IRC Logs for #crux Monday, 2011-08-29

pitillogood morning01:09
*** lasso|qt has joined #crux02:02
*** vee has joined #crux02:13
veeromster, wait wut lol02:14
*** lasso|qt has quit IRC02:27
*** vee has quit IRC02:27
*** lasso|qt has joined #crux02:52
cruxbot[opt.git/2.7]: dnsmasq: update to 2.5803:05
cruxbot[opt.git/2.7]: dovecot: update to 2.0.1403:05
cruxbot[opt.git/2.7]: [notify] masqmail: update to 0.2.3003:05
cruxbot[opt.git/2.7]: squid: update to 3.1.1503:05
*** s_warlock has joined #crux03:09
*** s_warlock has joined #crux03:11
*** s_warlock has quit IRC03:33
*** s_warlock has joined #crux03:39
*** s_warlock has quit IRC03:53
*** aubic has joined #crux05:37
*** slashbeast has quit IRC07:45
*** slashbeast has joined #crux07:51
teK_tilman / jue  is this relevant to us: ?08:23
Romsteri couldn't care less about Oracle i'll keep using it until Open JDK is fixed. also OpenJDK hasn't got all the extensions that the Oracle JDK/JRE has.08:33
teK_it's a lawyer thing so we may have to care.08:34
Romsterremove it or leave it at 26, i'll just keep my own copy and install it from there so it's not in the distro then.08:36
jaegerIt'd probably be safe to keep a port of it if the port requires you to agree to the EULA08:37
jaegerdon't know for sure, though08:37
Romsteralso mpd and libraries need a version bump teK_08:37
rmullI keep getting internal compiler error (segfault) with gcc 4.5.1 and an errorless memtest. Does anyone have any tips? My cpu is a Via Nano. I can't compile a new gcc or icu because of this error, but most other packages work fine.08:52
jaegerI'm not familiar with via nano but I remember some older via c3 or the like CPUs that wouldn't run things compiled with -march=i686 properly08:57
jaegerI wonder if it's something like that08:57
rmullHmm, perhaps. I'm using x86_64, but I had problems like this when compiling some things on a i686 gentoo installation, same platform08:58
rmullThe Via Nano isn't recognized by gcc march=native yet08:58
rmullBut I'm not sure if that's the issue.08:58
Romsteryeah use gcc 4.5.308:59
Romster4.5.1 is buggy08:59
teK_what about 4.6.1?08:59
Romsteri'm on 4.6.1 but it does break a few things you need to alter some CFLAGS/CXXFLAGS09:00
Romsteras it's more strict.09:00
Romsterjsut make a cross compile tool chain to make the new via nano tool chain?09:01
rmullRomster: Okay, I'll do that. I have a spare machine I can use to build the tools.09:03
Romsterthat way you don't break anything else with the bugs in that tool chain.09:04
Romsteryou might make new bugs though can't guarantee you wont make some subtle bug. but small price to pay for a really broken tool chain.09:05
Romsterthough sometimes if you got dodgy compiler flags you can get some strange bugs compiling too.09:06
rmull4.5.1 is the gcc that is distributed with crux 2.7, so that's not good if it's buggy09:06
jaegerI built a cross-compile toolchain this weekend for i686 -> x86_64 multilib, what a pain09:06
Romsteri had some bug with -O2 but wasn't there in -O009:07
rmulljaeger: Any writeup?09:07
Romsteryeah they arn't fun to build.09:07
jaegerrmull: not yet, though I did upload a tarball09:07
Romsterrmull, if you sysup you'll find it's now at 4.5.309:07
Romster4,5,1 was available when 2.7 was released.09:08
rmullYeah, but I can't build it because my 4.5.1 doesn't work :P09:08
Romsteractaully there is another idea.09:08
rmulljaeger: cool09:08
jaegerrmull: I can extract the newest package from the updated ISO if you want it09:08
Romstergrab jaeger's iso and pkgadd -u all of core ports.09:08
jaegeror that :)09:08
Romsterrebuild kernel reboot.09:08
rmullOh, that would be easy. Link to ISO?09:08
jaegerthat would make more sense since more than gcc has been updated09:09
jaeger or rsync://
rmullThanks so much. I'll try this out tonight.09:09
Romsterbe less painful than dong a tool chain.09:10
rmullFor sure09:10
Romsterjaeger, did you write up on how you did that tool chain. still curious i got so far with mine and i jsut haven't been motivated to complete it yet.09:11
jaegernot yet, I was thinking about making up some scripts for it though. I did upload the toolchain itself if you want to play with it09:12
Romsteri may do. but i really am keen on how to build one. i got stage 1 done then i just haven't got to doing stage 2.09:13
* Romster likes to tinker and learn.09:13
Romsterunless you included the scripts in that tarball?09:14
jaegerit's just the built toolchain09:16
*** aarchvile has joined #crux09:17
jueteK_: no sure, but as long we do not provide a package it should be save to keep the port09:37
jueteK_: let's wait for the next release (if any) to see if a download without user-interaction still works09:39
jueteK_: anyway, thanks for the hint09:40
Romsters/save/safe :)09:59
jueoops, thx Romster :)10:02
Romsterno problem. seems to be a common type for non-native English speakers.10:05
Romstererr typo... here i go again..10:05
cruxbot[core.git/2.7]: util-linux-ng: update to 2.2010:45
cruxbot[core-x86_64.git/2.7]: util-linux-ng: updated to 2.2011:07
*** Rotwang has joined #crux11:10
*** Evil_Bob has joined #crux11:27
*** lasso|qt has quit IRC12:04
*** vee has joined #crux12:20
veemornin peeps12:21
veei see what you did there12:21
tilmangreeting you?12:21
veenever mind.... :(12:21
veeblahr, college starts today blahr12:22
Rotwanghi ppl13:12
Rotwangso we are writing build system for our component in gnu make13:13
RotwangI told these guys it is going to be ugly as fuck13:14
Rotwangbut no one wants to listen to poor Rotwang13:14
tilmanthats because you're an unexperienced noob13:14
Rotwangwhat hurts me the most is that most likely I'm going to be maintaining this stuff later13:14
jaegerso change it to something else yourself :)13:15
RotwangI probably will13:16
Rotwanganyway, creating buildsystem in pure make is going to end in disaster, always13:18
Rotwangit is like writing web browser in asm13:18
tilmanRotwang: check out the sources for Quagmire ( for some neat _gnu_ make tricks13:42
RotwangI see13:43
*** vee has quit IRC14:02
*** seanw has quit IRC14:11
*** seanw has joined #crux14:19
*** vee has joined #crux14:31
*** vee has quit IRC15:02
*** Evil_Bob has quit IRC16:06
*** seanw has quit IRC16:17
*** Rotwang has quit IRC16:21
*** y3llow has quit IRC16:21
*** y3llow has joined #crux16:21
*** seanw has joined #crux16:24
*** vee has joined #crux17:46
*** j^2 has quit IRC18:09
*** Ovim-Obscurum has joined #crux18:18
veelibfm package is also failing to build O.o18:19
*** Ovim|afk_nA has quit IRC18:21
thrice`don't give too many details :>18:29
veelol im known not to18:48
*** aarchvile has quit IRC19:19
rmulljaeger: I'm getting build failures on opt/gnutls. x86_64, gcc 4.5.3. Errors are like: asm/appro-aes-x86-64:2021: Error: bad register name '%rdx' and it goes on to list all the CPU registers with similar error message, different line number19:30
rmullAs the maintainer, do you have any thoughts?19:30
rmullI have CFLAGS="-m64" for that19:30
rmullI haven't spen much time looking itno it but I'll do so now19:32
jaegersorry, no idea on that. I'm not the maintainer19:38
jaegerlooks like I was the original packager19:44
rmullNo worries, I'll sort it out19:48
*** lasso|qt has joined #crux20:01
*** lasso|qt has quit IRC20:08
rmulljaeger: Correct me if I'm wrong, but it seems like the gcc in your 2.7 upgrade ISO creates 32-bit executables20:10
rmullIs that expected?20:10
rmullOh, wha the hell, it's not a x86_64 iso. I'm sorry.20:37
rmullI guess I'm back to my original problem of not having an x86_64 4.5.3 gcc20:37
*** Dudde has quit IRC21:38
*** Dudde has joined #crux21:39
jaegerrmull: I'm sorry, I didn't realize you were looking for x86_64 =(22:19
jaegerNo idea if frinnst has an updated version of that22:19
jaegerI could build gcc and upload it somewhere if you like, on my x86_64 install22:20
jaegerI should have realized you didn't want 32-bit since you were asking about x86_64 gnutls, it didn't click for some reason22:22
jaegerwell, it's building. I'll upload it when it's done and you can grab it if you like22:58

Generated by 2.11.0 by Marius Gedminas - find it at!