2014-05-01 10:31:45 hi 2014-05-01 12:03:04 how can I check, which files a package does install? 2014-05-01 12:22:38 so guys, when are we switching o libressl? 2014-05-01 13:17:31 barthalion: did they make official release yet? 2014-05-01 13:17:47 shafire: apk info -L package 2014-05-01 13:18:27 apk info -L bridge-utils bridge-utils-1.5-r1 contains: 2014-05-01 13:18:30 it's empty? 2014-05-01 13:18:43 is the package installed? 2014-05-01 13:18:46 yes 2014-05-01 13:18:57 other packages show files 2014-05-01 13:19:02 yes, then i think its empty 2014-05-01 13:19:26 heh 2014-05-01 13:19:27 yes 2014-05-01 13:19:32 looks like its empty 2014-05-01 13:22:24 that seems wrong.. 2014-05-01 13:23:36 should it not provide brctl? 2014-05-01 13:23:40 I thought... 2014-05-01 13:23:47 i thought so too 2014-05-01 14:05:18 by the way: any nfs news? 2014-05-01 17:54:47 hi 2014-05-01 17:54:54 does alpine has a dhcp server? 2014-05-01 18:45:46 several 2014-05-01 18:45:59 shafire: dnsmasq, dhcpd are two 2014-05-02 07:41:36 good morning 2014-05-02 07:41:43 ncopa: are you around? 2014-05-02 07:56:58 hi 2014-05-02 07:57:08 clandmeter: pong 2014-05-02 07:57:11 hi 2014-05-02 07:57:45 can we add mongodb again wihtout v8 support? 2014-05-02 07:58:01 i see in the commit messages its a pile of sh*t :) 2014-05-02 08:00:15 ncopa: the build currently breaks on http://git.alpinelinux.org/cgit/aports/tree/testing/mongodb/uclibc-libc-version.patch 2014-05-02 08:01:16 and musl doesnt have gnu/libc-version.h (like uclibc) 2014-05-02 08:04:26 hm 2014-05-02 08:04:51 you can do #define gnu_get_libc_version(x) "musl libc" 2014-05-02 08:04:53 or simlar 2014-05-02 08:04:57 unconditionally 2014-05-02 08:05:04 but its a good question really 2014-05-02 08:05:26 we should ask in #musl 2014-05-02 08:09:42 or you could do: "unknown" 2014-05-02 08:12:41 ncopa: https://github.com/mongodb/mongo/blob/master/src/mongo/util/processinfo_linux2.cpp 2014-05-02 08:12:46 thats where it breaks 2014-05-02 08:14:46 we should add support for /etc/system-release or what its called 2014-05-02 08:15:40 what do i need to use to check for musl? __MUSL__? 2014-05-02 08:15:50 thats the problem you cant 2014-05-02 08:15:55 there is no way to check it 2014-05-02 08:16:08 and its by design 2014-05-02 08:16:49 http://wiki.musl-libc.org/wiki/FAQ#Q:_why_is_there_no_MUSL_macro_.3F 2014-05-02 08:19:32 sounds like forcing a real fix instead of ducktape 2014-05-02 08:20:01 yes 2014-05-02 08:20:04 but in this case 2014-05-02 08:20:08 it is used for information 2014-05-02 08:20:27 you want know what libc, kernel, OS was the app built for 2014-05-02 08:21:17 http://sprunge.us/cZAe 2014-05-02 08:21:33 you could all the patch alpine-musl.patch or similar 2014-05-02 08:21:40 or just alpine.patch 2014-05-02 08:21:48 i think we should implement /etc/system-release 2014-05-02 08:49:34 ncopa: your patch gives me http://sprunge.us/TgPc 2014-05-02 08:50:15 my typo 2014-05-02 08:50:17 its either 2014-05-02 08:50:22 #ifdef __GLIBC__ 2014-05-02 08:50:23 or 2014-05-02 08:50:28 #if defined (__GLIBC__) 2014-05-02 08:50:41 my bad sorry 2014-05-02 09:00:26 ncopa: http://sprunge.us/DcLP 2014-05-02 09:04:44 dunno 2014-05-02 09:05:42 sounds like something breaks with musl malloc 2014-05-02 09:05:49 or clashes 2014-05-02 09:06:05 do they have a malloc.h or are they talking about the system malloc? 2014-05-02 09:06:23 they check for malloc 2014-05-02 09:07:51 --use-system-tcmalloc=USE-SYSTEM-TCMALLOC 2014-05-02 09:08:07 not sure that part is related. 2014-05-02 09:10:18 i dunno either 2014-05-02 09:10:26 but i didnt add that switch, so i guess it should use something internal. 2014-05-02 09:10:40 but musl's malloc is claimed to be competitive with jemalloc 2014-05-02 09:10:47 so i think its ok to just use it 2014-05-02 10:07:51 musl does not support replacing system malloc 2014-05-02 10:07:56 it will break horribly run time if you try to do it 2014-05-02 10:08:46 i believe musl does not implement mallinfo 2014-05-02 10:08:49 that's why the error 2014-05-02 10:09:12 clandmeter: sorry, I haven't tried 2014-05-02 10:09:16 uclibc does? 2014-05-02 10:09:18 clandmeter: bulding mongo even in Arch is a pita 2014-05-02 10:13:27 barthalion: that pkbuild doesnt look that horrible. 2014-05-02 10:14:13 but it uses tcmalloc from system (which we dont have) 2014-05-02 10:15:31 you know, looping builds, issues with boost rebuilds, c++ driver location changing every release 2014-05-02 10:15:45 up to point it's missing now, because it's theoretically separated 2014-05-02 10:15:53 but in practice doesn't build against newest mongodb 2014-05-02 10:23:36 fabled: is it hard/usefull to add mallinfo? 2014-05-02 10:23:54 or should i just give up on mongodb. 2014-05-02 10:29:19 i wonder what it needs it for 2014-05-02 10:29:23 it's basically just statistics 2014-05-02 10:30:44 fabled: https://github.com/mongodb/mongo/blob/master/src/mongo/util/processinfo_linux2.cpp#L384 2014-05-03 08:58:55 ERROR: libmount-2.24.2-r1: trying to overwrite lib/libmount.so.1.1.0 owned by util-linux-2.24.2-r0. 2014-05-03 08:58:55 ERROR: libmount-2.24.2-r1: trying to overwrite lib/libmount.so.1 owned by util-linux-2.24.2-r0. 2014-05-03 08:58:58 ncopa, ^^^ 2014-05-03 09:02:34 Hi 2014-05-03 09:02:41 How can I use ipv6 dhcp in alpine? 2014-05-03 09:03:16 shafire, i have not done dhcp. kernel has built-in-support for ipv6 stateless auto config, and it works well 2014-05-03 09:03:51 fabled: this one: iface eth0 inet6 manual 2014-05-03 09:03:52 ? 2014-05-03 09:29:05 fabled: mh, I tried Fedora Live Desktop, it works there well with my dhcpv6 server... 2014-05-03 09:37:17 not sure if we have dhcpv6 client 2014-05-03 09:37:25 stateless link auto config is different 2014-05-03 09:55:11 something like this? http://wide-dhcpv6.sourceforge.net/ 2014-05-03 21:47:13 morning 2014-05-03 21:47:42 anyone here used alpine-iso since ncopa's changes to mkinitfs? 2014-05-03 21:48:21 because the alpine-iso repo's Makefile produces a broken symlink for grsec to grsec-vmlinuz (which isn't there) 2014-05-03 22:40:21 o 2014-05-03 22:40:34 are you being put into an emergency shell at boot fbettag? 2014-05-03 22:45:23 Mp5shooter: nah i'm trying to build an iso 2014-05-03 22:45:29 oh 2014-05-03 22:45:53 but i build a workaround, instead of the ln -sf being called, i made a shellskript which fixes the symlink 2014-05-03 22:45:58 not nice but works for now 2014-05-03 22:46:39 tomorrow i'll finish my apkovl http service, which will generate apkovl's depending on the system's purpose or delivering their stored ones (on the fly) 2014-05-03 22:47:52 also probably nobody builds .img files with the alpine-iso scripts, since it'S totally broken 2014-05-03 22:48:22 mcopy -i ... should've been mcopy -si (so you have an actual apk repository on the memory stick as welll as syslinux config) 2014-05-03 22:58:16 88.159.20.184 appears to be down 2014-05-03 23:00:44 (nl.alpinelinux.org) 2014-05-03 23:01:01 mom 2014-05-03 23:01:11 works for me 2014-05-03 23:01:18 but of course we're 4 hops over :P 2014-05-03 23:01:59 2 packets transmitted, 2 received, 0% packet loss, time 1002ms 2014-05-03 23:01:59 rtt min/avg/max/mdev = 7.742/7.812/7.883/0.113 ms 2014-05-05 06:08:09 morning 2014-05-05 06:08:48 fabled: didnt we have a recent apk-tools feature that allows packages overwrite files from others if origin is the same? 2014-05-05 06:09:03 ncopa, yes, it's in apk-tools.git but not in edge yet 2014-05-05 06:09:24 simply running apk fix should resolve your util-linux issue though 2014-05-05 06:09:32 yes 2014-05-05 06:09:39 assuming it's same source package 2014-05-05 06:09:44 i dunno if i should cherrypick the apk tools change or add replaces to util-linux 2014-05-05 06:09:44 ok 2014-05-05 06:09:47 i cherry pick it 2014-05-05 06:09:50 thanks 2014-05-05 06:10:03 we must get the v3.0 builders up asap 2014-05-05 06:10:30 builders for 3.0-stable 2014-05-05 08:24:49 fabled: does that mean, if i split a package, i dont have to use replaces anymore? 2014-05-05 08:25:11 clandmeter, correct. as long as the split is within same APKBUILD 2014-05-05 08:25:33 if you split to two APKBUILDs (e.g. by using upstream split packages), then you'll need replaces still 2014-05-05 08:26:06 ok got it. (and makes sence) 2014-05-05 09:07:06 binutils does not build without bash 2014-05-05 09:07:11 https://sourceware.org/bugzilla/show_bug.cgi?id=16370 2014-05-05 09:07:13 its biting us too 2014-05-05 09:07:25 seems to affect both uclibc and musl libc 2014-05-05 09:07:27 busybox ash 2014-05-05 09:08:13 i cannot reporoduce it on fedora though... 2014-05-05 10:27:48 seems like there are a lot of pinholes to fix... 2014-05-05 11:13:34 yeah 2014-05-06 04:50:15 ncopa, ping 2014-05-06 04:50:44 so, I am trying very hard to set up alpine 2.7.6 2014-05-06 04:50:53 and it is failing just as hard... 2014-05-06 04:51:09 `apk update` on any of the default mirrors says "BAD signature" 2014-05-06 04:51:20 I finally found that mirrors.centarra.com will at least pass the signature check 2014-05-06 04:51:48 now, the mtools apk on the CD causes apk to throw `bus error` 2014-05-06 04:51:52 I am not entirely sure why 2014-05-06 04:52:09 when I tried to `apk add --force mtools`... http://f.cl.ly/items/2g262y1U3S170c2x0V3c/arf464.png 2014-05-06 04:52:13 it purged every package from the system... 2014-05-06 04:52:19 this is just not going very well 2014-05-06 04:53:34 'bus error' is due to running out of ram 2014-05-06 04:53:54 sounds like you are trying to install all the stuff to ram on tmpfs install? 2014-05-06 04:54:14 it is during `setup-alpine` 2014-05-06 04:54:19 so I suppose yes 2014-05-06 04:54:33 rather, I received `bus error` while running `setup-alpine` 2014-05-06 04:54:58 run 'setup-disk' first if you want to install to a disk 2014-05-06 04:55:08 okay 2014-05-06 04:55:39 this still does not explain why `apk add` removed every package though o.O 2014-05-06 04:57:20 http://f.cl.ly/items/201V2O1E3b3q3Q0l093i/arf478.png well, setup-disk won't even /run/... 2014-05-06 04:57:33 sfdisk and lvm2 are missing 2014-05-06 04:57:44 I am using http://dl-5.alpinelinux.org/alpine//v2.7/releases/x86_64/alpine-xen-2.7.6-x86_64.iso 2014-05-06 05:01:30 mmm 2014-05-06 05:02:53 maybe setup-alpine is the right thing to do after all. it's been a long time since i've done hard disk install. 2014-05-06 05:03:06 i run my alpines in run-from-ram setup 2014-05-06 06:21:21 Elizacat: pong 2014-05-06 07:04:04 TheWilfox here i wrote an small wiki article when installing it on my ALIX 3D3 (256MB ram) i did not need to do any of such stuff. and it was x86 http://wiki.alpinelinux.org/wiki/Setting_up_LVM_on_GPT-labeled_disks 2014-05-06 07:50:16 ncopa, so apparently the APKINDEX is corrupt on all the mirros 2014-05-06 07:50:18 *mirrors 2014-05-06 07:53:59 oh? 2014-05-06 07:54:21 that indicates that main mirror is corrupt 2014-05-06 07:54:27 what arch? 2014-05-06 07:54:44 what package(s)? 2014-05-06 07:55:04 er binutils problem 2014-05-06 07:55:20 we have 2 options: 1) add bash as build dep 2014-05-06 07:55:39 2) repackage binutils with newer version of autoconf (run make dist) 2014-05-06 07:55:47 and use our own tarball 2014-05-06 07:56:02 i dont think we want pull in autoconf/automake as build dep for binutils 2014-05-06 07:56:31 problem with option 1 is that we need crosscompile bash too when bootstrapping new arch 2014-05-06 07:57:02 i think we can workaround that by generating our own tarball with newer autoconf 2014-05-06 07:57:10 https://sourceware.org/bugzilla/show_bug.cgi?id=16370 2014-05-06 07:57:57 fabled: any opinion of what do do with binutils? ^^^ option 1 or option 2? 2014-05-06 07:58:15 i'd say use bash as makedepend for now 2014-05-06 08:03:15 just makedepend on bash probably 2014-05-06 08:03:27 though bootstrap gets slightly more complicated then 2014-05-06 08:03:31 hope they fix it 2014-05-06 08:04:25 just using newer autoconf should fix it 2014-05-06 08:05:11 ok, lets makedepend on bash for now 2014-05-06 08:05:33 i think also implementing LINENO in busybox ash should workaround it 2014-05-06 08:05:52 what does LINENO do? 2014-05-06 08:06:04 current line number? 2014-05-06 08:06:10 $LINENO is a special shel var showing current line number 2014-05-06 08:06:14 used for debugging 2014-05-06 08:06:29 autoconf will try detect if the shell has support for it 2014-05-06 08:06:45 and if shell has not support for it, it wil generate a configure.lineno with sed 2014-05-06 08:06:52 where it ijects all the line numbers 2014-05-06 08:06:59 and re-exec that 2014-05-06 08:07:17 the problem only happens with the generated configure.lineno 2014-05-06 10:50:20 hi@all 2014-05-06 10:50:42 ncopa: anything wrong with my patch-mail or just no time at the moment? 2014-05-06 11:01:47 its more no time 2014-05-06 11:01:52 what patch are you thinking of? 2014-05-06 11:02:35 quassel? 2014-05-06 11:02:35 quassel update and group-fix 2014-05-06 11:03:18 looking at it now 2014-05-06 11:03:23 the /run vs /var/run 2014-05-06 11:03:45 is that change necessary? 2014-05-06 11:04:07 i think we will have to change all init scripts to use /run 2014-05-06 11:04:14 but i dont think I'll reach that beofer 3.0 2014-05-06 11:04:34 i c 2014-05-06 11:04:53 but it doesnt matter that much i think 2014-05-06 11:05:01 should be not a problem 2014-05-06 11:05:07 also, i think Id like to merge the patches together into one 2014-05-06 11:05:12 but i can do that 2014-05-06 11:05:25 as you wish :-) 2014-05-06 11:05:46 oh 2014-05-06 11:05:51 fatal: cannot convert from ISO-8859-15 to UTF-8 2014-05-06 11:06:19 ? 2014-05-06 11:07:39 i think its a musl issue 2014-05-06 11:08:06 bah no 2014-05-06 11:08:13 i trie git am the mail itself 2014-05-06 11:08:16 my bad 2014-05-06 11:13:03 done. thanks! 2014-05-06 11:13:09 sorry for the delay 2014-05-06 11:14:48 clandmeter: re #2843 2014-05-06 11:14:59 ncopa: cool, thx 2014-05-06 11:15:24 doesnt it work if you do: lxc-create ... -- --arch=i486 ? 2014-05-06 11:17:03 clandmeter: with your patch --arch=x86 will give i686 and --arch=i686 will give x86 2014-05-06 11:19:50 does google-analytics.com on git.a.o help or useful ? 2014-05-06 11:23:37 might be clandmeter knows 2014-05-06 11:24:05 fabled: re #2846, op to cherry pick update-ca-certificates for v2.7? 2014-05-06 11:24:26 ncopa, yes please 2014-05-06 12:07:09 ncopa regarding lxe i can try it right now. 2014-05-06 12:08:57 i think i will downgrade firefox to 24.5.0esr in v2.7 2014-05-06 12:09:08 we shoudl try aim for esr releases of firefox 2014-05-06 14:12:07 ncopa: Now that it is May, any thoughts on when 3.0 release might occur? 2014-05-06 14:19:52 i havent get up the official builders yet... 2014-05-06 14:19:58 i was supposed to do that yesterday 2014-05-06 14:20:00 or today 2014-05-06 14:20:43 i'm gonna do v2.7.7 now 2014-05-06 14:20:53 and then do full v3.0 focus 2014-05-06 14:47:25 urgh.. 2014-05-06 14:47:39 i just realized i should have asked rnalrd to upgrade virt kernel too 2014-05-06 14:47:43 ..again.. 2014-05-06 14:48:00 I should ad that to my release checklist 2014-05-06 14:50:40 :) 2014-05-06 20:39:02 i have hier an call stack: http://sprunge.us/jPHI (see end of dmesg) 2014-05-06 20:40:04 on 2.7.6 2014-05-07 09:27:56 flashcache users: flashcache 3rd party module does not build against 3.14 kernel 2014-05-07 09:28:11 and I see no commits in github that indicates that they fixed it 2014-05-07 09:28:26 https://github.com/facebook/flashcache/commits/master 2014-05-07 09:28:41 is it ok if i remove flashcache kernel module? 2014-05-07 09:49:53 ok to remove open-vm-tools? 2014-05-07 09:53:57 ncopa: why? 2014-05-07 09:54:10 kaniini: ipfw-grsec seems to lack some files and checksums 2014-05-07 09:54:28 StarWarsFan: because there seem to be no 3.14 kernel support 2014-05-07 09:54:33 StarWarsFan: do you use it? 2014-05-07 09:54:51 on 2.7 yes 2014-05-07 09:54:55 ok 2014-05-07 09:55:10 should try fix it then i suppose 2014-05-07 09:55:14 i thought nobody used it 2014-05-07 09:55:18 but for me it's not a show stopper 2014-05-07 09:55:26 nfs is way more important on edge... ;-) 2014-05-07 09:55:29 yeah 2014-05-07 09:55:34 i need fix that too 2014-05-07 09:55:52 the dmesg thingy 2014-05-07 09:55:59 looks like you are running nfs in a container? 2014-05-07 09:56:09 hopefully 3.14. kernel fixes those issues 2014-05-07 09:56:32 there are still problem with userspace daemons 2014-05-07 09:56:41 in fact its a double container 2014-05-07 09:56:48 oh :) 2014-05-07 09:57:24 ubuntu-xen < alpine-edge-64 < alpine-edge-x86-lxc 2014-05-07 09:57:42 nfs on the lxc container with x86 works 2014-05-07 09:57:53 but not on edge-64 2014-05-07 09:58:18 does the x86 container give any dmesg errors? 2014-05-07 10:14:57 ncopa: yes, it does 2014-05-07 10:14:58 http://sprunge.us/AGHe 2014-05-07 10:33:56 yesterday i have had an have hier an call trace with 2.7.6 : http://sprunge.us/jPHI (see end of dmesg) 2014-05-07 10:36:37 got a new router TP Link MR3020 :-) 2014-05-07 10:43:53 that my call trace seems from watchdog. 2014-05-07 10:44:09 what is build/mobilebox ? 2014-05-07 11:31:00 crow: looks like some kind of kernel issue 2014-05-07 11:32:41 rebooting brb 2014-05-07 11:50:58 on edge i have some issues after i followed the upgrade procedure. While running apk I get: 2014-05-07 11:51:04 rror relocating /lib/libcrypto.so.1.0.0: __ctype_b: symbol not found 2014-05-07 11:51:15 Error relocating /lib/libcrypto.so.1.0.0: __ctype_tolower: symbol not found 2014-05-07 11:51:24 how can i fix that? 2014-05-07 11:53:05 sounds like you have a mix of musl/uclibc 2014-05-07 11:53:18 do you have apk.static installed? 2014-05-07 11:53:49 run: apk.static upgrade -U --no-self-upgrade -a 2014-05-07 12:52:32 ncopa yea but how to debug, i mean i saw it first time yesterday, i am now on 2.7.7 2014-05-07 12:53:29 dont know 2014-05-07 12:53:42 see if git log has anything new for you net driver 2014-05-07 12:53:53 or if google say somethingabout htat message 2014-05-07 12:59:54 ok 2014-05-07 14:48:16 ok build-3-0-x86 is up and running now 2014-05-07 14:48:28 i'm bootstrapping build-3-0-x86_64 too 2014-05-07 14:54:03 ok 2014-05-07 14:54:12 build-3-0-x86_64 is up and running too now 2014-05-07 18:53:40 ncopa: hmm 2014-05-07 18:53:42 ncopa: will fix 2014-05-07 21:15:11 Evening 2014-05-07 21:31:08 n8@all 2014-05-08 08:15:45 GCC 4.9.0 would make to v3.0 ? 2014-05-08 11:21:51 ncopa, AL has wifi direct support ? 2014-05-08 11:22:24 not that i know 2014-05-08 11:22:28 what is wifi direct? 2014-05-08 11:22:33 CONFIG_IWLWIFI_P2P=y 2014-05-08 11:22:39 http://fabiobaltieri.com/2013/02/14/hp-wi-fi-direct-mouse-on-linux/ 2014-05-08 11:24:10 with wifi direct one does not need AP or wifi router to connect 2014-05-08 11:24:38 ncopa-desktop:~$ grep P2P /boot/config-grsec 2014-05-08 11:24:39 ncopa-desktop:~$ 2014-05-08 11:24:41 client acts as AP 2014-05-08 11:25:45 vkrishn: btw, i enable zram in erknel 2014-05-08 11:26:09 thanks, saw the bugs.a.o 2014-05-08 11:26:27 i'd like kill flashcache 2014-05-08 11:26:42 what is current states of zcache and zswap, are the stable ? 2014-05-08 11:26:49 they* 2014-05-08 11:26:54 they are in mainline kernel not in staging yes 2014-05-08 11:26:59 afaik 2014-05-08 11:27:24 i havent tested them 2014-05-08 11:28:00 I have not tested them 2014-05-08 11:28:15 ok 2014-05-08 11:29:51 should I add bug.a.o for CONFIG_IWLWIFI_P2P=y, wifi direct ? 2014-05-08 11:31:03 do you need it? 2014-05-08 11:31:43 I just got a wifi dongle that day softAP support on windows, so thought of AL support 2014-05-08 11:31:45 :) 2014-05-08 11:31:50 that has* 2014-05-08 11:32:06 ok 2014-05-08 11:32:10 I'll look into it 2014-05-08 11:32:51 thanks 2014-05-08 11:34:49 Whats the diff from NFC? 2014-05-08 11:35:48 if I am correct my android phone used wifi-direct to create hotspot, which I current connected to :) 2014-05-08 11:35:49 uses* 2014-05-08 11:36:14 nfc is similar to bluethooth - very short distances 2014-05-08 11:36:48 they are very good for security point when you want to exchange security codes 2014-05-08 11:37:11 without wide snooping area 2014-05-08 11:38:09 like software that change files by tapping/touching together two mobiles 2014-05-08 11:38:41 wifi-direct for p2p games maybe 2014-05-08 11:40:12 where can i find iwlwifi in menuconfig 2014-05-08 11:40:16 but with wifi-direct and available security with it, it may just replace nfc + bluetooth 2014-05-08 11:41:22 hmm.. goto check, would read the docs and report, did not do much findings 2014-05-08 11:45:45 its an intel thingy 2014-05-08 11:46:26 Intel Wireless WiFi Next Gen AGN 2014-05-08 11:47:05 i dont know whow to enable the P2P support 2014-05-08 11:47:55 would also check if it has full support on linux 2014-05-08 11:53:23 I think monitor with wi-di support also uses wifi-direct 2014-05-08 11:57:22 DLNA devices now have support for them 2014-05-08 12:02:39 http://www.wi-fi.org/discover-wi-fi/wi-fi-direct - nice explanations 2014-05-08 12:09:57 I think Bluetooth would stay because of its low power consumption 2014-05-08 12:28:52 android also calls it as wifi-tethering, portable hotspot 2014-05-08 12:50:09 /lib/modules/3.9.6-64/kernel/drivers/staging/ozwpan/ozwpan.ko , still in staging, not checked latest state 2014-05-08 14:17:19 apk is totally bricked on v3.0 builders 2014-05-08 14:17:22 i wonder why.. 2014-05-08 14:17:35 $ apk 2014-05-08 14:17:35 Killed 2014-05-08 14:17:56 [4837048.037745] PAX: From 84.202.203.3: execution attempt in: , 00000000-00013000 00000000 2014-05-08 14:18:14 i wonder if the linker is broken 2014-05-08 14:18:15 or gcc 2014-05-08 14:18:17 or something 2014-05-08 16:22:13 Question: Is it true saying that a .apk is nothing more than a *gziped tar package*? 2014-05-08 16:22:44 mostly true, yes 2014-05-08 16:22:48 I can tar -zxvf .apk 2014-05-08 16:23:55 we don't use any hipser compression, just like in every distro a package is a compressed tarball 2014-05-08 16:24:17 s/hipser/hipster/ 2014-05-08 16:25:25 "just like in every distro a package is a compressed tarball" great i didnt know that. I got to know that AL works like this because i am reading the old apk-tools ( shellscript based version ) 2014-05-08 16:25:43 barthalion: for me it is a new information! 2014-05-08 21:29:03 localhost:~# apk fix 2014-05-08 21:29:05 ERROR: unsatisfiable constraints: 2014-05-08 21:29:07 flashcache-grsec (missing): 2014-05-08 21:29:09 required by: world[flashcache-grsec] 2014-05-08 21:29:11 hmmph 2014-05-08 21:34:20 kaniini: it was removed due to the 3.14 kernel bump 2014-05-09 05:47:29 uggedal: yes, i looked into it and found that porting flashcache to 3.14 was really not a good use of my time 2014-05-09 06:09:33 anytime this year would be great 2014-05-09 06:09:51 the 3.0 builders are broken 2014-05-09 06:10:13 i should probably take them offline and fix them 2014-05-09 06:11:38 kaniini: is enhanceio a replacement for flashcache? 2014-05-09 06:11:44 does it need 3rd party kernel mods? 2014-05-09 06:18:31 fabled: 3.0 builders' apk-tools is totally broken 2014-05-09 06:18:40 does not even print help text 2014-05-09 06:18:42 http://sprunge.us/fQaj 2014-05-09 06:18:54 i supsect there something with our linker or so 2014-05-09 06:19:13 running the dynamic linked from edge works 2014-05-09 06:20:19 what id did was 1) create new container 2) build bintutils, musl and gcc 3)replace with local built toolchain 4) build world 2014-05-09 06:20:32 the apk-tools built with that seems to be broke 2014-05-09 06:21:54 whatta 2014-05-09 06:22:06 if I rebuild apk-tools once again, it works 2014-05-09 06:48:21 Ha! 2014-05-09 06:48:38 someone sent me suggestions on a new logo 2014-05-09 06:50:14 and a mockup on new website design 2014-05-09 06:56:44 and i think its really good 2014-05-09 07:36:25 can you publish it someplace to have a look at? 2014-05-09 07:37:00 http://bugs.alpinelinux.org/issues/2882 2014-05-09 07:37:18 i think its like 10x better than what we have today 2014-05-09 07:37:30 i think the logo could be simplified a bit 2014-05-09 07:37:55 so its more like the icon style on the website mockup 2014-05-09 07:38:08 maybe only 1 or 2 colors 2014-05-09 07:45:46 I uploaded some of my older ideas too 2014-05-09 08:27:20 moin 2014-05-09 08:28:06 i did not see the point why to change the logo 2014-05-09 08:28:20 the big advantage of the current logo is its simplicity 2014-05-09 08:28:46 the new suggestion looks cool but this is multicolor stuff 2014-05-09 08:28:59 which makes it difficult to handle on all places 2014-05-09 08:29:47 think about very small icons or "real" stuff i. e. for merchandising 2014-05-09 08:30:14 in fact as less colors as possible should be used IMHO 2014-05-09 08:30:52 an example of the current draft in 2/4/8 colors would be interesting... 2014-05-09 08:31:08 yes, thats what i told the guy 2014-05-09 08:31:11 fewer colors 2014-05-09 08:31:16 maybe only 1 color 2014-05-09 08:31:35 but i really like the concept 2014-05-09 08:31:42 the hexagon is like a bolt 2014-05-09 08:31:44 i've tried to get some shirts for our team with our logo and this was a real problem 2014-05-09 08:31:47 = engineering 2014-05-09 08:31:50 depending on the amount of colors... 2014-05-09 08:32:20 we need variants 2014-05-09 08:32:29 another question for me is: why a new logo? 2014-05-09 08:32:33 we need a one color variant 2014-05-09 08:32:42 ack 2014-05-09 08:33:14 because while the current one "works", i think its no great 2014-05-09 08:47:14 the current logo was made in a hurry 2014-05-09 08:47:25 he just drawed something to get going 2014-05-09 08:47:48 it was back when we had the main site in mediawiki 2014-05-09 08:48:58 and i think with 3.0 release, its a good time to make a refresh 2014-05-09 09:00:08 yeah. going to musl is pretty much changing many things. 2014-05-09 09:00:13 if we get new logo now is the time. 2014-05-09 10:56:31 new logo is amazing 2014-05-09 11:04:53 barthalion: good that its not only me who thinks so :) 2014-05-09 11:09:05 i got some military vibes from the ^^ in a badge looking hexagon 2014-05-09 11:09:49 but i yes, i think it looks nice. i think it just needs color simplifications 2014-05-09 11:25:46 ack 2014-05-09 11:32:33 the Alpine font is actually not a font either, its handmade 2014-05-09 11:33:06 which is *good* 2014-05-09 11:33:14 it means it will be unique 2014-05-09 12:00:52 whatta 2014-05-09 12:00:56 apk died again 2014-05-09 12:01:34 i wonder what happened 2014-05-09 12:03:10 fabled: I have a problem. 2014-05-09 12:03:18 when bootstrapping aports, and rebuild apk-tools 2014-05-09 12:03:30 apk breaks 2014-05-09 12:03:37 does not even print help text 2014-05-09 12:03:50 I have no idea what or why it happens 2014-05-09 12:03:58 apk works from edge 2014-05-09 12:04:13 but when boottrapping toolchain and rebuild apk, it appears to break badly 2014-05-09 12:04:53 huh 2014-05-09 12:05:07 build-3-0-x86_64:~$ apk 2014-05-09 12:05:07 Killed 2014-05-09 12:06:46 maybe i cna reproduce it on x86 builder 2014-05-09 12:06:56 i havent started the rebuild of world there yet 2014-05-09 12:07:08 so i have a fresh gcc toolchain 2014-05-09 12:07:17 and the old apktools 2014-05-09 12:10:05 rebuilding all the apk-tools deps 2014-05-09 12:10:53 i remember we had to do -nopie on apk 2014-05-09 12:10:58 i wonder if its related 2014-05-09 12:11:23 ok 2014-05-09 12:11:24 yes 2014-05-09 12:11:29 i have it reproduced now 2014-05-09 12:15:33 what i do is: rebuild musl binutils gcc make patch and all their deps, then apk upgrade -U -a, then build all apk-tools deps and apk-tools itself 2014-05-09 12:15:39 and apk ugprade 2014-05-09 12:15:42 and it goes boom 2014-05-09 12:16:04 It could be me who does someting wrong too... but I dont know what 2014-05-09 12:27:51 speaking of generating website, i think this is pretty cool: https://github.com/jroimartin/sw 2014-05-09 12:28:01 its the suckess generator 2014-05-09 12:28:08 suckless.org generator 2014-05-09 12:28:22 its a markdown -> html implementation in awk 2014-05-09 12:28:37 and a shell script that generates the index and menus 2014-05-09 12:28:50 i think its brilliant 2014-05-09 12:33:49 two commits, doesn't sound good 2014-05-09 12:35:14 its only one commit 2014-05-09 12:35:18 he borbably wrote it 2014-05-09 12:35:21 make it run 2014-05-09 12:35:22 used it 2014-05-09 12:35:34 and figured that "maybe i should store it in git" 2014-05-09 12:35:53 in any case, its a trivial script 2014-05-09 12:36:37 its baiscally a 100 lines shell script 2014-05-09 12:36:46 + a .md -> .html converter 2014-05-09 12:36:51 we could use anything for that part 2014-05-09 12:37:11 i just like the simplicity of it 2014-05-09 12:37:25 and that i can store the actual content in git 2014-05-09 14:00:58 alacerda wants to have a meeting about advocating/marketing alpine 2014-05-09 14:01:19 maybe we should set up a separate irc channel for that? 2014-05-09 14:01:54 The subj would be what else we can do to promote AL 2014-05-09 14:02:29 i think step 1 would be set up a "marketing" team or subproject 2014-05-09 14:02:40 or at least an IRC channel 2014-05-09 14:04:25 A marketing team would be great. 2014-05-09 14:04:54 the problem is, i suspect nobody wants to do the *work* :) 2014-05-09 14:05:50 alacerda: i suppose one of the questions are: who do we want to reach? 2014-05-09 14:05:58 who do we want should notice alpine linux? 2014-05-09 14:06:35 i would say: the kind of people who would contribute back 2014-05-09 14:07:30 That is another question that i would like to know the answer 2014-05-09 14:07:45 it seems the AL goal had changed 2014-05-09 14:08:01 ok? 2014-05-09 14:08:46 its no longer a small/simple/secure OS? 2014-05-09 14:43:19 ncopa: i missunderstood this: http://alpinelinux.org/about in the topic "Why the Name Alpine?" 2014-05-09 14:43:35 that is why i said that the goal changed 2014-05-09 14:47:47 "Alpine" originally stood for A Linux Powered Integrated Network Engine. The idea was that the distribution would be focused on networking, and be a tiny "engine" or framework, which larger systems could be built upon. 2014-05-09 14:47:47 Today, "Alpine" is nothing more than a name, despite continuing to live up to its original name. As an example, here are a few of Alpine's achievements and real-world uses: 2014-05-09 14:48:06 The first open-source implementation of Cisco's DMVPN, called OpenNHRP, was written for Alpine Linux. 2014-05-09 14:48:06 In addition to its use as a firewall or router system, Alpine Linux is also used in a number of installations as the basis for enterprise servers, running such software as PostgreSQL, Postfix, Asterisk, Kamailio, and being used for iSCSI SANs. It is the little engine that could. 2014-05-09 14:48:10 we could add xen to that 2014-05-09 14:48:22 "Alpine Linux is a community-developed operating system designed for x86 Routers, Firewalls, VPNs, VoIP and servers." 2014-05-09 14:48:32 that sentence could probably be clarified a bit too 2014-05-09 15:19:34 so your not after desktop enthusiast? 2014-05-09 15:46:11 lalala http://makthepla.net/blog/=/plesk-sso-xxe-xss 2014-05-09 15:47:25 just in case somebody is using plesk: http://makthepla.net/blog/=/plesk-sso-xxe-xss 2014-05-09 15:47:31 argh 2014-05-10 13:43:14 bah 2014-05-10 13:46:38 ? 2014-05-10 13:48:12 hmm 2014-05-10 13:50:21 this is rather annoying to have to work semi-blind with 2.7 2014-05-10 13:51:16 i don't get it. lua-pc works on lua5.2 on edge; but not 2.7 2014-05-10 14:03:08 duh 2014-05-10 14:03:11 cflags 2014-05-12 06:27:58 morning 2014-05-12 07:55:22 morning 2014-05-12 09:23:04 ncopa, for wifi P2P, just a latest wpa_supplicant would make it work of some kernel settings/ module also ? 2014-05-12 09:24:26 also wpa_supplicant needs to be compiled with options supporting P2P 2014-05-12 09:25:09 CONFIG_WPS=y 2014-05-12 09:25:11 CONFIG_WPS2=y 2014-05-12 09:25:11 CONFIG_READLINE=y 2014-05-12 09:25:11 CONFIG_AP=y 2014-05-12 09:25:11 CONFIG_P2P=y 2014-05-12 09:29:19 oh.. maybe it was wpa_supplicant config 2014-05-12 09:29:33 i dont know what it needed from kernel to make it work 2014-05-12 09:29:58 i thought there was some *_P2P config in kernel that needed to be enabled but i could not find out how 2014-05-12 09:31:02 ncopa, small favour, if you have a AL running on your laptop, could you pls try 2014-05-12 09:31:09 iw list 2014-05-12 09:31:38 presently CONFIG_WPS=y , and CONFIG_READLINE=y are already enabled 2014-05-12 09:33:28 I am interested in "Supported interface modes:" 2014-05-12 09:33:56 you find the full config of wpa_supplicant here: http://git.alpinelinux.org/cgit/aports/tree/main/wpa_supplicant/APKBUILD#n41 2014-05-12 09:34:10 im starting up my laptopn 2014-05-12 09:36:36 sudo iw list only gives me the iw help 2014-05-12 09:36:44 on my knoppix with wpa_supplicant 1.0, https://dpaste.de/Tix1 2014-05-12 09:37:23 though Supported TX / RX shows P2P support 2014-05-12 09:37:43 sudo iw list gives me this: http://sprunge.us/MMLd 2014-05-12 09:37:48 which looks wrong... 2014-05-12 09:37:57 but not in section "Supported interface modes:" 2014-05-12 09:40:02 sudo iw wlan0 info 2014-05-12 09:41:10 http://sprunge.us/THIM 2014-05-12 09:48:00 iw phy #<-- gives me same result as `iw list` 2014-05-12 09:48:24 yup it does :) 2014-05-12 09:48:29 only help text 2014-05-12 09:49:00 then something not ok with wpa_supplicant ? or some pkg missing ? 2014-05-12 09:49:08 dunno 2014-05-12 09:49:16 wpa_supplicant is installed and is running 2014-05-12 09:49:24 like nl80211 driver 2014-05-12 09:49:26 i am ssh to it via wifi 2014-05-12 09:50:00 1615 root 0:00 /sbin/wpa_supplicant -B -c/etc/wpa_supplicant/wpa_supplicant.conf -iwlan0 2014-05-12 10:12:31 just wondering if it should be CONFIG_DRIVER_NDIS=y or CONFIG_DRIVER_NDISWRAPPER=y ? 2014-05-12 10:29:30 LXQT if out 2014-05-12 10:32:21 nothing interesting yet 2014-05-12 10:32:28 less features than in LXDE, still buggy 2014-05-12 10:32:41 and frankly, I don't see major progress since razorqt 2014-05-12 10:43:35 should be interesting after couple of releases 2014-05-12 10:43:44 razorqt or lxqt could be missing link for creating small comm devices (diy), eg camera comm on house door 2014-05-12 10:44:51 or maybe use web interface 2014-05-12 13:10:11 i wonder ho lxqt memory usage is compared to lxc with glib 2014-05-12 13:10:16 lxde* 2014-05-12 13:10:37 its easy to blow up memory usage with c++ 2014-05-12 13:11:10 oh and with glib, its impossible to create a failsafe application 2014-05-12 13:11:29 as in handle out-of-memory in a propre way 2014-05-12 13:59:06 looks important, http://wireless.kernel.org/en/developers/Regulatory/CRDA 2014-05-12 14:12:26 this page http://wireless.kernel.org/en/developers/Documentation/nl80211 , says wpa_supplicant (with -Dnl80211) 2014-05-13 05:59:41 morning 2014-05-13 05:59:52 fabled! congrats with the apk pie fix! 2014-05-13 06:52:35 kaniini: audacious-3.5 seems to depend on libguess-1.2 but i can only see libguess-1.1? 2014-05-13 06:52:41 http://atheme.org/projects/libguess.html 2014-05-13 06:52:52 github does not seem to have any 1.2 tag either 2014-05-13 07:19:52 ncopa, i was in vacation but i've read the 3d about alpine logo 2014-05-13 07:19:56 where can i find the mokup? 2014-05-13 07:23:28 http://bugs.alpinelinux.org/issues/2882 2014-05-13 07:23:41 thanks! 2014-05-14 06:07:31 morning 2014-05-14 06:08:03 I want to bump llvm to 3.4.1 so that I can play around with rust 2014-05-14 06:08:23 if mesa builds with llvm 3.4.1 is it safe to push or should mesa be tested runtime first? 2014-05-14 06:11:04 the llvm patches needs some changes, I take it's ok to remove uclibc support? 2014-05-14 06:25:34 uggedal, yes, on edge we don't really care about uclibc anymore. 2014-05-14 06:25:45 if it's needed in 2.7-stable, then uclibc needs to stay. 2014-05-14 06:59:26 i looked at llvm a month ago or so 2014-05-14 06:59:35 i dont think its difficult to rebase the patch 2014-05-14 06:59:44 its just a little work 2014-05-14 07:22:50 ncopa: I've already rebased the patch 2014-05-14 07:23:30 my question was whether 'mesa compiles with llvm 3.4.1' is a good enough test 2014-05-14 07:23:36 yeah 2014-05-14 07:23:43 good enough test 2014-05-14 07:24:06 i think clang is able to compile helloworld.c is good enough 2014-05-14 07:24:25 thanks for doing it 2014-05-14 07:24:31 i wanted 3.4 too 2014-05-14 07:25:23 will mesa need a revision bump for rebuilding against newest llvm (not sure what mesa is using llvm for...)? 2014-05-14 07:28:11 i think not 2014-05-14 07:28:22 you can check with 2014-05-14 07:28:31 cd main/llvm && checkapk 2014-05-14 07:28:42 check if any of the libs has changed SONAME 2014-05-14 07:29:01 if anything in mesa links to any of the libs from llvm 2014-05-14 07:29:14 and the lib has changed SONAME, then you need to rebuild 2014-05-14 07:29:50 i think mesa has drivers that uses llvm for the gpu 2014-05-14 07:29:54 probably amd 2014-05-14 07:30:05 there might be a software renderer that uses llvm too iirc 2014-05-14 07:30:09 llvmpipe or so 2014-05-14 07:30:28 http://www.mesa3d.org/llvmpipe.html 2014-05-14 07:37:45 thanks! mesa-dri-ati depends on so:libLLVM-3.3.so which is so:libLLVM-3.4.1.so now 2014-05-14 07:38:43 oh, they even did soname change on the .1 release... 2014-05-14 07:38:45 ok 2014-05-14 07:39:12 we need to rebuild everything that depends on so:libLLVM-3.3.so 2014-05-14 07:39:51 http://sprunge.us/fLeV 2014-05-14 07:39:55 looks like mesa only 2014-05-14 07:40:11 just bumping the pkgrel on mesa should be enough 2014-05-14 07:40:16 yep 2014-05-14 07:44:23 hmm, why did build-edge-x86 build mesa first? 2014-05-14 07:51:06 WARNING: python: Skipped due to previous build failure 2014-05-14 07:51:06 WARNING: llvm: Skipped due to missing dependencies 2014-05-14 07:51:07 WARNING: mesa: Skipped due to missing dependencies 2014-05-14 07:51:41 is that from edge? 2014-05-14 07:51:57 http://bld1.alpinelinux.org/buildlogs/build-edge-x86.log 2014-05-14 07:52:30 we have it on "continue even if it fails" currently 2014-05-14 07:52:46 we should switch back to break and halt on error 2014-05-14 07:53:08 did I need to hold off on pushing the mesa revbump until llvm was built (if python worked that is) or do the builders build the commits in order? 2014-05-14 07:53:10 you find the individual build logs in http://bld1.alpinelinux.org/buildlogs/build-edge-x86/ 2014-05-14 07:53:16 no 2014-05-14 07:53:27 builders should dected build order 2014-05-14 07:53:52 http://bld1.alpinelinux.org/buildlogs/build-edge-x86/main/python/python-2.7.6-r4.log 2014-05-14 07:53:59 its the /dev/shm issue 2014-05-14 07:56:01 I'm restarting build-edge-x86 2014-05-14 07:57:04 ok 2014-05-14 07:57:08 python is building now 2014-05-14 07:57:19 good :) 2014-05-14 07:59:10 oaa 2014-05-14 07:59:17 looks like wonr commit message 2014-05-14 07:59:54 yeah.. 2014-05-14 07:59:55 bah 2014-05-14 09:28:24 my laptop does not boot the 3.14.4 kernel 2014-05-14 09:28:43 i think its my initramfs that is missing framebuffer/kms/something 2014-05-14 09:36:34 files from v2.5.4-286-g579552c uploaded 2014-05-14 09:45:57 ncopa edge? 2014-05-14 09:46:47 yup 2014-05-14 09:49:49 but i suspect its due to UEFI or simlar 2014-05-14 09:49:53 i think my desktop works 2014-05-14 09:49:59 rebooting... brb 2014-05-14 09:51:49 ncopa dont have such new laptop (with uefi) was thinking to put on my old IBM X60S alpine edge 2014-05-14 09:52:49 i'd expect that to work fairly well 2014-05-14 09:53:05 unless you depend on binary-only drivers 2014-05-14 09:54:03 brb 2014-05-14 09:57:18 my desktop works. no problem 2014-05-14 09:57:24 with lvm2 2014-05-14 09:57:48 rnalrd: i have upgraded grsec kernel in v2.7 (and edge) 2014-05-14 10:00:03 x60s is have an intel gpu "Intel Graphics Media Accelerator 950" which should be supported with the xf86-video-intel 2014-05-14 10:00:25 ncopa crypted lvm2 ? 2014-05-14 10:00:46 uncrypted 2014-05-14 10:00:59 ncopa: for 3.0, could we try to get cinnamon working? 2014-05-14 10:01:49 cinnamon is the gtk3? 2014-05-14 10:02:02 and mate is the gtk2? 2014-05-14 10:02:08 i never remember :) 2014-05-14 10:02:27 yes 2014-05-14 10:03:16 i added it to aports some time ago, but couldnt get it working. 2014-05-14 10:05:12 ncopa what DE you use with alpine edge? 2014-05-14 10:05:18 or VM 2014-05-14 10:05:27 erm wm 2014-05-14 10:07:30 xfce 2014-05-14 10:36:55 ncopa thank you, just to know what could i test and expect to work 2014-05-14 11:07:45 my lappy booted up with 3.14.4 kernel 2014-05-14 11:07:49 dunno what happened 2014-05-14 11:08:57 ncopa: your return question was someting like an answer? :) 2014-05-14 11:10:30 my laptop does not boot the 3.14.4 kernel 2014-05-14 11:10:30 i think its my initramfs that is missing framebuffer/kms/something 2014-05-14 11:10:35 it was 3.13 2014-05-14 11:10:38 3.14.2 2014-05-14 11:10:43 re cinnamon 2014-05-14 11:10:47 i'm okish with it 2014-05-14 11:10:55 but i doubt i will have time to work on it 2014-05-14 11:11:37 ok 2014-05-14 11:12:29 well whenever you have time and could make it only work, i can try to clean up the rest. (its probably something stupid im missing). 2014-05-14 13:07:06 ncopa, for wifi p2p/direct, pls have a look at http://wireless.kernel.org/en/developers/p2p/howto 2014-05-14 13:07:15 I can generate a patch for wpa_supplicant (untested) 2014-05-14 13:07:15 or add to bugs.a.o ? 2014-05-14 13:08:15 sure 2014-05-14 13:08:17 hm 2014-05-14 13:08:23 we have a circular dep problem 2014-05-14 13:08:36 avahi -> gtk -> cups -> avahi 2014-05-14 13:08:40 or similar 2014-05-14 13:11:41 what code handles circular deps ? 2014-05-14 13:11:42 if any 2014-05-14 13:12:32 nothing 2014-05-14 13:12:39 we try to avoid circular deps 2014-05-14 13:12:57 but seems like cups-1.7.2 absolutely wants avahi 2014-05-14 13:13:05 dirsvc.o: In function `cupsdStopBrowsing': 2014-05-14 13:13:06 dirsvc.c:(.text+0x49e): undefined reference to `dnssdDeregisterAllPrinters' 2014-05-14 13:13:06 dirsvc.o: In function `cupsdStartBrowsing': 2014-05-14 13:13:06 dirsvc.c:(.text+0x47a): undefined reference to `dnssdRegisterAllPrinters' 2014-05-14 13:18:48 hey ncopa 2014-05-14 13:19:01 idk if this is normal.. but I just upgraded my edge box 2014-05-14 13:19:06 which was already on musl 2014-05-14 13:19:09 and it's giving me all this again 2014-05-14 13:19:16 http://puu.sh/8LVvF.txt 2014-05-14 13:22:38 what does lddtree.sh do ? 2014-05-14 13:23:18 Mp5shooter: yes should be okish 2014-05-14 13:23:26 ok 2014-05-14 13:23:37 you can check your /etc/mkinitfs/features.d files 2014-05-14 13:23:38 to be sure 2014-05-14 13:23:45 there are some .apk-new there likely 2014-05-14 13:23:53 compare them 2014-05-14 13:24:11 the /lib/* and /usr/lib/* are no longer needed 2014-05-14 13:24:18 oh there are 2014-05-14 13:24:23 it should automatically pull in the deps now 2014-05-14 13:27:36 did apk fix mkinitfs and no errors now after replacing the files 2014-05-14 13:27:40 thx 2014-05-14 13:27:58 good 2014-05-14 13:29:01 I was afraid I wouldn't be able to boot again 2014-05-14 13:29:02 haha 2014-05-14 14:12:57 is progs from open-mesh.org important ? batman ? 2014-05-14 15:20:34 main/monit update to 5.8.1 pls 2014-05-14 18:09:02 Hi! Anybody knows where the alpine image for RPI is? 2014-05-14 18:09:17 Raspberry PI 2014-05-14 18:19:21 Never mind - found it 2014-05-14 19:48:07 uia : ) 2014-05-15 06:30:09 kaniini: did you git add musl-locking-fix.patch? 2014-05-15 06:30:20 I cannot see it in the commit: http://git.alpinelinux.org/cgit/aports/commit/?id=38778fca7461a8afc34109a1e5efc150113ef716 2014-05-15 06:54:46 kaniini: sha512sum: can't open 'musl-locking-fix.patch': No such file or directory 2014-05-15 07:53:32 ncopa: sorry about that 2014-05-15 07:54:10 np 2014-05-15 08:43:27 hm 2014-05-15 08:43:34 libiberty disappeared 2014-05-15 08:43:45 i think binutils no longer ships the libiberty.a 2014-05-15 08:45:26 we ship it in gcc for a while 2014-05-15 08:46:29 and that makes sense: https://gcc.gnu.org/ml/gcc/2013-03/msg00301.html 2014-05-15 09:33:54 algitbot: build 2.6-stable 2014-05-15 09:46:38 seems like bld1.a.o has problems... 2014-05-15 09:59:49 make -j 2014-05-15 09:59:49 interesting 2014-05-15 09:59:56 while building kernel... 2014-05-15 10:00:01 gives interesting results 2014-05-15 10:00:09 eats up your memory 2014-05-17 17:50:23 I'm quite surprised, I thought that we have wayland in repos 2014-05-17 17:59:14 ok, it builds 2014-05-17 18:17:45 build failure in 3 2014-05-17 18:17:46 2 2014-05-17 18:17:47 1 2014-05-17 18:20:51 has anyone used the dnscrypt-proxy package? 2014-05-17 18:21:16 algitbot: http://cl.jroo.me/z3/p/O/M/d/a.aaa-Happy-Snake.png 2014-05-17 18:23:55 I'm getting better, only 23 days since last contribution 2014-05-17 19:25:05 Howdy! 2014-05-17 19:25:29 ncopa fixed som updates on 'testing/motion' yesterday. 2014-05-17 19:25:59 The updates seems to be available in testing/motion on 'edge' repository for x86 and x86_64 2014-05-17 19:26:10 But in armfs it seems to be missing 2014-05-17 19:26:30 Does the builder for 'armfs' need some manual hands-on to build a package? 2014-05-17 19:27:20 s/armfs/armhf/g 2014-05-17 21:19:22 mhavela: no 2014-05-17 21:19:30 it probably means it doesnt build on arm 2014-05-18 08:34:05 Thanks for the feedback clandmeter. 2014-05-18 08:34:20 I try to build it on my RPI to see if I can figure out why it doesn't build. 2014-05-18 08:34:24 hi mhavela 2014-05-18 08:35:09 its an asumption im having not a fact. 2014-05-18 08:35:20 Do you know if the build logs for arm is somewhere accessable (without ssh access)? 2014-05-18 08:35:24 i know the builders continiu even if a build has failed. 2014-05-18 08:35:38 I think your asumption is logic - the thought slipped my mind too 2014-05-18 08:36:10 s/slipped/crossed/g 2014-05-18 08:37:57 i never seen a log from arm 2014-05-18 08:38:05 fabled takes care of that. 2014-05-18 09:49:12 I think it was a bit optimistic to start building stuff on my RPI :-/ 2014-05-18 09:49:12 It doesn't have that much memory :-) 2014-05-18 10:45:39 mhavela: :) 2014-05-18 10:45:49 I had the same problem 2014-05-18 10:45:59 so i wanted to use nfs, but it/was broken 2014-05-18 14:53:24 Building stuff on my RPI doesn't seem like a option. Too little resources. 2014-05-18 15:08:54 clandmeter: RE nfs. Do you remember if it was the client or serverside that was broken? 2014-05-19 06:20:49 Good morning. 2014-05-19 06:21:39 ncopa: I think I found out that 'v4l-utils' is breaking testing/motion for armhf 2014-05-19 06:22:09 ok? 2014-05-19 06:22:13 But I couldn't fully verify it because my RPI doesn't have resources to build it (too little memory) 2014-05-19 06:22:55 Should I update the APKBUILD as i /think/ it would work - and send it to you (even if I haven't verified if it works)? 2014-05-19 06:24:35 mhavela: you could sprunge a diff here? 2014-05-19 06:24:50 and exaplain why you think it will fix things 2014-05-19 06:24:53 Sure 2014-05-19 06:27:33 ncopa: http://sprunge.us/LfUI 2014-05-19 06:28:12 I think there is some libraries (or something else) missing in armhf for the v4l-utils stuff to work 2014-05-19 06:28:25 which breaks build 2014-05-19 06:28:54 mhavela, i think you got it the wrong way around 2014-05-19 06:29:17 the dependency is broke. 2014-05-19 06:29:28 you cannot expect to build the package if dependency failed 2014-05-19 06:29:35 we need fix the dependency to build first 2014-05-19 06:32:30 seems like is.gd is down 2014-05-19 06:32:47 its what we use for publishing the short url to git commits 2014-05-19 06:33:08 i wonder if i should register an app key to use goo.gl instead 2014-05-19 06:35:19 fabled: Thanks for the input. I only have a RPI (with too little resources to build anything on) 2014-05-19 06:35:29 so I couldn't fix it :-/ 2014-05-19 06:35:48 So I am in the hands of someone else to fix it - sorry 2014-05-19 07:08:03 ncopa: did you fix nfs on musl? 2014-05-19 07:08:11 and of course good morning 2014-05-19 07:08:12 clandmeter: not yet :-/ 2014-05-19 07:08:15 its on my list 2014-05-19 07:08:16 morning :) 2014-05-19 07:44:03 re 2014-05-19 07:44:05 morning 2014-05-19 11:13:47 do we have sysconf(_SC_PAGESIZE)? 2014-05-19 11:13:52 i think we do... 2014-05-19 13:00:38 was reading about Tizen, wondering if it would be nice to have it in AL 2014-05-19 13:00:57 along with qt-project.org/wiki/Tizen 2014-05-19 14:35:06 ncopa: whats the proper way to fix missing error.h? 2014-05-19 14:35:33 clandmeter: not 100% sure 2014-05-19 14:35:40 option 1: add error.h to musl 2014-05-19 14:36:06 option 2: patch app to use err.h and corresponding functions 2014-05-19 14:36:26 option 3: patch app to provide the functionality 2014-05-19 14:36:32 i think i have done option 3 a couple of times 2014-05-19 14:37:05 i saw a patch for option 2 some place recently and remember thinking "oh can you do that to..." 2014-05-19 14:39:03 i guess you dont remember where that was? 2014-05-19 14:44:49 guessing main/ipsec-tools ? 2014-05-19 14:46:34 new kamailio v4.1.3 available 2014-05-19 14:49:57 do we official page for procedure to accept hardware(donated) for testing, since we are moving towards arm ? 2014-05-19 14:50:03 do we have* 2014-05-19 14:51:04 intel SoC :) 2014-05-19 14:51:22 intel soc arm? 2014-05-19 14:51:52 I mean both 2014-05-19 14:52:26 intel has also plans for SoCs 2014-05-19 14:52:27 I think already launched 2014-05-19 14:52:43 well there is not *official* donation application or whatever. 2014-05-19 14:53:03 we are not an organisation, so everything you donate is based on person. 2014-05-19 14:53:09 http://www.anandtech.com/show/7305/intel-announces-quark-soc-a-tiny-soc-for-tiny-devices 2014-05-19 14:53:49 Hardware can be send to people who will manage the infra. 2014-05-19 14:54:57 then some official(ahem!) addresses could be published on wiki page 2014-05-19 14:55:05 ncopa: i saw you mention to fabled about error.h. you think thats difficult to add? 2014-05-19 14:55:24 vkrishn: we prefer not to (my guess) 2014-05-19 14:55:39 ok 2014-05-19 14:55:53 atleast i dont think any of the members wants to have their address on our wiki 2014-05-19 14:56:11 vkrishn: well, I guess everything goes through the alpine-devel or ncopa anyway 2014-05-19 14:56:16 usually both 2014-05-19 14:57:23 if you want to donate hardware for some reason, its better probably to dicsuss this reason our our ML 2014-05-19 14:57:32 if some interested party would like to donate then ? 2014-05-19 14:57:47 we also have an infra ML 2014-05-19 14:58:28 vkrishn: what do you mean with "some interested party" and why would they want to do that? 2014-05-19 15:00:00 any person or organization 2014-05-19 15:00:37 you are not talking about a specific case? 2014-05-19 15:02:40 I mean for fast rolling releases, it could be handy to test on hardwares and publish 2014-05-19 15:03:20 quite possible this could lead to a new department in AL 2014-05-19 15:04:28 well, let's see if anything lika that will happen 2014-05-19 15:04:32 I mostly thinking of cheap, low end SoCs 2014-05-19 15:04:50 and then worry about making someone responsible for handling this 2014-05-19 15:05:34 like DIY routers, sensors... etc 2014-05-19 15:06:03 untill now, hardware has never been an issue, manpower is rare though. 2014-05-19 15:08:18 hmm 2014-05-19 15:10:26 couple of days back I remembered by forgotten old linksys wrtgs, installed openwrt + kamailio (voip for family/neighbor 2014-05-19 15:10:52 :-), now can voip call using smartphones 2014-05-19 15:26:21 anything similar to "parprouted" 2014-05-19 15:26:22 ? 2014-05-20 06:08:12 Good morning 2014-05-20 06:11:35 I have a suggestion on 'abuild' improvement/feature: I would like some way to tell abuild to save output into some logfile. (basically same thing as 'abuild -r | tee /var/log/abuild/${pkg}.log') 2014-05-20 06:12:01 It would be great if this "logging" feature could be enabled/disabled by using ~/.abuild/abuild.conf 2014-05-20 06:13:00 I agree that this is not a high prio thing - because I can always add the 'tee' stuff to achieve what I want. 2014-05-20 06:13:32 Hope you all get a good day 2014-05-20 06:14:28 --- 2014-05-20 06:15:25 Question (sorry for bothering you with these Q's - but it's been a while since I built stuff - and things has changed since then/and I have forgotten things)... 2014-05-20 06:15:46 Om my build box - I do 'abuild -r' to build a package 2014-05-20 06:15:57 It ends up in ~/packages/... 2014-05-20 06:16:35 Now, how can I make this available to 'apk add' without needing to point to the absolute path? 2014-05-20 06:17:01 adding '~/packages' into /etc/apk/repositories doesn't seem to be a option 2014-05-20 06:17:12 What am I missing? 2014-05-20 06:35:50 mhavela: ~/packages would resolve to /root/packages 2014-05-20 06:36:08 (if its resolved that is) 2014-05-20 06:36:38 mhavela: my repositories has: /home/ncopa/packages/main 2014-05-20 06:36:59 sudo sh -c 'readlink -f ~' 2014-05-20 06:37:19 gives /root 2014-05-20 06:38:55 DOH - I had tested the /home/mhavela/packages in /etc/apk/repositories - but it complained missing APKINDEX.tar.gz 2014-05-20 06:38:55 I missed adding the '/main' in there /o\ 2014-05-20 06:39:03 Sorry for the noise! 2014-05-20 06:39:12 Adding /home/mhavela/packages/main works \o/ 2014-05-20 22:21:02 clandmeter, http://www.minnowboard.org/meet-minnowboard-max/ 2014-05-20 22:21:06 :-) 2014-05-21 06:29:29 fabled: any chance of an updated rpi image? 2014-05-21 06:29:42 rpi/arm 2014-05-21 06:33:36 i've been waiting for us to redo the image builders 2014-05-21 06:33:56 so we could make the rpi images be auto-created from tags like the .iso's are now 2014-05-21 06:34:06 ncopa, ^^^ 2014-05-21 06:34:14 in fact - that's probably the only blocker for 3.0-release 2014-05-21 06:45:35 i dunno who the arm images should be... 2014-05-21 06:45:52 or how i can test them in qemu 2014-05-21 06:46:27 should we provide disk images that you can dd? 2014-05-21 06:46:37 or just boot directly from qemu 2014-05-21 06:46:45 or do we want tar.gz images... 2014-05-21 06:46:51 i think we want tar.gz 2014-05-21 06:46:58 you need to choose bootloader based on the hardware anyway 2014-05-21 06:47:05 ok 2014-05-21 06:47:13 then it should be relatively simple... 2014-05-21 06:47:13 u-boot ships n+1 bootloaders; one per each board type 2014-05-21 06:48:00 rpi is special: no bootloader there. we just need to create specific files in the FS. so tar.gz works perfect there too. 2014-05-21 08:07:50 please git pull git://git.alpinelinux.org/fab/aports 2014-05-21 08:08:07 there are some update for existing packages 2014-05-21 08:08:53 thanks 2014-05-21 08:19:01 vkrishn) clandmeter, http://www.minnowboard.org/meet-minnowboard-max/ <-- looks nice 2014-05-21 08:45:36 pulled. thanks! 2014-05-21 09:00:26 ncopa: https://twitter.com/gregkh/status/468985929231654912 2014-05-21 09:01:31 nice! 2014-05-21 09:02:37 i just found: https://github.com/fabaff/mqtt-panel 2014-05-21 09:02:51 fabian_a: I'd like somethign similar for our build servers 2014-05-21 09:03:02 the build server status is posted as mqtt messages 2014-05-21 09:03:47 i have been thinking of using mqtt.js and websocket -> tcp socket bridge 2014-05-21 09:04:28 mqtt-panel is pretty simple but was enough for my use case 2014-05-21 09:04:44 yes, looks very simple 2014-05-21 09:04:50 what i have been looking for 2014-05-21 09:05:02 i wonder if its possible to do it without nodejs? 2014-05-21 09:05:15 that has been the most problematic part for me 2014-05-21 09:06:16 hmmm, i think lighttpd is able to handle websockets 2014-05-21 09:06:24 with a custom patch yes 2014-05-21 09:06:39 nginx has it out of the box if I'm not mistaken 2014-05-21 09:06:39 i think its possible with a plugin for uwsgi too 2014-05-21 09:06:52 no, barthalion its only proxy of websockets 2014-05-21 09:06:55 ah 2014-05-21 09:07:00 it nots a wesocket to tcp bridge 2014-05-21 09:07:22 but i wonder if i should just do it with nodejs 2014-05-21 09:07:25 looks very simple 2014-05-21 09:08:47 there are only a few lines of code because socket.io is doing the "heavy lifting" 2014-05-21 09:09:02 yeah 2014-05-21 09:09:44 ther eis a lua nodejs like thingy 2014-05-21 09:09:53 i wonder if that could be used instead 2014-05-21 09:10:04 luvit 2014-05-21 09:11:52 let me take a look at luvit.io 2014-05-21 09:12:12 we still need mqtt.js for the server thingy... 2014-05-21 09:14:29 how can I consume the mqtt messages from the build servers? 2014-05-21 09:15:30 i mean, is it possible to subscribe to those messages? 2014-05-21 09:15:42 mosquitto_sub -h msg.alpinelinux.org -t 'build/#' -v 2014-05-21 09:16:19 the topic is build/ 2014-05-21 09:16:32 messages are like: 120/225 917/1141 testing/par2cmdline 0.4-r1 2014-05-21 09:16:39 or: upgrading system 2014-05-21 09:16:49 thanks 2014-05-21 09:17:05 i pust them as retained messages 2014-05-21 09:17:15 because then you can always connect and get the current status 2014-05-21 09:38:19 ncopa: mongoose also supports websockets (never tried it though) 2014-05-21 09:38:44 cpan is down 2014-05-21 09:38:58 doh 2014-05-21 09:40:17 i think i can just copy the source archive 2014-05-21 09:40:39 unless you have other idea 2014-05-21 09:41:42 SOME DEPENDENCIES WERE MISSING. 2014-05-21 09:41:42 CORE missing dependencies: 2014-05-21 09:41:42 Mail::Header >= 2.12 ...MISSING 2014-05-21 09:42:59 mm 2014-05-21 09:43:11 not missing here 2014-05-21 09:43:31 Mail::Header >= 2.12 ...found 2014-05-21 09:43:44 is perl-mail-header in depends= 2014-05-21 09:43:45 ? 2014-05-21 09:44:04 it's not a package 2014-05-21 09:44:15 is part of perl-mail-tools i think 2014-05-21 09:44:40 mm 2014-05-21 09:45:28 i have perl-mail-tools 2.13, with 2.7 repos 2014-05-21 09:45:51 but 2.7 has actually 2.11 2014-05-21 10:06:07 barthalion: i added os-release file 2014-05-21 10:17:17 nice, thanks 2014-05-21 11:07:49 algitbot: build 2.7-stable 2014-05-21 14:08:07 maybe I should just tag rc1 2014-05-21 14:08:12 3.0.0_rc1 2014-05-21 14:29:09 ncopa, for latest realtek wifi drivers if not yet packaged can be downloaded from, 2014-05-21 14:29:16 http://218.210.127.131/downloads/downloadsView.aspx?Langid=1&PNid=21&PFid=48&Level=5&Conn=4&DownTypeID=3&GetDown=false&Downloads=true 2014-05-21 14:29:36 or github.com/pvaret/rtl8192cu-fixes 2014-05-21 14:49:23 urgh, i think we already have too many 3rd party drivers... 2014-05-21 14:58:08 the original download has some nice P2P examples too 2014-05-22 08:17:08 ncopa: maybe mosquitto-php (https://github.com/mgdm/Mosquitto-PHP/) can be an option instead of nodejs 2014-05-22 08:28:54 fabian_a: any reason py-flask is in testing? 2014-05-22 08:37:09 uggedal: no reason...so far nobody really wanted py-flask in main ;-) 2014-05-22 08:38:54 uggedal: if you have tested it and it works, then feel free to move it to main 2014-05-22 08:39:01 make sure you grab all deps too 2014-05-22 08:39:15 oh, it should have a maintainer too before going to main 2014-05-22 08:58:47 I can maintain py-flask and some of its extensions 2014-05-22 08:58:59 I use it on a pretty busy site in production 2014-05-22 09:00:00 fabian_a: mind if I take maintainership of py-flask and some of its extensions? 2014-05-22 09:07:07 sounds like a good idea to me 2014-05-22 09:30:12 any news regarding nfs? 2014-05-22 09:31:03 uggedal: not at all 2014-05-22 09:43:12 ncopa: its not clear to me why you are rejecting #2843 2014-05-22 09:43:38 clandmet1r: to set i486 arch 2014-05-22 09:43:46 you use --arch i486 2014-05-22 09:43:55 sure 2014-05-22 09:44:05 also the patch 2014-05-22 09:44:06 but your alpine template was broken 2014-05-22 09:44:12 no 2014-05-22 09:44:20 are you 100% sure? 2014-05-22 09:44:25 no :) 2014-05-22 09:44:32 but i think your patch broke it 2014-05-22 09:44:36 i486 is not a valid arch 2014-05-22 09:44:42 not for lxc 2014-05-22 09:44:47 its not? 2014-05-22 09:44:50 hm.. 2014-05-22 09:45:00 Valid options are x86, i686, x86_64, amd64 2014-05-22 09:45:23 http://manpages.ubuntu.com/manpages/saucy/man5/lxc.conf.5.html 2014-05-22 09:46:10 looks like master does have them now. 2014-05-22 09:46:44 https://github.com/lxc/lxc/commit/bb8d8207c36402ba1282081963b3dcd3356452ec 2014-05-22 09:48:19 but i think your patch is backwards 2014-05-22 09:48:37 what it does is repalce --arch i486 with x86 2014-05-22 09:48:45 and x86 will become i686 2014-05-22 09:48:50 so its not possible to set i486 2014-05-22 09:49:49 what else should it be? 2014-05-22 09:50:14 if you set i486 its not accepted 2014-05-22 09:52:31 hm 2014-05-22 09:52:34 looks like you are right 2014-05-22 09:52:53 it will use arch x86_&4 2014-05-22 09:53:46 the config will set i486, but it will not run as i486 2014-05-22 09:54:40 so lxc.arch should be x86 in config? 2014-05-22 10:01:35 clandmeter: i'd prefer this: http://sprunge.us/eBMU 2014-05-22 10:01:46 --arch x86 -> lxc.arch=x86 2014-05-22 10:01:56 --arch i486 -> lxc.arch=x86 2014-05-22 10:02:06 --arch i686 -> lxc.arch=i686 2014-05-22 10:02:53 or maybe we should have it reject the invalid i486? 2014-05-22 10:04:04 http://sprunge.us/DXYc 2014-05-22 10:12:37 ncopa: cant we use https://github.com/lxc/lxc/commit/bb8d8207c36402ba1282081963b3dcd3356452ec ? 2014-05-22 11:13:25 clandmeter: yeah, thats better 2014-05-22 11:26:06 StarWarsFan: not had time for nfs yet sorry :-( 2014-05-22 11:59:43 was there any eforts to make iso a hybrid one? (to be able to dd to USB Thumdrive) and boot from it 2014-05-22 12:03:54 I dont think so 2014-05-22 12:10:03 ncopa: ok 2014-05-22 12:10:12 was more a reminder... ;-) 2014-05-22 12:34:25 clandmeter over pxe it should work? 2014-05-22 12:34:41 should be 2014-05-22 12:34:44 but 2014-05-22 12:34:48 nfs is broken 2014-05-22 12:35:35 im also waiting for nfs. we have to make ncopa happy and hope he can fix it :) 2014-05-22 12:36:22 crow: making a bootable usb is pretty easy 2014-05-22 12:37:16 ncopa i will give you such app :) http://youtu.be/1hpU_Neg1KA 2014-05-22 12:38:19 clandmeter you mean with unetbootin ? 2014-05-22 12:46:01 no just copy contents to fat32 usb 2014-05-22 12:46:15 and install syslinux bootloader 2014-05-22 12:46:31 you can unpack the .iso with 7zip 2014-05-22 12:46:53 syslinux used to provide a tool named isohybrid 2014-05-22 12:47:03 barthalion: i looked at it 2014-05-22 12:47:20 the problem is that vfat cannot be resized 2014-05-22 12:47:47 so there would be a fixed amount of space, regardless the size of usb 2014-05-22 12:48:11 we want be able to store the configs on the usb 2014-05-22 12:50:09 files from v2.5.4-293-gcef8b69 uploaded 2014-05-22 13:14:44 algitbot: build master 2014-05-22 13:17:31 algitbot: build master 2014-05-22 13:26:50 ncopa: nice! 2014-05-22 13:27:05 almost time for a party 2014-05-22 13:29:14 yeah 2014-05-22 13:29:30 i should probably look at nfs again :-/ 2014-05-22 13:29:38 its non-trivial 2014-05-22 13:31:15 code is crap? 2014-05-22 13:31:49 yeah... 2014-05-22 13:32:16 i only use it on the pi, cause of logical reasons. 2014-05-22 13:34:44 i'm using nfs on our build-boxes, so without nfs, no edge-builds are possible 2014-05-22 13:42:42 i intend to use it on our official buildboxes too 2014-05-22 13:42:45 for sharing the distfiles 2014-05-22 13:42:48 so we need it 2014-05-22 14:04:46 ncopa: is it totally borked or just server or client side? 2014-05-22 14:05:25 server is broken 2014-05-22 14:05:30 i havent tested client 2014-05-22 14:05:41 i'd kinda expect client work 2014-05-22 14:07:21 i remember locking was not possible 2014-05-22 14:10:52 http://imgur.com/YA15IJf 2014-05-22 14:11:25 thats from tobias 2014-05-22 14:11:36 this too: http://imgur.com/u6HwAry 2014-05-22 14:13:02 my build-boxes are nfs-clients 2014-05-22 14:13:11 the server is running on ubuntu at the moment 2014-05-22 14:14:50 ncopa: nice 2014-05-22 14:15:05 StarWarsFan: im able to mount nfs share 2014-05-22 14:15:22 x86_64? 2014-05-22 14:15:39 arm 2014-05-22 14:15:44 ah ok 2014-05-22 14:15:56 edge-x86 is working 2014-05-22 14:16:03 edge-x86_64 is broken 2014-05-22 14:16:19 but even on x86 there are a lot of messages on dmesg 2014-05-22 14:16:34 i dont even know what the minium configuration is for nfs client 2014-05-22 14:17:01 i only started rpc.statd 2014-05-22 14:30:27 the logos look nice, but can there be a mockup webpage with exact image size of logo that would like be put on actual pages ? 2014-05-22 14:30:41 that would likely* 2014-05-22 14:32:20 I usually make use of a wiki for quick mockup of pages, that can also be edited by others 2014-05-22 14:32:26 example pmwiki 2014-05-22 15:11:37 I could set up pmwiki with basic layout template for editing, but would need logo sized image 2014-05-24 15:07:03 hi all. I've a (stupid) problem between lxc host <-> guest. Two interfaces macvlan with bridge doesn't talk each other. From the host, i'm able to dump packets. 2014-05-24 15:07:21 i think that is a stupid problem, but still i'm not able to figure out 2014-05-24 16:59:03 Umh..that's why: eth1@if10: 2014-05-24 16:59:05 M-DOWN 2014-05-24 16:59:14 how can set it up? 2014-05-25 21:05:38 hi@all 2014-05-25 21:05:46 is there something wrong with the edge repo? 2014-05-25 21:05:51 Installing mysql-common (5.5.36-r1) 2014-05-25 21:05:57 ERROR: mysql-common-5.5.36-r1: BAD signature 2014-05-25 21:23:34 I can trigger the rebuild 2014-05-25 21:24:01 it's not the first time when signature is bad without good reason 2014-05-25 21:32:08 StarWarsFan: ↑ 2014-05-25 21:36:22 ncopa: I'll probably shift more towards infra guy here and look for a way to better maintain mirrors 2014-05-25 21:46:58 barthalion: the pkg was broken on nl.a.o? 2014-05-26 05:40:01 anyone opposed to using musl malloc in stead of jemalloc for testing/redis and move testing/jemalloc to unmaintained? 2014-05-26 05:55:31 hmm, nevermind. redis-cli segfaults and I was under the wrong impression that it was jemalloc's fault 2014-05-26 06:03:21 barthalion: thx a lot, try'n again... 2014-05-26 06:03:34 moin btw. :-) 2014-05-26 06:08:13 morning 2014-05-26 06:09:04 barthalion: i'd be happy to have you around, even if your're moving to infra 2014-05-26 06:09:36 uggedal: we replace jemalloc with musl malloc in firefox/xulrunner 2014-05-26 06:09:54 what i have heard, musl malloc is competitive with jemalloc 2014-05-26 06:34:14 ncopa: some days ago you asked for open-vm-tools 2014-05-26 06:34:28 but actually it's not on the edge repo 2014-05-26 06:34:33 any problems there? 2014-05-26 06:36:04 yes, it does not build with musl 2014-05-26 06:39:13 hm 2014-05-26 06:41:25 i also spent some hours on friday trying to figure out nfs 2014-05-26 06:41:34 but i didnt find anything 2014-05-26 06:41:39 well, i fond another segfault 2014-05-26 06:41:58 but i dont have any idea really what/where/why things goes wrong 2014-05-26 06:42:04 ACTION send's mental support to ncopa ;-) 2014-05-26 06:43:11 thanks... 2014-05-26 07:44:00 ncopa: yeah, saw that musl malloc was used in xulrunner. only thing talking in favor for jemalloc for redis is that musl does not have malloc_size 2014-05-26 07:44:26 when libc malloc does not have malloc_size redis pads all allocations with its own size accounting info 2014-05-26 07:44:45 which in theory could mean more fragmantation when allocating small objects 2014-05-26 07:51:26 i'd suggest you mention it in #musl 2014-05-26 08:03:45 hm, btw, what doesn't build on musl currently? 2014-05-26 08:45:24 ncopa: it would be nice to have "lastsync" file to easily check if mirror is up to date 2014-05-26 08:51:24 do you have any example how that is imlpemented? 2014-05-26 08:52:47 we (as in Arch) just echo unix time to such file when files are copied to the main rsync server 2014-05-26 09:18:53 git pull git://git.alpinelinux.org/fab/aports 2014-05-26 09:19:02 there are a couple of updates 2014-05-26 09:21:02 thanks 2014-05-26 09:52:07 files from v2.5.4-296-gf1580d6 uploaded 2014-05-26 12:40:44 ncopa: im seeing lots of locale warnings with perl when setting up a new container (ubuntu in this case) 2014-05-26 12:40:52 is there a way to fix this? 2014-05-26 12:45:35 ubuntu container on alpine linux host or opposite? 2014-05-26 12:56:04 ncopa: ubuntu in container 2014-05-26 12:56:20 2.7 is the host 2014-05-26 12:56:43 ncopa: http://sprunge.us/JDeS 2014-05-26 12:59:51 hm 2014-05-26 12:59:52 dunno 2014-05-26 13:23:46 clandmeter: try running with LANG=en_US.UTF-8 lxc whatever 2014-05-26 13:24:08 or: export LANG=en_US.UTF-8 2014-05-26 13:24:31 i think in musl its set to C.UTF-8 2014-05-26 13:24:52 ok, but what is the real solution? 2014-05-26 13:24:53 its en.utf8 here on musl 2014-05-26 13:25:21 clandmeter: either patch the ubuntu template or set LANG before creating the ubuntu container 2014-05-26 13:25:45 shouldnt we have that by default? 2014-05-26 13:26:10 this line sets locale in the container to that of $LANG on the host: https://github.com/lxc/lxc/blob/master/templates/lxc-ubuntu.in#L106 2014-05-26 13:28:15 oh ok, the error is from within the chroot. 2014-05-26 13:29:30 yup 2014-05-26 14:35:03 ok to disable wtmp in openssh to fix bug #2744 ? 2014-05-26 14:37:04 seems musl does not implement wtmp, but defines WTMP_FILE to /dev/null/wtmp and openssh configure enables it 2014-05-26 14:37:33 I'd rather see it fixed in musl/openssh 2014-05-26 14:40:27 well, wtmp is not standardized 2014-05-26 14:43:09 just like a lot of other things used by software we provide 2014-05-26 14:53:16 so openssh trying to write to /dev/null is better than disabling the feature ;) 2014-05-26 14:54:08 i think aboriginal also disabled wtmp 2014-05-26 14:54:45 no, it's just better to get it reported somewhere 2014-05-26 14:54:55 and then disable it if it's needed 2014-05-26 16:51:50 ACTION is verifying ACF on 3.0_rc1 and found something weird 2014-05-26 16:52:33 with 3.0, when you run "/etc/init.d/xxxx describe" you now get an extra line stating "* cgroup_cleanup: no description" 2014-05-26 16:52:50 what is this new "cgroup_cleanup" command in rc? 2014-05-26 16:54:58 well, seems it kills all of the processes in the service's cgroup 2014-05-26 16:55:05 ACTION answered his own question 2014-05-26 16:55:35 just makes the ACF rc status page even busier :/ 2014-05-26 18:23:17 ACTION wonders if the release tarballs stopped working 2014-05-26 18:23:34 or if they need to be manually enabled for new git repos 2014-05-26 18:23:50 ACTION pushed tags to acf-jquery-mobile and acf-skins-jqmobile 2014-05-26 18:23:56 but no new tarballs created 2014-05-26 18:38:31 ncopa: ↑ 2014-05-26 18:41:44 looks like the "post-receive" hook was not added to these archives :( 2014-05-26 18:43:57 copied in the hook, now trying to generate the tarball manually 2014-05-26 19:11:39 boy, that was a pain 2014-05-26 19:12:01 ACTION missed a '/' when creating the tarball and had to do it again 2014-05-26 19:12:28 hopefully the hook will work in the future even though the ownership is slightly different 2014-05-26 19:17:08 vkrishn: ^^ I pushed the jqmobile skin from two months ago to testing 2014-05-26 19:17:26 still want to get it all working, but we're running out of time for alpine 3.0 2014-05-26 19:19:31 any idea why there is no ruby-minitest directory in git? 2014-05-26 19:19:47 though, there is http://nl.alpinelinux.org/alpine/edge/main/x86_64/ruby-minitest-2.0.0_p353-r2.apk 2014-05-26 19:21:39 hamletmun: it's a subpackage of ruby 2014-05-26 19:23:08 http://git.alpinelinux.org/cgit/aports/tree/main 2014-05-26 19:24:26 I can't find ruby-minitest directory with apkbuild for minitest ruby gem 2014-05-26 19:24:54 subpackages are split from the base one 2014-05-26 19:25:08 the base package for ruby-minitest is ruby 2014-05-26 19:25:11 http://git.alpinelinux.org/cgit/aports/tree/main/ruby/APKBUILD 2014-05-26 19:25:47 oh... 2014-05-26 19:27:03 if I want to use different version of minitest gem, what would be the best? 2014-05-26 19:28:10 create ruby-minitest or ruby-gollum_rails-minitest? 2014-05-26 19:29:30 probably the latter 2014-05-26 19:29:46 (btw, now I understand why minitest package shows 2.0.0_p353, instead of gem version, 4.3.2) 2014-05-26 19:29:58 thank you, barthalion 2014-05-26 19:30:30 provide= allows to overwrite files installed by other package 2014-05-26 19:30:54 err, replaces= 2014-05-26 19:30:59 but probably provides= will do 2014-05-26 19:31:04 :-D 2014-05-26 19:31:20 I'm not really into packaging here lately 2014-05-26 19:32:56 but I see that minitest gem doesn't conflict with ruby's minitest, so it will be ok without any magic 2014-05-26 19:33:28 do replaces/provides work with subpackages like minitest? 2014-05-26 19:34:57 depends what do you want to achieve, but generally speaking yes 2014-05-26 21:08:40 tdtrask, nice, thanks 2014-05-26 21:08:56 got busy with new hardware and my web related work 2014-05-26 21:09:27 would try get to acf testing 2014-05-26 21:21:11 vkrishn: believe me, I understand being busy :) 2014-05-26 23:54:03 ncopa: can you please move the acf-unbound repository from my home directory to gitroot? thanks. 2014-05-26 23:58:02 I set up the symlink for post-receive, but I could not set the ownership properly 2014-05-26 23:58:07 so, please check that too 2014-05-27 06:13:00 morning 2014-05-27 06:13:07 tdtrask: should be done now 2014-05-27 06:14:14 i think I will upgrade the git.a.o box to v3.0 2014-05-27 06:17:50 heh 2014-05-27 06:18:07 i upgraded git.a.o to v3.0 between uggedals commits 2014-05-27 06:18:37 which is kinda cool, because its practically a reinstall due to switch to musl 2014-05-27 06:20:13 what happened with the pkg move? 2014-05-27 06:20:24 it failed? 2014-05-27 06:20:45 I never remember where the build logs are... :) 2014-05-27 06:21:18 http://git.alpinelinux.org/cgit/aports/commit/?id=3d22bbc8b283ae942136be5aef653bbb5cb08c8e 2014-05-27 06:21:23 it made it to the git atleast 2014-05-27 06:21:30 i assume it worked 2014-05-27 06:21:40 http://bld1.alpinelinux.org/buildlogs/ 2014-05-27 06:21:43 for x86 2014-05-27 06:21:46 http://bld2.alpinelinux.org/buildlogs/ 2014-05-27 06:21:49 for x86_64 2014-05-27 06:23:33 hm 2014-05-27 06:23:36 yeah... 2014-05-27 06:23:40 what happened to it.. 2014-05-27 06:27:24 hm 2014-05-27 06:28:53 ncopa: can we have a central point for logs? 2014-05-27 06:29:10 yes, when I have set up nfs 2014-05-27 06:29:11 or doest hat depepnd on nfs? 2014-05-27 06:29:14 :) 2014-05-27 06:29:28 there are 2 different build server boxes 2014-05-27 06:29:41 3 i guess? 2014-05-27 06:29:48 3 with arm yes 2014-05-27 06:30:03 looks like mqtt message is not sent from git.a.o now 2014-05-27 06:30:09 or builders does not pick up... 2014-05-27 06:30:39 if we can centralise it, i can add that jslogtail again and add some menu to switch logs. 2014-05-27 06:32:21 ncopa: what do you think about build failure notifications? 2014-05-27 06:32:36 fatal: unable to connect to git.alpinelinux.org: 2014-05-27 06:32:36 git.alpinelinux.org[0: 91.220.88.36]: errno=Connection refused 2014-05-27 06:32:39 ha 2014-05-27 06:35:29 git-daemon is separate package now... 2014-05-27 06:35:43 so the git-daemon was gone... 2014-05-27 06:36:55 here we go... 2014-05-27 06:37:01 sorry about that uggedal 2014-05-27 06:37:14 py-wtforms is there now 2014-05-27 06:42:58 clandmeter: re build failure notifications 2014-05-27 06:43:01 i'm open to ideas 2014-05-27 06:43:07 how do we want it to work? 2014-05-27 06:43:17 we have 2 options iguess 2014-05-27 06:43:19 what we have now is per-package build logs 2014-05-27 06:43:30 so the notification should probably tell where the build log is 2014-05-27 06:43:44 smtp or jabber? 2014-05-27 06:44:24 mqtt? 2014-05-27 06:44:55 how do i get a notification? 2014-05-27 06:44:55 ncopa: thanks 2014-05-27 06:44:58 ok lets say smtp, to what email do we send the notificaion? 2014-05-27 06:45:02 uggedal: np 2014-05-27 06:45:11 i guess commiter 2014-05-27 06:45:31 and the author of given commit 2014-05-27 06:45:39 of given package 2014-05-27 06:45:40 ues 2014-05-27 06:45:41 yes 2014-05-27 06:45:43 (if it's different) 2014-05-27 06:45:50 makes sense 2014-05-27 06:47:20 i think i'd be happy with notification here in irc 2014-05-27 06:47:38 but email makes sense 2014-05-27 06:47:58 and it could include the build log 2014-05-27 06:48:22 yes, or just a link if its still on disk. 2014-05-27 06:48:38 we save all build logs now 2014-05-27 06:48:51 i havent done any cleanup cron job yet 2014-05-27 06:49:13 658.7M /var/cache/distfiles/buildlogs/ 2014-05-27 06:49:25 but looks like we need one soonish... 2014-05-27 06:49:33 or gzip them? 2014-05-27 06:50:30 would be nice if we could save the log untill its gets rebuild. 2014-05-27 06:51:01 yes 2014-05-27 06:51:03 that we do 2014-05-27 06:51:16 what we need is something that removes old logs 2014-05-27 06:51:28 package foo-1.0 etst upgraded to foo-1.1 2014-05-27 06:51:35 then we should remove foo-1.0 build log 2014-05-27 06:51:37 can you save them as pkgname? 2014-05-27 06:51:45 thats what i we do now 2014-05-27 06:51:53 pkgname-version 2014-05-27 06:53:05 oh you mean that we skip the version number? 2014-05-27 06:53:06 hm 2014-05-27 06:53:20 then we can only keep one generation of the log 2014-05-27 06:53:22 yes, or a hook to clean it per commit 2014-05-27 06:53:45 i'd rather do cleaning 2014-05-27 06:54:27 i'll do logs later 2014-05-27 06:54:36 i think i want build server status webpage first 2014-05-27 07:07:34 re 2014-05-27 07:07:39 morning@all 2014-05-27 07:08:16 morning 2014-05-27 14:31:44 Guys, who among you have one sec to explain me how macvlan bridge lxc work? If i have a bonded interface on the host, and the container in his config has lxc.network.link=$mybondedinterface, why doesn't the host<->guest talk each other? 2014-05-27 14:33:28 fcolista: it looks like you need create a macvlan interface on the host: http://noyaudolive.net/2012/05/09/lxc-and-macvlan-host-to-guest-connection/ 2014-05-27 14:35:24 so 2014-05-27 14:35:37 i've a bonded interface on the host 2014-05-27 14:35:42 let's say bond0.100 2014-05-27 14:35:55 then macvlan bridge interface, macvlan100 2014-05-27 14:36:05 the ip address should go to macvlan100 2014-05-27 14:36:09 on the host 2014-05-27 14:36:27 and then, in the guest, lxc.network.link=macvlan100 2014-05-27 14:37:46 i'm wondering what's the advantage of macvlan bridge VS bridge. "Just" to allow the evantual lxc containers to talk each other 2014-05-27 14:42:09 Also, when is used: lxc.network.ipv4 = $IP_ADDRESS ? Why should we use that rather than configure /etc/network/interface ? 2014-05-28 06:59:22 fabled: is it possible for you to spin an rpi image with recent kernel? 2014-05-28 06:59:43 I want to install a DAC, but current image doesnt have the driver. 2014-05-28 07:00:04 i'm busy for the morning. if i get everything done, i can take a look at it afternoon. 2014-05-28 07:00:12 define "recent" kernel 2014-05-28 07:00:30 did you modify the kernel for rpi with your latest release? 2014-05-28 07:00:36 i mean the config 2014-05-28 07:01:12 or just use rpi's default kernel conf? 2014-05-28 07:02:55 The IQaudIO Pi-DAC’s drivers are now built into the Raspberry Pi linux (3.10.y / 3.14.y) available on github/raspberrypi or for our own code changes / documents etc. check our own github/iqaudio page. 2014-05-28 07:03:00 mmm.. can't remember 2014-05-28 07:03:22 i probably took rpi defconfig long ago; and then just upgraded without doing rpi defconfig anymore 2014-05-28 07:03:34 should probably take latest 3.14.x and reset defconfig 2014-05-28 07:04:39 I guess you build the kernel on your wanboard (because of initrd)? 2014-05-28 07:53:13 hello 2014-05-28 07:53:35 trying to create a alpine domU 2014-05-28 07:57:04 I'm following to domU wiki on alpinelinux 2014-05-28 07:57:43 when booting the iso, i end up without a /lib/modules directory 2014-05-28 07:58:17 have you seen this before? alpine linux 2.7.8 full iso 2014-05-28 08:00:23 sounds like it does not find the modloop image 2014-05-28 08:02:22 i am running alpine-xen as my dom0 and that one had /.modloop mounted 2014-05-28 08:02:43 can i mount it manually in my domU? 2014-05-28 08:10:48 i'll try with alpine-mini 2014-05-28 08:16:19 same with alpine-mini 2014-05-28 08:23:21 it was a typo! apologies for the stupidity 2014-05-28 08:57:00 so, now that I have my domU working, and setup-alpine installed onto a device (xvda), I'm strugging to boot it. should I still be using pygrub bootloader? do i need kernel and ramdisk and extra values in my alpine xl file? 2014-05-28 08:58:06 ncopa: have you seen strange behaviour of cron on musl? 2014-05-28 09:01:26 ScrumpyJack: I don't really use xen myself so i dont know really 2014-05-28 09:01:39 would be nice if you share what you figure out 2014-05-28 09:01:43 maybe update the wiki? 2014-05-28 09:01:47 clandmeter: no? 2014-05-28 09:02:04 ncopa: http://sprunge.us/TKcO 2014-05-28 09:02:07 thats my cron file 2014-05-28 09:02:17 but the first custom line (at 8) runs at 6 2014-05-28 09:03:30 May 28 06:00:01 mpd cron.info crond[1627]: crond: USER root pid 17329 cmd mpc clear && mpc load Radio538 && mpc play && mpc volume 80 2014-05-28 09:03:40 sounds like it is running at UTV time 2014-05-28 09:03:43 UTC time 2014-05-28 09:04:02 how do you set your timezone? 2014-05-28 09:04:20 with your script 2014-05-28 09:04:36 i think its running ok now that i check it 2014-05-28 09:04:39 it just logs faulty 2014-05-28 09:05:01 log is probably just writing time as UTC 2014-05-28 09:05:05 which i think is ok 2014-05-28 09:06:34 well thats not happening on uclibc boxes afaik 2014-05-28 09:06:47 it uses the tz time 2014-05-28 09:06:48 maybe restart syslog? 2014-05-28 09:08:56 ncopa: its weird, when i restart it shows correct time, but it also restarts cron which shows utc after the restart http://sprunge.us/DaZL 2014-05-28 09:09:47 lol 2014-05-28 09:09:53 thats weird... :) 2014-05-28 09:11:01 i think its openrc that is logging that syslog restart 2014-05-28 09:11:33 while crond is likely writing itself 2014-05-28 09:11:45 so its to find out how they write to log, and how it differs 2014-05-28 09:12:00 if i use logger its also utc 2014-05-28 09:12:16 so its maybe something busybox des 2014-05-28 09:12:18 does* 2014-05-28 09:12:29 and if you reboot the machine? 2014-05-28 09:13:16 i think ive done that before. (its our radio so i prefer not to reboot it now) :) 2014-05-28 09:16:30 ok... i dont know why it behaves like that 2014-05-28 09:16:36 maybe in #musl they know? 2014-05-28 11:41:16 ncopa: will a commit without changes to pkgver and pkgrel build? 2014-05-28 11:41:43 I don't think so 2014-05-28 11:42:07 basically I want to change the maintainer comment 2014-05-28 11:42:35 I guess that nothing will happen 2014-05-28 11:42:35 it will not rebuild anything 2014-05-28 11:42:43 but you actually need bump pkgrel 2014-05-28 11:42:54 because the apkbuild format is a bit stupid 2014-05-28 11:43:05 the comment about maintainer will actually go into the apk 2014-05-28 11:43:20 hmm, I don't need me listed as maintainer in the apk until a potential upgrade in the future 2014-05-28 11:43:32 maintainer should not been a comment. it should been a variable. eg: maintainer="...." 2014-05-28 11:43:51 just for easier tracking of what packages I maintain (find -name APKBUILD | xargs fgrep -l Uggedal) 2014-05-28 11:44:06 i do something similar 2014-05-28 11:44:21 I guess your list is huge :) 2014-05-28 11:44:30 :D 2014-05-28 11:45:08 1349 actually... 2014-05-28 11:45:21 ncdev-edge-x86_64:~/aports$ grep 'Maintainer.*Natanael' main/*/APKBUILD testing/ 2014-05-28 11:45:21 */APKBUILD | wc -l 2014-05-28 11:45:21 1349 2014-05-28 11:45:21 ' 2014-05-28 11:45:34 :) 2014-05-28 11:45:47 :) 2014-05-28 11:47:43 ncopa: I guess you won't mind me maintaining redis then 2014-05-28 11:47:53 please go ahead :) 2014-05-28 11:48:07 and a big thanks for doing so 2014-05-28 11:48:39 what 2014-05-28 12:44:00 keymaps doesnt seem to start up on boot. is that me? 2014-05-28 12:44:21 i can start it manually after boot finishes 2014-05-28 12:46:30 rc-update add keymaps ? 2014-05-28 12:47:15 no matter where i put it in openrc, it doesn't start 2014-05-28 12:47:48 rc-update tells me it should boot in "default" 2014-05-28 12:47:52 xen domu? 2014-05-28 12:47:59 dom0 2014-05-28 12:48:07 keyword noopenvz noprefix nouml novserver noxenu 2014-05-28 12:48:40 ah 2014-05-28 12:48:47 when you boot, early, doesnt openrc dectect the dom0 as domu? 2014-05-28 12:48:57 so noxenu would stop it? 2014-05-28 12:49:05 what i think 2014-05-28 12:49:14 before the xen fs is mounted 2014-05-28 12:49:19 it is detected as domu 2014-05-28 12:49:35 and noxenu tells openrc, this should not run in xenu 2014-05-28 12:49:48 so if you set rc_sys in /etc/rc.conf 2014-05-28 12:50:01 to xen0 or what openrc calls it 2014-05-28 12:50:15 yeah 2014-05-28 12:50:21 rc_sys="xen0" 2014-05-28 12:50:34 will tell openrc that it is xen0 2014-05-28 12:50:38 i think that should work 2014-05-28 12:50:39 can i edit init.d/keymaps? (boot from usb stick) 2014-05-28 12:50:50 i think not 2014-05-28 12:51:06 i think it will be overwritten when it extracts the apks 2014-05-28 12:51:27 thats what it looked like when i tried earlier 2014-05-28 12:51:28 but the real problem is that openrc thinks you are a domu and not a dom0 2014-05-28 12:51:54 so i think the fix is to manually tell openrc that you are a dom0 2014-05-28 12:52:23 how can i tell what rc_sys I am in now? 2014-05-28 12:53:19 rc --sys 2014-05-28 12:53:58 once the xen services has started and xen filesystem is mounted, openrc will correctly detect dom0 2014-05-28 12:54:11 but i think i have seen domU in early boot message 2014-05-28 12:54:43 just looking 2014-05-28 12:56:47 no mention of domU 2014-05-28 12:57:02 anyway,setting rc_sys worked, thanks 2014-05-28 12:59:03 no i just need to figure out how to use LVM with XEN :) 2014-05-28 13:35:29 clandmeter, sorry, i didn't get time for rpi. i'll try to update it tm. 2014-05-28 14:08:02 files from v2.5.4-297-gfb15cd2 uploaded 2014-05-28 14:10:53 files from v2.5.4-297-gfb15cd2 uploaded 2014-05-28 14:36:29 really struggling with alpine-xen. does the person who maintain it hang out here? 2014-05-28 14:36:58 ScrumpyJack: kaniini is maintaining it afaik 2014-05-28 14:37:58 thanks, I'll look out for him/her 2014-05-28 14:38:27 might be royger can help you too 2014-05-28 15:17:26 fabled: no problem. 2014-05-28 15:58:57 ok, trying to install rc2 :-) 2014-05-28 15:59:32 on "acer aspire one" 2014-05-28 16:00:55 first try with the mini-edition failed 2014-05-28 16:01:03 the box did not boot after sys-installation 2014-05-28 16:01:16 message was something like "warning: no configuration file" 2014-05-28 16:01:28 and then just the boot prompt 2014-05-28 16:01:42 no trying the the full rc2 2014-05-28 16:02:15 -no +now 2014-05-28 16:02:26 hm, same message: 2014-05-28 16:02:33 SYSLINUX ... 2014-05-28 16:02:45 WARNING: No configuration file found 2014-05-28 16:02:48 boot: 2014-05-28 16:02:56 with the blinking cursor at the end 2014-05-28 16:03:00 any ideas? 2014-05-28 16:11:07 ok, if i boot the install-disk again 2014-05-28 16:11:18 and mount the used partitions somewhere, 2014-05-28 16:11:45 then the /boot on the disk on the box contains only two files: 2014-05-28 16:11:52 ldlinux.c32 2014-05-28 16:11:56 ldlinux.sys 2014-05-28 16:12:03 and the folder "lost+found" 2014-05-28 16:12:08 that's it 2014-05-28 18:52:52 ScrumpyJack: what is the problem you are having? 2014-05-28 21:34:47 n8@all 2014-05-29 06:01:12 morning 2014-05-29 06:03:37 what's the best way to have a mount point preserved across reboots in a USB based install with lbu saves to the USB stick? I want to keep /a for example, but if I include /a it will save to the apkovl /a and all it's contents no? 2014-05-29 06:13:20 ScrumpyJack, yes, just add the mount point in /etc/fstab 2014-05-29 06:20:19 but will that actually create the mount point? 2014-05-29 06:20:59 oh, you mean you need the empty directory to appear 2014-05-29 06:21:02 mmm 2014-05-29 06:21:05 yes 2014-05-29 06:21:10 include , exclude /* 2014-05-29 06:22:08 ok, cool. makes sense 2014-05-29 06:22:40 what does lbu look at? /etc/lbu/include? 2014-05-29 06:24:25 ScrumpyJack, in new(ish) alpines the list is at /etc/apk/protected_paths.d/lbu.list 2014-05-29 06:26:07 got it. the globbing goes in there. thanks 2014-05-29 06:27:30 yes, the globbing and auditing is built-in into apk-tools 2014-05-29 06:28:04 lbu is more of a front end that manages the lists, and constructs the apkovl from apk-tools' filelist 2014-05-29 06:36:09 morning 2014-05-29 06:36:21 ScrumpyJack: it should be enough to just add it to /etc/fstab 2014-05-29 06:36:30 it will create all dirs there if missing 2014-05-29 06:36:55 really? ok, i'll try both 2014-05-29 06:41:17 i'm stuggling to reboot an alpine domU once it's been installed (setup-alpine). I'm not sure what the xl file should look like post install. does anyone have one i can see? 2014-05-29 07:17:23 http://paste2.org/IP4eF63L 2014-05-29 07:26:59 i just can't seem to get alpine to mount its boot media that it created inside the image 2014-05-29 07:45:49 Moinmoin 2014-05-29 08:17:51 ScrumpyJack: i think you should set root=/dev/xvda too 2014-05-29 08:18:07 extra="modules=sd-mod,usb-storage,ext4 console=hvc0 root=/dev/xvda" 2014-05-29 08:18:10 or similar 2014-05-29 08:40:36 ah ok, I'll try that. 2014-05-29 08:41:02 any idea on how to investigate this issue? 2014-05-29 08:41:03 daemon.warn bgpd[2238]: stream_read_try: read failed on fd 16: Connection reset by peer 2014-05-29 08:41:29 if i restart bgp error disappears and routes are announcend correctly 2014-05-29 08:42:17 it has error on fd 16, but from /proc/$PID/task/$PDI/fd there's no fd 16 2014-05-29 08:42:30 maybe is 15 because startes from 0 2014-05-29 08:54:18 still can't mount root :( 2014-05-29 09:00:21 got it finally! 2014-05-29 09:01:39 i had to copy the kernel and ramdisk out of the domU and into the dom0. is that standard? 2014-05-29 09:06:52 i suppose you could apk add the kernel within the domu 2014-05-29 09:06:59 and use that instead 2014-05-29 09:12:21 ncopa: im not sure i get that. 2014-05-29 09:15:21 http://paste2.org/IP4eF63L 2014-05-29 09:15:31 You appear to use the kernel on the host 2014-05-29 09:15:35 to boot the domu 2014-05-29 09:15:54 i think it should be possible to make xen boot the kernel inside the domu image 2014-05-29 09:15:58 it's the kernal from the iso that was used to build the domU 2014-05-29 09:16:09 yes, i would really prefer that 2014-05-29 09:16:27 and i don't want to use pygrub (which doesn't seem to work anyway) 2014-05-29 09:17:13 i wonder if its possible with pvgrub? 2014-05-29 09:17:21 i dont really know how that is done 2014-05-29 09:21:19 ok, now i have a alpine-linux dom0 and an alpine-linux domU. time for a debian and sles domU! thanks for all your help! 2014-05-29 09:36:58 ncopa: tried the same as yesterday on vmware workstation 2014-05-29 09:37:02 but got the same result 2014-05-29 09:37:08 rc3 is not booting 2014-05-29 09:37:43 i wonder if i'm doing something wrong... 2014-05-29 09:38:04 i mean rc2 of course... 2014-05-29 10:31:12 ncopa: i can confirm, setup-alpine has issues with boot. 2014-05-29 10:31:19 StarWarsFan: ^ 2014-05-29 10:35:24 ah ok, thx 2014-05-29 10:37:32 it seems setup-alpine doesnt install any packages 2014-05-29 10:37:51 except baselayout or something 2014-05-29 10:38:48 fabled: are you setting up auto releases for rpi? 2014-05-29 10:40:30 clandmeter, i'm waiting ncopa to do the x86/x86_64 release rework first 2014-05-29 10:40:37 i'm doing onetime rpi image now, though 2014-05-29 10:50:58 clandmeter, try http://dev.alpinelinux.org/~tteras/alpine-edge-140529-armhf.rpi.tar.gz 2014-05-29 10:51:15 and yes, i hope we get it automated real soon now (tm) 2014-05-29 12:48:33 fabled: nice, ill try it tomorrow when i got access to my rpi 2014-05-29 12:48:50 holidays around here 2014-05-29 13:41:15 fabled: we have make img 2014-05-29 13:41:19 in alpine-iso 2014-05-29 13:41:25 dunno if that helps for arm 2014-05-29 13:52:51 not much 2014-05-29 13:53:05 i do have patch for the makefile to make rpi images - it's not polished though 2014-05-29 13:53:19 and did not build everything right after recent changes - i had to fix one thing manually 2014-05-29 13:53:31 doing wandboard/beaglebone/other images will be trickier 2014-05-29 15:08:45 ok, so let's try rc3 now... 2014-05-29 15:19:42 great, vmware test vm boots now :-) 2014-05-29 15:19:58 trying bare metal test later, some other work to do now... 2014-05-29 15:24:40 any idea on what can i check when BGP returns "Bad Peer AS (2)" ? If BGP is restarted, routes comes up. BGP AS is correct. RFC states "If the Autonomous System field of the OPEN message is unacceptable, 2014-05-29 15:24:40 then the Error Subcode MUST be set to Bad Peer AS. The determination 2014-05-29 15:24:40 of acceptable Autonomous System numbers is outside the scope of this 2014-05-29 15:24:40 protocol." 2014-05-29 15:25:09 other routers has the same AS and works without problems. 2014-05-29 15:36:02 for info, i now have alpine-xen dom0, alpine domU and debian domU but not managed to boot them using pygrub. to boot them, had to copy their respective kernels and ramdisks into the dom0. I'll have to find a way to boot those without pygrub of copying files to dom0 2014-05-30 06:00:09 re 2014-05-30 06:00:10 moin 2014-05-30 07:07:44 morning 2014-05-30 07:11:07 morning 2014-05-30 07:34:18 i've added gnu tar and tar-doc from the apk repros and i got get a man page for tar. 2014-05-30 07:34:23 I'll poke around 2014-05-30 07:35:54 ncopa: around? 2014-05-30 07:36:10 morning 2014-05-30 07:36:13 morning 2014-05-30 07:36:14 clandmeter: hi 2014-05-30 07:36:25 should mdev start after modloop? 2014-05-30 07:36:45 thats a good question 2014-05-30 07:37:02 i just started new rpi image 2014-05-30 07:37:16 and mdev init complains it cannot cd to lib/modules 2014-05-30 07:37:22 looks like no 2014-05-30 07:37:28 need dev 2014-05-30 07:38:23 tar-doc don't contain man pages, only info pages 2014-05-30 07:38:33 after starting mdev, it says modprobe: cannot cd to lib/modules 2014-05-30 07:38:54 hm 2014-05-30 07:39:23 i suppose there a stuff on USB that triggers this 2014-05-30 07:39:28 i excited about the arm port too :) 2014-05-30 07:46:48 ncopa: what do you meam with stuff on usb? 2014-05-30 07:57:53 clandmeter: http://bugs.alpinelinux.org/issues/1481 2014-05-30 07:58:55 i think the only thing that might try load modules is usb stuff 2014-05-30 07:59:07 so mdev starts, it tirggers usb hotplug 2014-05-30 07:59:15 and it actually finds things on usb 2014-05-30 07:59:20 and try load the drivers for those 2014-05-30 07:59:28 but since modloop is not yet loaded, it fails 2014-05-30 08:00:52 but mdev must start before modloop 2014-05-30 08:01:04 otherwise can we not know that there are antying in /dev 2014-05-30 08:01:25 and without anything in /dev, we cannot mount the modloop (i think...) 2014-05-30 08:02:57 will check after coffe 2014-05-30 08:11:40 seems localmount also mounts nfs filesystems 2014-05-30 09:27:38 how would i got about creating a HVM guest in alpine-xen? 2014-05-30 09:28:24 i mean, there is no console in HVM, so i need X and VNC in my alpine dom0 right? 2014-05-30 09:42:05 ncopa: kmod does not provide the default symlinks like modprobe. Is that a feature or a bug? 2014-05-30 09:48:13 dunno 2014-05-30 09:56:01 then ill just add it. 2014-05-30 12:25:22 I am reconfiguring the 3.0 builders to halt on errors now 2014-05-30 12:27:51 also edge builders 2014-05-30 12:28:06 ncopa: you fixed nfs? 2014-05-30 12:28:12 no :-( 2014-05-30 12:28:24 well, i think client should work 2014-05-30 12:28:31 so why is the issue closed? 2014-05-30 12:29:00 i closed all 'resolved' 2014-05-30 12:29:06 in a bulk 2014-05-30 12:29:12 what issuenumber was it? 2014-05-30 12:29:17 but, in any case 2014-05-30 12:29:20 http://bugs.alpinelinux.org/issues/2842#change-9454 2014-05-30 12:29:20 it does not segfault 2014-05-30 12:29:29 so the segfault issue is resolved 2014-05-30 12:29:52 i think we should create a new issue for 'hangs on mount' 2014-05-30 12:30:49 or does it still segfault? 2014-05-30 12:30:58 I think we can create a new issue 2014-05-30 12:31:05 will check... 2014-05-30 12:31:08 since there are multiple issues in nfs-utils 2014-05-30 12:31:37 another question: why is the version number no longer used on /boot/vmlinuz... and /boot/initramfs... on 3.0? 2014-05-30 12:31:56 because i want to be able to have /boot on vfat 2014-05-30 12:32:02 makes it more difficult to maintain if running on xen 2014-05-30 12:32:10 and symlinks does not work on vfat 2014-05-30 12:32:14 because if i copy the kernels to the host, i must choose the proper version 2014-05-30 12:32:42 so i can't use different kernels per default just by copying them... 2014-05-30 12:32:49 i see 2014-05-30 12:32:52 hm 2014-05-30 12:33:02 its to late to fix that for 3.0 anyways 2014-05-30 12:33:20 i should tag 3.0 today 2014-05-30 12:37:37 i c 2014-05-30 12:38:30 ok, here's dmesg after upgrading my edge-64 box and rebooting it: 2014-05-30 12:38:31 http://sprunge.us/BagX 2014-05-30 12:39:05 and nfs-client is still not woring 2014-05-30 12:39:06 +k 2014-05-30 12:39:13 [ 5.726515] rpcbind[1072]: segfault at 5a539070 ip 000070d75a529e34 sp 00007c5ff5dd3448 error 4 in ld-musl-x86_64.so.1[70d75a4de000+81000] 2014-05-30 12:39:21 i dont have time to fix it today 2014-05-30 12:39:29 and i really really want 3.0 out 2014-05-30 12:39:33 :-) 2014-05-30 12:39:35 understand 2014-05-30 12:39:47 i'm "shooting the engineer and just shipping the product" 2014-05-30 12:39:55 should i file a bug? 2014-05-30 12:40:01 yes please 2014-05-30 12:40:05 and set target 3.0.1 2014-05-30 12:40:11 ok 2014-05-30 12:40:15 thanks 2014-05-30 12:40:21 and thanks for your patience 2014-05-30 12:47:51 http://bugs.alpinelinux.org/issues/2972 2014-05-30 13:03:59 ncopa: why does localmount want to mount nfs shares when network is not up yet? 2014-05-30 13:04:07 or is there something ican do to prevent that? 2014-05-30 13:05:17 dunno 2014-05-30 13:05:24 looks like its not supposed to do it 2014-05-30 13:05:36 looks like there is a net_fs_list 2014-05-30 13:05:43 dunno where/what sets it 2014-05-30 13:06:13 it could also be some strange sed expr that no longer work that triggers it 2014-05-30 13:33:58 \o/ 2014-05-30 13:34:09 no more open issues with target 3.0 2014-05-30 13:34:53 maybe we should go over those issues that are older than 2 years 2014-05-30 13:34:58 and just close them 2014-05-30 13:35:18 "closed due to inactivity. please create new issue if this is still a problem" 2014-05-30 13:35:21 or so... 2014-05-30 13:44:47 ncopa: i have another issue (issue day today) 2014-05-30 13:45:12 the rpi has no rtc. but hwclock init does not support no rtc. 2014-05-30 13:45:29 i disabled hwclock boot initd, but it still runs after reboot. 2014-05-30 13:45:36 hm 2014-05-30 13:45:50 dunno 2014-05-30 13:46:08 if i remove it from boot level, should it still run? 2014-05-30 13:46:19 google rpi and gentoo/openrc 2014-05-30 13:46:28 it might be something else pulls it in as dependency 2014-05-30 13:46:32 believe me i did :D 2014-05-30 13:46:50 grep hwclock /etc/init.d/* 2014-05-30 13:46:55 already did 2014-05-30 13:47:08 i am making 3.0 release now 2014-05-30 13:47:22 please dont git push anything for now 2014-05-30 13:48:00 oh.. wait.. 2014-05-30 13:48:10 i need disable all package that does not build... :-/ 2014-05-30 14:14:38 fun 2014-05-30 14:14:44 here comes my broken packages list :) 2014-05-30 14:48:15 argh! 2014-05-30 14:48:24 i will not reach the 3.0 release today :-( 2014-05-30 14:48:26 :-(( 2014-05-30 15:01:55 ok 2014-05-30 15:02:01 there was no 3.0.0 release today 2014-05-30 15:02:10 but i did rc4 instead 2014-05-30 15:02:14 see you on monday 2014-05-30 15:02:34 oh 2014-05-30 15:02:39 edge is broken now... 2014-05-30 15:02:45 we need clean up testing 2014-05-30 15:03:01 don't worry, I'll fix it today or tommorow 2014-05-30 15:03:03 it fails when a testing package fails to build 2014-05-30 15:03:10 i think there are *many* 2014-05-30 15:03:32 we need to either: fix or move to unmaintained 2014-05-30 15:03:38 but i gotta go now 2014-05-30 18:05:11 bah, should have read the backscroll 2014-05-30 18:33:17 why the hell it even touch my commits 2014-05-30 18:33:27 I push it to master for a reason 2014-05-30 18:34:25 and what has happened to build logs? 2014-05-30 18:35:52 ok, screw it 2014-05-30 18:36:18 ping me if anyone knows anything 2014-05-30 19:19:43 running 3.0.0_rc3, postgresql refuses to start 2014-05-30 19:20:04 localhost:~# /etc/init.d/postgresql start 2014-05-30 19:20:04 * Caching service dependencies ... [ ok ] 2014-05-30 19:20:05 * Starting PostgreSQL ... 2014-05-30 19:20:05 pg_ctl: could not start server 2014-05-30 19:20:05 Examine the log output. 2014-05-30 19:20:06 * start-stop-daemon: failed to start `/usr/bin/pg_ctl' 2014-05-30 19:20:08 * Check the log for a possible explanation of the above error. 2014-05-30 19:20:10 * /var/lib/postgresql/9.3/data/postmaster.log [ !! ] 2014-05-30 19:20:13 * ERROR: postgresql failed to start 2014-05-30 19:20:21 and in /var/lib/postgresql/9.3/data/postmaster.log: 2014-05-30 19:20:46 invalid option -- - 2014-05-30 19:20:46 Try "postgres --help" for more information. 2014-05-30 19:21:17 that's using postgresql-9.3.4-r0 2014-05-30 19:27:11 the conf.d and init.d files look good to me, so I'm wondering if it's an rc change that broke it 2014-05-30 19:29:48 have you tried to start it manually? 2014-05-30 19:31:51 no 2014-05-30 19:33:19 it usually gives better insight what's wrong than relying on openrc 2014-05-30 19:33:51 ACTION is trying to figure out the proper command to do so ... 2014-05-30 19:38:02 localhost:/$ /usr/bin/pg_ctl start -s -w -t 10 -l /var/lib/postgresql/9.3/data/postmaster.log -D /var/lib/postgresql/9.3/data -o --data-directory=/var/lib/postgresql/9.3/data 2014-05-30 19:38:02 pg_ctl: could not start server 2014-05-30 19:38:02 Examine the log output. 2014-05-30 19:38:12 not too helpful :( 2014-05-30 19:40:09 the above command is what start-stop-daemon uses underneath (if I'm not mistaken) 2014-05-30 19:45:31 seems to be the exact same command (aside from path differences) as run on alpine-2.5 with postgresql-9.2, and it works there 2014-05-30 19:45:53 and the init.d files are identical and the conf.d files virtually the same 2014-05-30 19:46:11 ACTION is leaning towards saying postgresql on edge is broken 2014-05-30 19:51:03 ACTION created a new issue for postgresql: https://bugs.alpinelinux.org/issues/2984 2014-05-30 22:04:41 n8@all 2014-05-31 00:09:48 has anyone noticed 2014-05-31 00:10:26 that on alpine edge/musl when you use netstat -l or something it shows the port number by default now instead of the name like http for 80 etc 2014-05-31 00:21:29 also to anyone using nginx, have you had problems with nginx starting to listen on 0.0.0.0 and not the IP you tell it? 2014-05-31 04:34:49 hmm, perl-extutils-pkgconfig built according to buildlogs, but was not uploaded 2014-05-31 04:35:29 which breaks perl-glib, which in turn breaks a bunch of packages 2014-05-31 08:37:55 moved erlang and all deps to unmaintained until someone looks at fixing erlang on musl 2014-05-31 08:49:00 we should actually ditch all testing… 2014-05-31 08:49:48 it would be much easier to just keep it clean instead of trying to clean it 2014-05-31 09:04:42 barthalion: at least ditch ports in testing not touched in X months 2014-05-31 09:33:09 there should probably be a policy that packages are only to stay in testing for 30 days or something 2014-05-31 11:19:15 ncopa: moved lua-lmd5 to unmaintained even though build fix is trivial. did not seem maintained 2014-05-31 17:12:35 uggedal: I'll move everything older than 6 months to unmaintained 2014-05-31 23:28:10 anyone using roundcubemail package on musl edge? 2014-05-31 23:30:18 hey barthalion do you mind upgrading roundcubemail package if you have a chance?