IRC Logs for #crux-devel Sunday, 2014-06-15

*** Amnesia has quit IRC00:28
*** Amnesia has joined #crux-devel00:40
jaegerdid clang get put back in for llvm 3.4.1?01:04
jaegerIt looks that way, just wondered if I missed something01:11
Romsteri wondered that myself05:14
jaegergot some updates for llvm-32, mesa3d-32, and some deps that I'll push tomorrow. Going to sleep soon05:32
jaegerpedja: I'll push the nvidia patch at the same time, doesn't seem to have any adverse affects in my tests today05:33
Romsteri tried getting llvm-32 fixed up but i kept failing so i left it.05:33
jaegerbuild issues or something else? I was able to build it and mesa3d-32 and wine but only tested a few things05:34
Romsterjust build issues when bumping versions05:34
jaegerok05:34
Romsterthe whole llvm with 32bit confuses me...05:34
Romsterllvm by default has x86 and x86_64 as it's a vritual machine from byte code to machine code...05:35
Romsterand clang rust what have you just does the pre-processing stuff.05:35
Romstercan't see why llvm cna't be multilib05:35
Romsterlike gcc is05:36
jaegerno idea, I didn't look into it deeply. I don't use it directly so I just care if it works without problems, not why :)05:36
Romsteronly thing i can really make of it is it need the 32bit libs. that llvm-32 provides.05:36
Romsteryeah it works so its fine. but i still am puzzled why.05:38
jaegerhrmm, broken symlink into the work dir. I'll fix that in the morning, too tired now05:41
Romsterk05:48
*** rexich has joined #crux-devel08:43
*** heroux has quit IRC09:10
*** heroux has joined #crux-devel09:11
pedjajaeger: ok.10:08
*** rexich has quit IRC10:42
*** Amnesia has quit IRC10:50
*** rexich has joined #crux-devel10:56
*** rexich has left #crux-devel ()10:58
*** rexich has joined #crux-devel11:00
*** rexich has quit IRC11:01
*** rexich has joined #crux-devel11:01
*** Amnesia has joined #crux-devel11:03
*** rexich has quit IRC14:32
juejaeger: would be nice to have a "CONFIG_CRYPTO_XTS=m" in our default config cause it's needed by cryptsetup16:16
jaegeris module sufficient?16:17
jueyes, at least for me16:17
jaegerok, I'll add it16:17
juethx, it's works fine if you have only a crypted /home or something like that16:19
frinnsti didnt bump dhcpcd since i dont use it16:20
jaegerno problem16:20
jueat all I'm very pleased with 3.1, it's very stable and hassle-free for me :)16:20
frinnstyeah. the only thing i suspect we will have questions regarding is the interface names16:21
jaegerIt's been good for me so far, too. I'd definitely like to do another ISO build sometime before release, though, since things keep changing (mesa3d, firefox, etc.)16:21
frinnstmesa3d was reverted though :)16:23
frinnsti filed a bug regarding the autodetection, and its been fixed16:23
juefrinnst: maybe we should document the 'touch /etc/udev/rules.d/80-net-name-slot.rules' trick somewhere in the handbook16:23
jaegerOr just link to the docs so that people have to actually think a bit :)16:24
frinnstyes, or atleast really point it out in the release announcement16:24
frinnstand explain how the naming works16:24
jue+116:25
frinnstim not really sure how the "pN" works16:25
frinnstsorry, "sN"16:25
frinnstenpX is quite obvious once you look at lspci outputs16:25
jaegerhttp://cgit.freedesktop.org/systemd/systemd/tree/src/udev/udev-builtin-net_id.c#n2016:25
jaegerit's commented there16:25
juefrinnst: dhcpcd works fine for me, will commit it later16:25
jaegerand that's linked from http://www.freedesktop.org/wiki/Software/systemd/PredictableNetworkInterfaceNames/16:26
frinnstyeah. including that url in the announcement should do the trick16:27
jaegerSomething like: "please read this document to understand network device names in 3.1"16:27
frinnstnext week is my final week at work before my 2 week vacation, so i'll have extra time16:28
frinnst\o/16:28
jaegerI'm going to be out of town from the 18th to 25th for what that's worth16:28
juejaeger: are you able to do another rc before that?16:29
jaegerI think so. I'll start a build after I finish up my updates for llvm-32, etc.16:30
jueok, will commit dhcpcd, kmod and man-pages soon16:31
jaegerHave we made an official decision regarding clang in llvm?16:31
jueyes, I think so -> build llvm without it16:36
jaegerI was thinking that was what we decided but noticed llvm 3.4.1 has clang in it16:37
jaegerthough now that I'm looking more closely it seems that it's removed in 3.1 but left in 3.016:37
jueoops16:37
jueyeah, that was a 3.1-only change16:37
jaegerok, that's fine for me16:37
jaegerAnyone have any changes pending that will result in long compile times for things that are on the ISO?16:39
jueno, just committed everything important16:45
jaegerok16:52
Romsteri'd make that change you suggested to rc.d/net jaeger ?17:10
Romsterand also comment that "ip a" is your friend to find what the interface name is.17:10
Romsteroh and the contrib.rsync file needs a sed from 3.0 to 3.117:11
Romsteri know it's contrib.rsync.inactive on the iso...17:11
Romsterbut those that have contrib.rsync wont have it on 3.117:11
jaegerIt'd probably be slightly more future-proof to do something like "if contrib is enabled, copy the new inactive file over the old active one" instead of sed... just in case hostnames or paths change in the future17:29
jaegerI guess it doesn't really matter since we have control over the ISO anyway17:30
jaegerok, started a bootstrap. also going to push a few compat-32 updates today19:17

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