-hda-limit-40bit-dma-for-amd-hdmi-controllers-linux-3-4-107/   ???? luke-jr different things in Linux do use 64 DMA limited to 40bit dma
12:57 < oiaohm> luke-jr: so not using 40 DMA for a card that wants it would be driver bug.
12:57 < sh1ro> games are like phones or skype, a degree of proprietary which i must allow to maintain a social life
12:57 < luke-jr> oiaohm: it's something related to how Linux does IOMMU setup on POWER IIRC
12:58 < luke-jr> oiaohm: on x86, I don't think Linux even uses the IOMMU for drivers
12:58 < oiaohm> luke-jr: there is devices on Power that are also 64 dma limited to 40bit dma
12:58 < sh1ro> power would be my pick  for a workstation or server though... or maybe sparc, sparc was cool for a while
12:58 < oiaohm> luke-jr: IOMMU can do a limited mode with a mask.
12:59 < oiaohm> luke-jr: but you do have to use the mode larger and scale it back.
12:59 < sh1ro> luke-jr: will you be the one guy trying to get navi working on ppc next month?
13:01 < luke-jr> sh1ro: what?
13:02 < oiaohm> luke-jr: could be something in the power IOMMU driver where it not properly accepting masking down of mode.
13:02 < epicmetal> azarus: so don't bother with Alpine unless I want to run musl?
13:02 < azarus> epicmetal: alpine only uses musl, yes
13:02 < epicmetal> oh
13:02 < epicmetal> I thought it did both for some reason
13:02 < azarus> earlier on it had uClibc
13:03 < azarus> and it shouldn't matter anyway unless you run precompiled (bleh) stuff
13:04 < epicmetal> I don't at home, but I might at work
13:05 < sh1ro> luke-jr: new amd gpu which will hopefully work better than fronteir did
13:05 < luke-jr> oiaohm: there's a fix supposedly
13:05 < luke-jr> sh1ro: I don't care to mess with new GPUs
13:05 < luke-jr> especially not ones that require signed firmware
13:07 < MrElendig> luke-jr: so you mean any gpu made after ~2010
13:07 < MrElendig> :p
13:08 < luke-jr> MrElendig: no
13:08 < luke-jr> pretty sure AMD worked with unsigned firmware until just recently
13:10 < sh1ro> can't you mitigate any firmwarefuckery with iommu
13:10 < luke-jr> sh1ro: to a limited extent, perhaps, but the GPU has access to everything displayed on the screen..
13:10 < luke-jr> besides, there's nothing I need a newer GPU for
13:11 < oiaohm> luke-jr: IOMMU x86 drivers have to be used with Linux x86 when you are messing around video card pass though and the like.     So there should be a fix at some point.
13:11  * sh1ro does, mining, hash cracking and gaming in that order
13:12 < luke-jr> oiaohm: but in that scenario, the GPU interacts with the guest as if it was just a normal IOMMU-less system
13:12 < oiaohm> sh1ro: if you are sure iommu is functioning correctly you can be sure that you video card firmware cannot go off and like tell you disc controller to delete disc or the like without OS approval.
13:12 < luke-jr> oiaohm: in this case, Linux is trying to map the memory with the high bit set on the address
13:12 < luke-jr> which puts it above 40-bit
13:13 < oiaohm> luke-jr: that is when things are going to turn horrible with driver built with sub 40 address and iommu possible not masking to down to correct address range.
13:13 < luke-jr> oiaohm: but you can't be sure it isn't saving private info and leaking it in watermarks for screenshot exposure
13:13 < afidegnum> hello, i can't get my monitor info using get-edit | parse-edit any alternate command?
13:13 < luke-jr> of course, that's a problem with any non-free firmware, not just signed
13:14 < MrElendig> afidegnum: which info?
13:14 < MrElendig> afidegnum: get-edid and parse-edid works, unless you have broken edid
13:15 < oiaohm> luke-jr: think if the card can dma card to card as you could in older systems with IOMMU.   Graphics card screen caputring and sending it out by network use to be possible.
13:15 < MrElendig> no idea what get-edit is
13:15 < sh1ro> guess you'll be waiting for intel gpus to upgrade then
13:15 < oiaohm> luke-jr: so IOMMU improves things a little.
13:15 < MrElendig> just reading the manual for the monitor tend to work too
13:15 < MrElendig> sh1ro: intel requires firmware blobs
13:15 < luke-jr> sh1ro: TBD if Intel will be any better
13:15 < oiaohm> luke-jr: that is of course if the IOMMU it self does not have a flaw
13:16 < luke-jr> MrElendig: someone has a discrete Intel card?
13:16 < MrElendig> their igps requires blobs, you can be 100% sure their dedicated card, if they ever release it, will too
13:16 < luke-jr> igps?
13:16 < MrElendig> https://en.wikipedia.org/wiki/Graphics_processing_unit#Integrated_graphics
13:16 < luke-jr> integrated Intel GPUs don't require blobs, though..
13:17 < MrElendig> luke-jr: they do
13:17 < MrElendig> luke-jr: since quite some time ago
13:17 < luke-jr> not uploaded by the OS, at least
13:17 < MrElendig> luke-jr: distributed in linux-firmware
13:17 < luke-jr> MrElendig: my last x86 system was Haswell, and it didn't.
13:17 < azarus> i915 firmware, right?
13:18 < jim> Triffid_Hunter, just saw your explanation... thanks
13:18 < MrElendig> luke-jr: skylake and up
13:18 < oiaohm> luke-jr: just because you did not load blobs does not mean they did not get loaded by the firmware/efi
13:18 < MrElendig> (some older chips too)
13:18  * azarus is happy with his ivy bridge
13:19 < MrElendig> gentoo docs on it: https://wiki.gentoo.org/wiki/Intel#Firmware
13:19 < oiaohm> luke-jr: you see this a bit on intel hardware were people are so happy they got their comptuer working without using firmware blobs completely missing that the bios/efi did it for them.
13:20 < luke-jr> tbh I was expecting ME did it
13:20  * MrElendig have little faith in the state of the gnu/linux support for the intel dedicated card if it ever comes out too
13:21 < oiaohm> luke-jr: ME gets it firmware images out the motherboard firmware image other than it backup firmware that intel has added a 30 min timer to shutdown to.
13:23 < afidegnum> after startx "unable to connect to x server: connection refused"
13:23 < afidegnum> what could be the cause?
13:23 < azarus> afidegnum: lots, post Xorg.0.log
13:24 < system16> linux is immune to ratting right ?
13:25 < azarus> ratting?
13:25 < system16> yes
13:25 < system16> search this in youtube : ratting scaammers
13:25 < system16> scammers*
13:25 < MrElendig> what does linux has to do with your hair style?
13:26 < MrElendig> system16: still doesn't make any sense at all
13:26 < system16> wait...
13:27 < ThePendulum> 'lo
13:27 < searedvandal> the bloodsport of dogs catching rats?
13:27 < system16> no
13:27 < searedvandal> or telling on people?
13:27 < MrElendig> only hit on yt is some crappy indian videos about making fun of tech support scams
13:27 < system16> its a app that makes trojans
13:27 < MrElendig> I don't see how that is relevant in any way
13:28 < ThePendulum> I'm mounting an sshfs resource with this in fstab: pi@192.168.178.20:/mnt/elements /mnt/elements fuse.sshfs defaults,_netdev,allow_other,IdentityFile=/home/niels/.ssh/id_rsa 0 0
13:28 < system16> https://www.getsafeonline.org/protecting-yourself/avoiding-ratting/
13:28 < system16> rat : remote access trojan
13:28 < ThePendulum> however, it won't show up in boot, but will once I run mount -a again, although it says the resource is busy; it'll show up anyway
13:28 < ThePendulum> *on/after boot
13:28  * MrElendig wouldn't trust any webpage that has the url "getsafeonline"
13:28 < luke-jr> system16: Linux is just an OS. It can't protect you from yourself.
13:29 < MrElendig> you literally can't make a OS immune to trojans
13:29 < system16> what if someone emails me a remote access trojan ?
13:29 < searedvandal> but no, there are RAT malware that are cross platform, including linux
13:29 < luke-jr> system16: then don't run it, duh
13:29 < MrElendig> you can make it hard for the attackers, but you can't make it immune
13:30 < MrElendig> system16: get a real job instead of working for a spyware company? :p
13:30 < system16> actually rat is a good think in the right hands
13:30 < system16> thing*
13:30 < luke-jr> …
13:30 < MrElendig> there is never "the right hands"
13:30 < luke-jr> Intel ME built into every CPU is a RAT I guess
13:31 < MrElendig> it's not in every intel cpu
13:31 < luke-jr> every Intel CPU*
13:31 < system16> u know.. there is a guy that always bullied me. he has a pc...
13:31 < MrElendig> and for most of those where it is available it is disabled by default
13:31 < system16> i know his email address
13:31 < luke-jr> MrElendig: nope
13:31 < nidefawl> Wouldn't it be easy to proof if there was network traffic from Intel ME?
13:32 < peetaur2> wouldn't moving your machine within 1 Schwarzschild radius of a dense mass make it immune to trojans?
13:32 < system16> its time to get revenge... im gonna rat him
13:32 < luke-jr> nidefawl: lack of network traffic doesn't mean it isn't there
13:32  * MrElendig turns system16 over to the police
13:32 < luke-jr> system16: confessing to a crime on IRC is pretty dumb
13:32 < jim> system16, one sec
13:32 < nidefawl> no traffic means no traffic?
13:32 < MrElendig> breaking the law to get revenge is about the most stupid thing you can do
13:32 < searedvandal> system16, how thoughtful of you admitting to a crime on irc
13:32 < luke-jr> nidefawl: no traffic means it's not being used
13:32 < system16> MrElendig, just jk
13:32 < system16> lol
13:32 < MrElendig> but then again, my trolldar is getting a strong responce anyway
13:33 < searedvandal> it's way past maximum here
13:33 < searedvandal> I need to adjust range
13:33 < system16> anyway so remote access trojans cant run on linux ?
13:33 < searedvandal> they can
13:34 < nidefawl> They ran even better
13:34 < luke-jr> system16: even if you really are kidding, it's still stupid because you are now the prime suspect if he gets infected by anyone
13:34 < searedvandal> if they are written to do so
13:34 < nidefawl> run*
13:34 < peetaur2> just make remote access impossible, like what I suggested
13:34 < luke-jr> system16: Linux can run anything you decide to run..
13:34 < system16> luke-jr, i bluffed. idk his email
13:34 < luke-jr> system16: irrelevant
13:35 < searedvandal> system16, why so worried about RATs?
13:35 < iflema> snakes
13:35 < nidefawl> cats
13:35 < jim> system16, so it's going to be enough, that you have a place to come and not be bullied or whatever... no name calling, as long as I'm here... AND... don't ask people here to do anything like attacking, and it's better to learn what you can here and do stuff with it, put it tward your resume
13:35 < luke-jr> system16: if you have an Intel or AMD PC, you already have a RAT you can't remove, so why worry? ;)
13:35 < system16> because it can remotely control ur pc
13:36 < system16> delete files...access photos... thats why im worried
13:36 < nidefawl> They steal your dogecoin
13:36 < iflema> squirrel
13:36 < jim> system16, one thing, please don't use abbreviations... Please spell out u as you... it would help folks who are here and are new english speakers, some don't hear u as a rhyme for you
13:36 < system16> oh ok
13:37 < jim> thanks
13:37 < luke-jr> system16: that's unavoidable with Intel/AMD-based computers
13:37 < system16> maybe its not a bad idea to install an antivirus solution on linux ?
13:37 < azarus> there are basically none
13:37 < searedvandal> system16, I see. anyways, there are RATs that are written to run on pretty much any OS. There will always be risks.
13:38 < luke-jr> system16: antivirus is basically snake oil
13:38 < iflema> snakes
13:38 < luke-jr> it doesn't really protect you
13:38 < system16> its a scary thing tho. even worst than wannacry
13:38 < nidefawl> They are
13:38 < system16> though*
13:38 < azarus> worse*
13:39 < azarus> it's*
13:39 < system16> oh
13:39 < MrElendig> as long as there are humans using the computers, trojans will be a thing
13:39 < MrElendig> period
13:39 < luke-jr> system16: it's basically impractical to avoid for common folk, so just assume you're always being monitored
13:39 < searedvandal> no need for antivirus, just don't execute random things you find on the internet / get emailed
13:39 < nidefawl> Get a proper mail provider
13:39 < nidefawl> Do regular OS and browser updates
13:39 < jim> and contraceptives :)
13:40 < system16> yeah those scammers got ratted because they opened that trojan file
13:40 < luke-jr> searedvandal: that won't avoid ME/PSP
13:40 < searedvandal> luke-jr, nothing will avoid that
13:40 < system16> I use gmail
13:40 < azarus> get a computer without ME/PSP
13:40 < azarus> system16: one of the worst email providers
13:40 < system16> why ?
13:40 < azarus> they read every email
13:40 < peetaur2> best to just assume any mail provider is evil than try to find a non-evil one
13:40 < azarus> them being google
13:40 < searedvandal> system16, but if you're so worried, the only "anti malware" software I would consider is clamav
13:40 < luke-jr> searedvandal: Talos/POWER will, but even then, there's a hole in the OS update mechanism
13:41 < system16> thanks searedvandal
13:41 < peetaur2> same works for politics
13:41 < system16> i cant use hotmail lol
13:41 < azarus> hotmail is just about as bad
13:41 < jim> clamav, doesn't that do antivirus for win machines?
13:41 < system16> azarus, what provider do you use ?
13:41 < azarus> system16: a paid provider
13:42 < system16> like?
13:42 < azarus> there are many
13:42 < azarus> just look for one that suits your needs
13:42 < luke-jr> system16: most ISPs include email hosting
13:42 < jim> system16, what do you really want to know? :P
13:42 < nidefawl> gmail, they have 2factor authentication
13:42 < luke-jr> and with ISP email, at least you're paying them for the service, so have a justification to legally expect privacy
13:42 < jim> do ISPs have ESPs?
13:43 < nidefawl> I have a bad experience with small mail providers, they are misconfured, accept phishing mails from the wrong server
13:43 < nidefawl> misconfigured*
13:43 < luke-jr> nidefawl: [11:40:29] <azarus> system16: one of the worst email providers
13:43 < system16> luke-jr, not in where i live lol
13:43 < nidefawl> At least they don't accept wrong servers mails
13:44 < luke-jr> system16: if you really need private email nobody else can read, learn to use PGP; even hosting yourself won't protect you
13:44 < luke-jr> nidefawl: there are no "wrong servers mails"
13:44 < nidefawl> I've seen @paypal.de mails with legit headers that were a scam attempt in a friends inbox on a friends inbox
13:44 < system16> luke-jr, nah im fine with gmail
13:45 < system16> isnt it illegal for google to read emails ?
13:45 < azarus> you agreed to their terms of use
13:45 < MrElendig> system16: no
13:45 < MrElendig> sidenote: google have cut down on the email reading
13:45 < system16> it did not say "hey we have the right to read your emails"
13:46 < nidefawl> Its on page 216, you read that page?
13:46 < MrElendig> system16: it does say that
13:46 < qman__> for one, SMTP is neither secure nor private as-is, so anyone along the path of transit can already read your emails
13:47 < MrElendig> https://policies.google.com/privacy/example/collect-information
13:47 < system16> well according to you guys i should go to a jungle and disconnect myself from the world
13:48 < jim> system16, no... but, you -should- consider becoming proactive
13:49 < qman__> on the one hand, yes, everyone is trying to track you and any amount of usage of the internet will get some of your information out there
13:49 < BluesKaj>  system16 the paranoia you see here is merely users exaggerating the vulnerabilities ,because unfortunately they do exist , but few know how to really act on them
13:49 < MrElendig> you should just do something more productive with your time
13:49 < qman__> but on the other hand, maybe not broadcast all your information to the world
13:49 < system16> MrElendig, im trying to learn autocad inventor...
13:50 < jim> system16, do you need autolisp?
13:51 < system16> what is it ?
13:51 < jim> its a scripting language that autocad uses
13:51 < system16> no. autocad is complicated in its default settings. i dont wanna make things harder/
13:51 < system16> .
13:52 < system16> anyway thanks for your help guys. bye
13:52 < jim> lisp is pretty easy... I have a video
13:53 < system16> on youtube ?
13:53 < jim> yeah
13:53 < system16> oh great.
13:53 < system16> do you have a link ?
13:53 < jim> wanna look at the first one?
13:53 < jim> yeah
13:54 < jim> just a sec
13:54 < system16> yeah
13:56 < jim> it's actually a series of 20 videos... but the first one teaches you all of lisp (well really all of a dialect of lisp called scheme) in like 20 mins (the vid is an hour tho)
13:56 < jim> are you running linux now?
13:57 < jim> if so, probably be easy to get you a copy of the scheme interpreter so you can run the examples
13:57 < system16> jim autocad wont run in linux
13:57 < system16> but i have linux in a VM
13:57 < jim> but scheme will
13:58 < jim> it's pretty close to lisp, not too different
13:58 < system16> jim but i have installed autocad on my windows pc
13:58 < jim> plus...
13:59 < jim> this series of videos will show you most of the important concepts in programming
13:59 < system16> great
13:59 < jim> here's the video: https://www.youtube.com/watch?v=2Op3QLzMgSY
13:59 < system16> thanks
14:00 < Misha_B> you just have to get used to all the parentheses
14:00 < jim> it's also been called the best intro to CS on the planet
14:01 < azarus> i also call myself the best person in the world
14:01 < limbo_> how do I write to file, or exit vim without a : key?
14:01 < azarus> ZZ
14:02 < azarus> writes and quits vim
14:02 < jim> you don't have :?
14:02 < peetaur2> random VPS things have broken keyboard mappings so lots of chars are missing... : is a common one, also | (arg!)
14:03 < Misha_B> go into another terminal and type killall vim
14:03 < azarus> ZQ if you want to quit without writing
14:03 < jim> system16, which linux do you have on your vm?
14:03 < system16> ubuntu and linux mint
14:04 < jim> ok, one sec
14:04 < system16> oh and i also have ubuntu installed on a usb stick
14:04 < limbo_> jim: yes.; and :. the key is broken, so I disabled it.
14:05 < peetaur2> azarus: what about if it says  "E45: 'readonly' option is set (add ! to override)"
14:07 < jim> limbo_, can you put it on a different key?
14:07 < peetaur2> I banged keys until eventually hitting !! made it give me a :.! and then I could hit backspace to get : and finish wtih :wq!
14:08 < jim> system16, ok, I found lots of scheme interpreters you can use if you decide you want to do some of the things it's showing you
14:08 < Misha_B> now to avoid that problem entirely in the future
14:08 < Misha_B> just do sudo apt install emacs
14:09 < jim> guile-1.8 (you should have that available, or) guile-2.0
14:09 < jim> emacs is a pretty good editor, so are others
14:10 < azarus> Misha_B: he will have the same problem with emacs
14:10 < Misha_B> not many commands require ':'
14:11 < Misha_B> just do C-x C-s to save, C-x C-c to quit
14:12 < azarus> the right way to fix this is to fix the keyboard layout
14:12 < jim> well show him how to run the tutorial :) then at least he will know what C- and M- are :)
14:14 < jim> system16, does it look interesting?
14:15 < Misha_B> well C- is ctrl, M- is alt. Although I was mostly joking about installing emacs just because you're : key is broken. The obvious solution is to repair or buy a new keyboard. accessing emacs tutorial requires C-h t (which is holding control while pressing h, then releasing ctrl and pressing t) .
14:15 < Misha_B> your*
14:17 < limbo_> Misha_B: It's the built in keyboard in a netbook
14:17 < jim> Misha_B, thanks for explaining to him...
14:17 < limbo_> I'd just use nano if there's no solution to this.
14:17 < jim> limbo_, does the netbook have usb?
14:17 < limbo_> yes, I don't want to have to plug in a keyboard every time I use vim though,
14:17 < jim> nano is a good one too
14:18 < azarus> limbo_: is this in xorg or out of xorg?
14:18 < ZetFury> test ^^
14:18 < L00P3X> Sorry, i have a folder whit directories and subdirs.. i would just put all files in it to one single dir. could you help me whit a comand? catfish do not let me copy things
14:18 < Misha_B> I wonder, you could probably use a virtual keyboard. or rebind a key chord to simulate :
14:19 < jim> limbo_, I can understand that... still, you can get the key back if you do that :)
14:19 < limbo_> azarus: just terminal.
14:19 < azarus> limbo_: are other keys correct?
14:19 < limbo_> headless home server.
14:19 < limbo_> yes
14:19 < jim> this netbook is?
14:20 < limbo_> yes
14:20 < jim> oh, then of course you could ssh from a different machine
14:20 < limbo_> jim: I'm trying to fix my inability to ssh
14:21 < jim> limbo_, oh ok, maybe one of us can help you with that
14:21 < sirvictory4> ls
14:21 < limbo_> fail2ban is blocking me, I think. Despite me being on the same /24
14:21 < jim> ls: no such file or dir
14:21 < sirvictory4> hehe
14:21 < sirvictory4> whoops
14:22 < limbo_> From multiple machines though, so it might not be fail2ban doing it.
14:22 < Gringonar> Hey guyz
14:22 < jim> hi
14:22 < Pentode> hi
14:23 < Gringonar> Thanks again for helping me out
14:23 < Misha_B> L00P3X: for i in $(find . -not -type d); do cp $i .; done;
14:24 < Gringonar> I got some things fixed but pacman gives https://paste.debian.net/1032397/
14:25 < Gringonar> lspci no longer hangs with modproble.blacklist=nouveau, but still havent gotten an DE running :(
14:25 < Pentode> Gringonar, can you do a ls -la on 48x48/icons and paste it?
14:26 < Pentode> ../ even
14:26 < Pentode> or both, or the whole tree.
14:26 < Gringonar> yeah sure give me a moment
14:27 < Pentode> or just try doing chmod -R 755 /usr/share/icons/ and see if that helps
14:27 < jim> Gringonar, are you running the arch dist?
14:28 < L00P3X> Misha_B do i should run it in the folder to get al files in it?
14:29 < L00P3X> thank you
14:29 < Gringonar> yes
14:29 < Gringonar> its the only modern distro that works atm
14:30 < Gringonar> "modern"well anno 2017/2018 anyway
14:30 < Pentode> oh you were that guy i remember you, lol.
14:30 < Pentode> did you ever try those two suggestions? disabling uefi and or booting the usb media with the hard disk removed / windows efi partition gone?
14:31 < L00P3X> Like magic. thank you <3 whit love
14:31 < jim> at least turn secure boot off
14:32 < Pentode> he tried that
14:32 < Pentode> among other things, it's being stubborn
14:33 < Gringonar> yeah now the ethernet is being ... even tho it work perfectly before
14:33 < Pentode> maybe you should return that laptop and get one more suitable for linux
14:34 < reynierpm> hi there, I did know this channel is only for Linux related q/a and I should not ask anything else but I need some help with Vagrant and maybe someone here has the experience or knowledge to help, could you try to help here: https://stackoverflow.com/questions/51210481/how-to-properly-setup-a-vagrantfile-to-avoid-user-not-found-issues ???
14:35 < Gringonar> Pentode you could be right, but nothing in the price range compares :(
14:35 < Pentode> :(
14:36 < jim> Gringonar, well is the ethernet cable connected? also is the link light on?
14:40 < Gringonar> the lights are hard to but think so it says its waiting for carrier
14:41 < Gringonar> ok replugging was enough lol
14:41 < jim> Gringonar, maybe the cable isn't secure?
14:42 < Gringonar> it clicked so maybe i didnt push it far enough the first time anyway its fixed
14:43 < Gringonar> i have to rearange some logs stuff so i can make new ones
14:43 < jim> ok, good... can you ping something by name? howbout: ping yahoo.com
14:44 < Pentode> i think he has one of those firmware glob atheros controllers ;p
14:45 < Gringonar> You are right
14:45 < Gringonar> its qualcom atrheros
14:45 < Gringonar> Killer series w/e that means lol
14:45 < Gringonar> sounds fancy ^^
14:45 < Pentode> lol
14:46 < jim> Gringonar, do you have a wireless access point?
14:47 < Gringonar> yes but not tried wireless yet
14:48 < Gringonar> im more into getting graphics up n running atm
14:48 < Gringonar> the cable is a minor inconvenience
14:49 < jim> Gringonar, ok... in any case, your net is working with the ethernet and the cable
14:51 < Gringonar> yep ty for your suggestion
14:52 < Gringonar> ill try to get wireless working at some point i hear that can be difficult proces as well. not looking forward to it :(
14:53 < jim> it's true, if you're ok with wired, that's fine
14:57 < Psi-Jack> Ohhh. Killer card. Those are pretty silly and more so annoying as heck.
15:00 < zetheroo> how can I tell what DE a user is using (via terminal)?
15:00 < sh1ro> that DE will have a process with a somewhat identifiable name
15:01 < epicmetal> zetheroo: maybe $XDG_CURRENT_DESKTOP
15:01 < epicmetal> Although I swear it had a different name on Arch (currently Xubuntu)
15:01 < epicmetal> I recall a *SESSION* variable
15:01 < Pentode> Psi-Jack, yeah. They are a real nuisance
15:01 < epicmetal> $DESKTOP_SESSION
15:01 < zetheroo> I am logged into the users PC as a different user (the administrative user)
15:02 < zetheroo> this is Ubuntu fwiw
15:02 < zetheroo> is the DE noted in a file somewhere ?
15:02 < stevendale> echo $XDG_CURRENT_DESKTOP > DE | cat DE
15:02 < stevendale> This worked for me
15:03 < Psi-Jack> wut?
15:03 < Psi-Jack> Why echo it to a file and cat that?
15:03 < zetheroo> stevendale: returns blank
15:05 < triceratux> https://www.ostechnix.com/test-100-linux-and-unix-operating-systems-online-for-free/
15:08 < L00P3X> thanks again o/
15:08 < Gringonar> the wireless is a Intel Wireless-AC 9560
15:08 < Psi-Jack> just echo $XDG_CURRENT_DESKTOP /should/ show the current DE if your distro has XDG stuff properly setup as expected.
15:08 < zetheroo> is Enlightenment a DE or WM?
15:08 < Psi-Jack> Enlightenment is a WM.
15:09 < Psi-Jack> Kinda a pseudo DE in its own way, but incomplete.
15:09 < Gringonar> and /Adwaita/48x48/ is not present in /usr/share/icons
15:09 < zetheroo> So if the distro comes with Gnome and Enlightenment is installed does that change everything the same way installed KDE would?
15:09 < Gringonar> so it was never created
15:10 < Psi-Jack> "change everything?"
15:11 < zetheroo> well if I switch from Gnome to KDE it's not only the look that changes, but the apps change, the system utilities change ...
15:11 < stevendale> Yeah GNOME and KDE come with different apps preinstalled zetheroo
15:11 < FightingFalcon> What are good optimal values for sysctl.conf?
15:11 < zetheroo> like if I install Ubuntu and Kubuntu on two separate systems and compare them ... problems and solutions in the GUI are incredibly different
15:11 < luke-jr> oiaohm: fwiw, this is the patch that supposedly fixes the 40-bit DMA thing: https://paste.pound-python.org/show/nZAWnXntAFm38JBoSkXK/
15:12 < luke-jr> Psi-Jack: ever figure out about how Solus has lower memory use?
15:12 < Psi-Jack> luke-jr: It's just better optimized. :)
15:12 < luke-jr> Psi-Jack: but how?
15:12 < zetheroo> stevendale: right. But Enlightenment, is that just a change of look and feel? Does it still use the Gnome apps and system utils etc ...?
15:13 < peetaur2> FightingFalcon: I always set  kernel.sysrq=1  and the rest depends on your goal
15:13 < Psi-Jack> Well, it doesn't aggressively try to do -O3 for starters. Mostly does -O2.
15:13 < peetaur2> (and kernel.sysrq=1 will do nothing if you don't know what it does()
15:13 < luke-jr> everyone does -O2 though?
15:13 < FightingFalcon> peetaur2 what happens when its 1 or 2?
15:13 < kubast2_> How can I snapshot btrfs rootfs into subvolume without live usb?
15:13 < kubast2_> Do I just do it? Or?
15:13 < peetaur2> FightingFalcon: https://www.kernel.org/doc/html/latest/admin-guide/sysrq.html
15:14 < kubast2_> It will create @snapshot-date folder in pwd?
15:14 < peetaur2> I consider it essential to have REISUB work... when you need it and it's not enabled, you can lose data (like a hard power off..you should never lose data, but bugs happen, and this reduces the chance of that to about 0)
15:15 < Psi-Jack> Heh. I really don't know. Hard to really determine the cause, just the effect. :)
15:15 < peetaur2> kubast2_: just snapshot it
15:16 < peetaur2> if you have a microsoft mentality (that nothing syncs properly), then you also must freeze all io...get all programs to complete whatever they're working on (get to the point in between where it's safe to interrupt...where it syncs, in case it doesn't already at every instant like every sane program should), then it's a safe time to snap
15:19 < Gringonar> i have no idea where the failed to extract error comes from. i assume it creates the folder before trying to extract to it eitherwise it seems dumb
15:21 < triceratux> Gringonar: is this still original vanilla archlinux with no video & no network ?
15:22 < Gringonar> This is what lspci says about my laptop https://bpaste.net/show/6f9ac7594983
15:22 < Gringonar> I didnt add any packages, but the netwerk (ethernet) works out of the box
15:23 < Gringonar> network
15:24 < Gringonar> its the one i got from the torrent here https://www.archlinux.org/download/
15:24 < Gringonar> i'm not sure if thats (vanilla)
15:26 < cousteau> diff between /etc/shadow and /etc/shadow- ?
15:26 < triceratux> if i were in your position id look thru the output of this http://pastebin.centos.org/910191/raw/
15:26 < cousteau> nvm; shadow- seems to be a backup
15:30 < FightingFalcon> is there a command to see how many open sockets my server has?
15:34 < papl> I setup an ssh account jail for sftp transfers, but I my ssh/ftp clients can't do anything because their shell is set to nologin
15:34 < peetaur2> FightingFalcon: lsof -Pni, or ss
15:35 < papl> What I followed instructions in https://wiki.archlinux.org/index.php/SFTP_chroot
15:37 < peetaur2> instead you have to check the command... like this script as the shell https://bpaste.net/show/d119afcbb033  and  instead of the default internal sftp server, in sshd set  Subsystem sftp /usr/lib/openssh/sftp-server
15:38 < peetaur2> ^ papl
15:38 < Gringonar> triceratux uhm i download the script n renamed the ext to .sh but uhm how to i execute it run Script.sh not working :(
15:38 < peetaur2> (should probably trim out the obviously incomplete debug stuff)
15:39 < triceratux> Gringonar: itll need a chmod 755 so its executable
15:39 < twainwek> u+x is sufficient
15:39 < Gringonar> the file has x permision as far as i can tell with ls -la
15:40 < thebigj> I am trying to install and use LVM in my GNU/Linux distribution. This is my architecture and distribution https://dpaste.de/dkV1/raw Here is the output I get while installing the lvm package https://dpaste.de/OMON/raw
15:40 < thebigj> The package is installed but I am getting error or warning when I run "vgcreate --help" https://dpaste.de/X4X8/raw
15:40 < Gringonar> its rwxr-xr-x permission
15:40 < thebigj> Can anyone guideme what things I am doing wrong?
15:40 < thebigj> I am on Scaleway cloud instance
15:41 < Gringonar> u permission?
15:41 < twainwek> Gringonar: and what does "not working" mean
15:41 < Psi-Jack> thebigj: Did you contact Scaleway for support?
15:41 < jhodrien> thebigj: There's a service not running.  lvm2-lvmetad.service
15:42 < thebigj> jhodrien: I agree this is a problem of service. Let me try it once again.
15:42 < PlasmaHH> Hi, I have some remotely located antique system (2.6.18) that even after 3>drop_caches has 22.8G cached memory and all of its 50G swap full, but /proc/*/smaps show not a single byte being in swap.. any idea where that memory might have gone?
15:42 < triceratux> Gringonar: did you remember to say ./Script.sh ? it wont magically find scripts in the current dir
15:42 < Gringonar> zsh: command not found: run
15:42 < Gringonar> im sure im doing it wrong
15:43 < thebigj> jhodrien: Hey, thanks!
15:43 < Gringonar> i did run Script.sh
15:43 < thebigj> jhodrien: I did systemctl start lvm2-lvmetad.service
15:43 < thebigj> jhodrien: and it is running!
15:43 < thebigj> jhodrien: Thanks!
15:43 < triceratux> Gringonar: you have to say "./" in front of it
15:44 < Gringonar> ok
15:44 < ShotokanZH> is it possible to force chroot only on sftp and leave it out of it if connecting via ssh?
15:44 < Psi-Jack> sftp runs via ssh, so.. no.
15:45 < ShotokanZH> *sighs*
15:46 < jhodrien> But then why would you?
15:46 < jhodrien> If they can ssh in and access outside the jail, it's not really a jail.
15:46 < Psi-Jack> ^
15:46 < Psi-Jack> Now, care to explain the /real/ problem you're trying to solve?
15:46 < stevendale> Am drinking Apple & Blackcurrent juice
15:46 < stevendale> It's nice
15:47 < ShotokanZH> jhodrien, actually that's untrue
15:48 < Psi-Jack> ShotokanZH: No, what he said is completely true.
15:48 < peetaur2> ShotokanZH: you can sorta do it with apparmor
15:48 < ShotokanZH> jhodrien, i have a device, think at it as a linux router, and i have given to an admin user a limited shell to enable wifi, reboot the device and do little things
15:49 < ShotokanZH> jhodrien, the user can't do anything else except those commands i created for him (i'm using python cmd module)
15:49 < jhodrien> Can't do anything except those commands because you're using ssh keys to enforce this, or some other mechanism?
15:49 < ShotokanZH> but i want the user to be able to connect via sftp to upload a firmware
15:50 < ShotokanZH> jhodrien, he's presented with a custom shell written in python, he's not given the full bash shell
15:50 < Psi-Jack> Hmm. In that case, why not tftp?
15:50 < jhodrien> Provide them with two usernames?
15:51 < ShotokanZH> Psi-Jack, the client is requesting sftp..
15:51 < ShotokanZH> jhodrien, i would do that, but there can't seem to be a way to disable sftp for a single user without breaking everything
15:52 < ShotokanZH> jhodrien, also, it's not a so elegant solution
15:52 < jhodrien> Match
15:52 < jhodrien> Oh sorry, I see your point.
15:53 < ShotokanZH> jhodrien, also you can't specify any Subsystem option in Match
15:53 < jhodrien> Does sftp work with your overridden shell then?
15:53 < ShotokanZH> jhodrien, no sadly
15:53 < drzacek> Hello there
15:53 < drzacek> I'm having problem with kernel module
15:55 < drzacek> I want to load a module on the boot time. I put the module name in /etc/modules, but it doesn't seem to work. In dmesg I don't see any message regarding the module, but lsmod doesn't show it as active
15:55 < post-factum> drzacek: distro?
15:55 < jhodrien> Then I'm confused.  sftp doesn't work with the custom shell, and you can force a user to only use sftp via Match.
15:55 < drzacek> debian
15:56 < jhodrien> Doesn't that get you what you want with two usernames?
15:56 < post-factum> drzacek: what's in /etc/modules?
15:57 < post-factum> drzacek: iirc, anyway, it is advised to use /etc/modules-load.d/ instead nowadays
15:58 < drzacek> post-factum, /etc/modules-load.d/modules.conf is a symlink for /etc/modules (or other way around)
15:59 < drzacek> uio_netx custom_dpm_addr=0xB0000 custom_dpm_len=0x4000 custom_irq=0
15:59 < ShotokanZH> jhodrien, no i'm sorry i confused your question, sftp does work with the custom shell
15:59 < peetaur2> drzacek: I think that's where after initramfs is done and you get to the normal rootfs, it should load those.... and I don't know how that works anymore (depends on distro?) but what I know works is if you load it in the initramfs instead...put it in /etc/initramfs-tools/modules instead, and then run update-initramfs -u
15:59 < post-factum> drzacek: ok, but then you've got confused with modprobe.d
15:59 < ShotokanZH> jhodrien, so the user admin can use the limited shell and the 'unlimited' sftp (and that's bad)
15:59 < jhodrien> But you could use ForceCommand or something to railroad him a different way?
16:00 < post-factum> drzacek: /etc/modules should contain the list of modules only, and /etc/modprobe.d contains those options
16:00 < peetaur2> oh yeah that line looks like a modprobe.d line, not modules
16:00 < drzacek> now this is the answer I was looking for
16:00 < drzacek> I knew there was something with the params, because it works when I start without parameters
16:00 < post-factum> ShotokanZH: hmm, your nick has reminded me something… are you from UA?
16:00 < ShotokanZH> post-factum, nope sorry, europe
16:01 < post-factum> ok then
16:02 < ShotokanZH> jhodrien, yes but then i'd still be stuck at forcing the usage of two users
16:03 < Gringonar> triceratux i get zsh: ./Info: /bin/bash^M: bad interpreter: No such file or directory
16:03 < jhodrien> That's because it's a dos file format file.
16:03 < jhodrien> dos2unix
16:03 < Gringonar> i named it Info.sh
16:03 < post-factum> no matter how you name it…
16:04 < triceratux> control-m ?
16:05 < triceratux> that definitely looks like a squirreled shell environment. youll never get video or wifi to work if you cant run scripts
16:06 < peetaur2> Gringonar: dos2unix Info.sh
16:06 < Gringonar> im sure im using syntax wrong
16:08 < Gringonar> there is something in vim that sets the script to to unix ":set fileformat=unix" cant find it for nano im not sure if arch even has vim preinstalled
16:08 < Gringonar> it seems a farmatting error wth carage returns
16:08 < Gringonar> formatting*
16:08 < peetaur2> any distro that doesn't have vi or vim is silly .... sometimes they just have vi (which is usually even actually vim)
16:09 < Gringonar> dos2unix is preinstalled?
16:10 < peetaur2> no, usually not
16:11 < peetaur2> maybe tr can do it....   tr -d '\r' Info.sh > Info2.sh
16:11 < drzacek> okay, so I put this line "options uio_netx myparams=myvalues" in /et/modprobe.d/myfile.conf, didn't work either :(
16:11 < Gringonar> oh i read something like that didnt know what it meant
16:11 < peetaur2> *maybe tr can do it....   tr -d '\r' < Info.sh > Info2.sh
16:12 < peetaur2> dos is \r\n and linux is \n ....so deleting \r is almost always the same
16:12 < Gringonar> What does  tr -d '\r' Info.sh > Info2.sh""do?
16:12 < peetaur2> (and dos2unix will butcher files based on similar assumptions, so it doesn't mean it's worse ;))
16:12 < peetaur2> it fails because apparently tr doesn't want a file in args.... so add another < in there
16:12 < Gringonar> its a formatting on the fly thing?
16:13 < peetaur2> says delete CR / carriage return...and tr doesn't write the file, instead writes on terminal...so > will write it to another file
16:13 < peetaur2> tr is a simple "filter" which is a program that is barely useful on its own, but can be part of more complex scripts....you pipe a bunch of things together and you can shape data formats
16:14 < peetaur2> a filter is a superior way to do things compared to reading a whole file before processing anything ...makes piping many commands together very sane
16:14 < Gringonar> tr means trim?
16:14 < peetaur2> translate
16:14 < Gringonar> ok
16:15 < ShotokanZH> Gringonar: dos2unix Info.sh
16:15 < ShotokanZH> or tr -d $'\r' < infile > outfile
16:15 < peetaur2> he indirectly said he doesn't have dos2unix
16:17 < Gringonar> I havent installed any packages since it gives odd errors
16:18 < Gringonar> tric recommanded to use that script to probably get some information that allows me to determine why
16:19 < Gringonar> i have to add $?
16:19 < stevendale> You're an odd error
16:19 < stevendale> xD
16:20 < peetaur2> the $ is needed for programs that don't have their own escapes... not tr
16:20 < Gringonar> i can live with that comment
16:21 < Gringonar> it says "Only one string may be given without squeezing repeats"
16:22 < peetaur2> what is "it"?
16:22 < Gringonar> Try tr --help for more information
16:22 < Gringonar> tr probably
16:23 < peetaur2> sounds like busybox
16:24 < peetaur2> is it archlinux? then probably not busybox
16:24 < Gringonar> I should try "tr -ds '\r' Info.sh > Info2.sh""?
16:24 < peetaur2> debian based tends to torture their users with busybox
16:24 < Gringonar> yes Arch
16:24 < peetaur2> no....that was my first one with the error...you need a < in there
16:24 < peetaur2> tr -ds '\r' < Info.sh > Info2.sh
16:25 < peetaur2> and don't know about the -s
16:25 < Gringonar> its the squeeze reoeats option
16:25 < peetaur2> or if you prefer   cat Info.sh | tr -d '\r' > Info2.sh
16:25 < Gringonar> -d deletes -s squeezes repeats
16:25 < peetaur2> I think it's the <
16:26 < peetaur2> forget the squeeze thing... it's because you gave it -d plus 2 args... '\r' and 'Info.sh' but it wants just 1 ... with < it's not an arg any more
16:26 < Gringonar> ok the extra < did not give an error
16:26 < Gringonar> ty for helping with silly syntax stuff :)
16:27 < Gringonar> Unfortunatly it didnt fix the script
16:28 < Gringonar> ill do some more googling
16:28 < peetaur2> you could also just rewrite the script by hand
16:28 < peetaur2> then it won't have CRLF
16:29 < Gringonar> ok guess ill try that
16:31 < TyrfingMjolnir> Is there a way to manually trigger a script the way crontab triggers a script with output to stdout and stderr?
16:33 < triceratux> Gringonar: that original zsh message you got is really weird. there shouldnt be any bash or ^m involved in any of this. i was able to copy/paste that text from the web & execute it no problem
16:33 < triceratux> Gringonar: so thats starting to look like a separate issue. you may need to just get the video / wifi working
16:34 < peetaur2> TyrfingMjolnir: if you are troubleshooting cron, best to just change the time/copy the line with another time
16:35 < Gringonar> to get the video working i have to install the nvidia driver and remove nouveau?
16:36 < Gringonar> cause nouveau is buggy enough to cause lspci to hang
16:36 < triceratux> Gringonar: iirc thats what Dagmar was telling you the other day. what happened when you tried ?
16:36 < peetaur2> I doubt even nouveau, the lousiest of lousy drivers, can cause lspci to hang
16:37 < triceratux> https://superuser.com/questions/1299730/arch-linux-install-lspci-call-hangs
16:37 < azarus> nouveau isn't lousy because the devs are lousy, but because nvidia is a dick
16:39 < tw> Is it possible to make a unix socket only rw by descendent processes and processes in the creator's session? I looked at netns, but I only want to limit the one socket.
16:39 < triceratux> actually theres so much of this im glad i run intel graphics https://www.google.com/search?q=nouveau+lspci+hang
16:40 < tw> ignore the second half of that--just descendent processes would be sufficient.
16:41 < peetaur2> still trying to disbelieve nouveau can hang lspci...how is that possible? :D
16:42 < Gringonar> i got errors about aer
16:42 < Gringonar> i disabled it with pci-noaer but not yet
16:43 < triceratux> peetaur2: isnt that grizzly ? i discovered that working on this install. the solution is always to tidy up the drivers or just blacklist nouveau & so noones fixing the problem
16:44 < noonien> for some reason a script created a really long path, and `rm -rf` doesn't seem to be able to remove it, i'm given a "File name too long" error
16:44 < noonien> how can i get around this?
16:44 < noonien> pwd
16:45 < noonien> oops
16:45 < noonien> it appears to be 8078 characters long(including path separators)
16:47 < peetaur2> does it say  bash: ....  or mv: ... in the error?
16:47 < peetaur2> if it's bash, then don't use a glob
16:47 < peetaur2> or a long name
16:48 < peetaur2> but what can that mean....write a python script to move it?
16:48 < peetaur2> or you can remove the directory instead
16:48 < Gringonar> im no gud at pacman maybe but pacman -S Nvidia, should work dont it?
16:48 < Psi-Jack> "gud?"
16:49 < noonien> it says `rm: cannot remove '...': File name too long`
16:49 < noonien> i'm trying to remove it, not move it
16:50 < peetaur2> delete parent dir, or try some tool like ncdu or midnight commander, or write a python script or C program to delete it
16:50 < e36freak> O.o how long is this name
16:51 < peetaur2> if all else fails, I'm sure debug2fs can do it, but it might leave some side effects
16:51 < e36freak> are we talking like ARG_MAX? because heh
16:51 < noonien>  i'm not passing it as an argument
16:51 < e36freak> try unlink
16:51 < peetaur2> e36freak: he said 8078 chars including path separators
16:51 < e36freak> oh
16:51 < peetaur2> noonien: why do path separators matter? cd there first
16:51 < e36freak> can you cd into the dir and not give the full path
16:51 < Gringonar> I dont get it it keeps say target not found no matter what a Querry
16:52 < Gringonar> saying... I*
16:52 < noonien> ok, so i got `foo/bar/bar/bar/bar/bar/bar/bar/bar/...`, i'm in `foo` and doing `rm -rf bar`
16:52 < peetaur2> for such a trashy looking dir, I'm sure rm -rf --one-file-system foo  is fine
16:53 < peetaur2> (without --one-file-system, maybe it'll follow another mount, like a bind mount to / and delete more than you want)
16:53 < noonien> i tried going as deep as in can, and doing a rm -rf bar, that didn't work either
16:53 < noonien> i'll try one filesystem
16:53 < e36freak> same error?
16:53 < peetaur2> adding one filesystem will not make it delete more or unfail
16:53 < noonien> yeah, same error
16:53 < e36freak> find bar -depth -delete
16:53 < peetaur2> maybe try find....      find -xdev foo -delete
16:54 < e36freak> actually -depth might be implied there
16:54 < peetaur2> (where -xdev is like --one-file-system)
16:54 < noonien> hehe, find fails also
16:54 < peetaur2> (in most commands, this is simply -x, but these have long ugly versions of that)
16:54 < e36freak> with what error?
16:54 < noonien> it fails with file name too long, then with directory not empty
16:54 < peetaur2> maybe mv will help.... cd foo; mv bar/bar bla; rmdir bar
16:54 < noonien> hmm
16:54 < noonien> just a sec
16:54 < peetaur2> then   mv bla/bar bluh; rmdir bla
16:55 < peetaur2> could also use *
16:55 < peetaur2> mv bla/* bla2; rmdir bla; mv bla2/* bla3; rmdir bla2; ....
16:56 < peetaur2> er no that isn't right either if * matches more...but you get the idea I think
16:56 < peetaur2> mv bar deleteme; mv deleteme/* ./; rmdir deleteme
16:57 < peetaur2> which fs is that btw? the shell is basically supposed to support everything the fs does except long *commands*, not *paths*
16:57 < peetaur2> so I'm guessing it's one of those mega filesystems like zfs or btrfs
16:58 < triceratux> https://computingforgeeks.com/easiest-way-to-install-nvidia-3d-graphics-acceleration-driver-on-archlinux/ ?
16:58 < noonien> ext4, but it's in a docker container, don't think that matters though
16:58 < noonien> haha
16:58 < noonien> i cd-ed to the last possible repetion, and rmdir seems to work
16:59 < noonien> rm -rf does not though
16:59 < noonien> `while rmdir foo; do cd ..; done` should do the trick
17:00 < triceratux> wut ? "Note: Based On wiki, propietary driver not work on EFI system. Only Bios that work." http://www.archlinuxuser.com/2013/01/installing-nvidia-driver-on-archlinux.html
17:00 < peetaur2> there are bars and foos
17:00 < noonien> wohoo, fixed
17:01 < peetaur2> just buy AMD
17:01 < triceratux> just run swagarch
17:01 < peetaur2> in the past radeon worked fine, fglrx not so well... now amdgpu replaces both and works fine
17:05 < luke-jr> peetaur2: depends
17:31 < halftroll> Ok, is it possible that I put 127.0.0.1 facebook.com www.facebook.com in /etc/hosts and when I enter into facebook using my browser it leads me to the real facebook instead of localhost?
17:31 < Naan> hey if my spotify doesn't have a spotify.desktop entry how can i change what options it loads with everytime i click it from my taskbar on ubuntu (using cinnamon)
17:33 < ayecee> halftroll: yes
17:39 < searedvandal> halftroll, what are you trying to do? block facebook ads?
17:42 < Naan> oh I can use an alias
17:46 < Gringonar> Whatever i try to install it gives could not extract something not on everything tho. is that normal?
17:46 < Naan> poggers i fixed it
17:47 < BenderRodriguez> question
17:47 < BenderRodriguez> vmlinuz that's the kernel binary itself right
17:47 < BenderRodriguez> that's being decompressed and then executed as the first process on the OS
17:47 < BenderRodriguez> ?
17:47 < azarus> BenderRodriguez: vmlinuz is usually the kernel image
17:49 < azarus> the kernel itself is not PID 1
17:49 < azarus> usually /sbin/init is
17:49 < triceratux> Gringonar: it depends. are you still running the livesession or are you booting from the results of an "install" ? if the former you may have insufficient resources to get pacman to process all packages
17:50 < ][_R_][> BenderRodriguez: The kernel itself isn't a process, init is the first process.  Though certain parts of the kernel will spawn internal processes (they're the ones listed with [] around them)
17:50 < BenderRodriguez> azarus: but isn't the kernel image that's being executed and spawning off child processes itself a process?
17:50 < BenderRodriguez> so technically that would make init the second process
17:50 < BenderRodriguez> and the kernel the first?
17:51 < azarus> /sbin/init is the first process started by the kernel, after it has been loaded into RAM.
17:51 < triceratux> BenderRodriguez: the kernel is just a binary blob. its not even loaded by the runtime linker like all other processes. its just spooled into memory by the bootloader & given control
17:52 < Pusteblume> It's me again.
17:52 < tw> BenderRodriguez: Terminology problem there. process is used to specifically refer to userspace. there are a number of kernel threads executing prior to pid 1 but those aren't given pids.
17:52 < halftroll> searedvandal: I am trying so that I put www.facbeook.com in my browser and it leads me to localhost insteand of real facebook
17:53 < searedvandal> halftroll, that's the opposite of what you said before. but yeah, if you put that in your hosts file, facebook.com should lead you to localhost.
17:54 < Gringonar> yes live
17:55 < Gringonar> how can i tell if it has insicient resources?
17:55 < Gringonar> it has 16GB memory
17:56 < Pusteblume> something i have not seen in comments for ages: DJB's qmail. does that mean its dead?
17:56 < Gringonar> insufficient*
17:56 < WhiteDevil> so should i go back to Cprogramming or learn python
17:57 < WhiteDevil> C programming means everything is basiclly more tedious
17:57 < WhiteDevil> not so much but alot more than python i think
17:57 < Pusteblume> there is a ##programming channel where people discuss this all the time
17:57 < azarus> WhiteDevil: what do you want to accomplish?
17:57 < WhiteDevil> i want to write security tools
17:57 < WhiteDevil> exploits
17:57 < azarus> ah, count me out
17:58 < triceratux> Gringonar: its probably trying to download to /var/cache/pacman/pkg which is probably allocated in ram. but that does sound like a lot of ram even for pacman so there may be something else up
17:58 < WhiteDevil> but i also wanna do normal programming
17:58 < WhiteDevil> gui with C is hard
17:58 < WhiteDevil> but then i am thinking windows platform
17:58 < WhiteDevil> maybe on linux GUI programming is easier in C
17:58 < Pusteblume> looks like the person really really wants to talk about it here
17:58 < Gringonar> which log would tell me something?
17:59 < Gringonar> i think dmesg isnt show it
17:59 < Gringonar> Showing*
17:59 < triceratux> Gringonar: your advantage is that youre not running antergos or manjaro but rather bonafide arch. what happens when you ask in #archlinux ?
17:59 < Gringonar> ill dig some more through it
18:00 < Gringonar> i guess i can try now that its getting more specific
18:00 < Naan> hi do I have to make a desktop entry if I want to run an application with a specific argument every tome
18:00 < Naan> time*
18:00 < Gringonar> sorry to bother :(
18:01 < Naan> I'm using ubuntu and cinnamon
18:01 < Gringonar> i dont wanne flood multiple channels lol
18:01 < triceratux> Gringonar: theres a /var/log/pacman.log. this is a ton of fun actually but at some point it may be so arch-specific that only they will have a clue
18:09 < Naan> never mind I found the desktop entry it was in some snap location
18:10 < Gringonar> i looked at pacman logs
18:10 < Gringonar> they dont tell too much other than it failed
18:10 < Gringonar> let me try looking at the new one
18:26 < foo> I do grep on a ASCII file and it says "Binary file my.log matches" ... but file my.log shows  ASCII text, with very long lines ... du -sh my.log shows 34MB. Any reason for this?
18:26 < ayecee> yes
18:26 < ayecee> probably to do with those long lines
18:26 < ayecee> binary file detection is heuristic
18:26 < foo> ahh. anyway to get around this, I'll check man page
18:27 < ayecee> i think there's an option in the manpage to treat it as non-binary regardless what grep thinks
18:27 < ayecee> ah, -a
18:28 < foo> ayecee: "Treat all files as ASCII text." - ahh. I was searching for "binary" - great, thank you!
18:40 < ayecee> heh. so did i, and i was lucky that the -a option was immediately before the --binary-files option
18:42 < ayecee> document peripheral vision is handy
18:44 < foo> agreed, thank you ayecee !
18:47 < nekoseam> are there any good networkmanager alternatives?
18:50 < kubast2_> nekoseam netctl works well for cli for me
18:50 < kubast2_> Has no applets but it has one od the better wifi cli tools
18:51 < kubast2_> How to click sysrq?
18:51 < kubast2_> Tried alt+prntscrn+r and alt+prntscrn r
18:51 < BenediktXVII> Hello everybody. I need help to setup my servers maintenance tasks for when I am on holidays. Stuff like synchronization of sessions etc... Can I do that with the at command ?
18:52 < ayecee> i don't see why not
18:53 < triceratux> oh nooooes the dedoimedo guy had problems with mtp on fedora rofl https://www.dedoimedo.com/computers/fedora-28-kde.html
18:54 < BenediktXVII> In fact i am confused about the date syntax. Like I want to run a job on friday the 12th at 5pm. How would I write the date and time within the command for that ? I am looking at the man page, but it's confusing.
18:54 < ayecee> at "friday the 12th at 5pm"
18:54 < ayecee> believe it or not
18:55 < nekoseam> Manjaro is absolutely rotten...I'm getting GTK errors everywhere. I guess taking the easy way (xfce + arch) isn't the best option. Perhaps I should go with normal Arch + openbox or fedora netinst
18:55 < BenediktXVII> nice
18:55 < ayecee> there's more formal ways to write it, but it can figure out what you mean most of the time
18:56 < MrChinasky> Ok guys i think i really fucked up here. I've set a bad mountpoint to my external harddrive in my /etc/fstab/. Now my / fs is in Read Only, and i can't edit back fstab. Runnig 'sudo mount -o remount,rw /' returns 'mount: can't find PARTUUID=90a83158-560d-48ee-9de9-40c51d93c287'. Am I screwed?
18:56 < nekoseam> Their GUI program for managing kernels and such completely shit on me after a single update (including the falkon browser. Besides that the release cyccle is horrid. You never know when a large update is coming for you to prepare for
18:56 < ayecee> mind the language please
18:57 < MrChinasky> ayecee: :O ! My bad.
18:57 < UserUS> nekoseam, no. You can use recovery mode to mount over the drive and redirect the UID path to it's original state'
18:57 < ayecee> MrChinasky: could comment out the line for / in fstab, then run the command again
18:58 < nekoseam> UserUS: ?
18:58 < ayecee> the problem there is that fstab doesn't reflect what's mounted at / currently
18:58 < MrChinasky> ayecee: I'd love to, but all files are now Read Only :/
18:58 < ayecee> ah right
18:58 < UserUS> necrogami, sorry that was meant for mrchinasky
18:59 < MrChinasky> UserUS: sounds painfull ahahah. Any sources to share?
18:59 < ayecee> MrChinasky: what does mount say is mounted on / now?
18:59 < nekoseam> ayecee: I'm not exactly sure how the word "sh*t" could be taboo in 2018...
18:59 < nekoseam> but alright
19:00 < BenediktXVII> ayecee, it says invalid date format
19:00 < ayecee> nekoseam: it's not so much the actual word but the emotive content without information
19:01 < ayecee> BenediktXVII: eh. 2018-07-12 5pm then
19:01 < MrChinasky> ayecee: https://pastebin.com/kPxAWE2t
19:01 < UserUS> MrChinasky, haha it is a bit but not too bad. http://www.sudo-juice.com/how-to-change-the-uuid-of-a-linux-partition/
19:01 < ayecee> MrChinasky: okay, now what does mount say is mounted?
19:01 < MrChinasky> UserUS: <3
19:01 < ayecee> :P
19:01 < ayecee> eh, i guess this works.
19:02 < ayecee> mount /dev/mmcblk0p2 / -o remount,rw
19:02 < ayecee> mount also has an option to use an fstab file at a different location, fwiw.
19:03 < UserUS> Can't the grub repair tool can help fix a messed up fstab file too?
19:04 < MrElendig> no
19:04 < ayecee> can it? doesn't seem like something a grub repair tool would do
19:04 < MrElendig> your $EDITOR will fix that
19:04 < UserUS> I just remember fixing this issue starting with the grub repair tool, then redoing the uuid
19:04 < UserUS> idk maybe I messed multiple things up at the time haha wouldn't surprise me
19:04 < ayecee> heh
19:04 < triceratux> nekoseam: arch is the best. but you need something like swagarch to make it work https://swagarch.gitlab.io/
19:08 < ayecee> nekoseam: hehe, so i guess one could say that that's a shitty opinion.
19:14 < nekoseam> triceratux: I'll probably try Fedora netinst since it's easy to install but plops you into the CLI where it's relatively easy to install what you need and is an independent distro without any external repositories
19:14 < MrChinasky> ayecee: You're the real MVP! Thanks
19:15 < nekoseam> I believe dnf is written in pyth*n though
19:16 < BenediktXVII> Thanks a lot for everything. Have a nice evening. :)
19:16 < MrElendig> which is not a problem
19:16 < triceratux> f29 will see the final demise of yum ftw
19:16 < ayecee> we hardly knew ye
19:17 < nekoseam> Portage - slower than basically every other source based package manager, written in python. dnf - slower than basically every other package manager, written in python. Hmm...
19:17 < nekoseam> binary package manager*
19:18 < MrElendig> nekoseam: it is not because it is written in python
19:18 < MrElendig> (and there are plenty of slower package managers)
19:18 < nekoseam> Well speed tests have shown that Python is overall slower than C
19:18 < MrElendig> nekoseam: most of the time is spent doing IO
19:19 < ayecee> how much slower do those speed tests say it is
19:19 < ayecee> is it less than 1%
19:19 < MrElendig> dnf is slow due to design, not the language it was implemented in
19:19 < ayecee> or are the speed tests fictional
19:19 < MrElendig> ayecee: anything from 5 to 5 million times slower than C, depending on how you tailor the test to make python look bad :p
19:20 < ayecee> 5 times slower, that's 80% slower?
19:20 < ayecee> shocking
19:21 < nekoseam> Also give me 1 binary package manager that's slower than dnf. I've done direct speed tests with roughly  the same mirror speed of most popular package managers and dnf is slower than ALL of them
19:21 < MrElendig> yum :p
19:21 < MrElendig> that suse thing
19:21 < nekoseam> Portage is slower than sourcemage's package manager, void's package manager, crux's package manager, ect.
19:21 < MrElendig> whatever it is called
19:21 < triceratux> zypper
19:21 < nekoseam> But that could be justified by its functionality
19:21 < MrElendig> the old one
19:22 < MrElendig> anyway, dnf really is slow due to all the work it does, not because of the language
19:22 < MrElendig> funnily most code is like tha
19:22 < MrElendig> even C code
19:22 < ayecee> interpreter apologia
19:23 < nekoseam> mmhm...yet zypper has the same functionality as dnf except for an unused dependency cleaner and is around 7 seconds faster...
19:23 < MrElendig> also
19:23 < MrElendig> dnf is mostly in c++ now?
19:23 < nekoseam> I believe it's mostly written in Python with some C slowly coming in. Not entirely sure
19:23 < MrElendig> nekoseam: you would have to test with the exact same db, at the same state of cache etc
19:24 < ayecee> only if you wanted a meaningful comparison
19:24 < ayecee> who has time for that
19:24 < MrElendig> well comparable db anyway
19:24 < ayecee> as long as the results justify my preconceptions, that's good enough.
19:24 < MrElendig> both of local and remote packages
19:25 < bls> my building is going to blow up if I can't install this package in under 15 seconds! my package manager must be performant!
19:25 < MrElendig> a bigger issue with dnf than speed is the ram useage
19:25 < nekoseam> Never done memory tests
19:26 < nekoseam> How much does it use?
19:26 < MrElendig> it scales really bad with db size
19:26 < bls> to the point the OOM kicks in every time you have to install something?
19:26 < ayecee> hate it when that happens
19:27 < MrElendig> there are some fancy graphs, let's see if I can find them again
19:27 < MrElendig> bls: in limited containers/sbc/etc yes
19:27 < bls> now that's a legit gripe then
19:28 < MrElendig> urgh can't find the mem ones, but this is semi-relevant: https://rpm-software-management.github.io/announcement/2018/03/22/dnf-3-announcement/
19:28 < MrElendig> when it comes to the raw speed
19:30 < triceratux> https://lwn.net/Articles/750238/
19:31 < system16> FUCK
19:31 < MrElendig> that points out why dnf is slow
19:31 < ayecee> :o
19:31 < MrElendig> it syncs the db on *every* command
19:31 < system16> i accidentally erased my ubuntu
19:31 < system16> wow
19:32 < searedvandal> how does one do that on accident?
19:32 < system16> 3 hours of hard work
19:32 < MrElendig> one of the reasons for why it is slow anyway
19:32 < system16> GONE
19:32 < ayecee> searedvandal: right-click delete vm
19:32 < MrElendig> system16: just restore from your backup
19:32 < bls> heh, they actually considered rust
19:32 < system16> i dont have a backup
19:32 < system16> its a usb
19:32 < searedvandal> should go quicker second time around
19:32 < system16> i installed ubuntu on a usb drive
19:32 < system16> wow
19:32 < system16> im such an idiot
19:32 < searedvandal> ayecee, that's a pretty deliberate act
19:33 < ayecee> thought it was a different vm
19:33 < system16> i thought im formatting my other usb drive
19:33 < ayecee> is the accident part
19:33 < ayecee> see, pretty similar
19:33 < searedvandal> system16, just start all over again
19:33 < searedvandal> since you've already done everything once, it shouldn't take you 3 hours this time
19:33 < system16> searedvandal, lol its not that EZ
19:34 < system16> easy*
19:34 < searedvandal> it should be
19:34 < system16> i cant un-format a dive right ?
19:35 < ayecee> depends how you formated the dive
19:35 < system16> quick format
19:35 < system16> fat32
19:35 < luke-jr> also depends on what it was before
19:35 < system16> ext4
19:36 < ayecee> you could maybe unformat the dive in that case, but there isn't an easy way to do it.
19:36 < searedvandal> lots of different tools you can try, but my guess is that none of them would be any faster than reinstalling
19:36 < system16> so reinstalling ubuntu on it is easier ?
19:36 < ayecee> much easier
19:36 < ayecee> also pretty cheap life lesson tbh
19:36 < luke-jr> if you have no personal data to rescue, definitely prefer a reinstall
19:36 < searedvandal> installing ubuntu is pretty straight forward
19:37 < luke-jr> if you DO have personal data to rescue, clone the drive first
19:37 < system16> yes. but not on a usb drive
19:37 < luke-jr> (and next time remember to make backups)
19:37 < system16> i cant clone a usb. can i ?
19:37 < ayecee> sure, why ot
19:37 < ayecee> same as cloning a hard drive
19:37 < system16> windows wont recognize a ext4 usb drive
19:37 < system16> oh like clonezilla ?
19:38 < ayecee> yes, or even imageusb in windows
19:38 < searedvandal> a usb drive isn't much more than a storage device in a different packaging
19:38 < Dagmar> ...and being very removable, there's not much reason to use ext4 on them
19:38 < MrElendig> they are more simplistic than some
19:38 < markasoftware> how does Clonezilla compare to any live CD with `dd` available?
19:38 < ayecee> favorably
19:39 < system16> Dagmar, ubuntu cant be installed on a NTFS file system. it must be ext4
19:39 < system16> afaik
19:39 < Dagmar> I wasn't talking about NTFS
19:39 < ayecee> so what were you talking about
19:40 < Dagmar> exFAT
19:40 < system16> exfat ?
19:40 < ayecee> come on man
19:40 < eletious> oh lord
19:40 < MrElendig> good luck using that for /
19:40 < Dagmar> But you're already apparently phoning it in by using Ubuntu in the first place
19:40 < MrElendig> (ntfs would actually be better)
19:40 < Dagmar> Why even bother installing it
19:40 < system16> ubuntu cant be installed on that
19:40 < MrElendig> use f2fs or ext4
19:40 < markasoftware> probably not possible through the graphical installer, but it is certainly possible to install Linux under NTFS
19:41 < ayecee> does ubuntu's installer do f2fs?
19:41 < MrElendig> markasoftware: with some limitations
19:41 < system16> Dagmar, i use it when $hit hits the fan
19:41 < MrElendig> would have to set up user mapping etc
19:41 < MrElendig> and not use caps
19:41 < system16> like salvaging files from a destroyed OS etc .
19:41 < ayecee> and hope that no one found out
19:42 < luke-jr> it makes far more sense to just use ext4 on the USB drive.
19:42 < searedvandal> if its a usb stick, just create a liveusb with persistence
19:42 < ayecee> just
19:42 < system16> searedvandal, i did that
19:42 < system16> but now its gone
19:45 < MrElendig> f2fs can be beneficial on crappy drives with bad wear leveling
19:46 < searedvandal> system16, use the linuxlive usb creator if you're in windows, that lets you set persistence. quick and easy.
19:47 < system16> https://www.linuxliveusb.com/ ?
19:48 < bls> why not just perform a regular install?
19:48 < searedvandal> system16, that's the one. a quick and easy tool for setting up a persistent ubuntu liveusb in windows
19:48 < ayecee> why not just print more money
19:48 < system16> thats an option too
19:48 < ayecee> (google's first autocomplete for "why not just")
19:48 < searedvandal> bls, sure, that's what I would do. I'm just providing him with options.
19:48 < system16> last time I clicked on something else and I selected my usb drive as a target for installation
19:49 < bls> don't really understand how LiveCD with persistence on harddrive/USB didn't die and instead morphed into LiveUSB with persistence on the same USB
19:49 < searedvandal> there are a multitude of ways to achieve the goal of having a persistent live usb stick
19:50 < ayecee> which one to just use though
19:50 < system16> thanks
19:50 < searedvandal> I have a multi boot usb which loops the isos I need, that's my solution.
19:50 < ayecee> how did you just set that up
19:51 < searedvandal> ./createmultiboot.sh
19:51 < ayecee> XD
19:51  * triceratux queues up a proactive image copy of his principal liveusb
19:52 < searedvandal> if you wanna see fancy multiboot grub.cfg, ask triceratux
19:52 < system16> my friends say that installing an OS on a usb WILL increase wear and tear
19:52 < searedvandal> he got a pretty sexy grub.cfg
19:52 < system16> and thats notgood
19:52 < ayecee> oh no
19:52 < ayecee> you might have to eventually replace it
19:52 < searedvandal> usb sticks cost like next to nothing these days?
19:52 < bls> how is live USB going to be any better or worse than a normal install?
19:53 < ayecee> less writes, probably
19:53 < searedvandal> probably not
19:53 < triceratux> searedvandal: youre very kind. i did it all by hand so im not familiar with todays more elaborate tools. i know unetbootin is a mistake at all times tho
19:53 < bls> what if it does more writes? or is there a how many writes switch that only live distros know how to flip?
19:53 < searedvandal> triceratux, I did mine manually as well, but I only keep a handful of isos on it, so it's far from an achievement compared to yours
19:54 < searedvandal> I read some documentation and the arch wiki page on it and went from there
19:56 < triceratux> my grub.cfg is largely based on stuff from the archwiki as well. but they were doing a better job documenting distros other than their own so they scaled it way down
19:56 < wm4> how do you make usb keyboard NKRO work on Linux?
19:56 < ayecee> what is a NKRO
19:57 < wm4> https://en.wikipedia.org/wiki/NKRO#n-key_rollover
19:57 < ayecee> what indicates that it's not working
19:58 < ayecee> or, what should it do when it works
19:58 < eletious> wm4: Typically you get a keyboard that "supports" it and then find out that it only supports it over PS/2, and then you buy a USB -> PS/2 adapter
19:59 < wm4> eletious: fairly sure that's not the case here
19:59 < wm4> ayecee: when I hold down 6 keys, the 7th key isn't registered according to xev
20:00 < eletious> wm4: Did you buy a mech? What keyboard do you have?
20:00 < markasoftware> wm4: that's usually not osmething you do in the OS, but i could be wrong
20:00 < wm4> eletious: HV-KB395L
20:00 < wm4> markasoftware: ?
20:01 < eletious> wm4: I would be absolutely astonished if a havit kb had n-key rollover over USB
20:01 < ayecee> wm4: possibly related - someone reports they have a CoolaerMaster XTi and the nkro works only in windows, because the keyboard itself doesn't really support it and it's emulated by the windows driver.
20:01 < markasoftware> wm4: usually that's just part of the keyboard. It doesn't matter whether you're in Windows, linux, mac, etc.
20:02 < wm4> markasoftware: it could be that Linux is only using the boot protocol
20:05 < searedvandal> as far as I know nkro is over ps/2, with 6kro being the limitation over usb.
20:05 < markasoftware> wm4: did it work under Windows?
20:06 < searedvandal> ah, that is the boot mode
20:06 < bls> and from the link you posted: "To get full n-key rollover, HID report protocol must be implemented on both keyboard and computer." so does both the keyboard and Linux claim to implement the HID report protocol?
20:06 < eletious> it's possible that the kb has a weird driver... that's a common thing i guess
20:06 < markasoftware> oh yeah i remember, doesn't nkro over usb require multiple virtual devices or something?
20:07 < markasoftware> https://linustechtips.com/main/topic/40375-how-does-nkro-over-usb-work/?do=findComment&comment=522367
20:15 < phinxy> I'd like to migrate to a new harddisk because I got a few XFS corrupt sectors.
20:15 < phinxy> Could I just DD the entire thing to the new disk?
20:15 < phinxy> I have no LVM enabled
20:16 < bls> you'd end up copying the corruption
20:16 < phinxy> I have a feeling its corrupt on the video-media-entertainment partition only
20:18 < MrElendig> cp -a
20:19 < MrElendig> sidenote: when you are done, run a smart longtest
20:20 < bls> and mount it read-only while you're running that cp to prevent further corruption until you can try to repair the FS
20:29 < seven-eleven> what do I have to setup so I can retrieve a kerberos ticket from a windows AD to my linux host?
20:29 < zenix_2k2> so first i am not sure if this is the right place to ask this or not but on Linux is there a limitation for how many file descriptors can i have ?
20:30 < Sitri> zenix_2k2: MAX_INT
20:30 < peetaur> zenix_2k2: also /etc/security/limit.conf can limit it
20:30 < zenix_2k2> sitri: you mean echo $MAX_INT ?
20:30 < peetaur> seven-eleven: some pam module thing probably
20:30 < seven-eleven> what I tried: 1) install AD on windows, setup DNS with reverse lookup PTR, enabled winRM with firewall rules 2) added the windows DNS server IP to /etc/resolv.conf, installed kerberos on the linux host, configured krb5.conf like so http://termbin.com/wjhe, then tried `kinit administrator@VM-WIN01.TEST01.`, did I miss something?
20:30 < bls> seven-eleven: have you read over https://access.redhat.com/documentation/en-us/red_hat_enterprise_linux/7/html/windows_integration_guide/introduction  ?
20:31 < Sitri> zenix_2k2: no
20:32 < zenix_2k2> so what do you mean by that ?
20:32 < `Guest00000> in bash, does & apply only to whole pipeline or only one command in a pipeline?
20:33 < Sitri> A file-descriptor is a signed int.  -1 is a value that indicates an error, thus only positive integers are valid FDs.
20:35 < zenix_2k2> wait, so you mean that every file that i create on an OS has its own FD ?
20:35 < seven-eleven> bls, oh I read shortly about that, but I wasn't sure if i really had to directly add the linux host to the windows AD, i thought for a kerberos ticket retrieval the linux host just needs to know about the AD dns server :S
20:35 < bls> oh, so you only want krb tickets
20:35 < sadasaulna> zenix_2k2: open files
20:35 < zenix_2k2> sorry i am a bit new to these FDs stuffs
20:36 < zenix_2k2> sadasaulna: well yes, open files... but when i open a file, how can my program know where is my file ?
20:36 < sadasaulna> no worries, i'm no expert either
20:36 < zenix_2k2> there has to be some sort of ID or flags or something that marks the file's location right ?>
20:36 < zenix_2k2> ?*
20:36 < seven-eleven> bls, yeah
20:36 < sadasaulna> yeah its called an inode
20:37 < seven-eleven> peetaur, i think i need the pam module only when windows AD hosts want to connect to my linux host :S
20:37 < zenix_2k2> so where is that "inode" stored ?
20:37 < sadasaulna> its part of the filesystem metadata
20:37 < zenix_2k2> so i can't read them ?
20:37 < peetaur> seven-eleven: you could look at how to set up samba as a server, and somewhere in there it'll have the linux client config
20:37 < seven-eleven> bls, i get this error when trying to retrieve a ticket: kinit: KDC reply did not match expectations while getting initial credentials
20:37 < sadasaulna> well, yeah, if you look at the raw disk you could eem them
20:37 < bls> seven-eleven: been forever since I used kinit and manually handled the keys, mainly use k5start these days for a key-only setup
20:37 < sadasaulna> see
20:39 < zenix_2k2> sadasaulna: but what is so different between an inode and a file descriptor ?
20:39 < sadasaulna> zenix_2k2: this is how hard links work.  You have two or more files pointing to the same inode
20:39 < ddoobb> Can you keep the same /home when moving to a new distro?
20:39 < bls> ah, so you haven't even gotten to the point of a valid/compatible configuration
20:39 < Sitri> zenix_2k2: FDs are unique per process.  Each process will automatically have FD 0, 1 and 2 (stdin, stdout and stderr respsectively) automatically opened.  inodes are filesystem metadata, you can see them with ls -i, and find out how many are used with df -i.  A pair of hard-linked files will share an inode.
20:39 < seven-eleven> http://dpaste.com/3MRWM7J
20:39 < sadasaulna> zenix_2k2: the file descriptor is the in memory structure I think that you use to do operations on a open file, it is a "handle" if you will
20:39 < sadasaulna> i think that is right, but someone more experience may wish to chime in
20:39 < bls> ddoobb: you can try. as long as the versions of software used aren't wildly different so the config files are incompatible, it should be fine
20:39 < seven-eleven> peetaur, you mean a krb5.conf example?
20:40 < peetaur> I don't know...I never had such a setup, just hear people talk about doing it with samba and using the linux as clients too
20:40 < sadasaulna> and they don't have to be normal files, they could be pipes or sockets
20:41 < zenix_2k2> Sitri: so not that every file has their own inode then
20:41 < seven-eleven> peetaur, yeah that'd be a full integration to the AD which I dont want, and I think it shouldn't be needed if you just want a kerberos ticket :-)
20:41 < sadasaulna> well every file has at least one inode, but a filename is seperate from the file.  So multiple file*names* can point to the same *file*
20:42 < Sitri> zenix_2k2: Every file (as in the data of a file) has its own inode.  Every filename points to an inode that's in use.
20:42 < Sitri> Subtle difference
20:43 < bls> seven-eleven: looks like you're going to have to figure out how to crank up the verbosity client and server side to figure out why either one feels the other isn't configured properly
20:43 < sadasaulna> zenix_2k2: if you want to learn more, the original descriptions of UNIX from back in the 70s are around as PDFs and they do an excellent job of explaining it all.
20:43 < sadasaulna> i'll see if I can find a link
20:44 < zenix_2k2> yea, i think reading one of them will make things clearer
20:44 < zenix_2k2> though it is a bit clearer now :P
20:44 < zenix_2k2> but is this only appliable to Linux ?
20:44 < zenix_2k2> i mean this concept of Inodes and FDs
20:44 < Sitri> Any unix-like
20:45 < sadasaulna> zenix_2k2: completely a UNIX thing, going back to the very first UNIX
20:46 < phogg> I love getting calls from "Windows technical support" trying to tell me I have a problem with my computer.
20:46 < phogg> I always waste as much of their time as possible.
20:46 < zenix_2k2> sadasaulna: anyway have you found the link
20:46 < zenix_2k2> or you can give me some keywords cause i am suck at searching
20:47 < TaZeR> hey guys, is it possible to have the default app used to launch for example .magnet links from the terminal different then the program that would be used otherwise?
20:47 < bls> TaZeR: yes
20:47 < TaZeR> besides just altering the browser to use another program and setting the default as the one u want open by terminal
20:47 < phogg> TaZeR: since a .magnet file is not executable it would not, by default, be executable unless you invoked the program by hand
20:47 < phogg> of course you could set up a binfmt handler for it. That would be fun
20:47 < sadasaulna> you might try searching for a PDF of "The UNIX Programming Environment" but i'm not sure if that is the one or not.  I have a copy somewhere, they talk about the design of the OS and how it works (the original authors of UNIX)
20:48 < bls> oh, I was thinking this was a question about default program for extension/mime type
20:48 < phogg> TaZeR: Wait, what are you asking? First you open the terminal, then you type in the name of the program you want to run and pass it a .magnet file. Right?
20:48 < TaZeR> i use a script for example that searches torrent sites and lets me choose which to download by pressing the letter 0,1,2,3 etc i would like it to open in aria2c instead of my default gui torrent app
20:49 < phogg> TaZeR: if you are clicking on a .magnet file in a file manager and asking about what program will open at that time then it depends on the file manager. Firefox, for example, can be configured with its own set of mime handlers separate from the platform, so for that browser you can have a different handler than you would get from your file manager.
20:49 < bls> zenix_2k2: check out the docs on: http://doc.cat-v.org/unix/
20:50 < phogg> it all depends on exactly what programs you're using (file manager, browser). Tl;dr "maybe"
20:50 < TaZeR> so how can i set the handler for instances of magnet links called by a terminal program?
20:50 < TaZeR> im using guake, and the program is called pirate-get
20:50 < bls> "terminal program" is a bit too generic for a reasonable response
20:50 < fndreic> hello
20:51 < fndreic> How can i get polish signs on linux tty ?
20:51 < TaZeR> and when i tell pirate-get to download a torrent for me which it uses magnet link, i would like the handler to be aria2c
20:51  * ayecee waits for the punch line
20:51 < fndreic> now when i try to do polish sign i get "?"
20:51 < TaZeR> as for browser and file manager the way it is now is fine\
20:52 < zenix_2k2> bls: is that document gonna fully explains everything in UNIX ?
20:52 < bls> fndreic: you can try to load a polish console font, but if there's not one, you're better off launching a graphical environment that has better font support
20:52 < zenix_2k2> or at least some critical parts of it
20:52 < fndreic> bls: nah i wanna using just tty
20:52 < bls> zenix_2k2: no. I really doubt you're going to find a single document that describes every facet of unix
20:52 < fndreic> i remember i had polish signs before
20:52 < fndreic> in tty
20:52 < fndreic> but i forgot how i do that
20:52 < fndreic> before ;p
20:52 < zenix_2k2> bls: well then how about some critical part ? like some usually encountered ones
20:53 < TaZeR> is my question not understandable? should i restate?
20:53 < TaZeR> or nobody knows
20:53 < bls> fndreic: man setfont and see if there's one in the default location for your distro
20:54 < bls> TaZeR: that all comes down to how this pirate-get program works, and I doubt you're going to get much support for it here as it sounds pretty questionable
20:54 < fndreic> bls how to use that setfont ?
20:54 < zenix_2k2> bls: "vital parts" i meant :P
20:54 < TaZeR> its just a program that searches piratebay for torrents...
20:54 < bls> fndreic: no clue, I just start a GUI environment instead of mucking about with console fonts
20:54 < TaZeR> so your saying its based on the way the program calls for the magnet link to open and theres no standard i can set right
20:55 < bls> TaZeR: there's a standard, but that program might not follow it
20:55 < sadasaulna> zenix_2k2: search "Design of the UNIX Operating System" you should see a pdf link a few results down
20:55 < ddoobb> Any distro recommendation for someone who's only used ubuntu?
20:55 < sadasaulna> its an original AT&T doc
20:55 < fndreic> Okay ill try to configure that somehow but. If someone here know how can i load polish signs for tty ill be glad for help
20:55 < TaZeR> ok i see, can i try it anyway and find out? like can i set it to just use aria2c for any .magnet linked opened in guake?
20:55 < esselfe> ddoobb: openSuse
20:56 < TaZeR> while still maintaining my tixati gui program for anything else
20:56 < bls> TaZeR: the closest thing to a standard way is going to be xdg-open, you can configure it and which programs it uses
20:56 < zenix_2k2> sadasaulna: ok then
20:57 < bls> zenix_2k2: the page I linked covers some vital parts
20:57 < TaZeR> ive heard of xdg-open before i remember using it to set a different picture viewer before, i iwill try that thank you bls
20:57 < sadasaulna> zenix_2k2: that should set you up with the fundamentals.  Obviously UNIX and Linux kernels have gotten a lot more sophisticated since then, but a surprising amount of the fundamentals are still relevant today
20:57 < bls> "Program Design in the UNIX Environment" "A Research UNIX Reader: Annotated Excerpts from the Programmer’s Manual" "McIlroy’s pipes"
20:58 < bls> the page also has links to some of the original man pages
20:58 < esselfe> ddoobb: or pick one from the distrowatch.com listing
20:58 < sadasaulna> that should be plenty to be getting on with
20:59 < zenix_2k2> sadasaulna: but are Linux and UNIX quite... closed to each other ? cause i am afraid that i can't do any practice with UNIX because i am using Ubuntu
21:00 < zenix_2k2> and actually the purest UNIX isn't used anymore, they use something that are based on it instead but i am not sure that Linux is UNIX-based or not
21:00 < bls> zenix_2k2: unix is a specification that linux implements a large majority of
21:00 < sadasaulna> zenix_2k2: yes it will be relevant, and its much better to understand fundamentals first
21:01 < zenix_2k2> so Linux is UNIX-based after all ?
21:01 < sadasaulna> zenix_2k2: Its UNIX-like
21:01 < sadasaulna> UNIX is a trademark so they can't call it UNIX per se
21:01 < zenix_2k2> so what is different between "UNIX-like" and "UNIX-based" ?
21:01 < zenix_2k2> they seem quite the same to me
21:02 < koala_man> Linux shares no source code with anything traditionally called UNIX
21:02 < Psi-Jack> Linux doesn't use any of the UNIX code, for starters.
21:02 < sadasaulna> Well for a start it meant you couldnt originally use code from the originally UNIX you had to rewrite it
21:02 < MrElendig> GNU is Not Unix
21:02 < sadasaulna> but because of standards like POSIX and such, they are very alike
21:03 < MrElendig> if gnu/linux actually used unix code they would be sued even worst by sco than they already are
21:03 < hexnewbie> Although GNU/Linux is probably not POSIX-compliant, even when POSIXLY_CORRECT is set
21:03 < zenix_2k2> ok now this is a lot to take in
21:04 < sadasaulna> take your time zenix_2k2 - rome wasn't built in a day etc
21:04 < koala_man> GNU/Linux is mostly POSIX compliant, but is not POSIX certified in any way
21:04 < zenix_2k2> but seriously i still can't tell what is so different between UNIX-liked and UNIX-based
21:04 < zenix_2k2> those seem quite the same
21:04 < MrElendig> https://en.wikipedia.org/wiki/SCO%E2%80%93Linux_disputes
21:05 < zenix_2k2> only the terminology that seems different
21:05 < MrElendig> zenix_2k2: how they are developed is one thing
21:05 < MrElendig> read the cathedral versus the bazaar
21:05 < sadasaulna> In every day use it doesn't really matter much.  Most software written for one UNIX can be ported to another UNIX unless the code is heavily tied into some specific kernel feature
21:05 < zenix_2k2> the what ? it doesn't seem to be a sort of technology's terminology
21:06 < zenix_2k2> sadasaulna: well i do know that it doesn't matter but i am just a bit curious
21:06 < koala_man> zenix_2k2: no, you can't tell the difference in day-to-day use. however, you'd quickly discover the difference if you were licensing and selling it and starting getting cease and desist letters
21:07 < koala_man> it's like clone vs fork
21:07 < sadasaulna> zenix_2k2: if you're curious about UNIX reading those old docs will help, plus try learning some of the history of the BSD projects too, and perhaps even stuff like Minux and GNU/Hurd - all stuff to chew on if you're interested in OSes
21:08 < ddoobb> How can I see which distros are more popular?
21:08 < sadasaulna> distrowatch
21:08 < hexnewbie> zenix_2k2: -like means the two things are alike, -based means one is descended from the other. Like birds are dinosaur-based, but definitely not dinosaur-like. Or like dolphins are fish-like, but not fish-based.
21:09 < esselfe> ddoobb: there are 3 major kind of distros, debian-based, red-hat-based and source-based. The only difference is really from the community they have
21:09 < zenix_2k2> well thank to that analogy, i know that birds are dinosaur-based
21:09 < esselfe> ddoobb: you really have to try them and judge by yourself what suits you best
21:10 < ddoobb> Which category is arch-linux
21:10 < esselfe> ddoobb: in term of software they're all the same
21:10 < MrElendig> https://upload.wikimedia.org/wikipedia/commons/1/1b/Linux_Distribution_Timeline.svg
21:10 < sadasaulna> ddoobb: bleeding edge and expects you to RTFM
21:10 < esselfe> arch is source-based which implies lots of compilation time
21:10 < MrElendig> esselfe: arch is not source based
21:11 < sadasaulna> esselfe: Gentoo is source based
21:11 < ddoobb> Is Manjaro noob friendly?
21:11 < MrElendig> ddoobb: don't use manjaro
21:11 < zenix_2k2> hexnewbie: oh so "based" means that one things is born from "another thing" and "liked" means that one things looks like "another thing" right ?
21:11 < esselfe> I tried arch once and it compiles source code
21:11 < hexnewbie> zenix_2k2: That's the gist of it
21:11 < MrElendig> ddoobb: it has no community and does stupid things like deletes your /lib
21:11 < Sapphirus> Arch is binary. It can be hybrid, depending upon how you configure it.
21:11 < MrElendig> esselfefake news
21:12 < MrElendig> esselfe: fake news*
21:12 < zenix_2k2> well ok then
21:12 < MrElendig> ddoobb: https://forum.manjaro.org/t/usr-lib-modules-getting-deleted-on-boot/49984
21:12 < MrElendig> https://gitlab.manjaro.org/ste74/kernel-alive/issues/1
21:14 < MrElendig> ddoobb: no distro is noob friendly btw, I strongly suggest you stop being a noob
21:14 < MrElendig> ddoobb: https://www.urbandictionary.com/define.php?term=Noob
21:14 < hexnewbie> I've seen backup scripts with that kind of bash scripting bug. It's... fascinating.
21:14 < MrElendig> ddoobb: being a newbie on the other hand is fine
21:14 < triceratux> 80% of users want linux that works, whether theyre n00bs or not
21:15 < drbeavi5> ubuntu, solus, fedora, mint. all pretty basic and easy to install. i prefer void these days but im a fanboy
21:15 < MrElendig> drbeavi5: specially see the "II. Defining 'Noob' " bit
21:15 < MrElendig> er... ddoobb ^
21:15 < sadasaulna> i like void too from what i've seen of it so far
21:15 < ddoobb> drbeavi solus is ubuntu based?
21:15 < ddoobb> drbeavi5
21:15 < drbeavi5> yeah its a pretty cool distro
21:15 < MrElendig> drbeavi5: all distroes these days are easy to install, relatively speaking
21:16 < sadasaulna> but then I hate systemd so i have a pretty limited choice of distro
21:16 < sacules> wasn't solus made from scratch?
21:16 < drbeavi5> i've never actually used solus, ive just heard about it being relatively simple/easy to install, run etc.
21:16 < Psi-Jack> And maintain. Yes. :)
21:17 < Psi-Jack> sacules: Yes it is.
21:18 < Psi-Jack> And yes, I will attest that Solus is a pretty awesome distro.
21:18 < triceratux> Psi-Jack: the dedoimedo guy went completely off on fedora 28 rofl https://www.dedoimedo.com/computers/fedora-28-kde.html "ZOMG, Linux is amazing and everything works. Best thing ever. If it does not work for you, it's your fault. You're an idiot. If it works for you, you are smart. For example, if you plug your phone into the computer, and then open the file manager, and you get a red error, this is your fault. It means you
21:18 < triceratux> haven't read the ideology manifesto. And if you don't want to manually fix these problems, it means you are a noob and you do not understand how it works."
21:19 < Psi-Jack> Umm... Smack some sense into that ingrate.
21:19 < triceratux> lol
21:19 < sacules> lmao what an insecure jerk
21:20 < sadasaulna> he's being sarcastic, whats wrong with that?
21:21 < triceratux> in all fairness he was triggered by the mtp. thatll do it toya. he has yet to find the comforting ways of thunar
21:21 < Psi-Jack> That's.... Not sarcasm. That's complete and total ignominious stupidity.
21:22 < Guy1524> hey guys, in /proc/pid/maps, what are the unlabeled address spaces?
21:22 < sacules> i'm still bothered that thunar doesn't cut long file names
21:22 < ddoobb> Any distro that's both rolling release and stable?
21:22 < MrElendig> arch
21:22 < ultr4_l4s3r> Solus
21:22 < sacules> uuhhhh fedora and void I guess?
21:22 < Psi-Jack> ddoobb: Solus.
21:23 < MrElendig> fedora is just quasi-rolling
21:23 < triceratux> ddoobb: altlinux, tumbleweed, swagarch
21:23 < Psi-Jack> Heh.. I can't even consider swagarch "stable" when they've had gpg key issues for weeks now, unfixed.
21:24 < sadasaulna> MrElendig: depends on your definition of stable.  Nothing rolling release is going to be much use in a production environment
21:24 < Psi-Jack> More so, the reason why they do, and what's constantly breaking it is the use of tmpfs on the pacman gpg keys home.
21:24 < triceratux> ddoobb: what youre really looking for is monthly or weekly snapshots rather than minute-by-minute updates. that gives you the choice of whether you want to pretened its stable or not
21:25 < MrElendig> sadasaulna: eh, rolling release is fine in production if you know what you are doing
21:25 < ddoobb> I just don't want to do a fresh install every Ubuntu LTS release.
21:25 < sadasaulna> MrElendig: doubtful
21:25 < MrElendig> sadasaulna: most people seem to just blindly trust debian/ubuntu/centos/whatever repos though
21:25 < MrElendig> and never run test boxes, their own mirrors and so on
21:25 < DLange> ddoobb: you can upgrade from one LTS to the next in-place
21:26 < MrElendig> and then they try to run a rolling release distro the same way and then they end up crying
21:26 < MrElendig> or they try to hold some packages back and then break everything due to that :p
21:26 < triceratux> Psi-Jack: i uppraded to swagarch 1807 & the gpg stuff has gone away for me. i think he knew he had some config issues & ironed them out. its true its just one guy & a nonexistent community but ive seen worse
21:26 < sadasaulna> MrElendig: yeah once you're running your own mirrors thats different, because you have complete control over the upgrades
21:26 < MrElendig> rolling release is not the reason for why you probably shouldn't run arch in production
21:27 < sadasaulna> in production you gotta do a lot of testing of new software before it gets live
21:27 < MrElendig> support availability, security auditing etc are more major factors
21:27 < triceratux> swagarch is like the mint of arch. sure it looks a lot better. that doesnt mean its not profoundly broken in new ways
21:27 < sadasaulna> lol triceratux
21:28 < MrElendig> you don't run RHEL for the quality of the distro, you run it for the support
21:28 < DLange> and the quality of the distro
21:28 < sadasaulna> MrElendig: i've run Debian in production, no support, but I can trust it to work
21:28 < DLange> 'cause these guys know what they are doing
21:28 < DLange> while some "single guy distros" ... don't
21:29 < Psi-Jack> Heh.
21:29 < Psi-Jack> DLange: In SwagArch's case, definitely has some 'single guy distro' issues. :)
21:29 < ultr4_l4s3r> Void Linux has some "zero guy distro" issues
21:29 < sacules> lmao
21:29 < MrElendig> sadasaulna: but there are 5000 companies willing to sell support for debian when stuff breaks
21:30 < sadasaulna> I admit I was kind of put off by Void having its founder just run off and they had no control over the domain, irc channels and stuff
21:30 < triceratux> ultr4_l4s3r: ^^ ya beat me to it rofl
21:30 < sacules> sadasaulna: they regained control of the irc channels tho
21:30 < sadasaulna> MrElendig: we never needed it, i know some people want that comfort but we never used support from any 3rd party company
21:31 < sadasaulna> sacules: I know but it leeaves a bad feeling in the mouth
21:31 < sacules> yeah, dunno what happened to him but they're slowy recovering
21:31 < sacules> devs are pretty active and nice
21:31 < ultr4_l4s3r> Yeah, seems like they sorted everything out now.
21:31 < sadasaulna> yeah it seems like a nice project
21:32 < sacules> i finally switched to it some days ago, no regrets tbh
21:32 < sadasaulna> i'm out of the IT scene now, don't work in it anymore, never going to work in it anymore so I could care less about production now anyhow
21:32 < sadasaulna> for me Linux and UNIX is just for funs because I enjoy using it
21:32 < ddoobb> I'm going to try out Solus. Hope I don't regret it. Thanks everyone.
21:33 < ultr4_l4s3r> Good luck and have fun!
21:34 < Psi-Jack> Hehe. Nice. :D
21:39 < ddoobb> Anyone has an opinion on GNOME vs Budgie?
21:39 < ayecee> why
21:39 < Psi-Jack> Well, Budgie is pretty nice, actually.
21:40 < Psi-Jack> It's basically Gnome with Budgie as the window-manager/shell instead of Gnome shell.
21:40 < ultr4_l4s3r> Yup, now that icon grouping is implemented too Budgie is very comfortable to use
21:41 < sadasaulna> Anyone used RebeccaBlackOS?
21:41 < searedvandal> compared to gnome, budgie is amazing
21:42 < ultr4_l4s3r> sadasaulna, I think the last time I tried to use it in a VM it didn't boot, then I gave up ^^
21:42 < triceratux> sadasaulna: sure have. for more than a day at a time. dont mind it at all. one of the best ways to get a good look at wayland
21:42 < sadasaulna> yeah i've never used wayland so i'm interested to try it
21:42 < ultr4_l4s3r> For that purpose it should be a good choice
21:43 < sadasaulna> i mean you got like it just for the name rite?  :)
21:43 < ultr4_l4s3r> That's a big plus too ^^
21:43 < triceratux> sadasaulna: i run it on the metal from the iso. its pretty actively updated actually but i havent booted it since before ubuntu 17.10
21:44 < sadasaulna> i'm gonna give it a go over the weekend I think, got nothing better to do
21:44 < ultr4_l4s3r> go for it, can't hurt ;)
21:46 < Bazkitkaise> why can't I type
21:46 < Bazkitkaise> lol nm
21:46 < Bazkitkaise> oops
21:47 < TheWild> hello
21:47 < fndreic> hello, i set my font of choice in /etc/vconsole.conf but this not load my font when i restart linux
21:47 < fndreic> and i must do it manually by setfontcommand
21:48 < fndreic> what can i do to load it automatically ?
21:53 < sacules> font for what?
21:54 < fndreic> sacules: tty
21:54 < fndreic> i set it in vconsole.conf but this doesnt load at system start
21:54 < TheWild> there's one thing on Linux which makes me happy today. The rendering of truetype fonts that are supposed to be pixel fonts (PxPlus, GohuFont Mono etc.). On Windows it was kind of fked - it tried to antialias them anyway... or maybe it was in some settings.
21:54 < sacules> maybe you have to run fc-cache first?
21:54 < esselfe> fndreic: I put it in my bash_profile, but I know there's better place to do it
21:55 < fndreic> esselfe: yes and i search that better place ;p
21:55 < fndreic> btw where u have that bash profile ?
21:55 < TheWild> even FreeType didn't work quite right with them, but I suppose it was some old version.
21:55 < fndreic> sacules: hmm how to load that first ?
21:55 < sacules> run "fc-cache" on a terminal
21:55 < esselfe> in the home directory with a test to see if it's a tty (the tty command)
21:58 < fndreic> esselfe: what ?
21:58 < fndreic> i dont understand u
21:59 < Psi-Jack> u isn't here to be understood anyway.
22:00 < esselfe> fndreic: here's my line, not what I thought '[ "$TERM" = "linux" ] && setfont GohaClassic-16'
22:01 < fndreic> esselfe: line for what ?
22:01 < esselfe> for ~/.bash_profile
22:01 < fndreic> and why TERM=linux ?
22:01 < fndreic> what for ?
22:01 < esselfe> that's so the tty has correct font at startup
22:03 < esselfe> fndreic: in normal/GUI environment, TERM is xterm or xterm-256color
22:03 < esselfe> tty is "linux"
22:03 < fndreic> esselfe: yes
22:03 < fndreic> but in gui i wanna urxvt
22:05 < Zexaron> hello
22:07 < Zexaron> got a failing HDD, after 2 months of purchase
22:07 < Zexaron> need to get data off NTFS partition, smart looks weird but not sure how bad
22:09 < compdoc> that sucks
22:11 < CyberManifest> http://dpaste.com/1TRDQKC  vs  http://dpaste.com/35A0M7F for Linux Mint 19 on a MacBook Pro for mostly genreal Home use with caveats of exploring, programming, testing, and hosting.
22:13 < fndreic> How to mount lvm partition ?
22:13 < notmike> `mount`
22:15 < searedvandal> 'mount /dev/<volume_group>/<logical_volume> /<mountpoint>'
22:17 < fndreic> searedvandal: but what to do before that mount ?
22:18 < searedvandal> fndreic, how to set up lvm from scratch?
22:18 < searedvandal> or what do you mean?
22:18 < fndreic> searedvandal: no, i have OS on lvm partition already
22:18 < fndreic> now i wanna mount it
22:19 < fndreic> when i run live cd
22:19 < searedvandal> mounting should be as simple as the line I wrote above?
22:19 < peetaur> fndreic: lvm pvscan; vgs; vgchange -ay vgname; lvs; mount /dev/vgname/lvname /somewhere
22:19 < boblamont> I'm using the "top" command via ssh on my shared hosting server. In my cpanel usage graph, I'm using "excessive" amounts of cpu, in top, it's only showing 0.3% usage, and the only processes showing up are top and bash. Does that sound right? There's tons of stuff going on in my local top.
22:20 < fndreic> thanks peetaur
22:23 < ksk> boblamont: sounds like the machine you ssh to is not the machine cpanels runs on (some hosters have like jumphosts you ssh to, which have your web contents embedded)
22:24 < ksk> I dont know about cpanel at all, and I dont think this channel covers it.
22:24 < peetaur> boblamont: it's possible to hide processes from other users...maybe you happen to have such a setup and need to run top as root
22:24 < ksk> ,v cpanel
22:27 < ksk> oh nevermind about bein on topic, this is ##linux after all.
22:28 < boblamont> They claim in their documentation that logging in this way allows monitoring of cpu usage... but it does seem odd that it's only showing what I'm doing in that specific sessiI figure on goign to #cpanel, but figured I should make sure there isn't just some flag on top I was missing
22:28 < boblamont> That got smushed together a bit.
22:29 < searedvandal> contact customer support
22:29 < boblamont> Oh god, not that again.
22:29 < searedvandal> what? that's what I would do if things didn't match their documentation
22:29 < searedvandal> after all, I'm assuming you pay for this service
22:29 < boblamont> I know. I have.
22:30 < searedvandal> and what did they say?
22:30 < boblamont> They're nice, but they don't seem to know much about things that contradict the documentation.
22:30 < searedvandal> sounds like a rather unprofessional host, in my humble opinion
22:30 < e36freak> then they need to find out
22:31 < e36freak> "i'll look into that and call you back" would be acceptable
22:31 < TRS-80_mobile> nobody knows anything anymore, that's why I do everything myself (as much as possible)
22:31 < searedvandal> just say you'll change host if they can't answer a simple question on why their documentation doesn't match reality
22:32 < boblamont> Yeah, I use the help chat... so they go off and try to figure it out. It's worth the wait when they do well.
22:32 < searedvandal> never accept "I don't know" as an answer when you're a paying customer
22:34 < boblamont> Well, sometimes an "I don't know, I'll go ask someone else" or "I'll switch you to someone else who does more with X" might be more helpful than someone enthusiastically trying to help with an area they've never run across before.
22:34 < searedvandal> should be easier for them to answer than us that don't know what patches they've applied to the kernel, how their infrastructure is laid out and things like that. on a general level I would assume cpanel and top to show somewhat the same information, if both programs have access to the same information
22:36 < searedvandal> boblamont, sure. but that's not the same as "I don't know" and nothing more than that. they should always be able to provide an answer, either directly or through escalating your ticket to the right person
22:37 < searedvandal> I've worked years at first line support, and believe me, I would never get away with just saying I don't know and not escalate to someone who knew
22:38 < boblamont> They never cut off with I don't know. They keep at it, but since I'm doing help chat, I'm sort of stuck there while they keep trying to figure it out.
22:41 < CyberManifest> http://dpaste.com/1TRDQKC  vs  http://dpaste.com/35A0M7F for Linux Mint 19 on a MacBook Pro for mostly genreal Home use with caveats of exploring, programming, testing, and hosting ?  Which One?
22:43 < acf_> hello
22:43 < acf_> so I recently upgraded from Debian stable (stretch) to testing (buster)
22:44 < acf_> which lead to a kernel upgrade from 4.9.x to 4.16.x
22:44 < phinxy> there is no going back now!!! mu-ha-ha
22:44 < acf_> haha yep
22:44 < acf_> fortunately can still boot the old kernel
22:44 < acf_> because the new one seems to randomly freeze up
22:44 < peetaur> try 4.14 ... it's the best kernel
22:45 < acf_> I've enabled magic sysrq and tried to do some debugging, but that didn't go very well
22:45 < acf_> I mean I'm hoping to use stock debian kernel
22:45 < peetaur> the latest long term on https://www.kernel.org/category/releases.html expecially if maintained by GKH and especially if not maintained by someone from canonical is usually the best
22:46 < phinxy> typo or did you say canonical is bad for the kernel?
22:47 < peetaur> they tend to make more mistakes in the kernels that get cleared up eventually
22:47 < fndreic> How to set up fonts to load automatically when system boots ?
22:47 < phinxy> but theyre downstream, how do they influence a long term kernel?
22:47 < peetaur> I'm sure cannonical and any huge corp is great for the kernel, just that until it matures, there can be more bugs depending on who does it
22:48 < acf_> 4.16 is stable though, yeah?
22:48 < acf_> I understand if I want something super stable probably 4.14 is best, but I'd kind of like to know what's breaking
22:48 < peetaur> 4.16 is EOL upstream...so you rely entirely on the downstream to update it
22:48 < searedvandal> 4.16 is eol
22:48 < acf_> oh it is?
22:49 < searedvandal> 4.17 is the stable upstream now
22:49 < phinxy> Can you give me a example of who is upstream from 4.17 ?
22:50 < peetaur> and when I said latest, I meant like if it was 4.14.0 to 4.14.3 or some arbitrary small number, maybe you still stick to the previous (unless some reason..like ryzen didn't like kernels before 4.9 before 4.9 was the best kernel)
22:50 < peetaur> 3 is like 3 weeks of patches
22:50 < searedvandal> phinxy, was that question for me?
22:50 < phinxy> searedvandal• the channel in general
22:51 < searedvandal> upstream, kernel.org
22:51 < acf_> looks like the kernel package in Debian testing is still 4.16
22:51 < acf_> I guess I can wait for them to update it or something..
22:53 < peetaur> phinxy: upstream means wherever they get it...so talking about your machine, downstream is the bottom... your distro..and upstream is kernel.org; the distro probably has more weird patches their customers asked for, security hotfix patches, etc. some features (like BFQ scheduler on archlinux), but I find upstream has a better coverage of certain backported fixes (but in some cases the ubuntu kernels are the best for that..like vanilla 4.9 had
22:53 < peetaur> stable bcache, but ubuntu's 4.4 has them backported)
22:53 < peetaur> (and I find ubuntu's 4.4 kernel is exceptional...and 3.13 was crap)
22:53 < peetaur> (which are the standard ones in 16.04, and 14.04 respectively)
22:57 < NGC3982> in linux, can't i use img src in a html script to any normal path on my system?
22:58 < Psi-Jack> Well, you.. can, to a file:///
22:59 < NGC3982> oh
22:59 < phinxy> three forward slashes for safety, one extra cant hurt.
23:01 < phinxy> aha - the last one is for root / ?
23:01 < NGC3982> something like: <img src="file:///home/henrik/Dropbox/Apps/Tasker/planter1.jpg">
23:01 < bls> for an absolute path, yes. can also use relative
23:03 < NGC3982> hm, that's odd.
23:04 < NGC3982> still showing broken pic.
23:04 < NGC3982> i need to check permissions..
23:09 < jim> you can also look at the error log of the web server, might have something there
23:12 < autopsy> Usually the log will say permission denied on a file access if it's a permissions issue.
23:12 < bls> you can't serve a file like that via a webserver
23:12 < autopsy> ENOACCES Or similar.
23:13 < autopsy> Yeah bls.
23:20 < jeffree> my SMART data reports "power on hours" as only having changed by 3 hours since yesterday. Does this mean it's measuring total read and write time rather than power on time?
23:21 < bls> is it a "green" disk that can shut itself off on inactivity?
23:21 < jeffree> it's an nvme drive
23:22 < jeffree> so, I guess that's a yes?
23:23 < bls> don't really know that the fact it's an nvme drive implies that it does the powersaving that some of the bigger spinning disks use
23:23 < jeffree> I wouln't guess it uses much power when it's not reading or writing, no idea
23:23 < bls> this is more of a question for ##hardware
23:23 < jeffree> ok
23:24 < acf_> I've noticed my power on hours is quite small as well
23:24 < acf_> but I assumed it might not count time when the computer is in suspend?
23:24 < NGC3982> bls: i think i understand. i do not want my web server to actually reach into / like that.
23:24 < jeffree> acf_: for me, it's been actually on
23:24 < NGC3982> so ill simply keep the files in the given folder using the index.html.
23:25 < bls> NGC3982: file:// refers to paths on a disk local to the system the browser loading the HTML file is running on
23:25 < acf_> interesting..
23:43 < jim> NGC3982, so, it only allows your browser to access files on the machine the -browser- is running on, if that also happens to be where the -server- is running, then yes, it can access those files as well. but, some person who's not you and not using the server machine (instead, their own machine), file/// will only access their own files
23:43 < NGC3982> i see.
23:44 < NGC3982> i made a cronjob copying <what i need> to the same location as index.html.
23:45 < jim> so, obviously you want to control physical access to the server machine from others... but other than that, your server would only serve up urls it knows how to serve
23:46 < jim> it's a very good bet that file:/// doesn't touch any server at all
23:46 < NGC3982> ok
23:47 < bls> NGC3982: and if you want to only copy the files when something changes instead of on a timed cadence, checkout incron and/or syncthing
23:48 < NGC3982> i use incron for a lot of things. this is a camera importing pictures to Dropbox every two minutes. i set the cronjob for every minute, since it should be a 24/7 job.
23:49 < NGC3982> although i am noticing that dropbox is the bottle neck here. the synchronisation against the local server is really slow.
23:49 < NGC3982> http://hydroponics.henjoh.se/ <- also, bad auto focus galore.
23:54 < arora> hey, in libre calc, is there a way to check current date with the dates in a row and highlight the one that matches?
23:55 < bls> arora: try #libreoffice if no one in here knows
23:58 < phogg> arora: use conditional formatting for that
23:59 < phogg> arora: select the range of dates, open the Format menu, choose Conditional Formatting and then Condition. For condition 1 choose "cell value is" then "equal to" and put in the formula: =today()
23:59 < phogg> arora: then choose a style to apply
--- Log closed Sat Jul 07 00:00:28 2018