IRC Logs for #crux Saturday, 2016-08-06

*** pejman has quit IRC00:39
*** Romster has quit IRC00:46
*** Romster has joined #crux00:49
*** admincomplex has quit IRC01:01
*** admincomplex has joined #crux01:03
*** heroux has quit IRC01:06
*** pejman has joined #crux01:09
*** onodera has quit IRC01:09
*** heroux has joined #crux01:14
*** heroux has quit IRC01:34
*** heroux has joined #crux01:58
*** heroux has quit IRC02:03
*** Romster has quit IRC02:04
*** Romster has joined #crux02:05
*** heroux has joined #crux02:05
*** heroux has quit IRC02:16
*** heroux has joined #crux02:24
*** _________mavric6 has quit IRC02:28
*** _________mavric6 has joined #crux02:29
*** heroux has quit IRC02:31
*** heroux has joined #crux02:46
*** heroux has quit IRC03:07
*** Romster has quit IRC03:23
*** Romster has joined #crux03:24
*** heroux has joined #crux03:36
*** heroux has quit IRC03:43
*** heroux has joined #crux03:52
marakuhello! so I rewrote the script for getting package build times (https://raw.githubusercontent.com/maraku/bin/master/pkg)04:05
*** tilman_ has joined #crux04:05
marakufor now it justs gets build times of packages from prt-get's logs and if you're building a package you can check how long it's been building04:06
*** Romster has quit IRC04:08
*** Romster has joined #crux04:08
*** Romster has quit IRC04:19
*** Romster has joined #crux04:24
*** Romster has quit IRC04:33
*** Romster has joined #crux04:34
*** Romster has quit IRC04:45
*** Romster has joined #crux04:47
*** tragique has joined #crux04:55
*** lounge has joined #crux04:57
*** Romster has quit IRC04:58
*** Romster has joined #crux04:58
*** pejman has quit IRC05:01
*** pejman has joined #crux05:07
*** Romster has quit IRC05:14
*** Romster has joined #crux05:19
*** Romster has quit IRC05:27
*** Romster has joined #crux05:27
*** darfo has quit IRC05:32
*** Romster has quit IRC05:36
*** Romster has joined #crux05:36
*** Romster has quit IRC05:45
*** Romster has joined #crux07:36
*** nebg has joined #crux07:42
*** nebg has left #crux ("Leaving")07:43
*** tragique has quit IRC08:19
*** nebg has joined #crux09:38
*** nebg has left #crux ("Leaving")09:39
*** Na3iL has joined #crux10:57
*** Na3iL has quit IRC11:08
cruxbot[opt.git/3.2]: fakeroot: fix for glibc 2.2411:29
cruxbot[opt.git/3.2]: clex: fix man-page location11:29
cruxbot[opt.git/3.2]: fakeroot: forgot to add patch11:33
*** onodera has joined #crux12:52
*** onodera has quit IRC13:45
*** Na3iL has joined #crux15:00
*** nebg has joined #crux15:30
*** nebg has left #crux ("Leaving")15:30
cruxbot[contrib.git/3.2]: hiredis: initial import15:34
*** nebg has joined #crux15:41
*** nebg has left #crux ("Leaving")15:42
frinnstmissing some relevant information in the hiredis Pkgfile :)15:54
teK_got pulled away by my landlord, had to fix shit on his windows 10 install16:14
teK_(the windows icon stopped working, wtf)16:14
teK_planned on sneaking the fix in with the next commit for redis-py16:14
teK_;)16:14
frinnstah, yeah win10 have had issues with the startmenu not working and crap like that16:16
cruxbot[contrib.git/3.2]: redis-py: initial import16:17
teK_I installed classic shell and screwed with the fucking registry and permissions for SYSTEM (!!!) for one key so it will always display the list of users at the login screen without logging in the last user16:17
teK_superior GUI my ass16:17
frinnst"windows is much easier"16:17
frinnstgod gtk3 3.20 looks like shit16:18
teK_oh and the checkbox for 'users have to logon using user and pow' is ticked.. in a dialouge opened by ' control userpasswords2 '16:18
teK_example?16:18
teK_http://i1-news.softpedia-static.com/images/news2/gtk-plus-3-20-toolkit-to-include-the-windows-theme-on-all-platforms-500556-2.jpg ?16:18
frinnstthey completely broke themes16:19
frinnstand firefox-gtk3 looks like shit16:19
teK_100px top bar. wtf :)16:20
frinnsti dont understand how people are ok with thtat16:20
frinnstscreen space is fucking valuable16:20
teK_% xrandr | head -1 | cut -f2 -d, current 3840 x 216016:22
teK_btw I looked, again, into the github content-disposition header thingy16:22
frinnstcurrent 5760 x 120016:23
teK_was there a final verdict on how to proceed?16:23
frinnstoh, what did you find out?16:23
teK_the VM is only on one screen atm ;)16:23
frinnst..16:23
teK_yeah basically they are using this header so GUI browsers will use a sane filename16:23
frinnstwget lacked proper support for it, or what was the issue? i forget16:24
frinnstalso we need to consider how curl handles it16:24
teK_so we are screwed in the sense that we cannot parse the header and internally redirect md5/signature checking to the right 'file' (compared to whats in our source array)16:24
teK_wget and curl do support the header with according siwtches, as far as their man pages ar concerned16:24
frinnsthow do you suggest we resolve the md5/sig issue? any ideas?16:26
teK_nothing that's not 'hackish'16:27
frinnstyuck16:27
teK_as you want to check also if the file has been downloaded previously you'd have to reconnect to gh to find the 'right' filename16:27
teK_which sucks and rules parsing / using the header out at least in  a general solution16:27
teK_my take is, that currently only github is serving files like /v[0-9]\.[0-9](\.[0-9])*.tar.gz..16:30
frinnstyup16:30
teK_so we could check for that while looping throug source() and transparently redirect file operations (download, sigcheck and extraction) to hide this madness16:31
teK_it's not beautiful but fixes the issue and should be doable pretty pretty easily16:31
jaegerwget and curl have support for it but it was experimental or something16:39
jaegermight be worth a try, would be simpler than manually fixing the names16:40
teK_how is this supposed to work16:40
teK_consider this.16:40
teK_1) download the file with pkgmk -do16:40
teK_2) later run pkgmk16:40
teK_2a) pkgmk will check source() files16:40
jaegeryou'd still have to map names somewhere but it might be less work, not sure16:41
teK_2b) pkgmk has to invoke wget/curl for each and every file to find out its 'real' name on this16:41
jaegerhaven't planned it out or anything16:41
teK_3) this sucks :p16:41
teK_I think that using external info via http or some file should be avoided and the solution above is far more clear and not hard to implement :]16:41
jaegeralternatively we could finally implement a "rename this source archive in the source array"16:41
jaegerfeature16:41
teK_do I hear POC or didnt happen?16:42
*** lounge has quit IRC16:42
jaegerwhich would cover more than just github, though github is the big offender16:42
teK_only offender?16:42
teK_(for now ...)16:42
jaegerI'm sure there have been other instances where archives coincidentally had the same name outside github16:43
jaegerI guess I'd ask "why would we just do a special case for github when we could proactively solve it for any site?"16:43
teK_that's clearly an advantage of this approach, yes. My approach works transparently16:44
teK_which characters except / are not allowed in filenames?16:46
teK_hm. None as it seems (\0 is not an otpion :P)16:46
teK_windows does not allow : and >, which would be good candidates as separators16:47
jaegerWould probably make sense to use a multiple-character splitter like || or := or similar16:47
jaegerNot a perfect solution but might cut down on edge cases16:48
teK_so we could for something like that: source=(https://github.com/redis/hiredis/archive/v${version}.tar.gz:hiredis-${version}.tar.gz)16:48
teK_I wanted to argue that you cannot download files with : in them to windows anyway..16:48
teK_they still might popup as you can rename them with content disposition :P this thing is insane16:48
tilmancould potentially add a new variable (eval'd by pkgmk) to enable this feature16:48
jaegerI'd probably suggest putting the target name first since URLs can be messy as hell16:49
tilmanfor backwards compatibility :]16:49
jaegerBy default curl and wget will ignore the content-disposition header, at least currently16:49
teK_I hink postfix is more intuitive as in  mv old new. (after all, this is not asm :P)16:50
teK_(intel asm, iirc)16:50
teK_tilman: agreed16:50
jaegerI'm thinking of ugly URLs that could have our split characters in them, not intuitive rating :)16:50
teK_on the other hand.. thins still will break (i.e. not be downloadable for older pkgmk versions)16:50
teK_jaeger: doesnt matter, because we CAN assume (mandate) that $SPLITCHAR is not in the target filename16:51
teK_so its simply URL=${source[$i]%:*} and TARGET=${source[$i]##:*} if splitchar=:16:52
jaegerassumption is generally a bad idea but I suppose it could force the last match if multiples are found, for some reason16:52
teK_but dont take my word on the array syntax :-)16:52
teK_s/assumption/policy/  the target file names are controlled by _us_16:52
jaegerfine, fine, I'll shut up about it16:53
teK_can we bill github for this? Please?16:53
jaegerheh16:53
teK_hehe16:53
teK_charge them for crimes against the digitial humanity and the free folk of common sense16:53
tilmanyou can come up with policy all day long, but if it sucks people will still be annoyed :p16:53
teK_do you think I care? :p16:54
tilmanthough in this case it should be ok16:54
teK_I think so, too.16:54
teK_we have 47 ports in opt and contrib using github in source().16:56
tilmanAchievement unlocked: using grep on /usr/ports16:57
tilmangnihihi16:57
tilmanbtw, using a new variable to enable this feature would also fix the problem where you try to use the rename thing with an old pkgmk. not sure you mentioned that earlier16:58
joacimjaeger: there's a reason behind that ass in assumption16:58
*** lounge has joined #crux16:58
teK_tilman: being a CRUX dev in 2016 also unlocked. gnihihihi16:58
tilman:[16:58
teK_how, tilman? It will see a url like http://github.com/rules/maybe/not/a.tar.gz:b.tar.gz and hit a 40416:59
tilmanbrainfart16:59
tilmannvm16:59
teK_frinnst17:01
teK_honey17:01
*** Na3iL has quit IRC17:03
rmullAnyone here tried the chat software 'tox' ?17:08
*** blueness has quit IRC17:17
*** blueness has joined #crux17:20
frinnstasdf17:22
teK_we need to distribute keys and finish up the signing stuff17:23
teK_I event comitted my patches, cool.17:24
teK_who's up for updating the man page accordingly? :x17:24
frinnstjue, jaeger and I already have the core key iirc17:25
teK_great. one done, two more to go17:25
frinnstim a bit distracted today sorry17:25
teK_sure :)17:25
frinnstafk again17:26
teK_answer my query NAO17:26
frinnstasdfd17:27
tilmansigned ports are happening?17:27
teK_sure17:27
tilmanholy crap17:27
teK_patch is comitted. Stuff is tested. Only missing thing is proper docu17:27
*** blueness has quit IRC17:27
*** ileach has joined #crux17:27
teK_https://crux.nu/gitweb/?p=tools/pkgutils.git;a=summary FYI17:27
*** blueness has joined #crux17:30
jueif no one else volunteers I can try to update the man-pages in about two weeks but I would strongly prefer someone with better english skills17:34
cruxbot[core.git/3.2]: libmpfr: update to 3.1.4-p417:36
teK_maybe I can get around it earlier. What I am always willing to do, is proof reading if you are concerned with the wording17:37
*** fengshaun has quit IRC18:04
*** fengshaun has joined #crux18:06
*** heroux has quit IRC18:39
*** heroux has joined #crux18:48
*** lounge has left #crux ("WeeChat 1.4")20:22
*** cippp has joined #crux20:44
*** jue has quit IRC20:47
*** ileach has quit IRC21:18
*** onodera has joined #crux21:49
cruxbot[contrib.git/3.2]: redis: updated redis.conf22:43
cruxbot[contrib.git/3.2]: redis: update .md5sum22:44
pedjaavidemux has an..interesting...build system23:24
teK_hey its fun when upstream switches build systems23:30
teK_'23:30
pedjaI can only build it as a regular user, fakeroot breaks it, root build breaks it.fun, fun, fun.23:46
pedjaI wouldn't have bothered, but I want to combine 18 videos into one (Faith No More concert)23:49
pedjawill it choke on webm?we will see :)23:49

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!