IRC Logs for #crux-devel Thursday, 2014-03-27

*** frinnst has joined #crux-devel01:49
*** Workster has quit IRC03:50
*** Workster has joined #crux-devel03:53
*** Workster has quit IRC03:53
*** Workster has joined #crux-devel03:53
*** Workster has quit IRC06:31
*** Workster has joined #crux-devel08:28
*** Workster has quit IRC08:28
*** Workster has joined #crux-devel08:28
*** heroux has quit IRC09:10
*** heroux has joined #crux-devel09:28
*** heroux has quit IRC10:18
*** heroux has joined #crux-devel10:19
*** heroux has quit IRC10:47
*** heroux has joined #crux-devel15:04
*** heroux has quit IRC15:27
*** heroux has joined #crux-devel15:34
*** heroux has quit IRC16:14
*** heroux has joined #crux-devel16:21
jueRomster: finally fixed ruby on 3.1, sorry for the delay ;)17:32
*** jue has joined #crux-devel17:45
juefrinnst: no, never tried the betas18:28
juefrinnst: probably better to stick with our current version for 3.1?18:28
jaegerprobably so18:46
jaegerjue: what's your opinion on removing tcp_wrappers in the future?18:46
jaegeranyone run into icu failing in 3.1 with an undeclared identifier '__float128'?19:06
jaegersearching seems to point to clang not supporting __float128, wonder if it's using clang since it's installed19:08
jaegerah, that seems promising. forced gcc/g++ and it built19:08
jaegerDid anyone have any comments/questions about my suggestion to remove clang from llvm? I don't think we should ship it19:09
juejaeger: +1 for removing clang, it's easy to build it in a separate port and we don't need it for mesa19:28
jaegerthat's my thought as well19:28
jaegerand so far it causes an issue or two :)19:28
jaegeralso, looks like Romster kept that in the llvm-32 port from my patch so we have a bit of a mismatch between llvm and llvm-32 in 3.119:29
jueand building it on older boxes is at least very time consuming19:33
jaegertrue, heh19:33
juejaeger: wrt libwrap, I think it's still useful in LANs, but I know many people run in problem with it19:33
juebut at all im 50/50, iit's not a problem for me to remove it19:34
jaegerPersonally I'd rather just use iptables host-based firewalls. My complaint about tcp_wrappers isn't really about having trouble with it but how it's a bit inconsistent in terms of what invocation solves the issue at hand19:34
jaegernot a strong objection, though. If the majority of people want it  to stay it's no problem19:36
juesure, iptables is the way to go for greater things, but to protect only one service a simple host.allow entry is IMO the easiest way19:36
juejaeger: what are others like fedora or debian is doing here?19:37
jaegergood question, I'll look around19:38
jaegermy ubuntu 12.04 workstation here at work has libwrap installed but doesn't seem to actually link it19:39
jaegerlibnetsnmp* seems to be the only stuff on this system linking it19:40
*** rexich has joined #crux-devel19:43
jaegerI'll install a fedora VM to give that a look as well19:43
juegreat, thanks19:45
juejaeger: I've to admit that I didn't found the time to test your test-release yet, but, promised, I'll do it over the weekend ;)19:46
jaegerNo problem, I don't really feel like we're rushed19:46
jaegerespecially since I'd like to remove clang before the 3.1 release if nobody objects19:47
*** rexich has quit IRC20:01
frinnstI dont really feel like tcpwrappers adds anything. If i want security i'd use iptables20:22
frinnstand yeah, clang seems pointless20:22
jaegerfedora has a lot of stuff linked against tcp_wrappers but the hosts.allow and hosts.deny files are empty aside from comments21:38
jaegermaybe the better solution is just to ship a commented blank hosts.deny21:46
Workstercan someone test llvm 3.4 agenst icu please and +1 remove clang23:22
Worksterand make sure mesa3d is happy with it some options are now default and don't need to be secifyed in llvm 3.4 like --enable-experimental-targets=R600 is now enabled in 3.423:23
Worksterbbl copier to work on23:23

Generated by 2.11.0 by Marius Gedminas - find it at!