IRC Logs for #crux-devel Thursday, 2013-05-09

*** __mavrick61 has quit IRC02:48
*** __mavrick61 has joined #crux-devel02:49
juefrinnst: whois07:23
juefrinnst: any yeah, as he left the project he said that he will continue to pay and own the domain07:24
juebtw, good morning ;)07:25
jueprologic: sorry, but -1 for the idea to host subdomains of crux.nu on external host from me07:38
jueprologic: what's the idea behind packages?07:38
prologicjue, -1 for what reason?08:16
prologicre packages.crux.nu - it would be a good idea I think going forward (at least for me - others might find it useful too) to host "clean" packages for i686/x86_64/arm/etc08:16
juebecause we do not have controll over the subdomains if they are hosted external08:17
prologicI was thinking of building a simple web application where folks can "publish" built packages for a specific architecture/version with enough metadata to be useful by a tool that consumes it (e.g: a pkg-get rewrite/port/improvement)08:17
jueofficially we do not provided binary packages besides our iso08:18
prologicof course08:18
prologicI wouldn't expect any of you to support it :)08:18
prologicI'm going to build the system08:18
prologicI'll end up building one anyway08:18
prologicso I'm just thinking of ways of sharing it08:18
juesure, feel free to do that08:18
jueif you provide binaries you need strict dependency checking, it's impossible to have optional deps, that's the main reason why I doubt that it is a good idea08:22
juebut well, let's see how it will work for you ;)08:23
Romsteri have an idea on how to solve the optional deps issue. but it's still in planning stage.08:25
juewithout metadata withing the packages the whole idea of having binary packages will end in something like pkg-get08:28
juebut why do we need it?08:29
Romsteri dunno i was just gonna toy with binary packages for those with low end systems as a fun project i got a decent distcc setup now.08:41
Romsterplus it'll double up in testing which packages break on what ports.08:42
Romsteri'm planning on recording the md5sums of each built package then make a metadata file for each possible build that changes the built package. record that keep every different built package as a md5sum directory, use a tool to fetch the required packages based on the users installed packages.08:44
Romsternow that i've said that i'm sure prologic will beat me to it -_-08:44
juebut keep in mind that without pkg-get you cannot run pre/post stuff08:44
Romsteroh i nearly overlooked that bit... thanks for reminding me.08:45
Romsterkeep them files and pkg-get can run those before and after pkgadd as per normal08:45
Romsteri'd never ask for it to be official, just an idea i'm gonna toy with.08:46
jueyeah, sure, but it's not so easy as it might look at the first sight ;)08:47
Romsteroh on another note i've brought up a i686 core prots to 3.0 standards off your modified ports out of git and a few changes myself. i've only offered the built packages of core though08:47
Romsteroh everything looks easy until you try to do it.08:48
Romsteri'll find something i over looked08:48
Romsteri could do a i586 set but dunno how many are on that hardware now.08:48
Romsteri do have one i586 machine here doing nothing.08:49
Romsterhaven't decided if i wanna toss it or use it.08:49
Romsterk6-II 450MHz 386MB ram.08:49
teK_toiss it ;)08:54
Romsteri was thinking that.08:56
Romsterpretty old.08:56
Romsterthen again ig ot some 486's i'm usign one for a real dos machine, perhaps the i586 can be a backup if those 486's die.08:56
prologicjue, I completely agree - optional dependencies do not work and I  likely will not support such packages09:01
teK_toss the crap Romster09:01
Romsteri'm such a packrat..... salavages the ram and psu and probably toss the rest.09:01
teK_hahaha09:02
prologicjue, my plan is to build something as simple as pkg-get with it's sync09:02
teK_get more efficient man :)09:02
prologicbut as a web service09:02
prologicI'll build upon the same ideas os pkg-get09:02
Romsterfunny thing is i got a 68x cdr in that thing. no joke it says 68x but i doubt it reads/burns any faster than a 48x does.09:02
Romsterbut damn it can spin up loud.09:02
RomsterteK_, i'd keep the ram for someone else pc, ebay it or something09:04
Romsterand the psu will be handy for my 48609:04
prologicjue, the main two features I want to build upon pkg-get are09:04
prologica) publishing "clean" binary packages from others09:04
prologicb) no depending on perl - I was planning on rewriting/porting it to Python and building it as a portable executable either via shedskin or cython09:05
teK_romster toss all the crap09:26
prologicYeah Romster loose the junk :)10:00
prologicstop hoarding :)10:00
prologicworthless electrically inefficient slow old crap :)10:01
prologicspeaking of which my desktop is a POS10:01
jaegerRomster: reminds me of a kenwood 72x that destroyed discs :)13:39
jaegeryay, mate 1.6 moved away from gconf13:41
Romsterthak god and what too dconf?13:48
jaegeryeah, dconf. which is somewhat better13:49
Romstera 72x kenwood... thought i had the fastest rotational speed cd-rw drive.13:49
jaegerI don't think they made that one for very long since it was actually dangerous13:50
jaegerI saw a video of little shards of CD coming out of one of them13:51
jaegermaybe it was fake, though, don't know13:51
Romsteryeah the paper work said never to use any disc with any defects or cracks in a cd, for risk of exploding into shards13:52
jaegerjue: Looks like nfs-utils might have changed upstream, new md5sum14:51
jaegerRomster: the fastest one I can recall ever personally owning was 48x and it never actually got to 48x in use... more like 24-28x max14:53
jaegereven if I wrote to the very end of the usable space14:53
juejaeger: thanks, the old tarball was created with the wrong version string, 1.2.7 instead 1.2.8, that's fixed now15:24
jaegerAh, ok15:26

Generated by irclog2html.py 2.11.0 by Marius Gedminas - find it at mg.pov.lt!