IRC Logs for #crux Monday, 2018-06-25

betanbhi, I can't find the documentation for Pkgfile, like how could I change how pkgmk extract the tarball00:00
betanbI mean, is build() the only function on an Pkgfile?00:02
betanbor can I override some steps, like how to download the tarballs?00:03
rmullfrinnst: Thanks, that caught be by surprise00:12
frinnstyeah, shame00:13
frinnstbetanb: check out the sources.00:13
betanboh, nice, turns out pkgmk is actually quite simple.00:18
frinnstlike most crux stuff :-)00:41
*** frinnst has quit IRC00:52
*** frinnst has joined #crux01:02
*** tilman_ has joined #crux01:17
*** _________mavric6 has quit IRC02:21
*** _________mavric6 has joined #crux02:22
*** smolboye has joined #crux04:37
*** parlos has joined #crux04:46
*** betanb has quit IRC05:06
*** parlos has quit IRC05:07
*** abenz has joined #crux05:20
*** parlos has joined #crux06:24
*** abenz has quit IRC06:56
*** john_cephalopoda has joined #crux07:10
*** abenz has joined #crux07:18
*** workodera has joined #crux07:27
*** de11 has joined #crux07:27
*** rarz has joined #crux07:36
*** jue has joined #crux07:59
*** jue has quit IRC07:59
*** jue has joined #crux07:59
*** jue has quit IRC08:07
john_cephalopodaJun 21 12:48:14 hera sshd[30397]: Disconnecting invalid user admin port 22534: Too many authentication failures [preauth]08:34
john_cephalopodaJun 21 19:30:35 hera sshd[32528]: Disconnected from authenticating user root port 10181 [preauth]08:34
john_cephalopodaOh, the joys of running a server.08:35
ryuojohn_cephalopoda: if it's public facing, i normally disable password logins. that solves most of these drive by attacks.08:38
ryuoIt's a lot harder to brute-force a private key than a mere password.08:39
john_cephalopodaI do that, too.08:40
john_cephalopodaThus it fails at preauth state already.08:40
*** abenz has quit IRC09:02
*** parlos has quit IRC09:06
*** parlos has joined #crux09:13
*** de11 has quit IRC09:48
*** rarz has quit IRC10:30
john_cephalopoda <-- Hehe10:32
*** samathy has joined #crux10:48
*** btxd4-yaaan has joined #crux10:51
*** btxd4 has quit IRC10:54
*** samathy has quit IRC10:58
*** samathy has joined #crux10:58
pedjais moving ssh port from 22 worth it these days?11:02
*** samathy has quit IRC11:10
*** samathy has joined #crux11:13
*** samathy has quit IRC11:25
*** samathy has joined #crux11:25
pedjaso, risc-v isn't as open as people hoped. isn't the point of it to be royalty-free, so anyone can make it?11:29
Romsterroccat new keyboard cherry mx brown switches, gonna take a bit to get used to.11:30
pedjapics, please :)11:30
Romsteri've only ever had membrane cheap keyboards11:30
Romsteri haven't customised it yet11:30
Romsterneed to figure out the linux software for the lights etc. which i know exists11:31
pedjaRomster, which model? I am guessing Suora?11:32
Romsterand 3 monitor setup xorg but i've only made it 2 just now.11:32
RomsterRoccat Ryos FX, Cherry MX Brown, RGB LED11:33
Romsterand i've had a roccat xtd krone mouse for 3 years already11:33
pedja1600g? damn, that's a weapon11:34
Romsteryeah it is heavy11:35
Romsteri dabble in games so it made sense for the ryos i guess11:37
Romsterand i can easy shift the keyboard with my mouse button11:37
pedjaare you using this ?11:40
Romsteri will be i did package it ages ago in romster/11:42
Romsterfor someone else to use and then i finally decided to get my own11:42
Romsterthe things i could so with the scripting, didn't really think of that.11:43
Romsterthough i haven't messed with lua11:43
pedjaI still have some Lua scripts from conky days11:48
pedjanot written from scratch, mind you, modifies stuff from the Internets :)11:51
pedjacan't see the point of conky, really, since I rarely see the desktop anyway11:53
pedjait's either xfce-terminal or some app at full screen11:53
Romsteryeah i dunno i got conky but it's mostly covered11:53
Romstertend to use htop more than anything else.11:54
Romsterand only run it when needed11:54
Romsteri like the feel of this keyboard already11:54
Romster i see this and then there is the sourceforge site11:55
pedjawhen you go mech, you can't go back, to quote frinnst :)11:55
Romsterwe'll see if that sticks, but so far yeah i think frinnst is right.11:56
Romsterbut this feels so alian for me as the key spacing is different and the caps lock key does something else.11:57
frinnstmech ruins everything else11:57
Romsterhaving 3 monitors ruins everything else too.11:58
frinnstI make do with 2 at home. I could use 4 at work :/11:59
pedja'make do' #firstWorldProblems12:01
Romsterbut finally back slash is where it's meant to be above the enter key!12:06
dbrookepedja: I run ssh on a non-standard port and it does reduce the log clutter, though I have had to change the port recently as it was discovered and seemed to get into some scripts12:06
Romsteri always hated the larger enter key and stupid placement of the back slash key next to the backspace key.12:07
frinnstRomster: what are you on about, thats the *correct* placement :-)12:07
dbrookefrinnst: no, and I'm not telling you 8-)12:07
frinnstheh since you said it was used by scripts I assumed 2222 might be it :-)12:08
Romsterwell it was to the left of the backspace key now it is under it!12:08
dbrookeuh-oh, keyboard layout wars ...12:08
Romsterlol dbrooke12:08
frinnstare you using it at home or work romster?12:09
Romsterat home12:09
frinnstyou will probably be alright for this week. but a weekend of heavy use will ruin your work keyboard for you :-)12:09
Romsterprobably but linux doesn't use back slash that much but i do use the pipe shift lots. and windows only uses the back slash for shares.12:11
frinnsti mean the switches12:11
Romsterah yeah i might have to get a keyboard for work too.12:11
Romsterbut the issue is i work on several notebooks and desktops.12:12
Romsterone keyboard isn't enough even with a kvm for 4 machines12:12
Romsteri could have like 5 usb keyboards and mice around the work area.12:13
dbrookefrinnst: it did end in 22 but I suspect someone found it by scanning and it ended up in some kind of cracker database since it suddenly started getting attacked from many different addresses12:13
Romstersetup denyhosts?12:13
Romsterargh now i have the issue of losing my mouse pointer.12:13
Romsterknew i would have this12:14
dbrookeit'll probably take another few years for it to be found again 8-)12:14
Romsteroh man12:14
dbrookeI'm not worried about security as it's locked down12:14
pedjajust add 2 whenever bot scanning spikes :)12:16
dbrookewell, I actually added 2000 to the previous non-standard number, so it still ends in 2212:19
Romstermasker it use ssh keys and lock down the passwords12:25
Romsterdoubtful anything will get past it then12:25
Romstermake it port knock to open?12:25
*** abenz has joined #crux12:25
pedjawouldn't even with keys bots still try to login, and fail at pre-auth?12:27
cruxbot[compat-32.git/3.4]: nss-32: 3.37.3 -> 3.3812:27
Romsteryeah fail at invalid password or pre-auth12:33
*** abenz has quit IRC12:33
Romsterthe other thing you could do is use the geoip iptables module to limit what countries can use it12:33
Romsteror even use the isp netmask range12:34
frinnstor ipv6 only12:35
cruxbot[contrib.git/3.4]: [notify] gsmartcontrol: 0.8.7 -> 1.1.3 dependency change gtkmm -> gtkmm312:57
cruxbot[contrib.git/3.4]: atkmm: update source12:57
cruxbot[contrib.git/3.4]: mm-common: 0.9.10 -> 0.9.1212:57
pedjaipv6? yeah, I asked my ISP about that. 'why do you want it? you're not a business'13:14
dbrookeit's key only and specific user, it was on port 22 on IPv6 which didn't get hit but I changed it to be the same port on both protocols to simplify my client config13:14
pedjaRomster, geoip iptables module is an interesting idea13:15
*** de11 has joined #crux13:16
rmullI use fail2ban and it works pretty well13:17
dbrookesince it runs mostly public services I don't even have iptables configured13:17
pedjacan geoip module do white-listing?13:18
pedjaprobably would be my guess :)13:19
dbrookeI found fail2ban not too effective as the attack rate is low and I was never brave enough to block on a single fail (though using keys I shouldn't be able to mess up and lock myself out)13:19
Romsteri have noticed some attacks were so low rate that i had a monthly ipset table of hashes that got added to.13:23
Romsterto stop those13:23
pedjarate-limiting ssh looks to be the another thing people are doing13:32
pedjasecurity is hard, apparently13:34
Romsterif 10 or more in a month of failed attempts and no successful attempts to reset that count then block them outright.13:34
Romsterit's so much easier to pull the plug than to deal with it all.13:34
Romsteror hide it over openvpn or something13:35
pedjanever used vpn, but I did try to make strongswan port13:40
Romsteri did think about a vpn from my phone to home but i never got any further than that thought.13:42
pedjaspeaking of vpn, is wireguard merged upstream yet?13:42
Romsterno idea13:42
Romsterand ugh looked at keyboard more it has a windows key for the menu on it... looks intently at finding a replacement keycap for that.13:43
Romsterwonder if i can reassign it for something else.13:43
pedjaI use it to switch languages13:43
Romsteri only know english and i am sure the lot of you here think i don't even know that lol13:44
pedjaI'll let you native speakers fight it out :)13:45
Romsteri did try to learn devork but then it got annoying going back to qwerty at work so i gave up on that.13:47
Romsternot a big deal i'll figure something out for it.13:48
pedjawell, if you are that annoyed with it, perhaps
Romsterthis keyboard has RGB though but that's a nice option for notebook keyboards14:00
dbrookemy keyboard has Tux keys (which I use as the modifier key for dwm) instead of Windows keys
john_cephalopodaI use the Win-key as modifier key in i3wm. It has been used so extensively that the logo is already gone.14:11
dbrookegoing that way here but not as bad as some others14:12
Romsterjohn_cephalopoda, lol nice14:17
*** samathy has quit IRC14:25
*** samathy has joined #crux14:25
*** abenz has joined #crux15:15
*** de11 has quit IRC15:26
*** abenz has quit IRC15:26
*** workodera has quit IRC15:31
*** abenz has joined #crux15:55
pedjaIntel's comment: 'No big deal, nothing to see here' :)15:58
*** abenz has quit IRC16:00
*** SiFuh_ has quit IRC16:03
*** SiFuh__ has quit IRC16:03
*** SiFuh has joined #crux16:13
*** SiFuh_ has joined #crux16:13
pedjagrr. mounting usb sticks from Thunar is broken again.16:17
*** samathy has quit IRC16:20
*** samathy has joined #crux16:22
*** abenz has joined #crux16:35
*** abenz has quit IRC16:40
*** samathy has quit IRC17:00
*** onodera has joined #crux17:17
*** onodera has quit IRC17:17
*** onodera has joined #crux17:17
pedjaheh. usb stick I tried to update to 3.4 with boots fine using qemu, but wouldn't when I tried on the actual hardware17:52
*** tsaop has joined #crux18:00
tsaopRomster: there's a problem with the qt5 port,see the .footprint diff:;a=blobdiff;f=qt5/.footprint;h=9151ab89a325ff434376b052e03ca9f116ecc08d;hp=68d8266c766127b8aa69a3f301285f0dd4d958d2;hb=a21902a61e176561a09fdb0da829d30ec7acb1f0;hpb=1d691dd733b2c945b41cc356be60c58345c58f3718:01
*** smolboye has quit IRC18:02
tsaopRomster: a lot of cmake files got axed. To fix this you have to comment out the private includes fix in the Pkgfile, since it has been fixed with Qt 5.11.118:02
*** smolboye has joined #crux18:03
pedjayup, I tried to build Stellarium, which is using cmake to find qt5. it is broken :)18:37
pedjadamn, and I just updated qt5 this morning18:38
pedjathanks, tsaop18:39
pedjahm, dd transfer speed falls to ~6Mb/s after 1Gb or so of writing to usb stick. time to get a new one, I guess18:43
frinnstbadblocks is fun18:47
pedja'mkfs.vfat -c' too18:55
pedjaI forgot to pass verbose to it, and now I wait in anticipation18:57
ryuopedja: usb sticks? they're usually super slow.18:58
pedjaryuo, I know, but from ~80-100MB/s to 6 :) ?19:02
tsaopnow, if only could I solve the PyQt5 5.11.1 compilation problem19:05
tsaopI already symlinked the sip private copy in PyQt5 folder19:06
pedjatsaop, what's the issue?19:06
pedjaI yet have to try it, qgis3 depends on it, iirc19:06
ryuopedja: lmao.19:14
ryuopedja: what do you use it for? booting stuff?19:14
pedjayes. I am trying to make multiboot usb19:15
ryuopedja: i've started using iodd2531 myself. put a 2.5" SSD or HDD inside and you can directly boot from virtual CD drive that uses ISO files directly.19:15
ryuoBut, then again, i use it because i got tired of USB preparation.19:16
ryuowith this device, i just copy the ISO over and i'm done.19:17
ryuoI just need to tell it which ISO to use.19:17
ryuoAnd, it works on any ISO from what i've seen.19:17
ryuoi don't need to fuck with special software to make them USB bootable.19:17
pedjaryuo, nice. yes, that looks like a more elegant solution.19:20
tsaoppedja: it errors out at something not defined in QtWebEngine19:20
tsaopI checked the includes and it appears to be defined19:20
*** deep42thought has joined #crux19:22
deep42thoughtHi, I'm struggling to get texlive compiled again after the poppler problem - currently it complains about "../../../texk/web2c/pdftexdir/ error: invalid conversion from 'const char*' to 'char*'"19:29
deep42thoughtACTION tries his rebuild-all-dependencies hammer :-)19:33
frinnstpedja: the speed decrease is pretty normal. the first gb is cached and the rest is written directly to disk20:01
frinnstor, flash rather20:01
frinnstdepends on the options you pass to dd20:01
deep42thoughthrmm, still the same :-/20:03
frinnsttry this port:
deep42thought... this will take a while (download is only 2MB/s here)20:10
frinnstits groovy2shoes's port20:11
deep42thoughtok, thx, I'll report back if it worked :-)20:14
pedjafrinnst, dd and dd_rescue report similar speeds, and the behaviour is pretty similar with both. oh, well20:17
pedjaI see PyQt is PITA as usual20:21
tsaoppedja: eh, it spits out errors20:22
tsaopwill wait a few days to copy what other distros do20:22
pedjanever got that far. private sip module copy? wtf20:23
tsaoppedja: someone linked it here maybe yesterday20:24
tsaophold on20:24
tsaopI don't know whether to make a package with just the private copy or to simply depend on sip + symlink20:25
tsaopsee my sip Pkgfile:
pedjaI would do the first thing20:27
tsaopmore like sip-pyqt5?20:27
pedjayeah, that :)20:28
tsaopbut first I must make pyqt5 compile20:28
deep42thoughtalso fails :-(20:39
pedjano wonder 'from PyQt5 import sip' didn't work. wrong python.20:40
*** onodera has quit IRC20:59
*** pedja has quit IRC21:05
*** pedja has joined #crux21:06
deep42thoughtin case someone can help, here's my log:
*** girls has joined #crux21:11
pedjatsaop, my py3qt5 build fails with 'unable to create /usr/lib/python3.6/site-packages/PyQt5-5.11.1.dist-info'21:19
tsaoppedja: using my Pkgfile?21:19
pedjano. let me see it21:20
tsaop note that it also fails21:20
tsaopmake sure you have created a private copy of sip into PyQt5 before using it21:21
pedjaI am using your sip port.21:22
pedjathe only difference between yours and mine pyqt5 Pkgfile is '--no-sip-files'21:23
pedjaand mine fails at install stage, afaict21:23
pedjaI had the same issue with my sip port, Makefile is broken and tries to install dist-info outside DESTDIR21:33
*** smolboye has quit IRC21:33
pedjaI disabled it, let's see if it builds now. it will take a while, thou21:34
*** tsaop has quit IRC21:36
groovy2shoesdeep42thought, argh, I know what the problem is21:56
deep42thoughtoh, wanna know!21:56
groovy2shoesthe configure script for zziplib is expecting /bin/sh to be more bash-like than the default dash...  there's an environment variable that can tell the script to use bash instead (I'd linked /bin/sh to /bin/bash on my system and forgot about it; now I've switched it back and testing a small change to set the env before configuring)22:01
groovy2shoesthat was an infuriating one to debug, because all those errors from GCC are created by a shell script, generated by configure, that uses sed to change the source files...  dash doesn't like the way quotes are escaped in the generated script, which is apparently left implementation-defined by the POSIX spec for sh, though it seems to me that just about every bourne-style shell follows the ksh behavior *except* dash ¯\_(ツ)_/¯22:06
pedjagroovy2shoes, 'CONFIG_SHELL=/bin/bash ./configure...' ?22:09
groovy2shoesI updated the tarball at the link with the new Pkgfile, looks like it's compiling okay, but it'll be a while before I'll know whether or not it finishes22:09
pedjaah :) that bit me a few times22:10
groovy2shoesI wasn't sure it would work because the problematic script is generated haphazardly by a custom "module" tossed into the source tree, not by autoconf or configure per se, so I wasn't sure it'd respect CONFIG_SHELL, but it looks like it does  :)22:13
deep42thought"MiscXS.c: loadable library and perl binaries are mismatched (got handshake key 0xdb80080, needed 0xde00080)"22:19
deep42thoughtlooks like I need to rebuild some perl thingy?22:19
groovy2shoesyeah, but it's not clear to me which thingy you need to rebuild from just that line22:20
deep42thoughtmakeinfo is owned by texinfo (from my repo), I just rebuilt that - let's see22:27
john_cephalopodaVarious perl things are kinda broken.22:27
deep42thoughtjohn_cephalopoda: hmm, what do you mean by that?22:32
groovy2shoesyeah, I don't have makeinfo  ¬_¬  but that does look like a likely culprit for the problem22:32
deep42thoughtgroovy2shoes: yes, looks, like I solved that problem (have 14k lines of log now, when last time it failed at line 10k)22:33
groovy2shoesthe latest Perl update broke a lot of things that implicitly (or incidentally) depend on it due to ABI changes22:33
deep42thoughtshouldn't revdep point out these issues?22:34
groovy2shoesprobably not22:35
deep42thoughtah, right, perl isn't haskell :-)22:35
deep42thoughtbtw: it's executing the last "find ... -delete" now :-)22:36
deep42thoughtthanks groovy2shoes!22:36
groovy2shoesthe binary Perl modules get loaded at runtime with dlopen(), so there's no evidence in the ELF headers that they're needed22:36
groovy2shoesnp!  thanks for letting me know about the issue  :)22:36
*** girls has left #crux ()22:37
groovy2shoesmight help to get a list of the files from /usr/lib/perl5/site_perl/{5.24,5.26} and look for files in the old dir that aren't present in the new one, to get some clues as to which packages might need to be rebuilt against the new perl libs22:38
deep42thought hmmm22:47
groovy2shoesp5-xml-parser at least22:49
deep42thoughtp5-xml-simple it is22:49
groovy2shoesyou'll want to look inside those linux-thread-multi subdirs, too22:49
deep42thoughtyes, I'm doing that currently22:50
deep42thought/usr/lib/perl5/site_perl/5.24/CPAN/ is from /usr/ports/contrib/p5-cpan-distnameinfo22:52
deep42thoughtwhich has the wrong footprint :-)22:53
groovy2shoesprobably lots of stuff tosses files into the CPAN dir (basically perl's hackage)22:54
deep42thoughtno, was actually only those two packages22:54
deep42thoughtnow the 5.24 dir is gone completely :-)22:54
groovy2shoesiirc, some of those files come from the perl package itself22:54
groovy2shoesnice  :D22:54
deep42thought`prt-get 5.24` gives a list of all packages that need to be updated ;-)22:55
groovy2shoeswell, sometimes you can wind up with perl modules in there that aren't parts of a package, if you have a habit of installing things willy-nilly  ¬_¬22:57
deep42thoughtyeah, ok22:57
deep42thoughtbut wrapping something in a Pkgfile is so painless, that I tend to put everything into one22:58
groovy2shoesprobably a good habit22:58
deep42thoughtok, time for my bed - good night and thanks again!22:58
groovy2shoesbut sometimes it's easier to just do `cpan Foo::Bar` and let it handle its own dependencies22:58
*** deep42thought has quit IRC22:59
*** z7f0 has joined #crux23:14
*** TheCephalopod has joined #crux23:30
*** john_cephalopoda has quit IRC23:34

Generated by 2.14.0 by Marius Gedminas - find it at!