--- Log opened Sun Jun 24 00:00:00 2018 --- Day changed Sun Jun 24 2018 00:00 < Dagmar> execsnoop being the most useful of those right now 00:00 < Dagmar> You're _learning_ 00:00 < Henry151> right execsnoop would be really helpful for figuring out why i can't install bcc-tools, but I need to first install bcc-tools to have execsnoop 00:00 < Dagmar> Six month of this and you'll be ready to strangle reporters talking smack about street view cars eavesdropping on people's porn-surfing habit 00:06 < Loshki> Henry151: this is why engineers with short attention spans need faster machines. Tell your boss... 00:10 < pnbeast> Henry151, tell your boss I need a faster machine, too. 00:11 < jaziz> are brown switch keyboards too loud for an open office environment? 00:12 < dviola> I hate open offices 00:12 < jaziz> why? 00:12 < Umeaboy> Use Libreoffice then. :) 00:12 < jaziz> looooooool 00:13 < dviola> I have worked with a few manchildren in open offices and they distract me with their noise 00:13 < dviola> I want to work not listen to their bullshit all day 00:13 < Umeaboy> Buy bubble wrapper and wrap them in it. :) 00:13 < dviola> is that clear? or should I go on? 00:13 < jaziz> is clear, I understand 00:13 < storge> the place i work is nearly silent, even with 5 people in the room. 00:14 < jaziz> you could also get noise cancelling headphones though 00:14 < storge> people are just quiet 00:14 < jaziz> would probably work like a charm 00:14 < jaziz> lol you guys would absolutely hate me in my office 00:14 < storge> should i go ahead and hate you now? 00:15 < jaziz> hmmmmmmmmmmmmm 00:15 < jaziz> not before you tell me whether brown switches would be too loud haha 00:15 < storge> anyway the good thing about open offices is that you can have more people see when you masturbate 00:16 < jaziz> lolol 00:16 < dviola> jaziz: answering your question, I do have a filco with brown switches and it's ok :P 00:16 < jaziz> but is it okay with the people arouuuund you haha 00:16 < storge> i think typing sounds are soothing, so i would say get the loudest ones 00:16 < dviola> jaziz: no idea, I suppose :P 00:16 < jaziz> loool 00:16 < jaziz> nice 00:16 < jaziz> maybe I can get dampeners 00:17 < alexandre9099> Dagmar, where should i add the disk to be passed? 00:17 < alexandre9099> disk and fs pool do not seem to work 00:18 < alexandre9099> Format of device '/dev/sdb' does not match the expected format 'dos' 00:18 < dviola> jaziz: no, I don't hate anyone that I work with, but I had this coworker that would put very loud music in our office, it was just impossible to work this way 00:18 < alexandre9099> mount: /var/lib/libvirt/images/abc123: /dev/sdb already mounted or mount point busy. 00:19 < jaziz> music in an office is extraordinarily obnoxious 00:19 < dviola> indeed 00:19 < jaziz> personally I just speak loudly and troll around 00:19 < jaziz> which might be equally obnoxious 00:19 < storge> might be? 00:19 < jaziz> should prob tone it down 00:20 < storge> yes, you can start by toning it down right here right now 00:20 < jaziz> I'm a little excited lolol, I'll dial it down next week most likely 00:20 < jaziz> just started an internship 00:21 < alexandre9099> hmm https://wiki.archlinux.org/index.php/PCI_passthrough_via_OVMF#Physical_Disk.2FPartition 00:21 < alexandre9099> i'll try thi 00:22 < Henry151> i like http://arrlexamreview.appspot.com/ 00:22 < Henry151> oops wrong window 00:52 < hesperaux> greetings. I've got a hardware raid6 and I am about to add it to my system. It's just a blank raid6 array presenting as /dev/sdb. I'm debating on whether or not I should create /dev/sdb1 to fill the array and create an ext4 fs on it, or if I should use LVM w/ LUKS 00:53 < hesperaux> I'd like to use LVM with LUKS, but I'd like some advice on whether or not to pvcreate the entire disk /dev/sdb, or if I should create /dev/sdb1 and then pvcreate. My gut says to do it on the whole disk, because if I expand the array later (tomorrow) it will be easy to expand the filesystem using LVM tools and not require a reboot etc. 00:53 < hesperaux> However, some comments online say that not creating a /dev/sdb1 is a bad idea because it's bad. Thoughts? 00:57 < stumpys> is there a program that allows mere humans to modify iptables? 00:57 < Dan39> stumpys: umm... iptables? 00:57 < BlueShark> You missed humans. 00:57 < Dan39> firewalld? 00:58 < stumpys> Dan39, firewalld eh? will take a look 00:58 < phogg> stumpys: No. There are many programs that write the rules for you so you don't have to learn to use iptables, but you won't get the same level of control. 00:58 < phogg> stumpys: if you find iptables intimidating I have good news! You can use nft instead! 00:58 < Dan39> hesperaux: i saw no luks, plain dm-crypt, then lvm on top 00:58 < Dan39> haha phogg 00:58 < phogg> of course nft is almost as obtuse and far less thoroughly documented via Google searches 00:58 < stumpys> phogg, I don't think I need really detailed filtering 00:59 < alexandre9099> hi again, i would like to do a kind of serial to gamepad converter (to use with this amazing arduino code i made https://gitlab.com/alex9099/segaread/blob/master/arduinocode/segaread.ino ), but i have no idea where to start, i'm targeting linux, anyone can point me where to look at? i tought about linux modules, but maybe that is too low level 00:59 < stumpys> nft, not familar with that, though wasn't familar with iptables for that matter 00:59 < hesperaux> Dan39, seems a little more complicated than just using lvm with luks built in - are you suggesting also to create a /dev/sdb1, or just use /dev/sdb with no partitions? 00:59 < triceratux> mere humans modify iptables all the time. they just dont manage to save them 01:00 < phogg> stumpys: nft is a new and more powerful front end to kernel packet filtering; it's recommended by the in-kernel netfilter maintainers as the tool of choice. 01:00 < phogg> triceratux: sure we do 01:00 < phogg> stumpys: what kind of rule do you want? iptables isn't that hard 01:00 < Dan39> hesperaux: well lvm on sdb is fine, you dont need sdb1 i think, im not an expert though. i'm just saying for encryption to hae a point to me, i dont want people to easily know im using linux (like if you pvcreate first) or using encryption (like if you LUKs the drive first) 01:00 < stumpys> phogg, thanks. I will check it out 01:00 < stumpys> phogg, my situ is I have a vm and just want to allow imap traffic, and one website to go in/out 01:01 < stumpys> phogg, block everything else 01:01 < phogg> stumpys: learning *that* much iptables is trivially easy (learning it with nft is just as easy) 01:01 < tds> phogg: do you have any recommended starting documentation for nft? I should really learn it at some point 01:01 < stumpys> phogg, well if nft is the newer thing then perhaps I should take a look at it. 01:01 < hesperaux> Dan39, I see your point from a security standpoint..hmm. I think it's pretty easy for an expert to tell if it's encrypted though just by reading some raw data right? 01:01 < phogg> stumpys: you want to set INPUT policy to DROP, OUTPUT to ACCEPT and then add a rule to the input chain for each imap port 01:02 < Dan39> nft is for the new nftables btw, its like the kernel replacement for iptables i think, just fyi 01:02 < stumpys> phogg, alright, will study up... thx 01:02 < Dan39> hesperaux: no 01:02 < tds> I thought iptables is just a frontend for netfilter, iptables and nft both use the same kernel module? 01:03 < Dan39> maybe im wrong tds 01:03 < Dan39> you might be right now that i think 01:03 < Dan39> derp 01:03 < Dan39> hesperaux: the whole point of encryption is you shouldnt be able to tell it apart from random data 01:04 < hesperaux> Dan39, so if I were to search the array for a mime type for example, and didn't find it, all I could conclude is that the disk has random data or is encrypted - I don't think the data on fresh disks is random, is it? 01:04 < Dan39> well thats a bad way of saying it 01:04 < Dan39> not the whole point 01:04 < Dan39> dunno why i said that haha 01:04 < hesperaux> Dan39, I know what you meant, it's fine 01:04 < Dan39> but yes, encryption should look like random data i believe 01:04 < phogg> tds: as usual the arch wiki is a good reference https://wiki.archlinux.org/index.php/nftables 01:04 < hesperaux> aside from being able to determine that a specific scheme of encryption was used, I don't see the risk 01:04 < Dan39> hesperaux: right, iirc HDDs usually come zero'ed 01:05 < alexandre9099> don't X forwarding need a gpu on the server? 01:05 < tds> phogg: heh, people in #Netfilter grumbled about the arch iptables page when I mentioned it there, is the nft one better? 01:05 < phogg> tds: nft allows you to update and delete specific rules without reloading a whole ruleset. Ever crashed a server by running iptables -L? 01:05 < Dan39> hesperaux: but you could just say "oh yea, that was an old windows install my mom used and i wiped it with random data" 01:05 < Dan39> alexandre9099: no 01:05 < hesperaux> for plausible deniability, which is kinda what you're talking about, a non-blank disk with random data is likely encrypted - the question is how. So in my scenario, if I used luks on lvm, they could guess easily that it's luks and work from there; as opposed to just saying "well it's encrypted somehow probably" 01:05 < phogg> tds: I don't say arch has the best advice, but for a quick overview it's fine. 01:05 < hesperaux> Dan39, true 01:06 < alexandre9099> Dan39, hmm, so X forwarding would not work at all with 3d applications? 01:06 < Dan39> hesperaux: i think like you said, there is some plausible deniability 01:06 < Dan39> alexandre9099: im not sure, i dont think i've tried, let me see now haha 01:06 < tds> phogg: yeah, it looks like a nice starting overview, thanks :) 01:06 < hesperaux> gotcha - thanks for the advice. I'll think about that 01:06 < phogg> tds: there are a few tutorials easily found on google which I had no trouble following. Of course you can also go straight to the horse's mouth: https://wiki.nftables.org 01:07 < hesperaux> Dan39, will dm-crypting the disk allow me to expand its capacity? It's got a raid6 underneath 01:07 < Dan39> hesperaux: well i think you would dm-crypt the entire drive, then lvm the dm 01:07 < Dan39> then it's flexible 01:07 < hesperaux> Dan39, right. I was just wondering if expanding the physical disk would piss off dm-crypt or something 01:08 < tds> phogg: hmm, that doesn't load, odd 01:08 < Dan39> hesperaux: what do you mean, expanding it with the RAID controller? i have no idea 01:08 < Dan39> good question 01:08 < Dan39> test it? 01:08 < phogg> tds: yeah I was just noticing that. Seems like a few sites are inaccessible right now... maybe a common VM host is having trouble 01:08 < hesperaux> ok i'll look into that 01:08 < Dan39> create a 100MB file.img and test? 01:08 < phogg> or maybe the internet censors are finally coming for our guns 01:09 < hesperaux> Dan39, good idea 01:09 < phogg> tds: I can't even ping netfilter.org right now either 01:09 < tds> ah yeah, same 01:10 < tds> oh well, someone else's problem :) 01:10 < hesperaux> brb, I think some fried chicken is in order 01:10 < Dan39> alexandre9099: 3d doesnt seem to work to me, i tried glxgears and got "Error: couldn't get an RGB, Double-buffered visual" 01:11 < Dan39> xclock works fine, so i know the X forwarding is working 01:11 < tds> I've been meaning to actually automate the config for my routers for a while as well, I guess I could switch them to nftables at that point 01:12 < alexandre9099> Dan39, i see, maybe it is kind of some cpu processing, so there is no room for 3d :D 01:12 < phogg> tds: if you like hand rolling your firewall config and don't use management tools the nft syntax is so, so much nicer 01:13 < Dan39> alexandre9099: im sure it's possible though, or juse use something like VNC 01:13 < tds> yeah, at the moment all my stuff is plain iptables, I have a standard ruleset applied to every server but still manage the rules on the routers pretty much manually :/ 01:13 < alexandre9099> *i was forgetting that there is minitool partition for windows, goes to install trough some installer that found online... then i get asked if i want opera browser... ffs, i just was a program, not two, ah, the beauty of gnu+linux* 01:13 < alexandre9099> :D 01:14 < tds> or at least the standard config on a router is "go and copy the rules from another one and modify them a bit" 01:14 < alexandre9099> Dan39, yeah, but then a gpu would be needed 01:14 < __builtin> hi 01:14 < alexandre9099> on the server 01:14 < Dan39> alexandre9099: research how steam-link does it 01:14 < __builtin> how can I make a read() from a socket block until the full amount of request bytes is read? 01:14 < phogg> tds: I wrote my own scripts around iptables years ago and use them because I know them. I'm trying to convert currently, but even though nft is nicer there's a lot of scenarios I have to support before I can switch 01:15 < Dan39> __builtin: i dont think sockets are usually programmed that way. just read the bytes into a buffer until you get tot he end of your message, based on length or some delimiter 01:15 < Dan39> but im no C expert <_< 01:15 < __builtin> hmm, I don't think I made my question clear 01:16 < __builtin> I want it so that when I call read(sockfd, buf, len), it always returns len bytes, never less 01:16 < alexandre9099> Dan39, my point was if x forwarding required a gpu on the server :) i know that there is vnc, AFAIK steam *in home stream* is some kind of vnc on steroids :D 01:16 < phogg> __builtin: I assume you are trying to avoid a block within your application code 01:16 < __builtin> no, blocking is exactly what I want 01:16 < __builtin> I want the program to block until the `len' bytes are read from the socket 01:17 < Dan39> __builtin: well the man page for read() says it doesnt work like that 01:17 < Dan39> soooo 01:17 < phogg> __builtin: I am not really an expert but I don't think you can do that. You need a higher order abstraction. 01:17 < Dan39> we cant wave our magic wands and change the way read() has worked for 50 years :P 01:18 < phogg> we can make a new read_myway() 01:23 < phogg> __builtin: if you look at e.g. https://stackoverflow.com/questions/666601/what-is-the-correct-way-of-reading-from-a-tcp-socket-in-c-c the accepted answer has a simple read() wrapper that does what you want, more or less. 01:24 < phogg> you could still spin in the while loop forever since he doesn't check for and abort when the socket closes without sending all data 01:25 < Dan39> also FYI, there is ##c channel if thats what you want 01:26 < cluelessperson> * * * * * root /var/www/update.sh > /var/www/run.log 01:26 < phogg> yeah, better to ask experts and not people like me who only sometimes delve in to C 01:26 < cluelessperson> I don't understand why this cronjob fails 01:26 < cluelessperson> run.log is epmty 01:26 < cluelessperson> no errors 01:27 < phogg> cluelessperson: your update.sh probably depends on an environment which is not set 01:27 < phogg> that's the most common reason 01:27 < aBound> I'm an expert in eating. :P 01:27 < phogg> cluelessperson: it could also be non-executable (but I assume you checked that) 01:27 < Dan39> cluelessperson: also, you're not capturing stderr 01:27 < cluelessperson> https://hastebin.com/ejijulibug.bash 01:27 < phogg> cluelessperson: pastebin update.sh if you want more detailed help 01:27 < Dan39> > /path/to/log 2>&1 01:27 < phogg> ah nvm 01:27 < cluelessperson> this is what it is, and it's executable 01:27 < cluelessperson> I'll try stderr 01:27 < phogg> ugh, hastebin 01:27 < phogg> why are the awful pastebins so popular? 01:28 < aBound> Probably cause they've been using it for a period of time. 01:28 < pnbeast> I'm not sure it's any explanation, but I'd like to note the most popular OSs at this point. 01:28 < cluelessperson> yeah, I just don't have a better one 01:28 < cluelessperson> I'm willing to break the habit 01:29 < Dan39> https://ptpb.pw 01:29 < cluelessperson> that's mnemonic 01:29 < phogg> cluelessperson: termbin.com, gist.github.com 01:29 < phogg> cluelessperson: when in doubt paste.debian.net is not run by malicious idiots 01:30 < phogg> I always liked sprunge.us 01:30 < cluelessperson> phogg: are hastebin malicious? 01:30 < cluelessperson> I'd instantly stop if I thought they were 01:30 < Dan39> i use to, but someone said it's bad because it is closed-source :P 01:30 < phogg> cluelessperson: Maybe and maybe not. I'd say "unfriendly" and won't go there. Too much JS nonsense. 01:31 < phogg> give me something I can view in lynx 01:31 < Dan39> well, it was, i see it is on github now 01:31 < phogg> Dan39: yeah 01:31 < Dan39> oh wait, maybe it's not 01:31 < aBound> Github danger will robinson. 01:31 < Dan39> thats only a client 01:32 < __builtin> phogg: I think using recv() with the MSG_WAITALL flag does what I want 01:32 < phogg> Dan39: doesn't look like just a client to me. The license is obnoxious, though. 01:32 < phogg> __builtin: I'm so excited that you're on this adventure. Have fun! Don't forget to read the fine print,. 01:33 < __builtin> thanks :) I'd hardly qualify it as an adventure, though 01:34 < phogg> the problem with cute licenses like WTFPL is that they're exactly like a BSD license but on shakier legal ground 01:35 < Dan39> __builtin: good find, like i said, i just read the read() manual, im no C expert, so i didnt even remember recv() existed :P 01:35 < __builtin> thanks for the help, Dan39! 01:35 < Dan39> phogg: oh, maybe 01:36 < Dan39> __builtin: really do join ##c if you have more questions or maybe ##posix 01:36 < phogg> Dan39: I am no python expert, but yeah it's the server too (otherwise why reference WSGI?) 01:36 * __builtin used to hang out in ##c and all a crapload of other channels 01:36 < Tech_8> hi 01:36 < Dan39> phogg: i think you're right. such a small :D 01:36 < phogg> Dan39: so small it's probably bug free, too 01:37 < __builtin> too much data bandwidth to and from my quassel core, though 01:37 < Dan39> __builtin: iirc, last time i was in ##c asking questions about sockets they send me to ##posix 01:37 < phogg> I like that 23 of its 159 lines are the man page. 01:37 < phogg> Tech_8: hi 01:38 < Tech_8> hi phogg 01:38 * phogg plays virtual jim 01:38 < phogg> Tech_8: Hi. Did you have a question? 01:38 < Tech_8> no 01:39 < phogg> Ah, okay. I'll go back to sleep. 01:39 < Tech_8> we can chat though 01:39 < phogg> Nope. I'm too busy procrastinating. The deal is that you guys give me problems and I avoid the work I'm supposed to be doing. If I'm not working on something else I have work for real. 01:40 < phogg> cluelessperson: anyway, the most common solution to your type of problem is to manually set PATH at the to of your script to the value you expect it to have. 01:41 < phogg> cluelessperson: in any case a shell script should not have a .sh extension (or any at all). 01:41 < cluelessperson> the cron job is not producing output to run.log 01:41 < cluelessperson> phogg: why should I set PATH ? 01:41 < cluelessperson> that doesn't make sense 01:41 < phogg> cluelessperson: because PATH in cron scripts is very sparse by default 01:41 < phogg> without seeing your script I can't be sure that's your problem, but it is the most common one 01:42 < cluelessperson> https://hastebin.com/ejijulibug.bash 01:42 < phogg> cluelessperson: I already said I'm not going to visit hastebin. 01:42 < cluelessperson> https://hastebin.com/raw/ejijulibug.bash 01:42 < phogg> I will wget a raw, I guess 01:42 < Loshki> cluelessperson: alternatively to setting PATH, you can use absolute file names for anything you execute, a recommended practice anyway. 01:42 < cluelessperson> :P 01:42 < Dan39> just wait phogg 01:43 < phogg> what directory is journalctl normally installed in? 01:43 < phogg> cluelessperson: you need to check your script: shellcheck.net 01:43 < phogg> great zeus, man! It's 2018 and you're still using *PRE STANDARD SUBSHELLS* 01:43 < Dan39> https://ptpb.pw/Joi7 01:44 < Dan39> there ya go 01:44 < phogg> Dan39: you are a treasure 01:44 < Dan39> just for you and your safe space phogg <3 01:44 < phogg> but I can stomach a noninteractive experience with hastebin 01:44 < Dan39> next time ill give you a free tampon with it 01:44 < phogg> Anybody remember what year backticks were officially obsoleted? 01:45 < Dan39> they are called ABOVE_TAB jeez 01:45 < phogg> prime, technically 01:45 < Tech_8> phogg: what do you do at your job? 01:46 < __builtin> backticks are obsolete? 01:46 < __builtin> as in `command`? 01:46 < Dan39> __builtin: yes 01:46 < Dan39> use $(command) instead 01:46 < __builtin> dang, first time I've heard of that 01:46 < Dan39> and usually "$(command)" 01:46 < __builtin> I've almost never used $(command) 01:47 < phogg> __builtin: yes, they have been obsolete for a long time 01:47 < Dan39> ill admit i still use backticks sometimes for quick one-offs :P 01:47 < __builtin> I'm assuming it was because there was no easy way to nest them? 01:47 < Tech_8> whats a backtick? 01:47 < Dan39> Tech_8: ABOVE_TAB 01:48 < Tech_8> Dan39: what do you mean? 01:48 < Dan39> look at your keyboard 01:48 < Dan39> the key above the tab :P 01:48 < Tech_8> ~ 01:48 < Tech_8> taht? 01:48 < Dan39> -_- 01:48 < Dan39> without holding shift 01:48 < Tech_8> ` 01:48 < Dan39> yes 01:48 < Tech_8> you used to use that as a command? 01:49 < Dan39> -_- 01:49 < phogg> __builtin: that and there are certain semantic differences which make $() better, see http://pubs.opengroup.org/onlinepubs/9699919799/utilities/sh.html 01:49 < phinxy> I have no luck compiling a frambuffer driver. Its one error after another about definitions missing, and if the .h is included at the top of the .c, another set of definitions are missing instead. Clearly this method of fixing the errors is not working for me. Other things are also suggesting the order of of compilation is wrong, or what would you call it when things go in wrong order? 01:50 < __builtin> the order of includes should not matter 01:51 < phinxy> not in the .c file itself? ok. 01:51 < Dan39> cluelessperson: anyways... that shell script looks awful, what are you actually trying to do? 01:52 < Loshki> phinxy: clearly, it was never compiled on your particular setup. which means it was never tested there either. You can spend a lot of time making something compile and that can just be the beginning of your problems. Consider your options. 01:52 < Loshki> Been there. Abandoned that. 01:54 < phogg> __builtin: anyway, in 1992 SUS specified both and recommended $(). They still do that. 01:54 < phogg> or maybe it was 1993 01:54 < phogg> regardless, people still using `` today (who are not writing for older shells like solaris /bin/sh) need to stop 01:55 < cluelessperson> the cron job is not producing output to run.log 01:55 < Tech_8> give me an example using ` 01:55 < Tech_8> `command? 01:55 < Dan39> cluelessperson: what do you actually want to do 01:56 < cluelessperson> Dan39: I want to run this shell script as a cron job and have it run 01:56 < cluelessperson> it runs fine when I run it 01:56 < Dan39> but why? 01:56 < cluelessperson> what it's doing is very obvious 01:56 < phogg> __builtin: sorry, that was the wrong link before. Here's the right one: http://pubs.opengroup.org/onlinepubs/007908799/xcu/chap2.html#tag_001_006_003 01:56 < Dan39> print a hashrate to a file? 01:57 < __builtin> phogg: thanks 01:57 < Dan39> but why? 01:57 < cluelessperson> Dan39: it just runs once a minute to grab data from logs and export to files for inclusion to a web page 01:57 < Dan39> maybe if you tell us, we can find a better utility that already does what you want 01:57 < cluelessperson> it's just a minute update 01:57 < __builtin> I remember the last time I tried making anything complex with a bash script 01:57 < phogg> cluelessperson: none of your commands write to stdout 01:57 < phogg> __builtin: I do complex bash scripts every week. 01:58 < cluelessperson> phogg: they should update the > pretty_report.txt and > pretty_hashrate.txt 01:58 < phogg> Sometimes every day. 01:58 < phogg> cluelessperson: Do those files get new mtimes? 01:58 < __builtin> I spent most of my time trying to work around that issue with backticks 01:59 < Dan39> cluelessperson: i feel like there has to be a better way to get the hashrate from your mining service than parsing the journal logs every time 01:59 < __builtin> it was a script that was supposed to ssh into a server, and edit the .ssh/authorized_keys file to remove a key from itself upon use 01:59 < phogg> cluelessperson: I did basic cleanup without altering the behavior: http://termbin.com/bw29 02:00 < phogg> cluelessperson: you will need to alter PATH if journalctl is not in /bin 02:01 < phogg> __builtin: shell quoting is weird, but once you get your head around it you stop having problems like that. 02:01 < phogg> it's just not like other things 02:04 < Dan39> giggity 02:05 < phinxy> Is there a small risk that the CPU is overheated and destroyed with a bad kernel? 02:06 < cluelessperson> phogg: bios should have critical shutoff 02:06 < cluelessperson> so should cpu 02:06 < cluelessperson> more likely it would just lower the lifetime 02:06 < cluelessperson> and shut off 02:10 < phogg> phinxy: If your heatsink and fans are working there should be no way for that to happen. As a last resort x86 CPUs will turn off to avoid death, but it should never come to that. 02:11 < phogg> phinxy: Once upon a time a server room where I worked lost AC and the temperature rose to 140 degrees F. It was like that for a few hours before we got things under control. All of the servers stayed on. No CPUs failed during that time, though some power cables did start to melt. 02:12 < phogg> phinxy: compared to that a kernel that tries can't kill a CPU with heat. There are a few instructions which can brick a few CPUs, though. 02:12 < Dan39> like that TV show with the implausibly hot programmer chick? :P 02:14 < phogg> Dan39: I've known some hot programmer chicks. Also, which show? 02:14 < Dan39> i think it was Halt and catch fire 02:14 < phogg> never saw it 02:14 < Dan39> was decent, i enjoyed it :P 02:22 < cluelessperson> the cronjob thing should just give a decent error message 02:23 < m1n> anyone recommend a keylogger (linux)? 02:24 < Dan39> cluelessperson: is cron throwing an error into your root mail maybe? 02:25 < Dan39> like if root is not allowed to execute /var/www/update.sh 02:25 < Dan39> or not allowed to write to that file... 02:26 < Dan39> cluelessperson: add some echo's to your script so you know the output is being captured 02:26 < phinxy> What are the variables referenced in C files like: #if defined(CONFIG_ION_ROCKCHIP)? 02:27 < Dan39> phinxy: those are processed by the compiler i know, how the variables are set im not sure 02:30 < mgolisch> no they are not 02:36 < mgolisch> all the things with # infront are preprocessor directives, the c compiler is never gona see any of that 02:36 < Loshki> phinxy: those are MACROS, defined and used by the C preprocessor before the compiler proper even sees it. Usually there will be a switch you can set on the compiler to show you this output. 02:37 < mgolisch> you will probably find a #define directive that declares this macro somewhere in the source 02:38 < rcf> phinxy: see above, but in your specific case(kernel module) they are set for a given kernel at build time. 02:39 < Dan39> mgolisch: thats what i was getting at.. they aren't C code that is compiled at all, they are processed before actual compilation happens, just bad terminology on my part :P 03:09 < rappet> http://gnu.church/ 03:11 < Dan39> ha 03:13 < TRS-80> lol 03:14 < lessthan0> is freeBSD less popular now? 03:14 < lessthan0> what would stallman use? 03:15 < cluelessperson> How can I test a cron file? 03:15 < lessthan0> there are people trash talking the freeBSD team on twitter calling them racist sexist or something 03:15 < phinxy> lessthan0• Trisquel OS? 03:15 < cluelessperson> so far, I still can't get this to run 03:15 < cluelessperson> it runs fine in my environment 03:15 < cluelessperson> I attempted setting path to equal mine 03:15 < cluelessperson> still nothing. :/ 03:15 < Dan39> he uses some libre linux distro im sure, on a like 90s laptop before intel added the management stuff to cpu 03:17 < lessthan0> I already read this https://www.gnu.org/distros/free-distros.html 03:18 < lessthan0> but not sure where to go from here 03:18 < lessthan0> I am a lot less interested in the GUI and the package managers 03:18 < revel> Straight to http://www.linuxfromscratch.org/ 03:18 < lessthan0> ah yes a classic 03:19 < lessthan0> I was introduced to LFS many years ago 03:19 < lessthan0> when it was over my head 03:19 < lessthan0> I am slowly getting my skills up 03:20 < lessthan0> also on my mind is what should I use for bare metal distro to run docker? 03:21 < lessthan0> I see almost everyone is on some RH or suse etc.. 03:21 < rappet> openbsd is the best linux distro 03:21 < rascul> no way, it's reactos 03:22 < lessthan0> so is it open source but not GNU free? 03:22 < lessthan0> is it not worth analyzing the differences? 03:22 < rappet> lessthan0: Yes openbsd is linux without gnu and without systemd 03:22 < rascul> also without linux ;) 03:22 < mgolisch> no its not 03:22 < mgolisch> yeah that 03:23 < rappet> also without decent wifi drivers 03:23 < lessthan0> did yall hear about the void linux guy went MIA after publicly talking about the director of the FAA? 03:24 < phogg> lessthan0: pics or it didn't happen 03:24 < phinxy> #ReactOSmasterrace 03:24 < lessthan0> and they lost some of the source and lost his passwords so the team is kinda stuck doing a rollback or just giving up 03:24 < phogg> I'm still waiting for our new Haiku overlords 03:25 < rappet> Templeos best routing distro 03:25 < phogg> the difference is that I am not joking 03:26 < triceratux> lubuntu 18.10 lxqt. resistance is futile 03:27 < cluelessperson> cron jobs are so frustrating 03:27 < victorqueiroz> WTF? https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=900399 03:27 < cluelessperson> completely silent failure 03:27 < dviola> Terry A. Davis called Linus a "nigger" because he didn't write his own compiler: https://www.youtube.com/watch?v=B5YokNW7tIs 03:28 < el> that's not really worth quoting 03:29 < Sitri> cluelessperson: cronjob output is mailed to the user 03:30 < lessthan0> victorqueiroz do you think the failure was a hardware failure caused by heat? 03:30 < phogg> cluelessperson: your script, at least the version I saw, also is not supposed to output anything 03:30 < Dan39> i told him to check his mail like an hour ago <_, 03:31 < dviola> el: I know, and sorry if you are offended, but I still want to point out that the author of TempleOS is this crazy in case someone is looking at it 03:32 < dviola> I know he suffers from mental illness, so it's not entirely his fault 03:32 < cluelessperson> Sitri: that's stupid 03:32 < cluelessperson> phogg: yes, but it's magically failing 03:33 < Dan39> ah, saturdays in ##linux 03:33 < Dan39> i think ever saturday TempleOS is brought up and the N word repeated a few times for good measure 03:33 < phogg> lessthan0: you could try reading the bug report where it explains *exactly* why 03:34 < phogg> cluelessperson: having cronjob output mailed to the job owner isn't stupid, it's standard practice going back decades. You can always configure it to be silent instead, helping nobody. 03:34 < phogg> cluelessperson: There is **NO** magic. 03:34 < phogg> cluelessperson: If you are seeing unexpected results you are expecting the wrong thing or asking for the wrong thing. Are your executables executable? Are the timestamps on the files you are redirecting into changing? 03:35 < Matviy> Noob Q, if I'm trying to write a basic Linux driver on Ubuntu, I need to download the kernel source tree for the same kernel version I'm running, and setup the same config build parameters, correct? 03:35 < Dan39> writing drivers is a noob question? :| 03:35 < rascul> it is now 03:35 < phogg> Matviy: You would need at least the kernel headers to build a module. 03:35 < cluelessperson> phogg: still stupid. The fact that it is run is outputted to syslog (system log?) but none of the error output is 03:36 < Matviy> KK ty 03:36 < revel> Might as well just build and use your own kernel instead of trying to make a module for your current one. 03:36 < cluelessperson> and if email isn't set up, it fails silently 03:36 < phogg> cluelessperson: neither is output to syslog 03:36 < phogg> cluelessperson: a Unix box with no mail capability? Now THAT is stupid. 03:36 < Matviy> True, but I'm trying to practice building both in-tree and out of tree drivers 03:36 < cluelessperson> phogg: this is in syslog Jun 24 01:36:01 snowday1 CRON[24165]: (root) CMD (/var/www/update.sh > /var/www/run.log 2>&1) 03:36 < phogg> Matviy: *building* out of tree modules is a lot easier than writing them 03:37 < cluelessperson> phogg: relying on EMAIL for notifications is incredibly stupid 03:37 < phogg> cluelessperson: recording the *fact* that it ran is *not* the same as recording output 03:37 < phogg> cluelessperson: Stupid in what way? Email is a messaging protocol, the output is a message. Very sane. 03:37 < Dan39> cluelessperson: add an echo to script so you know you get some output for now 03:37 < iflema> cluelessperson: grab a windows trial 03:38 < cluelessperson> phogg: the logic is simple. By default, systems don't have email. By defalut, they have a syslog where system logs are output. That a system event is logged as being run, but only emailing the output is illogical 03:38 < cluelessperson> because email isn't available BY DEFAULT. 03:38 < phogg> cluelessperson: All we're doing is bouncing a stream of text from cron to the local MTA and in to a local mail spool. Any other option would require a pointless custom procotol. 03:38 < phogg> protocol* 03:38 < cluelessperson> just because that's how it's been done doesn't mean it's sane. 03:38 < phogg> cluelessperson: By default *all Unix systems have email*, so your point is wrong. 03:38 < cluelessperson> phogg: or, just output to syslog 03:38 < cluelessperson> like *everything else* 03:39 < phogg> cluelessperson: log messages are typically "just one line" but cron output is usually far from that. Possibly many kilobytes. You do not want that in syslog. 03:39 < Dan39> phogg makes a good point 03:39 < cluelessperson> then have /var/log/cron/job 03:39 < cluelessperson> point is, it's sane to LOG it 03:39 < cluelessperson> not email 03:39 < phogg> cluelessperson: Again, smarter people than you have designed it this way *for a reason*. If you don't see the virtue you just haven't thought about it. 03:40 < phogg> cluelessperson: Email is a perfectly reasonable solution. A log message that is not itself an email message would not have enough information. 03:40 < triceratux> cluelessperson: & then smarter peeps than that invented systemd timers 03:40 < victorqueiroz> lessthan0: Not sure, I think the guy might be right. unless both computers were in a very hot environment, this problem was caused by memcheck 03:41 < cluelessperson> phogg: "smarter people" would not send an email for cron jobs that can run every single minute 03:41 < phogg> cluelessperson: You're frustrated because you have a problem you can't figure out and you're lashing out, blaming the tools. Don't do that. Instead locate the problem and fix it. 03:41 < lessthan0> ok I read the entire bug report now 03:41 < cluelessperson> "smarter people" would not limit cron jobs to 1 minute periods. 03:41 < lessthan0> I understand it 03:41 < cluelessperson> "smarter people" wouldn't consider emails a better choice than LOGS 03:41 < lessthan0> I have a job doing PCB layout 03:41 < phogg> cluelessperson: there's nothing wrong with sending an email every minute. It's a blob of text into a local file, just like a log. 03:41 < Dan39> shut up about the email vs logs thing 03:41 < lessthan0> they should have never tied the outputs to VCC or GND 03:41 < phogg> cluelessperson: Ever had a cron job run away and dump ONE GIGABYTE of output? I have. 03:42 < Dan39> #lessthanthree 03:42 < lessthan0> and you never connect and output to a regulator since they are basically the same thing 03:42 < cluelessperson> phogg: sounds like an extreme failure of whomever wrote the thing. 03:42 < cluelessperson> that doesn't mean it's magically more sane to email logs instead of logging 03:42 < phogg> cluelessperson: It's a failure, but not that extreme. It happens more often than you'd think. 03:42 < cluelessperson> you're insane if you think email is better than simple logs 03:42 < lessthan0> there must be a resistor somewhere to dissipate heat and limit the load if two regulators are going to do battle 03:42 < cluelessperson> There are cases where Emails/notifications are required, needed by people 03:43 < phogg> cluelessperson: you are hung up on your concept of what a log is and what the output of a program is. Program stdout is not necessarily a loggable message. 03:43 < cluelessperson> but by DEFAULT having cron jobs send emails, is stupid 03:43 < cluelessperson> literally stupid 03:43 < phogg> cluelessperson: No. You are wrong. Just wrong. It's the opposite of stupid. It's clever and extremely nice. 03:43 < cluelessperson> For example, I want to know when certain events happen on my system 03:43 < cluelessperson> So I configure emails to be sent when those occur 03:43 < phogg> cluelessperson: The alternative would be to invent some other, custom also-not-syslog protocol. In no sane world would cron output appear directly in syslog. THAT would be stupid. 03:44 < cluelessperson> phogg: or just output to the syslog!!! 03:44 < sauvin> Am I wrong in my impression that systemd has cron-like functionality with subsecond resolution? 03:44 < phogg> cluelessperson: You can say that as many times as you like and it would still be wrong. 03:44 < cluelessperson> phogg: you can twist my words over and over, it just makes you look like a liar. 03:44 < phogg> sauvin: "timers" - yes, or so I hear. I say just write a daemon at that point. 03:44 < Dan39> sauvin: it has cron-like functionality im pretty sure with timers, no idea on the resolution 03:44 * cluelessperson looks up systemd timers 03:44 < Dan39> SHUT UP cluelessperson 03:45 < phogg> cluelessperson: I am not twisting your words. I am telling you: in order to get the behavior you need for recording cron output you need something other than syslog. 03:45 < lessthan0> there is no down side to leaving an unused output floating 03:45 < cluelessperson> well, now the cronjob magically silently fails 03:45 < cluelessperson> it runs fine when I run it myself 03:45 < lessthan0> it will consume less power anyway since the load is 0 amps 03:45 < phogg> cluelessperson: You have a DEAD SIMPLE problem with EASY troubleshooting steps. You do not need systemd timers. You do not need to go back to the 80s and change how cron has *always worked*. You just need to fix your problem. 03:46 < sauvin> What's the problem? 03:46 < Dan39> ^^^ 03:46 < phogg> sauvin: his cron job "silently fails" 03:46 < cluelessperson> phogg: if systemd actually sanely outputs errors instead of failing silently, then yes, that will work where cron fails. 03:46 < sauvin> What's the cron job? 03:46 < cluelessperson> phogg: I cannot troubleshoot nothing. 03:46 < phogg> sauvin: https://hastebin.com/raw/ejijulibug.bash 03:46 < phogg> cluelessperson: Yes you CAN 03:46 < phogg> cluelessperson: do you want to know how? I will tell you if you ask. 03:47 < phogg> cluelessperson: sulking and raving about cron is not helping YOU 03:47 < phogg> doesn't do anything to me 03:47 < sauvin> What's the cron job invocation line in the crontab? 03:47 < cluelessperson> https://hastebin.com/raw/efamovudiz 03:48 < cluelessperson> phogg: If you know how to troubleshoot cron, I'd love to know your solution. Thanks for the help regardless. 03:48 < phogg> sauvin: * * * * * root /var/www/update.sh > /var/www/run.log 03:48 < phogg> cluelessperson: this is **NOT** a cron problem 03:48 < rascul> this is in /etc/crontab? 03:48 < cluelessperson> rascul: /etc/cron.d/ 03:48 < rascul> close enough 03:48 < phogg> cluelessperson: First: Is your job running at all? If the timestamp on run.log changes every minute it is. Your entry in syslog seems to suggest it is. 03:49 < sauvin> How are you determining that it's failing? 03:49 < phogg> his original complaint was "no output in run.log" 03:49 < cluelessperson> /var/www/pretty_report.py > /var/www/pretty_report.txt 03:49 < cluelessperson> pretty_report.txt is not updated 03:49 < phogg> cluelessperson: is the timestamp on that file updated? 03:49 < cluelessperson> no 03:49 < phogg> cluelessperson: what about the timestamp on the file you redirect to earlier in the script? 03:49 < jamithy> break it down into small chunks, that's how you can troubleshoot it :) 03:50 < cluelessperson> -rw-r--r-- 1 root root 0 Jun 24 01:50 run.log 03:50 < sauvin> And you say it runs fine if you run it on a CLI? 03:50 < cluelessperson> sauvin: yes :P 03:50 < phogg> what directory is journalctl normally installed in? 03:50 < rascul> /usr/bin/journalctl here 03:51 < sauvin> When you run it at a CLI, do you do so as root? 03:51 < cluelessperson> sauvin: yes. :P 03:52 < lessthan0> the more I think about this lenovo bug the more I think they had problems with both hardware and software engineers failing to use best practices. even if the hardware is designed wrong, they can validate the software that reserves memory. 03:52 < phogg> cluelessperson: Is /var/www/pretty_report.py executable? 03:52 < lessthan0> and it would never die 03:52 < rascul> is pretty_hashrate.txt updated? 03:52 < Dan39> cluelessperson: what distro? 03:52 < lessthan0> like if they ran memtest for example 03:53 < cluelessperson> phogg: yes :) 03:53 < cluelessperson> Dan39: Debian Stretch? 03:53 < phogg> lessthan0: I solidly blame the hardware engineers. Software is supposed to suck and fail and break. 03:53 < Dan39> cluelessperson: are you asking me or telling me? o_O 03:54 < phogg> cat /etc/os-release 03:54 < HarriedItGuy> hello cluelessperson and all --- remember that cron does not have path setup the same as from the CLI, so you need to specify the full path to the programs you wish to run. 03:54 < cluelessperson> Dan39: PRETTY_NAME="Debian GNU/Linux 9 (stretch)" - 4.9.0-6-amd64 03:54 < phogg> or just set PATH manually 03:55 < HarriedItGuy> e.g. -- /bin/grep -- not just grep 03:55 < Dan39> HarriedItGuy: yea, we already had him add PATH= 03:55 < lessthan0> phogg I would dissagree and I will support that with why: if your power and thermal management are controlled in software or firmware then you will always have the possibility to catch fire or at least die from bad software. 03:55 < sauvin> cluelessperson, where is that environment variable set? 03:55 < phogg> HarriedItGuy: everything he needs is in /bin or /usr/bin, and I know /bin is in PATH by default. He just won't set PATH, though. 03:55 < cluelessperson> phogg: I took your suggestion from earlier. I set the path in the cron file itself. should I set it in the shell script? 03:55 < HarriedItGuy> sorry -- I missed that getting food .... 03:55 < b1sh0p> phogg: Testing is essential. 03:55 < phogg> cluelessperson: Yes, in your script. At the top. 03:55 < lessthan0> the solution is to have analog computers controlling thermals and power management 03:55 < jim> one way... you can set PATH in the cron script 03:56 < lessthan0> it would never happen at the cost and speed of development that they expect from this years model of laptop 03:56 < phogg> you can set PATH in the crontab file, but each job might need different things so it's better to do it in the script 03:56 < sauvin> What phogg said. 03:56 < phogg> and depending on how you tried to do it you might or might not have done it correctly 03:56 < Dan39> cluelessperson: my first step would be to add an echo line to the top of my script to make sure it is executing and can write to the log 03:57 < jim> what sauvin said 03:57 < lessthan0> it would also not be controlled in software so it would have absolutely no ability to change power management profiles at all ever. 03:58 < lessthan0> only one power profile set at the factory. any changes require hardware engineers and pcb or part changes. 03:58 < sauvin> When I test cron jobs (because I'm worthless with cron), I'll have cron call a script, set the script itself up with needed variables and suchlike, I'll explicitly call out anything by fully qualified path/filename, and I'll echo each line to some splat file to make sure what's happening is what I intended. 03:58 < lessthan0> it would be a nightmare and basically impossible in reality. 03:59 < phogg> sauvin: I used to be like that, but then one day it hit me: If I know what tools I'm using I can save typing by just setting PATH once. 03:59 < cluelessperson> Dan39: set -x ? :) 03:59 < jim> yep that's usually6 how you test stuff 03:59 < phogg> sauvin: of course there's a slight chance of someone dropping a malicious binary ahead of my expected one in the PATH, but if he has root I'm screwed anyway. 03:59 < Dan39> cluelessperson: ?? what? 04:00 < cluelessperson> phogg: it works now, it looks like it was path 04:00 < phogg> cluelessperson: Dude. Do you remember the FIRST thing I said? 04:00 < cluelessperson> you suggested that earlier, but putting it in the cron file didn't seem to work 04:00 < phogg> cluelessperson: five seconds after you asked? 04:00 < triceratux> what jim said 04:00 < Dan39> no i didn't 04:00 < lessthan0> yes they could have left the MCU outputs floating and it would have prevented the bug. but it can not prevent all power management bugs that may occur in software. all software needs to go through validation/ 04:00 < Dan39> id maybe suggest set -v. not exactly sure what set -x does, but it looks similar 04:01 < cluelessperson> I discovered I can hold a pin low with my raspberry pi apartment lock and override the doorlock's unlcok. :D 04:01 < Dan39> in this case, i only suggested an echo 04:01 < jim> triceratux did it.... now we have an inescapable loop! 04:02 < cluelessperson> phogg: you said to add path at the top. Didn't seem to help with the cron file, but did with the script itself. 04:02 < lessthan0> and in this case, proper validation would have definitely prevented the hardware bug from being an issue. so this is why software validation is everything and everything. 04:02 < cluelessperson> am I missing something else? 04:02 < Dan39> -x looks nice 04:02 < cluelessperson> phogg: thanks for the help regardless 04:02 < phogg> cluelessperson: I guess it took 17 minutes for me to specifically say "Set PATH in your script" 04:02 < sauvin> cluelessperson, apart from cron having no environment by default? :D 04:02 < cluelessperson> phogg: Sorry for being a PITA. :) 04:02 < triceratux> one of these days we will have to send someone back to the 80s to fix this once & for all ;) 04:03 < phogg> triceratux: I believe Paul Vixie is still alive. You can go beat him up if it makes you feel better. 04:04 < Dan39> shouldnt he have gotten a command not found error in the run.log file? 04:04 < rascul> not if it was printed to stderr 04:05 < Dan39> he was capturing stderr too 04:05 < jim> cluelessperson, well, you set up a cron job (or many) by editing your crontab... it's there that you would make sure that each executable you run in the cron script can be found.... OR... you fully qualify each call to an executable, not like this: cat kitty.txt ,,, but like this: /usr/bin/cat kitty.txt 04:05 < Dan39> * * * * * root /var/www/update.sh > /var/www/run.log 2>&1 04:05 < rascul> oh i didn't see the 2>&1 at the end 04:05 < Dan39> fguj: dont be PMing me bro, ask channel for help 04:06 < sauvin> fguj, you can't speak in the channel until you identify to nickserv. Ask for help with that in #freenode. 04:06 < jim> and, you would -always- fully qualify -all- executables in scripts run as root. there shouldn't be a choice there, just doo eet 04:06 < phogg> cluelessperson: as Dan39 keeps mentioning you really should record stderr somewhere, too. 04:07 < Dan39> he was... 04:07 < Dan39> he should have gotten an error in run.log 04:07 < Dan39> ¯\_(ツ)_/¯ 04:08 < rascul> it's a secret error 04:09 < phogg> mission accomplished! I have avoided real work all day. I will roll again tomorrow. 04:10 < pnbeast> When I was three, mom took me to the grocery store and I made a "secret error" around the corner from the snacks aisle. The manager said we couldn't come back any more. 04:10 < revel> Thanks for telling us. 04:14 < cluelessperson> Dan39: :P 04:15 < tomleb> Hi, I'm running rsync daemon and I'm using password authentication for sending/fetching files, but listing modules don't require any password. Is there an option to require password for module listing ? 04:16 < Dan39> tomleb: why rsync daemon instead of over ssh? 04:17 < Dan39> cluelessperson: seriously, i just tested it, worked fine for me, got command not found error into log 04:18 < cluelessperson> Dan39: ? :/ * * * * * root /var/www/update.sh > /var/www/run.log 2>&1 04:18 < cluelessperson> I mean, it seems to work now 04:18 < tomleb> No particular reason, nas4free can run rsync as daemon and I can add modules from the ui. So I'm using that 04:18 < cluelessperson> but it didn't give error ealrier 04:18 < Dan39> cluelessperson: well, i dont put a space after the 1st >, but yea, similar line 04:18 < Dan39> * * * * * root /home/dan/test.sh >/tmp/cronlog.txt 2>&1 04:19 < cluelessperson> I don't think the space matters 04:19 * cluelessperson could be wrong 04:19 < Dan39> in /tmp/cronlog.txt i get /home/dan/test.sh: line 3: blahblahlbah: command not found 04:19 < Dan39> cluelessperson: i agree, it has always worked fine with the space before 04:19 < Evidlo> remember the days when android supported USB mass storage mode? 04:20 < Dan39> Evidlo: it doesn't anymore? 04:20 < Evidlo> no they dropped that years ago. theres only MTP and some camera mode now 04:21 < Dan39> why... :( 04:21 < Evidlo> I used to be able to just mount the device straight to my filesystem 04:23 < Dan39> cluelessperson: FYI, -s This option will direct Cron to send the job output to the system log using syslog(3). This is useful if your system does not have sendmail(8), installed or if mail is disabled. 04:23 < Dan39> though i have cronie installed, debian might be vixie-cron, not positive it has -s 04:24 < `7hr34t_hvntr> su -l user causes you to switch to that users home directory and also take on any environmental variables that user has set, correct? 04:25 < spare> login pulls /etc/profile /etc/bashrc and ~/.profile /.bashrc or whatever the shell calls by default 04:25 < mous> `7hr34t_hvntr: in general yes 04:25 < copart> Dan39: the error on line3 is inside test.sh 04:26 < Dan39> thanks captain obvious 04:27 < `7hr34t_hvntr> thanks 04:27 < `7hr34t_hvntr> searching for switches on google doesnt have expected results 04:28 < pnbeast> You could check a man page and get something slightly more authoritative than random google results 04:29 < `7hr34t_hvntr> do people talk regex in here 04:29 < `7hr34t_hvntr> i see there is a regex channel 04:29 < Dan39> sure 04:30 < Dan39> not saying we'll be good as regex channel though 04:30 < `7hr34t_hvntr> i created a named group, then after it put some chars that must follow it. it seems to behave just like a lookahead, and extracts only my named group 04:30 < Dan39> makes sense... 04:30 < `7hr34t_hvntr> but its not a lookahead, so im wondering if there are some technical differences between this hackish implicit lookahead and an explicit lookahead 04:31 < Dan39> good question 04:32 < `7hr34t_hvntr> i try 04:32 < Dan39> i know matched grouped can get confusing, like with multiple nested 04:33 < `7hr34t_hvntr> regex says lookahead is useful for when you dont want to create a capture group 04:33 < sauvin> Which regular expression engine? 04:33 < `7hr34t_hvntr> pcre 04:33 < Dan39> `7hr34t_hvntr: so sounds like they are about the same 04:33 < Dan39> `7hr34t_hvntr: do some timing tests for us :P 04:34 < `7hr34t_hvntr> maybe when i get to the point that i dont rely on prayer to get my regex to compile 04:36 < sauvin> Sheah, whenever I think I need to use nested groups, I write test cases and print the results. They're... confusing. 04:38 < `7hr34t_hvntr> nested capture group? 04:39 < `7hr34t_hvntr> so creating alternate regex's but then also assigning them to separate capture groups 04:40 < `7hr34t_hvntr> so you could match on IP addresses but then create some nested groups to determine if the ip address is RFC1918 or not 04:44 < `7hr34t_hvntr> that only halfway makes sense, i'll just read up on it 05:14 < eh4> hi there I want to run this command /bin/flink run -p 20 -m cdc-hplx070-20.prod.wart.com:45660 /storage-cdc/wanauria/collector-0.0.1-SNAPSHOT.jar \ --run_date 2018-06-01 -Dconfig.file=/home/jvwa/flink-1.4.2-bin/flink-1.4.2/application.conf 05:14 < eh4> i tried this date +%Y-%m-`%d-1` 05:14 < fire2199> some knowledge about msmtp? 05:14 < eh4> the goal is i now whenever i run this command, the date should always be current date - 1 05:16 < copart> eh4: hard part is when the day is 1 05:17 < eh4> yeah copart 05:17 < pnbeast> You know you can't really back in time, even with Linux, right? 05:17 < alexey-nemovff> lol 05:17 < eh4> pnbeast: right 05:18 < triceratux> yep thats why i always execute scripts from the crontab instead of inline code so i can do some sane logging & set the path if necessary 05:18 < eh4> copart: hw abt the other days except 1 05:19 < siwica> What is the best way to recursively copy a directory somewhere else and subsequently set file permissions given that `install` doesn't have a -r flag? 05:19 < siwica> A combination of cp/find/install seems rather verbose 05:20 < pnbeast> siwica, set them how? You know about rsync, yes? 05:20 < copart> echo $(perl -w -e '@t=localtime(time-86400);printf "%d-%02d-%02d",$t[5]+1900,$t[4]+1,$t[3];') 05:20 < siwica> pnbeast: In a Makefile... 05:20 < [1]n1x> ™£∞º≠∑®†Ω¡ 05:20 < siwica> pnbeast: Basically I just want to install a python package via make. 05:20 < revel> APL? 05:21 < copart> eh4: got that almost from google, but changed to your format, seems to work 05:21 < eh4> copart: why perl 05:21 < [1]n1x> revel 05:21 < copart> because doing bath in BASH is ugly 05:21 < copart> math in BASH 05:22 < pnbeast> siwica, oh, beats me. I guess I'd try to copy someone else's working version! 05:23 < lessthan0> I see perl is not included in some distros now that makes me :( 05:23 < siwica> pnbeast: The problem seems to be pretty common. Isn't there a standard program capable of doing it? 05:23 < copart> you could do in BASH if you use the posix date time and subtract seconds, then use the date command again using that DT in your + format 05:23 < lessthan0> I really hate tab spaced programming languages like py 05:23 < lessthan0> there was absolutely nothing wrong with scripting in perl 05:24 < eh4> copart: does it work if the date becomes 1? 05:24 < pnbeast> lessthan0, citation needed. (tm) 05:24 < copart> eh4: it should since it is creating a datetime and usbtracking 86400 seconds, which is one day 05:24 < lessthan0> I can't remember the distro but I remember being angry about it 05:25 < eh4> sure 05:25 < copart> I do like Python more then perl... but both serve their own 05:26 < lessthan0> py is expanded to support embedded and compiled. perl is just perl. 05:26 < pnbeast> lessthan0, python does not need tabs AFAIK. One of us is wrong or you're conflating python's absurd white-space desire with the use of tabs. 05:26 < lessthan0> I call it tab spacing you call it white spacing 05:27 < lessthan0> programming tutorials I was reading said the IDE would through error flags if the spaces were removed 05:27 < lessthan0> *throw flags 05:27 < zlg> Is this channel a good place to ask for software recommendations? 05:27 < pnbeast> zlg, this channel is a good place to discuss Linux. 05:27 < lessthan0> and maybe software that runs on linux maybe 05:28 < zlg> pnbeast: That could be a looot of things, hence why I asked :) 05:28 < eh4> pnbeast: also why doing math with bash is horrible? 05:28 < eh4> any thoughts? 05:28 < pnbeast> zlg, it is a lot of things. "hence why" is redundant. 05:29 < copart> lessthan0 and pnbeast: blocks need to be indented... be it spaces or tabs.... 05:29 < pnbeast> copart, yes, as I indicated, above. 05:30 < copart> pnbeast sorry 05:30 < pnbeast> There's nothing to be sorry about. 05:30 < copart> if you ever want a fun read, read PEP8.... about how to style Python 05:31 < lessthan0> I will add it to the reading list thanks 05:31 < copart> it def will not make you like python more, lol 05:32 < sauvin> eh4, because math isn't bash's primary concern. 05:32 < lessthan0> I just googled and I find credible info that says python programs can not be made without the new line or carriage return if the program has more than one instruction 05:32 < lessthan0> that is why I loved perl so much 05:32 < copart> eh4: integer math works... but still have to be carefull... floats are impossible, unless you pipe 05:33 < lessthan0> you could do something on one line or multiple lines, you could indent or not indent at will 05:33 < zlg> copart: did they ever make a program that'll format it for you? As strict and dogmatic as PEP8 is, they'd be better served by an auto-styler like that. I think Go has `gofmt` for its own style. 05:33 < copart> lessthan0: I agree 100%, perl is great for oneliners 05:33 < lessthan0> using the ; instead of \n 05:34 < copart> zlg: I do not know, but PyCharm will highlight in some fashion every single infraction until you fix it 05:34 < copart> lessthan0: Python allows ;, its optional 05:35 < zlg> copart: ouch. Well I know which IDE to stay away from now. 05:35 < lessthan0> also interesting is that the python IDE will break and unbreak the code as you add spaces. even number of leading spaces = good (odd = bad) 05:36 < lessthan0> so you could have 9 spaces and you need to delete one to make 8 spaces (4 double spaces) but you can't count them that easy because they are invisible 05:36 < zlg> Even C and C++ can be fed through something like `astyle` so nobody has to worry about formatting problems. 05:36 < lessthan0> and it breaks the compilers 05:36 < Matviy> If uname -r says 4.15.0-20-generic, does that mean that ubuntu is using the stock kernel with no modifications? 05:37 < lessthan0> so it creates a real problem on monochromatic displays 05:37 < Matviy> Or do the ubuntu distro guys make kernel changes? 05:37 < zlg> Matviy: a lot of distros ship their own distro-specific patches, but they also tend to have a "vanilla" kernel package. I can't speak for Ubuntu but Gentoo certainly has both. 05:37 < jim> lessthan0, your python --version says? 05:38 < copart> Eclipse has a great C++ style corrector 05:38 < lessthan0> this was when I tried the official 2.7 IDE under win 7 05:38 < pnbeast> lessthan0, I don't like Python's spacing requirements, but in real life I just use vim and never have to deal with it except when I want to run a long line for a print statement or something odd like that. Why are you typing leading spaces? 05:38 < lessthan0> like a year ago 05:39 < jim> lessthan0, ohh :) how are things now, under linux? 05:39 < Pentode> man what a crazy night. O_O 05:39 < lessthan0> now I have perl and I can relive my childhood 05:39 < zlg> I never got around to learning Perl.. 05:39 * Pentode tells the bartender to make it a triple 05:40 < lessthan0> I can have both versions of py on linux installed by default 05:40 < lessthan0> if I need it but I probably wont 05:40 < jim> you mean py2 and py3? 05:40 < zlg> lessthan0: You'd be surprised, if you're talking py2 vs py3. 05:40 < lessthan0> 2.7 and 3 05:40 < zlg> There's a ton of stuff still using py2 05:40 < lessthan0> yes that is why I prefer 2.7 05:41 < Pentode> reliving my childhood would be something like trs-80 extended color basic 05:41 < lessthan0> in theory there should never be a use case that requires me to use py 3 05:41 < zlg> lessthan0: except for language features that 2.7 doesn't have... 05:41 < jim> I moved to py3 (and 3.6) awhile back 05:41 < lessthan0> I strictly avoid it until I have some embedded device that only does py 3 05:41 < lessthan0> haven't had that problem yet 05:41 < zlg> There's still `import x from __future__` or whatever, but that's dependent on Python devs. 05:41 < jim> do you follow Raymond Hettinger on twitter? 05:41 < M3rd> ... 05:43 < `7hr34t_hvntr> if im in sh and i do python -c ....pty.spawn("/bin/bash") 05:43 < `7hr34t_hvntr> ps aux is going to show python as running with that command line right 05:43 < ananke> `7hr34t_hvntr: try it and see 05:43 < TRS-80> Pentode: I'll join you in a drink 05:44 < lessthan0> Pentode are you a guitar player? hi-fi builder? tube enthusiast? 05:44 < `7hr34t_hvntr> and the sh wouldn't be reflected in ps auxxx with that python command line right 05:44 < zlg> yes, ps (+ *top and friends) will show the command issued to start a process. 05:45 < jim> well pentodes are generally signal tubes (not power tubes) 05:45 < Pentode> lessthan0, all of the above ;) 05:45 < lessthan0> I am a guitar pedal designer 05:45 < Pentode> im good at copying pedal designs ;0 05:45 < lessthan0> I sell kits of my original designs 05:45 < Pentode> ive never been able to design a pedal that i liked even after six months of tweaking, lol 05:46 < Pentode> you have a website? 05:46 < lessthan0> never advanced to the level of amp technician 05:46 < jim> lessthan0, have you been to your local makespace? 05:46 < lessthan0> www.eatyourguitar.com 05:46 < lessthan0> I am the makespace 05:46 < Pentode> lessthan0, it's all a breeze. designing a good pedal is not always a piece of cake either. 05:47 < jim> I should design enough pedals to make a whole cake! 05:47 < lessthan0> it takes time 05:48 < lessthan0> and you have to be able to put yourself in the feedback loop of development 05:48 < Pentode> lol 05:48 < lessthan0> you are evaluating and problem solving 05:48 < jim> lessthan0, who knows, you might find folks to sell kits to at your local make/hackerspace 05:49 < lessthan0> so if you don't know what guitar sounds you want or like and you don't know how to create that then you are screwed 05:49 < Pentode> i never got too involved in guitar amplification but i used to do a lot of simple hifi tube power amps 05:49 < lessthan0> it does well on the internet but not so good in real life 05:49 < jim> lessthan0, but more than one player prefers to just plug into an amp 05:49 < lessthan0> if I moved to austin I could sling guitar pedal kits and soldering iron on the corner to passers by 05:50 < Pentode> the thing is, amplifying / modulating a signal well is a lot different than purposely distorting / altering a signal in a way that's pleasant.. 05:51 < lessthan0> ironically I prefer to go straight in to a tube amp 05:51 < Pentode> moving to austin soon.. gonna be a guitar pedal tycoon... 05:51 < jim> what tube amps do you prefer? 05:51 < lessthan0> really? 05:51 < lessthan0> trainwreck 05:51 < zlg> man, I think I solved my issue. I was trying to get local mail filtering without an e-mail server. imapfilter filters remotely, but not locally... so I just put it before mbsync in my mail syncing script so that sorting happens remotely before I sync locally. Now I don't get duplicate e-mail. 05:52 < lessthan0> fender blackface 05:52 < Pentode> not really ;p 05:52 < lessthan0> zvex nano is not so bad 05:52 < lessthan0> I think we should also talk cabs because an amp head is just an amp head 05:53 < jim> agreed 05:53 < lessthan0> there is really no marshall that I like exactly 05:53 < Pentode> i've never been a marshall fan 05:54 < lessthan0> but if I had a http://www.diezel.ch/en/products/vh4.php I could get decent marshal sounds 05:54 * jim personally prefers ampegs, but not to lift 05:54 < lessthan0> I like a 4x12 with two different speakers in an x pattern 05:54 < lessthan0> ampegs are great on bass 05:55 < jim> yep 05:55 < jim> :) 05:55 < lessthan0> fender bassman is great for guitar and bass 05:55 < lessthan0> jensen speakers 05:55 < rasknikoff> 8-) 05:55 * jim plays a 5 string fretless 05:55 < lessthan0> but the ampegs have these fans in them that make it a live only thing 05:56 < lessthan0> I like a GK combo for small practice rig 05:56 < jim> transistor amp? 05:57 < lessthan0> I was thinking the tube amp can't remember 05:57 < lessthan0> its really big 05:57 < jim> oh ok, I had a gk transistor amp, sold it to someone who ran it as a bass amp and pretty much destroyed it :) 05:58 < lessthan0> your right it is solid state 05:58 < lessthan0> O just looked it up 05:58 < Evidlo> ahhhh, I've wasted an hour just trying to backup photos from my android 05:58 < lessthan0> the GK was solid state also 05:58 < lessthan0> but sounded amazing 05:58 < lessthan0> nice cab nice speaker nice ss amp 05:58 < sh1ro> you'd know if you had a tube amp it'll be drawing more power than your tower pc and cost as much 05:59 < Pentode> lessthan0, have you ever heard / tried any volition amps? 05:59 < jim> ok, we probably gotta move this discussion elsewhere 05:59 < lessthan0> nope 05:59 < Pentode> yeah probably 05:59 < Pentode> lol 05:59 < TRS-80> >not using FolderSync to automatically back up your photos every night to a self hosted NextCloud instance 05:59 < Pentode> it's all his fault 05:59 < sh1ro> no, just install linux on your amp 05:59 < lessthan0> join #guitar 05:59 < sh1ro> or better yet on your gibson 05:59 < jim> I didn't say it was anyone's fault :) 05:59 < lessthan0> agreed to move it 05:59 < Pentode> i know i did ;P 06:00 < Dan39> whats with that 1 negative FolderSync review lol 06:00 < jim> let's see if someone's there 06:00 < Dan39> "This app is good in theory & worked fine for a while to sync my phone photos with my Google Drive. But then one day 1.5GB of photos disappeared from my phone! After some digging around I found a sync history item which showed "0 deletions" in the summary, but clicking on "local file deletions" showed it had deleted hundreds of files! Given this is supposed to be an app to improve the resilience of my phone & 06:00 < Dan39> photo collection it's a pretty fundamental flaw!This app is good in theory & worked fine for a while to sync my phone photos with my Google Drive. But then one day 1.5GB of photos disappeared from my phone! After some digging around I found a sync history item which showed "0 deletions" in the summary, but clicking on "local file deletions" showed it had deleted hundreds of files! Given this is supposed to be 06:00 < Dan39> an app to improve the resilience of my phone & photo collection it's a pretty fundamental flaw!" 06:00 < Dan39> kind of worrying 06:01 < Dan39> woops double pasted 06:01 < Psi-Jack> rasknikoff: It's midnight. 06:01 < sh1ro> ._. wonder if the freenode ops would let me have zerocool for hacking a self tuning gibson to run linux 06:03 < pnbeast> I'll give you twenty bucks for a self-tuning Gibson. It doesn't even need to run Linux. 06:03 < Evidlo> are you guys hacking your gibsons? 06:04 < rasknikoff> ok its 12 here 06:05 < rasknikoff> hackers was a c000l flick 06:06 < Psi-Jack> Yeah, feck off dannylee. :p 06:08 < sauvin> Psi-Jack, for additional giggles, do a 'dict feckless'. 06:09 < [R]> is that anything like freckleless? 06:09 < cmj> crash/override 06:09 < sauvin> [R], get a dictionary. 06:09 < [R]> too heavy 06:10 < sh1ro> pnbeast: try 3k 06:10 < sauvin> Then remand yourself to the kiddies' table until you learn some English. 06:10 < [R]> lol 06:10 < pnbeast> sh1ro, okay, 30. 06:10 < Psi-Jack> sauvin: Heh 06:11 < sh1ro> ok but but i get to group zerocool and opal because reasons 06:16 < eh4> hello, i want to download just a particular file from git repo instead of downloading everything. tried curl -o 06:16 < eh4> but somehow the file is downloaded with html tags 06:17 < sh1ro> raw.github.com 06:17 < sh1ro> assuming microsoft didn't cock that subdomain up already 06:18 < Pentode> sh1ro, what, you think they would leave anything alone? lol 06:18 < mgolisch> why not? 06:20 < mgolisch> they didnt change alot with linkedin 06:21 < sh1ro> linked in was pretty lame to begin with they didn't have to do much to get it to their standard of lameness 06:26 < granttrec> if I have a script with #!/usr/bin/en python, how do I get it to run as python 2 06:26 < [R]> granttrec: what is your python binary called? 06:29 < jim> probably better to do #!/usr/bin/env python2.7 06:29 < granttrec> [R]: whats that? 06:30 < [R]> whats what... 06:30 < jim> that is a pronoun 06:30 < granttrec> [R]: the binary, isn't python interpreted I am not sure where the binary file would be 06:31 < [R]> python is interpreted by a program... 06:31 < [R]> so whats your python2 called... 06:31 < granttrec> Python 2.7.12 06:32 < [R]> thats the version... 06:32 < jim> can you do ls /usr/bin/python* 06:32 < jim> see what you have available 06:32 < granttrec> lrwxrwxrwx 1 root root 9 May 4 18:17 /usr/bin/python -> python2.7 06:33 < [R]> well, sounds like 'python' is python 2... so what makes you think you ahve to changea nything in your script? 06:33 < granttrec> just getting some errors for a project that I know has to use python 2 so that was my first thought 06:34 < [R]> lol 06:34 < jim> granttrec, ok, so you could probably do either one of these: #!/usr/bin/python2.7 or #!/usr/bin/env python2.7 06:35 < [R]> which is going to change a grand total of... nothing 06:35 < jim> that will make sure your python is version 2.7 06:35 < jim> yep, not gonna change a thing unless you install python3 too 06:36 < [R]> no sane dist is going t ocahnge what /usr/bin/python points to 06:36 < [R]> and also, no sane dist doesn't have python3 and python2 to begin wtih 06:36 < jim> unless in some years the python devs decide to deprecate python2 06:37 < [R]> a) thats been done ages ago 06:37 < [R]> b) no sane dist is going to cahgne /usr/bin/python out from under you 06:37 < jim> (and since it's pretty easy to move to python3, might be closer than you think) 06:37 < [R]> a NEW dist version might change it 06:37 < [R]> but nothing sane will jsut magically cahnge it 06:38 < jim> whether they do or don't shouldn't matter, if you're independent of it 06:38 * [R] shakes his hea 06:38 < [R]> d 06:39 < jim> that is, the programmer can choose to select the specific version he wants, and should do so 06:41 < jim> to take your statement to the next logical step, installing python3 but -not- installing python2 should not install an executable called python 06:42 < iflema> lol 06:42 < jim> maybe debian uses the /etc/alternatives system 06:43 < [R]> [09:36:42] <[R]> and also, no sane dist doesn't have python3 and python2 to begin wtih 06:43 < [R]> and debian doesn't use alternatives for python 06:43 < [R]> because... 06:43 < [R]> [09:36:04] <[R]> no sane dist is going t ocahnge what /usr/bin/python points to 06:44 < [R]> and it also shows that you ahve ABSOULTELY NO IDEA what alternatives even is 06:44 < [R]> because he showed python was directly pointing to python2.7 06:44 < [R]> if it was altenatives, it would point to /etc/alternatives 06:44 < [R]> you REALLY should quit while you're ahead 06:45 < eh4> hello there, is there a way to download a single file from the git repo 06:46 < eh4> i tried using curl -o 06:46 < eh4> but the file is getting downloaded with html tags 06:46 < eh4> any other better way of doing this 06:46 < [R]> eh4: click on the 'raw' link on the page 06:46 < sh1ro> eh4: raw.github.com that was for you dude 06:49 < eh4> sh1ro: i dont see raw.github.com. we have our private github 06:51 < Evidlo> eh4: go to the file you want in a browser. There is a "Raw" button at the top that will give you the link 06:52 < eh4> Thanks Evidlo 06:52 < Evidlo> also, does my message show up as "eh4 (IRC):"? 06:53 < eh4> nooo 07:07 < pantato> anyone know if anbox is supposed to work at all in ubuntu 14.04 ? 07:08 < [R]> 14.04 is old as dirt 07:08 < [R]> i doubt anything is 07:08 < [R]> haha 07:09 < pnbeast> I heard that several devs at Canonical said "Let's fix it so that anbox won't work in 2018, say, around summer time.". 07:10 < pnbeast> This was in 2012, so I don't know if it's true. It could just be a rumor. 07:10 < [R]> you're a rumor 07:11 * pnbeast gently palpates himself in the shower to check for rumors. 07:11 < pantato> haha it just crashes 07:12 < pantato> oh well i guess i have a reason to upgrade now... 07:12 < pantato> my PPA's tho 07:12 < [R]> the fact that its old as dirt wasn't reason enough? 07:12 < pantato> [R]: no 07:12 < [R]> lol 07:12 < pantato> why would I care if everything works? 07:12 < [R]> lol 07:13 < pantato> ? 07:13 < pantato> something i'm missing? 07:13 < pantato> pls tell 07:14 < [R]> its going to be eol very soon anyway 07:14 < [R]> cutting it a little close 07:14 < storge> does ubuntu ppa have a backports-like function/repo? 07:15 < [R]> a what? 07:16 < storge> like debian backports, a repo that provides security updates to older packages 07:16 < [R]> there is backports in ubuntu the exact same as there is in debain 07:16 < storge> because that makes 'why would i care if everything works' a bit more understandable 07:17 < storge> good 07:18 < Barcode> Anybody have time for a quick xmodmap question? 07:19 < pnbeast> Barcode, don't ask to ask. 07:19 < storge> can i ask a second question after this one? 07:20 < pnbeast> Okay, since it's you, storge. 07:20 < pantato> can i have a big brother? 07:21 < storge> pantato: i'll be your big brother, but we have to keep it on the downlow on account of my record 07:21 < pnbeast> pantato, do you have a nice house, a good stereo, at least two popular gaming consoles and a Mercedes? 07:22 < quint> I'm trying to set up rsync to work in it's plain un-tunneled mode. I keep getting "rsync: read error: Connection reset by peer (104) 07:22 < quint> " 07:22 < [R]> quint: and the command you are running... 07:22 < storge> take it up with your peers 07:22 < quint> rsync --daemon --config rsyncd.conf 07:22 * pnbeast has no peers. 07:22 < quint> lemme post the config one sec 07:23 < pantato> pnbeast: lmao i actually do 07:24 < pantato> pnbeast: all of those things 07:24 < quint> https://paste.linux.community/view/raw/b7f3ebbc 07:24 < pnbeast> Well, I'm a way better big brother than storge. 07:24 < [R]> quint: and the command you are running to get the error... 07:25 < pantato> pnbeast: i don't want you to be my big brother though because i can't respect someone who plays console instead of pc 07:25 < storge> hey hey now 07:25 < quint> the command i'm running: rsync -avr Pictures/ rsync://192.168.0.103/incoming 07:25 < [R]> and you're not using ssh beccause... 07:25 < quint> ever so slightly faster and there are a buttload of files 07:26 < [R]> well, its not faster, because itts not working 07:26 < pantato> lol [R] 07:26 < pantato> quint: try filezilla 07:26 < pantato> xD 07:26 < quint> but how come that doesn't work? 07:27 < quint> Is something misconfigured? The uid and gid are correct to the appropriate user, and the daemon is running as root 07:27 < storge> what's 'incoming' 07:27 < quint> the destination on the server 07:27 < quint> or so I think.. 07:28 < pantato> quint: what's the error? 07:29 < quint> pantato: https://paste.linux.community/view/raw/8ffcbc13 07:29 < storge> have you googled that exact error? 07:30 < quint> I did, but it seems that everyone has network related issues 07:30 < quint> In my case, the default port is unencumbered by firewalls 07:31 < [R]> you'd be done by now if you were just using ssh 07:31 < pantato> quint: why is your config file /home/*user*/incoming ? shouldn't it have the actual user name? or did you just put that there for security? 07:31 < quint> redacted it 07:32 < pantato> i'm with [R] on this one. Do what needs to be done and find out why rsync is screwing up later 07:32 < storge> if it's /home/user/incoming, shouldn't the command path be more than //ip/incoming? 07:32 < storge> like //ip:/home/user/incoming or something 07:35 < quint> storge: @ERROR: Unknown module 'home' 07:35 < dnanib> quint: you should use the --log-file=/tmp/rsyncd.log to log something on the server side too 07:37 < pantato> this sounds like a nightmare. I'm begging you to use ssh 07:37 < quint> I already am, but I really just want to know what the heck that's all about 07:37 < quint> I know I've done this before 07:38 < storge> to push: 07:38 < storge> rsync [OPTION...] SRC... rsync://[USER@]HOST[:PORT]/DEST 07:38 < storge> are you doing something like that? 07:39 < [R]> push it good 07:39 < [R]> push it real good 07:40 < Pentode> more cowbell 07:41 < pantato> [R]: thanks for that. now i've got salt n peppa up in my brain 07:41 < quint> storge: yep 07:42 < storge> rsync -avr Pictures/ rsync://192.168.0.103/incoming ...does not look like what the man page says 07:43 < storge> rsync [OPTION...] SRC... rsync://[USER@]HOST[:PORT]/DEST 07:43 < storge> see? different 07:45 < quint> But there is no user in this case 07:45 < pantato> quint: ?? 07:45 < storge> huh? 07:46 < storge> so there's just a net-facing filesystems called incoming right in the root of 192.168.0.103 ? 07:46 < pnbeast> Users always mess up my machines. Respect, quint! 07:47 < pantato> lol 07:47 < pantato> you guys always crack me up 07:47 < quint> In plain rsync, a user doesn't need to be specified. That's not a root dir, it's a module specified in the config file. 07:47 < quint> BOOM. 07:48 < quint> I dunno I'll just dick around with it. 07:48 < quint> I know I did this exact thing before and I'm sure I'll have a laugh when I find out i did something wrong 07:48 < dnanib> quint: I run rsync the way you are attempting to. It works just fine. Most likely a permission issue. You should enable logging at the daemon end to check. 07:49 < quint> dnanib: that's what I'm leaning towards, 07:50 < storge> i blame society 08:02 < cmj> rsync -avc 08:05 < dnanib> rsync -a doesn't honour destination permissions and might be probematic in some use-cases. Been bitten by that once 08:06 < [R]> its not honorable? 08:11 < dnanib> Let us say its honour is malleable based on how you interact with it. 08:15 < cmj> this is why i'm single 08:16 < pnbeast> You're touching yourself at McDonalds, again? 08:16 < Matviy> Is TempleOS debian based? 08:17 < [R]> Matviy: well what does the wikipedia page say? 08:17 < pnbeast> Matviy, it's a pretty even mix of Debian and Win 3.11, but with a BSD TCP stack. 08:17 < sh1ro> it's written from scratch by a lunitic 08:18 < [R]> sh1ro: lunatic... or misunderstood? 08:18 < pnbeast> lol 08:18 < Matviy> pnbeast: that sounds pretty crazy hmm 08:18 < sh1ro> both? 08:18 < pnbeast> Matviy, doesn't it? 08:19 < sh1ro> the guy's a genius he's just nuttier than squirl shit 08:20 < quint> He has some interesting ideas 08:20 < sh1ro> thath is not an untrue statement 08:22 < sh1ro> iirc he the whole OS runs kernel mode and is like LSD jesus oregon trail 08:23 < sauvin> It's "non-preemptive multitasking, multi-cored, public domain, open source, ring-0 only, single address space, non-networked" and "uses an interface similar to a mixture of DOS and Turbo C". 08:23 < sauvin> Sounds like a gut-wrenching retronightmare to me. 08:24 < sh1ro> oh yeah but all by one person which is undeniably impressive 08:24 < sauvin> Yes, if it works, it's quite an accomplishment. 08:26 < sh1ro> it works but i've only known someone to have used it seriously for a drinking game 08:28 < quint> So yeah. Never figured out the permissions thing with rsync. That's absolutely frustrating, but oh well. 08:28 < Barcode> Wait. Am I reading this correctly. "Holy C"? 08:28 < quint> The copy is in progress, anyway. 08:31 < sauvin> Barcode, yes, you read that right. I'm all "Holy crap!" 08:35 < pnbeast> Is that related to The Dude and the pope in the woods? 08:43 < notmike> Is that on-topic? Asking for a friend. 08:48 < Barcode> sauvin: Have you watched the video on the TempleOS webside? It's on another level, man. 08:50 < sauvin> No, I haven't. I value my sanity. 08:52 < sandman13> should one run load balacner, and upstream for load balancer on the same machine? Without any containers. Just plain services 08:58 < hendrix> Barcode: yea. that man is living in 2050 already. it's the os of the future 09:10 < notmike> Temple OS was clearly designed by a Divine intellect 09:13 < Barcode> He is "The smartest programmer who ever lived" 09:13 < pnbeast> He could parallel park a train? 09:14 < Barcode> If you give him a bible and ten years he could. 09:18 < Pentode> lol 09:20 < notmike> Smartest programmer who ever lived 09:20 * Pentode never knew cout was niggerlicious 09:20 < Pentode> learn something new every day o_O 09:24 < iodev> Barcode: don't mock, there is no "Thou shalt make a 16-bit operating system", how he came to this project is completely absurd. 09:28 < Dagmar> Now, now. It's not nice to pick on the mentally disabled. 09:30 < sh1ro> we're not talking about adobe devs? 09:32 < Dagmar> You got any other explanation as to why it took him so long to do something college kids do in a semester for credit? 09:32 < Dagmar> ...and why the @#$@# it doesn't have process isolation? 09:33 < Dagmar> "Security" on that apparently comes down to "God will protect my data" 09:33 < sh1ro> there's no networking and it doesn't run other programs 09:37 < storge> he IDE that comes with TempleOS supports several features, such as embedding images in code. 09:38 < storge> whut 09:38 < storge> A file can have, for example, a spinning 3D model of a tank as a comment in source code. 09:38 < storge> whut 09:39 < kma> I am working on Centos7 , i stopped "sdb" through echo 1 > /sys/block/sdb/device/power/delete , how can I rerun it ? 09:42 < Nilby> Does anyone know the reasoning why cp won't ever overwrite file, even with -f ? 09:44 < sauvin> Nilby, I can't imagine any circumstance under which it would be logical to copy a directory structure into a *file* with exactly one exception: if is actually a mountpoint to some other structure of a type not markedly dissimilar to an fs. 09:47 < jim> pantato, https://forum.filezilla-project.org/viewtopic.php?f=2&t=48441 09:50 < Nilby> Hmmm. Well my case is I just want to re-use the name for a copy of directory and I don't care about the contents. Even with --remove-destination it refuses. I just have to rm and then cp -a . I just don't understand the reason to restrict it even with -f. 09:50 < Nilby> (thanks for answering though :) 09:53 < Nilby> I suppose the designers thought it must be a mistake, so don't do it. 09:57 < oiaohm> Nilby: that one where directory will not over write file. That could be in vfs as well. https://www.tldp.org/LDP/khg/HyperNews/get/fs/vfstour.html Look at inode operations. 09:57 < storge> jim: interesting, thanks 09:57 < oiaohm> Nilby: file and directory are two different commands on file systems. So attempting to perform a directory command on a file should fail. 09:58 < oiaohm> Nilby: so the copy program would require special logic for that case. 09:59 < Nilby> sudo sudo cp -a -f --clobber --remove-destination ==> "Thou shall not pass!" 10:00 < storge> sudo sudo. sounds important. 10:01 < pingfloyd> maybe add in a few su for extra spice while at it 10:01 < MrElendig> and 5 pkexec 10:02 < pingfloyd> Nilby: why are you trying to overwrite a directory with a file? 10:03 < Nilby> pingfloyd, It's the other way around. I'm trying to overwrite a file with a directory. 10:03 < Nilby> It's just the file is junk and I want to re-use that name for a copy of the directory. 10:03 < hexnewbie> It's little known, but saying sudo three times grants you super sudo, which is the equivalent of yes | -fff --yes 10:04 < pingfloyd> how come you can't rm the file and then cp the dir? 10:04 < sauvin> Super sudo is what gets the nerds and the geeks some really yummy babes! 10:04 < MrElendig> also turns your hair a spiky yellow 10:05 < pingfloyd> rm /path/to/file && cp -a /path/to/sourcedir /path/to/dest 10:05 < Nilby> I can. I'm just wondering what's reasoning why "cp" can't do it for me. 10:05 < hexnewbie> If you're super nerdy, though, it just makes your computer gain sentience 10:05 < pingfloyd> Nilby: because cp is for copying and not removing 10:05 < sauvin> Think I'd rather have the babes. "sentience" doesn't mean what you seem to think. 10:05 < quickquick> good morning =) 10:05 < pingfloyd> there's a tool for removing files and dir 10:06 < storge> grep -i beyond /usr/src/linux-4.17.2/arch/mips/pci/ops-bridge.c 10:06 < storge> (fun for kids) 10:06 < jim> hi 10:06 < Pentode> hihi 10:06 < Nilby> I agreee. Except for the very normal case of cp , it "removes" just fine. 10:06 < sandman13> if you want cp to do that for you, use mv 10:07 < hendrix> chaining many sudos, su's and pkexecs makes command's powerlevel over 9000 10:07 < storge> two sudos and a su = developer mode 10:07 < hexnewbie> Nilby: Because it's an ambiguous destructive situation. Overwriting a file is common, but anything more complex is beyond the scope of cp. Typically you use rsync in many of those situations, even when cp can do them, I'm almost sure rsync had an option to handle overwriting files with directories and vice versa (perhaps --delete-before? although it has failed for me in that situation sometimes) 10:07 < pingfloyd> with cp, it's more that it is overwrite by default instead of prompting you or refusing by default 10:09 < pingfloyd> also, if the user is trying to overwrite a file with a dir, the only safe and smart assumption is that they don't realize they're trying to overwrite a file with a dir (something that nobody ever does)> 10:09 < Nilby> It's true. Maybe I'm just asking for too much ambiguous destruction power. 10:09 < storge> i think you should write the cp authors and interrogate them as to why they didn't enable their tool to perform arcane functions that you want 10:09 < pingfloyd> as a programmer you're forced to make assumptions at some point 10:09 < pingfloyd> you want error on the side of caution 10:09 < storge> to err 10:10 < pingfloyd> but you don't want to be so bad about it that you become like Windows where it annoys you with second guessing. 10:10 < oiaohm> Nilby: also funny part is you really do need to talk to the cp developers because this could be a simple coder error not including the extra code. 10:10 < oiaohm> Nilby: you test case applications sometimes you miss really odd things. 10:10 < pingfloyd> yeah, you never know 10:11 < storge> i blame society. society made them code cp that way. 10:11 * storge punches fist into palm 10:11 < oiaohm> storge: over the years there have been some really strange cp errors. 10:11 < pingfloyd> saying this seems like a corner case, that a user would need it to be able to act that way, is kind of an understatement. 10:12 < pingfloyd> they need it to where -i has a "yes to all" 10:12 < storge> they might also want to avoid accidental lunacy like cp / ~/foo 10:12 < pingfloyd> like -i is handy for sometimes catching oversights, but it tends to get in the way too much 10:13 < hexnewbie> storge: Why, no. Your lunacy is totally to my liking. 10:13 < quickquick> can somebody help a beginner with dumb questions in pm please? :/ 10:13 < storge> quickquick: ask here 10:13 < Nilby> Yeah. I'm gonna hop in the time machine back to 1970 bell labs and argue this dang issue. Then we'll see who has the power to do stupid directory copying! 10:13 < Dagmar> Sure. We don't do pm's tho 10:13 < Dagmar> Dumb questions are generally not a big deal 10:13 < storge> cp /dev/zero ~/foo 10:14 < pingfloyd> Nilby: except we use gnu cp 10:14 < storge> [SATIRE, don't run that or course] 10:14 < pingfloyd> Nilby: what if it turns out that sysv cp supported it? 10:14 < Nilby> Even better. I'm sure RMS will argue this. 10:15 < storge> thank the gods for rms 10:15 < Nilby> pingfloyd, That would be very interesting. 10:15 < oiaohm> Nilby: if you read over the man page for cp 'if an existing destination file cannot be opened, remove it and try again (this option is ignored when the -n option is also used)' this is force notice it does not mention anything about directories. 10:15 < oiaohm> Nilby: it seams like a problem no one considered. 10:15 < hexnewbie> quickquick: PMs are at the detriment to both parties - the helper is singled out and has to give more of their time and attention, and the helpee may be abused by a total stranger giving them destructive commands. Even if they don't, advice is best when verified by many :) 10:15 < storge> many eyes 10:15 < pingfloyd> I think the unix argument is, "it's rm's job" 10:16 < Dagmar> This way, everyone can see if you're being given bad advice 10:16 < [R]> hexnewbie: kinky 10:16 < hexnewbie> storge: We've got cameras in every corner, and even the bathrooms, so I think we got the open source eyes covered. 10:16 < storge> [R]: pm? let's emote 10:16 < pingfloyd> Nilby: watch it backfire and instead they make it so cp can never clobber at all. 10:16 < [R]> storge: haa 10:16 < quickquick> thank you storage and Dagmar :) i want to be anonymous so i did a script that: 1-kill processes like xchat firefox etc... 2-stop networking&wifi cards 3-macchange 4-change hostname 5-clean dhcp 6-clean iptables 7-start networking&wifi cards 8- launch OpenVPN and connect to VPN 10:17 < pingfloyd> because they remember "it's rm's job" 10:17 < quickquick> buy i see dns leaks 10:17 < oiaohm> pingfloyd: cp already has a flag for that for files. 10:17 < quickquick> what i have to do? change resolv.conf? 10:17 < Dagmar> Yes 10:17 < pingfloyd> oiaohm: I know. 10:17 < quickquick> or more steps? 10:17 < Dagmar> It *has* to be directing DNS queries through the VPN 10:17 < oiaohm> pingfloyd: please note I said for files. The documentation does not say anything about directories. 10:17 < oiaohm> pingfloyd: lets say faith. 10:17 < pingfloyd> oiaohm: I'm saying, watch his time travel plan backfire, and they decide to remove such flags. 10:17 < storge> clean iptables?? 10:18 < Dagmar> Basically, standard rule for connecting to a network is that you *must* take on the DNS scope equal to their network or things like "DNS leaks" happen 10:18 < [R]> Dagmar: do you like to go "through" a vpn? 10:18 < Dagmar> Since you're connecting to * through the VPN, it means the correct thing to do is take _their_ DNS settings exclusively 10:18 < hexnewbie> Nilby: Without having looked at cp's code (recently or ever), I assume it mostly makes stat(), read(), open() and write() calls. It probably overwrites files by writing over them, not by deleting them. So adding unlink() calls to a copy command may seem unreasonable 10:18 < [R]> storge: you know, with like a cloth 10:18 < storge> pingfloyd: we already know the time travel plan backfired or it would be different 10:18 < nekoseam> yay 10:19 < pingfloyd> storge: that's true 10:19 < Dagmar> [R]: No the major eff-up that happens _frequently_ is that people will still be using their ISP's DNS server settings 10:19 < Nilby> cp --> "Not my job, Bro. Go ask rm." 10:19 < hexnewbie> Well, hopefully fstatat() and openat(), cause else it would be a security hole. 10:19 < Psi-Jack> Anyone here happen to know what NetworkManager needs in order to actually properly setup a hostapd hotspot? 10:19 < pingfloyd> storge: I just went back and invented Pnix before Unix and now you're all running Pnix-likes. 10:19 < Dagmar> Psi-Jack: "not be involved" 10:20 < storge> [R]: iptables --flush --wipe --closethelid 10:20 * Psi-Jack rolls his eyes. 10:20 < Dagmar> I'm not even kidding 10:20 < [R]> storge: shouldn't you wipe then flush? 10:20 * [R] giggles 10:20 < Dagmar> NetworkManager is not even a little bit your friend when it comes to running an AP 10:20 < storge> [R]: i like to leave a surprise on the handle for the next bloke 10:20 < [R]> storge: ROFl 10:21 < pingfloyd> you're that dirty guy 10:21 < storge> [R]: do it right before lunch and see how many people get sick. 10:21 < Psi-Jack> Well, I'm still trying to regardless. I've tested and verified though that I can setup an AP with create_ap, and it works like a charm on this ath9k mPCIe I have in my laptop for testing this. 10:22 < Dagmar> You people are the very reason I'm glad hax0ring has given me the bladder of a large dromedary 10:22 < pingfloyd> they'll all end up suing Jack in The Box 10:22 < storge> "you people"? 10:22 < Psi-Jack> I'm trying to figure out what it takes, though, to get NM to set up the hotspot to understand it. See if it can as it should be able to, and what it requires to get it to. 10:22 < sauvin> He means the "Great Unwashed". 10:22 < Pentode> yeah who are you people 10:22 < storge> oh, well that's me. 10:22 < Pentode> no we _know_ what he means 10:23 < Dagmar> Psi-Jack: Simpler and cleaner to just set up hostapd and dhcpd as services, and tell NM to leave that interface the hell alone for the duration 10:23 < Pentode> ;p 10:23 < sauvin> I was being polite. :D 10:23 < Psi-Jack> Not helpful, Dagmar. 10:25 < Dagmar> You'll find this basically repeated in several wikis 10:25 < Dagmar> It literally just gets underfoot 10:25 < Dagmar> Also, hostapd and dhcpd are actually pretty easy to set up, even with WPA2 10:26 < Psi-Jack> Yeah, like I said, I've gotten it setup with create_ap already for testing that side of things. 10:26 < Psi-Jack> I'm trying to first-hand test the NM side of things. 10:28 < Dagmar> It's also kind of outside of the scope of NM, but with systemd being taken seriously I suppose it's also in scope with whales 10:29 < pingfloyd> what about networkd? 10:29 < pingfloyd> systemd should steal Apple's old slogan 10:30 < domhnall> SystemBSD 10:30 < Psi-Jack> Actually, NM itself does have support to run an AP in it. :p 10:30 < Psi-Jack> So, it's well within scope. 10:30 < Dagmar> good luck with that then 10:30 < pingfloyd> doesn't nm use wpa_supplicant 10:30 < pingfloyd> for that 10:30 < Dagmar> wpa_supplicant won't present an AP to clients 10:30 < pingfloyd> nm here, depends on wpa_supplicant 10:31 < Psi-Jack> I believe it does. and dnsmasq as well. 10:31 < Dagmar> You *need* hostapd to make the wireless card transmit as an AP, otherwise you're stuck with ad-hoc mode 10:32 < pingfloyd> ham-fist mode 10:32 < Dagmar> I literally set these things up on RPis in like, ten minutes flat 10:33 < Dagmar> Image Sd card, boot Pi, login, enable hostapd and dhcpd, five minutes of typing, done. 10:33 < pingfloyd> as in use them as wifi aps? 10:33 < Dagmar> Yes. 10:33 < Psi-Jack> Heh, that seems pretty silly to me. heh 10:33 < pingfloyd> that's a great idea 10:34 < Dagmar> RIght so, you know how the Nintendo DS has this "StreetPass" functionality? 10:34 < pingfloyd> for something useful to do with those 10:34 < Dagmar> When your DS gets near another DS they trade magic cookies or something, and you get occasional perks in games for it? 10:34 < Dagmar> ...which in Japan works because high pop density, but in the US... Nearly never 10:35 < dgurney> no, that's the 3DS 10:35 < Dagmar> So they did a thing where if you took the unit near basically, _A McDonalds_ it would pick up data from the past _six_ people who were there 10:35 < pingfloyd> probably would work great in Seoul SK too 10:35 < pingfloyd> they're all game junkies there too 10:35 < Dagmar> The StreetPass Relay thing. It's entire thing is based on the mac address of the AP your unit talks to 10:36 < Dagmar> So, you can go to wigle.net, dig up the MAC addresses for wherever the heck you like, and spoof them with an RPi. 10:37 < Dagmar> On a lark I'd written a MAC cycler and was sending 3DS notes to a friend whod moved to Jersey 10:37 < Psi-Jack> Hmm. I don't even see where/how NM is even trying to start the WNIC in AP mode. grr. 10:37 < Dagmar> heh 10:37 < pingfloyd> https://howtoraspberrypi.com/create-a-wi-fi-hotspot-in-less-than-10-minutes-with-pi-raspberry/ 10:37 < pingfloyd> is that the method you use? 10:38 < Dagmar> Probably. It's made incredibly simple on the Pi's simply because it's 100% guaranteed that the wifi chip on there is supported by hostapd 10:38 < Dagmar> There's only like a million wireles dongles out there that _fail_ at that 10:38 < Dagmar> It's a major PITA when you're trying to actually set up your own AP 10:39 < pingfloyd> it's kind of back to basics 10:40 < Dagmar> Heh. I didnt' even know they had a web GUI 10:40 < pingfloyd> yeah, the web gui sounds like bloat 10:40 < MrElendig> buy a proper ap, plug it in, done 10:40 < MrElendig> (and get much better performance) 10:40 < Dagmar> Then wait for the binary firmware patches to come out to remove the backdoors they just can never seem to resist putting in 10:41 < MrElendig> Dagmar: just don't but dlink trash :p 10:41 < Dagmar> I can't even trust NetGear anymore 10:41 < pingfloyd> seems like they always come with routers these days 10:41 < MrElendig> I never trusted netgear 10:41 < MrElendig> ubiq/mikrotik/etc 10:42 < MrElendig> netgear are generally just accidentally malicious instead of intentionally malicious though 10:42 < MrElendig> unlike cisco/dlink/belkin 10:44 < MrElendig> also supports auto-update now which is nice, but sadly not enabled out of the box 10:45 < MrElendig> would help with people never bothering to update their router firmware 10:45 < Psi-Jack> heh 10:45 < Psi-Jack> Netgear... I've brushed my feet of Netgear. 10:45 < MrElendig> their small switches are quite nice 10:45 < Psi-Jack> I never used D-Link, but I also never will. 10:47 < MrElendig> dlink had a backdoor and then instead of removing it they just "hid" it behind port knocking 10:47 < MrElendig> which took about half a day to be discovered 10:49 < Pentode> oh thats cute 10:50 < Pentode> did they really think that would go unnoticed? 10:50 < Psi-Jack> So hence. Port Knocking is not security. Hahah 10:51 < jim> In case any of you use FileZilla, the main author again packs a malware installer ... https://forum.filezilla-project.org/viewtopic.php?f=2&t=48441 10:51 < Pentode> nice. 10:52 < Pentode> i was using it on an older macbook since it was the best thing i could find that'd still run on 10.7 10:52 < V7> Hey all 10:52 < Pentode> hi 10:52 < V7> Made all like in post was written 10:52 < V7> https://ubuntuforums.org/showthread.php?t=1188099 10:52 < V7> Got: sg: failed to crypt password with previous salt: Invalid argument 10:53 < V7> After executing a script 10:54 < V7> Is it because of sg msut be ran by root ? 10:54 < dgurney> filezilla is a great program, but why does the developer insist on using junk bundle installers on Windows? to me it just seems like a way to tarnish your reputation for no reason 10:56 < SuperSeriousCat> For money of course 10:56 < V7> dgurney: This program is free. He spent a time to make it and also he might staying on air to listen for a bugs and so on. It's quiet usually that some poeple wants money from their projects. So, as a result, he added ads. 10:56 < Pentode> they make it hard to resist. its tough making money as a share/free-ware developer. 10:56 < Pentode> heck it's tough in general 10:57 < s_gautam> how do I make the cursor blink in xterm? 10:57 < V7> Of course, this is Open Source, but ... I have no ads on Linux :) Window's installer doesn't bother me. 10:58 < dgurney> yes, but ads don't need to be terribly implemented 10:58 < V7> dgurney: indeed 10:59 < dgurney> I have no problem with ads, so long as they are implemented right 11:00 < SuperSeriousCat> If they are just at the installers, it is not terrible implemented. Just a 1 time thing 11:01 < dgurney> triggering anti-malware or antivirus is terrible 11:01 < SuperSeriousCat> You dont need that kind of programs anymore. It is just bloat 11:01 < nekoseam> why do so many distros ship with ugly flat themes instead of themes that attempt to put detail into it? 11:01 < SuperSeriousCat> The Windows one and common sense is more then good enough 11:02 < dgurney> you're saying that you don't need an antivirus program, and then you say that another antivirus program is good enough 11:02 < dgurney> I agree with you fully, but that's a bit of a silly way of saying it 11:32 < nai> hi, i noticed that my partitions' device names are in the wrong order, i.e. /dev/sda6 comes before /dev/sda5 on disk. not that this is too much of a problem, but i'd like to know how it's possible? i thought those device names were assigned at boot time? 11:37 < nai> does it just mean that the partitions have been written in this order in the partition table? 11:38 < nai> yeah i think that explains it. 11:46 < Matviy> ffs how long does it take to compile the kernel 11:46 < dgurney> depends 11:47 < dgurney> if you use a distro config, it can take hours, but if you use a small custom one, usually some minutes 11:47 < dgurney> naturally, it also depends on what you're compiling on 12:02 < kubast2> Hey why does free -h shows 100MB more being used than sys/block/zram0/original data size file ? 12:03 < kuri0> how do i reset mokmanager keys ? sudo mokutil --reset doesn't do anything 12:04 < kubast2> As in used in swap 12:05 < kubast2> Compressed data size is around 150MB and it matches used and it matches total-used-zramCompressedSize = Free 12:06 < kubast2> But the ammount of swap and ammount of original size of zram doesn't count up 12:06 < V7> Is it okay that /etc/sudoers is not writable via root after resume after systemctl suspend ? 12:09 < V7> It says read only file system 12:10 < atmx> erzak [~Superzak@a83-162-211-144.adsl.xs4all.nl] has joined ##linux 12:10 < atmx> [11:05] <-- | Drakonan [~Drakonan@184-156-111-145.dyn.centurytel.net] has quit (Read error: Connection reset by peer) 12:10 < atmx> [11:05] <-- | hotwire007 [~hotwire00@49.207.61.76] has quit (Ping timeout: 256 seconds) 12:10 < atmx> [11erzak [~Superzak@a83-162-211-144.adsl.xs4all.nl] has joined ##linux 12:10 < atmx> [11:05] <-- | Drakonan [~Drakonan@184-156-111-145.dyn.centurytel.net] has quit (Read error: Connection reset by peer) 12:10 < atmx> [11:05] <-- | hotwire007 [~hotwire00@49.207.61.76] has quit (Ping timeout: 256 seconds) 12:10 < atmx> [11erzak [~Superzak@a83-162-211-144.adsl.xs4all.nl] has joined ##linux 12:10 < atmx> [11:05] <-- | Drakonan [~Drakonan@184-156-111-145.dyn.centurytel.net] has quit (Read error: Connection reset by peer) 12:10 < atmx> [11:05] <-- | hotwire007 [~hotwire00@49.207.61.76] has quit (Ping timeout: 256 seconds) 12:22 < Pentode> ...and? 12:25 < V7> Anyone ? :) 12:28 < kuri0> how do i reset mokmanager keys ? sudo mokutil --reset doesn't do anything 12:45 < fire2199> what's the channel for c language? 12:49 < Pentode> #C? 13:17 < rcf> V7: that implies that nothing else on the filesystem is writable, so it's not going to be OK for too long. 13:19 < domhnall> fire2199: ##programming is pretty decent 13:23 < brutser> hi, in my laptop there is "Qualcomm Atheros QCA8172" but the link is going down the whole time and i am losing the connection. The laptop is lenovo g505, not the newest laptop, but ok. how can i troubleshoot the eth problem? 13:23 < fire2199> domhnall : i found ##c 13:25 < domhnall> fire2199: great 13:26 < brutser> just small output to give idea: https://pastebin.com/7TKRU3j1 13:29 < domhnall> brutser: wired or wireless connection? but maybe - check the cable(s), ping and ping6 test (looking for packet loss), and/or check if interfaces are sharing same ip, i.e eth0 and wlan0 both have 10.x.x.XXX. 13:30 < brutser> wired connection: cables already tried 4 others, new ones, ping no packet loss, interface just goes down 13:30 < domhnall> hm, what's your manager? 13:31 < domhnall> that pastebin, can you repaste to dpaste or other, please. 13:33 < brutser> domhnall: http://dpaste.com/2QBRR4J 13:35 < domhnall> brutser: that's just pci info. and how long does the connection last before dying? 13:36 < domhnall> hmm..this might be your issue, enp1s0: renamed from eth0 13:37 < brutser> domhnall: yes i thought about this too, but the laptop got a default install with debian stretch, i got nothing else configured left it as default 13:40 < domhnall> is eth0 or enp3s0 in /sys/class/net ? 13:41 < brutser> enp1s0 13:43 < BluesKaj> Hiyas all 13:46 < Pentode> brutser, it may seem silly but have you checked the cable? does ifconfig show any dropped packets, errors or overruns? 13:47 < brutser> Pentode: yes cabled replaced 4 times with brand new ones 13:47 < domhnall> brutser, dpaste - dmesg |grep eth0 (ommit your macaddress too please) 13:47 < domhnall> then same for enp 13:48 < brutser> domhnall: ok sec 13:50 < brutser> domhnall: http://dpaste.com/06SCRPS 13:50 < domhnall> brutser, my 'guess', you might have two network managers running. NetworkManager and networking might be conflicting. One is winning. 13:51 < brutser> domhnall: yes i was also thinking that direction, how can i solve this? just uninstall every networkmanager i can find installed? 13:51 < Pentode> oh yeah that is very possible 13:51 < Pentode> that'll cause some link flapping ;p 13:52 < Pentode> _usually_ things like the network manager applet will indicate something else is managing the network but perhaps not always. 13:52 < brutser> Pentode: you mean uninstalling the nw managers , or what we talked before? 13:52 < Pentode> maybe remove any network managers you arent using 13:53 < Pentode> yeah, get rid of whatever you aren't using to manage the network 13:53 < domhnall> brutser, nah, no need to uninstall, just stop the dang thing. like stopping NetworkManager. 13:53 < brutser> domhnall: how can i see which managers or other processes are trying to control the network config? 13:54 < domhnall> 'services --status-all' 13:56 < domhnall> + is active 13:56 < brutser> http://dpaste.com/1NTA8JV 13:57 < brutser> i guess only the network-manager should be disabled then? 14:02 < domhnall> that's the safest one. you can also stop it from starting automatically on startup. use systemctl disable, or stop depending on how you want to handle it. 14:03 < domhnall> brutser, this is all for troubleshooting though. 14:03 < domhnall> could be hw related afaik. 14:08 < pankaj> I am designing web pages. How to download and use fonts in linux for it. I generally prefer microsoft fonts but I am open to open choices also. I also tried typecatcher but later realised that most of the fonts in typecatcher had no differences. 14:15 < pankaj> Hello, Is their anybody here? 14:16 < rcf> Maybe. 14:17 < compdoc> who wants to know? 14:18 < Pentode> pankaj, you can just stuff ttf fonts in ~/.fonts/ 14:18 < Dr_Willis> I just got here. 14:20 < domhnall> Isn't that what they always say in movies before a jump scare? 14:21 < Dr_Willis> Been Playing Subnautica all week, i think i am Immune to jump-scares now. 14:26 < domhnall> pankaj, those tff Pentode mentioned http://fontarchive.com/html/circular-fonts-1.html , not sure how safe to dl from link but, titles starting point. 14:26 < Pentode> if he has a windows installation he could just yank them from there also. 14:26 < domhnall> ... meant to post this link http://fontarchive.com/ 14:27 < Pentode> bah @ windows. i don't miss having to mess with windows95, lol. 14:28 < Pentode> im trying to get this old laptop running again and a fresh install of win95 just decides it's going to hang un-provoked at boot for seemingly no reason, lol 14:28 < compdoc> just as well, win95 being so horror show 14:28 < Pentode> and of course it's idea of verbose is a black screen with a cursor ;p 14:29 < domhnall> pankaj, even better one https://fonts.google.com/ 14:33 < domhnall> last one and im done https://fontawesome.com/?from=io 14:37 < domhnall> ^this is most distro's repositories 14:38 < domhnall> s/is *in 14:38 < pankaj> domhnall: OK 14:40 < birkoff> how can I print a binary file to stdout with all its unique characters ready to be copy-pasted ? 14:41 < birkoff> on netcat 14:46 < domhnall> why are trying to print binary files, usually you'd open them. 14:46 < MrElendig> birkoff: you can't 14:47 < MrElendig> your terminal will interpet the binary data in a non-expected way 14:47 < MrElendig> also there can be characters that are non-printing etc 14:48 < MrElendig> birkoff: what is the actual use case here? 14:50 < kuri0> how do i reset mokmanager keys ? sudo mokutil --reset doesn't do anything 14:50 < pankaj> domhnall: I am downloading fonts form fonts.awesome. I rarely know how to use it. Like I downloaded google fonts and when opening they get installed automatically but in this case I suspect what to do? 14:52 < MrElendig> pankaj: somewhat distro dependent 14:53 < MrElendig> if you have a up to date fontconfig then dump them in ~/.local/share/fonts 14:54 < pankaj> MrElendig: Is their not any standard procedure common to all linux destros. 14:56 < MrElendig> pankaj: no, because a lot of them still uses an old fontconfig 14:56 < MrElendig> or do their own thing just because *coughDEBIANcough* 14:56 < MrElendig> read the docs for your distro 14:57 < pankaj> MrElendig: OK 14:59 < rootkea> Hi! Is there any way to listen NPR's live radio station "All Songs 24/7 Music" in Quod Libet? ref: https://www.npr.org/music/allsongs247/ 14:59 < rootkea> I'm thinking about adding an Internet station in Quod Libet. Only if I could figure out the stream url :( 15:00 < sandman13> if anyone here has decent experience with load balancing, could you please tell if it's a good idea to run load balancer, and the upstream service on the same machine 15:00 < pankaj> MrElendig: I got ttf-font-awesome for my destro. 15:01 < MrElendig> pankaj: if you got an actual package then just install it using your package manager 15:05 < Dagmar> sandman13: That depends on how serious you are about uptime 15:17 < LissajousPattern> I am so happy right now with linux. 15:18 < rootkea> Update - I used Firefox's Dev tools (Network monitor to be precise) and found the url! Now listening to live NPR station in Quod Libet :) 15:20 < domhnall> rootkea ++ 15:21 < mox1> What's the C++ questions IRC channel? 15:22 < MrElendig> /q alis help list 15:29 < Dan39> LissajousPattern: please do tell us more 15:32 < LissajousPattern> Dan39, well... I just really like it is all. 15:32 < Dan39> wow, such a good reason! 15:33 < Dan39> -_- 15:33 < LissajousPattern> and my neighbors cat's name is Tux. 15:33 < LissajousPattern> and the fact that I can make A.I. fart on command using linux 15:34 < NGC3982> i want to use my old android phone to create a live feed on my web site. any suggestions? 15:34 < NGC3982> i already have the "picture part" where a tasker profile saves a bunch of pictures. 15:34 < Dan39> i bet it's been done before NGC3982 15:34 < NGC3982> yes. 15:34 < LissajousPattern> its being done every day. 15:35 < NGC3982> that's cool. 15:35 < Dan39> LissajousPattern does it on his back 15:35 < NGC3982> so the suggestion is that someone else can answer my query? ;) 15:36 < Dagmar> Or you could just use Google 15:36 < Dan39> oooh, google!? is that new? 15:36 < NGC3982> sure i can. but why do you think i came here in the first place? 15:36 < LissajousPattern> hahaha 15:36 < Dagmar> Because you're an entitled jerk? 15:36 < LissajousPattern> and google is for googlers 15:36 < Dan39> to tempt us to sin by working on the sabbath! 15:37 < LissajousPattern> thou shalt not 15:37 < NGC3982> Dagmar: what is your problem? if you find that ##linux does not allow general discussions about how to do things even though some else has already done it, id suggest you explain that to me instead. 15:38 < Dan39> damn son 15:38 < LissajousPattern> NGC3982, what kind of experience do you have with web development? 15:38 < LissajousPattern> or linux for that matter? 15:38 < Dan39> LissajousPattern: i would want a live feed, not picture updates 15:39 < NGC3982> LissajousPattern: more linux than web development. i have an apache based site that does not contain much. 15:39 < Dan39> get some kind of a video casting server going IMO 15:39 < NGC3982> somewhere in the back of my head i think motion did perhaps have some direct-link cajigger to apache 15:39 < LissajousPattern> well it sounds like a cool challenge I already have a template that I am messing with 15:39 < NGC3982> cool! 15:40 < Dan39> cajiggeR? is that a technical term? 15:40 < NGC3982> absolutely not :-p. 15:40 < Dan39> was seriously just checking 15:40 < NGC3982> i dont really need it to be a "video" since the purpose is to monitor hydroponic plants. 15:41 < Dan39> oh 15:41 < NGC3982> Dan39: ah, sorry. i think the word can be used to describe "a thing that i really do not understand". 15:41 < Dan39> well why not just have it overwrite a file that is in your document root? 15:42 < NGC3982> i guess so. i tried that last time, but i noticed that the phone (with the tasker app) was not keen on snapping a picture every minute. it got super hot and did not seem to want to sleep in between photos. 15:42 < NGC3982> but that is of course an android problem. 15:42 < LissajousPattern> so yeah this is easy 15:42 < Dan39> NGC3982: right, that sucks 15:43 < Dan39> NGC3982: is using an old android phone for the challenge? 15:43 < Dan39> instead of buying a $5 webcam 15:43 < LissajousPattern> ha 15:43 < NGC3982> it would be cool to use a cell phone since i already have them, but yes ..if they are so cheap why not. 15:43 < NGC3982> hah, i haven't really thought about it. 15:43 < Dan39> right 15:43 < LissajousPattern> dude why have you not won yet? 15:44 < Dan39> the phone project seems fun 15:44 < paulcarroty> Dan39, which software help you to use your phone as webcam? 15:44 < Dan39> but like you said, the picture a minute thing heating your phone up may be an issue that you cannot easily fi 15:44 < Dan39> x 15:44 < Dan39> might have to root phone and update ROM 15:44 < NGC3982> Dan39: it sure does, but in this case the linux part seems more fun that the hardware part. 15:45 < Dan39> whats the phone model? 15:45 < NGC3982> galaxy s7 15:45 < Dan39> if it has an easy root, id try that, might only take 5 minutes 15:45 < Dan39> oooh 15:45 < Dan39> um aren't the galaxy's like unrootable? 15:45 < NGC3982> it might already be rooted. 15:45 < NGC3982> no, don't think so. 15:45 * NGC3982 checks. 15:46 < LissajousPattern> dude I think I may know a cool hack to try for it 15:46 < LissajousPattern> there are apps that let you use your phone as a web server 15:47 < NGC3982> oh? 15:47 < LissajousPattern> which if you have access would also allow you to gain the path to your camera 15:47 < paulcarroty> Linux can't directly connect to phone camera 15:47 < LissajousPattern> ha says who? 15:48 < Dan39> paulcarroty, that's who. 15:48 < LissajousPattern> I guess 15:48 < Dan39> if he is right, i don't know though. 15:48 < NGC3982> paulcarroty: my initial thought was to simply snap a picture every minute, write it directly to the samba share and overwrite the current file. 15:48 < Dan39> im pretty sure there are interfaces to the camera some how :P 15:49 < paulcarroty> Android has their own camera API 15:49 < NGC3982> maybe i should stick to that solution right now, and try to fix the actual camera issue instead. 15:49 < NGC3982> maybe ill learn something. 15:49 < LissajousPattern> learn all the things 15:49 < paulcarroty> you can share photos with samba/nfs, but nothing like realtime video 15:50 < Dan39> NGC3982: galaxy s7 isn't that old, right? you should be able to do this without that issue i feel, but im no expert 15:50 < Dan39> paulcarroty: he doesn't want realtime, just 1 pic a minute 15:51 < NGC3982> Dan39: i think ill continue with reseting the phone, installing tasker, set the picture-snap profile and simply try again. 15:51 < LissajousPattern> well use ipwebcam or some shit 15:51 < paulcarroty> then it's better to use any android automation tool 15:51 < NGC3982> -> tasker. <3 15:51 < LissajousPattern> and view it through the browser and then embed the url in the site 15:51 * NGC3982 tries 15:51 < NGC3982> thanks guys. 15:52 < NGC3982> im not going to try what i did last time and charge the phone with the pi. 15:53 < LissajousPattern> so turns out you can embed any ip cam into html with like a single line of code. 15:54 < takeme> how can i delete attr_val 15:54 < takeme> fw_setenv test 1234 15:54 < MrElendig> takeme: context missing 15:54 < takeme> how can i delte test? 15:54 < takeme> MrElendig: ? 15:54 < NGC3982> LissajousPattern: tell me more. 15:55 < LissajousPattern> I just did 15:55 < MrElendig> takeme: http://www.catb.org/esr/faqs/smart-questions.html 15:55 < MrElendig> takeme: you have not provided any context so we have no idea what you are talking about 15:55 < takeme> can i know what is uboot? 15:55 < sandman13> Dagmar: so it can be done? 15:56 < MrElendig> takeme: https://en.wikipedia.org/wiki/Das_U-Boot 15:56 < LissajousPattern> even h.264 based 15:56 < LissajousPattern> dude look at a google 15:56 < LissajousPattern> wtf 15:56 < NGC3982> yes, i think i found something cool 15:56 < NGC3982> thanks 15:57 < LissajousPattern> yes computers are cool... are computers still cool? 15:57 < NGC3982> in my job, someone recently said that programmers are the new flight attendants 15:57 < NGC3982> so perhaps not 15:58 < LissajousPattern> haha the new flight attendants. I have a good friend who is a programmer. 15:58 < takeme> ERROR: Invalid configuration of endpoints. <—— do you know what it mean? MrElendig ? 15:58 < MrElendig> not without any context no 16:02 < Li> is there any way to correct/complete wrong downloaded iso file? 16:04 < MrElendig> Li: binary patch 16:04 < mawk> if you have the original file maybe 16:04 < MrElendig> will be just as fast to just get a new copy 16:04 < mawk> if it's just completing it yeah, you download the remaining bytes 16:04 < mawk> but you need to know where the good part ends 16:05 < MrElendig> you could read eg 1meg, checksum it, compare against upstream, repeat until done 16:05 < MrElendig> but really, waste of time 16:05 < MrElendig> unless your distro provides torrents, if so, just feed your broken one to the torrent client and it will take care of it 16:06 < NGC3982> that ip webcam thing was of course the best option since it is already working. 16:06 < NGC3982> .. 16:07 < NGC3982> -.. 16:10 < Li> so the answer is No unless torrent was used to download the file 16:11 < mawk> no 16:11 < mawk> you can use the existing broken file with a matching torrent file 16:14 < xryuu> Hello all. I have a HDD that I use with WD My Cloud. Is there any way to read the dmesg file from that HDD? 16:15 < takeme> what is DFU ? 16:16 < MrElendig> xryuu: take it out and plug it into something else 16:16 < MrElendig> (wd my cloud is pretty horrid btw) 16:16 < MrElendig> make sure it does not have direct access to the internet 16:17 < xryuu> MrElendig: Well, I installed debian into it. Upgraded to 9, no internet. At the moment I have that HDD connected to my pc so I can mount the raid and see everything but how do I use the dmesg command to read that specific log on the sdb? 16:18 < MrElendig> if it is running debian then you can do whatever you like 16:19 < xryuu> You didn't understand me, let me explain a bit better 16:19 < MrElendig> xryuu: see the journalctl man page, you can tell it to read a file instead of the current log 16:19 < MrElendig> xryuu: your original question did not mention that you had installed debian and moved the disk :p 16:21 < xryuu> Yeah let me ask again properly! I have a HDD with debian installed in it. I can't boot it directly from my pc. After upgrading to Debian 9 I don't have internet since eth0 doesn't work anymore. I'm trying to find a log file in that specific HDD(/sdb) to see what it has converted it to. Is there any way to do that? 16:22 < TheWild> hello 16:23 < twainwek> hi 16:23 < TheWild> on linux, is this possible to allocate block of memory (malloc) in a way that once allocated, it keeps its *physical* location? 16:24 < DLange> xryuu: what filesystem is on sdb1 and sdb2? 16:24 < MrElendig> TheWild: depending on the hardware "maybe" 16:24 < DLange> (BTW making a raid out of two partitions on one disk doesn't make any sense) 16:25 < xryuu> DLange: linux_r. Yeah I know but that's how a specific hardware operates.. 16:26 < DLange> xryuu: linux_r is not a filesystem, EXT4 is, or BTRFS or FAT32 or ... 16:27 < xryuu> DLange: then there is none but I can mount it and see the contents 16:27 < DLange> well, then check /var/log for what you are looking for 16:27 < DLange> dmesg itself is memory only but what it stores, it stores in /var/log 16:29 < hexnewbie> xryuu: Is your problem that the computer doesn't boot, or that it lacks eth0 when you boot it? If only the latter, see ‘ip link’ for a list of interfaces 16:30 < raindev> Anyone here has been reading dotslashlinux and has any idea why it has disappeared? 16:30 < raindev> https://web.archive.org/web/20180118062855/https://www.dotslashlinux.com/ 16:30 < hexnewbie> Not booting and missing eth0 are unrelated problems, although they can both happen if you're on UEFI (esp. upgrading from BIOS), and you're upgrading from a system before predictable interface names (eth0) 16:30 < xryuu> hexnewbie: can I pm you? 16:30 < xryuu> so I don't spam with the question/explaination 16:30 < hexnewbie> xryuu: No, why would you do that? 16:31 < Pentode> raindev, sadly sites like this disappear regularly. it costs money to host stuffs. 16:32 < Pentode> it's especially hard when you are ad-free and rely on donations. :/ 16:32 < DLange> raindev: https://www.reddit.com/r/DOTSLASHLINUX/comments/8pvjeg/website_down/ 16:33 < raindev> Pentode, there are ways to host stuff for free (e.g. GitHub pages or something similar). But there are other costs of course. 16:33 < raindev> The Patreon page of the project is gone too. 16:34 < Pentode> true. i guess he just wanted to do his own thing 16:34 < raindev> DLange, thanks for the link, didn't know there was a subreddit. 16:34 < xryuu> hexnewbie: Ok so here it goes again. I've got a WD My Cloud 4TB, I installed Debian on it, upgraded to 9 and lost my ethernet connection. Because of the partitioning etc that My Cloud uses I think that I can't boot it from my pc so I only mount it as a secondary hdd(/sdb). Now the question is how can I find the eth0, ens33, enp3s0 etc used without a monitor and only from that specific HDD 16:39 < hexnewbie> xryuu: As DLange suggested, look in ./var/log/kern.log where . is the mountpoint where you mounted it 16:39 < hexnewbie> xryuu: If a network card was detected it will be visible there. 16:40 < hexnewbie> Unrelated, but does the word cloud even mean anything nowadays? How is a computer in the local network a cloud? 16:40 < DLange> $20 more expensive than not labeled "cloud" 16:41 < DLange> just like "cloud hosting" vs "hosting" :) 16:41 < xryuu> Will look into it thank you. Well, I don't even care for the cloud. I got it for free and I'm trying to experiment with it so 16:41 < ananke> hexnewbie: it's related to the service WD has, not the local disk 16:43 < WhiteDevil> hello 16:43 < WhiteDevil> i installed by RHEL again 16:43 < WhiteDevil> and i remember that i would use irc on root 16:43 < WhiteDevil> and someone told me perhaps thats why eitheer my computer gets infected with a trojan 16:44 < WhiteDevil> and it keeps shutting down my system 16:44 < WhiteDevil> or it could be a a heat issue 16:44 < Dan39> WhiteDevil: i doubt using irc on root caused you to get a trojan unless you accepted and executed a DCC file download 16:45 < Dan39> or you are using some old un-updated irc client with an exploit 16:45 < hexnewbie> WhiteDevil: Using computer as root is a bad practice, but heat or memory is much more likely than someone serving you malware. 16:45 < WhiteDevil> i did a yum install package name 16:45 < Dan39> but of course it is always bad practice to run stuff as root :p 16:46 < WhiteDevil> I see because if it a hardware fault 16:46 < hexnewbie> Well, except for admin stuff like ‘yum install’. That you do run as root. 16:46 < WhiteDevil> what tools would i use to check for it 16:46 < hexnewbie> WhiteDevil: memtest86 16:46 < WhiteDevil> k 16:46 < Dan39> hexnewbie: well even that would be preferred to run as a user then only elevate commands when needed, no? :P 16:46 < WhiteDevil> but this problem dosnt occure when i run debian 16:47 < Dan39> that seems to be the trend 16:47 < hexnewbie> WhiteDevil: There might be a package for RHEL that will add it to your grub menu. There are ISOs out there. Overheating can be checked with sensors (package name probably lm-sensors) 16:47 < Dan39> WhiteDevil: RHEL might have an older driver cuainsg issues 16:47 < Dan39> causing 16:47 < Dan39> iirc debian 9 is a few years newer than RHEL 7 16:47 < hexnewbie> WhiteDevil: Kernel errors preceding the crash may be visible in /var/log/kern.log 16:47 < WhiteDevil> driver seems like the likely cause 16:48 < WhiteDevil> okay ill look into that log file 16:48 < xryuu> There is no kern file in /var/log 16:49 < Dan39> WhiteDevil: if there is nothing in logs you might have to do something like sit on a console with increased logging output and wait for it to crash, and see if that gives any errors or you can see what it was doing 16:50 < hexnewbie> xryuu: syslog or messages may carry the message, but if you removed rsyslog and you're logging in the journal, you may need to run journalctl --root=/path/where/you/mounted/the/debian 16:51 < xryuu> hexnewbie: no journal founds. journalctl --root=/mnt/hdd (mnt/hdd/ is where I mounted) 16:52 < hexnewbie> xryuu: Is there anything in /mnt/hdd/var/log/journal/ 16:53 < xryuu> hexnewbie: not even the folder journal.. 16:53 < Dan39> xryuu: ffs just pastebin `tree /var/log/` 16:53 < Dan39> well, /mnt/hdd/var/log/ 16:55 < xryuu> Dan39: https://pastebin.com/ygmSuaEk 16:58 < hexnewbie> xryuu: The interface name may be in /mnt/hdd/var/log/dmesg then 16:58 < xryuu> hexnewbie: it's eth0 but it's not getting an ip! 16:59 < Dan39> xryuu: well there aren't many files, just look in the them -_- 16:59 < hexnewbie> xryuu: Are you sure the var/log/dmesg file is recent enough? Run stat on it. It may be that the device didn't boot at all, and that's an old file from the last succesfful boot. 17:00 < Dan39> and they have timestamps usually... -_- 17:01 < xryuu> hexnewbie: Yeah it's an old one, you're right.. 17:01 < Dan39> bummer 17:07 < Pusteblume> i am trying to find a tutorial for "ufw" similar to the iptables one by frozentux. especially the forwarding documentation is interesting to me. could somebody point me there? 17:10 < DLange> Pusteblume: https://gist.github.com/kimus/9315140 17:11 < Pusteblume> DLange, thank you! 17:11 < DLange> actually even https://help.ubuntu.com/lts/serverguide/firewall.html.en is quite useful 17:18 < Pentode> lol, setting up this old machine is a hoot. windows95 can't even update the mouse cursor while it's accessing the floppy drive. now thats multitasking. 17:19 < Pentode> all of this just to play a single game o_O 17:19 < doors_1> what kind of machine is still running that version of Windows? 17:19 < doors_1> Can't you just use an emulator? 17:19 < Pentode> yeah but what fun is that? 17:19 < Pentode> plus i need to use my spaceorb 360, which is an old serial game controller 17:20 < Pentode> we're going to have a little retro lan party for old time sake 17:20 < Li> thanks MrElendig & mawk 17:20 < Pentode> apparently my machines one of the fastest. a whopping amd k6-2 at 380mhz ;) 17:22 < Pentode> driving me nuts though im on the third install of windows. what i've learned: just don't even try getting usb working in 95 17:22 < pankaj> I have a NTFS formatted external SSD. It has only one partition. What is the best way to protect it from other users in case they have access to it physically. 17:23 < Pentode> use luks and encrypt it? 17:24 < hexnewbie> pankaj: Do you want to protect it while you can still access it from both Windows and GNU/Linux? Do you want to use it from multiple computers? The more yeses you get, the more difficult it would be. 17:27 < hexnewbie> Pentode: Waiting out for Windows 98 to come out would still be easier than trying Linux 2.3 or following linux-usb.sourceforge.net, I think. Unless the time travel rules allow you to pick a modern operating system with you. 17:27 < pankaj> hexnewbie: My brother has windows and I am a linux user. The drive is NTFS formatted. I want that the windows one should not be able to access particular folders that is important to my brother. 17:28 < hexnewbie> pankaj: That somehow omitted the answer to the questions. Would you use the disk from both those computers? 17:28 < Pentode> hexnewbie, lol ;p 17:29 < Pentode> if memory serves me, which it probably doesnt, linux used to run pretty good on my p120 \ 17:29 < Pentode> whopping 64mb of ram at the time, too. which was _expencive_ 17:29 < dgurney> modern linux should *run* alright 17:29 < hexnewbie> pankaj: Do you want one folder encrypted to be accessible from only one computer? 17:30 < dgurney> but good luck doing any work on it ;P 17:30 < Pentode> X11R6, fvwm and a cool 120 million cycles per second. what more do you need? 17:30 < hexnewbie> pankaj: Because Windows *can* do that, but your Linux computer won't be able to read it. 17:30 < Pentode> dgurney, yeah i had an old slackware distro running on this thing. it was definitely usable and really pretty light after trimming the kernel down a bit 17:30 < Pentode> this machine has 160mb of ram though 17:31 < Pentode> it would have all fit in 16, though with much swapping. ;) 17:31 < pankaj> hexnewbie: Do, I have to encrypt those folders 17:31 < hexnewbie> pankaj: If you want a directory only visible from your GNU/Linux machine, then you shrink the NTFS and create a LUKS container. 17:31 < Pentode> it booted in with an xserver at about 38-40mb 17:31 < Pentode> not too shabby 17:32 < pankaj> hexnewbie: I just want that he should not be able to access the files he want (as he has different folder for his own stuff in the same partition) without my permission. 17:32 < hexnewbie> pankaj: Yes. Of course. How else would you protect from physical access? Although LUKS (for GNU/Linux) would be unnecessary, as I guess ecryptfs and encfs are also options (minding the warnings about encfs security issues)./ 17:33 < pankaj> hexnewbie: OK. Shrinking. But I have to search about this LUKS 17:33 < hexnewbie> pankaj: OK, so that story got changed again. You want to protect your files (GNU/Linux) from your brother (Windows)? 17:33 < pankaj> hexnewbie: Yes, He is so terrible. I own that SSD but he fights if I do not give it to him. 17:34 < pankaj> hexnewbie: So, even if he had the SSD he would not be able to access his files without my permission and that is what I want. 17:34 < hexnewbie> pankaj: I'd just buy a second SSD and LUKS-encrypt it. But baring that, I do have an encfs encrypted directory on UDF. I guess you can have one on NTFS. Your brother could theoretically still attack it (encfs has issues if someone has write access to the data), but ecryptfs and LUKS won't have theissue 17:38 < pankaj> hexnewbie: OK. So, I noted some important terms -> "LUKS", "encfs", "encryptfs". I will certainly research on them and if found appropriate I will experiment with them with my pendrive. 17:39 < acresearch> does gopher servers run on linux? 17:42 < sagax> hi all! 17:42 < sagax> how to remove remote ascii from bitbucket? 17:43 < pnbeast> You have to put images into the bitbucket until it overflows, washing out the ASCII. It's the only way. 17:44 < hexnewbie> Could simply pour in UTF-8 cuneiform if images are too expensive. 17:45 < mawk> takeme: some USB protocol for uploading new firmwares to embedded devices 17:48 < mawk> which I'm presently implementing into a bootloader for some ARM card 17:48 < kerframil> acresearch: potentially. with gofish, for example. 17:53 < pepm1nper> I don't know where else to ask this even though it is a shell question: what does `. command` mean? @_@ 17:54 < revel> afaik, `.` is the same as `source` 17:55 < taaperotassu> I am trying to use hugo to publish in my /var/www/htm I already used sudo chmod -R 765 /var/www/ yet it doesnt work? Same error comes │Error: Error copying static files: mkdir /var/www: read-only file system 17:55 < pepm1nper> that's so strange. revel have you ever used ranger? 17:55 < revel> No. 17:55 < kurahaupo> pepm1nper: "source" is a new synonym for "." 17:56 < kurahaupo> taaperotassu: that's not a permission error. 17:57 < pepm1nper> in order to get ranger to change your directory after you quit with q or :q, one should use . ranger or source ranger then it is possible to navigate to the desired directory, quit, and you will be in the directory. Otherwise, your directory will remain unchanged. I am not sure what craziness is going on here. Many strange. 17:57 < pepm1nper> I looked at src: This embedded bash script can be executed by sourcing this file. ... @_@ 17:58 < offline> hi 17:58 < taaperotassu> kurahaupo: But I thought that means it doesnt have write permissions? 17:59 < kurahaupo> pepm1nper: if a script is sourced its constituent commands are executed by the current shell rather than a separate shell as would be the case if you ran it as a command 17:59 < kurahaupo> taaperotassu: no, it means the filesystem as a whole is not writable 17:59 < pepm1nper> right I undrestand; still strange 18:00 < pepm1nper> it's wrapped in 4 double quotes and a colon on top and 3 double quotes on bottom 18:00 < taaperotassu> kurahaupo: You mean if I put write permissions to those folders it still doesnt help? 18:00 < taaperotassu> maybe only way is to use other locations? 18:01 < kurahaupo> taaperotassu: you won't be able to change the permissions on the directories or anything else 18:01 < taaperotassu> kurahaupo: why so? I used sudo chmod and sudo chown and they worked? 18:01 < kurahaupo> taaperotassu: better would be to find out why it's mounted read-only 18:02 < kurahaupo> taaperotassu: ok, check you are writing where you think you are 18:02 < taaperotassu> ok 18:04 < kurahaupo> taaperotassu: make a paste of a transcript from your terminal that includes the chmod & chown commands, and whatever copy command, and pwd just to show your current directory for good measure 18:06 < kurahaupo> taaperotassu: does hugo actual write to the same host? 18:09 < taaperotassu> │-rwxrw-r-x is okay and the host is same 18:10 < taaperotassu> Tho not sure if /var is the problem as its │drwxr-xr-x 14 root root 18:10 < taaperotassu> if it needs the whole path to be read write i 18:12 < takeme> mawk: thanks can i know my question? 18:12 < takeme> it is too old question so i can’t remember it 18:14 < maryo> I am trying to start a service using nohup, but as soon as I start the service using nohup it is getting stopped. Just looking for some pointers to narrow down the issue. Any help would be appreciated. -> https://dpaste.de/BLq7 18:15 < pnbeast> maryo, AFAIK, nohup is typically used for programs you run interactively, from a shell. It's not intended for services, which should drop their controlling terminal by themselves. 18:17 < xinming> Anyone ehre familiar with multipath and raid1? 18:18 < pnbeast> xinming, don't ask to ask: ask your real question. 18:23 < xinming> what are the differences between RAID1 and multipath? If I have 2 iscsi disks attached, I create a raid1 over these 2 iscsi disks, It seems they *does* the same thing. Or, anything I missed? 18:23 < dunnousernamefn> Hey, can I use an ARM device's UART as a serial port and open up a terminal on that? 18:23 < xinming> I googled about this, and I can't know what the advantage over software raid1 vs multipath 18:27 < dunnousernamefn> So it seems if they're actually /dev/tty's it should work 18:40 < takeme> export IIOD_REMOTE=usb:3.6.5 <— what it mean? 18:41 < pnbeast> takeme, that's how some shells (bash, for example) set an environmental variable. "IIOD_REMOTE" is the value, etc. 18:41 < rindolf> takeme: hi 18:41 < rindolf> takeme: what pnbeast said 18:41 < pnbeast> takeme, er, it's the *variable*. 18:42 < pnbeast> NOT what pnbeast said! 18:42 < takeme> i set it every times when i turn on computer? 18:42 < pnbeast> Why? 18:42 < takeme> it is too tired work 18:43 < takeme> to set it every times 18:43 < takeme> there is no way to set it permently? 18:44 < takeme> permenant 18:44 < takeme> ly 18:44 < pnbeast> Of course there is. Read the docs for your shell. 18:45 < pnbeast> Is it for you, or every user on the computer? If it's for you, it probably gets set in your .bashrc file or your .bash_profile file, if it's for everyone, then maybe in /etc some place. 18:46 < takeme> .bashrc or .bash_profile 18:46 < takeme> pnbeast: thanks good guy 18:46 < takeme> 8) 18:47 < pnbeast> You must be talking about rindolf or someone else. 18:47 < takeme> rindolf: ? 18:47 < takeme> i idon’t know rindolf who is he or she? 18:47 < rindolf> takeme: i am me 18:48 < takeme> ah 18:48 < rindolf> takeme: see http://www.shlomifish.org/me/rindolf/ 18:52 < takeme> .bashrc have many contents, but .bash_profile don’t exist. so i can make .bash_profile, and i save it? pnbeast ? 18:53 < mawk> takeme: you asked about DFU 18:53 < mawk> so I answered 18:53 < takeme> ah thanks mawk :) 18:53 < takeme> but answer is too fast 18:53 < mawk> well dfu is also a file format that encodes information on where to put the firmware in memory 18:54 < mawk> also using the dfu protocole you can read the device memory, erase it, etc 18:54 < takeme> ah 18:54 < takeme> good 18:54 < takeme> where can i learn about dfu? 18:54 < mawk> you can try ST documents 18:55 < takeme> ST? 18:55 < takeme> no idea can you give me link? 18:55 < mawk> https://www.st.com/content/ccc/resource/technical/document/application_note/6a/17/92/02/58/98/45/0c/CD00264379.pdf/files/CD00264379.pdf/jcr:content/translations/en.CD00264379.pdf 18:56 < takeme> only 23 page? 18:56 < takeme> can i learn it with only 23 page? 18:57 < sagax> hi all 18:57 < mawk> you'll see 18:57 < sagax> how can i add fossil scm to redmine? 18:57 < sagax> like as git/hg scm 18:57 < lnnb> hmm i recognize one of those words, source control management ? 18:58 < takeme> pnbeast: when i put export IIOD_REMOTE=usb:3.6.5 into .bash_profile, it don’t works. what’s wrong? 18:58 < sagax> yes, fossil it's scm 18:58 < NGC3982> Dan39: http://hydroponics.henjoh.se/ 18:59 < takeme> pnbeast: you lied to me? 19:00 < mawk> lol 19:04 < Dan39> NGC3982: live now!? 19:04 < Dan39> NGC3982: so how are you doing it? 19:04 < Dan39> looks great 19:04 < takeme> mawk: there is a more link related? 19:05 < NGC3982> Dan39: ip webcam, and then embed it in html. 19:05 < Dan39> hahaha 19:05 < Dan39> not a phone then? 19:05 < NGC3982> Dan39: yes, the phone. it has heat issues, but in the quality you see the phone is not more than luke warm 19:05 < Dan39> wow, it works with just and i do not need better quality than that. also, live video as a bonus. 19:05 < NGC3982> yes 19:05 < Dan39> NGC3982: what software is on phone? 19:05 < NGC3982> probably -highly- unsecure. 19:06 < NGC3982> the app 'ip webcam' 19:06 < Dan39> ha 19:06 < mawk> just google it takeme 19:06 < takeme> ah ok 19:06 < takeme> thanks 19:06 < Dan39> it's like a streaming jpeg when i open the source directly, interesting 19:06 < mawk> I don't have more info that you have 19:06 < NGC3982> Dan39: so, thanks for the advice. this will work out nicely. 19:06 < mawk> I just read the ST document and that's all 19:06 < takeme> ok thanks mawk 19:07 < takeme> mawk: can i ask question? 19:07 < Dan39> i didn't really give much advice :P but sometimes ya know... just rubber ducky :P 19:07 < mawk> yes if you want 19:07 < takeme> mawk: i add export IIOD_REMOTE=usb:3.6.5 to .bash_profile. but it don’t work. how can i do ? mawk ? 19:09 < mawk> why not .bashrc ? 19:09 < takeme> mawk: i have no time because mom call me 19:09 < takeme> ah 19:09 < takeme> there is many contents already 19:10 < takeme> so i made bash_profile newly 19:11 < takeme> mawk: i can’t wait anymore give me answer 19:13 < rascul> if it's in ~/.bash_profile, and bash is your default shell, logging out of all active logins and logging back in should have the desired effect, as bash will read this file upon login 19:13 < pikaro> I use a VPN with an iptables killswitch (disallow Internet traffic except on tun*) for privacy reasons but would like to circumvent that with a second browser profile sometimes. I don't have an external server in my country available for proxying, so I'd like to implement that using the same machine. is that possible? and if so, what would be the general concept? 19:14 < mawk> I don't think bash_profile is sourced everytime takeme 19:14 < takeme> rascul: then i must use .bashrc? 19:14 < mawk> I'd put the exports at the top of bashrc, before the [[ $- == *i* ]] || exit 0 19:14 < rascul> .bashrc is read and executed when bash is started as a non login shell 19:14 < mysall> it is not by default. Some flavors of linux add a source command to bashrc which sources bash_profile on start up 19:15 < rascul> mysall that's because the distros do weird things 19:15 < rascul> it is by default, as per bash docs 19:15 < mysall> rascul: that's what I was saying 19:15 < rascul> i guess i don't understand what you're saying 19:16 < rascul> or something 19:16 < rascul> https://www.gnu.org/software/bash/manual/bashref.html#Bash-Startup-Files 19:17 < mysall> I'm saying that by default bashrc is what gets run by default when using bash. But some linux distros add a line to bashrc which sources bash_profile as well. 19:17 < rascul> it depends on how bash is called 19:17 < mysall> yes 19:18 < rascul> do you have an example of a distro which does this? 19:22 < rascul> debian will source ~/.bashrc from ~/.profile by default, but not the other way around 19:24 < pnbeast> ZOMG, what? I lied?! 19:24 * pnbeast weeps out of shame and guilt. 19:26 < rascul> pnbeast i think that's a lie 19:27 < MrElendig> man bash, see the INVOCATION section 19:28 < pnbeast> I prefer to think of it as performance art. 19:28 < pnbeast> Don't tell him to read docs, MrElendig, he'll just call you a liar! 19:29 < notmike> MrElendig: unban me fam 19:29 < revel> MrElendig: ban me pls 19:30 < notmike> MrElendig got in his feelings and banned me for personal reasons from #archlinux. 19:30 < notmike> It's like the least arch thing a person could do. 19:31 < revel> lol what 19:32 < roxlu> hi, I just learned that the source-address of a socket can change after I bound it. I'm wondering if there is a way to prevent this? (but still use INADDR_ANY). 19:48 < takeme> i want to learn linux including shell, but i have no method how can i do it? can you give me link for that? 19:49 < Pentode> learn by doing 19:49 < MrElendig> https://www.edx.org/course/introduction-linux-linuxfoundationx-lfs101x-1 19:49 < MrElendig> takeme: ^ 19:49 < MrChinasky> takeme: You should start with being comfortable with cmds such as 'ls' 'cd' 'mkdir' 'cp' 'mv 'grep' 19:50 < takeme> MrChinasky: i know it already 19:51 < sauvin> You should also hang out *here*, among other things. 19:52 < Pentode> a good way to familiarize is to learn how to write scripts for whatever shell you use. maybe following some tutorials and learn a language (like C) 19:53 < kamura> takeme: when I was getting into linux I used i3 exclusivly for 6+months 19:53 < kamura> you learn prety quick when you've got no other options but to do it 19:57 < granttrec> I cannot seem to find the bash functions in this repository https://github.com/dylanaraps/pure-bash-bible yet there are referenced in the test script, does anyone know how the functions are stored 19:58 < sagax> i look for scm online services for svn/cvs/bzr/darcs like github/bitbucket/gitlab 19:58 < Pentode> granttrec, what functions are you talking about? 19:58 < sagax> if they exist 19:59 < Pentode> granttrec, i see the tutorial giving multiple example functions that do various things. 19:59 < MrElendig> good luck finding a darcs one 19:59 < MrElendig> the rest is easy 19:59 < granttrec> for example in test.sh the function "trim_string" is defined but how is being called? 19:59 < MrElendig> by calling it? 20:00 < MrElendig> trim_string arg1 .. 20:00 < granttrec> MrElendig: from where tho thats what I want to know 20:00 < Pentode> they are being defined then called from an argument passed as test.sh is executed it appears. 20:00 < MrElendig> you do it whenever you watn to use it 20:00 < Pentode> as MrElendig said 20:00 < MrElendig> want* 20:01 < MrElendig> it is not magically called anywhere 20:03 < granttrec> MrElendig: ok, that means those functions are passed as paramters? so they are not in the repo? 20:09 < Pentode> it's reading and parsing code chunks from the README.md itself 20:09 < Pentode> interesting. 20:11 < Pentode> then it appears to be testing the various results from the code blocks themselves. i guess it's some kind of crazy exercise? 20:11 * Pentode shrugs 20:14 < granttrec> Pentode: is that in the while loop bit? I was starting to suspect as much 20:14 < Pentode> yeah 20:15 < granttrec> The examples are really usefull and I want to use then to test another shell, so I guess Ill have to parse the readme too 20:15 < Pentode> you can just call the functions any way you want really 20:16 < Pentode> though i guess some portions are dependent on the information in readme.md 20:16 < Pentode> the ones _in_ readme.md however you could just put them in your script, then call them somewhere inside it. 20:17 < Pentode> or pass an argument and call them that way 20:17 < Pentode> using $1 $2, etc. 20:18 < Pentode> you could even put a bunch of them in your bashrc if you wanted, then call them right from your shell. 20:39 < Reventlov> Hi. 20:39 < subash> hello 20:39 < Reventlov> I'm trying to understand the SGI flag in the iwlwifi driver: https://github.com/torvalds/linux/blob/master/drivers/net/wireless/intel/iwlwifi/mvm/rs.c#L3299 20:40 < Reventlov> According to https://github.com/torvalds/linux/blob/master/drivers/net/wireless/intel/iwlwifi/fw/api/rs.h#L458 it should be the bit 13 20:41 < Reventlov> yet, converting the 0x400D019, 0x400D018, 0x400D017, 0x4005007, 0x4009006, and 0x4005005 20:42 < Reventlov> to arrays of bit does not give the expected result (true, true, true, false, false, false) 20:47 < Dan39> Reventlov: try ##kernel 20:52 < Reventlov> Ok, thanks for the hint 20:59 < fr0tzed> guys, how do i remoce the hybernate file of win10 20:59 < fr0tzed> ik trying to use chntpw 20:59 < fr0tzed> With windows 7 i can do mount -t ntfs-3g -o remove-hiberfile 21:01 < lnnb> .votekick 21:02 < micrex22> fr0tzed you need to disable the hybrid boot option so that hyberfil.sys isn't perpetually in use 24/7 21:03 < micrex22> fr0tzed because windows is so poorly coded along with NTFS, microsoft opted on just having the OS perpetually operate in a hybrid-hibernation mode. 21:03 < Mag-en00b> (context: rowhammer attack): am i correct in assuming that if a virtual machine's backing memory (physical memory) is contiguous and unique to that virtual machine (not shared), then rowhammer attack cannot cause havoc in the host machine? 21:04 < fr0tzed> micrex22, how if i do not have the pswd 21:05 < micrex22> fr0tzed you need to use something like ntpasswd or another SAM database reset program to boot the machine in and blank whatever local account there is 21:05 < micrex22> this does not work on microsoft cloud accounts 21:05 < Mag-en00b> next question (assuming above is true): how on earth do i accomplish this with qemu (with kvm), as it seems that there are no mechanisms besides reserving hugepages on kernel startup and using that memory 21:06 < Mag-en00b> .. and even that seems to be unsure (there is too much and too much outdated information regarding linux virtual memory handling) 21:08 < micrex22> Mag-en00b I'd be interested to know as well 21:10 < fr0tzed> micrex22, found a easier way. boot till logon page, then hold shift and shutdown. This way it will apparently clear the hibernate file 21:11 < Sapphirus> Mag-en00b, to answer your first question, yes, there could be the potential for vm escape. 21:11 < Sapphirus> Rowhammer is a hardware bug, like Meltdown, so there's no reason it wouldn't affect a virtual environment. 21:12 < Sapphirus> More easily mitigated by having memory which isn't vulnerable. 21:15 < jim> Mag-en00b, you said you want to accomplish something, what is it? 21:16 < fr0tzed> micrex22, it worked, managed to clear the password. thanks though 21:16 < jim> I looked at like 50 scrollups, only found you in first 3 or so 21:17 < lessthan0> https://github.com/xoreaxeaxeax/ 21:17 < lessthan0> this guy has some awesome stuff for researching VM escape and hardware architechure bugs 21:18 < lessthan0> he showed a proof of concept of code that runs differently in qemu because qemu is based on incorrect documentation provided by intel 21:19 < lessthan0> in theory they can't really patch qemu if they don't know what the CPU will do in every case 21:20 < lessthan0> the cpu is like a black box and qemu can't test something that breaks if it never breaks. things don't break in qemu because the app devs are using the same incorrect docs that the qemu devs get from intel 21:22 < lessthan0> qemu passes the malicious code to the cpu and there is an unexpect breakout 21:23 < lessthan0> because intel refuses to document the x86 instruction set %100 unless you are an intel partner 21:23 < pnbeast> No one expects the Spa^H^H^H QEMU breakout! 21:24 < lessthan0> it would be better if intel developed a container os 21:24 < lessthan0> then they don't have to worry about NDA 21:24 < mawk> I wouldn't use a container OS that isn't open source 21:24 < lessthan0> and they can be lazy while at the same time being secure 21:25 < lessthan0> qemu is open source and it breaks because of a proprietary CPU with secret intel partner NDA crap 21:25 < lessthan0> so what have you gained from open source? 21:26 < lessthan0> if everyone has to lock down qemu by making some proprietary version from source then there is a lot of duplication of effort 21:27 < solidfox> vaults in kde are lacking a lot of features 21:27 < solidfox> such as deleting the vault, or exporting it 21:31 < feuerbach> Is there a way to make 'perf' (or a similar command) output the number of executed AVX instructions? 21:32 < lessthan0> are you benchmarking or debugging? 21:34 < feuerbach> somewhere in between I guess :) I'm following this tutorial on matrix multiplication https://gist.github.com/nadavrot/5b35d44e8ba3dd718e595e40184d03f0 21:35 < feuerbach> "One way to verify this claim is to use performance counters. The two relevant counters are: the number of executed AVX instructions, and the total number of cycle." 21:35 < feuerbach> The author uses XCode for this, and I'm trying to do the same on Linux 21:38 < mawk> yes 21:38 < mawk> then use the performance counters ? 21:41 < feuerbach> right, how do I do that? 21:42 < feuerbach> I tried to use perf, but I don't see anything AVX-related in `perf list` 21:42 < mawk> you'll maybe have to do it manually then 21:43 < mawk> let me search that 21:44 < mawk> or you can maybe add performance events to perf, with the correct number from the intel manual 21:44 < mawk> what processor do you have feuerbach ? 21:44 < mawk> exactly 21:45 < feuerbach> Intel(R) Core(TM) i7-6500U CPU @ 2.50GHz 21:47 < mawk> well there are a lot of event types related to AVX 21:47 < mawk> for instance there is "Counts the number of packed SSE, AVX, AVX2, and AVX-512 micro-ops (both floating point and integer) except for loads (memory-to-register mov-type micro-ops), packed byte and word multiplies." 21:48 < feuerbach> d'oh. I was searching case-sensitively. Thanks! 21:48 < mawk> my quote is from the intel manual feuerbach , not perf 21:48 < mawk> I didn't find much related to avx from perf list myself 21:49 < feuerbach> well I see it in perf-list now 21:49 < mawk> ah 21:49 < mawk> you have a more recent perf than I have then maybe 21:49 < feuerbach> e.g. fp_arith_inst_retired.128b_packed_double 21:49 < mawk> yeah 21:49 < mawk> indeed 21:49 < mawk> I'm reading stuff such as "FP_ARITH_INST_RETIRED.512B_PACKED_DOUBLE" in the intel manual 21:50 < mawk> but only 128 bits instructions in perf list 21:50 < mawk> not 256 or 512 21:50 < mawk> if you want them you may have to add them manually using the codes 21:51 < feuerbach> mine has 256 but not 512 21:51 < mawk> FP_ARITH_INST_RETIRED.256B_PACKED_DOUBLE is event number 0xC7, umask 0x10 21:51 < mawk> and FP_ARITH_INST_RETIRED.256B_PACKED_SINGLE is event number 0xC7, umask 0x20 21:51 < mawk> if you want to add them to perf 21:52 < feuerbach> cool, thanks mawk 21:52 < mawk> and 512 double is 0xC7/0x40, and 512 single is 0xC7/0x80 21:56 < mawk> feuerbach: to get 512b you can use for instance this stat command: perf stat -e 'cpu/event=0xc7,umask=0x40,name=FP_ARITH_INST_RETIRED.512B_PACKED_DOUBLE/' -e 'cpu/event=0xc7,umask=0x80,name=FP_ARITH_INST_RETIRED.512B_PACKED_SINGLE/' 21:58 < feuerbach> I think I may not even have 512-bit instructions. In /proc/cpuinfo flags, I have avx and avx2, but nothing similar to avx-512 (I don't know what the exact flag name should be) 21:58 < mawk> right 21:58 < mawk> so if you already have 128 and 256 bits in perf list, it's fine 21:59 < feuerbach> maybe that's why perf list is not showing 512 events, because it sees that they are not supported? 22:02 < mawk> yeah maybe 22:22 < phinxy> If a driver is compiling a .o object sucessfully but then linking fails with undefined references, what could be the reason? 22:23 < lopid> installation lacks something 22:25 < Loshki> phinxy: it means someone is making a call to a library which you have forgotten to link in. What is the exact text of the error message? 22:27 < maryo> what does this mean https://dpaste.de/yes7 [priv] and notty ? 22:28 < phinxy> Could it mean the headers are there for the compilation to succeed, but the .c with the definition is not found? 22:31 < lopid> it means your sshd is using privilege separation 22:35 < angelo_ts> hi, was looking for a linux ide with syntax check, tested eclipse, works but seems too much as an ide, tested atom + linter, seems not working as expected, is there any other alternative ? 22:38 < hp> Hey can I set a high value(6GB for 2GB ram device) for /dev/zram0 swap could it end with an OOM/kernel panic? Afaik at least looking at free -h ,while ram used by zram0 doesn't count towards used ram it reduces down the used ram. I have a question what happens if the /dev/zram0 device isn't full ,but the ram is. Does kernel crash? Does a loop of trying to put a compressed page happen? Does a program Out Of Memories or just decides not to swap it in? Does a kernel 22:38 < hp> delete something from /dev/zram0 swap or it doesn't (since the device isn't full ,even though it can't put anything due to lack of ram) ? 22:40 < hp> I just used zram with high success and got 25%/around 4.0 compression ratios with lz0 for my use case ,I would like to set a high value to get advantage from times where high compression ratios are possible 22:42 < hp> but I don't want to have an unstable system if the pages that swap in aren't easilly compressable and high zram swap disksize would cause issues 22:44 < chey> im really getting annoyed trying to ssh into the raspi 22:44 < hp> well guess I will find out today 22:45 < chey> connection refused iv been at it all day 22:45 < chey> hello peeps hope your all good 22:47 < storge> hp: you're creating 2G zram devices? what's the 6G you mention? 22:48 < storge> did he leave already right after asking all that 22:56 < granttrec> what happens when you name a function the same name as a command, like echo() { echo "hello world"; } 22:56 < granttrec> how would you call the function echo? 22:56 < camsn0w> ^^ 22:56 < camsn0w> I think they mean naming the method as one of the keywords 22:57 < camsn0w> I'd guess that it would cause an error 22:57 < kerframil> granttrec: it will recurse, with exactly the adverse results that one would expect. what you actually want in that case is: echo() { command echo ...; } 22:58 < camsn0w> fish catches it as a "illegal command name" 22:58 < granttrec> kerframil: very cool 22:58 < granttrec> camsn0w: why not allow recursion, can be useful 22:59 < camsn0w> A lot of the time it's unintended I guess 22:59 < kerframil> infinite recursion generally isn't helpful 22:59 < camsn0w> lol 22:59 < camsn0w> it can be fun though 23:00 < kerframil> well, in so far as understanding the consequences, sure :) 23:00 < camsn0w> One time I used a vm and put "bash" in my fish_conf and fish in my bashrc 23:01 < camsn0w> I wish I recorded it 23:01 < tips-fedora> I'm trying to condense a shell script I wrote to replace my display manager. It works by detecting DEs in /bin but it is about 200 lines of if statements. How would I go about condensing it? 23:02 < camsn0w> Kinda depends on how you wrote it and what is contained in the statements 23:02 < camsn0w> maybe use a switch 23:03 < tips-fedora> it basically goes "does this desktop environment exist? if so echo it's name out onto the screen, store the count variable into the desktop variable, then count++ and continue until you find another desktop" 23:04 < tips-fedora> then it goes "okay now enter the number of the desktop you want, does it match with the numnbers on screen? if not run the script again" 23:07 < storge> cool 23:08 < TingPing> tips-fedora, store the valid DEs in a table 23:09 < TingPing> tips-fedora, a better solution entirely is probably to look in /usr/share/xsessions instead of looking for binaries? 23:11 < tips-fedora> so write a function to scan for DEs, then store them in a table? I could look in xsessions but I also want to include tmux and any virtual machines I have. I guess I could just add them explicitly 23:13 < TingPing> at some point a shell script may not be the best format :P 23:13 < phinxy> Loshki• The driver I'm trying to compile depends on some ION Memory Manager. That is located in /staging/. I ticked the CONFIG_STAGING and oldconfig now asks me about it. This way of adding the driver is probably more correct than adding #include " I probably could figure something out in C, but then I would have to restart 23:15 < TheNH813> I need help with cpu frequency scaling. 23:15 < TheNH813> I have cpufreq_ondemand loaded, but cpupower is not reporting anything. 23:16 < TheNH813> It returns a statement like this: analyzing CPU 0: no or unknown cpufreq driver is active on this CPU 23:16 < TheNH813> I guess I could always change the settings on the motherboard config, but I'd like to be able to set the options from the os. 23:17 < TheNH813> Also, the folder /sys/devices/system/cpu/cpu*/cpufreq dosen't exist. 23:17 < TheNH813> Any advice? 23:18 < streuner> how's your recommended size for root partition on 1 TB drive? 23:18 < Loshki> phinxy: yours is a link-time error, first thing to check is a missing library. Did you say which symbol is missing? That will help us find it. 23:18 < streuner> i am going to install arch linux 23:18 < streuner> gentoo takes a lot of my time for compilation 23:18 < Industrial> Hi. I can probably use rsync best to keep 2 directories in sync, right? 23:18 < tips-fedora> Gentoo is nice but if you have a budget CPU it'll be painfull to install 23:18 < Industrial> I want to do things like avoid re-writes of unchanged files, etc 23:19 < TheNH813> Yes. Rsync is meant for that. https://linux.die.net/man/1/rsync 23:19 < Industrial> What I would like to know is: Is there a way to (re-)trigger the rsync from the Source to the Target directory when a file inside the Source directory (recursively) changes? 23:20 < TheNH813> Hmmm. 23:20 < Industrial> I want it to be continuous. 23:20 < TheNH813> Interesting question. 23:20 < streuner> tips-fedora, i have quad-core intel core i5 6300-hq 23:20 < Industrial> Like if it were a symlink but it isn't, since Docker doesn't support symlinks :/ 23:20 < phinxy> http://termbin.com/6zpm that was the error. It looks to link alright now.. No need to investigate further 23:21 < kerframil> Industrial: with inotify-tools, though if that's the requirement, event-driven invocations of rsync may not be the best fit for your problem 23:21 < TheNH813> Industrial: It could be scheduled to check, say every 5 minutes with a script. Otherwise it will require watching for all writes to the drive. 23:21 < tips-fedora> streuner: that should work okay though. I have a bunch of old computers with budget processors like Pentiums and Celerons so gentoo won't work for me 23:21 < tips-fedora> Even compiling emacs takes like 30 minutes 23:22 < Industrial> Basically I have a dir of common code I need to distribute to my different services' directories such that the Docker Compose Volume mount will pick up the file write and put it inside the running docker container through the volume 23:22 < streuner> tips-fedora, what's your recommended root partition size for arch linux? 23:22 < kerframil> tips-fedora: in fairness, you don't have to build on the old computers in question. that said, pacman is about as fast it gets. 23:22 < Industrial> I guess I could write a Node.js program for it 23:22 < Industrial> (probably uses inotify) 23:22 < tips-fedora> steruner: what's the size of your disk? 23:22 < TheNH813> streuner: Arch is a good idea. Especially because it installs pretty much nothing. Build you own OS, choose the DE adn everything. I like Arch because of that very reason. 23:23 < phinxy> LINK vmlinuz, is this close to the end? 23:23 < streuner> tips-fedora, 1 TB 23:24 < TheNH813> streuner: My install on my desktop is on a 16GB SSD, but it's a bit of a squeeze. I'd say maybe 30GB-100GB is reasonable. 23:24 < tips-fedora> steruner: I have an 80G disk with debian installed on it. the root partition is 30G and I haven't had any problems 23:25 < tips-fedora> streuner: I'd probably go with 100-200G depending of if you want lots of programs 23:25 < tips-fedora> the root partition also stores log files so bigger is sometimes more helpful 23:25 < TheNH813> Or if you're storing lots of data in /var. 23:25 < TheNH813> Ex: A web server. 23:25 < streuner> tips-fedora, is it a good idea make dual boot arch with windows 10? 23:25 < streuner> does windows 10 work well in virtualbox? 23:26 < TheNH813> It works well, unless you need graphical acceleration. 23:26 < streuner> i need win10 for visual studio and unity3d 23:26 < cactus00000> streuner: yes, i use qemu. 23:26 < kerframil> streuner: it's stable but the graphical acceleration isn't great 23:26 < TheNH813> Install Windows first if you plan on a hard install. Otherwise Windows might nuke grub during installation. 23:26 < streuner> for linux, unreal engine is better than unity3d 23:27 < kerframil> streuner: certainly, I would recommend disabling animation effects and the like 23:27 < tips-fedora> streuner: I don't recommend dual booting with Windows 10 because it has been known to overwrite people's GRUB. VIrtualbox should work fine. You can configure it to be closer to your actual hardware if you do a little work 23:27 < streuner> TheNH813, how can I encrypt win10 installation (assuming use open-source software)? 23:27 < TheNH813> VeraCrypt can do that. 23:27 < streuner> i don't trust bitlocker 23:28 < TheNH813> It's basically the successor of TrueCrypt. 23:28 < kerframil> regarding the grub overwriting issue, the solution is to use UEFI booting if possible 23:28 < streuner> is veracrypt enough good as cryptsetup? 23:28 < kerframil> veracrypt has been audited 23:28 < TheNH813> Cryptsetup actually supports veracrypt drives as well. 23:28 < TheNH813> So yes, they're quite compatible. 23:28 < cactus00000> ubuntu is too cute, wifi just stopped working after a week of uptime, gosh i wonder if it is the 10$ wifi card :o/ 23:29 < tips-fedora> streuner: I just use LUKS. If your threat model is the NSA nothing will be safe, but if you want the people physically around you to stay out, LUKS and booting into a command prompt will stop them 23:29 < tips-fedora> afterall, normies can't startx if they don't even know what Xorg is 23:29 < TheNH813> Well, Windows can't exactly boot off LUKS. 23:29 < streuner> https://aur.archlinux.org/packages/unity-editor/ 23:30 < TheNH813> But that's not a bad idea, not starting X. 23:30 < kerframil> another advantage of UEFI booting is that the veracrypt UEFI executable is much faster than the one that runs in real-mode 23:31 < TheNH813> UEFI is nice. Especially if you turn ultra fast boot on. 23:31 < tips-fedora> I'm actually in tmux right now. Xorg is too resource intensive for my budget CPUs lol 23:31 < TheNH813> lol 23:31 < TheNH813> Hey, if it works for you, that's what counts. 23:31 < streuner> how's difference between uefi and bios? 23:32 < TheNH813> UEFI uses bootable modules stored in a parition, BIOS uses the boot sector of a drive. 23:32 < TheNH813> To put it simply. But there are many other HUGE differences. 23:33 < TheNH813> UEFI also supports signed kernels, so only kernels with a key stored in the motherboard will boot. IT's called secure boot. 23:33 < streuner> uefi is more secure than bios? 23:33 < TheNH813> I would say so. It's a bit of a pain to get Linux booted on sometimes because of it though. 23:34 < TheNH813> Most distros have support for secure boot now. 23:36 < TheNH813> The thing about UEFI that's way different is all the motherboard looks for is a partition that is of the 'ESP' type (basically fat32 with a custom header), and looks for a 'boot' subdirectory to load files from. 23:38 < jim> TheNH813, yeah, it's mostly because grub2 has support for uefi boots 23:38 < TheNH813> jim: Yeah. 23:38 < Dan39> TheNH813: "custom header"? is ESP more than just a partition type? 23:39 < Dan39> i know it has to be FAT32 23:39 < TheNH813> Well, the partition ID is different, even though the underlyign filesystem is fat32. 23:39 < Dan39> right 23:39 < streuner> i want to use windows mainly for fl studio, visual studio and unity3d 23:39 < Dan39> so i wouldn't really call it a custom header, its just a different partitiont ype id 23:40 < TheNH813> Yeah, that's a better way to put it. 23:40 < Dan39> its a regular FAT32 fs 23:40 < Dan39> which is the nice thing about it 23:40 < Dan39> there is no longer magic custom data spots or magic headers 23:41 < Dan39> its a standard partition with standard fs and just a standard partition type set 23:41 < kerframil> Dan39: while ESP has a specialised partition type code, in practice the firmware doesn't necessarily care 23:41 < Dan39> kerframil: was going to say that next, iirc i read that many firmwares aren't even that strict 23:41 < TheNH813> streuner: I'd use a real install of Windows for a DAW and graphics acceleration heavy applications. 23:41 < kerframil> Dan39: they key thing is that it's a FAT32 filesystem that has an EFI/ sub-directory 23:41 < TheNH813> I had choppy performance for FL Studio in a VM. It actually worked better in Wine, surprisingly. 23:42 < Ameisen> hrmm 23:42 < Ameisen> 2m 35s isn't bad for building the kernel. 23:42 < TheNH813> That's REALLY fast. 23:42 < kerframil> Dan39: no magical real-mode code required within the first 440 bytes of the first sector either 23:42 < Ameisen> it's also a very trimmed down kernel 23:42 < TheNH813> I will say one thing, I like the fact my Lenovo laptop respects subdirectories of efi/ 23:43 < Dan39> kerframil: you're preaching to the choir here, thats what i was just telling the other guy lol 23:43 < Ameisen> but yeah, that was the real build time 23:43 < Ameisen> should build faster with this kernel, actually 23:43 < TheNH813> It reports Grub and Windows Boot Manager as seperate entires in the boot menu itself. 23:43 < TheNH813> Which means that Windows can overwrite /boot/efi/boot all it wants, because /boot/efi/arch will always be found. 23:44 < Dan39> TheNH813: nice, i noticed some mobos nicely detect and display the different bootable efi files automatically, while one i dealt with recently i had to use efibootmgr to get it booting 23:44 < TheNH813> Dan39: Exactly. Some motherboards aren't very kind. 23:44 < kerframil> Dan39: oh, ok 23:44 < kerframil> Dan39: sorry! 23:45 < TheNH813> Some older HP motherboards in specific are extra pesky. 23:46 < Dan39> the situation with secure boot is interesting too, how Microsoft is pretty much the one in control 23:46 < TheNH813> Yeah. 23:46 < Dan39> at least we got shim signed, but i feel like that is less secure than having grub signed and booting straight to grub 23:47 < Dan39> or systemd-boot, whatever 23:48 < TheNH813> I think I'l limit my desktop's cpu frequency in the motherboard config. Can't get the governor modules to work. 23:49 < TheNH813> I want it to produce less heat as all it does is idle and serve files to my laptop. 23:50 < TheNH813> And I occasionally offload tasks to it. 23:50 < TheNH813> gonna reboot and change some of the turbo settings to get the desired effect. 23:50 < Dan39> TheNH813: i often see BIOS settings to allow the os to throttle cpu. sure that isnt turned off? 23:53 < Dan39> oh well 23:58 < mavorsa> is it just me or is every single modern gtk theme these days abhorrently bland, boring, and a copy of the exact same thing? 23:58 < swine_> does anyone here have experience with MAAS? 23:59 < mavorsa> "i've got this AMAZING idea for a gtk theme guys!........ it will be flat, the buttons will have a slight gradient, and the window control buttons will have a _, a square, and an x in the corner! it will come in either plain white, or plain dark grey!!!!" - 95% of all gtk theme designers. --- Log closed Mon Jun 25 00:00:12 2018