IRC Logs for #crux-devel Wednesday, 2014-06-11

jaegerfrinnst: nspr in 3.1 looks to be 4.10.5 and firefox wants 4.10.601:16
*** c0x has quit IRC01:19
*** mike_k has joined #crux-devel05:30
frinnstI am still having trouble with mesa3d 10.2.1. vdpau autodetection seems broken06:10
frinnstconfigure: WARNING: unrecognized options: --with-llvm-shared-libs07:24
frinnsti'll dig around in it a bit today i think07:24
juefrinnst: from the release notes: Renamed --with-llvm-shared-libs to --enable-llvm-shared-libs ;)07:33
teK--frinnst: how did you like the last episode of GoT?07:36
teK--you have to get the accent raiight, you cuuunt07:41
teK--I was a little sad, though :\07:41
frinnstyeah but nobody's safe07:42
frinnstmaybe we should be careful with the spoilers :)07:42
teK--I was sad for all the men and women that died in that episode, you heartless creature!07:44
frinnstmesa no longer seems to depend on xorg-libxvmc. and fails to autodetect vdpau :(07:50
frinnstlibxvmc is also default=auto, maybe it also fails to detect08:01
Amnesiaff 30.0 fucks up your fontsize..08:07
Amnesiaso how's crux 3.1 progressing:)?08:11
Romster3.1 works fine here but i have not tried it on my desktop yet08:29
juefrinnst: you know that 10.2.1 is still a development release?08:43
frinnstuh, no? :)08:44
frinnstah fuck08:44
frinnstfrinnst strikes again!08:44
frinnstshoot first, ask questions later08:44
frinnstso, revert then i guess08:45
frinnstmore trouble than it's worth anyways08:45
jueyeah, it is. Normaly x.x.1 is the first stable release, but this time it fixes only a build problem, so 10.2.2 should be the first stable08:46
frinnstso, lets revert for now then08:46
jueno objection from my side08:47
frinnsti've filed a bug regarding vdpau autodetection during build. hopefully they can fix it for the next release08:48
Romsterwhy would they do that and not have it as a 10.3.x odd number people08:48
Romsteror even do 10.1.90 and 10.1.91 92 93 94 95 etc before releasing 10.2.0 like xorg does.08:49
Romsterman why do projects not keep to a version convention08:49
frinnsti'll push the flash security update11:01
Romsterthere is always a security issue in flash, regardless.11:06
*** rexich has joined #crux-devel13:16
*** mike_k_ has joined #crux-devel13:29
*** mike_k has quit IRC13:29
*** sammi` has joined #crux-devel13:48
*** rexich has quit IRC14:15
*** erdic has joined #crux-devel18:57
jaegerAnyone recommend a decent online C programming course? I have a couple books but I'm curious if there's a good online one as well.19:45
*** mike_k_ has quit IRC19:58
*** frinnst has quit IRC20:20
*** frinnst has joined #crux-devel20:22
*** frinnst has quit IRC20:22
*** frinnst has joined #crux-devel20:22
teK--maybe coursera can help you20:24
jaegerthanks, I'll take a look at both of those20:31
teK--I really think that both practicing and the classics from K&R and the Programming in UNIX environments will help you go pro ;)20:32
teK--to me C still is the most beautiful language out there; I don't know why but coding C makes me feel best, compared with Java, JS, Python etc.20:32
jaegerPython is what I prefer to quick work at work but I've been wanting to use C for something more involved. Since I've been wanting to tweak/add some functionality to pkgutils and prt-get I figured I'd at least give it a shot there.20:34
jaegerI have some experience with C over the years but nothing big20:34
*** novak has quit IRC20:39
teK--I always found that C++ was radically different, even bad (I dislike it) compared to C..20:40
teK--and prt-get is written in C++20:41
*** novak has joined #crux-devel20:41
jaegerI'm not a big fan of it, myself20:41
jaegerMy plan was to attempt to implement a new tool that combines functionality of pkgutils and prt-get in C20:42
jaegersome work was already done in that vein with pkgutils6 but it's abandoned as far as I know20:42
teK--I'd volunteer to do that with you (in C)20:45
teK--I'd love to replace the _zoo_ of tools by at least one package..20:45
jaegerok, will keep that in mind :) I'm going to be out of town for a week starting next wednesday but I'll try to get started with something soon20:46
teK--I'll be somewhat low on time until mid July because of (final!) exams and job interviews but I'd be happy to talk through ideas etc.20:54
jaegerat the very least I think it would be nice to have:20:54
jaeger1) all of pkgutils' current functionality20:54
jaeger2) all of prt-get's current functionality20:54
jaeger3) notification/prompting for newly added deps when doing sysup20:54
teK--I have to do pkginfo -l pkgutil/prt-get to judge your statement first :o20:54
teK--pkgutil != prt-utils20:55
jaegerpossibly also a new package database format with some metadata in it20:55
teK--if it's justified + backwards compatibility is kept: why not20:55
teK--imho we should support sepen's mpup to include single ports from foreign repositories20:56
teK--and a nice interface for adding things 'automatically'20:56
jaegerthat starts crossing into ports/*up territory20:56
jaegerwhich might be nice but I'm not sure should go into a pkgutils rewrite20:57
teK--like:  prt-get portadd rsync://xxxx.yyy:..20:57
jaegerAh, for manipulating prt-get.conf style options?20:57
teK--and prt-get portdbsearch ..20:57
teK--yes but somehow it would be nice no having to pull single ports into mpup.conf by hand, either20:57
jaegerperhaps something to add for later20:58
teK--and yes that's ports(8) business, somewhat20:58
jaegerI definitely want to focus on pkgutils/prt-get first20:58
teK--I talked to some folks in #crux and they proposed a more community-driven portdb, too. Which I kinda liked as it's rather complicated to add repositories/ports, inspect port versions and report (via flyspray) outdated packages20:59
teK--yes, focusing helps getting things done ;)20:59
jaegerindeed :)20:59
jaegerI've got no objections to the portdb work they're doing so far... will be interesting to see if it balloons or goes off track21:00
teK--I still don't see me installing go on the current installation :D21:00
teK--which is a matter of its own21:00
teK--I don't dare updating that kernel, for example21:00
teK--s/updating/\1 and rebooting/21:01 is an odd beast at the moment, since it's not 64-bit hardware21:02
teK--hehe yes21:02
teK--at least user facing services are really really uptodate21:02
teK--* mostly21:02
jaegerI actually have a nice server I'd be happy to donate to replace the current hardware but I think shipping it to sweden could be expensive :)21:04
teK--I am far more concerned about administrative delays and gaps21:05
teK---rw-r--r-- 1 tek  users 201M [2014/06/11] Spaceballs.mkv21:05
teK---rw-r--r-- 1 tek  users 707M [2005/08/17] Spaceballs.mpg21:05
teK--by my son's demand :P21:06
teK--took this laptop <10 minutes to convert21:06
teK--he's watching it for the 8th or 9th time now21:06
teK--anyway.. sooner or later we will have to figure things out for our box21:07
teK--a friend of mine works at hosting company in Nuernberg so maybe we could go that route, too21:09
jaegerAnother thing I'd probably do is get rid of "depinst" and just make "install" do what depinst does now21:28
teK--yes, one could always provide a command to install the given port only21:30
teK--but depinst should be default, depupdate configurable21:31
jaegerAre there any ports you can think of off the top of your head where you'd prefer install to depinst?21:31
teK--but not all ports are as acurate as core/opt/contrib wrt dependencies21:32
jaegerI've also considered lightening up on the dependency checking and strict footprint adherence but that's something we'll have to test and play with to see if it's good or bad21:32
teK--and have the normal install won't hurt, just think of a port already built etc.. why check (if you remember as a user) if all deps were installed if you simply removed it 2 minutes ago21:32
teK--what aspect of footprints?21:33
jaegerright now unless you use -if new files cause a failure. that sort of thing. The user might prefer to have it go ahead and install those but just store that info to be printed at the end of the process or something like that21:33
jaegerit's subjective, just something that's crossed me mind while thinking about this21:34
teK--I'd also want to make verbosity configurable21:35
teK--I usually have pkgmk run build >/dev/null21:35
teK--and only enable it if something with my ports goes wrong during updates21:36
teK--I see quite some warnings from configure and make, though :D21:36
teK--parallel downloads in the background would be fancy, too21:40
*** deus_ex has joined #crux-devel22:06
*** pitillo has quit IRC22:15
*** pitillo has joined #crux-devel22:15

Generated by 2.11.0 by Marius Gedminas - find it at!